diff --git a/.clang-format b/.clang-format new file mode 100644 index 000000000..45232b80e --- /dev/null +++ b/.clang-format @@ -0,0 +1,161 @@ +--- +Language: Cpp +AlignAfterOpenBracket: Align +AlignArrayOfStructures: Left +AlignConsecutiveAssignments: AcrossComments +AlignConsecutiveBitFields: AcrossComments +AlignConsecutiveDeclarations: AcrossComments +AlignConsecutiveMacros: AcrossComments +# AlignConsecutiveShortCaseStatements: AcrossComments +AlignEscapedNewlines: Left # LeftWithLastLine +AlignOperands: Align +AlignTrailingComments: + Kind: Always + OverEmptyLines: 1 +AllowAllArgumentsOnNextLine: true +AllowAllParametersOfDeclarationOnNextLine: false +# AllowBreakBeforeNoexceptSpecifier: OnlyWithParen +AllowShortBlocksOnASingleLine: Never +AllowShortCaseLabelsOnASingleLine: false +AllowShortFunctionsOnASingleLine: Inline +AllowShortIfStatementsOnASingleLine: Never +AllowShortLambdasOnASingleLine: Inline +AllowShortLoopsOnASingleLine: false +AlwaysBreakBeforeMultilineStrings: true +BinPackArguments: true +BinPackParameters: true # OnePerLine +BitFieldColonSpacing: Both +BreakBeforeBraces: Custom # Attach +BraceWrapping: + AfterCaseLabel: true + AfterClass: false + AfterControlStatement: false + AfterEnum: false + AfterFunction: false + AfterNamespace: false + AfterObjCDeclaration: false + AfterStruct: false + AfterUnion: false + AfterExternBlock: false + BeforeCatch: false + BeforeElse: false + BeforeLambdaBody: false + BeforeWhile: false + IndentBraces: false + SplitEmptyFunction: false + SplitEmptyRecord: false + SplitEmptyNamespace: false +# BreakAdjacentStringLiterals: true +BreakAfterAttributes: Never +BreakBeforeBinaryOperators: None +BreakBeforeInlineASMColon: OnlyMultiline +BreakBeforeTernaryOperators: false +# BreakBinaryOperations: Never +BreakConstructorInitializers: AfterColon +# BreakFunctionDefinitionParameters: false +BreakInheritanceList: AfterComma +BreakStringLiterals: true +# BreakTemplateDeclarations: Yes +ColumnLimit: 120 +CommentPragmas: '^ IWYU pragma:' +CompactNamespaces: false +ConstructorInitializerIndentWidth: 4 +ContinuationIndentWidth: 4 +Cpp11BracedListStyle: false +DerivePointerAlignment: false +DisableFormat: false +EmptyLineBeforeAccessModifier: Leave +EmptyLineAfterAccessModifier: Never +ExperimentalAutoDetectBinPacking: false +FixNamespaceComments: true +IncludeBlocks: Regroup +IncludeCategories: + - Regex: '^<.*\.h>' + Priority: 1 + SortPriority: 0 + - Regex: '^<.*' + Priority: 2 + SortPriority: 0 + - Regex: '.*' + Priority: 3 + SortPriority: 0 +IncludeIsMainRegex: '([-_](test|unittest))?$' +IncludeIsMainSourceRegex: '' +IndentAccessModifiers: false +IndentCaseBlocks: true +IndentCaseLabels: true +IndentExternBlock: NoIndent +IndentGotoLabels: false +IndentPPDirectives: AfterHash +IndentWidth: 4 +IndentWrappedFunctionNames: false +InsertBraces: true # NOTE: may lead to incorrect formatting +InsertNewlineAtEOF: true +JavaScriptQuotes: Leave +JavaScriptWrapImports: true +KeepEmptyLinesAtTheStartOfBlocks: false +LambdaBodyIndentation: Signature +LineEnding: LF +MacroBlockBegin: '' +MacroBlockEnd: '' +MaxEmptyLinesToKeep: 1 +NamespaceIndentation: None +ObjCBinPackProtocolList: Auto +ObjCBlockIndentWidth: 4 +ObjCSpaceAfterProperty: true +ObjCSpaceBeforeProtocolList: true +PPIndentWidth: -1 +PackConstructorInitializers: CurrentLine +PenaltyBreakAssignment: 2 +PenaltyBreakBeforeFirstCallParameter: 1 +PenaltyBreakComment: 300 +PenaltyBreakFirstLessLess: 120 +PenaltyBreakString: 1000 +PenaltyBreakTemplateDeclaration: 10 +PenaltyExcessCharacter: 1000000 +PenaltyReturnTypeOnItsOwnLine: 200 +PointerAlignment: Middle +QualifierAlignment: Left +#QualifierOrder: ['static', 'inline', 'friend', 'constexpr', 'const', 'volatile', 'type', 'restrict'] +RawStringFormats: + - Language: Cpp + Delimiters: + - cc + - CC + - cpp + - Cpp + - CPP + - 'c++' + - 'C++' + CanonicalDelimiter: '' +ReferenceAlignment: Middle +ReflowComments: false # IndentOnly +SeparateDefinitionBlocks: Always +SortIncludes: CaseInsensitive +SortUsingDeclarations: LexicographicNumeric +SpaceAfterCStyleCast: true +SpaceAfterLogicalNot: false +SpaceAfterTemplateKeyword: true +SpaceBeforeAssignmentOperators: true +SpaceBeforeCpp11BracedList: false +SpaceBeforeCtorInitializerColon: true +SpaceBeforeInheritanceColon: true +SpaceBeforeParens: ControlStatements +SpaceBeforeRangeBasedForLoopColon: true +SpaceInEmptyBlock: false +SpaceInEmptyParentheses: false +SpacesBeforeTrailingComments: 2 +SpacesInAngles: Never +SpacesInContainerLiterals: true +SpacesInLineCommentPrefix: + Minimum: 1 + Maximum: -1 +SpacesInParentheses: false +SpacesInSquareBrackets: false +SpaceBeforeSquareBrackets: false +Standard: c++17 +TabWidth: 4 +UseTab: Never +WhitespaceSensitiveMacros: ['STRINGIZE'] +... + diff --git a/.clang-tidy b/.clang-tidy index 952c0cca8..310c3d182 100644 --- a/.clang-tidy +++ b/.clang-tidy @@ -17,8 +17,10 @@ Checks: > -clang-analyzer-security.insecureAPI.DeprecatedOrUnsafeBufferHandling, performance-*, portability-*, + -portability-simd-intrinsics, misc-*, -misc-const-correctness, -misc-non-private-member-variables-in-classes, -misc-no-recursion, + -misc-use-anonymous-namespace, FormatStyle: none diff --git a/.devops/cpu.Dockerfile b/.devops/cpu.Dockerfile new file mode 100644 index 000000000..8d020f16c --- /dev/null +++ b/.devops/cpu.Dockerfile @@ -0,0 +1,81 @@ +ARG UBUNTU_VERSION=22.04 + +FROM ubuntu:$UBUNTU_VERSION AS build + +RUN apt-get update && \ + apt-get install -y build-essential git cmake libcurl4-openssl-dev + +WORKDIR /app + +COPY . . + +RUN cmake -S . -B build -DGGML_BACKEND_DL=ON -DGGML_NATIVE=OFF -DGGML_CPU_ALL_VARIANTS=ON -DLLAMA_CURL=ON -DCMAKE_BUILD_TYPE=Release && \ + cmake --build build -j $(nproc) + +RUN mkdir -p /app/lib && \ + find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +## Base image +FROM ubuntu:$UBUNTU_VERSION AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +COPY --from=build /app/lib/ /app + +### Full +FROM base AS full + +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3 \ + python3-pip \ + && pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] diff --git a/.devops/cuda.Dockerfile b/.devops/cuda.Dockerfile new file mode 100644 index 000000000..974dd78a8 --- /dev/null +++ b/.devops/cuda.Dockerfile @@ -0,0 +1,94 @@ +ARG UBUNTU_VERSION=22.04 +# This needs to generally match the container host's environment. +ARG CUDA_VERSION=12.6.0 +# Target the CUDA build image +ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION} + +ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} + +FROM ${BASE_CUDA_DEV_CONTAINER} AS build + +# CUDA architecture to build for (defaults to all supported archs) +ARG CUDA_DOCKER_ARCH=default + +RUN apt-get update && \ + apt-get install -y build-essential cmake python3 python3-pip git libcurl4-openssl-dev libgomp1 + +WORKDIR /app + +COPY . . + +RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \ + export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \ + fi && \ + cmake -B build -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ + cmake --build build --config Release -j$(nproc) + +RUN mkdir -p /app/lib && \ + find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +## Base image +FROM ${BASE_CUDA_RUN_CONTAINER} AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +COPY --from=build /app/lib/ /app + +### Full +FROM base AS full + +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3 \ + python3-pip \ + && pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] diff --git a/.devops/full-cuda.Dockerfile b/.devops/full-cuda.Dockerfile deleted file mode 100644 index d5acd35e2..000000000 --- a/.devops/full-cuda.Dockerfile +++ /dev/null @@ -1,33 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG CUDA_VERSION=12.6.0 -# Target the CUDA build image -ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_CUDA_DEV_CONTAINER} AS build - -# CUDA architecture to build for (defaults to all supported archs) -ARG CUDA_DOCKER_ARCH=default - -RUN apt-get update && \ - apt-get install -y build-essential cmake python3 python3-pip git libcurl4-openssl-dev libgomp1 - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -# Use the default CUDA archs if not specified -RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \ - export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \ - fi && \ - cmake -B build -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release -j$(nproc) && \ - cp build/bin/* . - -ENTRYPOINT ["/app/.devops/tools.sh"] diff --git a/.devops/full-musa.Dockerfile b/.devops/full-musa.Dockerfile deleted file mode 100644 index 34ba856d3..000000000 --- a/.devops/full-musa.Dockerfile +++ /dev/null @@ -1,26 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG MUSA_VERSION=rc3.1.0 -# Target the MUSA build image -ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-devel-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_MUSA_DEV_CONTAINER} AS build - -RUN apt-get update && \ - apt-get install -y build-essential cmake python3 python3-pip git libcurl4-openssl-dev libgomp1 - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -RUN cmake -B build -DGGML_MUSA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release -j$(nproc) && \ - cp build/bin/* . - -ENTRYPOINT ["/app/.devops/tools.sh"] diff --git a/.devops/full-rocm.Dockerfile b/.devops/full-rocm.Dockerfile deleted file mode 100644 index df496bcd2..000000000 --- a/.devops/full-rocm.Dockerfile +++ /dev/null @@ -1,50 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -# This needs to generally match the container host's environment. -ARG ROCM_VERSION=5.6 - -# Target the CUDA build image -ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete - -FROM ${BASE_ROCM_DEV_CONTAINER} AS build - -# Unless otherwise specified, we make a fat build. -# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 -# This is mostly tied to rocBLAS supported archs. -ARG ROCM_DOCKER_ARCH="\ - gfx803 \ - gfx900 \ - gfx906 \ - gfx908 \ - gfx90a \ - gfx1010 \ - gfx1030 \ - gfx1100 \ - gfx1101 \ - gfx1102" - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -# Set nvcc architecture -ENV AMDGPU_TARGETS=${ROCM_DOCKER_ARCH} -# Enable ROCm -ENV GGML_HIPBLAS=1 -ENV CC=/opt/rocm/llvm/bin/clang -ENV CXX=/opt/rocm/llvm/bin/clang++ - -# Enable cURL -ENV LLAMA_CURL=1 -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev - -RUN make -j$(nproc) - -ENTRYPOINT ["/app/.devops/tools.sh"] diff --git a/.devops/full.Dockerfile b/.devops/full.Dockerfile deleted file mode 100644 index 2a06f82b7..000000000 --- a/.devops/full.Dockerfile +++ /dev/null @@ -1,25 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -FROM ubuntu:$UBUNTU_VERSION AS build - -RUN apt-get update && \ - apt-get install -y build-essential python3 python3-pip git libcurl4-openssl-dev libgomp1 - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -ENV LLAMA_CURL=1 - - -RUN make -j$(nproc) - -ENV LC_ALL=C.utf8 - -ENTRYPOINT ["/app/.devops/tools.sh"] diff --git a/.devops/intel.Dockerfile b/.devops/intel.Dockerfile new file mode 100644 index 000000000..af783f5e9 --- /dev/null +++ b/.devops/intel.Dockerfile @@ -0,0 +1,91 @@ +ARG ONEAPI_VERSION=2025.0.0-0-devel-ubuntu22.04 + +## Build Image + +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build + +ARG GGML_SYCL_F16=OFF +RUN apt-get update && \ + apt-get install -y git libcurl4-openssl-dev + +WORKDIR /app + +COPY . . + +RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \ + echo "GGML_SYCL_F16 is set" \ + && export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \ + fi && \ + echo "Building with dynamic libs" && \ + cmake -B build -DGGML_NATIVE=OFF -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DLLAMA_CURL=ON ${OPT_SYCL_F16} && \ + cmake --build build --config Release -j$(nproc) + +RUN mkdir -p /app/lib && \ + find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +### Full +FROM base AS full + +COPY --from=build /app/lib/ /app +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3 \ + python3-pip \ + && pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/lib/ /app +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/lib/ /app +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] + diff --git a/.devops/llama-cli-cann.Dockerfile b/.devops/llama-cli-cann.Dockerfile index db5ba2f25..02dce501c 100644 --- a/.devops/llama-cli-cann.Dockerfile +++ b/.devops/llama-cli-cann.Dockerfile @@ -1,6 +1,6 @@ ARG ASCEND_VERSION=8.0.rc2.alpha003-910b-openeuler22.03-py3.8 -FROM cosdt/cann:$ASCEND_VERSION AS build +FROM ascendai/cann:$ASCEND_VERSION AS build WORKDIR /app @@ -22,11 +22,11 @@ ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/runtime/lib64/stub:$LD_LIBRARY_PATH RUN echo "Building with static libs" && \ source /usr/local/Ascend/ascend-toolkit/set_env.sh --force && \ - cmake -B build -DGGML_CANN=ON -DBUILD_SHARED_LIBS=OFF && \ + cmake -B build -DGGML_NATIVE=OFF -DGGML_CANN=ON -DBUILD_SHARED_LIBS=OFF && \ cmake --build build --config Release --target llama-cli # TODO: use image with NNRT -FROM cosdt/cann:$ASCEND_VERSION AS runtime +FROM ascendai/cann:$ASCEND_VERSION AS runtime COPY --from=build /app/build/bin/llama-cli /llama-cli ENV LC_ALL=C.utf8 diff --git a/.devops/llama-cli-cuda.Dockerfile b/.devops/llama-cli-cuda.Dockerfile deleted file mode 100644 index b75163b94..000000000 --- a/.devops/llama-cli-cuda.Dockerfile +++ /dev/null @@ -1,37 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG CUDA_VERSION=12.6.0 -# Target the CUDA build image -ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION} -# Target the CUDA runtime image -ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_CUDA_DEV_CONTAINER} AS build - -# CUDA architecture to build for (defaults to all supported archs) -ARG CUDA_DOCKER_ARCH=default - -RUN apt-get update && \ - apt-get install -y build-essential git cmake - -WORKDIR /app - -COPY . . - -# Use the default CUDA archs if not specified -RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \ - export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \ - fi && \ - cmake -B build -DGGML_CUDA=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release --target llama-cli -j$(nproc) - -FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime - -RUN apt-get update && \ - apt-get install -y libgomp1 - -COPY --from=build /app/build/ggml/src/libggml.so /libggml.so -COPY --from=build /app/build/src/libllama.so /libllama.so -COPY --from=build /app/build/bin/llama-cli /llama-cli - -ENTRYPOINT [ "/llama-cli" ] diff --git a/.devops/llama-cli-intel.Dockerfile b/.devops/llama-cli-intel.Dockerfile deleted file mode 100644 index 79dba06a7..000000000 --- a/.devops/llama-cli-intel.Dockerfile +++ /dev/null @@ -1,28 +0,0 @@ -ARG ONEAPI_VERSION=2024.1.1-devel-ubuntu22.04 - -FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build - -ARG GGML_SYCL_F16=OFF -RUN apt-get update && \ - apt-get install -y git - -WORKDIR /app - -COPY . . - -RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \ - echo "GGML_SYCL_F16 is set" && \ - export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \ - fi && \ - echo "Building with static libs" && \ - cmake -B build -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx \ - ${OPT_SYCL_F16} -DBUILD_SHARED_LIBS=OFF && \ - cmake --build build --config Release --target llama-cli - -FROM intel/oneapi-basekit:$ONEAPI_VERSION AS runtime - -COPY --from=build /app/build/bin/llama-cli /llama-cli - -ENV LC_ALL=C.utf8 - -ENTRYPOINT [ "/llama-cli" ] diff --git a/.devops/llama-cli-musa.Dockerfile b/.devops/llama-cli-musa.Dockerfile deleted file mode 100644 index b5696794f..000000000 --- a/.devops/llama-cli-musa.Dockerfile +++ /dev/null @@ -1,30 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG MUSA_VERSION=rc3.1.0 -# Target the MUSA build image -ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-devel-ubuntu${UBUNTU_VERSION} -# Target the MUSA runtime image -ARG BASE_MUSA_RUN_CONTAINER=mthreads/musa:${MUSA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_MUSA_DEV_CONTAINER} AS build - -RUN apt-get update && \ - apt-get install -y build-essential git cmake - -WORKDIR /app - -COPY . . - -RUN cmake -B build -DGGML_MUSA=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release --target llama-cli -j$(nproc) - -FROM ${BASE_MUSA_RUN_CONTAINER} AS runtime - -RUN apt-get update && \ - apt-get install -y libgomp1 - -COPY --from=build /app/build/ggml/src/libggml.so /libggml.so -COPY --from=build /app/build/src/libllama.so /libllama.so -COPY --from=build /app/build/bin/llama-cli /llama-cli - -ENTRYPOINT [ "/llama-cli" ] diff --git a/.devops/llama-cli-rocm.Dockerfile b/.devops/llama-cli-rocm.Dockerfile deleted file mode 100644 index e60c747bd..000000000 --- a/.devops/llama-cli-rocm.Dockerfile +++ /dev/null @@ -1,45 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -# This needs to generally match the container host's environment. -ARG ROCM_VERSION=5.6 - -# Target the CUDA build image -ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete - -FROM ${BASE_ROCM_DEV_CONTAINER} AS build - -# Unless otherwise specified, we make a fat build. -# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 -# This is mostly tied to rocBLAS supported archs. -ARG ROCM_DOCKER_ARCH="\ - gfx803 \ - gfx900 \ - gfx906 \ - gfx908 \ - gfx90a \ - gfx1010 \ - gfx1030 \ - gfx1100 \ - gfx1101 \ - gfx1102" - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -# Set nvcc architecture -ENV AMDGPU_TARGETS=${ROCM_DOCKER_ARCH} -# Enable ROCm -ENV GGML_HIPBLAS=1 -ENV CC=/opt/rocm/llvm/bin/clang -ENV CXX=/opt/rocm/llvm/bin/clang++ - -RUN make -j$(nproc) llama-cli - -ENTRYPOINT [ "/app/llama-cli" ] diff --git a/.devops/llama-cli-vulkan.Dockerfile b/.devops/llama-cli-vulkan.Dockerfile deleted file mode 100644 index 9b0dad8bf..000000000 --- a/.devops/llama-cli-vulkan.Dockerfile +++ /dev/null @@ -1,27 +0,0 @@ -ARG UBUNTU_VERSION=jammy - -FROM ubuntu:$UBUNTU_VERSION AS build - -# Install build tools -RUN apt update && apt install -y git build-essential cmake wget libgomp1 - -# Install Vulkan SDK -RUN wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | apt-key add - && \ - wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list && \ - apt update -y && \ - apt-get install -y vulkan-sdk - -# Build it -WORKDIR /app -COPY . . -RUN cmake -B build -DGGML_VULKAN=1 && \ - cmake --build build --config Release --target llama-cli - -# Clean up -WORKDIR / -RUN cp /app/build/bin/llama-cli /llama-cli && \ - rm -rf /app - -ENV LC_ALL=C.utf8 - -ENTRYPOINT [ "/llama-cli" ] diff --git a/.devops/llama-cli.Dockerfile b/.devops/llama-cli.Dockerfile deleted file mode 100644 index 7f741aa46..000000000 --- a/.devops/llama-cli.Dockerfile +++ /dev/null @@ -1,23 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -FROM ubuntu:$UBUNTU_VERSION AS build - -RUN apt-get update && \ - apt-get install -y build-essential git - -WORKDIR /app - -COPY . . - -RUN make -j$(nproc) llama-cli - -FROM ubuntu:$UBUNTU_VERSION AS runtime - -RUN apt-get update && \ - apt-get install -y libgomp1 - -COPY --from=build /app/llama-cli /llama-cli - -ENV LC_ALL=C.utf8 - -ENTRYPOINT [ "/llama-cli" ] diff --git a/.devops/llama-server-cuda.Dockerfile b/.devops/llama-server-cuda.Dockerfile deleted file mode 100644 index a40e24205..000000000 --- a/.devops/llama-server-cuda.Dockerfile +++ /dev/null @@ -1,42 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG CUDA_VERSION=12.6.0 -# Target the CUDA build image -ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION} -# Target the CUDA runtime image -ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_CUDA_DEV_CONTAINER} AS build - -# CUDA architecture to build for (defaults to all supported archs) -ARG CUDA_DOCKER_ARCH=default - -RUN apt-get update && \ - apt-get install -y build-essential git cmake libcurl4-openssl-dev - -WORKDIR /app - -COPY . . - -# Use the default CUDA archs if not specified -RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \ - export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \ - fi && \ - cmake -B build -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release --target llama-server -j$(nproc) - -FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime - -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev libgomp1 curl - -COPY --from=build /app/build/ggml/src/libggml.so /libggml.so -COPY --from=build /app/build/src/libllama.so /libllama.so -COPY --from=build /app/build/bin/llama-server /llama-server - -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/llama-server" ] diff --git a/.devops/llama-server-intel.Dockerfile b/.devops/llama-server-intel.Dockerfile deleted file mode 100644 index 9c355b664..000000000 --- a/.devops/llama-server-intel.Dockerfile +++ /dev/null @@ -1,34 +0,0 @@ -ARG ONEAPI_VERSION=2024.1.1-devel-ubuntu22.04 - -FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build - -ARG GGML_SYCL_F16=OFF -RUN apt-get update && \ - apt-get install -y git libcurl4-openssl-dev - -WORKDIR /app - -COPY . . - -RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \ - echo "GGML_SYCL_F16 is set" && \ - export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \ - fi && \ - echo "Building with dynamic libs" && \ - cmake -B build -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DLLAMA_CURL=ON ${OPT_SYCL_F16} && \ - cmake --build build --config Release --target llama-server - -FROM intel/oneapi-basekit:$ONEAPI_VERSION AS runtime - -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev curl - -COPY --from=build /app/build/bin/llama-server /llama-server - -ENV LC_ALL=C.utf8 -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/llama-server" ] diff --git a/.devops/llama-server-musa.Dockerfile b/.devops/llama-server-musa.Dockerfile deleted file mode 100644 index 193a6d77c..000000000 --- a/.devops/llama-server-musa.Dockerfile +++ /dev/null @@ -1,35 +0,0 @@ -ARG UBUNTU_VERSION=22.04 -# This needs to generally match the container host's environment. -ARG MUSA_VERSION=rc3.1.0 -# Target the MUSA build image -ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-devel-ubuntu${UBUNTU_VERSION} -# Target the MUSA runtime image -ARG BASE_MUSA_RUN_CONTAINER=mthreads/musa:${MUSA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} - -FROM ${BASE_MUSA_DEV_CONTAINER} AS build - -RUN apt-get update && \ - apt-get install -y build-essential git cmake libcurl4-openssl-dev - -WORKDIR /app - -COPY . . - -RUN cmake -B build -DGGML_MUSA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ - cmake --build build --config Release --target llama-server -j$(nproc) - -FROM ${BASE_MUSA_RUN_CONTAINER} AS runtime - -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev libgomp1 curl - -COPY --from=build /app/build/ggml/src/libggml.so /libggml.so -COPY --from=build /app/build/src/libllama.so /libllama.so -COPY --from=build /app/build/bin/llama-server /llama-server - -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/llama-server" ] diff --git a/.devops/llama-server-rocm.Dockerfile b/.devops/llama-server-rocm.Dockerfile deleted file mode 100644 index 8553af75b..000000000 --- a/.devops/llama-server-rocm.Dockerfile +++ /dev/null @@ -1,54 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -# This needs to generally match the container host's environment. -ARG ROCM_VERSION=5.6 - -# Target the CUDA build image -ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete - -FROM ${BASE_ROCM_DEV_CONTAINER} AS build - -# Unless otherwise specified, we make a fat build. -# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 -# This is mostly tied to rocBLAS supported archs. -ARG ROCM_DOCKER_ARCH="\ - gfx803 \ - gfx900 \ - gfx906 \ - gfx908 \ - gfx90a \ - gfx1010 \ - gfx1030 \ - gfx1100 \ - gfx1101 \ - gfx1102" - -COPY requirements.txt requirements.txt -COPY requirements requirements - -RUN pip install --upgrade pip setuptools wheel \ - && pip install -r requirements.txt - -WORKDIR /app - -COPY . . - -# Set nvcc architecture -ENV AMDGPU_TARGETS=${ROCM_DOCKER_ARCH} -# Enable ROCm -ENV GGML_HIPBLAS=1 -ENV CC=/opt/rocm/llvm/bin/clang -ENV CXX=/opt/rocm/llvm/bin/clang++ -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -# Enable cURL -ENV LLAMA_CURL=1 -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev curl - -RUN make -j$(nproc) llama-server - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/app/llama-server" ] diff --git a/.devops/llama-server-vulkan.Dockerfile b/.devops/llama-server-vulkan.Dockerfile deleted file mode 100644 index 93c5e0c26..000000000 --- a/.devops/llama-server-vulkan.Dockerfile +++ /dev/null @@ -1,31 +0,0 @@ -ARG UBUNTU_VERSION=jammy - -FROM ubuntu:$UBUNTU_VERSION AS build - -# Install build tools -RUN apt update && apt install -y git build-essential cmake wget - -# Install Vulkan SDK and cURL -RUN wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | apt-key add - && \ - wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list && \ - apt update -y && \ - apt-get install -y vulkan-sdk libcurl4-openssl-dev curl - -# Build it -WORKDIR /app -COPY . . -RUN cmake -B build -DGGML_VULKAN=1 -DLLAMA_CURL=1 && \ - cmake --build build --config Release --target llama-server - -# Clean up -WORKDIR / -RUN cp /app/build/bin/llama-server /llama-server && \ - rm -rf /app - -ENV LC_ALL=C.utf8 -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/llama-server" ] diff --git a/.devops/llama-server.Dockerfile b/.devops/llama-server.Dockerfile deleted file mode 100644 index 02accc85e..000000000 --- a/.devops/llama-server.Dockerfile +++ /dev/null @@ -1,29 +0,0 @@ -ARG UBUNTU_VERSION=22.04 - -FROM ubuntu:$UBUNTU_VERSION AS build - -RUN apt-get update && \ - apt-get install -y build-essential git libcurl4-openssl-dev - -WORKDIR /app - -COPY . . - -ENV LLAMA_CURL=1 - -RUN make -j$(nproc) llama-server - -FROM ubuntu:$UBUNTU_VERSION AS runtime - -RUN apt-get update && \ - apt-get install -y libcurl4-openssl-dev libgomp1 curl - -COPY --from=build /app/llama-server /llama-server - -ENV LC_ALL=C.utf8 -# Must be set to 0.0.0.0 so it can listen to requests from host machine -ENV LLAMA_ARG_HOST=0.0.0.0 - -HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] - -ENTRYPOINT [ "/llama-server" ] diff --git a/.devops/musa.Dockerfile b/.devops/musa.Dockerfile new file mode 100644 index 000000000..bfd7fc1c1 --- /dev/null +++ b/.devops/musa.Dockerfile @@ -0,0 +1,108 @@ +ARG UBUNTU_VERSION=22.04 +# This needs to generally match the container host's environment. +ARG MUSA_VERSION=rc3.1.0 +# Target the MUSA build image +ARG BASE_MUSA_DEV_CONTAINER=mthreads/musa:${MUSA_VERSION}-devel-ubuntu${UBUNTU_VERSION} + +ARG BASE_MUSA_RUN_CONTAINER=mthreads/musa:${MUSA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} + +FROM ${BASE_MUSA_DEV_CONTAINER} AS build + +# MUSA architecture to build for (defaults to all supported archs) +ARG MUSA_DOCKER_ARCH=default + +RUN apt-get update && \ + apt-get install -y \ + build-essential \ + cmake \ + python3 \ + python3-pip \ + git \ + libcurl4-openssl-dev \ + libgomp1 + +COPY requirements.txt requirements.txt +COPY requirements requirements + +RUN pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt + +WORKDIR /app + +COPY . . + +# Use the default MUSA archs if not specified +RUN if [ "${MUSA_DOCKER_ARCH}" != "default" ]; then \ + export CMAKE_ARGS="-DMUSA_ARCHITECTURES=${MUSA_DOCKER_ARCH}"; \ + fi && \ + cmake -B build -DGGML_NATIVE=OFF -DGGML_MUSA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \ + cmake --build build --config Release -j$(nproc) + +RUN mkdir -p /app/lib && \ + find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +## Base image +FROM ${BASE_MUSA_RUN_CONTAINER} AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +COPY --from=build /app/lib/ /app + +### Full +FROM base AS full + +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3 \ + python3-pip \ + && pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] diff --git a/.devops/nix/package.nix b/.devops/nix/package.nix index 5d7d7ea5a..043c4364b 100644 --- a/.devops/nix/package.nix +++ b/.devops/nix/package.nix @@ -31,6 +31,7 @@ # Increases the runtime closure size by ~700M useMpi ? false, useRocm ? config.rocmSupport, + rocmGpuTargets ? builtins.concatStringsSep ";" rocmPackages.clr.gpuTargets, enableCurl ? true, useVulkan ? false, llamaVersion ? "0.0.0", # Arbitrary version, substituted by the flake @@ -126,9 +127,9 @@ effectiveStdenv.mkDerivation (finalAttrs: { }; postPatch = '' - substituteInPlace ./ggml/src/ggml-metal.m \ + substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \ --replace '[bundle pathForResource:@"ggml-metal" ofType:@"metal"];' "@\"$out/bin/ggml-metal.metal\";" - substituteInPlace ./ggml/src/ggml-metal.m \ + substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \ --replace '[bundle pathForResource:@"default" ofType:@"metallib"];' "@\"$out/bin/default.metallib\";" ''; @@ -173,7 +174,7 @@ effectiveStdenv.mkDerivation (finalAttrs: { (cmakeBool "GGML_NATIVE" false) (cmakeBool "GGML_BLAS" useBlas) (cmakeBool "GGML_CUDA" useCuda) - (cmakeBool "GGML_HIPBLAS" useRocm) + (cmakeBool "GGML_HIP" useRocm) (cmakeBool "GGML_METAL" useMetalKit) (cmakeBool "GGML_VULKAN" useVulkan) (cmakeBool "GGML_STATIC" enableStatic) @@ -188,7 +189,7 @@ effectiveStdenv.mkDerivation (finalAttrs: { ] ++ optionals useRocm [ (cmakeFeature "CMAKE_HIP_COMPILER" "${rocmPackages.llvm.clang}/bin/clang") - (cmakeFeature "CMAKE_HIP_ARCHITECTURES" (builtins.concatStringsSep ";" rocmPackages.clr.gpuTargets)) + (cmakeFeature "CMAKE_HIP_ARCHITECTURES" rocmGpuTargets) ] ++ optionals useMetalKit [ (lib.cmakeFeature "CMAKE_C_FLAGS" "-D__ARM_FEATURE_DOTPROD=1") diff --git a/.devops/nix/python-scripts.nix b/.devops/nix/python-scripts.nix index 392e9ffe4..56ea18278 100644 --- a/.devops/nix/python-scripts.nix +++ b/.devops/nix/python-scripts.nix @@ -34,7 +34,7 @@ let # server tests openai - behave + pytest prometheus-client ]; in diff --git a/.devops/rocm.Dockerfile b/.devops/rocm.Dockerfile new file mode 100644 index 000000000..a8088ea00 --- /dev/null +++ b/.devops/rocm.Dockerfile @@ -0,0 +1,113 @@ +ARG UBUNTU_VERSION=24.04 + +# This needs to generally match the container host's environment. +ARG ROCM_VERSION=6.3 +ARG AMDGPU_VERSION=6.3 + +# Target the CUDA build image +ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete + +### Build image +FROM ${BASE_ROCM_DEV_CONTAINER} AS build + +# Unless otherwise specified, we make a fat build. +# List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 +# This is mostly tied to rocBLAS supported archs. +# gfx803, gfx900, gfx1032, gfx1101, gfx1102,not officialy supported +# gfx906 is deprecated +#check https://rocm.docs.amd.com/projects/install-on-linux/en/docs-6.2.4/reference/system-requirements.html + +#ARG ROCM_DOCKER_ARCH='gfx803,gfx900,gfx906,gfx908,gfx90a,gfx942,gfx1010,gfx1030,gfx1032,gfx1100,gfx1101,gfx1102' +ARG ROCM_DOCKER_ARCH=gfx1100 + +# Set nvcc architectured +ENV AMDGPU_TARGETS=${ROCM_DOCKER_ARCH} +# Enable ROCm +# ENV CC=/opt/rocm/llvm/bin/clang +# ENV CXX=/opt/rocm/llvm/bin/clang++ + +RUN apt-get update \ + && apt-get install -y \ + build-essential \ + cmake \ + git \ + libcurl4-openssl-dev \ + curl \ + libgomp1 + +WORKDIR /app + +COPY . . + +RUN HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ + cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=$ROCM_DOCKER_ARCH -DCMAKE_BUILD_TYPE=Release -DLLAMA_CURL=ON \ + && cmake --build build --config Release -j$(nproc) + +RUN mkdir -p /app/lib \ + && find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +## Base image +FROM ${BASE_ROCM_DEV_CONTAINER} AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +COPY --from=build /app/lib/ /app + +### Full +FROM base AS full + +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3-pip \ + python3 \ + python3-wheel\ + && pip install --break-system-packages --upgrade setuptools \ + && pip install --break-system-packages -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] diff --git a/.devops/tools.sh b/.devops/tools.sh index 24dcfd350..9a86e6ea0 100755 --- a/.devops/tools.sh +++ b/.devops/tools.sh @@ -8,11 +8,11 @@ arg1="$1" shift if [[ "$arg1" == '--convert' || "$arg1" == '-c' ]]; then - python3 ./convert_hf_to_gguf.py "$@" + exec python3 ./convert_hf_to_gguf.py "$@" elif [[ "$arg1" == '--quantize' || "$arg1" == '-q' ]]; then - ./llama-quantize "$@" + exec ./llama-quantize "$@" elif [[ "$arg1" == '--run' || "$arg1" == '-r' ]]; then - ./llama-cli "$@" + exec ./llama-cli "$@" elif [[ "$arg1" == '--all-in-one' || "$arg1" == '-a' ]]; then echo "Converting PTH to GGML..." for i in `ls $1/$2/ggml-model-f16.bin*`; do @@ -20,11 +20,11 @@ elif [[ "$arg1" == '--all-in-one' || "$arg1" == '-a' ]]; then echo "Skip model quantization, it already exists: ${i/f16/q4_0}" else echo "Converting PTH to GGML: $i into ${i/f16/q4_0}..." - ./llama-quantize "$i" "${i/f16/q4_0}" q4_0 + exec ./llama-quantize "$i" "${i/f16/q4_0}" q4_0 fi done elif [[ "$arg1" == '--server' || "$arg1" == '-s' ]]; then - ./llama-server "$@" + exec ./llama-server "$@" else echo "Unknown command: $arg1" echo "Available commands: " diff --git a/.devops/vulkan.Dockerfile b/.devops/vulkan.Dockerfile new file mode 100644 index 000000000..cfc2162e3 --- /dev/null +++ b/.devops/vulkan.Dockerfile @@ -0,0 +1,88 @@ +ARG UBUNTU_VERSION=jammy + +FROM ubuntu:$UBUNTU_VERSION AS build + +# Install build tools +RUN apt update && apt install -y git build-essential cmake wget + +# Install Vulkan SDK and cURL +RUN wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | apt-key add - && \ + wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list && \ + apt update -y && \ + apt-get install -y vulkan-sdk libcurl4-openssl-dev curl + +# Build it +WORKDIR /app + +COPY . . + +RUN cmake -B build -DGGML_NATIVE=OFF -DGGML_VULKAN=1 -DLLAMA_CURL=1 && \ + cmake --build build --config Release -j$(nproc) + +RUN mkdir -p /app/lib && \ + find build -name "*.so" -exec cp {} /app/lib \; + +RUN mkdir -p /app/full \ + && cp build/bin/* /app/full \ + && cp *.py /app/full \ + && cp -r gguf-py /app/full \ + && cp -r requirements /app/full \ + && cp requirements.txt /app/full \ + && cp .devops/tools.sh /app/full/tools.sh + +## Base image +FROM ubuntu:$UBUNTU_VERSION AS base + +RUN apt-get update \ + && apt-get install -y libgomp1 curl\ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +COPY --from=build /app/lib/ /app + +### Full +FROM base AS full + +COPY --from=build /app/full /app + +WORKDIR /app + +RUN apt-get update \ + && apt-get install -y \ + git \ + python3 \ + python3-pip \ + && pip install --upgrade pip setuptools wheel \ + && pip install -r requirements.txt \ + && apt autoremove -y \ + && apt clean -y \ + && rm -rf /tmp/* /var/tmp/* \ + && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \ + && find /var/cache -type f -delete + +ENTRYPOINT ["/app/tools.sh"] + +### Light, CLI only +FROM base AS light + +COPY --from=build /app/full/llama-cli /app + +WORKDIR /app + +ENTRYPOINT [ "/app/llama-cli" ] + +### Server, Server only +FROM base AS server + +ENV LLAMA_ARG_HOST=0.0.0.0 + +COPY --from=build /app/full/llama-server /app + +WORKDIR /app + +HEALTHCHECK CMD [ "curl", "-f", "http://localhost:8080/health" ] + +ENTRYPOINT [ "/app/llama-server" ] diff --git a/.github/ISSUE_TEMPLATE/01-bug-low.yml b/.github/ISSUE_TEMPLATE/01-bug-low.yml deleted file mode 100644 index 54785854f..000000000 --- a/.github/ISSUE_TEMPLATE/01-bug-low.yml +++ /dev/null @@ -1,50 +0,0 @@ -name: Low Severity Bugs -description: Used to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches) -title: "Bug: " -labels: ["bug-unconfirmed", "low severity"] -body: - - type: markdown - attributes: - value: | - Thanks for taking the time to fill out this bug report! - Please include information about your system, the steps to reproduce the bug, - and the version of llama.cpp that you are using. - If possible, please provide a minimal code example that reproduces the bug. - - type: textarea - id: what-happened - attributes: - label: What happened? - description: Also tell us, what did you expect to happen? - placeholder: Tell us what you see! - validations: - required: true - - type: textarea - id: version - attributes: - label: Name and Version - description: Which executable and which version of our software are you running? (use `--version` to get a version string) - placeholder: | - $./llama-cli --version - version: 2999 (42b4109e) - built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu - validations: - required: true - - type: dropdown - id: operating-system - attributes: - label: What operating system are you seeing the problem on? - multiple: true - options: - - Linux - - Mac - - Windows - - BSD - - Other? (Please let us know in description) - validations: - required: false - - type: textarea - id: logs - attributes: - label: Relevant log output - description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks. - render: shell diff --git a/.github/ISSUE_TEMPLATE/010-bug-compilation.yml b/.github/ISSUE_TEMPLATE/010-bug-compilation.yml new file mode 100644 index 000000000..b85bf5741 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/010-bug-compilation.yml @@ -0,0 +1,87 @@ +name: Bug (compilation) +description: Something goes wrong when trying to compile llama.cpp. +title: "Compile bug: " +labels: ["bug-unconfirmed", "compilation"] +body: + - type: markdown + attributes: + value: > + Thanks for taking the time to fill out this bug report! + This issue template is intended for bug reports where the compilation of llama.cpp fails. + Before opening an issue, please confirm that the compilation still fails with `-DGGML_CCACHE=OFF`. + If the compilation succeeds with ccache disabled you should be able to permanently fix the issue + by clearing `~/.cache/ccache` (on Linux). + - type: textarea + id: commit + attributes: + label: Git commit + description: Which commit are you trying to compile? + placeholder: | + $git rev-parse HEAD + 84a07a17b1b08cf2b9747c633a2372782848a27f + validations: + required: true + - type: dropdown + id: operating-system + attributes: + label: Operating systems + description: Which operating systems do you know to be affected? + multiple: true + options: + - Linux + - Mac + - Windows + - BSD + - Other? (Please let us know in description) + validations: + required: true + - type: dropdown + id: backends + attributes: + label: GGML backends + description: Which GGML backends do you know to be affected? + options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan] + multiple: true + validations: + required: true + - type: textarea + id: info + attributes: + label: Problem description & steps to reproduce + description: > + Please give us a summary of the problem and tell us how to reproduce it. + If you can narrow down the bug to specific compile flags, that information would be very much appreciated by us. + placeholder: > + I'm trying to compile llama.cpp with CUDA support on a fresh install of Ubuntu and get error XY. + Here are the exact commands that I used: ... + validations: + required: true + - type: textarea + id: first_bad_commit + attributes: + label: First Bad Commit + description: > + If the bug was not present on an earlier version: when did it start appearing? + If possible, please do a git bisect and identify the exact commit that introduced the bug. + validations: + required: false + - type: textarea + id: command + attributes: + label: Compile command + description: > + Please provide the exact command you used to compile llama.cpp. For example: `cmake -B ...`. + This will be automatically formatted into code, so no need for backticks. + render: shell + validations: + required: true + - type: textarea + id: logs + attributes: + label: Relevant log output + description: > + Please copy and paste any relevant log output, including any generated text. + This will be automatically formatted into code, so no need for backticks. + render: shell + validations: + required: true diff --git a/.github/ISSUE_TEMPLATE/011-bug-results.yml b/.github/ISSUE_TEMPLATE/011-bug-results.yml new file mode 100644 index 000000000..1ccef0793 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/011-bug-results.yml @@ -0,0 +1,101 @@ +name: Bug (model use) +description: Something goes wrong when using a model (in general, not specific to a single llama.cpp module). +title: "Eval bug: " +labels: ["bug-unconfirmed", "model evaluation"] +body: + - type: markdown + attributes: + value: > + Thanks for taking the time to fill out this bug report! + This issue template is intended for bug reports where the model evaluation results + (i.e. the generated text) are incorrect or llama.cpp crashes during model evaluation. + If you encountered the issue while using an external UI (e.g. ollama), + please reproduce your issue using one of the examples/binaries in this repository. + The `llama-cli` binary can be used for simple and reproducible model inference. + - type: textarea + id: version + attributes: + label: Name and Version + description: Which version of our software are you running? (use `--version` to get a version string) + placeholder: | + $./llama-cli --version + version: 2999 (42b4109e) + built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu + validations: + required: true + - type: dropdown + id: operating-system + attributes: + label: Operating systems + description: Which operating systems do you know to be affected? + multiple: true + options: + - Linux + - Mac + - Windows + - BSD + - Other? (Please let us know in description) + validations: + required: true + - type: dropdown + id: backends + attributes: + label: GGML backends + description: Which GGML backends do you know to be affected? + options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan] + multiple: true + validations: + required: true + - type: textarea + id: hardware + attributes: + label: Hardware + description: Which CPUs/GPUs are you using? + placeholder: > + e.g. Ryzen 5950X + 2x RTX 4090 + validations: + required: true + - type: textarea + id: model + attributes: + label: Models + description: > + Which model(s) at which quantization were you using when encountering the bug? + If you downloaded a GGUF file off of Huggingface, please provide a link. + placeholder: > + e.g. Meta LLaMA 3.1 Instruct 8b q4_K_M + validations: + required: false + - type: textarea + id: info + attributes: + label: Problem description & steps to reproduce + description: > + Please give us a summary of the problem and tell us how to reproduce it. + If you can narrow down the bug to specific hardware, compile flags, or command line arguments, + that information would be very much appreciated by us. + placeholder: > + e.g. when I run llama-cli with -ngl 99 I get garbled outputs. + When I use -ngl 0 it works correctly. + Here are the exact commands that I used: ... + validations: + required: true + - type: textarea + id: first_bad_commit + attributes: + label: First Bad Commit + description: > + If the bug was not present on an earlier version: when did it start appearing? + If possible, please do a git bisect and identify the exact commit that introduced the bug. + validations: + required: false + - type: textarea + id: logs + attributes: + label: Relevant log output + description: > + Please copy and paste any relevant log output, including the command that you entered and any generated text. + This will be automatically formatted into code, so no need for backticks. + render: shell + validations: + required: true diff --git a/.github/ISSUE_TEMPLATE/019-bug-misc.yml b/.github/ISSUE_TEMPLATE/019-bug-misc.yml new file mode 100644 index 000000000..1904e31fd --- /dev/null +++ b/.github/ISSUE_TEMPLATE/019-bug-misc.yml @@ -0,0 +1,91 @@ +name: Bug (misc.) +description: Something is not working the way it should (and it's not covered by any of the above cases). +title: "Misc. bug: " +labels: ["bug-unconfirmed"] +body: + - type: markdown + attributes: + value: > + Thanks for taking the time to fill out this bug report! + This issue template is intended for miscellaneous bugs that don't fit into any other category. + If you encountered the issue while using an external UI (e.g. ollama), + please reproduce your issue using one of the examples/binaries in this repository. + - type: textarea + id: version + attributes: + label: Name and Version + description: Which version of our software is affected? (You can use `--version` to get a version string.) + placeholder: | + $./llama-cli --version + version: 2999 (42b4109e) + built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu + validations: + required: true + - type: dropdown + id: operating-system + attributes: + label: Operating systems + description: Which operating systems do you know to be affected? + multiple: true + options: + - Linux + - Mac + - Windows + - BSD + - Other? (Please let us know in description) + validations: + required: false + - type: dropdown + id: module + attributes: + label: Which llama.cpp modules do you know to be affected? + multiple: true + options: + - Documentation/Github + - libllama (core library) + - llama-cli + - llama-server + - llama-bench + - llama-quantize + - Python/Bash scripts + - Test code + - Other (Please specify in the next section) + validations: + required: false + - type: textarea + id: command + attributes: + label: Command line + description: > + Please provide the exact commands you entered, if applicable. For example: `llama-server -m ... -c ...`, `llama-cli -m ...`, etc. + This will be automatically formatted into code, so no need for backticks. + render: shell + validations: + required: false + - type: textarea + id: info + attributes: + label: Problem description & steps to reproduce + description: > + Please give us a summary of the problem and tell us how to reproduce it (if applicable). + validations: + required: true + - type: textarea + id: first_bad_commit + attributes: + label: First Bad Commit + description: > + If the bug was not present on an earlier version and it's not trivial to track down: when did it start appearing? + If possible, please do a git bisect and identify the exact commit that introduced the bug. + validations: + required: false + - type: textarea + id: logs + attributes: + label: Relevant log output + description: > + If applicable, please copy and paste any relevant log output, including any generated text. + This will be automatically formatted into code, so no need for backticks. + render: shell + validations: + required: false diff --git a/.github/ISSUE_TEMPLATE/02-bug-medium.yml b/.github/ISSUE_TEMPLATE/02-bug-medium.yml deleted file mode 100644 index a6285c6f0..000000000 --- a/.github/ISSUE_TEMPLATE/02-bug-medium.yml +++ /dev/null @@ -1,50 +0,0 @@ -name: Medium Severity Bug -description: Used to report medium severity bugs in llama.cpp (e.g. Malfunctioning Features but generally still useable) -title: "Bug: " -labels: ["bug-unconfirmed", "medium severity"] -body: - - type: markdown - attributes: - value: | - Thanks for taking the time to fill out this bug report! - Please include information about your system, the steps to reproduce the bug, - and the version of llama.cpp that you are using. - If possible, please provide a minimal code example that reproduces the bug. - - type: textarea - id: what-happened - attributes: - label: What happened? - description: Also tell us, what did you expect to happen? - placeholder: Tell us what you see! - validations: - required: true - - type: textarea - id: version - attributes: - label: Name and Version - description: Which executable and which version of our software are you running? (use `--version` to get a version string) - placeholder: | - $./llama-cli --version - version: 2999 (42b4109e) - built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu - validations: - required: true - - type: dropdown - id: operating-system - attributes: - label: What operating system are you seeing the problem on? - multiple: true - options: - - Linux - - Mac - - Windows - - BSD - - Other? (Please let us know in description) - validations: - required: false - - type: textarea - id: logs - attributes: - label: Relevant log output - description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks. - render: shell diff --git a/.github/ISSUE_TEMPLATE/05-enhancement.yml b/.github/ISSUE_TEMPLATE/020-enhancement.yml similarity index 97% rename from .github/ISSUE_TEMPLATE/05-enhancement.yml rename to .github/ISSUE_TEMPLATE/020-enhancement.yml index 58fca7318..02dd4f575 100644 --- a/.github/ISSUE_TEMPLATE/05-enhancement.yml +++ b/.github/ISSUE_TEMPLATE/020-enhancement.yml @@ -1,5 +1,5 @@ name: Enhancement -description: Used to request enhancements for llama.cpp +description: Used to request enhancements for llama.cpp. title: "Feature Request: " labels: ["enhancement"] body: diff --git a/.github/ISSUE_TEMPLATE/03-bug-high.yml b/.github/ISSUE_TEMPLATE/03-bug-high.yml deleted file mode 100644 index ff816b937..000000000 --- a/.github/ISSUE_TEMPLATE/03-bug-high.yml +++ /dev/null @@ -1,50 +0,0 @@ -name: High Severity Bug -description: Used to report high severity bugs in llama.cpp (e.g. Malfunctioning features hindering important common workflow) -title: "Bug: " -labels: ["bug-unconfirmed", "high severity"] -body: - - type: markdown - attributes: - value: | - Thanks for taking the time to fill out this bug report! - Please include information about your system, the steps to reproduce the bug, - and the version of llama.cpp that you are using. - If possible, please provide a minimal code example that reproduces the bug. - - type: textarea - id: what-happened - attributes: - label: What happened? - description: Also tell us, what did you expect to happen? - placeholder: Tell us what you see! - validations: - required: true - - type: textarea - id: version - attributes: - label: Name and Version - description: Which executable and which version of our software are you running? (use `--version` to get a version string) - placeholder: | - $./llama-cli --version - version: 2999 (42b4109e) - built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu - validations: - required: true - - type: dropdown - id: operating-system - attributes: - label: What operating system are you seeing the problem on? - multiple: true - options: - - Linux - - Mac - - Windows - - BSD - - Other? (Please let us know in description) - validations: - required: false - - type: textarea - id: logs - attributes: - label: Relevant log output - description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks. - render: shell diff --git a/.github/ISSUE_TEMPLATE/06-research.yml b/.github/ISSUE_TEMPLATE/030-research.yml similarity index 97% rename from .github/ISSUE_TEMPLATE/06-research.yml rename to .github/ISSUE_TEMPLATE/030-research.yml index 3ae4e9f8c..18975dbbf 100644 --- a/.github/ISSUE_TEMPLATE/06-research.yml +++ b/.github/ISSUE_TEMPLATE/030-research.yml @@ -1,5 +1,5 @@ name: Research -description: Track new technical research area +description: Track new technical research area. title: "Research: " labels: ["research 🔬"] body: diff --git a/.github/ISSUE_TEMPLATE/04-bug-critical.yml b/.github/ISSUE_TEMPLATE/04-bug-critical.yml deleted file mode 100644 index 7af42a80b..000000000 --- a/.github/ISSUE_TEMPLATE/04-bug-critical.yml +++ /dev/null @@ -1,50 +0,0 @@ -name: Critical Severity Bug -description: Used to report critical severity bugs in llama.cpp (e.g. Crashing, Corrupted, Dataloss) -title: "Bug: " -labels: ["bug-unconfirmed", "critical severity"] -body: - - type: markdown - attributes: - value: | - Thanks for taking the time to fill out this bug report! - Please include information about your system, the steps to reproduce the bug, - and the version of llama.cpp that you are using. - If possible, please provide a minimal code example that reproduces the bug. - - type: textarea - id: what-happened - attributes: - label: What happened? - description: Also tell us, what did you expect to happen? - placeholder: Tell us what you see! - validations: - required: true - - type: textarea - id: version - attributes: - label: Name and Version - description: Which executable and which version of our software are you running? (use `--version` to get a version string) - placeholder: | - $./llama-cli --version - version: 2999 (42b4109e) - built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu - validations: - required: true - - type: dropdown - id: operating-system - attributes: - label: What operating system are you seeing the problem on? - multiple: true - options: - - Linux - - Mac - - Windows - - BSD - - Other? (Please let us know in description) - validations: - required: false - - type: textarea - id: logs - attributes: - label: Relevant log output - description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks. - render: shell diff --git a/.github/ISSUE_TEMPLATE/07-refactor.yml b/.github/ISSUE_TEMPLATE/040-refactor.yml similarity index 95% rename from .github/ISSUE_TEMPLATE/07-refactor.yml rename to .github/ISSUE_TEMPLATE/040-refactor.yml index 3a68d3d53..b6e6ab36d 100644 --- a/.github/ISSUE_TEMPLATE/07-refactor.yml +++ b/.github/ISSUE_TEMPLATE/040-refactor.yml @@ -1,5 +1,5 @@ name: Refactor (Maintainers) -description: Used to track refactoring opportunities +description: Used to track refactoring opportunities. title: "Refactor: " labels: ["refactor"] body: diff --git a/.github/labeler.yml b/.github/labeler.yml index 89436740d..1b47bc968 100644 --- a/.github/labeler.yml +++ b/.github/labeler.yml @@ -3,19 +3,18 @@ Kompute: - changed-files: - any-glob-to-any-file: - ggml/include/ggml-kompute.h - - ggml/src/ggml-kompute.cpp + - ggml/src/ggml-kompute/** - README-kompute.md Apple Metal: - changed-files: - any-glob-to-any-file: - ggml/include/ggml-metal.h - - ggml/src/ggml-metal.cpp + - ggml/src/ggml-metal/** - README-metal.md SYCL: - changed-files: - any-glob-to-any-file: - ggml/include/ggml-sycl.h - - ggml/src/ggml-sycl.cpp - ggml/src/ggml-sycl/** - docs/backend/SYCL.md - examples/sycl/** @@ -27,8 +26,8 @@ Nvidia GPU: Vulkan: - changed-files: - any-glob-to-any-file: - - ggml/ggml_vk_generate_shaders.py - - ggml/src/ggml-vulkan* + - ggml/include/ggml-vulkan.h + - ggml/src/ggml-vulkan/** documentation: - changed-files: - any-glob-to-any-file: @@ -75,11 +74,7 @@ server: ggml: - changed-files: - any-glob-to-any-file: - - ggml/include/ggml*.h - - ggml/src/ggml*.c - - ggml/src/ggml*.cpp - - ggml/src/ggml*.h - - ggml-cuda/** + - ggml/** nix: - changed-files: - any-glob-to-any-file: diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md index 997c6d9d0..d9f5bdc23 100644 --- a/.github/pull_request_template.md +++ b/.github/pull_request_template.md @@ -1,7 +1 @@ - - -- [x] I have read the [contributing guidelines](https://github.com/ggerganov/llama.cpp/blob/master/CONTRIBUTING.md) -- Self-reported review complexity: - - [ ] Low - - [ ] Medium - - [ ] High +*Make sure to read the [contributing guidelines](https://github.com/ggerganov/llama.cpp/blob/master/CONTRIBUTING.md) before submitting a PR* diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index 16a52a3cc..e59cf9ab4 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -55,7 +55,12 @@ jobs: sysctl -a mkdir build cd build - cmake -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL_EMBED_LIBRARY=ON -DLLAMA_CURL=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF .. + cmake .. \ + -DLLAMA_FATAL_WARNINGS=ON \ + -DLLAMA_CURL=ON \ + -DGGML_METAL_USE_BF16=ON \ + -DGGML_METAL_EMBED_LIBRARY=ON \ + -DGGML_RPC=ON cmake --build . --config Release -j $(sysctl -n hw.logicalcpu) - name: Test @@ -113,7 +118,11 @@ jobs: sysctl -a # Metal is disabled due to intermittent failures with Github runners not having a GPU: # https://github.com/ggerganov/llama.cpp/actions/runs/8635935781/job/23674807267#step:5:2313 - cmake -B build -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL=OFF -DLLAMA_CURL=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF + cmake -B build \ + -DLLAMA_FATAL_WARNINGS=ON \ + -DLLAMA_CURL=ON \ + -DGGML_METAL=OFF \ + -DGGML_RPC=ON cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) - name: Test @@ -149,66 +158,6 @@ jobs: path: llama-${{ steps.tag.outputs.name }}-bin-macos-x64.zip name: llama-bin-macos-x64.zip - ubuntu-focal-make: - runs-on: ubuntu-20.04 - env: - LLAMA_NODE_AVAILABLE: true - LLAMA_PYTHON_AVAILABLE: true - - steps: - - name: Clone - id: checkout - uses: actions/checkout@v4 - - - name: Dependencies - id: depends - run: | - sudo apt-get update - sudo apt-get install build-essential gcc-8 - - - uses: actions/setup-node@v4 - with: - node-version: "20" - - - uses: actions/setup-python@v5 - with: - python-version: "3.11" - - - name: Build - id: make_build - env: - LLAMA_FATAL_WARNINGS: 1 - run: | - CC=gcc-8 make -j $(nproc) - - - name: Test - id: make_test - run: | - CC=gcc-8 make tests -j $(nproc) - make test -j $(nproc) - - ubuntu-focal-make-curl: - runs-on: ubuntu-20.04 - - steps: - - name: Clone - id: checkout - uses: actions/checkout@v4 - - - name: Dependencies - id: depends - run: | - sudo apt-get update - sudo apt-get install build-essential gcc-8 libcurl4-openssl-dev - - - name: Build - id: make_build - env: - LLAMA_FATAL_WARNINGS: 1 - LLAMA_CURL: 1 - run: | - CC=gcc-8 make -j $(nproc) - ubuntu-latest-cmake: runs-on: ubuntu-latest @@ -230,7 +179,7 @@ jobs: run: | mkdir build cd build - cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF + cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON -DGGML_RPC=ON cmake --build . --config Release -j $(nproc) - name: Test @@ -366,7 +315,7 @@ jobs: wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | sudo apt-key add - sudo wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list sudo apt-get update -y - sudo apt-get install -y build-essential vulkan-sdk + sudo apt-get install -y build-essential mesa-vulkan-drivers vulkan-sdk - name: Build id: cmake_build @@ -376,6 +325,12 @@ jobs: cmake -DGGML_VULKAN=ON .. cmake --build . --config Release -j $(nproc) + - name: Test + id: cmake_test + run: | + cd build + ctest -L main --verbose --timeout 900 + ubuntu-22-cmake-hip: runs-on: ubuntu-22.04 container: rocm/dev-ubuntu-22.04:6.0.2 @@ -394,15 +349,36 @@ jobs: - name: Build with native CMake HIP support id: cmake_build run: | - cmake -B build -S . -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIPBLAS=ON + cmake -B build -S . -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIP=ON cmake --build build --config Release -j $(nproc) - name: Build with legacy HIP support id: cmake_build_legacy_hip run: | - cmake -B build2 -S . -DCMAKE_C_COMPILER=hipcc -DCMAKE_CXX_COMPILER=hipcc -DGGML_HIPBLAS=ON + cmake -B build2 -S . -DCMAKE_C_COMPILER=hipcc -DCMAKE_CXX_COMPILER=hipcc -DGGML_HIP=ON cmake --build build2 --config Release -j $(nproc) + ubuntu-22-cmake-musa: + runs-on: ubuntu-22.04 + container: mthreads/musa:rc3.1.0-devel-ubuntu22.04 + + steps: + - name: Clone + id: checkout + uses: actions/checkout@v4 + + - name: Dependencies + id: depends + run: | + apt-get update + apt-get install -y build-essential git cmake libcurl4-openssl-dev + + - name: Build with native CMake MUSA support + id: cmake_build + run: | + cmake -B build -S . -DGGML_MUSA=ON + cmake --build build --config Release -j $(nproc) + ubuntu-22-cmake-sycl: runs-on: ubuntu-22.04 @@ -485,36 +461,6 @@ jobs: cmake -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON .. cmake --build . --config Release -j $(nproc) - # TODO: build with GGML_NO_METAL because test-backend-ops fail on "Apple Paravirtual device" and I don't know - # how to debug it. - # ref: https://github.com/ggerganov/llama.cpp/actions/runs/7131777249/job/19420981052#step:5:1124 - macOS-latest-make: - runs-on: macos-latest - - steps: - - name: Clone - id: checkout - uses: actions/checkout@v4 - - - name: Dependencies - id: depends - continue-on-error: true - run: | - brew update - - - name: Build - id: make_build - env: - LLAMA_FATAL_WARNINGS: 1 - run: | - GGML_NO_METAL=1 make -j $(sysctl -n hw.logicalcpu) - - - name: Test - id: make_test - run: | - GGML_NO_METAL=1 make tests -j $(sysctl -n hw.logicalcpu) - GGML_NO_METAL=1 make test -j $(sysctl -n hw.logicalcpu) - # TODO: build with GGML_METAL=OFF because test-backend-ops fail on "Apple Paravirtual device" and I don't know # how to debug it. # ref: https://github.com/ggerganov/llama.cpp/actions/runs/7132125951/job/19422043567?pr=4359#step:5:6584 @@ -569,6 +515,7 @@ jobs: mkdir build cd build cmake -G Xcode .. \ + -DGGML_METAL_USE_BF16=ON \ -DGGML_METAL_EMBED_LIBRARY=ON \ -DLLAMA_BUILD_EXAMPLES=OFF \ -DLLAMA_BUILD_TESTS=OFF \ @@ -599,6 +546,7 @@ jobs: mkdir build cd build cmake -G Xcode .. \ + -DGGML_METAL_USE_BF16=ON \ -DGGML_METAL_EMBED_LIBRARY=ON \ -DLLAMA_BUILD_EXAMPLES=OFF \ -DLLAMA_BUILD_TESTS=OFF \ @@ -626,15 +574,26 @@ jobs: run: | brew update + - name: Build llama.cpp with CMake + id: cmake_build + run: | + sysctl -a + mkdir build + cd build + cmake -G Xcode .. \ + -DGGML_METAL_USE_BF16=ON \ + -DGGML_METAL_EMBED_LIBRARY=ON \ + -DLLAMA_BUILD_EXAMPLES=OFF \ + -DLLAMA_BUILD_TESTS=OFF \ + -DLLAMA_BUILD_SERVER=OFF \ + -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64" + cmake --build . --config Release -j $(sysctl -n hw.logicalcpu) + sudo cmake --install . --config Release + - name: xcodebuild for swift package id: xcodebuild run: | - xcodebuild -scheme llama -destination "${{ matrix.destination }}" - - - name: Build Swift Example - id: make_build_swift_example - run: | - make swift + xcodebuild -scheme llama-Package -destination "${{ matrix.destination }}" windows-msys2: runs-on: windows-latest @@ -661,21 +620,6 @@ jobs: mingw-w64-${{matrix.env}}-cmake mingw-w64-${{matrix.env}}-openblas - - name: Build using make - shell: msys2 {0} - run: | - make -j $(nproc) - - - name: Clean after building using make - shell: msys2 {0} - run: | - make clean - - - name: Build using make w/ OpenBLAS - shell: msys2 {0} - run: | - make GGML_OPENBLAS=1 -j $(nproc) - - name: Build using CMake shell: msys2 {0} run: | @@ -694,7 +638,7 @@ jobs: cmake --build build --config ${{ matrix.build }} -j $(nproc) windows-latest-cmake: - runs-on: windows-2019 + runs-on: windows-latest env: OPENBLAS_VERSION: 0.3.23 @@ -705,23 +649,25 @@ jobs: matrix: include: - build: 'noavx-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF' - build: 'avx2-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON' - build: 'avx-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX2=OFF -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX2=OFF' - build: 'avx512-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX512=ON -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_AVX512=ON' - build: 'openblas-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_BLAS=ON -DBUILD_SHARED_LIBS=ON -DGGML_BLAS_VENDOR=OpenBLAS -DBLAS_INCLUDE_DIRS="$env:RUNNER_TEMP/openblas/include" -DBLAS_LIBRARIES="$env:RUNNER_TEMP/openblas/lib/openblas.lib"' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS -DBLAS_INCLUDE_DIRS="$env:RUNNER_TEMP/openblas/include" -DBLAS_LIBRARIES="$env:RUNNER_TEMP/openblas/lib/openblas.lib"' - build: 'kompute-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_KOMPUTE=ON -DKOMPUTE_OPT_DISABLE_VULKAN_VERSION_CHECK=ON -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_KOMPUTE=ON -DKOMPUTE_OPT_DISABLE_VULKAN_VERSION_CHECK=ON' - build: 'vulkan-x64' - defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_VULKAN=ON -DBUILD_SHARED_LIBS=ON' + defines: '-DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_VULKAN=ON' - build: 'llvm-arm64' - defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON' + defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON' - build: 'msvc-arm64' - defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-msvc.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DBUILD_SHARED_LIBS=ON' + defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-msvc.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON' + - build: 'llvm-arm64-opencl-adreno' + defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/opencl-arm64-release" -DGGML_OPENCL=ON -DGGML_OPENCL_USE_ADRENO_KERNELS=ON' steps: - name: Clone @@ -734,7 +680,7 @@ jobs: id: clone_kompute if: ${{ matrix.build == 'kompute-x64' }} run: | - git submodule update --init ggml/src/kompute + git submodule update --init ggml/src/ggml-kompute/kompute - name: Download OpenBLAS id: get_openblas @@ -763,6 +709,28 @@ jobs: run: | choco install ninja + - name: Install OpenCL Headers and Libs + id: install_opencl + if: ${{ matrix.build == 'llvm-arm64-opencl-adreno' }} + run: | + git clone https://github.com/KhronosGroup/OpenCL-Headers + cd OpenCL-Headers + mkdir build && cd build + cmake .. ` + -DBUILD_TESTING=OFF ` + -DOPENCL_HEADERS_BUILD_TESTING=OFF ` + -DOPENCL_HEADERS_BUILD_CXX_TESTS=OFF ` + -DCMAKE_INSTALL_PREFIX="$env:RUNNER_TEMP/opencl-arm64-release" + cmake --build . --target install + git clone https://github.com/KhronosGroup/OpenCL-ICD-Loader + cd OpenCL-ICD-Loader + mkdir build-arm64-release && cd build-arm64-release + cmake .. ` + -A arm64 ` + -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/opencl-arm64-release" ` + -DCMAKE_INSTALL_PREFIX="$env:RUNNER_TEMP/opencl-arm64-release" + cmake --build . --target install --config release + - name: Build id: cmake_build run: | @@ -792,7 +760,7 @@ jobs: - name: Test id: cmake_test # not all machines have native AVX-512 - if: ${{ matrix.build != 'msvc-arm64' && matrix.build != 'llvm-arm64' && matrix.build != 'kompute-x64' && matrix.build != 'vulkan-x64' && (matrix.build != 'avx512-x64' || env.HAS_AVX512F == '1') }} + if: ${{ matrix.build != 'msvc-arm64' && matrix.build != 'llvm-arm64' && matrix.build != 'llvm-arm64-opencl-adreno' && matrix.build != 'kompute-x64' && matrix.build != 'vulkan-x64' && (matrix.build != 'avx512-x64' || env.HAS_AVX512F == '1') }} run: | cd build ctest -L main -C Release --verbose --timeout 900 @@ -837,12 +805,33 @@ jobs: path: llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}.zip name: llama-bin-win-${{ matrix.build }}.zip - windows-latest-cmake-cuda: + ubuntu-latest-cmake-cuda: + runs-on: ubuntu-latest + container: nvidia/cuda:12.6.2-devel-ubuntu24.04 + + steps: + - name: Clone + id: checkout + uses: actions/checkout@v4 + + - name: Install dependencies + env: + DEBIAN_FRONTEND: noninteractive + run: | + apt update + apt install -y cmake build-essential ninja-build libgomp1 git + + - name: Build with CMake + run: | + cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=89-real -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined -DLLAMA_FATAL_WARNINGS=ON + cmake --build build + + windows-2019-cmake-cuda: runs-on: windows-2019 strategy: matrix: - cuda: ['12.2.0', '11.7.1'] + cuda: ['12.4', '11.7'] build: ['cuda'] steps: @@ -850,24 +839,83 @@ jobs: id: checkout uses: actions/checkout@v4 with: - fetch-depth: 0 + fetch-depth: 0 - - name: Install CUDA toolkit - id: cuda-toolkit - uses: Jimver/cuda-toolkit@v0.2.15 + - name: Install Cuda Toolkit 11.7 + if: ${{ matrix.cuda == '11.7' }} + run: | + mkdir -p "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" + choco install unzip -y + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cudart/windows-x86_64/cuda_cudart-windows-x86_64-11.7.99-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvcc/windows-x86_64/cuda_nvcc-windows-x86_64-11.7.99-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvrtc/windows-x86_64/cuda_nvrtc-windows-x86_64-11.7.99-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/libcublas/windows-x86_64/libcublas-windows-x86_64-11.7.4.6-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvtx/windows-x86_64/cuda_nvtx-windows-x86_64-11.7.91-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/visual_studio_integration/windows-x86_64/visual_studio_integration-windows-x86_64-11.7.91-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvprof/windows-x86_64/cuda_nvprof-windows-x86_64-11.7.101-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cccl/windows-x86_64/cuda_cccl-windows-x86_64-11.7.91-archive.zip" + unzip '*.zip' -d "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_cudart-windows-x86_64-11.7.99-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_nvcc-windows-x86_64-11.7.99-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_nvrtc-windows-x86_64-11.7.99-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\libcublas-windows-x86_64-11.7.4.6-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_nvtx-windows-x86_64-11.7.91-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\visual_studio_integration-windows-x86_64-11.7.91-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_nvprof-windows-x86_64-11.7.101-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\cuda_cccl-windows-x86_64-11.7.91-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" /E /I /H /Y + echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append + echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\libnvvp" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append + echo "CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8 + echo "CUDA_PATH_V11_7=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8 + + - name: Install Cuda Toolkit 12.4 + if: ${{ matrix.cuda == '12.4' }} + run: | + mkdir -p "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" + choco install unzip -y + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cudart/windows-x86_64/cuda_cudart-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvcc/windows-x86_64/cuda_nvcc-windows-x86_64-12.4.131-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvrtc/windows-x86_64/cuda_nvrtc-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/libcublas/windows-x86_64/libcublas-windows-x86_64-12.4.5.8-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvtx/windows-x86_64/cuda_nvtx-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_profiler_api/windows-x86_64/cuda_profiler_api-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/visual_studio_integration/windows-x86_64/visual_studio_integration-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvprof/windows-x86_64/cuda_nvprof-windows-x86_64-12.4.127-archive.zip" + curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cccl/windows-x86_64/cuda_cccl-windows-x86_64-12.4.127-archive.zip" + unzip '*.zip' -d "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_cudart-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvcc-windows-x86_64-12.4.131-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvrtc-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\libcublas-windows-x86_64-12.4.5.8-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvtx-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_profiler_api-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\visual_studio_integration-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvprof-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_cccl-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y + echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append + echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\libnvvp" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append + echo "CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8 + echo "CUDA_PATH_V12_4=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8 + + - name: Install ccache + uses: hendrikmuhs/ccache-action@v1.2 with: - cuda: ${{ matrix.cuda }} - method: 'network' - sub-packages: '["nvcc", "cudart", "cublas", "cublas_dev", "thrust", "visual_studio_integration"]' + key: ${{ github.job }}-${{ matrix.cuda }}-${{ matrix.build }} + + - name: Install Ninja + id: install_ninja + run: | + choco install ninja - name: Build id: cmake_build + shell: cmd run: | - mkdir build - cd build - cmake .. -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=ON -DGGML_RPC=ON - cmake --build . --config Release -j $((${env:NUMBER_OF_PROCESSORS} - 1)) -t ggml - cmake --build . --config Release -j ${env:NUMBER_OF_PROCESSORS} + call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Auxiliary\Build\vcvars64.bat" + cmake -S . -B build -G "Ninja Multi-Config" -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DGGML_RPC=ON + set /A NINJA_JOBS=%NUMBER_OF_PROCESSORS%-1 + cmake --build build --config Release -j %NINJA_JOBS% -t ggml + cmake --build build --config Release - name: Determine tag name id: tag @@ -896,10 +944,12 @@ jobs: name: llama-bin-win-cu${{ matrix.cuda }}-x64.zip - name: Copy and pack Cuda runtime + if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }} run: | - echo "Cuda install location: ${{steps.cuda-toolkit.outputs.CUDA_PATH}}" + echo "Cuda install location: ${{ env.CUDA_PATH }}" $dst='.\build\bin\cudart\' - robocopy "${{steps.cuda-toolkit.outputs.CUDA_PATH}}\bin" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll + robocopy "${{env.CUDA_PATH}}\bin" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll + robocopy "${{env.CUDA_PATH}}\lib" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll 7z a cudart-llama-bin-win-cu${{ matrix.cuda }}-x64.zip $dst\* - name: Upload Cuda runtime @@ -917,8 +967,8 @@ jobs: shell: bash env: - WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/7dff44ba-e3af-4448-841c-0d616c8da6e7/w_BaseKit_p_2024.1.0.595_offline.exe - WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel + WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/b380d914-366b-4b77-a74a-05e3c38b3514/intel-oneapi-base-toolkit-2025.0.0.882_offline.exe + WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel:intel.oneapi.win.dnnl:intel.oneapi.win.tbb.devel ONEAPI_ROOT: "C:/Program Files (x86)/Intel/oneAPI" steps: - name: Clone @@ -928,7 +978,8 @@ jobs: fetch-depth: 0 - name: Install - run: scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL + run: | + scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL - name: Build id: cmake_build @@ -947,25 +998,33 @@ jobs: echo "name=${SAFE_NAME}-b${BUILD_NUMBER}-${SHORT_HASH}" >> $GITHUB_OUTPUT fi - - name: Pack artifacts + - name: Build the release package id: pack_artifacts if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }} run: | echo "cp oneAPI running time dll files in ${{ env.ONEAPI_ROOT }} to ./build/bin" - cp "${{ env.ONEAPI_ROOT }}/mkl/latest/bin/mkl_sycl_blas.4.dll" ./build/bin + + cp "${{ env.ONEAPI_ROOT }}/mkl/latest/bin/mkl_sycl_blas.5.dll" ./build/bin cp "${{ env.ONEAPI_ROOT }}/mkl/latest/bin/mkl_core.2.dll" ./build/bin cp "${{ env.ONEAPI_ROOT }}/mkl/latest/bin/mkl_tbb_thread.2.dll" ./build/bin - cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/pi_win_proxy_loader.dll" ./build/bin - cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/pi_level_zero.dll" ./build/bin - cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/sycl7.dll" ./build/bin + cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/ur_adapter_level_zero.dll" ./build/bin + cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/ur_adapter_opencl.dll" ./build/bin + cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/ur_loader.dll" ./build/bin + cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/ur_win_proxy_loader.dll" ./build/bin + + cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/sycl8.dll" ./build/bin cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/svml_dispmd.dll" ./build/bin cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/libmmd.dll" ./build/bin cp "${{ env.ONEAPI_ROOT }}/compiler/latest/bin/libiomp5md.dll" ./build/bin + + cp "${{ env.ONEAPI_ROOT }}/dnnl/latest/bin/dnnl.dll" ./build/bin + cp "${{ env.ONEAPI_ROOT }}/tbb/latest/bin/tbb12.dll" ./build/bin + echo "cp oneAPI running time dll files to ./build/bin done" 7z a llama-${{ steps.tag.outputs.name }}-bin-win-sycl-x64.zip ./build/bin/* - - name: Upload artifacts + - name: Upload the release package if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }} uses: actions/upload-artifact@v4 with: @@ -996,12 +1055,17 @@ jobs: run: | & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version + - name: Install ccache + uses: hendrikmuhs/ccache-action@v1.2 + with: + key: ${{ github.job }} + - name: Build id: cmake_build run: | $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path) $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}" - cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIPBLAS=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON + cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON cmake --build build -j ${env:NUMBER_OF_PROCESSORS} windows-latest-cmake-hip-release: @@ -1016,6 +1080,8 @@ jobs: - name: Clone id: checkout uses: actions/checkout@v4 + with: + fetch-depth: 0 - name: Install id: depends @@ -1037,7 +1103,7 @@ jobs: run: | $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path) $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}" - cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIPBLAS=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON + cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON cmake --build build -j ${env:NUMBER_OF_PROCESSORS} md "build\bin\rocblas\library\" cp "${env:HIP_PATH}\bin\hipblas.dll" "build\bin\" @@ -1075,6 +1141,29 @@ jobs: - name: Checkout code uses: actions/checkout@v4 + - name: Build + id: cmake_build + run: | + sysctl -a + mkdir build + cd build + cmake -G Xcode .. \ + -DGGML_METAL_USE_BF16=ON \ + -DGGML_METAL_EMBED_LIBRARY=ON \ + -DLLAMA_BUILD_EXAMPLES=OFF \ + -DLLAMA_BUILD_TESTS=OFF \ + -DLLAMA_BUILD_SERVER=OFF \ + -DCMAKE_SYSTEM_NAME=iOS \ + -DCMAKE_OSX_DEPLOYMENT_TARGET=14.0 \ + -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml + cmake --build . --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO + sudo cmake --install . --config Release + + - name: xcodebuild for swift package + id: xcodebuild + run: | + xcodebuild -scheme llama-Package -destination 'generic/platform=iOS' + - name: Build Xcode project run: xcodebuild -project examples/llama.swiftui/llama.swiftui.xcodeproj -scheme llama.swiftui -sdk iphoneos CODE_SIGNING_REQUIRED=NO CODE_SIGN_IDENTITY= -destination 'generic/platform=iOS' build @@ -1102,35 +1191,16 @@ jobs: ./gradlew build --no-daemon -# freeBSD-latest: -# runs-on: macos-12 -# steps: -# - name: Clone -# uses: actions/checkout@v4 -# -# - name: Build -# uses: cross-platform-actions/action@v0.19.0 -# with: -# operating_system: freebsd -# version: '13.2' -# hypervisor: 'qemu' -# run: | -# sudo pkg update -# sudo pkg install -y gmake automake autoconf pkgconf llvm15 openblas -# gmake CC=/usr/local/bin/clang15 CXX=/usr/local/bin/clang++15 -j `sysctl -n hw.ncpu` - release: if: ${{ ( github.event_name == 'push' && github.ref == 'refs/heads/master' ) || github.event.inputs.create_release == 'true' }} runs-on: ubuntu-latest needs: - - ubuntu-focal-make - ubuntu-latest-cmake - - macOS-latest-make - macOS-latest-cmake - windows-latest-cmake - - windows-latest-cmake-cuda + - windows-2019-cmake-cuda - windows-latest-cmake-hip-release - macOS-latest-cmake-arm64 - macOS-latest-cmake-x64 @@ -1167,7 +1237,7 @@ jobs: - name: Create release id: create_release - uses: anzz1/action-create-release@v1 + uses: ggml-org/action-create-release@v1 env: GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} with: diff --git a/.github/workflows/docker.yml b/.github/workflows/docker.yml index a953cdac9..d71f1eb38 100644 --- a/.github/workflows/docker.yml +++ b/.github/workflows/docker.yml @@ -10,12 +10,10 @@ name: Publish Docker image on: - #pull_request: - push: - branches: - - master - paths: ['.github/workflows/docker.yml', '.devops/*.Dockerfile', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.cuh', '**/*.swift', '**/*.m', '**/*.metal'] - workflow_dispatch: # allows manual triggering, useful for debugging + workflow_dispatch: # allows manual triggering + schedule: + # Rebuild daily rather than on every push because it is expensive + - cron: '12 4 * * *' concurrency: group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} @@ -29,7 +27,6 @@ permissions: jobs: push_to_registry: name: Push Docker image to Docker Hub - #if: github.event.pull_request.draft == false runs-on: ubuntu-latest env: @@ -37,21 +34,14 @@ jobs: strategy: matrix: config: - - { tag: "light", dockerfile: ".devops/llama-cli.Dockerfile", platforms: "linux/amd64,linux/arm64" } - - { tag: "server", dockerfile: ".devops/llama-server.Dockerfile", platforms: "linux/amd64,linux/arm64" } - - { tag: "full", dockerfile: ".devops/full.Dockerfile", platforms: "linux/amd64,linux/arm64" } - - { tag: "light-cuda", dockerfile: ".devops/llama-cli-cuda.Dockerfile", platforms: "linux/amd64" } - - { tag: "server-cuda", dockerfile: ".devops/llama-server-cuda.Dockerfile", platforms: "linux/amd64" } - - { tag: "full-cuda", dockerfile: ".devops/full-cuda.Dockerfile", platforms: "linux/amd64" } - - { tag: "light-musa", dockerfile: ".devops/llama-cli-musa.Dockerfile", platforms: "linux/amd64" } - - { tag: "server-musa", dockerfile: ".devops/llama-server-musa.Dockerfile", platforms: "linux/amd64" } - - { tag: "full-musa", dockerfile: ".devops/full-musa.Dockerfile", platforms: "linux/amd64" } + # Multi-stage build + - { tag: "cpu", dockerfile: ".devops/cpu.Dockerfile", platforms: "linux/amd64,linux/arm64", full: true, light: true, server: true, freediskspace: false} + - { tag: "cuda", dockerfile: ".devops/cuda.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false} + - { tag: "musa", dockerfile: ".devops/musa.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false} + - { tag: "intel", dockerfile: ".devops/intel.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false} + - { tag: "vulkan", dockerfile: ".devops/vulkan.Dockerfile", platforms: "linux/amd64", full: true, light: true, server: true, freediskspace: false} # Note: the rocm images are failing due to a compiler error and are disabled until this is fixed to allow the workflow to complete - #- { tag: "light-rocm", dockerfile: ".devops/llama-cli-rocm.Dockerfile", platforms: "linux/amd64,linux/arm64" } - #- { tag: "server-rocm", dockerfile: ".devops/llama-server-rocm.Dockerfile", platforms: "linux/amd64,linux/arm64" } - #- { tag: "full-rocm", dockerfile: ".devops/full-rocm.Dockerfile", platforms: "linux/amd64,linux/arm64" } - - { tag: "light-intel", dockerfile: ".devops/llama-cli-intel.Dockerfile", platforms: "linux/amd64" } - - { tag: "server-intel", dockerfile: ".devops/llama-server-intel.Dockerfile", platforms: "linux/amd64" } + #- {tag: "rocm", dockerfile: ".devops/rocm.Dockerfile", platforms: "linux/amd64,linux/arm64", full: true, light: true, server: true, freediskspace: true } steps: - name: Check out the repo uses: actions/checkout@v4 @@ -59,10 +49,10 @@ jobs: fetch-depth: 0 # preserve git history, so we can determine the build number - name: Set up QEMU - uses: docker/setup-qemu-action@v2 + uses: docker/setup-qemu-action@v3 - name: Set up Docker Buildx - uses: docker/setup-buildx-action@v2 + uses: docker/setup-buildx-action@v3 - name: Log in to Docker Hub uses: docker/login-action@v2 @@ -82,26 +72,34 @@ jobs: # determine tag name postfix (build number, commit hash) if [[ "${{ env.GITHUB_BRANCH_NAME }}" == "master" ]]; then - TAG_POSTFIX="b${BUILD_NUMBER}" + TAG_POSTFIX="-b${BUILD_NUMBER}" else SAFE_NAME=$(echo "${{ env.GITHUB_BRANCH_NAME }}" | tr '/' '-') - TAG_POSTFIX="${SAFE_NAME}-${SHORT_HASH}" + TAG_POSTFIX="-${SAFE_NAME}-${SHORT_HASH}" fi - # list all tags possible - TAGS="" - TAGS="${TAGS}ghcr.io/${REPO_OWNER}/${REPO_NAME}:${{ matrix.config.tag }}," - TAGS="${TAGS}ghcr.io/${REPO_OWNER}/${REPO_NAME}:${{ matrix.config.tag }}-${TAG_POSTFIX}" - - echo "output_tags=$TAGS" >> $GITHUB_OUTPUT - echo "output_tags=$TAGS" # print out for debugging + if [[ "${{ matrix.config.tag }}" == "cpu" ]]; then + TYPE="" + else + TYPE="-${{ matrix.config.tag }}" + fi + PREFIX="ghcr.io/${REPO_OWNER}/${REPO_NAME}:" + FULLTAGS="${PREFIX}full${TYPE},${PREFIX}full${TYPE}${TAG_POSTFIX}" + LIGHTTAGS="${PREFIX}light${TYPE},${PREFIX}light${TYPE}${TAG_POSTFIX}" + SERVERTAGS="${PREFIX}server${TYPE},${PREFIX}server${TYPE}${TAG_POSTFIX}" + echo "full_output_tags=$FULLTAGS" >> $GITHUB_OUTPUT + echo "light_output_tags=$LIGHTTAGS" >> $GITHUB_OUTPUT + echo "server_output_tags=$SERVERTAGS" >> $GITHUB_OUTPUT + echo "full_output_tags=$FULLTAGS" # print out for debugging + echo "light_output_tags=$LIGHTTAGS" # print out for debugging + echo "server_output_tags=$SERVERTAGS" # print out for debugging env: GITHUB_BRANCH_NAME: ${{ github.head_ref || github.ref_name }} GITHUB_REPOSITORY_OWNER: '${{ github.repository_owner }}' - # https://github.com/jlumbroso/free-disk-space/tree/54081f138730dfa15788a46383842cd2f914a1be#example - name: Free Disk Space (Ubuntu) - uses: jlumbroso/free-disk-space@main + if: ${{ matrix.config.free_disk_space == true }} + uses: ggml-org/free-disk-space@v1.3.1 with: # this might remove tools that are actually needed, # if set to "true" but frees about 6 GB @@ -116,13 +114,59 @@ jobs: docker-images: true swap-storage: true - - name: Build and push Docker image (tagged + versioned) - if: github.event_name == 'push' + - name: Build and push Full Docker image (tagged + versioned) + if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.full == true }} uses: docker/build-push-action@v6 with: context: . push: true platforms: ${{ matrix.config.platforms }} # tag list is generated from step above - tags: ${{ steps.tag.outputs.output_tags }} + tags: ${{ steps.tag.outputs.full_output_tags }} file: ${{ matrix.config.dockerfile }} + target: full + provenance: false + # using github experimental cache + cache-from: type=gha + cache-to: type=gha,mode=max + # return to this if the experimental github cache is having issues + #cache-to: type=local,dest=/tmp/.buildx-cache + #cache-from: type=local,src=/tmp/.buildx-cache + + - name: Build and push Light Docker image (tagged + versioned) + if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.light == true }} + uses: docker/build-push-action@v6 + with: + context: . + push: true + platforms: ${{ matrix.config.platforms }} + # tag list is generated from step above + tags: ${{ steps.tag.outputs.light_output_tags }} + file: ${{ matrix.config.dockerfile }} + target: light + provenance: false + # using github experimental cache + cache-from: type=gha + cache-to: type=gha,mode=max + # return to this if the experimental github cache is having issues + #cache-to: type=local,dest=/tmp/.buildx-cache + #cache-from: type=local,src=/tmp/.buildx-cache + + - name: Build and push Server Docker image (tagged + versioned) + if: ${{ (github.event_name == 'push' || github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && matrix.config.server == true }} + uses: docker/build-push-action@v6 + with: + context: . + push: true + platforms: ${{ matrix.config.platforms }} + # tag list is generated from step above + tags: ${{ steps.tag.outputs.server_output_tags }} + file: ${{ matrix.config.dockerfile }} + target: server + provenance: false + # using github experimental cache + cache-from: type=gha + cache-to: type=gha,mode=max + # return to this if the experimental github cache is having issues + #cache-to: type=local,dest=/tmp/.buildx-cache + #cache-from: type=local,src=/tmp/.buildx-cache diff --git a/.github/workflows/editorconfig.yml b/.github/workflows/editorconfig.yml index ae86e9927..f02b7c219 100644 --- a/.github/workflows/editorconfig.yml +++ b/.github/workflows/editorconfig.yml @@ -23,5 +23,7 @@ jobs: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - - uses: editorconfig-checker/action-editorconfig-checker@main + - uses: editorconfig-checker/action-editorconfig-checker@v2 + with: + version: v3.0.3 - run: editorconfig-checker diff --git a/.github/workflows/nix-ci-aarch64.yml b/.github/workflows/nix-ci-aarch64.yml deleted file mode 100644 index 0da6acdf1..000000000 --- a/.github/workflows/nix-ci-aarch64.yml +++ /dev/null @@ -1,72 +0,0 @@ -name: Nix aarch64 builds - -on: - workflow_dispatch: # allows manual triggering - schedule: - # Rebuild daily rather than on every push because QEMU is expensive (e.g. - # 1.5h instead of minutes with the cold cache). - # - # randint(0, 59), randint(0, 23) - - cron: '26 12 * * *' - # But also rebuild if we touched any of the Nix expressions: - push: - branches: - - master - paths: ['**/*.nix', 'flake.lock'] - pull_request: - types: [opened, synchronize, reopened] - paths: ['**/*.nix', 'flake.lock'] - -concurrency: - group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} - cancel-in-progress: true - -# Fine-grant permission -# https://docs.github.com/en/actions/security-for-github-actions/security-guides/automatic-token-authentication#modifying-the-permissions-for-the-github_token -permissions: - # https://github.com/DeterminateSystems/nix-installer-action?tab=readme-ov-file#with-flakehub - id-token: write - contents: read - -jobs: - nix-build-aarch64: - runs-on: ubuntu-latest - steps: - - name: Checkout repository - uses: actions/checkout@v4 - - name: Install QEMU - # Copy-paste from https://github.com/orgs/community/discussions/8305#discussioncomment-5888654 - run: | - sudo apt-get update - sudo apt-get install -y qemu-user-static qemu-system-aarch64 - sudo usermod -a -G kvm $USER - - name: Install Nix - uses: DeterminateSystems/nix-installer-action@v9 - with: - github-token: ${{ secrets.GITHUB_TOKEN }} - extra-conf: | - extra-platforms = aarch64-linux - extra-system-features = nixos-test kvm - extra-substituters = https://llama-cpp.cachix.org https://cuda-maintainers.cachix.org - extra-trusted-public-keys = llama-cpp.cachix.org-1:H75X+w83wUKTIPSO1KWy9ADUrzThyGs8P5tmAbkWhQc= cuda-maintainers.cachix.org-1:0dq3bujKpuEPMCX6U4WylrUDZ9JyUG0VpVZa7CNfq5E= - - uses: DeterminateSystems/magic-nix-cache-action@v2 - with: - upstream-cache: https://${{ matrix.cachixName }}.cachix.org - - name: Set-up cachix to push the results to - uses: cachix/cachix-action@v13 - with: - authToken: '${{ secrets.CACHIX_AUTH_TOKEN }}' - name: llama-cpp - - name: Show all output paths - run: > - nix run github:nix-community/nix-eval-jobs - -- --gc-roots-dir gcroot - --flake - ".#packages.aarch64-linux" - - name: Build - run: > - nix run github:Mic92/nix-fast-build - -- --skip-cached --no-nom - --systems aarch64-linux - --flake - ".#checks.aarch64-linux" diff --git a/.github/workflows/nix-ci.yml b/.github/workflows/nix-ci.yml deleted file mode 100644 index 8ecbbe53b..000000000 --- a/.github/workflows/nix-ci.yml +++ /dev/null @@ -1,79 +0,0 @@ -name: Nix CI - -on: - workflow_dispatch: # allows manual triggering - push: - branches: - - master - pull_request: - types: [opened, synchronize, reopened] - -concurrency: - group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} - cancel-in-progress: true - -# Fine-grant permission -# https://docs.github.com/en/actions/security-for-github-actions/security-guides/automatic-token-authentication#modifying-the-permissions-for-the-github_token -permissions: - # https://github.com/DeterminateSystems/nix-installer-action?tab=readme-ov-file#with-flakehub - id-token: write - contents: read - -jobs: - nix-eval: - strategy: - fail-fast: false - matrix: - os: [ ubuntu-latest, macos-latest ] - runs-on: ${{ matrix.os }} - steps: - - name: Checkout repository - uses: actions/checkout@v4 - - name: Install Nix - uses: DeterminateSystems/nix-installer-action@v9 - with: - github-token: ${{ secrets.GITHUB_TOKEN }} - extra-conf: | - extra-substituters = https://llama-cpp.cachix.org https://cuda-maintainers.cachix.org - extra-trusted-public-keys = llama-cpp.cachix.org-1:H75X+w83wUKTIPSO1KWy9ADUrzThyGs8P5tmAbkWhQc= cuda-maintainers.cachix.org-1:0dq3bujKpuEPMCX6U4WylrUDZ9JyUG0VpVZa7CNfq5E= - - uses: DeterminateSystems/magic-nix-cache-action@v2 - with: - upstream-cache: https://${{ matrix.cachixName }}.cachix.org - - name: List all flake outputs - run: nix flake show --all-systems - - name: Show all output paths - run: > - nix run github:nix-community/nix-eval-jobs - -- --gc-roots-dir gcroot - --flake - ".#packages.$(nix eval --raw --impure --expr builtins.currentSystem)" - nix-build: - strategy: - fail-fast: false - matrix: - os: [ ubuntu-latest, macos-latest ] - runs-on: ${{ matrix.os }} - steps: - - name: Checkout repository - uses: actions/checkout@v4 - - name: Install Nix - uses: DeterminateSystems/nix-installer-action@v9 - with: - github-token: ${{ secrets.GITHUB_TOKEN }} - extra-conf: | - extra-substituters = https://llama-cpp.cachix.org https://cuda-maintainers.cachix.org - extra-trusted-public-keys = llama-cpp.cachix.org-1:H75X+w83wUKTIPSO1KWy9ADUrzThyGs8P5tmAbkWhQc= cuda-maintainers.cachix.org-1:0dq3bujKpuEPMCX6U4WylrUDZ9JyUG0VpVZa7CNfq5E= - - uses: DeterminateSystems/magic-nix-cache-action@v2 - with: - upstream-cache: https://${{ matrix.cachixName }}.cachix.org - - name: Set-up cachix to push the results to - uses: cachix/cachix-action@v13 - with: - authToken: '${{ secrets.CACHIX_AUTH_TOKEN }}' - name: llama-cpp - - name: Build - run: > - nix run github:Mic92/nix-fast-build - -- --skip-cached --no-nom - --flake - ".#checks.$(nix eval --raw --impure --expr builtins.currentSystem)" diff --git a/.github/workflows/nix-flake-update.yml b/.github/workflows/nix-flake-update.yml deleted file mode 100644 index 3a6a96e26..000000000 --- a/.github/workflows/nix-flake-update.yml +++ /dev/null @@ -1,22 +0,0 @@ -name: update-flake-lock -on: - workflow_dispatch: - schedule: - - cron: '0 0 * * 0' # runs weekly on Sunday at 00:00 - -jobs: - lockfile: - runs-on: ubuntu-latest - steps: - - name: Checkout repository - uses: actions/checkout@v4 - - name: Install Nix - uses: DeterminateSystems/nix-installer-action@main - - name: Update flake.lock - uses: DeterminateSystems/update-flake-lock@main - with: - pr-title: "nix: update flake.lock" - pr-labels: | - nix - pr-reviewers: philiptaron,SomeoneSerge - token: ${{ secrets.FLAKE_TOKEN }} diff --git a/.github/workflows/nix-publish-flake.yml b/.github/workflows/nix-publish-flake.yml deleted file mode 100644 index 2c3c1ebda..000000000 --- a/.github/workflows/nix-publish-flake.yml +++ /dev/null @@ -1,36 +0,0 @@ -# Make the flake discoverable on https://flakestry.dev and https://flakehub.com/flakes -name: "Publish a flake to flakestry & flakehub" -on: - push: - tags: - - "*" - workflow_dispatch: - inputs: - tag: - description: "The existing tag to publish" - type: "string" - required: true -jobs: - flakestry-publish: - runs-on: ubuntu-latest - permissions: - id-token: "write" - contents: "read" - steps: - - uses: flakestry/flakestry-publish@main - with: - version: "${{ inputs.tag || github.ref_name }}" - flakehub-publish: - runs-on: "ubuntu-latest" - permissions: - id-token: "write" - contents: "read" - steps: - - uses: "actions/checkout@v4" - with: - ref: "${{ (inputs.tag != null) && format('refs/tags/{0}', inputs.tag) || '' }}" - - uses: "DeterminateSystems/nix-installer-action@main" - - uses: "DeterminateSystems/flakehub-push@main" - with: - visibility: "public" - tag: "${{ inputs.tag }}" diff --git a/.github/workflows/python-lint.yml b/.github/workflows/python-lint.yml index a8d46f31d..ddfdf73b8 100644 --- a/.github/workflows/python-lint.yml +++ b/.github/workflows/python-lint.yml @@ -1,6 +1,13 @@ name: flake8 Lint -on: [push, pull_request] +on: + push: + branches: + - master + paths: ['.github/workflows/python-lint.yml', '**/*.py'] + pull_request: + types: [opened, synchronize, reopened] + paths: ['.github/workflows/python-lint.yml', '**/*.py'] concurrency: group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} diff --git a/.github/workflows/server.yml b/.github/workflows/server.yml index 699ac095d..671fe595c 100644 --- a/.github/workflows/server.yml +++ b/.github/workflows/server.yml @@ -76,20 +76,26 @@ jobs: run: | pip install -r examples/server/tests/requirements.txt - - name: Verify server deps - id: verify_server_deps + # Setup nodejs (to be used for verifying bundled index.html) + - uses: actions/setup-node@v4 + with: + node-version: '22.11.0' + + - name: Verify bundled index.html + id: verify_server_index_html run: | git config --global --add safe.directory $(realpath .) - cd examples/server - git ls-files --others --modified + cd examples/server/webui git status - ./deps.sh + npm ci + npm run build git status - not_ignored_files="$(git ls-files --others --modified)" - echo "Modified files: ${not_ignored_files}" - if [ -n "${not_ignored_files}" ]; then - echo "Repository is dirty or server deps are not built as expected" - echo "${not_ignored_files}" + modified_files="$(git status -s)" + echo "Modified files: ${modified_files}" + if [ -n "${modified_files}" ]; then + echo "Repository is dirty or server/webui is not built as expected" + echo "Hint: You may need to follow Web UI build guide in server/README.md" + echo "${modified_files}" exit 1 fi @@ -122,14 +128,14 @@ jobs: id: server_integration_tests run: | cd examples/server/tests - PORT=8888 ./tests.sh + ./tests.sh - name: Slow tests id: server_integration_tests_slow if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }} run: | cd examples/server/tests - PORT=8888 ./tests.sh --stop --no-skipped --no-capture --tags slow + SLOW_TESTS=1 ./tests.sh server-windows: @@ -180,11 +186,12 @@ jobs: run: | cd examples/server/tests $env:PYTHONIOENCODING = ":replace" - behave.exe --summary --stop --no-capture --exclude 'issues|wrong_usages|passkey' --tags llama.cpp + pytest -v -x - name: Slow tests id: server_integration_tests_slow if: ${{ (github.event.schedule || github.event.inputs.slow_tests == 'true') && matrix.build_type == 'Release' }} run: | cd examples/server/tests - behave.exe --stop --no-skipped --no-capture --tags slow + $env:SLOW_TESTS = "1" + pytest -v -x diff --git a/.gitignore b/.gitignore index 1092d097a..694f36e04 100644 --- a/.gitignore +++ b/.gitignore @@ -3,6 +3,7 @@ *.a *.bat *.bin +*.d *.dll *.dot *.etag @@ -17,6 +18,7 @@ *.metallib *.o *.so +*.swp *.tmp # IDE / OS @@ -103,6 +105,10 @@ examples/server/*.mjs.hpp !examples/sycl/*.bat !examples/sycl/*.sh +# Server Web UI temporary files +node_modules +examples/server/webui/dist + # Python /.venv @@ -133,3 +139,7 @@ poetry.toml # Test models for lora adapters /lora-tests + +# Local scripts +/run-vim.sh +/run-chat.sh diff --git a/.gitmodules b/.gitmodules index 5861d59cb..23ce5ff05 100644 --- a/.gitmodules +++ b/.gitmodules @@ -1,3 +1,3 @@ [submodule "kompute"] - path = ggml/src/kompute + path = ggml/src/ggml-kompute/kompute url = https://github.com/nomic-ai/kompute.git diff --git a/AUTHORS b/AUTHORS index 1bd36158a..2eb60806a 100644 --- a/AUTHORS +++ b/AUTHORS @@ -1,4 +1,4 @@ -# date: Wed Jun 26 19:36:34 EEST 2024 +# date: Thu Nov 28 20:46:15 EET 2024 # this file is auto-generated by scripts/gen-authors.sh 0cc4m @@ -7,6 +7,7 @@ 2f38b454 3ooabkhxtn <31479382+3ooabkhxtn@users.noreply.github.com> 44670 <44670@users.noreply.github.com> +65a <10104049+65a@users.noreply.github.com> AN Long AT Aarni Koskela @@ -19,20 +20,28 @@ Adithya Balaji AdithyanI Adrian Adrian Hesketh +Ahmad Tameem <113388789+Tameem-10xE@users.noreply.github.com> Ahmet Zeer AidanBeltonS <87009434+AidanBeltonS@users.noreply.github.com> +AidanBeltonS Aisuko +Akarshan Biswas Akarshan Biswas +Al Mochkin <14274697+amochkin@users.noreply.github.com> Albert Jin Alberto <57916483+albbus-stack@users.noreply.github.com> +Alberto Cabrera Pérez +Alberto Cabrera Pérez Alex Alex Azarov Alex Azarov Alex Klinkhamer Alex Klinkhamer Alex Nguyen +Alex O'Connell <35843486+acon96@users.noreply.github.com> Alex Petenchea Alex Renda +Alex Tuddenham <61622354+AlexsCode@users.noreply.github.com> Alex von Gluck IV Alexey Parfenov Ali Chraghi <63465728+alichraghi@users.noreply.github.com> @@ -45,18 +54,25 @@ AmirAli Mirian <37371367+amiralimi@users.noreply.github.com> Ananta Bastola Anas Ahouzi <112881240+aahouzi@users.noreply.github.com> András Salamon +Andreas (Andi) Kunar Andrei Andrew Canis Andrew Downing Andrew Duffy Andrew Godfrey +Andrew Minh Nguyen <40281306+amqdn@users.noreply.github.com> +Andy Salerno Andy Tai +Anthony Van de Gejuchte +Antonis Makropoulos Arik Poznanski +Armen Kaleshian Artem Artem Zinnatullin Artyom Lebedev Asbjørn Olling Ásgeir Bjarni Ingvarsson +Asghar Ghorbani Ashish <1856117+ashishdatta@users.noreply.github.com> Ashok Gelal <401055+ashokgelal@users.noreply.github.com> Ashraful Islam @@ -76,12 +92,16 @@ Ben Williams Benjamin Findley <39356821+Kartoffelsaft@users.noreply.github.com> Benjamin Lecaillon <84293038+blecaillon@users.noreply.github.com> Bernat Vadell +Bert Wagner Bingan <70050083+binganao@users.noreply.github.com> +Bjarke Viksøe <164612031+bviksoe@users.noreply.github.com> Bodo Graumann Bono Lv Borislav Stanimirov Branden Butler +Brandon Squizzato <35474886+bsquizz@users.noreply.github.com> Brian +Brian Cunnie Bruce MacDonald Bryan Honof CJ Pais @@ -90,32 +110,47 @@ Calvin Laurenson Cameron Cameron Kaiser Carolinabanana <140120812+Carolinabanana@users.noreply.github.com> +CarryFun <76023481+CarryFun@users.noreply.github.com> +Carsten Kragelund Jørgensen +CarterLi999 <664681047@qq.com> Casey Primozic Casey Primozic CausalLM <148736309+CausalLM@users.noreply.github.com> Cebtenzzre Chad Brewbaker +Changyeon Kim Chao Jiang +Charles Xu <63788048+chaxu01@users.noreply.github.com> +Charles Xu +Chen Xi +Chen Xi Cheng Shao +Chenguang Li <87689256+noemotiovon@users.noreply.github.com> Chris Elrod Chris Kuehl Christian Demsar Christian Demsar Christian Falch <875252+chrfalch@users.noreply.github.com> Christian Kögler +Christian Köhnenkamp Christian Zhou-Zheng <59622928+christianazinn@users.noreply.github.com> Clark Saben <76020733+csaben@users.noreply.github.com> Clint Herron +Conrad Kramer CrispStrobe <154636388+CrispStrobe@users.noreply.github.com> +Csaba Kecskemeti Cuong Trinh Manh DAN™ Damian Stewart +Dan Johansson <164997844+eddnjjn@users.noreply.github.com> +Dan Johansson Dane Madsen DaniAndTheWeb <57776841+DaniAndTheWeb@users.noreply.github.com> Daniel Bevenius Daniel Drake Daniel Hiltgen Daniel Illescas Romero +Daniel Kleine <53251018+d-kleine@users.noreply.github.com> Daniele <57776841+daniandtheweb@users.noreply.github.com> DannyDaemonic Dat Quoc Nguyen <2412555+datquocnguyen@users.noreply.github.com> @@ -129,19 +164,28 @@ David Pflug David Renshaw David Sommers <12738+databyte@users.noreply.github.com> David Yang +DavidKorczynski Dawid Potocki Dawid Wysocki <62249621+TortillaZHawaii@users.noreply.github.com> Dean Deins +Denis Spasyuk <34203011+dspasyuk@users.noreply.github.com> +Derrick T. Woolworth Deven Mistry <31466137+deven367@users.noreply.github.com> +Dibakar Gope Didzis Gosko +Diego Devesa +Diogo Teles Sant'Anna Djip007 Don Mahurin DooWoong Lee (David) Doomsdayrs <38189170+Doomsdayrs@users.noreply.github.com> +Dou Xinpeng <15529241576@163.com> +Dou Xinpeng <81913537+Dou-Git@users.noreply.github.com> Douglas Hanley Dr. Tom Murphy VII Ph.D <499244+tom7@users.noreply.github.com> Ebey Abraham +Echo Nolan Ed Lee Ed Lepedus Eddie-Wang @@ -151,10 +195,13 @@ Elbios <141279586+Elbios@users.noreply.github.com> Elton Kola Engininja2 <139037756+Engininja2@users.noreply.github.com> Equim +Eric Curtin +Eric Curtin Eric Sommerlade Eric Zhang <34133756+EZForever@users.noreply.github.com> Erik Garrison Erik Scholz +Esko Toivonen Ettore Di Giacinto Evan Jones Evan Miller @@ -166,19 +213,26 @@ FK Fabian Fabio R. Sluzala Faez Shakil +Faisal Zaghloul +Faisal Zaghloul +Fan Shupei FantasyGmm <16450052+FantasyGmm@users.noreply.github.com> +Farbod Bijary <110523279+farbodbj@users.noreply.github.com> Fattire <528174+fat-tire@users.noreply.github.com> Felix Finn Voorhees Firat +FirstTimeEZ <179362031+FirstTimeEZ@users.noreply.github.com> Folko-Ven <71110216+Folko-Ven@users.noreply.github.com> Foul-Tarnished <107711110+Foul-Tarnished@users.noreply.github.com> Francisco Melo <43780565+francis2tm@users.noreply.github.com> Frank Mai FrankHB +Frankie Robertson Fred Douglas <43351173+fredlas@users.noreply.github.com> Frederik Vogel Gabe Goodhart +Gabe Goodhart GainLee Galunid Gary Linscott @@ -187,11 +241,13 @@ Gavin Zhao Genkagaku.GPT Georgi Gerganov Gilad S +Gilad S. <7817232+giladgd@users.noreply.github.com> Giuseppe Scrivano GiviMAD Govlzkoy Guillaume "Vermeille" Sanchez Guillaume Wenzek +Guoliang Hua <32868157+nbcsm@users.noreply.github.com> Guoteng <32697156+SolenoidWGT@users.noreply.github.com> Gustavo Rocha Dias <91472747+gustrd@users.noreply.github.com> Haggai Nuchi @@ -213,11 +269,14 @@ Hong Bo PENG Hongyu Ouyang <96765450+casavaca@users.noreply.github.com> Howard Su Hua Jiang +Huang Qi Huawei Lin Hugo Roussel +Huifeng Ou <79071290+ho2103@users.noreply.github.com> Ian Bull Ian Bull Ian Scrivener +Icecream95 Ido S IgnacioFDM Igor Okulist @@ -226,11 +285,15 @@ Ilya Kurdyukov <59548320+ilyakurdyukov@users.noreply.github.com> Ionoclast Laboratories Isaac McFadyen IsaacDynamo <61521674+IsaacDynamo@users.noreply.github.com> +Ivan +Ivan Filipov <159561759+vanaka11@users.noreply.github.com> Ivan Komarov Ivan Stepanov JH23X <165871467+JH23X@users.noreply.github.com> +Jack Mousseau Jack Mousseau JackJollimore <130917767+JackJollimore@users.noreply.github.com> +Jaeden Amero Jaemin Son Jag Chadha Jakub N @@ -243,10 +306,14 @@ Jannis Schönleber Jared Van Bortel Jared Van Bortel Jason McCartney +Jason Stillerman Jean-Christophe Hoelt Jean-Michaël Celerier Jed Fox +Jeff Bolz +Jeffrey Morgan Jeffrey Quesnelle +Jeroen Mostert Jesse Jojo Johnson Jeximo Jhen-Jie Hong @@ -258,6 +325,9 @@ Jiří Podivín <66251151+jpodivin@users.noreply.github.com> Jiří Sejkora Joan Fontanals Joan Fontanals +João Dinis Ferreira +Joe Eli McIlvain +Joe Todd Johan Johannes Gäßler Johannes Rudolph @@ -274,7 +344,9 @@ Joyce Juan Calderon-Perez <835733+gaby@users.noreply.github.com> Judd Julius Arkenberg +Jun Hee Yoo Jun Jie <71215065+junnjiee16@users.noreply.github.com> +Junil Kim Junyang Lin Juraj Bednar Justin Parker @@ -292,12 +364,14 @@ Karthik Sethuraman Kasumi <90275229+kasumi-1@users.noreply.github.com> Kawrakow <48489457+ikawrakow@users.noreply.github.com> Keiichi Tabata +Keke Han Kenvix ⭐ Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com> Kevin Gibbons Kevin Ji <1146876+kevinji@users.noreply.github.com> Kevin Kwok Kevin Lo +Kevin Wang Kolen Cheung Konstantin Herud Konstantin Zhuravlyov @@ -315,22 +389,29 @@ LeonEricsson <70749762+LeonEricsson@users.noreply.github.com> Leonardo Neumann Li Tan Linwei Wang +Liu Jia <109258120+Septa2112@users.noreply.github.com> +Liu Jia LoganDark +Loïc Carrère LostRuins <39025047+LostRuins@users.noreply.github.com> Luciano Luo Tian Lyle Dean +M-A M. Yusuf Sarıgöz +Ma Mingfei Maarten ter Huurne Mack Straight Maël Kerbiriou MaggotHATE +Mahesh Madhav <67384846+heshpdx@users.noreply.github.com> Manuel <44313466+makuche@users.noreply.github.com> Marc Köhlbrugge Marco Matthies <71844+marcom@users.noreply.github.com> Marcus Dunn <51931484+MarcusDunn@users.noreply.github.com> Marian Cepok Mark Fairbairn +Mark Zhuang Marko Tasic Markus Tavenrath Martin Delille @@ -342,11 +423,15 @@ MasterYi1024 <39848311+MasterYi1024@users.noreply.github.com> Mateusz Charytoniuk Matheus C. França Matheus Gabriel Alves Silva +Mathieu Geli Mathieu Nayrolles +Mathijs Henquet Mathijs de Bruin Matt Clayton <156335168+mattjcly@users.noreply.github.com> Matt Pulver +Matt Stephenson Matteo Boschini <12133566+mbosc@users.noreply.github.com> +Matteo Mortari Mattheus Chediak Matthew Tejo Matvey Soloviev @@ -356,8 +441,10 @@ Maxime <672982+maximegmd@users.noreply.github.com> Maximilian Winter Meng Zhang Meng, Hengyu +Mengqing Cao Merrick Christensen Michael Coppola +Michael Francis Michael Hueschen Michael Kesper Michael Klimenko @@ -365,41 +452,57 @@ Michael Podvitskiy Michael Potter Michael de Gans Michaël de Vries +Michał Tuszyński Mihai Mike Mikko Juola Minsoo Cheong <54794500+mscheong01@users.noreply.github.com> +Minsoo Cheong Mirko185 Mirror Azure <54669636+MirrorAzure@users.noreply.github.com> +MistApproach <98988043+MistApproach@users.noreply.github.com> Miwa / Ensan <63481257+ensan-hcl@users.noreply.github.com> Mohammadreza Hendiani Mohammadreza Hendiani +Molly Sophia +MorganRO8 <47795945+MorganRO8@users.noreply.github.com> Murilo Santana Musab Gultekin Nam D. Tran <42194884+namtranase@users.noreply.github.com> Nathan Epstein +Natsu NawafAlansari <72708095+NawafAlansari@users.noreply.github.com> Nebula Neo Zhang <14088817+arthw@users.noreply.github.com> Neo Zhang Neo Zhang Jianyu Neuman Vong +Nexes the Old <124105151+Nexesenex@users.noreply.github.com> Nexesenex <124105151+Nexesenex@users.noreply.github.com> Niall Coates <1349685+Niall-@users.noreply.github.com> +Nicholai Tukanov +Nico Bosshard Nicolai Weitkemper Nicolás Pérez Nigel Bosch Niklas Korz +NikolaiLyssogor <59844691+NikolaiLyssogor@users.noreply.github.com> Nikolas <127742645+nneubacher@users.noreply.github.com> Nindaleth +OSecret <135510162+OLSecret@users.noreply.github.com> Oleksandr Nikitin Oleksii Maryshchenko Olivier Chafik Ondřej Čertík Ouadie EL FAROUKI +PAB +Pablo Duboue +Pascal Patry Patrice Ferlet Paul Tsochantaris +Pavel Zloi Pavol Rusnak +Paweł Wodnicki <151604+32bitmicro@users.noreply.github.com> Pedro Cuenca Peter Sugihara Phil H <5756783+phiharri@users.noreply.github.com> @@ -407,10 +510,15 @@ Philip Taron Phillip Kravtsov Pierre Alexandre SCHEMBRI Pierrick Hymbert +Pieter Ouwerkerk +Plamen Minev +Prashant Vithule <119530321+Vithulep@users.noreply.github.com> Przemysław Pawełczyk Qin Yue Chen <71813199+chenqiny@users.noreply.github.com> Qingyou Meng Qu Zongfu <43257352+yancaoweidaode@users.noreply.github.com> +R0CKSTAR +R0CKSTAR RJ Adriaansen Radoslav Gerganov Radosław Gryta @@ -419,11 +527,13 @@ Raj Hammeer Singh Hada Ralph Soika Rand Xie Randall Fitzgerald +Random Fly Reinforce-II Ren Xuancheng Rene Leonhardt <65483435+reneleonhardt@users.noreply.github.com> RhinoDevel Riceball LEE +Rich Dougherty Richard Kiss Richard Roberson Rick G <26732651+TheFlipbook@users.noreply.github.com> @@ -439,21 +549,30 @@ Robey Holderith Robyn Roger Meier Roland <14355895+rbur0425@users.noreply.github.com> +Romain Biessy Romain D <90720+Artefact2@users.noreply.github.com> Romain Neutron Roman Parykin Ron Evans Ron Jailall +Roni Ronny Brendel Ronsor Rowan Hart +Ruchira Hasaranga +Ruixin Huang <18860020911@163.com> Rune <43761327+Rune-AI@users.noreply.github.com> +RunningLeon +RunningLeon Ryan Landay Ryder Wishart Ryuei Rőczey Barnabás <31726601+An0nie@users.noreply.github.com> +SRHMorris <69468379+SRHMorris@users.noreply.github.com> +SXX SakuraUmi Salvador E. Tropea +Salvatore Mesoraca Sam Spilsbury Sami Farin <3876865+Safari77@users.noreply.github.com> Samuel Maynard @@ -463,23 +582,29 @@ Sebastián A SebastianApel <13675545+SebastianApel@users.noreply.github.com> Senemu <10880819+Senemu@users.noreply.github.com> Sergey Alirzaev +Sergio López Sergio López Sertaç Özercan <852750+sozercan@users.noreply.github.com> SeungWon Jeong <65549245+redlion0929@users.noreply.github.com> ShadovvBeast Shakhar Dasgupta +Shane A Shangning Xu <32517059+xushangning@users.noreply.github.com> +Shankar +Shanshan Shen <467638484@qq.com> Shijie <821898965@qq.com> Shintarou Okada Shouzheng Liu <61452103+lshzh-ww@users.noreply.github.com> Shouzheng Liu Shuichi Tsutsumi +Shupei Fan Sigbjørn Skjæret Simon Willison Siwen Yu Sky Yan Slaren <2141330+slaren@users.noreply.github.com> Slava Primenko +Small Grass Forest SoftwareRenderer <138734813+SoftwareRenderer@users.noreply.github.com> Someone Someone Serge @@ -491,12 +616,15 @@ Stefan Sydow Steffen Röcker Stephan Walter Stephen Nichols +Steve Bonds Steve Grubb Steven Prichard Steven Roussey Steward Garcia <57494570+FSSRepo@users.noreply.github.com> +StrangeBytesDev <141275258+StrangeBytesDev@users.noreply.github.com> Suaj Carrot <72162667+SuajCarrot@users.noreply.github.com> SuperUserNameMan +Sutou Kouhei Tai Duc Nguyen Taikono-Himazin Tameem <113388789+AhmadTameem@users.noreply.github.com> @@ -507,7 +635,9 @@ Theia Vogel Thérence <13496987+Royalphax@users.noreply.github.com> Thibault Terrasson Thomas Klausner +Thorsten Sommer Tim Miller +Tim Wang Timmy Knight Timothy Cronin <40186632+4imothy@users.noreply.github.com> Ting Lou @@ -517,24 +647,31 @@ Tom C Tom Jobbins <784313+TheBloke@users.noreply.github.com> Tomas Tomáš Pazdiora +Tony Wasserka <4840017+neobrain@users.noreply.github.com> Tristan Druyen Tristan Ross +Trivikram Kamat <16024985+trivikr@users.noreply.github.com> Tungsten842 <886724vf@anonaddy.me> Tungsten842 Tushar UEXTM.com <84163508+uextm@users.noreply.github.com> +Ujjawal Panchal <31011628+Ujjawal-K-Panchal@users.noreply.github.com> Ulrich Drepper Uzo Nweke Vaibhav Srivastav Val Kharitonov Valentin Konovalov Valentyn Bezshapkin <61702053+valentynbez@users.noreply.github.com> +Vali Malinoiu <0x4139@gmail.com> Victor Nogueira Victor Z. Peng +Viet-Anh NGUYEN (Andrew) +Vinesh Janarthanan <36610342+VJHack@users.noreply.github.com> Vlad Vladimir Vladimir Malyutin Vladimir Zorin +VoidIsVoid <343750470@qq.com> Volodymyr Vitvitskyi <72226+signalpillar@users.noreply.github.com> WangHaoranRobin <56047610+WangHaoranRobin@users.noreply.github.com> Weird Constructor @@ -551,15 +688,22 @@ Xiang (Kevin) Li Xiao-Yong Jin XiaotaoChen Xiaoyi Chen +Xie Yanbo Xingchen Song(宋星辰) +Xinpeng Dou <81913537+Dou-Git@users.noreply.github.com> Xuan Son Nguyen +Yaiko Yann Follet <131855179+YannFollet@users.noreply.github.com> Yaroslav Yazan Agha-Schrader Yiming Cui Yishuo Wang +Yoshi Suhara +Yoshi Suhara +Younes Belkada <49240599+younesbelkada@users.noreply.github.com> Yueh-Po Peng <94939112+y10ab1@users.noreply.github.com> Yui +Yuri Khrustalev Yusuf Kağan Hanoğlu Yuval Peled <31162840+Yuval-Peled@users.noreply.github.com> ZHAOKAI WANG @@ -568,6 +712,8 @@ Zay <95888118+isaiahbjork@users.noreply.github.com> Zenix Zhang Peiyuan Zheng.Deng <32841220+dengzheng-cloud@users.noreply.github.com> +Zhenwei Jin <109658203+kylo5aby@users.noreply.github.com> +Zhiyuan Li ZhouYuChen Ziad Ben Hadj-Alouane Ziang Wu <97337387+ZiangWu-77@users.noreply.github.com> @@ -581,6 +727,7 @@ alexpinel <93524949+alexpinel@users.noreply.github.com> alonfaraj alwqx amd-lalithnc +amritahs-ibm andrijdavid anon998 <131767832+anon998@users.noreply.github.com> anzz1 @@ -588,14 +735,18 @@ apaz apcameron <37645737+apcameron@users.noreply.github.com> arch-btw <57669023+arch-btw@users.noreply.github.com> arcrank +ardfork <134447697+ardfork@users.noreply.github.com> arlo-phoenix <140345165+arlo-phoenix@users.noreply.github.com> at8u <129688334+at8u@users.noreply.github.com> automaticcat +awatuna <23447591+awatuna@users.noreply.github.com> +b4b4o bandoti <141645996+bandoti@users.noreply.github.com> beiller bhubbb <79117352+bhubbb@users.noreply.github.com> bmwl bobqianic <129547291+bobqianic@users.noreply.github.com> +brucepro bryanSwk <93190252+bryanSwk@users.noreply.github.com> bsilvereagle bssrdf @@ -614,10 +765,14 @@ cpumaxx <163466046+cpumaxx@users.noreply.github.com> crasm crasm daboe01 +daghanerdonmez <44506702+daghanerdonmez@users.noreply.github.com> +daminho <37615795+daminho@users.noreply.github.com> david raistrick ddh0 ddpasa <112642920+ddpasa@users.noreply.github.com> deepdiffuser <112834445+deepdiffuser@users.noreply.github.com> +devojony <61173062+devojony@users.noreply.github.com> +ditsuke divinity76 dm4 dotpy314 <33351922+dotpy314@users.noreply.github.com> @@ -629,14 +784,18 @@ ebraminio eiery <19350831+eiery@users.noreply.github.com> eric8607242 fairydreaming <166155368+fairydreaming@users.noreply.github.com> +fengerhu1 <2748250768@qq.com> fraxy-v <65565042+fraxy-v@users.noreply.github.com> github-actions[bot] gliptic goerch grahameth <96447521+grahameth@users.noreply.github.com> +gtygo gwjr <502526+gwjr@users.noreply.github.com> h-h-h-h <13482553+h-h-h-h@users.noreply.github.com> hankcs +haopeng <657407891@qq.com> +hipudding hoangmit hongbo.mo <352280764@qq.com> hopkins385 <98618192+hopkins385@users.noreply.github.com> @@ -649,12 +808,14 @@ hxer7963 hydai iSma iacore <74560659+iacore@users.noreply.github.com> +icppWorld <124377669+icppWorld@users.noreply.github.com> igarnier intelmatt <61025942+intelmatt@users.noreply.github.com> iohub jacobi petrucciani <8117202+jpetrucciani@users.noreply.github.com> jaime-m-p <167997752+jaime-m-p@users.noreply.github.com> jameswu2014 <545426914@qq.com> +jdomke <28772296+jdomke@users.noreply.github.com> jiez <373447296@qq.com> jneem joecryptotoo <80373433+joecryptotoo@users.noreply.github.com> @@ -677,28 +838,35 @@ klosax <131523366+klosax@users.noreply.github.com> kunal-vaishnavi <115581922+kunal-vaishnavi@users.noreply.github.com> kunnis kuronekosaiko +kustaaya <58045274+kustaaya@users.noreply.github.com> kuvaus <22169537+kuvaus@users.noreply.github.com> kwin1412 <42286931+kwin1412@users.noreply.github.com> l3utterfly +laik ldwang le.chang leejet +leo-pony limitedAtonement liuwei-git <14815172+liuwei-git@users.noreply.github.com> lon <114724657+longregen@users.noreply.github.com> loonerin <132926317+loonerin@users.noreply.github.com> +ltoniazzi <61414566+ltoniazzi@users.noreply.github.com> luoyu-intel m3ndax maddes8cht <55592906+maddes8cht@users.noreply.github.com> makomk manikbhandari maor-ps <154728172+maor-ps@users.noreply.github.com> +matiaslin <45382001+matiaslin@users.noreply.github.com> +matteo mdrokz mgroeber9110 <45620825+mgroeber9110@users.noreply.github.com> minarchist mj-shifu <77107165+mj-shifu@users.noreply.github.com> mmyjona momonga <115213907+mmnga@users.noreply.github.com> +momonga <146910567+mmngays@users.noreply.github.com> moritzbrantner <31051084+moritzbrantner@users.noreply.github.com> mzcu nanahi <130121847+na-na-hi@users.noreply.github.com> @@ -716,8 +884,10 @@ omahs <73983677+omahs@users.noreply.github.com> oobabooga <112222186+oobabooga@users.noreply.github.com> opparco ostix360 <55257054+ostix360@users.noreply.github.com> +pculliton pengxin99 perserk +piDack <104877312+piDack@users.noreply.github.com> pmysl postmasters pudepiedj @@ -733,6 +903,7 @@ runfuture sandyiscool sasha0552 semidark +serhii-nakon <57632032+serhii-nakon@users.noreply.github.com> sharpHL <132747147+sharpHL@users.noreply.github.com> shibe2 singularity <12184989+singularity-s0@users.noreply.github.com> @@ -741,42 +912,55 @@ sjxx <63994076+ylsdamxssjxxdd@users.noreply.github.com> slaren <2141330+slaren@users.noreply.github.com> slaren snadampal <87143774+snadampal@users.noreply.github.com> +standby24x7 staviq stduhpf strawberrymelonpanda <152940198+strawberrymelonpanda@users.noreply.github.com> swittk takov751 <40316768+takov751@users.noreply.github.com> tarcey +tc-mb <157115220+tc-mb@users.noreply.github.com> texmex76 <40733439+texmex76@users.noreply.github.com> thement <40525767+thement@users.noreply.github.com> +thewh1teagle <61390950+thewh1teagle@users.noreply.github.com> tjohnman +toyer <2042519524@qq.com> tslmy ubik2 uint256_t uint256_t unbounded +uvos valiray <133289098+valiray@users.noreply.github.com> +vb vik viric vodkaslime <646329483@qq.com> vvhg1 <94630311+vvhg1@users.noreply.github.com> vxiiduu <73044267+vxiiduu@users.noreply.github.com> +wangshuai09 <391746016@qq.com> wbpxre150 <100937007+wbpxre150@users.noreply.github.com> whoreson <139810751+whoreson@users.noreply.github.com> woachk <24752637+woachk@users.noreply.github.com> wonjun Jang woodx <124784234+woodx9@users.noreply.github.com> +wwoodsTM <104587230+wwoodsTM@users.noreply.github.com> wzy <32936898+Freed-Wu@users.noreply.github.com> xaedes xaedes +xctan xloem <0xloem@gmail.com> yangli2 yuiseki +yuri@FreeBSD zakkor zhangkaihuo +zhentaoyu zhouwg <6889919+zhouwg@users.noreply.github.com> zhouwg zrm Ștefan-Gabriel Muscalu +杨朱 · Kiki 源文雨 <41315874+fumiama@users.noreply.github.com> +蕭澧邦 <45505768+shou692199@users.noreply.github.com> Нияз Гарифзянов <112617865+garrnizon@users.noreply.github.com> diff --git a/CMakeLists.txt b/CMakeLists.txt index ef0932a7b..a717a508f 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -46,6 +46,11 @@ if (WIN32) add_compile_definitions(_CRT_SECURE_NO_WARNINGS) endif() +if (MSVC) + add_compile_options("$<$:/utf-8>") + add_compile_options("$<$:/utf-8>") +endif() + # # option list # @@ -75,6 +80,7 @@ option(LLAMA_CURL "llama: use libcurl to download model from an URL" OFF) # Required for relocatable CMake package include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info.cmake) +include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/common.cmake) # override ggml options set(GGML_SANITIZE_THREAD ${LLAMA_SANITIZE_THREAD}) @@ -88,10 +94,6 @@ if (NOT DEFINED GGML_LLAMAFILE) set(GGML_LLAMAFILE_DEFAULT ON) endif() -if (NOT DEFINED GGML_AMX) - set(GGML_AMX ON) -endif() - if (NOT DEFINED GGML_CUDA_GRAPHS) set(GGML_CUDA_GRAPHS_DEFAULT ON) endif() @@ -140,7 +142,6 @@ set(LLAMA_INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR} CACHE PATH "Location o set(LLAMA_LIB_INSTALL_DIR ${CMAKE_INSTALL_LIBDIR} CACHE PATH "Location of library files") set(LLAMA_BIN_INSTALL_DIR ${CMAKE_INSTALL_BINDIR} CACHE PATH "Location of binary files") - # At the moment some compile definitions are placed within the ggml/src # directory but not exported on the `ggml` target. This could be improved by # determining _precisely_ which defines are necessary for the llama-config @@ -157,8 +158,11 @@ if (GGML_TARGET_DEFINES) list(APPEND GGML_TRANSIENT_DEFINES ${GGML_TARGET_DEFINES}) endif() get_target_property(GGML_LINK_LIBRARIES ggml LINK_LIBRARIES) - -set_target_properties(llama PROPERTIES PUBLIC_HEADER ${CMAKE_CURRENT_SOURCE_DIR}/include/llama.h) +# all public headers +set(LLAMA_PUBLIC_HEADERS + ${CMAKE_CURRENT_SOURCE_DIR}/include/llama.h + ${CMAKE_CURRENT_SOURCE_DIR}/include/llama-cpp.h) +set_target_properties(llama PROPERTIES PUBLIC_HEADER "${LLAMA_PUBLIC_HEADERS}") install(TARGETS llama LIBRARY PUBLIC_HEADER) configure_package_config_file( diff --git a/CMakePresets.json b/CMakePresets.json index ae45d60af..13bdd7907 100644 --- a/CMakePresets.json +++ b/CMakePresets.json @@ -24,11 +24,19 @@ "CMAKE_INSTALL_RPATH": "$ORIGIN;$ORIGIN/.." } }, - { "name": "debug", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Debug" } }, - { "name": "release", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Release" } }, - { "name": "reldbg", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "RelWithDebInfo" } }, - { "name": "static", "hidden": true, "cacheVariables": { "GGML_STATIC": "ON" } }, - { "name": "sycl_f16", "hidden": true, "cacheVariables": { "GGML_SYCL_F16": "ON" } }, + { "name": "debug", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Debug" } }, + { "name": "release", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "Release" } }, + { "name": "reldbg", "hidden": true, "cacheVariables": { "CMAKE_BUILD_TYPE": "RelWithDebInfo" } }, + { "name": "static", "hidden": true, "cacheVariables": { "GGML_STATIC": "ON" } }, + { "name": "sycl_f16", "hidden": true, "cacheVariables": { "GGML_SYCL_F16": "ON" } }, + { "name": "vulkan", "hidden": true, "cacheVariables": { "GGML_VULKAN": "ON" } }, + + { + "name": "x64-windows-llvm", "hidden": true, + "cacheVariables": { + "CMAKE_TOOLCHAIN_FILE": "${sourceDir}/cmake/x64-windows-llvm.cmake" + } + }, { "name": "arm64-windows-msvc", "hidden": true, @@ -57,25 +65,33 @@ } }, - { "name": "arm64-windows-llvm-debug" , "inherits": [ "base", "arm64-windows-llvm", "debug" ] }, - { "name": "arm64-windows-llvm-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg" ] }, - { "name": "arm64-windows-llvm+static-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg", "static" ] }, + { "name": "arm64-windows-llvm-debug", "inherits": [ "base", "arm64-windows-llvm", "debug" ] }, + { "name": "arm64-windows-llvm-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg" ] }, + { "name": "arm64-windows-llvm+static-release", "inherits": [ "base", "arm64-windows-llvm", "reldbg", "static" ] }, - { "name": "arm64-apple-clang-debug" , "inherits": [ "base", "arm64-apple-clang", "debug" ] }, - { "name": "arm64-apple-clang-release" , "inherits": [ "base", "arm64-apple-clang", "reldbg" ] }, - { "name": "arm64-apple-clang+static-release" , "inherits": [ "base", "arm64-apple-clang", "reldbg", "static" ] }, + { "name": "arm64-apple-clang-debug", "inherits": [ "base", "arm64-apple-clang", "debug" ] }, + { "name": "arm64-apple-clang-release", "inherits": [ "base", "arm64-apple-clang", "reldbg" ] }, + { "name": "arm64-apple-clang+static-release", "inherits": [ "base", "arm64-apple-clang", "reldbg", "static" ] }, - { "name": "arm64-windows-msvc-debug" , "inherits": [ "base", "arm64-windows-msvc", "debug" ] }, + { "name": "arm64-windows-msvc-debug", "inherits": [ "base", "arm64-windows-msvc", "debug" ] }, { "name": "arm64-windows-msvc-release", "inherits": [ "base", "arm64-windows-msvc", "reldbg" ] }, { "name": "arm64-windows-msvc+static-release", "inherits": [ "base", "arm64-windows-msvc", "reldbg", "static" ] }, - { "name": "x64-windows-msvc-debug" , "inherits": [ "base", "debug" ] }, + { "name": "x64-windows-llvm-debug", "inherits": [ "base", "x64-windows-llvm", "debug" ] }, + { "name": "x64-windows-llvm-release", "inherits": [ "base", "x64-windows-llvm", "release" ] }, + { "name": "x64-windows-llvm-reldbg", "inherits": [ "base", "x64-windows-llvm", "reldbg" ] }, + { "name": "x64-windows-llvm+static-release", "inherits": [ "base", "x64-windows-llvm", "reldbg", "static" ] }, + + { "name": "x64-windows-msvc-debug", "inherits": [ "base", "debug" ] }, { "name": "x64-windows-msvc-release", "inherits": [ "base", "reldbg" ] }, { "name": "x64-windows-msvc+static-release", "inherits": [ "base", "reldbg", "static" ] }, - { "name": "x64-windows-sycl-debug" , "inherits": [ "sycl-base", "debug" ] }, + { "name": "x64-windows-sycl-debug", "inherits": [ "sycl-base", "debug" ] }, { "name": "x64-windows-sycl-debug-f16", "inherits": [ "sycl-base", "debug", "sycl_f16" ] }, { "name": "x64-windows-sycl-release", "inherits": [ "sycl-base", "release" ] }, - { "name": "x64-windows-sycl-release-f16", "inherits": [ "sycl-base", "release", "sycl_f16" ] } + { "name": "x64-windows-sycl-release-f16", "inherits": [ "sycl-base", "release", "sycl_f16" ] }, + + { "name": "x64-windows-vulkan-debug", "inherits": [ "base", "vulkan", "debug" ] }, + { "name": "x64-windows-vulkan-release", "inherits": [ "base", "vulkan", "release" ] } ] } diff --git a/CODEOWNERS b/CODEOWNERS new file mode 100644 index 000000000..72d594b46 --- /dev/null +++ b/CODEOWNERS @@ -0,0 +1,11 @@ +# collaborators can optionally add themselves here to indicate their availability for reviewing related PRs + +/ci/ @ggerganov +/.devops/*.Dockerfile @ngxson +/examples/server/ @ngxson +/ggml/src/ggml-cuda/fattn* @JohannesGaessler +/ggml/src/ggml-cuda/mmq.* @JohannesGaessler +/ggml/src/ggml-cuda/mmv.* @JohannesGaessler +/ggml/src/ggml-cuda/mmvq.* @JohannesGaessler +/ggml/src/ggml-opt.cpp @JohannesGaessler +/ggml/src/gguf.cpp @JohannesGaessler diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 4c882c254..8d411982b 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -1,9 +1,10 @@ # Pull requests (for contributors) - Test your changes: - - Using the commands in the [`tests`](tests) folder. For instance, running the `./tests/test-backend-ops` command tests different backend implementations of the `ggml` library - - Execute [the full CI locally on your machine](ci/README.md) before publishing -- Optionally rate the complexity of your PR (i.e. `Review Complexity : Low`, `Review Complexity : Medium`, `Review Complexity : High`). This makes it easier for maintainers to triage the PRs + - Execute [the full CI locally on your machine](ci/README.md) before publishing + - Verify that the perplexity and the performance are not affected negatively by your changes (use `llama-perplexity` and `llama-bench`) + - If you modified the `ggml` source, run the `test-backend-ops` tool to check whether different backend implementations of the `ggml` operators produce consistent results (this requires access to at least two different `ggml` backends) + - If you modified a `ggml` operator or added a new one, add the corresponding test cases to `test-backend-ops` - Consider allowing write access to your branch for faster reviews, as reviewers can push commits directly - If your PR becomes stale, don't hesitate to ping the maintainers in the comments @@ -12,20 +13,111 @@ - Squash-merge PRs - Use the following format for the squashed commit title: ` : (#)`. For example: `utils : fix typo in utils.py (#1234)` - Optionally pick a `` from here: https://github.com/ggerganov/llama.cpp/wiki/Modules +- Consider adding yourself to [CODEOWNERS](CODEOWNERS) # Coding guidelines - Avoid adding third-party dependencies, extra files, extra headers, etc. - Always consider cross-compatibility with other operating systems and architectures - Avoid fancy-looking modern STL constructs, use basic `for` loops, avoid templates, keep it simple -- There are no strict rules for the code style, but try to follow the patterns in the code (indentation, spaces, etc.). Vertical alignment makes things more readable and easier to batch edit +- Vertical alignment makes things more readable and easier to batch edit - Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a` -- Naming usually optimizes for common prefix (see https://github.com/ggerganov/ggml/pull/302#discussion_r1243240963) +- Use sized integer types such as `int32_t` in the public API, e.g. `size_t` may also be appropriate for allocation sizes or byte offsets +- Declare structs with `struct foo {}` instead of `typedef struct foo {} foo` + - In C++ code omit optional `struct` and `enum` keyword whenever they are not necessary + ```cpp + // OK + llama_context * ctx; + const llama_rope_type rope_type; + + // not OK + struct llama_context * ctx; + const enum llama_rope_type rope_type; + ``` + + _(NOTE: this guideline is yet to be applied to the `llama.cpp` codebase. New code should follow this guideline.)_ + +- Try to follow the existing patterns in the code (indentation, spaces, etc.). In case of doubt use `clang-format` to format the added code +- For anything not covered in the current guidelines, refer to the [C++ Core Guidelines](https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines) - Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices - Matrix multiplication is unconventional: [`C = ggml_mul_mat(ctx, A, B)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means $C^T = A B^T \Leftrightarrow C = B A^T.$ ![matmul](media/matmul.png) +# Naming guidelines + +- Use `snake_case` for function, variable and type names +- Naming usually optimizes for longest common prefix (see https://github.com/ggerganov/ggml/pull/302#discussion_r1243240963) + + ```cpp + // not OK + int small_number; + int big_number; + + // OK + int number_small; + int number_big; + ``` + +- Enum values are always in upper case and prefixed with the enum name + + ```cpp + enum llama_vocab_type { + LLAMA_VOCAB_TYPE_NONE = 0, + LLAMA_VOCAB_TYPE_SPM = 1, + LLAMA_VOCAB_TYPE_BPE = 2, + LLAMA_VOCAB_TYPE_WPM = 3, + LLAMA_VOCAB_TYPE_UGM = 4, + LLAMA_VOCAB_TYPE_RWKV = 5, + }; + ``` + +- The general naming pattern is `_`, with `` being `_` + + ```cpp + llama_model_init(); // class: "llama_model", method: "init" + llama_sampler_chain_remove(); // class: "llama_sampler_chain", method: "remove" + llama_sampler_get_seed(); // class: "llama_sampler", method: "get_seed" + llama_set_embeddings(); // class: "llama_context", method: "set_embeddings" + llama_n_threads(); // class: "llama_context", method: "n_threads" + llama_adapter_lora_free(); // class: "llama_adapter_lora", method: "free" + ``` + + - The `get` `` can be omitted + - The `` can be omitted if not necessary + - The `_context` suffix of the `` is optional. Use it to disambiguate symbols when needed + - Use `init`/`free` for constructor/destructor `` + +- Use the `_t` suffix when a type is supposed to be opaque to the user - it's not relevant to them if it is a struct or anything else + + ```cpp + typedef struct llama_context * llama_context_t; + + enum llama_pooling_type llama_pooling_type(const llama_context_t ctx); + ``` + + _(NOTE: this guideline is yet to be applied to the `llama.cpp` codebase. New code should follow this guideline)_ + +- C/C++ filenames are all lowercase with dashes. Headers use the `.h` extension. Source files use the `.c` or `.cpp` extension +- Python filenames are all lowercase with underscores + +- _(TODO: abbreviations usage)_ + +# Preprocessor directives + +- _(TODO: add guidelines with examples and apply them to the codebase)_ + + ```cpp + #ifdef FOO + #endif // FOO + ``` + +# Documentation + +- Documentation is a community effort +- When you need to look into the source code to figure out how to use an API consider adding a short summary to the header file for future reference +- When you notice incorrect or outdated documentation, please update it + # Resources The Github issues, PRs and discussions contain a lot of information that can be useful to get familiar with the codebase. For convenience, some of the more important information is referenced from Github projects: diff --git a/Makefile b/Makefile index b9131eae5..19ae0d5f1 100644 --- a/Makefile +++ b/Makefile @@ -1,3 +1,7 @@ +ifndef LLAMA_MAKEFILE +$(error The Makefile build is deprecated. Use the CMake build instead. For more details, see https://github.com/ggerganov/llama.cpp/blob/master/docs/build.md) +endif + # Define the default target now so that it is always the first target BUILD_TARGETS = \ libllava.a \ @@ -18,6 +22,7 @@ BUILD_TARGETS = \ llama-infill \ llama-llava-cli \ llama-minicpmv-cli\ + llama-qwen2vl-cli\ llama-lookahead \ llama-lookup \ llama-lookup-create \ @@ -34,6 +39,7 @@ BUILD_TARGETS = \ llama-server \ llama-simple \ llama-simple-chat \ + llama-run \ llama-speculative \ llama-tokenize \ llama-vdot \ @@ -48,7 +54,6 @@ TEST_TARGETS = \ tests/test-backend-ops \ tests/test-chat-template \ tests/test-double-float \ - tests/test-grad0 \ tests/test-grammar-integration \ tests/test-grammar-parser \ tests/test-json-schema-to-grammar \ @@ -251,11 +256,11 @@ endif # Compile flags # -# keep standard at C11 and C++11 -MK_CPPFLAGS = -Iggml/include -Iggml/src -Iinclude -Isrc -Icommon +# keep standard at C11 and C++17 +MK_CPPFLAGS = -Iggml/include -Iggml/src -Iinclude -Isrc -Icommon -DGGML_USE_CPU MK_CFLAGS = -std=c11 -fPIC -MK_CXXFLAGS = -std=c++11 -fPIC -MK_NVCCFLAGS = -std=c++11 +MK_CXXFLAGS = -std=c++17 -fPIC +MK_NVCCFLAGS = -std=c++17 ifdef LLAMA_NO_CCACHE GGML_NO_CCACHE := 1 @@ -291,6 +296,7 @@ endif # some memory allocation are available on Linux through GNU extensions in libc ifeq ($(UNAME_S),Linux) MK_CPPFLAGS += -D_GNU_SOURCE + MK_LDFLAGS += -ldl endif # RLIMIT_MEMLOCK came in BSD, is not specified in POSIX.1, @@ -359,6 +365,10 @@ ifdef LLAMA_SERVER_SSL MK_LDFLAGS += -lssl -lcrypto endif +ifndef GGML_NO_CPU_AARCH64 + MK_CPPFLAGS += -DGGML_USE_CPU_AARCH64 +endif + # warnings WARN_FLAGS = \ -Wall \ @@ -436,6 +446,10 @@ ifeq ($(UNAME_M),$(filter $(UNAME_M),x86_64 i686 amd64)) MK_CFLAGS += -march=native -mtune=native HOST_CXXFLAGS += -march=native -mtune=native + # Usage AMX build test + #MK_CFLAGS += -march=graniterapids -mtune=graniterapids + #HOST_CXXFLAGS += -march=graniterapids -mtune=graniterapids + # Usage AVX-only #MK_CFLAGS += -mfma -mf16c -mavx #MK_CXXFLAGS += -mfma -mf16c -mavx @@ -523,70 +537,62 @@ ifndef GGML_NO_ACCELERATE # Mac OS - include Accelerate framework. # `-framework Accelerate` works both with Apple Silicon and Mac Intel ifeq ($(UNAME_S),Darwin) - MK_CPPFLAGS += -DGGML_USE_ACCELERATE -DGGML_USE_BLAS - MK_CPPFLAGS += -DACCELERATE_NEW_LAPACK - MK_CPPFLAGS += -DACCELERATE_LAPACK_ILP64 - MK_LDFLAGS += -framework Accelerate - OBJ_GGML += ggml/src/ggml-blas.o + MK_CPPFLAGS += -DGGML_USE_ACCELERATE -DGGML_USE_BLAS -DGGML_BLAS_USE_ACCELERATE + MK_CPPFLAGS += -DACCELERATE_NEW_LAPACK + MK_CPPFLAGS += -DACCELERATE_LAPACK_ILP64 + MK_LDFLAGS += -framework Accelerate + OBJ_GGML_EXT += ggml/src/ggml-blas/ggml-blas.o endif endif # GGML_NO_ACCELERATE -ifdef GGML_MUSA - CC := clang - CXX := clang++ - GGML_CUDA := 1 - MK_CPPFLAGS += -DGGML_USE_MUSA -endif - ifndef GGML_NO_OPENMP MK_CPPFLAGS += -DGGML_USE_OPENMP MK_CFLAGS += -fopenmp MK_CXXFLAGS += -fopenmp - ifdef GGML_MUSA - MK_CPPFLAGS += -I/usr/lib/llvm-10/include/openmp - MK_LDFLAGS += -L/usr/lib/llvm-10/lib - endif # GGML_MUSA endif # GGML_NO_OPENMP ifdef GGML_OPENBLAS - MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas) - MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas) - MK_LDFLAGS += $(shell pkg-config --libs openblas) - OBJ_GGML += ggml/src/ggml-blas.o + MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas) + MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas) + MK_LDFLAGS += $(shell pkg-config --libs openblas) + OBJ_GGML_EXT += ggml/src/ggml-blas/ggml-blas.o endif # GGML_OPENBLAS ifdef GGML_OPENBLAS64 - MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas64) - MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas64) - MK_LDFLAGS += $(shell pkg-config --libs openblas64) - OBJ_GGML += ggml/src/ggml-blas.o + MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas64) + MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas64) + MK_LDFLAGS += $(shell pkg-config --libs openblas64) + OBJ_GGML_EXT += ggml/src/ggml-blas/ggml-blas.o endif # GGML_OPENBLAS64 ifdef GGML_BLIS - MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_BLIS -I/usr/local/include/blis -I/usr/include/blis - MK_LDFLAGS += -lblis -L/usr/local/lib - OBJ_GGML += ggml/src/ggml-blas.o + MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_BLIS -I/usr/local/include/blis -I/usr/include/blis + MK_LDFLAGS += -lblis -L/usr/local/lib + OBJ_GGML_EXT += ggml/src/ggml-blas/ggml-blas.o endif # GGML_BLIS ifdef GGML_NVPL - MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_NVPL -DNVPL_ILP64 -I/usr/local/include/nvpl_blas -I/usr/include/nvpl_blas - MK_LDFLAGS += -L/usr/local/lib -lnvpl_blas_core -lnvpl_blas_ilp64_gomp - OBJ_GGML += ggml/src/ggml-blas.o + MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_NVPL -DNVPL_ILP64 -I/usr/local/include/nvpl_blas -I/usr/include/nvpl_blas + MK_LDFLAGS += -L/usr/local/lib -lnvpl_blas_core -lnvpl_blas_ilp64_gomp + OBJ_GGML_EXT += ggml/src/ggml-blas/ggml-blas.o endif # GGML_NVPL ifndef GGML_NO_LLAMAFILE - MK_CPPFLAGS += -DGGML_USE_LLAMAFILE - OBJ_GGML += ggml/src/llamafile/sgemm.o + MK_CPPFLAGS += -DGGML_USE_LLAMAFILE + OBJ_GGML_EXT += ggml/src/ggml-cpu/llamafile/sgemm.o endif ifndef GGML_NO_AMX MK_CPPFLAGS += -DGGML_USE_AMX - OBJ_GGML += ggml/src/ggml-amx.o ggml/src/ggml-amx/mmq.o + OBJ_GGML_EXT += ggml/src/ggml-cpu/amx/amx.o ggml/src/ggml-cpu/amx/mmq.o endif +# only necessary for the CPU backend files +MK_CPPFLAGS += -Iggml/src/ggml-cpu + ifdef GGML_RPC - MK_CPPFLAGS += -DGGML_USE_RPC - OBJ_GGML += ggml/src/ggml-rpc.o + MK_CPPFLAGS += -DGGML_USE_RPC + OBJ_GGML_EXT += ggml/src/ggml-rpc.o endif # GGML_RPC OBJ_CUDA_TMPL = $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-wmma*.cu)) @@ -601,41 +607,27 @@ else endif # GGML_CUDA_FA_ALL_QUANTS ifdef GGML_CUDA - ifdef GGML_MUSA - ifneq ('', '$(wildcard /opt/musa)') - CUDA_PATH ?= /opt/musa - else - CUDA_PATH ?= /usr/local/musa - endif - - MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include - MK_LDFLAGS += -lmusa -lmublas -lmusart -lpthread -ldl -lrt -L$(CUDA_PATH)/lib -L/usr/lib64 - MK_NVCCFLAGS += -x musa -mtgpu --cuda-gpu-arch=mp_21 --cuda-gpu-arch=mp_22 + ifneq ('', '$(wildcard /opt/cuda)') + CUDA_PATH ?= /opt/cuda else - ifneq ('', '$(wildcard /opt/cuda)') - CUDA_PATH ?= /opt/cuda - else - CUDA_PATH ?= /usr/local/cuda - endif + CUDA_PATH ?= /usr/local/cuda + endif - MK_CPPFLAGS += -DGGML_USE_CUDA -DGGML_CUDA_USE_GRAPHS -I$(CUDA_PATH)/include -I$(CUDA_PATH)/targets/$(UNAME_M)-linux/include - MK_LDFLAGS += -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L$(CUDA_PATH)/lib64 -L/usr/lib64 -L$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib -L$(CUDA_PATH)/lib64/stubs -L/usr/lib/wsl/lib - MK_NVCCFLAGS += -use_fast_math - endif # GGML_MUSA + MK_CPPFLAGS += -DGGML_USE_CUDA -DGGML_CUDA_USE_GRAPHS -I$(CUDA_PATH)/include -I$(CUDA_PATH)/targets/$(UNAME_M)-linux/include + MK_LDFLAGS += -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L$(CUDA_PATH)/lib64 -L/usr/lib64 -L$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib -L$(CUDA_PATH)/lib64/stubs -L/usr/lib/wsl/lib + MK_NVCCFLAGS += -use_fast_math - OBJ_GGML += ggml/src/ggml-cuda.o - OBJ_GGML += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) - OBJ_GGML += $(OBJ_CUDA_TMPL) + OBJ_GGML_EXT += ggml/src/ggml-cuda/ggml-cuda.o + OBJ_GGML_EXT += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) + OBJ_GGML_EXT += $(OBJ_CUDA_TMPL) ifdef LLAMA_FATAL_WARNINGS MK_NVCCFLAGS += -Werror all-warnings endif # LLAMA_FATAL_WARNINGS -ifndef GGML_MUSA ifndef JETSON_EOL_MODULE_DETECT MK_NVCCFLAGS += --forward-unknown-to-host-compiler endif # JETSON_EOL_MODULE_DETECT -endif # GGML_MUSA ifdef LLAMA_DEBUG MK_NVCCFLAGS += -lineinfo @@ -648,11 +640,7 @@ endif # GGML_CUDA_DEBUG ifdef GGML_CUDA_NVCC NVCC = $(CCACHE) $(GGML_CUDA_NVCC) else - ifdef GGML_MUSA - NVCC = $(CCACHE) mcc - else - NVCC = $(CCACHE) nvcc - endif # GGML_MUSA + NVCC = $(CCACHE) nvcc endif # GGML_CUDA_NVCC ifdef CUDA_DOCKER_ARCH @@ -661,10 +649,6 @@ else ifndef CUDA_POWER_ARCH MK_NVCCFLAGS += -arch=native endif # CUDA_DOCKER_ARCH -ifdef GGML_CUDA_FORCE_DMMV - MK_NVCCFLAGS += -DGGML_CUDA_FORCE_DMMV -endif # GGML_CUDA_FORCE_DMMV - ifdef GGML_CUDA_FORCE_MMQ MK_NVCCFLAGS += -DGGML_CUDA_FORCE_MMQ endif # GGML_CUDA_FORCE_MMQ @@ -673,20 +657,6 @@ ifdef GGML_CUDA_FORCE_CUBLAS MK_NVCCFLAGS += -DGGML_CUDA_FORCE_CUBLAS endif # GGML_CUDA_FORCE_CUBLAS -ifdef GGML_CUDA_DMMV_X - MK_NVCCFLAGS += -DGGML_CUDA_DMMV_X=$(GGML_CUDA_DMMV_X) -else - MK_NVCCFLAGS += -DGGML_CUDA_DMMV_X=32 -endif # GGML_CUDA_DMMV_X - -ifdef GGML_CUDA_MMV_Y - MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_MMV_Y) -else ifdef GGML_CUDA_DMMV_Y - MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_DMMV_Y) # for backwards compatibility -else - MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=1 -endif # GGML_CUDA_MMV_Y - ifdef GGML_CUDA_F16 MK_NVCCFLAGS += -DGGML_CUDA_F16 endif # GGML_CUDA_F16 @@ -695,12 +665,6 @@ ifdef GGML_CUDA_DMMV_F16 MK_NVCCFLAGS += -DGGML_CUDA_F16 endif # GGML_CUDA_DMMV_F16 -ifdef GGML_CUDA_KQUANTS_ITER - MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER) -else - MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2 -endif - ifdef GGML_CUDA_PEER_MAX_BATCH_SIZE MK_NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=$(GGML_CUDA_PEER_MAX_BATCH_SIZE) else @@ -724,15 +688,9 @@ define NVCC_COMPILE $(NVCC) -I. -Icommon -D_XOPEN_SOURCE=600 -D_GNU_SOURCE -DNDEBUG -DGGML_USE_CUDA -I/usr/local/cuda/include -I/opt/cuda/include -I/usr/local/cuda/targets/aarch64-linux/include -std=c++11 -O3 $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@ endef # NVCC_COMPILE else - ifdef GGML_MUSA -define NVCC_COMPILE - $(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -c $< -o $@ -endef # NVCC_COMPILE - else define NVCC_COMPILE $(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@ endef # NVCC_COMPILE - endif # GGML_MUSA endif # JETSON_EOL_MODULE_DETECT ggml/src/ggml-cuda/%.o: \ @@ -742,8 +700,8 @@ ggml/src/ggml-cuda/%.o: \ ggml/src/ggml-cuda/common.cuh $(NVCC_COMPILE) -ggml/src/ggml-cuda.o: \ - ggml/src/ggml-cuda.cu \ +ggml/src/ggml-cuda/ggml-cuda.o: \ + ggml/src/ggml-cuda/ggml-cuda.cu \ ggml/include/ggml-cuda.h \ ggml/include/ggml.h \ ggml/include/ggml-backend.h \ @@ -754,9 +712,9 @@ ggml/src/ggml-cuda.o: \ endif # GGML_CUDA ifdef GGML_VULKAN - MK_CPPFLAGS += -DGGML_USE_VULKAN - MK_LDFLAGS += $(shell pkg-config --libs vulkan) - OBJ_GGML += ggml/src/ggml-vulkan.o ggml/src/ggml-vulkan-shaders.o + MK_CPPFLAGS += -DGGML_USE_VULKAN + MK_LDFLAGS += $(shell pkg-config --libs vulkan) + OBJ_GGML_EXT += ggml/src/ggml-vulkan.o ggml/src/ggml-vulkan-shaders.o ifdef GGML_VULKAN_CHECK_RESULTS MK_CPPFLAGS += -DGGML_VULKAN_CHECK_RESULTS @@ -786,10 +744,10 @@ GLSLC_CMD = glslc _ggml_vk_genshaders_cmd = $(shell pwd)/vulkan-shaders-gen _ggml_vk_header = ggml/src/ggml-vulkan-shaders.hpp _ggml_vk_source = ggml/src/ggml-vulkan-shaders.cpp -_ggml_vk_input_dir = ggml/src/vulkan-shaders +_ggml_vk_input_dir = ggml/src/ggml-vulkan/vulkan-shaders _ggml_vk_shader_deps = $(echo $(_ggml_vk_input_dir)/*.comp) -ggml/src/ggml-vulkan.o: ggml/src/ggml-vulkan.cpp ggml/include/ggml-vulkan.h $(_ggml_vk_header) $(_ggml_vk_source) +ggml/src/ggml-vulkan.o: ggml/src/ggml-vulkan/ggml-vulkan.cpp ggml/include/ggml-vulkan.h $(_ggml_vk_header) $(_ggml_vk_source) $(CXX) $(CXXFLAGS) $(shell pkg-config --cflags vulkan) -c $< -o $@ $(_ggml_vk_header): $(_ggml_vk_source) @@ -801,12 +759,12 @@ $(_ggml_vk_source): $(_ggml_vk_shader_deps) vulkan-shaders-gen --target-hpp $(_ggml_vk_header) \ --target-cpp $(_ggml_vk_source) -vulkan-shaders-gen: ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp - $(CXX) $(CXXFLAGS) -o $@ $(LDFLAGS) ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp +vulkan-shaders-gen: ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp + $(CXX) $(CXXFLAGS) -o $@ $(LDFLAGS) ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp endif # GGML_VULKAN -ifdef GGML_HIPBLAS +ifdef GGML_HIP ifeq ($(wildcard /opt/rocm),) ROCM_PATH ?= /usr AMDGPU_TARGETS ?= $(shell $(shell which amdgpu-arch)) @@ -815,11 +773,7 @@ ifdef GGML_HIPBLAS AMDGPU_TARGETS ?= $(shell $(ROCM_PATH)/llvm/bin/amdgpu-arch) endif - GGML_CUDA_DMMV_X ?= 32 - GGML_CUDA_MMV_Y ?= 1 - GGML_CUDA_KQUANTS_ITER ?= 2 - - MK_CPPFLAGS += -DGGML_USE_HIPBLAS -DGGML_USE_CUDA + MK_CPPFLAGS += -DGGML_USE_HIP -DGGML_USE_CUDA ifdef GGML_HIP_UMA MK_CPPFLAGS += -DGGML_HIP_UMA @@ -832,13 +786,6 @@ endif # GGML_HIP_UMA HIPCC ?= $(CCACHE) $(ROCM_PATH)/bin/hipcc HIPFLAGS += $(addprefix --offload-arch=,$(AMDGPU_TARGETS)) - HIPFLAGS += -DGGML_CUDA_DMMV_X=$(GGML_CUDA_DMMV_X) - HIPFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_MMV_Y) - HIPFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER) - -ifdef GGML_CUDA_FORCE_DMMV - HIPFLAGS += -DGGML_CUDA_FORCE_DMMV -endif # GGML_CUDA_FORCE_DMMV ifdef GGML_CUDA_FORCE_MMQ HIPFLAGS += -DGGML_CUDA_FORCE_MMQ @@ -852,12 +799,12 @@ ifdef GGML_CUDA_NO_PEER_COPY HIPFLAGS += -DGGML_CUDA_NO_PEER_COPY endif # GGML_CUDA_NO_PEER_COPY - OBJ_GGML += ggml/src/ggml-cuda.o - OBJ_GGML += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) - OBJ_GGML += $(OBJ_CUDA_TMPL) + OBJ_GGML_EXT += ggml/src/ggml-cuda/ggml-cuda.o + OBJ_GGML_EXT += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) + OBJ_GGML_EXT += $(OBJ_CUDA_TMPL) -ggml/src/ggml-cuda.o: \ - ggml/src/ggml-cuda.cu \ +ggml/src/ggml-cuda/ggml-cuda.o: \ + ggml/src/ggml-cuda/ggml-cuda.cu \ ggml/include/ggml-cuda.h \ ggml/include/ggml.h \ ggml/include/ggml-backend.h \ @@ -872,72 +819,172 @@ ggml/src/ggml-cuda/%.o: \ ggml/src/ggml-common.h \ ggml/src/ggml-cuda/common.cuh $(HIPCC) $(CXXFLAGS) $(HIPFLAGS) -x hip -c -o $@ $< -endif # GGML_HIPBLAS +endif # GGML_HIP + +ifdef GGML_MUSA + ifeq ($(wildcard /opt/musa),) + MUSA_PATH ?= /usr/local/musa + else + MUSA_PATH ?= /opt/musa + endif + MUSA_ARCHITECTURES ?= 21;22 + + MK_CPPFLAGS += -DGGML_USE_MUSA -DGGML_USE_CUDA + MK_LDFLAGS += -L$(MUSA_PATH)/lib -Wl,-rpath=$(MUSA_PATH)/lib + MK_LDFLAGS += -lmusa -lmusart -lmublas + + ifndef GGML_NO_OPENMP + # For Ubuntu Focal + MK_CPPFLAGS += -I/usr/lib/llvm-10/include/openmp + MK_LDFLAGS += -L/usr/lib/llvm-10/lib + # For Ubuntu Jammy + MK_CPPFLAGS += -I/usr/lib/llvm-14/lib/clang/14.0.0/include + MK_LDFLAGS += -L/usr/lib/llvm-14/lib + endif # GGML_NO_OPENMP + + CC := $(MUSA_PATH)/bin/clang + CXX := $(MUSA_PATH)/bin/clang++ + MCC := $(CCACHE) $(MUSA_PATH)/bin/mcc + + MUSAFLAGS = -x musa -mtgpu + MUSAFLAGS += $(foreach arch,$(subst ;, ,$(MUSA_ARCHITECTURES)),--cuda-gpu-arch=mp_$(arch)) + +ifdef GGML_CUDA_FORCE_MMQ + MUSAFLAGS += -DGGML_CUDA_FORCE_MMQ +endif # GGML_CUDA_FORCE_MMQ + +ifdef GGML_CUDA_FORCE_CUBLAS + MUSAFLAGS += -DGGML_CUDA_FORCE_CUBLAS +endif # GGML_CUDA_FORCE_CUBLAS + +ifdef GGML_CUDA_F16 + MUSAFLAGS += -DGGML_CUDA_F16 +endif # GGML_CUDA_F16 + +ifdef GGML_CUDA_DMMV_F16 + MUSAFLAGS += -DGGML_CUDA_F16 +endif # GGML_CUDA_DMMV_F16 + +ifdef GGML_CUDA_PEER_MAX_BATCH_SIZE + MUSAFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=$(GGML_CUDA_PEER_MAX_BATCH_SIZE) +else + MUSAFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=128 +endif # GGML_CUDA_PEER_MAX_BATCH_SIZE + +ifdef GGML_CUDA_NO_PEER_COPY + MUSAFLAGS += -DGGML_CUDA_NO_PEER_COPY +endif # GGML_CUDA_NO_PEER_COPY + +ifdef GGML_CUDA_FA_ALL_QUANTS + MUSAFLAGS += -DGGML_CUDA_FA_ALL_QUANTS +endif # GGML_CUDA_FA_ALL_QUANTS + + OBJ_GGML_EXT += ggml/src/ggml-cuda/ggml-cuda.o + OBJ_GGML_EXT += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) + OBJ_GGML_EXT += $(OBJ_CUDA_TMPL) + +ggml/src/ggml-cuda/ggml-cuda.o: \ + ggml/src/ggml-cuda/ggml-cuda.cu \ + ggml/include/ggml-cuda.h \ + ggml/include/ggml.h \ + ggml/include/ggml-backend.h \ + ggml/src/ggml-backend-impl.h \ + ggml/src/ggml-common.h \ + $(wildcard ggml/src/ggml-cuda/*.cuh) + $(MCC) $(CXXFLAGS) $(MUSAFLAGS) -c -o $@ $< + +ggml/src/ggml-cuda/%.o: \ + ggml/src/ggml-cuda/%.cu \ + ggml/include/ggml.h \ + ggml/src/ggml-common.h \ + ggml/src/ggml-cuda/common.cuh + $(MCC) $(CXXFLAGS) $(MUSAFLAGS) -c -o $@ $< +endif # GGML_MUSA ifdef GGML_METAL - MK_CPPFLAGS += -DGGML_USE_METAL - MK_LDFLAGS += -framework Foundation -framework Metal -framework MetalKit - OBJ_GGML += ggml/src/ggml-metal.o + MK_CPPFLAGS += -DGGML_USE_METAL + MK_LDFLAGS += -framework Foundation -framework Metal -framework MetalKit + OBJ_GGML_EXT += ggml/src/ggml-metal/ggml-metal.o + +ifdef GGML_METAL_USE_BF16 + MK_CPPFLAGS += -DGGML_METAL_USE_BF16 +endif # GGML_METAL_USE_BF16 ifdef GGML_METAL_NDEBUG MK_CPPFLAGS += -DGGML_METAL_NDEBUG endif ifdef GGML_METAL_EMBED_LIBRARY - MK_CPPFLAGS += -DGGML_METAL_EMBED_LIBRARY - OBJ_GGML += ggml/src/ggml-metal-embed.o + MK_CPPFLAGS += -DGGML_METAL_EMBED_LIBRARY + OBJ_GGML_EXT += ggml/src/ggml-metal-embed.o endif endif # GGML_METAL ifdef GGML_METAL -ggml/src/ggml-metal.o: \ - ggml/src/ggml-metal.m \ +ggml/src/ggml-metal/ggml-metal.o: \ + ggml/src/ggml-metal/ggml-metal.m \ + ggml/src/ggml-metal/ggml-metal-impl.h \ ggml/include/ggml-metal.h \ ggml/include/ggml.h $(CC) $(CFLAGS) -c $< -o $@ ifdef GGML_METAL_EMBED_LIBRARY ggml/src/ggml-metal-embed.o: \ - ggml/src/ggml-metal.metal \ + ggml/src/ggml-metal/ggml-metal.metal \ + ggml/src/ggml-metal/ggml-metal-impl.h \ ggml/src/ggml-common.h @echo "Embedding Metal library" - @sed -e '/#include "ggml-common.h"/r ggml/src/ggml-common.h' -e '/#include "ggml-common.h"/d' < ggml/src/ggml-metal.metal > ggml/src/ggml-metal-embed.metal + @sed -e '/__embed_ggml-common.h__/r ggml/src/ggml-common.h' -e '/__embed_ggml-common.h__/d' < ggml/src/ggml-metal/ggml-metal.metal > ggml/src/ggml-metal/ggml-metal-embed.metal.tmp + @sed -e '/#include "ggml-metal-impl.h"/r ggml/src/ggml-metal/ggml-metal-impl.h' -e '/#include "ggml-metal-impl.h"/d' < ggml/src/ggml-metal/ggml-metal-embed.metal.tmp > ggml/src/ggml-metal/ggml-metal-embed.metal $(eval TEMP_ASSEMBLY=$(shell mktemp -d)) - @echo ".section __DATA, __ggml_metallib" > $(TEMP_ASSEMBLY)/ggml-metal-embed.s - @echo ".globl _ggml_metallib_start" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s - @echo "_ggml_metallib_start:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s - @echo ".incbin \"ggml/src/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s - @echo ".globl _ggml_metallib_end" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s - @echo "_ggml_metallib_end:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo ".section __DATA, __ggml_metallib" > $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo ".globl _ggml_metallib_start" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo "_ggml_metallib_start:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo ".incbin \"ggml/src/ggml-metal/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo ".globl _ggml_metallib_end" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s + @echo "_ggml_metallib_end:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s $(CC) $(CFLAGS) -c $(TEMP_ASSEMBLY)/ggml-metal-embed.s -o $@ @rm -f ${TEMP_ASSEMBLY}/ggml-metal-embed.s @rmdir ${TEMP_ASSEMBLY} endif endif # GGML_METAL -OBJ_GGML += \ - ggml/src/ggml.o \ - ggml/src/ggml-cpu.o \ - ggml/src/ggml-alloc.o \ - ggml/src/ggml-backend.o \ - ggml/src/ggml-quants.o \ - ggml/src/ggml-aarch64.o +DIR_GGML = ggml +DIR_LLAMA = src +DIR_COMMON = common + +OBJ_GGML = \ + $(DIR_GGML)/src/ggml.o \ + $(DIR_GGML)/src/ggml-alloc.o \ + $(DIR_GGML)/src/ggml-backend.o \ + $(DIR_GGML)/src/ggml-backend-reg.o \ + $(DIR_GGML)/src/ggml-opt.o \ + $(DIR_GGML)/src/ggml-quants.o \ + $(DIR_GGML)/src/ggml-threading.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu_cpp.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu-aarch64.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu-hbm.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu-quants.o \ + $(DIR_GGML)/src/ggml-cpu/ggml-cpu-traits.o \ + $(OBJ_GGML_EXT) OBJ_LLAMA = \ - src/llama.o \ - src/llama-vocab.o \ - src/llama-grammar.o \ - src/llama-sampling.o \ - src/unicode.o \ - src/unicode-data.o + $(DIR_LLAMA)/llama.o \ + $(DIR_LLAMA)/llama-vocab.o \ + $(DIR_LLAMA)/llama-grammar.o \ + $(DIR_LLAMA)/llama-sampling.o \ + $(DIR_LLAMA)/unicode.o \ + $(DIR_LLAMA)/unicode-data.o OBJ_COMMON = \ - common/common.o \ - common/arg.o \ - common/log.o \ - common/console.o \ - common/ngram-cache.o \ - common/sampling.o \ - common/build-info.o \ - common/json-schema-to-grammar.o + $(DIR_COMMON)/common.o \ + $(DIR_COMMON)/arg.o \ + $(DIR_COMMON)/log.o \ + $(DIR_COMMON)/console.o \ + $(DIR_COMMON)/ngram-cache.o \ + $(DIR_COMMON)/sampling.o \ + $(DIR_COMMON)/speculative.o \ + $(DIR_COMMON)/build-info.o \ + $(DIR_COMMON)/json-schema-to-grammar.o OBJ_ALL = $(OBJ_GGML) $(OBJ_LLAMA) $(OBJ_COMMON) @@ -993,7 +1040,6 @@ $(info I CXX: $(shell $(CXX) --version | head -n 1)) ifdef GGML_CUDA $(info I NVCC: $(shell $(NVCC) --version | tail -n 1)) CUDA_VERSION := $(shell $(NVCC) --version | grep -oP 'release (\K[0-9]+\.[0-9])') -ifndef GGML_MUSA ifeq ($(shell awk -v "v=$(CUDA_VERSION)" 'BEGIN { print (v < 11.7) }'),1) ifndef CUDA_DOCKER_ARCH @@ -1003,7 +1049,6 @@ endif # CUDA_POWER_ARCH endif # CUDA_DOCKER_ARCH endif # eq ($(shell echo "$(CUDA_VERSION) < 11.7" | bc),1) -endif # GGML_MUSA endif # GGML_CUDA $(info ) @@ -1040,224 +1085,78 @@ endif # Build libraries # -# ggml +# Libraries +LIB_GGML = libggml.so +LIB_GGML_S = libggml.a -ggml/src/ggml.o: \ - ggml/src/ggml.c \ - ggml/include/ggml.h - $(CC) $(CFLAGS) -c $< -o $@ +LIB_LLAMA = libllama.so +LIB_LLAMA_S = libllama.a -ggml/src/ggml-cpu.o: \ - ggml/src/ggml-cpu.c \ - ggml/include/ggml.h \ - ggml/src/ggml-common.h - $(CC) $(CFLAGS) -c $< -o $@ +LIB_COMMON = libcommon.so +LIB_COMMON_S = libcommon.a -ggml/src/ggml-alloc.o: \ - ggml/src/ggml-alloc.c \ - ggml/include/ggml.h \ - ggml/include/ggml-alloc.h - $(CC) $(CFLAGS) -c $< -o $@ +# Targets +BUILD_TARGETS += $(LIB_GGML) $(LIB_GGML_S) $(LIB_LLAMA) $(LIB_LLAMA_S) $(LIB_COMMON) $(LIB_COMMON_S) -ggml/src/ggml-backend.o: \ - ggml/src/ggml-backend.cpp \ - ggml/src/ggml-backend-impl.h \ - ggml/include/ggml.h \ - ggml/include/ggml-backend.h - $(CXX) $(CXXFLAGS) -c $< -o $@ +# Dependency files +DEP_FILES = $(OBJ_GGML:.o=.d) $(OBJ_LLAMA:.o=.d) $(OBJ_COMMON:.o=.d) -ggml/src/ggml-quants.o: \ - ggml/src/ggml-quants.c \ - ggml/include/ggml.h \ - ggml/src/ggml-quants.h \ - ggml/src/ggml-common.h - $(CC) $(CFLAGS) -c $< -o $@ +# Default target +all: $(BUILD_TARGETS) -ggml/src/ggml-aarch64.o: \ - ggml/src/ggml-aarch64.c \ - ggml/include/ggml.h \ - ggml/src/ggml-aarch64.h \ - ggml/src/ggml-common.h - $(CC) $(CFLAGS) -c $< -o $@ +# force c++ build for source file that have same name as c file +# Note: need this exception because `ggml-cpu.c` and `ggml-cpu.cpp` both produce the same obj/dep files +$(DIR_GGML)/%_cpp.o: $(DIR_GGML)/%.cpp + $(CXX) $(CXXFLAGS) -MMD -c $< -o $@ -ggml/src/ggml-blas.o: \ - ggml/src/ggml-blas.cpp \ - ggml/include/ggml-blas.h - $(CXX) $(CXXFLAGS) -c $< -o $@ +# Rules for building object files +$(DIR_GGML)/%.o: $(DIR_GGML)/%.c + $(CC) $(CFLAGS) -MMD -c $< -o $@ -ifndef GGML_NO_LLAMAFILE -ggml/src/llamafile/sgemm.o: \ - ggml/src/llamafile/sgemm.cpp \ - ggml/src/llamafile/sgemm.h \ - ggml/include/ggml.h - $(CXX) $(CXXFLAGS) -c $< -o $@ -endif # GGML_NO_LLAMAFILE +$(DIR_GGML)/%.o: $(DIR_GGML)/%.cpp + $(CXX) $(CXXFLAGS) -MMD -c $< -o $@ -ifndef GGML_NO_AMX -ggml/src/ggml-amx.o: \ - ggml/src/ggml-amx.cpp \ - ggml/include/ggml-amx.h - $(CXX) $(CXXFLAGS) -c $< -o $@ +$(DIR_LLAMA)/%.o: $(DIR_LLAMA)/%.cpp + $(CXX) $(CXXFLAGS) -MMD -c $< -o $@ -ggml/src/ggml-amx/mmq.o: \ - ggml/src/ggml-amx/mmq.cpp \ - ggml/src/ggml-amx/mmq.h \ - ggml/include/ggml.h - $(CXX) $(CXXFLAGS) -c $< -o $@ -endif +$(DIR_COMMON)/%.o: $(DIR_COMMON)/%.cpp + $(CXX) $(CXXFLAGS) -MMD -c $< -o $@ -ifdef GGML_RPC -ggml/src/ggml-rpc.o: \ - ggml/src/ggml-rpc.cpp \ - ggml/include/ggml-rpc.h - $(CXX) $(CXXFLAGS) -c $< -o $@ -endif # GGML_RPC - -$(LIB_GGML): \ - $(OBJ_GGML) +# Rules for building libraries +$(LIB_GGML): $(OBJ_GGML) $(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS) -$(LIB_GGML_S): \ - $(OBJ_GGML) +$(LIB_GGML_S): $(OBJ_GGML) ar rcs $(LIB_GGML_S) $^ -# llama - -src/unicode.o: \ - src/unicode.cpp \ - src/unicode.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -src/unicode-data.o: \ - src/unicode-data.cpp \ - src/unicode-data.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -src/llama.o: \ - src/llama.cpp \ - src/llama-impl.h \ - src/llama-vocab.h \ - src/llama-grammar.h \ - src/llama-sampling.h \ - src/unicode.h \ - include/llama.h \ - ggml/include/ggml-cuda.h \ - ggml/include/ggml-metal.h \ - ggml/include/ggml.h \ - ggml/include/ggml-alloc.h \ - ggml/include/ggml-backend.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -src/llama-vocab.o: \ - src/llama-vocab.cpp \ - src/llama-vocab.h \ - src/llama-impl.h \ - include/llama.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -src/llama-grammar.o: \ - src/llama-grammar.cpp \ - src/llama-grammar.h \ - src/llama-impl.h \ - src/llama-vocab.h \ - src/llama-sampling.h \ - include/llama.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -src/llama-sampling.o: \ - src/llama-sampling.cpp \ - src/llama-sampling.h \ - src/llama-impl.h \ - include/llama.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -$(LIB_LLAMA): \ - $(OBJ_LLAMA) \ - $(LIB_GGML) +$(LIB_LLAMA): $(OBJ_LLAMA) $(LIB_GGML) $(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS) -$(LIB_LLAMA_S): \ - $(OBJ_LLAMA) +$(LIB_LLAMA_S): $(OBJ_LLAMA) ar rcs $(LIB_LLAMA_S) $^ -# common - -common/common.o: \ - common/common.cpp \ - common/common.h \ - common/console.h \ - common/sampling.h \ - common/json.hpp \ - common/json-schema-to-grammar.h \ - include/llama.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/arg.o: \ - common/arg.cpp \ - common/arg.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/log.o: \ - common/log.cpp \ - common/log.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/sampling.o: \ - common/sampling.cpp \ - common/sampling.h \ - include/llama.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/console.o: \ - common/console.cpp \ - common/console.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/json-schema-to-grammar.o: \ - common/json-schema-to-grammar.cpp \ - common/json-schema-to-grammar.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -common/ngram-cache.o: \ - common/ngram-cache.cpp \ - common/ngram-cache.h - $(CXX) $(CXXFLAGS) -c $< -o $@ - -$(LIB_COMMON): \ - $(OBJ_COMMON) \ - $(LIB_LLAMA) \ - $(LIB_GGML) +$(LIB_COMMON): $(OBJ_COMMON) $(LIB_LLAMA) $(LIB_GGML) $(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS) -$(LIB_COMMON_S): \ - $(OBJ_COMMON) +$(LIB_COMMON_S): $(OBJ_COMMON) ar rcs $(LIB_COMMON_S) $^ -clean: - rm -vrf *.dot $(BUILD_TARGETS) $(TEST_TARGETS) - rm -rvf src/*.o - rm -rvf tests/*.o - rm -rvf examples/*.o - rm -rvf common/*.o - rm -rvf *.a - rm -rvf *.dll - rm -rvf *.so - rm -rvf *.dot - rm -rvf ggml/*.a - rm -rvf ggml/*.dll - rm -rvf ggml/*.so - rm -vrf ggml/src/*.o - rm -rvf ggml/src/llamafile/*.o - rm -rvf common/build-info.cpp - rm -vrf ggml/src/ggml-metal-embed.metal - rm -vrf ggml/src/ggml-cuda/*.o - rm -vrf ggml/src/ggml-cuda/template-instances/*.o - rm -vrf ggml/src/ggml-amx/*.o - rm -rvf $(BUILD_TARGETS) - rm -rvf $(TEST_TARGETS) - rm -f vulkan-shaders-gen ggml/src/ggml-vulkan-shaders.hpp ggml/src/ggml-vulkan-shaders.cpp - rm -rvf $(LEGACY_TARGETS_CLEAN) - find examples pocs -type f -name "*.o" -delete +# Include dependency files +-include $(DEP_FILES) + +# Clean generated server assets +clean-server-assets: + find examples/server -type f -name "*.js.hpp" -delete + find examples/server -type f -name "*.mjs.hpp" -delete + find examples/server -type f -name "*.css.hpp" -delete + find examples/server -type f -name "*.html.hpp" -delete + +# Clean rule +clean: clean-server-assets + rm -vrf $(BUILD_TARGETS) $(TEST_TARGETS) + rm -rvf *.a *.dll *.so *.dot + find ggml src common tests examples pocs -type f -name "*.o" -delete + find ggml src common tests examples pocs -type f -name "*.d" -delete # # Examples @@ -1283,6 +1182,11 @@ llama-infill: examples/infill/infill.cpp \ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) +llama-run: examples/run/run.cpp \ + $(OBJ_ALL) + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + llama-simple: examples/simple/simple.cpp \ $(OBJ_ALL) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) @@ -1456,20 +1360,14 @@ llama-server: \ examples/server/utils.hpp \ examples/server/httplib.h \ examples/server/index.html.hpp \ - examples/server/completion.js.hpp \ examples/server/loading.html.hpp \ - examples/server/deps_daisyui.min.css.hpp \ - examples/server/deps_markdown-it.js.hpp \ - examples/server/deps_tailwindcss.js.hpp \ - examples/server/deps_vue.esm-browser.js.hpp \ common/json.hpp \ - common/stb_image.h \ $(OBJ_ALL) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h %.hpp $<,$^) -Iexamples/server $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) $(LWINSOCK2) # Portable equivalent of `cd examples/server/public && xxd -i $(notdir $<) ../$(notdir $<).hpp`: -examples/server/%.hpp: examples/server/public/% Makefile +examples/server/%.hpp: examples/server/public/% FORCE Makefile @( export NAME=$(subst .,_,$(subst -,_,$(notdir $<))) && \ echo "unsigned char $${NAME}[] = {" && \ cat $< | od -v -t x1 -An | sed -E 's/([0-9a-fA-F]+)/0x\1, /g' && \ @@ -1507,6 +1405,14 @@ llama-minicpmv-cli: examples/llava/minicpmv-cli.cpp \ $(OBJ_ALL) $(CXX) $(CXXFLAGS) $< $(filter-out %.h $<,$^) -o $@ $(LDFLAGS) -Wno-cast-qual +llama-qwen2vl-cli: examples/llava/qwen2vl-cli.cpp \ + examples/llava/llava.cpp \ + examples/llava/llava.h \ + examples/llava/clip.cpp \ + examples/llava/clip.h \ + $(OBJ_ALL) + $(CXX) $(CXXFLAGS) $< $(filter-out %.h $<,$^) -o $@ $(LDFLAGS) -Wno-cast-qual + ifeq ($(UNAME_S),Darwin) swift: examples/batched.swift (cd examples/batched.swift; make build) @@ -1563,11 +1469,6 @@ tests/test-json-schema-to-grammar: tests/test-json-schema-to-grammar.cpp \ $(CXX) $(CXXFLAGS) -Iexamples/server -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) -tests/test-grad0: tests/test-grad0.cpp \ - $(OBJ_GGML) - $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) - $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) - tests/test-opt: tests/test-opt.cpp \ $(OBJ_GGML) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) @@ -1649,7 +1550,7 @@ llama-q8dot: pocs/vdot/q8dot.cpp ggml/src/ggml.o \ # Deprecated binaries that we want to keep around long enough for people to migrate to the new filenames, then these can be removed. # # Mark legacy binary targets as .PHONY so that they are always checked. -.PHONY: main quantize perplexity embedding server +.PHONY: FORCE main quantize perplexity embedding server # Define the object file target examples/deprecation-warning/deprecation-warning.o: examples/deprecation-warning/deprecation-warning.cpp diff --git a/Package.swift b/Package.swift index d3661d13c..01c996d24 100644 --- a/Package.swift +++ b/Package.swift @@ -2,49 +2,6 @@ import PackageDescription -var sources = [ - "src/llama.cpp", - "src/llama-vocab.cpp", - "src/llama-grammar.cpp", - "src/llama-sampling.cpp", - "src/unicode.cpp", - "src/unicode-data.cpp", - "ggml/src/ggml.c", - "ggml/src/ggml-cpu.c", - "ggml/src/ggml-alloc.c", - "ggml/src/ggml-backend.cpp", - "ggml/src/ggml-quants.c", - "ggml/src/ggml-aarch64.c", -] - -var resources: [Resource] = [] -var linkerSettings: [LinkerSetting] = [] -var cSettings: [CSetting] = [ - .unsafeFlags(["-Wno-shorten-64-to-32", "-O3", "-DNDEBUG"]), - .unsafeFlags(["-fno-objc-arc"]), - // NOTE: NEW_LAPACK will required iOS version 16.4+ - // We should consider add this in the future when we drop support for iOS 14 - // (ref: ref: https://developer.apple.com/documentation/accelerate/1513264-cblas_sgemm?language=objc) - // .define("ACCELERATE_NEW_LAPACK"), - // .define("ACCELERATE_LAPACK_ILP64") -] - -#if canImport(Darwin) -sources.append("ggml/src/ggml-metal.m") -resources.append(.process("ggml/src/ggml-metal.metal")) -linkerSettings.append(.linkedFramework("Accelerate")) -cSettings.append( - contentsOf: [ - .define("GGML_USE_ACCELERATE"), - .define("GGML_USE_METAL") - ] -) -#endif - -#if os(Linux) - cSettings.append(.define("_GNU_SOURCE")) -#endif - let package = Package( name: "llama", platforms: [ @@ -57,24 +14,6 @@ let package = Package( .library(name: "llama", targets: ["llama"]), ], targets: [ - .target( - name: "llama", - path: ".", - exclude: [ - "cmake", - "examples", - "scripts", - "models", - "tests", - "CMakeLists.txt", - "Makefile" - ], - sources: sources, - resources: resources, - publicHeadersPath: "spm-headers", - cSettings: cSettings, - linkerSettings: linkerSettings - ) - ], - cxxLanguageStandard: .cxx11 + .systemLibrary(name: "llama", pkgConfig: "llama"), + ] ) diff --git a/README.md b/README.md index 0378a674e..413a16422 100644 --- a/README.md +++ b/README.md @@ -4,7 +4,6 @@ [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT) [![Server](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml/badge.svg)](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml) -[![Conan Center](https://shields.io/conan/v/llama-cpp)](https://conan.io/center/llama-cpp) [Roadmap](https://github.com/users/ggerganov/projects/7) / [Project status](https://github.com/ggerganov/llama.cpp/discussions/3471) / [Manifesto](https://github.com/ggerganov/llama.cpp/discussions/205) / [ggml](https://github.com/ggerganov/ggml) @@ -26,7 +25,7 @@ Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others) ## Description The main goal of `llama.cpp` is to enable LLM inference with minimal setup and state-of-the-art performance on a wide -variety of hardware - locally and in the cloud. +range of hardware - locally and in the cloud. - Plain C/C++ implementation without any dependencies - Apple silicon is a first-class citizen - optimized via ARM NEON, Accelerate and Metal frameworks @@ -36,14 +35,17 @@ variety of hardware - locally and in the cloud. - Vulkan and SYCL backend support - CPU+GPU hybrid inference to partially accelerate models larger than the total VRAM capacity -Since its [inception](https://github.com/ggerganov/llama.cpp/issues/33#issuecomment-1465108022), the project has -improved significantly thanks to many contributions. It is the main playground for developing new features for the -[ggml](https://github.com/ggerganov/ggml) library. +The `llama.cpp` project is the main playground for developing new features for the [ggml](https://github.com/ggerganov/ggml) library. -**Supported models:** +
+Models Typically finetunes of the base models below are supported as well. +Instructions for adding support for new models: [HOWTO-add-model.md](docs/development/HOWTO-add-model.md) + +#### Text-only + - [X] LLaMA 🦙 - [x] LLaMA 2 🦙🦙 - [x] LLaMA 3 🦙🦙🦙 @@ -67,6 +69,7 @@ Typically finetunes of the base models below are supported as well. - [x] [Qwen models](https://huggingface.co/models?search=Qwen/Qwen) - [x] [PLaMo-13B](https://github.com/ggerganov/llama.cpp/pull/3557) - [x] [Phi models](https://huggingface.co/models?search=microsoft/phi) +- [x] [PhiMoE](https://github.com/ggerganov/llama.cpp/pull/11003) - [x] [GPT-2](https://huggingface.co/gpt2) - [x] [Orion 14B](https://github.com/ggerganov/llama.cpp/pull/5118) - [x] [InternLM2](https://huggingface.co/models?search=internlm2) @@ -79,6 +82,7 @@ Typically finetunes of the base models below are supported as well. - [x] [SEA-LION](https://huggingface.co/models?search=sea-lion) - [x] [GritLM-7B](https://huggingface.co/GritLM/GritLM-7B) + [GritLM-8x7B](https://huggingface.co/GritLM/GritLM-8x7B) - [x] [OLMo](https://allenai.org/olmo) +- [x] [OLMo 2](https://allenai.org/olmo) - [x] [OLMoE](https://huggingface.co/allenai/OLMoE-1B-7B-0924) - [x] [Granite models](https://huggingface.co/collections/ibm-granite/granite-code-models-6624c5cec322e4c148c8b330) - [x] [GPT-NeoX](https://github.com/EleutherAI/gpt-neox) + [Pythia](https://github.com/EleutherAI/pythia) @@ -95,10 +99,10 @@ Typically finetunes of the base models below are supported as well. - [x] [Jais](https://huggingface.co/inceptionai/jais-13b-chat) - [x] [Bielik-11B-v2.3](https://huggingface.co/collections/speakleash/bielik-11b-v23-66ee813238d9b526a072408a) - [x] [RWKV-6](https://github.com/BlinkDL/RWKV-LM) +- [x] [QRWKV-6](https://huggingface.co/recursal/QRWKV6-32B-Instruct-Preview-v0.1) +- [x] [GigaChat-20B-A3B](https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct) -(instructions for supporting more models: [HOWTO-add-model.md](./docs/development/HOWTO-add-model.md)) - -**Multimodal models:** +#### Multimodal - [x] [LLaVA 1.5 models](https://huggingface.co/collections/liuhaotian/llava-15-653aac15d994e992e2677a7e), [LLaVA 1.6 models](https://huggingface.co/collections/liuhaotian/llava-16-65b9e40155f60fd046a5ccf2) - [x] [BakLLaVA](https://huggingface.co/models?search=SkunkworksAI/Bakllava) @@ -109,8 +113,12 @@ Typically finetunes of the base models below are supported as well. - [x] [Mini CPM](https://huggingface.co/models?search=MiniCPM) - [x] [Moondream](https://huggingface.co/vikhyatk/moondream2) - [x] [Bunny](https://github.com/BAAI-DCAI/Bunny) +- [x] [Qwen2-VL](https://huggingface.co/collections/Qwen/qwen2-vl-66cee7455501d7126940800d) -**Bindings:** +
+ +
+Bindings - Python: [abetlen/llama-cpp-python](https://github.com/abetlen/llama-cpp-python) - Go: [go-skynet/go-llama.cpp](https://github.com/go-skynet/go-llama.cpp) @@ -131,321 +139,339 @@ Typically finetunes of the base models below are supported as well. - Java: [kherud/java-llama.cpp](https://github.com/kherud/java-llama.cpp) - Zig: [deins/llama.cpp.zig](https://github.com/Deins/llama.cpp.zig) - Flutter/Dart: [netdur/llama_cpp_dart](https://github.com/netdur/llama_cpp_dart) +- Flutter: [xuegao-tzx/Fllama](https://github.com/xuegao-tzx/Fllama) - PHP (API bindings and features built on top of llama.cpp): [distantmagic/resonance](https://github.com/distantmagic/resonance) [(more info)](https://github.com/ggerganov/llama.cpp/pull/6326) - Guile Scheme: [guile_llama_cpp](https://savannah.nongnu.org/projects/guile-llama-cpp) - Swift [srgtuszy/llama-cpp-swift](https://github.com/srgtuszy/llama-cpp-swift) - Swift [ShenghaiWang/SwiftLlama](https://github.com/ShenghaiWang/SwiftLlama) -**UI:** +
-Unless otherwise noted these projects are open-source with permissive licensing: - -- [MindWorkAI/AI-Studio](https://github.com/MindWorkAI/AI-Studio) (FSL-1.1-MIT) -- [iohub/collama](https://github.com/iohub/coLLaMA) -- [janhq/jan](https://github.com/janhq/jan) (AGPL) -- [nat/openplayground](https://github.com/nat/openplayground) -- [Faraday](https://faraday.dev/) (proprietary) -- [LMStudio](https://lmstudio.ai/) (proprietary) -- [Layla](https://play.google.com/store/apps/details?id=com.laylalite) (proprietary) -- [ramalama](https://github.com/containers/ramalama) (MIT) -- [LocalAI](https://github.com/mudler/LocalAI) (MIT) -- [LostRuins/koboldcpp](https://github.com/LostRuins/koboldcpp) (AGPL) -- [Mozilla-Ocho/llamafile](https://github.com/Mozilla-Ocho/llamafile) -- [nomic-ai/gpt4all](https://github.com/nomic-ai/gpt4all) -- [ollama/ollama](https://github.com/ollama/ollama) -- [oobabooga/text-generation-webui](https://github.com/oobabooga/text-generation-webui) (AGPL) -- [psugihara/FreeChat](https://github.com/psugihara/FreeChat) -- [cztomsik/ava](https://github.com/cztomsik/ava) (MIT) -- [ptsochantaris/emeltal](https://github.com/ptsochantaris/emeltal) -- [pythops/tenere](https://github.com/pythops/tenere) (AGPL) -- [RAGNA Desktop](https://ragna.app/) (proprietary) -- [RecurseChat](https://recurse.chat/) (proprietary) -- [semperai/amica](https://github.com/semperai/amica) -- [withcatai/catai](https://github.com/withcatai/catai) -- [Mobile-Artificial-Intelligence/maid](https://github.com/Mobile-Artificial-Intelligence/maid) (MIT) -- [Msty](https://msty.app) (proprietary) -- [LLMFarm](https://github.com/guinmoon/LLMFarm?tab=readme-ov-file) (MIT) -- [KanTV](https://github.com/zhouwg/kantv?tab=readme-ov-file)(Apachev2.0 or later) -- [Dot](https://github.com/alexpinel/Dot) (GPL) -- [MindMac](https://mindmac.app) (proprietary) -- [KodiBot](https://github.com/firatkiral/kodibot) (GPL) -- [eva](https://github.com/ylsdamxssjxxdd/eva) (MIT) -- [AI Sublime Text plugin](https://github.com/yaroslavyaroslav/OpenAI-sublime-text) (MIT) -- [AIKit](https://github.com/sozercan/aikit) (MIT) -- [LARS - The LLM & Advanced Referencing Solution](https://github.com/abgulati/LARS) (AGPL) -- [LLMUnity](https://github.com/undreamai/LLMUnity) (MIT) -- [Llama Assistant](https://github.com/vietanhdev/llama-assistant) (GPL) -- [PocketPal AI - An iOS and Android App](https://github.com/a-ghorbani/pocketpal-ai) (MIT) +
+UIs *(to have a project listed here, it should clearly state that it depends on `llama.cpp`)* -**Tools:** +- [AI Sublime Text plugin](https://github.com/yaroslavyaroslav/OpenAI-sublime-text) (MIT) +- [cztomsik/ava](https://github.com/cztomsik/ava) (MIT) +- [Dot](https://github.com/alexpinel/Dot) (GPL) +- [eva](https://github.com/ylsdamxssjxxdd/eva) (MIT) +- [iohub/collama](https://github.com/iohub/coLLaMA) (Apache-2.0) +- [janhq/jan](https://github.com/janhq/jan) (AGPL) +- [KanTV](https://github.com/zhouwg/kantv?tab=readme-ov-file) (Apache-2.0) +- [KodiBot](https://github.com/firatkiral/kodibot) (GPL) +- [llama.vim](https://github.com/ggml-org/llama.vim) (MIT) +- [LARS](https://github.com/abgulati/LARS) (AGPL) +- [Llama Assistant](https://github.com/vietanhdev/llama-assistant) (GPL) +- [LLMFarm](https://github.com/guinmoon/LLMFarm?tab=readme-ov-file) (MIT) +- [LLMUnity](https://github.com/undreamai/LLMUnity) (MIT) +- [LMStudio](https://lmstudio.ai/) (proprietary) +- [LocalAI](https://github.com/mudler/LocalAI) (MIT) +- [LostRuins/koboldcpp](https://github.com/LostRuins/koboldcpp) (AGPL) +- [MindMac](https://mindmac.app) (proprietary) +- [MindWorkAI/AI-Studio](https://github.com/MindWorkAI/AI-Studio) (FSL-1.1-MIT) +- [Mobile-Artificial-Intelligence/maid](https://github.com/Mobile-Artificial-Intelligence/maid) (MIT) +- [Mozilla-Ocho/llamafile](https://github.com/Mozilla-Ocho/llamafile) (Apache-2.0) +- [nat/openplayground](https://github.com/nat/openplayground) (MIT) +- [nomic-ai/gpt4all](https://github.com/nomic-ai/gpt4all) (MIT) +- [ollama/ollama](https://github.com/ollama/ollama) (MIT) +- [oobabooga/text-generation-webui](https://github.com/oobabooga/text-generation-webui) (AGPL) +- [PocketPal AI](https://github.com/a-ghorbani/pocketpal-ai) (MIT) +- [psugihara/FreeChat](https://github.com/psugihara/FreeChat) (MIT) +- [ptsochantaris/emeltal](https://github.com/ptsochantaris/emeltal) (MIT) +- [pythops/tenere](https://github.com/pythops/tenere) (AGPL) +- [ramalama](https://github.com/containers/ramalama) (MIT) +- [semperai/amica](https://github.com/semperai/amica) (MIT) +- [withcatai/catai](https://github.com/withcatai/catai) (MIT) + +
+ +
+Tools - [akx/ggify](https://github.com/akx/ggify) – download PyTorch models from HuggingFace Hub and convert them to GGML - [akx/ollama-dl](https://github.com/akx/ollama-dl) – download models from the Ollama library to be used directly with llama.cpp - [crashr/gppm](https://github.com/crashr/gppm) – launch llama.cpp instances utilizing NVIDIA Tesla P40 or P100 GPUs with reduced idle power consumption - [gpustack/gguf-parser](https://github.com/gpustack/gguf-parser-go/tree/main/cmd/gguf-parser) - review/check the GGUF file and estimate the memory usage -- [Styled Lines](https://marketplace.unity.com/packages/tools/generative-ai/styled-lines-llama-cpp-model-292902) (proprietary licensed, async wrapper of inference part for game development in Unity3d with prebuild Mobile and Web platform wrappers and a model example) +- [Styled Lines](https://marketplace.unity.com/packages/tools/generative-ai/styled-lines-llama-cpp-model-292902) (proprietary licensed, async wrapper of inference part for game development in Unity3d with pre-built Mobile and Web platform wrappers and a model example) -**Infrastructure:** +
+ +
+Infrastructure - [Paddler](https://github.com/distantmagic/paddler) - Stateful load balancer custom-tailored for llama.cpp - [GPUStack](https://github.com/gpustack/gpustack) - Manage GPU clusters for running LLMs - [llama_cpp_canister](https://github.com/onicai/llama_cpp_canister) - llama.cpp as a smart contract on the Internet Computer, using WebAssembly +- [llama-swap](https://github.com/mostlygeek/llama-swap) - transparent proxy that adds automatic model switching with llama-server + +
+ +
+Games -**Games:** - [Lucy's Labyrinth](https://github.com/MorganRO8/Lucys_Labyrinth) - A simple maze game where agents controlled by an AI model will try to trick you. -## Demo - -
-Typical run using LLaMA v2 13B on M2 Ultra - -``` -$ make -j && ./llama-cli -m models/llama-13b-v2/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:\nStep 1:" -n 400 -e -I llama.cpp build info: -I UNAME_S: Darwin -I UNAME_P: arm -I UNAME_M: arm64 -I CFLAGS: -I. -O3 -std=c11 -fPIC -DNDEBUG -Wall -Wextra -Wpedantic -Wcast-qual -Wdouble-promotion -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes -pthread -DGGML_USE_K_QUANTS -DGGML_USE_ACCELERATE -I CXXFLAGS: -I. -I./common -O3 -std=c++11 -fPIC -DNDEBUG -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wno-multichar -pthread -DGGML_USE_K_QUANTS -I LDFLAGS: -framework Accelerate -I CC: Apple clang version 14.0.3 (clang-1403.0.22.14.1) -I CXX: Apple clang version 14.0.3 (clang-1403.0.22.14.1) - -make: Nothing to be done for `default'. -main: build = 1041 (cf658ad) -main: seed = 1692823051 -llama_model_loader: loaded meta data with 16 key-value pairs and 363 tensors from models/llama-13b-v2/ggml-model-q4_0.gguf (version GGUF V1 (latest)) -llama_model_loader: - type f32: 81 tensors -llama_model_loader: - type q4_0: 281 tensors -llama_model_loader: - type q6_K: 1 tensors -llm_load_print_meta: format = GGUF V1 (latest) -llm_load_print_meta: arch = llama -llm_load_print_meta: vocab type = SPM -llm_load_print_meta: n_vocab = 32000 -llm_load_print_meta: n_merges = 0 -llm_load_print_meta: n_ctx_train = 4096 -llm_load_print_meta: n_ctx = 512 -llm_load_print_meta: n_embd = 5120 -llm_load_print_meta: n_head = 40 -llm_load_print_meta: n_head_kv = 40 -llm_load_print_meta: n_layer = 40 -llm_load_print_meta: n_rot = 128 -llm_load_print_meta: n_gqa = 1 -llm_load_print_meta: f_norm_eps = 1.0e-05 -llm_load_print_meta: f_norm_rms_eps = 1.0e-05 -llm_load_print_meta: n_ff = 13824 -llm_load_print_meta: freq_base = 10000.0 -llm_load_print_meta: freq_scale = 1 -llm_load_print_meta: model type = 13B -llm_load_print_meta: model ftype = mostly Q4_0 -llm_load_print_meta: model size = 13.02 B -llm_load_print_meta: general.name = LLaMA v2 -llm_load_print_meta: BOS token = 1 '' -llm_load_print_meta: EOS token = 2 '' -llm_load_print_meta: UNK token = 0 '' -llm_load_print_meta: LF token = 13 '<0x0A>' -llm_load_tensors: ggml ctx size = 0.11 MB -llm_load_tensors: mem required = 7024.01 MB (+ 400.00 MB per state) -................................................................................................... -llama_new_context_with_model: kv self size = 400.00 MB -llama_new_context_with_model: compute buffer total size = 75.41 MB - -system_info: n_threads = 16 / 24 | AVX = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | FMA = 0 | NEON = 1 | ARM_FMA = 1 | F16C = 0 | FP16_VA = 1 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 0 | VSX = 0 | -sampling: repeat_last_n = 64, repeat_penalty = 1.100000, presence_penalty = 0.000000, frequency_penalty = 0.000000, top_k = 40, tfs_z = 1.000000, top_p = 0.950000, typical_p = 1.000000, temp = 0.800000, mirostat = 0, mirostat_lr = 0.100000, mirostat_ent = 5.000000 -generate: n_ctx = 512, n_batch = 512, n_predict = 400, n_keep = 0 - - - Building a website can be done in 10 simple steps: -Step 1: Find the right website platform. -Step 2: Choose your domain name and hosting plan. -Step 3: Design your website layout. -Step 4: Write your website content and add images. -Step 5: Install security features to protect your site from hackers or spammers -Step 6: Test your website on multiple browsers, mobile devices, operating systems etc… -Step 7: Test it again with people who are not related to you personally – friends or family members will work just fine! -Step 8: Start marketing and promoting the website via social media channels or paid ads -Step 9: Analyze how many visitors have come to your site so far, what type of people visit more often than others (e.g., men vs women) etc… -Step 10: Continue to improve upon all aspects mentioned above by following trends in web design and staying up-to-date on new technologies that can enhance user experience even further! -How does a Website Work? -A website works by having pages, which are made of HTML code. This code tells your computer how to display the content on each page you visit – whether it’s an image or text file (like PDFs). In order for someone else’s browser not only be able but also want those same results when accessing any given URL; some additional steps need taken by way of programming scripts that will add functionality such as making links clickable! -The most common type is called static HTML pages because they remain unchanged over time unless modified manually (either through editing files directly or using an interface such as WordPress). They are usually served up via HTTP protocols – this means anyone can access them without having any special privileges like being part of a group who is allowed into restricted areas online; however, there may still exist some limitations depending upon where one lives geographically speaking. -How to -llama_print_timings: load time = 576.45 ms -llama_print_timings: sample time = 283.10 ms / 400 runs ( 0.71 ms per token, 1412.91 tokens per second) -llama_print_timings: prompt eval time = 599.83 ms / 19 tokens ( 31.57 ms per token, 31.68 tokens per second) -llama_print_timings: eval time = 24513.59 ms / 399 runs ( 61.44 ms per token, 16.28 tokens per second) -llama_print_timings: total time = 25431.49 ms -``` -
-
-Demo of running both LLaMA-7B and whisper.cpp on a single M1 Pro MacBook - -And here is another demo of running both LLaMA-7B and [whisper.cpp](https://github.com/ggerganov/whisper.cpp) on a single M1 Pro MacBook: - -https://user-images.githubusercontent.com/1991296/224442907-7693d4be-acaa-4e01-8b4f-add84093ffff.mp4 - -
- -## Usage - -Here are the end-to-end binary build and model conversion steps for most supported models. - -### Basic usage - -Firstly, you need to get the binary. There are different methods that you can follow: -- Method 1: Clone this repository and build locally, see [how to build](./docs/build.md) -- Method 2: If you are using MacOS or Linux, you can install llama.cpp via [brew, flox or nix](./docs/install.md) -- Method 3: Use a Docker image, see [documentation for Docker](./docs/docker.md) -- Method 4: Download pre-built binary from [releases](https://github.com/ggerganov/llama.cpp/releases) - -You can run a basic completion using this command: - -```bash -llama-cli -m your_model.gguf -p "I believe the meaning of life is" -n 128 - -# Output: -# I believe the meaning of life is to find your own truth and to live in accordance with it. For me, this means being true to myself and following my passions, even if they don't align with societal expectations. I think that's what I love about yoga – it's not just a physical practice, but a spiritual one too. It's about connecting with yourself, listening to your inner voice, and honoring your own unique journey. -``` - -See [this page](./examples/main/README.md) for a full list of parameters. - -### Conversation mode - -If you want a more ChatGPT-like experience, you can run in conversation mode by passing `-cnv` as a parameter: - -```bash -llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv - -# Output: -# > hi, who are you? -# Hi there! I'm your helpful assistant! I'm an AI-powered chatbot designed to assist and provide information to users like you. I'm here to help answer your questions, provide guidance, and offer support on a wide range of topics. I'm a friendly and knowledgeable AI, and I'm always happy to help with anything you need. What's on your mind, and how can I assist you today? -# -# > what is 1+1? -# Easy peasy! The answer to 1+1 is... 2! -``` - -By default, the chat template will be taken from the input model. If you want to use another chat template, pass `--chat-template NAME` as a parameter. See the list of [supported templates](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) - -```bash -./llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv --chat-template chatml -``` - -You can also use your own template via in-prefix, in-suffix and reverse-prompt parameters: - -```bash -./llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv --in-prefix 'User: ' --reverse-prompt 'User:' -``` - -### Web server - -[llama.cpp web server](./examples/server/README.md) is a lightweight [OpenAI API](https://github.com/openai/openai-openapi) compatible HTTP server that can be used to serve local models and easily connect them to existing clients. - -Example usage: - -```bash -./llama-server -m your_model.gguf --port 8080 - -# Basic web UI can be accessed via browser: http://localhost:8080 -# Chat completion endpoint: http://localhost:8080/v1/chat/completions -``` - -### Interactive mode - -> [!NOTE] -> If you prefer basic usage, please consider using conversation mode instead of interactive mode - -In this mode, you can always interrupt generation by pressing Ctrl+C and entering one or more lines of text, which will be converted into tokens and appended to the current context. You can also specify a *reverse prompt* with the parameter `-r "reverse prompt string"`. This will result in user input being prompted whenever the exact tokens of the reverse prompt string are encountered in the generation. A typical use is to use a prompt that makes LLaMA emulate a chat between multiple users, say Alice and Bob, and pass `-r "Alice:"`. - -Here is an example of a few-shot interaction, invoked with the command - -```bash -# default arguments using a 7B model -./examples/chat.sh - -# advanced chat with a 13B model -./examples/chat-13B.sh - -# custom arguments using a 13B model -./llama-cli -m ./models/13B/ggml-model-q4_0.gguf -n 256 --repeat_penalty 1.0 --color -i -r "User:" -f prompts/chat-with-bob.txt -``` - -Note the use of `--color` to distinguish between user input and generated text. Other parameters are explained in more detail in the [README](examples/main/README.md) for the `llama-cli` example program. - -![image](https://user-images.githubusercontent.com/1991296/224575029-2af3c7dc-5a65-4f64-a6bb-517a532aea38.png) - -### Persistent Interaction - -The prompt, user inputs, and model generations can be saved and resumed across calls to `./llama-cli` by leveraging `--prompt-cache` and `--prompt-cache-all`. The `./examples/chat-persistent.sh` script demonstrates this with support for long-running, resumable chat sessions. To use this example, you must provide a file to cache the initial chat prompt and a directory to save the chat session, and may optionally provide the same variables as `chat-13B.sh`. The same prompt cache can be reused for new chat sessions. Note that both prompt cache and chat directory are tied to the initial prompt (`PROMPT_TEMPLATE`) and the model file. - -```bash -# Start a new chat -PROMPT_CACHE_FILE=chat.prompt.bin CHAT_SAVE_DIR=./chat/default ./examples/chat-persistent.sh - -# Resume that chat -PROMPT_CACHE_FILE=chat.prompt.bin CHAT_SAVE_DIR=./chat/default ./examples/chat-persistent.sh - -# Start a different chat with the same prompt/model -PROMPT_CACHE_FILE=chat.prompt.bin CHAT_SAVE_DIR=./chat/another ./examples/chat-persistent.sh - -# Different prompt cache for different prompt/model -PROMPT_TEMPLATE=./prompts/chat-with-bob.txt PROMPT_CACHE_FILE=bob.prompt.bin \ - CHAT_SAVE_DIR=./chat/bob ./examples/chat-persistent.sh -``` - -### Constrained output with grammars - -`llama.cpp` supports grammars to constrain model output. For example, you can force the model to output JSON only: - -```bash -./llama-cli -m ./models/13B/ggml-model-q4_0.gguf -n 256 --grammar-file grammars/json.gbnf -p 'Request: schedule a call at 8pm; Command:' -``` - -The `grammars/` folder contains a handful of sample grammars. To write your own, check out the [GBNF Guide](./grammars/README.md). - -For authoring more complex JSON grammars, you can also check out https://grammar.intrinsiclabs.ai/, a browser app that lets you write TypeScript interfaces which it compiles to GBNF grammars that you can save for local use. Note that the app is built and maintained by members of the community, please file any issues or FRs on [its repo](http://github.com/intrinsiclabsai/gbnfgen) and not this one. - -## Build - -Please refer to [Build llama.cpp locally](./docs/build.md) - ## Supported backends | Backend | Target devices | | --- | --- | -| [Metal](./docs/build.md#metal-build) | Apple Silicon | -| [BLAS](./docs/build.md#blas-build) | All | -| [BLIS](./docs/backend/BLIS.md) | All | -| [SYCL](./docs/backend/SYCL.md) | Intel and Nvidia GPU | -| [MUSA](./docs/build.md#musa) | Moore Threads MTT GPU | -| [CUDA](./docs/build.md#cuda) | Nvidia GPU | -| [hipBLAS](./docs/build.md#hipblas) | AMD GPU | -| [Vulkan](./docs/build.md#vulkan) | GPU | -| [CANN](./docs/build.md#cann) | Ascend NPU | +| [Metal](docs/build.md#metal-build) | Apple Silicon | +| [BLAS](docs/build.md#blas-build) | All | +| [BLIS](docs/backend/BLIS.md) | All | +| [SYCL](docs/backend/SYCL.md) | Intel and Nvidia GPU | +| [MUSA](docs/build.md#musa) | Moore Threads MTT GPU | +| [CUDA](docs/build.md#cuda) | Nvidia GPU | +| [HIP](docs/build.md#hip) | AMD GPU | +| [Vulkan](docs/build.md#vulkan) | GPU | +| [CANN](docs/build.md#cann) | Ascend NPU | -## Tools +## Building the project -### Prepare and Quantize +The main product of this project is the `llama` library. Its C-style interface can be found in [include/llama.h](include/llama.h). +The project also includes many example programs and tools using the `llama` library. The examples range from simple, minimal code snippets to sophisticated sub-projects such as an OpenAI-compatible HTTP server. Possible methods for obtaining the binaries: -> [!NOTE] -> You can use the [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space on Hugging Face to quantise your model weights without any setup too. It is synced from `llama.cpp` main every 6 hours. +- Clone this repository and build locally, see [how to build](docs/build.md) +- On MacOS or Linux, install `llama.cpp` via [brew, flox or nix](docs/install.md) +- Use a Docker image, see [documentation for Docker](docs/docker.md) +- Download pre-built binaries from [releases](https://github.com/ggerganov/llama.cpp/releases) -To obtain the official LLaMA 2 weights please see the Obtaining and using the Facebook LLaMA 2 model section. There is also a large selection of pre-quantized `gguf` models available on Hugging Face. +## Obtaining and quantizing models -Note: `convert.py` has been moved to `examples/convert_legacy_llama.py` and shouldn't be used for anything other than `Llama/Llama2/Mistral` models and their derivatives. -It does not support LLaMA 3, you can use `convert_hf_to_gguf.py` with LLaMA 3 downloaded from Hugging Face. +The [Hugging Face](https://huggingface.co) platform hosts a [number of LLMs](https://huggingface.co/models?library=gguf&sort=trending) compatible with `llama.cpp`: -To learn more about quantizing model, [read this documentation](./examples/quantize/README.md) +- [Trending](https://huggingface.co/models?library=gguf&sort=trending) +- [LLaMA](https://huggingface.co/models?sort=trending&search=llama+gguf) -### Perplexity (measuring model quality) +You can either manually download the GGUF file or directly use any `llama.cpp`-compatible models from Hugging Face by using this CLI argument: `-hf /[:quant]` -You can use the `perplexity` example to measure perplexity over a given prompt (lower perplexity is better). -For more information, see [https://huggingface.co/docs/transformers/perplexity](https://huggingface.co/docs/transformers/perplexity). +After downloading a model, use the CLI tools to run it locally - see below. + +`llama.cpp` requires the model to be stored in the [GGUF](https://github.com/ggerganov/ggml/blob/master/docs/gguf.md) file format. Models in other data formats can be converted to GGUF using the `convert_*.py` Python scripts in this repo. + +The Hugging Face platform provides a variety of online tools for converting, quantizing and hosting models with `llama.cpp`: + +- Use the [GGUF-my-repo space](https://huggingface.co/spaces/ggml-org/gguf-my-repo) to convert to GGUF format and quantize model weights to smaller sizes +- Use the [GGUF-my-LoRA space](https://huggingface.co/spaces/ggml-org/gguf-my-lora) to convert LoRA adapters to GGUF format (more info: https://github.com/ggerganov/llama.cpp/discussions/10123) +- Use the [GGUF-editor space](https://huggingface.co/spaces/CISCai/gguf-editor) to edit GGUF meta data in the browser (more info: https://github.com/ggerganov/llama.cpp/discussions/9268) +- Use the [Inference Endpoints](https://ui.endpoints.huggingface.co/) to directly host `llama.cpp` in the cloud (more info: https://github.com/ggerganov/llama.cpp/discussions/9669) + +To learn more about model quantization, [read this documentation](examples/quantize/README.md) + +## [`llama-cli`](examples/main) + +#### A CLI tool for accessing and experimenting with most of `llama.cpp`'s functionality. + +-
+ Run in conversation mode + + Models with a built-in chat template will automatically activate conversation mode. If this doesn't occur, you can manually enable it by adding `-cnv` and specifying a suitable chat template with `--chat-template NAME` + + ```bash + llama-cli -m model.gguf + + # > hi, who are you? + # Hi there! I'm your helpful assistant! I'm an AI-powered chatbot designed to assist and provide information to users like you. I'm here to help answer your questions, provide guidance, and offer support on a wide range of topics. I'm a friendly and knowledgeable AI, and I'm always happy to help with anything you need. What's on your mind, and how can I assist you today? + # + # > what is 1+1? + # Easy peasy! The answer to 1+1 is... 2! + ``` + +
+ +-
+ Run in conversation mode with custom chat template + + ```bash + # use the "chatml" template (use -h to see the list of supported templates) + llama-cli -m model.gguf -cnv --chat-template chatml + + # use a custom template + llama-cli -m model.gguf -cnv --in-prefix 'User: ' --reverse-prompt 'User:' + ``` + +
+ +-
+ Run simple text completion + + To disable conversation mode explicitly, use `-no-cnv` + + ```bash + llama-cli -m model.gguf -p "I believe the meaning of life is" -n 128 -no-cnv + + # I believe the meaning of life is to find your own truth and to live in accordance with it. For me, this means being true to myself and following my passions, even if they don't align with societal expectations. I think that's what I love about yoga – it's not just a physical practice, but a spiritual one too. It's about connecting with yourself, listening to your inner voice, and honoring your own unique journey. + ``` + +
+ +-
+ Constrain the output with a custom grammar + + ```bash + llama-cli -m model.gguf -n 256 --grammar-file grammars/json.gbnf -p 'Request: schedule a call at 8pm; Command:' + + # {"appointmentTime": "8pm", "appointmentDetails": "schedule a a call"} + ``` + + The [grammars/](grammars/) folder contains a handful of sample grammars. To write your own, check out the [GBNF Guide](grammars/README.md). + + For authoring more complex JSON grammars, check out https://grammar.intrinsiclabs.ai/ + +
+ + +## [`llama-server`](examples/server) + +#### A lightweight, [OpenAI API](https://github.com/openai/openai-openapi) compatible, HTTP server for serving LLMs. + +-
+ Start a local HTTP server with default configuration on port 8080 + + ```bash + llama-server -m model.gguf --port 8080 + + # Basic web UI can be accessed via browser: http://localhost:8080 + # Chat completion endpoint: http://localhost:8080/v1/chat/completions + ``` + +
+ +-
+ Support multiple-users and parallel decoding + + ```bash + # up to 4 concurrent requests, each with 4096 max context + llama-server -m model.gguf -c 16384 -np 4 + ``` + +
+ +-
+ Enable speculative decoding + + ```bash + # the draft.gguf model should be a small variant of the target model.gguf + llama-server -m model.gguf -md draft.gguf + ``` + +
+ +-
+ Serve an embedding model + + ```bash + # use the /embedding endpoint + llama-server -m model.gguf --embedding --pooling cls -ub 8192 + ``` + +
+ +-
+ Serve a reranking model + + ```bash + # use the /reranking endpoint + llama-server -m model.gguf --reranking + ``` + +
+ +-
+ Constrain all outputs with a grammar + + ```bash + # custom grammar + llama-server -m model.gguf --grammar-file grammar.gbnf + + # JSON + llama-server -m model.gguf --grammar-file grammars/json.gbnf + ``` + +
+ + +## [`llama-perplexity`](examples/perplexity) + +#### A tool for measuring the perplexity [^1][^2] (and other quality metrics) of a model over a given text. + +-
+ Measure the perplexity over a text file + + ```bash + llama-perplexity -m model.gguf -f file.txt + + # [1]15.2701,[2]5.4007,[3]5.3073,[4]6.2965,[5]5.8940,[6]5.6096,[7]5.7942,[8]4.9297, ... + # Final estimate: PPL = 5.4007 +/- 0.67339 + ``` + +
+ +-
+ Measure KL divergence + + ```bash + # TODO + ``` + +
+ +[^1]: [examples/perplexity/README.md](examples/perplexity/README.md) +[^2]: [https://huggingface.co/docs/transformers/perplexity](https://huggingface.co/docs/transformers/perplexity) + +## [`llama-bench`](examples/llama-bench) + +#### Benchmark the performance of the inference for various parameters. + +-
+ Run default benchmark + + ```bash + llama-bench -m model.gguf + + # Output: + # | model | size | params | backend | threads | test | t/s | + # | ------------------- | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: | + # | qwen2 1.5B Q4_0 | 885.97 MiB | 1.54 B | Metal,BLAS | 16 | pp512 | 5765.41 ± 20.55 | + # | qwen2 1.5B Q4_0 | 885.97 MiB | 1.54 B | Metal,BLAS | 16 | tg128 | 197.71 ± 0.81 | + # + # build: 3e0ba0e60 (4229) + ``` + +
+ +## [`llama-run`](examples/run) + +#### A comprehensive example for running `llama.cpp` models. Useful for inferencing. Used with RamaLama [^3]. + +-
+ Run a model with a specific prompt (by default it's pulled from Ollama registry) + + ```bash + llama-run granite-code + ``` + +
+ +[^3]: [RamaLama](https://github.com/containers/ramalama) + +## [`llama-simple`](examples/simple) + +#### A minimal example for implementing apps with `llama.cpp`. Useful for developers. + +-
+ Basic text completion + + ```bash + llama-simple -m model.gguf + + # Hello my name is Kaitlyn and I am a 16 year old girl. I am a junior in high school and I am currently taking a class called "The Art of + ``` + +
-To learn more how to measure perplexity using llama.cpp, [read this documentation](./examples/perplexity/README.md) ## Contributing @@ -458,22 +484,21 @@ To learn more how to measure perplexity using llama.cpp, [read this documentatio - Make sure to read this: [Inference at the edge](https://github.com/ggerganov/llama.cpp/discussions/205) - A bit of backstory for those who are interested: [Changelog podcast](https://changelog.com/podcast/532) -## Other documentations +## Other documentation -- [main (cli)](./examples/main/README.md) -- [server](./examples/server/README.md) -- [jeopardy](./examples/jeopardy/README.md) -- [GBNF grammars](./grammars/README.md) +- [main (cli)](examples/main/README.md) +- [server](examples/server/README.md) +- [GBNF grammars](grammars/README.md) -**Development documentations** +#### Development documentation -- [How to build](./docs/build.md) -- [Running on Docker](./docs/docker.md) -- [Build on Android](./docs/android.md) -- [Performance troubleshooting](./docs/development/token_generation_performance_tips.md) +- [How to build](docs/build.md) +- [Running on Docker](docs/docker.md) +- [Build on Android](docs/android.md) +- [Performance troubleshooting](docs/development/token_generation_performance_tips.md) - [GGML tips & tricks](https://github.com/ggerganov/llama.cpp/wiki/GGML-Tips-&-Tricks) -**Seminal papers and background on the models** +#### Seminal papers and background on the models If your issue is with model generation quality, then please at least scan the following links and papers to understand the limitations of LLaMA models. This is especially important when choosing an appropriate model size and appreciating both the significant and subtle differences between LLaMA models and ChatGPT: - LLaMA: @@ -484,3 +509,6 @@ If your issue is with model generation quality, then please at least scan the fo - GPT-3.5 / InstructGPT / ChatGPT: - [Aligning language models to follow instructions](https://openai.com/research/instruction-following) - [Training language models to follow instructions with human feedback](https://arxiv.org/abs/2203.02155) + +#### References + diff --git a/Sources/llama/llama.h b/Sources/llama/llama.h new file mode 100644 index 000000000..41725880e --- /dev/null +++ b/Sources/llama/llama.h @@ -0,0 +1,4 @@ +#pragma once + +#include + diff --git a/Sources/llama/module.modulemap b/Sources/llama/module.modulemap new file mode 100644 index 000000000..d010555b1 --- /dev/null +++ b/Sources/llama/module.modulemap @@ -0,0 +1,5 @@ +module llama [system] { + header "llama.h" + link "llama" + export * +} diff --git a/ci/run.sh b/ci/run.sh index 21b62dd1e..77c32ce00 100755 --- a/ci/run.sh +++ b/ci/run.sh @@ -39,7 +39,7 @@ SRC=`pwd` CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=ON" if [ ! -z ${GG_BUILD_METAL} ]; then - CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON" + CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON -DGGML_METAL_USE_BF16=ON" fi if [ ! -z ${GG_BUILD_CUDA} ]; then @@ -326,17 +326,17 @@ function gg_run_open_llama_7b_v2 { ./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k ./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k - (time ./bin/llama-cli --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log - (time ./bin/llama-cli --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log - (time ./bin/llama-cli --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log - (time ./bin/llama-cli --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log - (time ./bin/llama-cli --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log - (time ./bin/llama-cli --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log - (time ./bin/llama-cli --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log - (time ./bin/llama-cli --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log - (time ./bin/llama-cli --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log - (time ./bin/llama-cli --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log - (time ./bin/llama-cli --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log + (time ./bin/llama-cli -no-cnv --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log + (time ./bin/llama-cli -no-cnv --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log (time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log (time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log @@ -460,17 +460,17 @@ function gg_run_pythia_1_4b { ./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k ./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k - (time ./bin/llama-cli --model ${model_f16} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log - (time ./bin/llama-cli --model ${model_q8_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log - (time ./bin/llama-cli --model ${model_q4_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log - (time ./bin/llama-cli --model ${model_q4_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log - (time ./bin/llama-cli --model ${model_q5_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log - (time ./bin/llama-cli --model ${model_q5_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log - (time ./bin/llama-cli --model ${model_q2_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log - (time ./bin/llama-cli --model ${model_q3_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log - (time ./bin/llama-cli --model ${model_q4_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log - (time ./bin/llama-cli --model ${model_q5_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log - (time ./bin/llama-cli --model ${model_q6_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log + (time ./bin/llama-cli -no-cnv --model ${model_f16} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log + (time ./bin/llama-cli -no-cnv --model ${model_q8_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_0} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_1} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q2_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q3_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q6_k} -ngl 99 -c 0 -s 1234 -n 64 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log (time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log (time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test_60} -ngl 99 -c 128 -b 128 --chunks 1 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log @@ -591,17 +591,17 @@ function gg_run_pythia_2_8b { ./bin/llama-quantize ${model_f16} ${model_q5_k} q5_k ./bin/llama-quantize ${model_f16} ${model_q6_k} q6_k - (time ./bin/llama-cli --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log - (time ./bin/llama-cli --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log - (time ./bin/llama-cli --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log - (time ./bin/llama-cli --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log - (time ./bin/llama-cli --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log - (time ./bin/llama-cli --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log - (time ./bin/llama-cli --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log - (time ./bin/llama-cli --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log - (time ./bin/llama-cli --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log - (time ./bin/llama-cli --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log - (time ./bin/llama-cli --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log + (time ./bin/llama-cli -no-cnv --model ${model_f16} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log + (time ./bin/llama-cli -no-cnv --model ${model_q8_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_0} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_0.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_1} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_1.log + (time ./bin/llama-cli -no-cnv --model ${model_q2_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q2_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q3_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q3_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q4_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q4_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q5_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q5_k.log + (time ./bin/llama-cli -no-cnv --model ${model_q6_k} -t 1 -ngl 99 -c 0 -s 1234 -n 256 --ignore-eos -p "I believe the meaning of life is" ) 2>&1 | tee -a $OUT/${ci}-tg-q6_k.log (time ./bin/llama-perplexity --model ${model_f16} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-f16.log (time ./bin/llama-perplexity --model ${model_q8_0} -f ${wiki_test} -t 1 -ngl 99 -c 2048 -b 512 --chunks 4 ) 2>&1 | tee -a $OUT/${ci}-tg-q8_0.log @@ -815,7 +815,10 @@ if [ -z ${GG_BUILD_LOW_PERF} ]; then ln -sfn ${mnt_models} ${SRC}/models-mnt # Create a fresh python3 venv and enter it - python3 -m venv "$MNT/venv" + if ! python3 -m venv "$MNT/venv"; then + echo "Error: Failed to create Python virtual environment at $MNT/venv." + exit 1 + fi source "$MNT/venv/bin/activate" pip install -r ${SRC}/requirements.txt --disable-pip-version-check diff --git a/cmake/common.cmake b/cmake/common.cmake new file mode 100644 index 000000000..0f54871e4 --- /dev/null +++ b/cmake/common.cmake @@ -0,0 +1,33 @@ +function(llama_add_compile_flags) + if (LLAMA_FATAL_WARNINGS) + if (CMAKE_CXX_COMPILER_ID MATCHES "GNU" OR CMAKE_CXX_COMPILER_ID MATCHES "Clang") + list(APPEND C_FLAGS -Werror) + list(APPEND CXX_FLAGS -Werror) + elseif (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC") + add_compile_options(/WX) + endif() + endif() + + if (LLAMA_ALL_WARNINGS) + if (NOT MSVC) + list(APPEND C_FLAGS -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes + -Werror=implicit-int -Werror=implicit-function-declaration) + + list(APPEND CXX_FLAGS -Wmissing-declarations -Wmissing-noreturn) + + list(APPEND WARNING_FLAGS -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function) + + list(APPEND C_FLAGS ${WARNING_FLAGS}) + list(APPEND CXX_FLAGS ${WARNING_FLAGS}) + + ggml_get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}) + + add_compile_options("$<$:${C_FLAGS};${GF_C_FLAGS}>" + "$<$:${CXX_FLAGS};${GF_CXX_FLAGS}>") + else() + # todo : msvc + set(C_FLAGS "" PARENT_SCOPE) + set(CXX_FLAGS "" PARENT_SCOPE) + endif() + endif() +endfunction() diff --git a/cmake/llama-config.cmake.in b/cmake/llama-config.cmake.in index f072b76a3..5c55bc6b8 100644 --- a/cmake/llama-config.cmake.in +++ b/cmake/llama-config.cmake.in @@ -3,18 +3,60 @@ set(LLAMA_BUILD_COMMIT @LLAMA_BUILD_COMMIT@) set(LLAMA_BUILD_NUMBER @LLAMA_BUILD_NUMBER@) set(LLAMA_SHARED_LIB @BUILD_SHARED_LIBS@) -set(GGML_BLAS @GGML_BLAS@) -set(GGML_CUDA @GGML_CUDA@) -set(GGML_METAL @GGML_METAL@) -set(GGML_HIPBLAS @GGML_HIPBLAS@) +set(GGML_STATIC @GGML_STATIC@) +set(GGML_NATIVE @GGML_NATIVE@) +set(GGML_LTO @GGML_LTO@) +set(GGML_CCACHE @GGML_CCACHE@) +set(GGML_AVX @GGML_AVX@) +set(GGML_AVX2 @GGML_AVX2@) +set(GGML_AVX512 @GGML_AVX512@) +set(GGML_AVX512_VBMI @GGML_AVX512_VBMI@) +set(GGML_AVX512_VNNI @GGML_AVX512_VNNI@) +set(GGML_AVX512_BF16 @GGML_AVX512_BF16@) +set(GGML_AMX_TILE @GGML_AMX_TILE@) +set(GGML_AMX_INT8 @GGML_AMX_INT8@) +set(GGML_AMX_BF16 @GGML_AMX_BF16@) +set(GGML_FMA @GGML_FMA@) +set(GGML_LASX @GGML_LASX@) +set(GGML_LSX @GGML_LSX@) +set(GGML_RVV @GGML_RVV@) +set(GGML_SVE @GGML_SVE@) + set(GGML_ACCELERATE @GGML_ACCELERATE@) -set(GGML_VULKAN @GGML_VULKAN@) +set(GGML_OPENMP @GGML_OPENMP@) +set(GGML_CPU_HBM @GGML_CPU_HBM@) +set(GGML_BLAS_VENDOR @GGML_BLAS_VENDOR@) + +set(GGML_CUDA_FORCE_MMQ @GGML_CUDA_FORCE_MMQ@) +set(GGML_CUDA_FORCE_CUBLAS @GGML_CUDA_FORCE_CUBLAS@) +set(GGML_CUDA_F16 @GGML_CUDA_F16@) +set(GGML_CUDA_PEER_MAX_BATCH_SIZE @GGML_CUDA_PEER_MAX_BATCH_SIZE@) +set(GGML_CUDA_NO_PEER_COPY @GGML_CUDA_NO_PEER_COPY@) +set(GGML_CUDA_NO_VMM @GGML_CUDA_NO_VMM@) +set(GGML_CUDA_FA_ALL_QUANTS @GGML_CUDA_FA_ALL_QUANTS@) +set(GGML_CUDA_GRAPHS @GGML_CUDA_GRAPHS@) + +set(GGML_HIP_UMA @GGML_HIP_UMA@) + set(GGML_VULKAN_CHECK_RESULTS @GGML_VULKAN_CHECK_RESULTS@) -set(GGML_VULKAN_DEBUG @GGML_VULKAN_DEBUG@) -set(GGML_VULKAN_MEMORY_DEBUG @GGML_VULKAN_MEMORY_DEBUG@) -set(GGML_VULKAN_VALIDATE @GGML_VULKAN_VALIDATE@) -set(GGML_SYCL @GGML_SYCL@) -set(GGML_OPENMP @GGML_OPENMP@) +set(GGML_VULKAN_DEBUG @GGML_VULKAN_DEBUG@) +set(GGML_VULKAN_MEMORY_DEBUG @GGML_VULKAN_MEMORY_DEBUG@) +set(GGML_VULKAN_SHADER_DEBUG_INFO @GGML_VULKAN_SHADER_DEBUG_INFO@) +set(GGML_VULKAN_PERF @GGML_VULKAN_PERF@) +set(GGML_VULKAN_VALIDATE @GGML_VULKAN_VALIDATE@) +set(GGML_VULKAN_RUN_TESTS @GGML_VULKAN_RUN_TESTS@) + +set(GGML_METAL_USE_BF16 @GGML_METAL_USE_BF16@) +set(GGML_METAL_NDEBUG @GGML_METAL_NDEBUG@) +set(GGML_METAL_SHADER_DEBUG @GGML_METAL_SHADER_DEBUG@) +set(GGML_METAL_EMBED_LIBRARY @GGML_METAL_EMBED_LIBRARY@) +set(GGML_METAL_MACOSX_VERSION_MIN @GGML_METAL_MACOSX_VERSION_MIN@) +set(GGML_METAL_STD @GGML_METAL_STD@) + +set(GGML_SYCL_F16 @GGML_SYCL_F16@) +set(GGML_SYCL_TARGET @GGML_SYCL_TARGET@) +set(GGML_SYCL_DEVICE_ARCH @GGML_SYCL_DEVICE_ARCH@) + @PACKAGE_INIT@ @@ -22,65 +64,111 @@ set_and_check(LLAMA_INCLUDE_DIR "@PACKAGE_LLAMA_INCLUDE_INSTALL_DIR@") set_and_check(LLAMA_LIB_DIR "@PACKAGE_LLAMA_LIB_INSTALL_DIR@") set_and_check(LLAMA_BIN_DIR "@PACKAGE_LLAMA_BIN_INSTALL_DIR@") -# Ensure transient dependencies satisfied - find_package(Threads REQUIRED) -if (APPLE AND GGML_ACCELERATE) - find_library(ACCELERATE_FRAMEWORK Accelerate REQUIRED) +set(_llama_transient_defines "@GGML_TRANSIENT_DEFINES@") +set(_llama_link_deps "") +set(_llama_link_opts "") +foreach(_ggml_lib ggml ggml-base) + string(REPLACE "-" "_" _ggml_lib_var "${_ggml_lib}_LIBRARY") + find_library(${_ggml_lib_var} ${_ggml_lib} + REQUIRED + HINTS ${LLAMA_LIB_DIR} + NO_CMAKE_FIND_ROOT_PATH + ) + list(APPEND _llama_link_deps "${${_ggml_lib_var}}") + message(STATUS "Found ${${_ggml_lib_var}}") +endforeach() + +foreach(backend amx blas cann cpu cuda hip kompute metal musa rpc sycl vulkan) + string(TOUPPER "GGML_${backend}" backend_id) + set(_ggml_lib "ggml-${backend}") + string(REPLACE "-" "_" _ggml_lib_var "${_ggml_lib}_LIBRARY") + + find_library(${_ggml_lib_var} ${_ggml_lib} + HINTS ${LLAMA_LIB_DIR} + NO_CMAKE_FIND_ROOT_PATH + ) + if(${_ggml_lib_var}) + list(APPEND _llama_link_deps "${${_ggml_lib_var}}") + set(${backend_id} ON) + message(STATUS "Found backend ${${_ggml_lib_var}}") + else() + set(${backend_id} OFF) + endif() +endforeach() + +if (NOT LLAMA_SHARED_LIB) + if (APPLE AND GGML_ACCELERATE) + find_library(ACCELERATE_FRAMEWORK Accelerate REQUIRED) + list(APPEND _llama_link_deps ${ACCELERATE_FRAMEWORK}) + endif() + + if (GGML_OPENMP) + find_package(OpenMP REQUIRED) + list(APPEND _llama_link_deps OpenMP::OpenMP_C OpenMP::OpenMP_CXX) + endif() + + if (GGML_CPU_HBM) + find_library(memkind memkind REQUIRED) + list(APPEND _llama_link_deps memkind) + endif() + + if (GGML_BLAS) + find_package(BLAS REQUIRED) + list(APPEND _llama_link_deps ${BLAS_LIBRARIES}) + list(APPEND _llama_link_opts ${BLAS_LINKER_FLAGS}) + endif() + + if (GGML_CUDA) + find_package(CUDAToolkit REQUIRED) + endif() + + if (GGML_METAL) + find_library(FOUNDATION_LIBRARY Foundation REQUIRED) + find_library(METAL_FRAMEWORK Metal REQUIRED) + find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) + list(APPEND _llama_link_deps ${FOUNDATION_LIBRARY} + ${METAL_FRAMEWORK} ${METALKIT_FRAMEWORK}) + endif() + + if (GGML_VULKAN) + find_package(Vulkan REQUIRED) + list(APPEND _llama_link_deps Vulkan::Vulkan) + endif() + + if (GGML_HIP) + find_package(hip REQUIRED) + find_package(hipblas REQUIRED) + find_package(rocblas REQUIRED) + list(APPEND _llama_link_deps hip::host roc::rocblas roc::hipblas) + endif() + + if (GGML_SYCL) + find_package(DNNL) + if (${DNNL_FOUND} AND GGML_SYCL_TARGET STREQUAL "INTEL") + list(APPEND _llama_link_deps DNNL::dnnl) + endif() + if (WIN32) + find_package(IntelSYCL REQUIRED) + find_package(MKL REQUIRED) + list(APPEND _llama_link_deps IntelSYCL::SYCL_CXX MKL::MKL MKL::MKL_SYCL) + endif() + endif() endif() -if (GGML_BLAS) - find_package(BLAS REQUIRED) -endif() - -if (GGML_CUDA) - find_package(CUDAToolkit REQUIRED) -endif() - -if (GGML_METAL) - find_library(FOUNDATION_LIBRARY Foundation REQUIRED) - find_library(METAL_FRAMEWORK Metal REQUIRED) - find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) -endif() - -if (GGML_VULKAN) - find_package(Vulkan REQUIRED) -endif() - -if (GGML_HIPBLAS) - find_package(hip REQUIRED) - find_package(hipblas REQUIRED) - find_package(rocblas REQUIRED) -endif() - -if (GGML_SYCL) - find_package(IntelSYCL REQUIRED) - find_package(MKL REQUIRED) -endif() - -if (GGML_OPENMP) - find_package(OpenMP REQUIRED) -endif() - - -find_library(ggml_LIBRARY ggml - REQUIRED - HINTS ${LLAMA_LIB_DIR}) - find_library(llama_LIBRARY llama REQUIRED - HINTS ${LLAMA_LIB_DIR}) - -set(_llama_link_deps "${ggml_LIBRARY}" "@GGML_LINK_LIBRARIES@") -set(_llama_transient_defines "@GGML_TRANSIENT_DEFINES@") + HINTS ${LLAMA_LIB_DIR} + NO_CMAKE_FIND_ROOT_PATH +) add_library(llama UNKNOWN IMPORTED) - set_target_properties(llama PROPERTIES INTERFACE_INCLUDE_DIRECTORIES "${LLAMA_INCLUDE_DIR}" INTERFACE_LINK_LIBRARIES "${_llama_link_deps}" + INTERFACE_LINK_OPTIONS "${_llama_link_opts}" INTERFACE_COMPILE_DEFINITIONS "${_llama_transient_defines}" IMPORTED_LINK_INTERFACE_LANGUAGES "CXX" IMPORTED_LOCATION "${llama_LIBRARY}" diff --git a/cmake/llama.pc.in b/cmake/llama.pc.in index 326acbb61..0b2b6bcfa 100644 --- a/cmake/llama.pc.in +++ b/cmake/llama.pc.in @@ -6,5 +6,5 @@ includedir=${prefix}/include Name: llama Description: Port of Facebook's LLaMA model in C/C++ Version: @PROJECT_VERSION@ -Libs: -L${libdir} -lllama +Libs: -L${libdir} -lggml -lggml-base -lllama Cflags: -I${includedir} diff --git a/cmake/x64-windows-llvm.cmake b/cmake/x64-windows-llvm.cmake new file mode 100644 index 000000000..0603d738f --- /dev/null +++ b/cmake/x64-windows-llvm.cmake @@ -0,0 +1,11 @@ +set( CMAKE_SYSTEM_NAME Windows ) +set( CMAKE_SYSTEM_PROCESSOR x86_64 ) + +set( CMAKE_C_COMPILER clang ) +set( CMAKE_CXX_COMPILER clang++ ) + +set( arch_c_flags "-march=native" ) + +set( CMAKE_C_FLAGS_INIT "${arch_c_flags}" ) +set( CMAKE_CXX_FLAGS_INIT "${arch_c_flags}" ) + diff --git a/common/CMakeLists.txt b/common/CMakeLists.txt index 5ab1ffa19..df1cdf9a5 100644 --- a/common/CMakeLists.txt +++ b/common/CMakeLists.txt @@ -2,6 +2,8 @@ find_package(Threads REQUIRED) +llama_add_compile_flags() + # Build info header # @@ -66,6 +68,8 @@ add_library(${TARGET} STATIC ngram-cache.h sampling.cpp sampling.h + speculative.cpp + speculative.h ) if (BUILD_SHARED_LIBS) @@ -77,12 +81,12 @@ set(LLAMA_COMMON_EXTRA_LIBS build_info) # Use curl to download model url if (LLAMA_CURL) find_package(CURL REQUIRED) - add_definitions(-DLLAMA_USE_CURL) + target_compile_definitions(${TARGET} PUBLIC LLAMA_USE_CURL) include_directories(${CURL_INCLUDE_DIRS}) find_library(CURL_LIBRARY curl REQUIRED) set(LLAMA_COMMON_EXTRA_LIBS ${LLAMA_COMMON_EXTRA_LIBS} ${CURL_LIBRARY}) endif () target_include_directories(${TARGET} PUBLIC .) -target_compile_features (${TARGET} PUBLIC cxx_std_11) +target_compile_features (${TARGET} PUBLIC cxx_std_17) target_link_libraries (${TARGET} PRIVATE ${LLAMA_COMMON_EXTRA_LIBS} PUBLIC llama Threads::Threads) diff --git a/common/arg.cpp b/common/arg.cpp index 7c5c5e5cd..dd10b6352 100644 --- a/common/arg.cpp +++ b/common/arg.cpp @@ -22,6 +22,11 @@ common_arg & common_arg::set_examples(std::initializer_list return *this; } +common_arg & common_arg::set_excludes(std::initializer_list excludes) { + this->excludes = std::move(excludes); + return *this; +} + common_arg & common_arg::set_env(const char * env) { help = help + "\n(env: " + env + ")"; this->env = env; @@ -37,6 +42,10 @@ bool common_arg::in_example(enum llama_example ex) { return examples.find(ex) != examples.end(); } +bool common_arg::is_exclude(enum llama_example ex) { + return excludes.find(ex) != excludes.end(); +} + bool common_arg::get_value_from_env(std::string & output) { if (env == nullptr) return false; char * value = std::getenv(env); @@ -119,28 +128,74 @@ std::string common_arg::to_string() { // utils // -static void common_params_handle_model_default(common_params & params) { - if (!params.hf_repo.empty()) { +static void common_params_handle_model_default( + std::string & model, + const std::string & model_url, + std::string & hf_repo, + std::string & hf_file, + const std::string & hf_token) { + if (!hf_repo.empty()) { // short-hand to avoid specifying --hf-file -> default it to --model - if (params.hf_file.empty()) { - if (params.model.empty()) { - throw std::invalid_argument("error: --hf-repo requires either --hf-file or --model\n"); + if (hf_file.empty()) { + if (model.empty()) { + auto auto_detected = common_get_hf_file(hf_repo, hf_token); + if (auto_detected.first.empty() || auto_detected.second.empty()) { + exit(1); // built without CURL, error message already printed + } + hf_repo = auto_detected.first; + hf_file = auto_detected.second; + } else { + hf_file = model; } - params.hf_file = params.model; - } else if (params.model.empty()) { - params.model = fs_get_cache_file(string_split(params.hf_file, '/').back()); } - } else if (!params.model_url.empty()) { - if (params.model.empty()) { - auto f = string_split(params.model_url, '#').front(); + // make sure model path is present (for caching purposes) + if (model.empty()) { + // this is to avoid different repo having same file name, or same file name in different subdirs + std::string filename = hf_repo + "_" + hf_file; + // to make sure we don't have any slashes in the filename + string_replace_all(filename, "/", "_"); + model = fs_get_cache_file(filename); + } + } else if (!model_url.empty()) { + if (model.empty()) { + auto f = string_split(model_url, '#').front(); f = string_split(f, '?').front(); - params.model = fs_get_cache_file(string_split(f, '/').back()); + model = fs_get_cache_file(string_split(f, '/').back()); } - } else if (params.model.empty()) { - params.model = DEFAULT_MODEL_PATH; + } else if (model.empty()) { + model = DEFAULT_MODEL_PATH; } } +const std::vector kv_cache_types = { + GGML_TYPE_F32, + GGML_TYPE_F16, + GGML_TYPE_BF16, + GGML_TYPE_Q8_0, + GGML_TYPE_Q4_0, + GGML_TYPE_Q4_1, + GGML_TYPE_IQ4_NL, + GGML_TYPE_Q5_0, + GGML_TYPE_Q5_1, +}; + +static ggml_type kv_cache_type_from_str(const std::string & s) { + for (const auto & type : kv_cache_types) { + if (ggml_type_name(type) == s) { + return type; + } + } + throw std::runtime_error("Unsupported cache type: " + s); +} + +static std::string get_all_kv_cache_types() { + std::ostringstream msg; + for (const auto & type : kv_cache_types) { + msg << ggml_type_name(type) << (&type == &kv_cache_types.back() ? "" : ", "); + } + return msg.str(); +} + // // CLI argument parsing functions // @@ -233,16 +288,19 @@ static bool common_params_parse_ex(int argc, char ** argv, common_params_context } } - postprocess_cpu_params(params.cpuparams, nullptr); + postprocess_cpu_params(params.cpuparams, nullptr); postprocess_cpu_params(params.cpuparams_batch, ¶ms.cpuparams); - postprocess_cpu_params(params.draft_cpuparams, ¶ms.cpuparams); - postprocess_cpu_params(params.draft_cpuparams_batch, ¶ms.cpuparams_batch); + + postprocess_cpu_params(params.speculative.cpuparams, ¶ms.cpuparams); + postprocess_cpu_params(params.speculative.cpuparams_batch, ¶ms.cpuparams_batch); if (params.prompt_cache_all && (params.interactive || params.interactive_first)) { throw std::invalid_argument("error: --prompt-cache-all not supported in interactive mode yet\n"); } - common_params_handle_model_default(params); + // TODO: refactor model params in a common struct + common_params_handle_model_default(params.model, params.model_url, params.hf_repo, params.hf_file, params.hf_token); + common_params_handle_model_default(params.vocoder.model, params.vocoder.model_url, params.vocoder.hf_repo, params.vocoder.hf_file, params.hf_token); if (params.escape) { string_process_escapes(params.prompt); @@ -251,7 +309,7 @@ static bool common_params_parse_ex(int argc, char ** argv, common_params_context for (auto & antiprompt : params.antiprompt) { string_process_escapes(antiprompt); } - for (auto & seq_breaker : params.sparams.dry_sequence_breakers) { + for (auto & seq_breaker : params.sampling.dry_sequence_breakers) { string_process_escapes(seq_breaker); } } @@ -297,6 +355,27 @@ static void common_params_print_usage(common_params_context & ctx_arg) { print_options(specific_options); } +static std::vector parse_device_list(const std::string & value) { + std::vector devices; + auto dev_names = string_split(value, ','); + if (dev_names.empty()) { + throw std::invalid_argument("no devices specified"); + } + if (dev_names.size() == 1 && dev_names[0] == "none") { + devices.push_back(nullptr); + } else { + for (const auto & device : dev_names) { + auto * dev = ggml_backend_dev_by_name(device.c_str()); + if (!dev || ggml_backend_dev_type(dev) != GGML_BACKEND_DEVICE_TYPE_GPU) { + throw std::invalid_argument(string_format("invalid device: %s", device.c_str())); + } + devices.push_back(dev); + } + devices.push_back(nullptr); + } + return devices; +} + bool common_params_parse(int argc, char ** argv, common_params & params, llama_example ex, void(*print_usage)(int, char **)) { auto ctx_arg = common_params_parser_init(params, ex, print_usage); const common_params params_org = ctx_arg.params; // the example can modify the default params @@ -322,14 +401,29 @@ bool common_params_parse(int argc, char ** argv, common_params & params, llama_e return true; } +static std::string list_builtin_chat_templates() { + std::vector supported_tmpl; + int32_t res = llama_chat_builtin_templates(nullptr, 0); + supported_tmpl.resize(res); + res = llama_chat_builtin_templates(supported_tmpl.data(), supported_tmpl.size()); + std::ostringstream msg; + for (auto & tmpl : supported_tmpl) { + msg << tmpl << (&tmpl == &supported_tmpl.back() ? "" : ", "); + } + return msg.str(); +} + common_params_context common_params_parser_init(common_params & params, llama_example ex, void(*print_usage)(int, char **)) { + // load dynamic backends + ggml_backend_load_all(); + common_params_context ctx_arg(params); ctx_arg.print_usage = print_usage; ctx_arg.ex = ex; std::string sampler_type_chars; std::string sampler_type_names; - for (const auto & sampler : params.sparams.samplers) { + for (const auto & sampler : params.sampling.samplers) { sampler_type_chars += common_sampler_type_to_chr(sampler); sampler_type_names += common_sampler_type_to_str(sampler) + ";"; } @@ -344,7 +438,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex * - if both {LLAMA_EXAMPLE_COMMON, LLAMA_EXAMPLE_*,} are set, we will prioritize the LLAMA_EXAMPLE_* matching current example */ auto add_opt = [&](common_arg arg) { - if (arg.in_example(ex) || arg.in_example(LLAMA_EXAMPLE_COMMON)) { + if ((arg.in_example(ex) || arg.in_example(LLAMA_EXAMPLE_COMMON)) && !arg.is_exclude(ex)) { ctx_arg.options.push_back(std::move(arg)); } }; @@ -407,26 +501,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex } } )); - add_opt(common_arg( - {"-td", "--threads-draft"}, "N", - "number of threads to use during generation (default: same as --threads)", - [](common_params & params, int value) { - params.draft_cpuparams.n_threads = value; - if (params.draft_cpuparams.n_threads <= 0) { - params.draft_cpuparams.n_threads = std::thread::hardware_concurrency(); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"-tbd", "--threads-batch-draft"}, "N", - "number of threads to use during batch and prompt processing (default: same as --threads-draft)", - [](common_params & params, int value) { - params.draft_cpuparams_batch.n_threads = value; - if (params.draft_cpuparams_batch.n_threads <= 0) { - params.draft_cpuparams_batch.n_threads = std::thread::hardware_concurrency(); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); add_opt(common_arg( {"-C", "--cpu-mask"}, "M", "CPU affinity mask: arbitrarily long hex. Complements cpu-range (default: \"\")", @@ -515,108 +589,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.cpuparams_batch.poll = value; } )); - add_opt(common_arg( - {"-Cd", "--cpu-mask-draft"}, "M", - "Draft model CPU affinity mask. Complements cpu-range-draft (default: same as --cpu-mask)", - [](common_params & params, const std::string & mask) { - params.draft_cpuparams.mask_valid = true; - if (!parse_cpu_mask(mask, params.draft_cpuparams.cpumask)) { - throw std::invalid_argument("invalid cpumask"); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"-Crd", "--cpu-range-draft"}, "lo-hi", - "Ranges of CPUs for affinity. Complements --cpu-mask-draft", - [](common_params & params, const std::string & range) { - params.draft_cpuparams.mask_valid = true; - if (!parse_cpu_range(range, params.draft_cpuparams.cpumask)) { - throw std::invalid_argument("invalid range"); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--cpu-strict-draft"}, "<0|1>", - "Use strict CPU placement for draft model (default: same as --cpu-strict)", - [](common_params & params, int value) { - params.draft_cpuparams.strict_cpu = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--prio-draft"}, "N", - string_format("set draft process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: %d)\n", params.draft_cpuparams.priority), - [](common_params & params, int prio) { - if (prio < 0 || prio > 3) { - throw std::invalid_argument("invalid value"); - } - params.draft_cpuparams.priority = (enum ggml_sched_priority) prio; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--poll-draft"}, "<0|1>", - "Use polling to wait for draft model work (default: same as --poll])", - [](common_params & params, int value) { - params.draft_cpuparams.poll = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"-Cbd", "--cpu-mask-batch-draft"}, "M", - "Draft model CPU affinity mask. Complements cpu-range-draft (default: same as --cpu-mask)", - [](common_params & params, const std::string & mask) { - params.draft_cpuparams_batch.mask_valid = true; - if (!parse_cpu_mask(mask, params.draft_cpuparams_batch.cpumask)) { - throw std::invalid_argument("invalid cpumask"); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"-Crbd", "--cpu-range-batch-draft"}, "lo-hi", - "Ranges of CPUs for affinity. Complements --cpu-mask-draft-batch)", - [](common_params & params, const std::string & range) { - params.draft_cpuparams_batch.mask_valid = true; - if (!parse_cpu_range(range, params.draft_cpuparams_batch.cpumask)) { - throw std::invalid_argument("invalid cpumask"); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--cpu-strict-batch-draft"}, "<0|1>", - "Use strict CPU placement for draft model (default: --cpu-strict-draft)", - [](common_params & params, int value) { - params.draft_cpuparams_batch.strict_cpu = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--prio-batch-draft"}, "N", - string_format("set draft process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: %d)\n", params.draft_cpuparams_batch.priority), - [](common_params & params, int prio) { - if (prio < 0 || prio > 3) { - throw std::invalid_argument("invalid value"); - } - params.draft_cpuparams_batch.priority = (enum ggml_sched_priority) prio; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--poll-batch-draft"}, "<0|1>", - "Use polling to wait for draft model work (default: --poll-draft)", - [](common_params & params, int value) { - params.draft_cpuparams_batch.poll = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); - add_opt(common_arg( - {"--draft"}, "N", - string_format("number of tokens to draft for speculative decoding (default: %d)", params.n_draft), - [](common_params & params, int value) { - params.n_draft = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_LOOKUP})); - add_opt(common_arg( - {"-ps", "--p-split"}, "N", - string_format("speculative decoding split probability (default: %.1f)", (double)params.p_split), - [](common_params & params, const std::string & value) { - params.p_split = std::stof(value); - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); add_opt(common_arg( {"-lcs", "--lookup-cache-static"}, "FNAME", "path to static lookup cache to use for lookup decoding (not updated by generation)", @@ -672,7 +644,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex [](common_params & params) { params.ctx_shift = false; } - ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_NO_CONTEXT_SHIFT")); + ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SERVER, LLAMA_EXAMPLE_IMATRIX, LLAMA_EXAMPLE_PERPLEXITY}).set_env("LLAMA_ARG_NO_CONTEXT_SHIFT")); add_opt(common_arg( {"--chunks"}, "N", string_format("max number of chunks to process (default: %d, -1 = all)", params.n_chunks), @@ -695,13 +667,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex [](common_params & params, const std::string & value) { params.prompt = value; } - )); + ).set_excludes({LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"--no-perf"}, string_format("disable internal libllama performance timings (default: %s)", params.no_perf ? "true" : "false"), [](common_params & params) { params.no_perf = true; - params.sparams.no_perf = true; + params.sampling.no_perf = true; } ).set_env("LLAMA_ARG_NO_PERF")); add_opt(common_arg( @@ -719,7 +691,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.prompt.pop_back(); } } - )); + ).set_excludes({LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"--in-file"}, "FNAME", "an input file (repeat to specify multiple files)", @@ -746,7 +718,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.prompt = ss.str(); fprintf(stderr, "Read %zu bytes from binary file %s\n", params.prompt.size(), value.c_str()); } - )); + ).set_excludes({LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"-e", "--escape"}, string_format("process escapes sequences (\\n, \\r, \\t, \\', \\\", \\\\) (default: %s)", params.escape ? "true" : "false"), @@ -805,15 +777,19 @@ common_params_context common_params_parser_init(common_params & params, llama_ex ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"-cnv", "--conversation"}, - string_format( - "run in conversation mode:\n" - "- does not print special tokens and suffix/prefix\n" - "- interactive mode is also enabled\n" - "(default: %s)", - params.conversation ? "true" : "false" - ), + "run in conversation mode:\n" + "- does not print special tokens and suffix/prefix\n" + "- interactive mode is also enabled\n" + "(default: auto enabled if chat template is available)", [](common_params & params) { - params.conversation = true; + params.conversation_mode = COMMON_CONVERSATION_MODE_ENABLED; + } + ).set_examples({LLAMA_EXAMPLE_MAIN})); + add_opt(common_arg( + {"-no-cnv", "--no-conversation"}, + "force disable conversation mode (default: false)", + [](common_params & params) { + params.conversation_mode = COMMON_CONVERSATION_MODE_DISABLED; } ).set_examples({LLAMA_EXAMPLE_MAIN})); add_opt(common_arg( @@ -867,7 +843,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex [](common_params & params) { params.warmup = false; } - ).set_examples({LLAMA_EXAMPLE_MAIN})); + ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"--spm-infill"}, string_format( @@ -883,155 +859,154 @@ common_params_context common_params_parser_init(common_params & params, llama_ex string_format("samplers that will be used for generation in the order, separated by \';\'\n(default: %s)", sampler_type_names.c_str()), [](common_params & params, const std::string & value) { const auto sampler_names = string_split(value, ';'); - params.sparams.samplers = common_sampler_types_from_names(sampler_names, true); + params.sampling.samplers = common_sampler_types_from_names(sampler_names, true); } ).set_sparam()); add_opt(common_arg( {"-s", "--seed"}, "SEED", - string_format("RNG seed (default: %d, use random seed for %d)", params.sparams.seed, LLAMA_DEFAULT_SEED), + string_format("RNG seed (default: %d, use random seed for %d)", params.sampling.seed, LLAMA_DEFAULT_SEED), [](common_params & params, const std::string & value) { - params.sparams.seed = std::stoul(value); + params.sampling.seed = std::stoul(value); } ).set_sparam()); add_opt(common_arg( - {"--sampling-seq"}, "SEQUENCE", + {"--sampling-seq", "--sampler-seq"}, "SEQUENCE", string_format("simplified sequence for samplers that will be used (default: %s)", sampler_type_chars.c_str()), [](common_params & params, const std::string & value) { - params.sparams.samplers = common_sampler_types_from_chars(value); + params.sampling.samplers = common_sampler_types_from_chars(value); } ).set_sparam()); add_opt(common_arg( {"--ignore-eos"}, "ignore end of stream token and continue generating (implies --logit-bias EOS-inf)", [](common_params & params) { - params.sparams.ignore_eos = true; - } - ).set_sparam()); - add_opt(common_arg( - {"--penalize-nl"}, - string_format("penalize newline tokens (default: %s)", params.sparams.penalize_nl ? "true" : "false"), - [](common_params & params) { - params.sparams.penalize_nl = true; + params.sampling.ignore_eos = true; } ).set_sparam()); add_opt(common_arg( {"--temp"}, "N", - string_format("temperature (default: %.1f)", (double)params.sparams.temp), + string_format("temperature (default: %.1f)", (double)params.sampling.temp), [](common_params & params, const std::string & value) { - params.sparams.temp = std::stof(value); - params.sparams.temp = std::max(params.sparams.temp, 0.0f); + params.sampling.temp = std::stof(value); + params.sampling.temp = std::max(params.sampling.temp, 0.0f); } ).set_sparam()); add_opt(common_arg( {"--top-k"}, "N", - string_format("top-k sampling (default: %d, 0 = disabled)", params.sparams.top_k), + string_format("top-k sampling (default: %d, 0 = disabled)", params.sampling.top_k), [](common_params & params, int value) { - params.sparams.top_k = value; + params.sampling.top_k = value; } ).set_sparam()); add_opt(common_arg( {"--top-p"}, "N", - string_format("top-p sampling (default: %.1f, 1.0 = disabled)", (double)params.sparams.top_p), + string_format("top-p sampling (default: %.1f, 1.0 = disabled)", (double)params.sampling.top_p), [](common_params & params, const std::string & value) { - params.sparams.top_p = std::stof(value); + params.sampling.top_p = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--min-p"}, "N", - string_format("min-p sampling (default: %.1f, 0.0 = disabled)", (double)params.sparams.min_p), + string_format("min-p sampling (default: %.1f, 0.0 = disabled)", (double)params.sampling.min_p), [](common_params & params, const std::string & value) { - params.sparams.min_p = std::stof(value); + params.sampling.min_p = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--xtc-probability"}, "N", - string_format("xtc probability (default: %.1f, 0.0 = disabled)", (double)params.sparams.xtc_probability), + string_format("xtc probability (default: %.1f, 0.0 = disabled)", (double)params.sampling.xtc_probability), [](common_params & params, const std::string & value) { - params.sparams.xtc_probability = std::stof(value); + params.sampling.xtc_probability = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--xtc-threshold"}, "N", - string_format("xtc threshold (default: %.1f, 1.0 = disabled)", (double)params.sparams.xtc_threshold), + string_format("xtc threshold (default: %.1f, 1.0 = disabled)", (double)params.sampling.xtc_threshold), [](common_params & params, const std::string & value) { - params.sparams.xtc_threshold = std::stof(value); + params.sampling.xtc_threshold = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--typical"}, "N", - string_format("locally typical sampling, parameter p (default: %.1f, 1.0 = disabled)", (double)params.sparams.typ_p), + string_format("locally typical sampling, parameter p (default: %.1f, 1.0 = disabled)", (double)params.sampling.typ_p), [](common_params & params, const std::string & value) { - params.sparams.typ_p = std::stof(value); + params.sampling.typ_p = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--repeat-last-n"}, "N", - string_format("last n tokens to consider for penalize (default: %d, 0 = disabled, -1 = ctx_size)", params.sparams.penalty_last_n), + string_format("last n tokens to consider for penalize (default: %d, 0 = disabled, -1 = ctx_size)", params.sampling.penalty_last_n), [](common_params & params, int value) { - params.sparams.penalty_last_n = value; - params.sparams.n_prev = std::max(params.sparams.n_prev, params.sparams.penalty_last_n); + if (value < -1) { + throw std::runtime_error(string_format("error: invalid repeat-last-n = %d\n", value)); + } + params.sampling.penalty_last_n = value; + params.sampling.n_prev = std::max(params.sampling.n_prev, params.sampling.penalty_last_n); } ).set_sparam()); add_opt(common_arg( {"--repeat-penalty"}, "N", - string_format("penalize repeat sequence of tokens (default: %.1f, 1.0 = disabled)", (double)params.sparams.penalty_repeat), + string_format("penalize repeat sequence of tokens (default: %.1f, 1.0 = disabled)", (double)params.sampling.penalty_repeat), [](common_params & params, const std::string & value) { - params.sparams.penalty_repeat = std::stof(value); + params.sampling.penalty_repeat = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--presence-penalty"}, "N", - string_format("repeat alpha presence penalty (default: %.1f, 0.0 = disabled)", (double)params.sparams.penalty_present), + string_format("repeat alpha presence penalty (default: %.1f, 0.0 = disabled)", (double)params.sampling.penalty_present), [](common_params & params, const std::string & value) { - params.sparams.penalty_present = std::stof(value); + params.sampling.penalty_present = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--frequency-penalty"}, "N", - string_format("repeat alpha frequency penalty (default: %.1f, 0.0 = disabled)", (double)params.sparams.penalty_freq), + string_format("repeat alpha frequency penalty (default: %.1f, 0.0 = disabled)", (double)params.sampling.penalty_freq), [](common_params & params, const std::string & value) { - params.sparams.penalty_freq = std::stof(value); + params.sampling.penalty_freq = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--dry-multiplier"}, "N", - string_format("set DRY sampling multiplier (default: %.1f, 0.0 = disabled)", (double)params.sparams.dry_multiplier), + string_format("set DRY sampling multiplier (default: %.1f, 0.0 = disabled)", (double)params.sampling.dry_multiplier), [](common_params & params, const std::string & value) { - params.sparams.dry_multiplier = std::stof(value); + params.sampling.dry_multiplier = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--dry-base"}, "N", - string_format("set DRY sampling base value (default: %.2f)", (double)params.sparams.dry_base), + string_format("set DRY sampling base value (default: %.2f)", (double)params.sampling.dry_base), [](common_params & params, const std::string & value) { float potential_base = std::stof(value); if (potential_base >= 1.0f) { - params.sparams.dry_base = potential_base; + params.sampling.dry_base = potential_base; } } ).set_sparam()); add_opt(common_arg( {"--dry-allowed-length"}, "N", - string_format("set allowed length for DRY sampling (default: %d)", params.sparams.dry_allowed_length), + string_format("set allowed length for DRY sampling (default: %d)", params.sampling.dry_allowed_length), [](common_params & params, int value) { - params.sparams.dry_allowed_length = value; + params.sampling.dry_allowed_length = value; } ).set_sparam()); add_opt(common_arg( {"--dry-penalty-last-n"}, "N", - string_format("set DRY penalty for the last n tokens (default: %d, 0 = disable, -1 = context size)", params.sparams.dry_penalty_last_n), + string_format("set DRY penalty for the last n tokens (default: %d, 0 = disable, -1 = context size)", params.sampling.dry_penalty_last_n), [](common_params & params, int value) { - params.sparams.dry_penalty_last_n = value; + if (value < -1) { + throw std::runtime_error(string_format("error: invalid dry-penalty-last-n = %d\n", value)); + } + params.sampling.dry_penalty_last_n = value; } ).set_sparam()); add_opt(common_arg( {"--dry-sequence-breaker"}, "STRING", string_format("add sequence breaker for DRY sampling, clearing out default breakers (%s) in the process; use \"none\" to not use any sequence breakers\n", - params.sparams.dry_sequence_breakers.empty() ? "none" : - std::accumulate(std::next(params.sparams.dry_sequence_breakers.begin()), - params.sparams.dry_sequence_breakers.end(), - std::string("'") + (params.sparams.dry_sequence_breakers[0] == "\n" ? "\\n" : params.sparams.dry_sequence_breakers[0]) + "'", + params.sampling.dry_sequence_breakers.empty() ? "none" : + std::accumulate(std::next(params.sampling.dry_sequence_breakers.begin()), + params.sampling.dry_sequence_breakers.end(), + std::string("'") + (params.sampling.dry_sequence_breakers[0] == "\n" ? "\\n" : params.sampling.dry_sequence_breakers[0]) + "'", [](const std::string& a, const std::string& b) { std::string formatted_b = (b == "\n") ? "\\n" : b; return a + ", '" + formatted_b + "'"; @@ -1040,51 +1015,51 @@ common_params_context common_params_parser_init(common_params & params, llama_ex static bool defaults_cleared = false; if (!defaults_cleared) { - params.sparams.dry_sequence_breakers.clear(); + params.sampling.dry_sequence_breakers.clear(); defaults_cleared = true; } if (value == "none") { - params.sparams.dry_sequence_breakers.clear(); + params.sampling.dry_sequence_breakers.clear(); } else { - params.sparams.dry_sequence_breakers.emplace_back(value); + params.sampling.dry_sequence_breakers.emplace_back(value); } } ).set_sparam()); add_opt(common_arg( {"--dynatemp-range"}, "N", - string_format("dynamic temperature range (default: %.1f, 0.0 = disabled)", (double)params.sparams.dynatemp_range), + string_format("dynamic temperature range (default: %.1f, 0.0 = disabled)", (double)params.sampling.dynatemp_range), [](common_params & params, const std::string & value) { - params.sparams.dynatemp_range = std::stof(value); + params.sampling.dynatemp_range = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--dynatemp-exp"}, "N", - string_format("dynamic temperature exponent (default: %.1f)", (double)params.sparams.dynatemp_exponent), + string_format("dynamic temperature exponent (default: %.1f)", (double)params.sampling.dynatemp_exponent), [](common_params & params, const std::string & value) { - params.sparams.dynatemp_exponent = std::stof(value); + params.sampling.dynatemp_exponent = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--mirostat"}, "N", string_format("use Mirostat sampling.\nTop K, Nucleus and Locally Typical samplers are ignored if used.\n" - "(default: %d, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)", params.sparams.mirostat), + "(default: %d, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)", params.sampling.mirostat), [](common_params & params, int value) { - params.sparams.mirostat = value; + params.sampling.mirostat = value; } ).set_sparam()); add_opt(common_arg( {"--mirostat-lr"}, "N", - string_format("Mirostat learning rate, parameter eta (default: %.1f)", (double)params.sparams.mirostat_eta), + string_format("Mirostat learning rate, parameter eta (default: %.1f)", (double)params.sampling.mirostat_eta), [](common_params & params, const std::string & value) { - params.sparams.mirostat_eta = std::stof(value); + params.sampling.mirostat_eta = std::stof(value); } ).set_sparam()); add_opt(common_arg( {"--mirostat-ent"}, "N", - string_format("Mirostat target entropy, parameter tau (default: %.1f)", (double)params.sparams.mirostat_tau), + string_format("Mirostat target entropy, parameter tau (default: %.1f)", (double)params.sampling.mirostat_tau), [](common_params & params, const std::string & value) { - params.sparams.mirostat_tau = std::stof(value); + params.sampling.mirostat_tau = std::stof(value); } ).set_sparam()); add_opt(common_arg( @@ -1100,7 +1075,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex try { if (ss >> key && ss >> sign && std::getline(ss, value_str) && (sign == '+' || sign == '-')) { const float bias = std::stof(value_str) * ((sign == '-') ? -1.0f : 1.0f); - params.sparams.logit_bias.push_back({key, bias}); + params.sampling.logit_bias.push_back({key, bias}); } else { throw std::invalid_argument("invalid input format"); } @@ -1111,9 +1086,9 @@ common_params_context common_params_parser_init(common_params & params, llama_ex ).set_sparam()); add_opt(common_arg( {"--grammar"}, "GRAMMAR", - string_format("BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '%s')", params.sparams.grammar.c_str()), + string_format("BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '%s')", params.sampling.grammar.c_str()), [](common_params & params, const std::string & value) { - params.sparams.grammar = value; + params.sampling.grammar = value; } ).set_sparam()); add_opt(common_arg( @@ -1127,7 +1102,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex std::copy( std::istreambuf_iterator(file), std::istreambuf_iterator(), - std::back_inserter(params.sparams.grammar) + std::back_inserter(params.sampling.grammar) ); } ).set_sparam()); @@ -1135,7 +1110,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex {"-j", "--json-schema"}, "SCHEMA", "JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object\nFor schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead", [](common_params & params, const std::string & value) { - params.sparams.grammar = json_schema_to_grammar(json::parse(value)); + params.sampling.grammar = json_schema_to_grammar(json::parse(value)); } ).set_sparam()); add_opt(common_arg( @@ -1255,18 +1230,28 @@ common_params_context common_params_parser_init(common_params & params, llama_ex ).set_env("LLAMA_ARG_NO_KV_OFFLOAD")); add_opt(common_arg( {"-ctk", "--cache-type-k"}, "TYPE", - string_format("KV cache data type for K (default: %s)", params.cache_type_k.c_str()), + string_format( + "KV cache data type for K\n" + "allowed values: %s\n" + "(default: %s)", + get_all_kv_cache_types().c_str(), + ggml_type_name(params.cache_type_k) + ), [](common_params & params, const std::string & value) { - // TODO: get the type right here - params.cache_type_k = value; + params.cache_type_k = kv_cache_type_from_str(value); } ).set_env("LLAMA_ARG_CACHE_TYPE_K")); add_opt(common_arg( {"-ctv", "--cache-type-v"}, "TYPE", - string_format("KV cache data type for V (default: %s)", params.cache_type_v.c_str()), + string_format( + "KV cache data type for V\n" + "allowed values: %s\n" + "(default: %s)", + get_all_kv_cache_types().c_str(), + ggml_type_name(params.cache_type_v) + ), [](common_params & params, const std::string & value) { - // TODO: get the type right here - params.cache_type_v = value; + params.cache_type_v = kv_cache_type_from_str(value); } ).set_env("LLAMA_ARG_CACHE_TYPE_V")); add_opt(common_arg( @@ -1433,28 +1418,42 @@ common_params_context common_params_parser_init(common_params & params, llama_ex else { throw std::invalid_argument("invalid value"); } } ).set_env("LLAMA_ARG_NUMA")); + add_opt(common_arg( + {"-dev", "--device"}, "", + "comma-separated list of devices to use for offloading (none = don't offload)\n" + "use --list-devices to see a list of available devices", + [](common_params & params, const std::string & value) { + params.devices = parse_device_list(value); + } + ).set_env("LLAMA_ARG_DEVICE")); + add_opt(common_arg( + {"--list-devices"}, + "print list of available devices and exit", + [](common_params &) { + printf("Available devices:\n"); + for (size_t i = 0; i < ggml_backend_dev_count(); ++i) { + auto * dev = ggml_backend_dev_get(i); + if (ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_GPU) { + size_t free, total; + ggml_backend_dev_memory(dev, &free, &total); + printf(" %s: %s (%zu MiB, %zu MiB free)\n", ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), total / 1024 / 1024, free / 1024 / 1024); + } + } + exit(0); + } + )); add_opt(common_arg( {"-ngl", "--gpu-layers", "--n-gpu-layers"}, "N", "number of layers to store in VRAM", [](common_params & params, int value) { params.n_gpu_layers = value; if (!llama_supports_gpu_offload()) { - fprintf(stderr, "warning: not compiled with GPU offload support, --gpu-layers option will be ignored\n"); - fprintf(stderr, "warning: see main README.md for information on enabling GPU BLAS support\n"); + fprintf(stderr, "warning: no usable GPU found, --gpu-layers option will be ignored\n"); + fprintf(stderr, "warning: one possible reason is that llama.cpp was compiled without GPU support\n"); + fprintf(stderr, "warning: consult docs/build.md for compilation instructions\n"); } } ).set_env("LLAMA_ARG_N_GPU_LAYERS")); - add_opt(common_arg( - {"-ngld", "--gpu-layers-draft", "--n-gpu-layers-draft"}, "N", - "number of layers to store in VRAM for the draft model", - [](common_params & params, int value) { - params.n_gpu_layers_draft = value; - if (!llama_supports_gpu_offload()) { - fprintf(stderr, "warning: not compiled with GPU offload support, --gpu-layers-draft option will be ignored\n"); - fprintf(stderr, "warning: see main README.md for information on enabling GPU BLAS support\n"); - } - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); add_opt(common_arg( {"-sm", "--split-mode"}, "{none,layer,row}", "how to split the model across multiple GPUs, one of:\n" @@ -1468,10 +1467,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex } else if (arg_next == "layer") { params.split_mode = LLAMA_SPLIT_MODE_LAYER; } else if (arg_next == "row") { -#ifdef GGML_USE_SYCL - fprintf(stderr, "warning: The split mode value:[row] is not supported by llama.cpp with SYCL. It's developing.\nExit!\n"); - exit(1); -#endif // GGML_USE_SYCL params.split_mode = LLAMA_SPLIT_MODE_ROW; } else { throw std::invalid_argument("invalid value"); @@ -1539,7 +1534,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex {"--lora"}, "FNAME", "path to LoRA adapter (can be repeated to use multiple adapters)", [](common_params & params, const std::string & value) { - params.lora_adapters.push_back({ std::string(value), 1.0 }); + params.lora_adapters.push_back({ std::string(value), 1.0, nullptr }); } // we define this arg on both COMMON and EXPORT_LORA, so when showing help message of export-lora, it will be categorized as "example-specific" arg ).set_examples({LLAMA_EXAMPLE_COMMON, LLAMA_EXAMPLE_EXPORT_LORA})); @@ -1547,7 +1542,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex {"--lora-scaled"}, "FNAME", "SCALE", "path to LoRA adapter with user defined scaling (can be repeated to use multiple adapters)", [](common_params & params, const std::string & fname, const std::string & scale) { - params.lora_adapters.push_back({ fname, std::stof(scale) }); + params.lora_adapters.push_back({ fname, std::stof(scale), nullptr }); } // we define this arg on both COMMON and EXPORT_LORA, so when showing help message of export-lora, it will be categorized as "example-specific" arg ).set_examples({LLAMA_EXAMPLE_COMMON, LLAMA_EXAMPLE_EXPORT_LORA})); @@ -1593,13 +1588,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.model = value; } ).set_examples({LLAMA_EXAMPLE_COMMON, LLAMA_EXAMPLE_EXPORT_LORA}).set_env("LLAMA_ARG_MODEL")); - add_opt(common_arg( - {"-md", "--model-draft"}, "FNAME", - "draft model for speculative decoding (default: unused)", - [](common_params & params, const std::string & value) { - params.model_draft = value; - } - ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); add_opt(common_arg( {"-mu", "--model-url"}, "MODEL_URL", "model download url (default: unused)", @@ -1608,19 +1596,35 @@ common_params_context common_params_parser_init(common_params & params, llama_ex } ).set_env("LLAMA_ARG_MODEL_URL")); add_opt(common_arg( - {"-hfr", "--hf-repo"}, "REPO", - "Hugging Face model repository (default: unused)", + {"-hf", "-hfr", "--hf-repo"}, "/[:quant]", + "Hugging Face model repository; quant is optional, case-insensitive, default to Q4_K_M, or falls back to the first file in the repo if Q4_K_M doesn't exist.\n" + "example: unsloth/phi-4-GGUF:q4_k_m\n" + "(default: unused)", [](common_params & params, const std::string & value) { params.hf_repo = value; } ).set_env("LLAMA_ARG_HF_REPO")); add_opt(common_arg( {"-hff", "--hf-file"}, "FILE", - "Hugging Face model file (default: unused)", + "Hugging Face model file. If specified, it will override the quant in --hf-repo (default: unused)", [](common_params & params, const std::string & value) { params.hf_file = value; } ).set_env("LLAMA_ARG_HF_FILE")); + add_opt(common_arg( + {"-hfv", "-hfrv", "--hf-repo-v"}, "/[:quant]", + "Hugging Face model repository for the vocoder model (default: unused)", + [](common_params & params, const std::string & value) { + params.vocoder.hf_repo = value; + } + ).set_env("LLAMA_ARG_HF_REPO_V")); + add_opt(common_arg( + {"-hffv", "--hf-file-v"}, "FILE", + "Hugging Face model file for the vocoder model (default: unused)", + [](common_params & params, const std::string & value) { + params.vocoder.hf_file = value; + } + ).set_env("LLAMA_ARG_HF_FILE_V")); add_opt(common_arg( {"-hft", "--hf-token"}, "TOKEN", "Hugging Face access token (default: value from HF_TOKEN environment variable)", @@ -1789,6 +1793,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.public_path = value; } ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_STATIC_PATH")); + add_opt(common_arg( + {"--no-webui"}, + string_format("Disable the Web UI (default: %s)", params.webui ? "enabled" : "disabled"), + [](common_params & params) { + params.webui = false; + } + ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_NO_WEBUI")); add_opt(common_arg( {"--embedding", "--embeddings"}, string_format("restrict to only support embedding use case; use only with dedicated embedding models (default: %s)", params.embedding ? "enabled" : "disabled"), @@ -1904,9 +1915,11 @@ common_params_context common_params_parser_init(common_params & params, llama_ex ).set_examples({LLAMA_EXAMPLE_SERVER})); add_opt(common_arg( {"--chat-template"}, "JINJA_TEMPLATE", - "set custom jinja chat template (default: template taken from model's metadata)\n" - "if suffix/prefix are specified, template will be disabled\n" - "only commonly used templates are accepted:\nhttps://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template", + string_format( + "set custom jinja chat template (default: template taken from model's metadata)\n" + "if suffix/prefix are specified, template will be disabled\n" + "list of built-in templates:\n%s", list_builtin_chat_templates().c_str() + ), [](common_params & params, const std::string & value) { if (!common_chat_verify_template(value)) { throw std::runtime_error(string_format( @@ -1939,17 +1952,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex params.simple_io = true; } ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_INFILL})); - add_opt(common_arg( - {"-ld", "--logdir"}, "LOGDIR", - "path under which to save YAML logs (no logging if unset)", - [](common_params & params, const std::string & value) { - params.logdir = value; - - if (params.logdir.back() != DIRECTORY_SEPARATOR) { - params.logdir += DIRECTORY_SEPARATOR; - } - } - )); add_opt(common_arg( {"--positive-file"}, "FNAME", string_format("positive prompts file, one prompt per line (default: '%s')", params.cvector_positive_file.c_str()), @@ -2048,5 +2050,197 @@ common_params_context common_params_parser_init(common_params & params, llama_ex } ).set_env("LLAMA_LOG_TIMESTAMPS")); + // speculative parameters + add_opt(common_arg( + {"-td", "--threads-draft"}, "N", + "number of threads to use during generation (default: same as --threads)", + [](common_params & params, int value) { + params.speculative.cpuparams.n_threads = value; + if (params.speculative.cpuparams.n_threads <= 0) { + params.speculative.cpuparams.n_threads = std::thread::hardware_concurrency(); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"-tbd", "--threads-batch-draft"}, "N", + "number of threads to use during batch and prompt processing (default: same as --threads-draft)", + [](common_params & params, int value) { + params.speculative.cpuparams_batch.n_threads = value; + if (params.speculative.cpuparams_batch.n_threads <= 0) { + params.speculative.cpuparams_batch.n_threads = std::thread::hardware_concurrency(); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"-Cd", "--cpu-mask-draft"}, "M", + "Draft model CPU affinity mask. Complements cpu-range-draft (default: same as --cpu-mask)", + [](common_params & params, const std::string & mask) { + params.speculative.cpuparams.mask_valid = true; + if (!parse_cpu_mask(mask, params.speculative.cpuparams.cpumask)) { + throw std::invalid_argument("invalid cpumask"); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"-Crd", "--cpu-range-draft"}, "lo-hi", + "Ranges of CPUs for affinity. Complements --cpu-mask-draft", + [](common_params & params, const std::string & range) { + params.speculative.cpuparams.mask_valid = true; + if (!parse_cpu_range(range, params.speculative.cpuparams.cpumask)) { + throw std::invalid_argument("invalid range"); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--cpu-strict-draft"}, "<0|1>", + "Use strict CPU placement for draft model (default: same as --cpu-strict)", + [](common_params & params, int value) { + params.speculative.cpuparams.strict_cpu = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--prio-draft"}, "N", + string_format("set draft process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: %d)\n", params.speculative.cpuparams.priority), + [](common_params & params, int prio) { + if (prio < 0 || prio > 3) { + throw std::invalid_argument("invalid value"); + } + params.speculative.cpuparams.priority = (enum ggml_sched_priority) prio; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--poll-draft"}, "<0|1>", + "Use polling to wait for draft model work (default: same as --poll])", + [](common_params & params, int value) { + params.speculative.cpuparams.poll = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"-Cbd", "--cpu-mask-batch-draft"}, "M", + "Draft model CPU affinity mask. Complements cpu-range-draft (default: same as --cpu-mask)", + [](common_params & params, const std::string & mask) { + params.speculative.cpuparams_batch.mask_valid = true; + if (!parse_cpu_mask(mask, params.speculative.cpuparams_batch.cpumask)) { + throw std::invalid_argument("invalid cpumask"); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"-Crbd", "--cpu-range-batch-draft"}, "lo-hi", + "Ranges of CPUs for affinity. Complements --cpu-mask-draft-batch)", + [](common_params & params, const std::string & range) { + params.speculative.cpuparams_batch.mask_valid = true; + if (!parse_cpu_range(range, params.speculative.cpuparams_batch.cpumask)) { + throw std::invalid_argument("invalid cpumask"); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--cpu-strict-batch-draft"}, "<0|1>", + "Use strict CPU placement for draft model (default: --cpu-strict-draft)", + [](common_params & params, int value) { + params.speculative.cpuparams_batch.strict_cpu = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--prio-batch-draft"}, "N", + string_format("set draft process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: %d)\n", params.speculative.cpuparams_batch.priority), + [](common_params & params, int prio) { + if (prio < 0 || prio > 3) { + throw std::invalid_argument("invalid value"); + } + params.speculative.cpuparams_batch.priority = (enum ggml_sched_priority) prio; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--poll-batch-draft"}, "<0|1>", + "Use polling to wait for draft model work (default: --poll-draft)", + [](common_params & params, int value) { + params.speculative.cpuparams_batch.poll = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE})); + add_opt(common_arg( + {"--draft-max", "--draft", "--draft-n"}, "N", + string_format("number of tokens to draft for speculative decoding (default: %d)", params.speculative.n_max), + [](common_params & params, int value) { + params.speculative.n_max = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_LOOKUP, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_DRAFT_MAX")); + add_opt(common_arg( + {"--draft-min", "--draft-n-min"}, "N", + string_format("minimum number of draft tokens to use for speculative decoding (default: %d)", params.speculative.n_min), + [](common_params & params, int value) { + params.speculative.n_min = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_LOOKUP, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_DRAFT_MIN")); + add_opt(common_arg( + {"--draft-p-split"}, "P", + string_format("speculative decoding split probability (default: %.1f)", (double)params.speculative.p_split), + [](common_params & params, const std::string & value) { + params.speculative.p_split = std::stof(value); + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE}).set_env("LLAMA_ARG_DRAFT_P_SPLIT")); + add_opt(common_arg( + {"--draft-p-min"}, "P", + string_format("minimum speculative decoding probability (greedy) (default: %.1f)", (double)params.speculative.p_min), + [](common_params & params, const std::string & value) { + params.speculative.p_min = std::stof(value); + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_DRAFT_P_MIN")); + add_opt(common_arg( + {"-cd", "--ctx-size-draft"}, "N", + string_format("size of the prompt context for the draft model (default: %d, 0 = loaded from model)", params.speculative.n_ctx), + [](common_params & params, int value) { + params.speculative.n_ctx = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_CTX_SIZE_DRAFT")); + add_opt(common_arg( + {"-devd", "--device-draft"}, "", + "comma-separated list of devices to use for offloading the draft model (none = don't offload)\n" + "use --list-devices to see a list of available devices", + [](common_params & params, const std::string & value) { + params.speculative.devices = parse_device_list(value); + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER})); + add_opt(common_arg( + {"-ngld", "--gpu-layers-draft", "--n-gpu-layers-draft"}, "N", + "number of layers to store in VRAM for the draft model", + [](common_params & params, int value) { + params.speculative.n_gpu_layers = value; + if (!llama_supports_gpu_offload()) { + fprintf(stderr, "warning: no usable GPU found, --gpu-layers-draft option will be ignored\n"); + fprintf(stderr, "warning: one possible reason is that llama.cpp was compiled without GPU support\n"); + fprintf(stderr, "warning: consult docs/build.md for compilation instructions\n"); + } + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_N_GPU_LAYERS_DRAFT")); + add_opt(common_arg( + {"-md", "--model-draft"}, "FNAME", + "draft model for speculative decoding (default: unused)", + [](common_params & params, const std::string & value) { + params.speculative.model = value; + } + ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_MODEL_DRAFT")); + + add_opt(common_arg( + {"-mv", "--model-vocoder"}, "FNAME", + "vocoder model for audio generation (default: unused)", + [](common_params & params, const std::string & value) { + params.vocoder.model = value; + } + ).set_examples({LLAMA_EXAMPLE_TTS, LLAMA_EXAMPLE_SERVER})); + + // model-specific + add_opt(common_arg( + {"--tts-oute-default"}, + string_format("use default OuteTTS models (note: can download weights from the internet)"), + [](common_params & params) { + params.hf_repo = "OuteAI/OuteTTS-0.2-500M-GGUF"; + params.hf_file = "OuteTTS-0.2-500M-Q8_0.gguf"; + params.vocoder.hf_repo = "ggml-org/WavTokenizer"; + params.vocoder.hf_file = "WavTokenizer-Large-75-F16.gguf"; + } + ).set_examples({LLAMA_EXAMPLE_TTS})); + return ctx_arg; } diff --git a/common/arg.h b/common/arg.h index a6700d323..49ab8667b 100644 --- a/common/arg.h +++ b/common/arg.h @@ -12,6 +12,7 @@ struct common_arg { std::set examples = {LLAMA_EXAMPLE_COMMON}; + std::set excludes = {}; std::vector args; const char * value_hint = nullptr; // help text or example for arg value const char * value_hint_2 = nullptr; // for second arg value @@ -53,9 +54,11 @@ struct common_arg { ) : args(args), value_hint(value_hint), value_hint_2(value_hint_2), help(help), handler_str_str(handler) {} common_arg & set_examples(std::initializer_list examples); + common_arg & set_excludes(std::initializer_list excludes); common_arg & set_env(const char * env); common_arg & set_sparam(); bool in_example(enum llama_example ex); + bool is_exclude(enum llama_example ex); bool get_value_from_env(std::string & output); bool has_value_from_env(); std::string to_string(); diff --git a/common/common.cpp b/common/common.cpp index 19674af15..a6f9252b2 100644 --- a/common/common.cpp +++ b/common/common.cpp @@ -2,6 +2,9 @@ #define _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING #endif +#include "ggml.h" +#include "gguf.h" + #include "common.h" #include "log.h" // Change JSON_ASSERT from assert() to GGML_ASSERT: @@ -18,6 +21,7 @@ #include #include #include +#include #include #include #include @@ -62,11 +66,29 @@ #ifdef __linux__ #include #elif defined(_WIN32) -#define PATH_MAX MAX_PATH +# if !defined(PATH_MAX) +# define PATH_MAX MAX_PATH +# endif #else #include #endif #define LLAMA_CURL_MAX_URL_LENGTH 2084 // Maximum URL Length in Chrome: 2083 + +// +// CURL utils +// + +using curl_ptr = std::unique_ptr; + +// cannot use unique_ptr for curl_slist, because we cannot update without destroying the old one +struct curl_slist_ptr { + struct curl_slist * ptr = nullptr; + ~curl_slist_ptr() { + if (ptr) { + curl_slist_free_all(ptr); + } + } +}; #endif // LLAMA_USE_CURL using json = nlohmann::ordered_json; @@ -536,12 +558,12 @@ std::string string_from(const struct llama_context * ctx, const struct llama_bat [](const unsigned char c) { return !std::isprint(c); }), detokenized.end()); - buf << "\n" << std::to_string(i) - << ":token '" << detokenized << "'" - << ":pos " << std::to_string(batch.pos[i]) - << ":n_seq_id " << std::to_string(batch.n_seq_id[i]) - << ":seq_id " << std::to_string(batch.seq_id[i][0]) - << ":logits " << std::to_string(batch.logits[i]); + buf << "\n" << std::to_string(i) + << ", token '" << detokenized << "'" + << ", pos " << std::to_string(batch.pos[i]) + << ", n_seq_id " << std::to_string(batch.n_seq_id[i]) + << ", seq_id " << std::to_string(batch.seq_id[i][0]) + << ", logits " << std::to_string(batch.logits[i]); } buf << " ]"; @@ -652,7 +674,17 @@ bool fs_validate_filename(const std::string & filename) { std::u32string filename_utf32; try { +#if defined(__clang__) + // disable C++17 deprecation warning for std::codecvt_utf8 +# pragma clang diagnostic push +# pragma clang diagnostic ignored "-Wdeprecated-declarations" +#endif std::wstring_convert, char32_t> converter; + +#if defined(__clang__) +# pragma clang diagnostic pop +#endif + filename_utf32 = converter.from_bytes(filename); // If the reverse conversion mismatches, it means overlong UTF-8 sequences were used, @@ -829,11 +861,11 @@ struct common_init_result common_init_from_params(common_params & params) { llama_model * model = nullptr; if (!params.hf_repo.empty() && !params.hf_file.empty()) { - model = common_load_model_from_hf(params.hf_repo.c_str(), params.hf_file.c_str(), params.model.c_str(), params.hf_token.c_str(), mparams); + model = common_load_model_from_hf(params.hf_repo, params.hf_file, params.model, params.hf_token, mparams); } else if (!params.model_url.empty()) { - model = common_load_model_from_url(params.model_url.c_str(), params.model.c_str(), params.hf_token.c_str(), mparams); + model = common_load_model_from_url(params.model_url, params.model, params.hf_token, mparams); } else { - model = llama_load_model_from_file(params.model.c_str(), mparams); + model = llama_model_load_from_file(params.model.c_str(), mparams); } if (model == NULL) { @@ -841,26 +873,28 @@ struct common_init_result common_init_from_params(common_params & params) { return iparams; } + const llama_vocab * vocab = llama_model_get_vocab(model); + if (params.reranking) { bool ok = true; - if (llama_token_bos(model) == LLAMA_TOKEN_NULL) { - LOG_WRN("%s: warning: model does not have a BOS token, reranking will not work\n", __func__); + if (llama_vocab_bos(vocab) == LLAMA_TOKEN_NULL) { + LOG_WRN("%s: warning: vocab does not have a BOS token, reranking will not work\n", __func__); ok = false; } - if (llama_token_eos(model) == LLAMA_TOKEN_NULL) { - LOG_WRN("%s: warning: model does not have an EOS token, reranking will not work\n", __func__); + if (llama_vocab_eos(vocab) == LLAMA_TOKEN_NULL) { + LOG_WRN("%s: warning: vocab does not have an EOS token, reranking will not work\n", __func__); ok = false; } - if (llama_token_sep(model) == LLAMA_TOKEN_NULL) { - LOG_WRN("%s: warning: model does not have a SEP token, reranking will not work\n", __func__); + if (llama_vocab_sep(vocab) == LLAMA_TOKEN_NULL) { + LOG_WRN("%s: warning: vocab does not have a SEP token, reranking will not work\n", __func__); ok = false; } if (!ok) { - llama_free_model(model); + llama_model_free(model); return iparams; } @@ -868,34 +902,40 @@ struct common_init_result common_init_from_params(common_params & params) { auto cparams = common_context_params_to_llama(params); - llama_context * lctx = llama_new_context_with_model(model, cparams); + llama_context * lctx = llama_init_from_model(model, cparams); if (lctx == NULL) { LOG_ERR("%s: failed to create context with model '%s'\n", __func__, params.model.c_str()); - llama_free_model(model); + llama_model_free(model); return iparams; } + if (params.ctx_shift && !llama_kv_cache_can_shift(lctx)) { + LOG_WRN("%s: KV cache shifting is not supported for this model, disabling KV cache shifting\n", __func__); + params.ctx_shift = false; + } + if (!params.control_vectors.empty()) { if (params.control_vector_layer_start <= 0) params.control_vector_layer_start = 1; - if (params.control_vector_layer_end <= 0) params.control_vector_layer_end = llama_n_layer(model); + if (params.control_vector_layer_end <= 0) params.control_vector_layer_end = llama_model_n_layer(model); const auto cvec = common_control_vector_load(params.control_vectors); if (cvec.n_embd == -1) { llama_free(lctx); - llama_free_model(model); + llama_model_free(model); return iparams; } - int err = llama_control_vector_apply(lctx, - cvec.data.data(), - cvec.data.size(), - cvec.n_embd, - params.control_vector_layer_start, - params.control_vector_layer_end); + int err = llama_apply_adapter_cvec( + lctx, + cvec.data.data(), + cvec.data.size(), + cvec.n_embd, + params.control_vector_layer_start, + params.control_vector_layer_end); if (err) { llama_free(lctx); - llama_free_model(model); + llama_model_free(model); return iparams; } @@ -903,33 +943,54 @@ struct common_init_result common_init_from_params(common_params & params) { // load and optionally apply lora adapters for (auto & la : params.lora_adapters) { - common_lora_adapter_container loaded_la; - loaded_la.path = la.path; - loaded_la.scale = la.scale; - loaded_la.adapter = llama_lora_adapter_init(model, la.path.c_str()); - if (loaded_la.adapter == nullptr) { + llama_adapter_lora_ptr lora; + lora.reset(llama_adapter_lora_init(model, la.path.c_str())); + if (lora == nullptr) { LOG_ERR("%s: failed to apply lora adapter '%s'\n", __func__, la.path.c_str()); llama_free(lctx); - llama_free_model(model); + llama_model_free(model); return iparams; } - iparams.lora_adapters.push_back(loaded_la); // copy to list of loaded adapters - } - if (!params.lora_init_without_apply) { - common_lora_adapters_apply(lctx, iparams.lora_adapters); + + la.ptr = lora.get(); + iparams.lora.emplace_back(std::move(lora)); // copy to list of loaded adapters } - if (params.sparams.ignore_eos && llama_token_eos(model) == LLAMA_TOKEN_NULL) { - LOG_WRN("%s: warning: model does not have an EOS token, ignoring --ignore-eos\n", __func__); - params.sparams.ignore_eos = false; + if (!params.lora_init_without_apply) { + common_set_adapter_lora(lctx, params.lora_adapters); + } + + if (params.sampling.ignore_eos && llama_vocab_eos(vocab) == LLAMA_TOKEN_NULL) { + LOG_WRN("%s: warning: vocab does not have an EOS token, ignoring --ignore-eos\n", __func__); + params.sampling.ignore_eos = false; + } + + if (params.sampling.ignore_eos) { + for (llama_token i = 0; i < llama_vocab_n_tokens(vocab); i++) { + if (llama_vocab_is_eog(vocab, i)) { + LOG_INF("%s: added %s logit bias = %f\n", __func__, common_token_to_piece(lctx, i).c_str(), -INFINITY); + params.sampling.logit_bias.push_back({i, -INFINITY}); + } + } + } + + if (params.sampling.penalty_last_n == -1) { + LOG_INF("%s: setting penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx)); + params.sampling.penalty_last_n = llama_n_ctx(lctx); + } + + if (params.sampling.dry_penalty_last_n == -1) { + LOG_INF("%s: setting dry_penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx)); + params.sampling.dry_penalty_last_n = llama_n_ctx(lctx); } if (params.warmup) { LOG_WRN("%s: warming up the model with an empty run - please wait ... (--no-warmup to disable)\n", __func__); std::vector tmp; - llama_token bos = llama_token_bos(model); - llama_token eos = llama_token_eos(model); + llama_token bos = llama_vocab_bos(vocab); + llama_token eos = llama_vocab_eos(vocab); + // some models (e.g. T5) don't have a BOS token if (bos != LLAMA_TOKEN_NULL) { tmp.push_back(bos); @@ -944,7 +1005,7 @@ struct common_init_result common_init_from_params(common_params & params) { if (llama_model_has_encoder(model)) { llama_encode(lctx, llama_batch_get_one(tmp.data(), tmp.size())); llama_token decoder_start_token_id = llama_model_decoder_start_token(model); - if (decoder_start_token_id == -1) { + if (decoder_start_token_id == LLAMA_TOKEN_NULL) { decoder_start_token_id = bos; } tmp.clear(); @@ -958,24 +1019,27 @@ struct common_init_result common_init_from_params(common_params & params) { llama_perf_context_reset(lctx); } - iparams.model = model; - iparams.context = lctx; + iparams.model.reset(model); + iparams.context.reset(lctx); return iparams; } -void common_lora_adapters_apply(struct llama_context * ctx, std::vector & lora_adapters) { - llama_lora_adapter_clear(ctx); - for (auto & la : lora_adapters) { +void common_set_adapter_lora(struct llama_context * ctx, std::vector & lora) { + llama_clear_adapter_lora(ctx); + for (auto & la : lora) { if (la.scale != 0.0f) { - llama_lora_adapter_set(ctx, la.adapter, la.scale); + llama_set_adapter_lora(ctx, la.ptr, la.scale); } } } -struct llama_model_params common_model_params_to_llama(const common_params & params) { +struct llama_model_params common_model_params_to_llama(common_params & params) { auto mparams = llama_model_default_params(); + if (!params.devices.empty()) { + mparams.devices = params.devices.data(); + } if (params.n_gpu_layers != -1) { mparams.n_gpu_layers = params.n_gpu_layers; } @@ -996,38 +1060,6 @@ struct llama_model_params common_model_params_to_llama(const common_params & par return mparams; } -static ggml_type kv_cache_type_from_str(const std::string & s) { - if (s == "f32") { - return GGML_TYPE_F32; - } - if (s == "f16") { - return GGML_TYPE_F16; - } - if (s == "bf16") { - return GGML_TYPE_BF16; - } - if (s == "q8_0") { - return GGML_TYPE_Q8_0; - } - if (s == "q4_0") { - return GGML_TYPE_Q4_0; - } - if (s == "q4_1") { - return GGML_TYPE_Q4_1; - } - if (s == "iq4_nl") { - return GGML_TYPE_IQ4_NL; - } - if (s == "q5_0") { - return GGML_TYPE_Q5_0; - } - if (s == "q5_1") { - return GGML_TYPE_Q5_1; - } - - throw std::runtime_error("Unsupported cache type: " + s); -} - struct llama_context_params common_context_params_to_llama(const common_params & params) { auto cparams = llama_context_default_params(); @@ -1062,8 +1094,8 @@ struct llama_context_params common_context_params_to_llama(const common_params & cparams.pooling_type = LLAMA_POOLING_TYPE_RANK; } - cparams.type_k = kv_cache_type_from_str(params.cache_type_k); - cparams.type_v = kv_cache_type_from_str(params.cache_type_v); + cparams.type_k = params.cache_type_k; + cparams.type_v = params.cache_type_v; return cparams; } @@ -1089,13 +1121,7 @@ struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const cpu_p #define CURL_MAX_RETRY 3 #define CURL_RETRY_DELAY_SECONDS 2 - -static bool starts_with(const std::string & str, const std::string & prefix) { - // While we wait for C++20's std::string::starts_with... - return str.rfind(prefix, 0) == 0; -} - -static bool curl_perform_with_retry(const std::string& url, CURL* curl, int max_attempts, int retry_delay_seconds) { +static bool curl_perform_with_retry(const std::string & url, CURL * curl, int max_attempts, int retry_delay_seconds) { int remaining_attempts = max_attempts; while (remaining_attempts > 0) { @@ -1119,9 +1145,9 @@ static bool curl_perform_with_retry(const std::string& url, CURL* curl, int max_ } static bool common_download_file(const std::string & url, const std::string & path, const std::string & hf_token) { - // Initialize libcurl - std::unique_ptr curl(curl_easy_init(), &curl_easy_cleanup); + curl_ptr curl(curl_easy_init(), &curl_easy_cleanup); + curl_slist_ptr http_headers; if (!curl) { LOG_ERR("%s: error initializing libcurl\n", __func__); return false; @@ -1135,11 +1161,9 @@ static bool common_download_file(const std::string & url, const std::string & pa // Check if hf-token or bearer-token was specified if (!hf_token.empty()) { - std::string auth_header = "Authorization: Bearer "; - auth_header += hf_token.c_str(); - struct curl_slist *http_headers = NULL; - http_headers = curl_slist_append(http_headers, auth_header.c_str()); - curl_easy_setopt(curl.get(), CURLOPT_HTTPHEADER, http_headers); + std::string auth_header = "Authorization: Bearer " + hf_token; + http_headers.ptr = curl_slist_append(http_headers.ptr, auth_header.c_str()); + curl_easy_setopt(curl.get(), CURLOPT_HTTPHEADER, http_headers.ptr); } #if defined(_WIN32) @@ -1149,8 +1173,7 @@ static bool common_download_file(const std::string & url, const std::string & pa #endif // Check if the file already exists locally - struct stat model_file_info; - auto file_exists = (stat(path.c_str(), &model_file_info) == 0); + auto file_exists = std::filesystem::exists(path); // If the file exists, check its JSON metadata companion file. std::string metadata_path = path + ".json"; @@ -1192,11 +1215,13 @@ static bool common_download_file(const std::string & url, const std::string & pa std::string etag; std::string last_modified; }; + common_load_model_from_url_headers headers; + { typedef size_t(*CURLOPT_HEADERFUNCTION_PTR)(char *, size_t, size_t, void *); auto header_callback = [](char * buffer, size_t /*size*/, size_t n_items, void * userdata) -> size_t { - common_load_model_from_url_headers *headers = (common_load_model_from_url_headers *) userdata; + common_load_model_from_url_headers * headers = (common_load_model_from_url_headers *) userdata; static std::regex header_regex("([^:]+): (.*)\r\n"); static std::regex etag_regex("ETag", std::regex_constants::icase); @@ -1333,17 +1358,17 @@ static bool common_download_file(const std::string & url, const std::string & pa } struct llama_model * common_load_model_from_url( - const char * model_url, - const char * path_model, - const char * hf_token, + const std::string & model_url, + const std::string & local_path, + const std::string & hf_token, const struct llama_model_params & params) { // Basic validation of the model_url - if (!model_url || strlen(model_url) == 0) { + if (model_url.empty()) { LOG_ERR("%s: invalid model_url\n", __func__); return NULL; } - if (!common_download_file(model_url, path_model, hf_token)) { + if (!common_download_file(model_url, local_path, hf_token)) { return NULL; } @@ -1354,9 +1379,9 @@ struct llama_model * common_load_model_from_url( /*.no_alloc = */ true, /*.ctx = */ NULL, }; - auto * ctx_gguf = gguf_init_from_file(path_model, gguf_params); + auto * ctx_gguf = gguf_init_from_file(local_path.c_str(), gguf_params); if (!ctx_gguf) { - LOG_ERR("\n%s: failed to load input GGUF from %s\n", __func__, path_model); + LOG_ERR("\n%s: failed to load input GGUF from %s\n", __func__, local_path.c_str()); return NULL; } @@ -1375,13 +1400,13 @@ struct llama_model * common_load_model_from_url( // Verify the first split file format // and extract split URL and PATH prefixes { - if (!llama_split_prefix(split_prefix, sizeof(split_prefix), path_model, 0, n_split)) { - LOG_ERR("\n%s: unexpected model file name: %s n_split=%d\n", __func__, path_model, n_split); + if (!llama_split_prefix(split_prefix, sizeof(split_prefix), local_path.c_str(), 0, n_split)) { + LOG_ERR("\n%s: unexpected model file name: %s n_split=%d\n", __func__, local_path.c_str(), n_split); return NULL; } - if (!llama_split_prefix(split_url_prefix, sizeof(split_url_prefix), model_url, 0, n_split)) { - LOG_ERR("\n%s: unexpected model url: %s n_split=%d\n", __func__, model_url, n_split); + if (!llama_split_prefix(split_url_prefix, sizeof(split_url_prefix), model_url.c_str(), 0, n_split)) { + LOG_ERR("\n%s: unexpected model url: %s n_split=%d\n", __func__, model_url.c_str(), n_split); return NULL; } } @@ -1408,14 +1433,14 @@ struct llama_model * common_load_model_from_url( } } - return llama_load_model_from_file(path_model, params); + return llama_model_load_from_file(local_path.c_str(), params); } struct llama_model * common_load_model_from_hf( - const char * repo, - const char * model, - const char * path_model, - const char * hf_token, + const std::string & repo, + const std::string & remote_path, + const std::string & local_path, + const std::string & hf_token, const struct llama_model_params & params) { // construct hugging face model url: // @@ -1429,32 +1454,111 @@ struct llama_model * common_load_model_from_hf( std::string model_url = "https://huggingface.co/"; model_url += repo; model_url += "/resolve/main/"; - model_url += model; + model_url += remote_path; - return common_load_model_from_url(model_url.c_str(), path_model, hf_token, params); + return common_load_model_from_url(model_url, local_path, hf_token, params); +} + +/** + * Allow getting the HF file from the HF repo with tag (like ollama), for example: + * - bartowski/Llama-3.2-3B-Instruct-GGUF:q4 + * - bartowski/Llama-3.2-3B-Instruct-GGUF:Q4_K_M + * - bartowski/Llama-3.2-3B-Instruct-GGUF:q5_k_s + * Tag is optional, default to "latest" (meaning it checks for Q4_K_M first, then Q4, then if not found, return the first GGUF file in repo) + * + * Return pair of (with "repo" already having tag removed) + * + * Note: we use the Ollama-compatible HF API, but not using the blobId. Instead, we use the special "ggufFile" field which returns the value for "hf_file". This is done to be backward-compatible with existing cache files. + */ +std::pair common_get_hf_file(const std::string & hf_repo_with_tag, const std::string & hf_token) { + auto parts = string_split(hf_repo_with_tag, ':'); + std::string tag = parts.size() > 1 ? parts.back() : "latest"; + std::string hf_repo = parts[0]; + if (string_split(hf_repo, '/').size() != 2) { + throw std::invalid_argument("error: invalid HF repo format, expected /[:quant]\n"); + } + + // fetch model info from Hugging Face Hub API + json model_info; + curl_ptr curl(curl_easy_init(), &curl_easy_cleanup); + curl_slist_ptr http_headers; + std::string res_str; + std::string url = "https://huggingface.co/v2/" + hf_repo + "/manifests/" + tag; + curl_easy_setopt(curl.get(), CURLOPT_URL, url.c_str()); + curl_easy_setopt(curl.get(), CURLOPT_NOPROGRESS, 1L); + typedef size_t(*CURLOPT_WRITEFUNCTION_PTR)(void * ptr, size_t size, size_t nmemb, void * data); + auto write_callback = [](void * ptr, size_t size, size_t nmemb, void * data) -> size_t { + static_cast(data)->append((char * ) ptr, size * nmemb); + return size * nmemb; + }; + curl_easy_setopt(curl.get(), CURLOPT_WRITEFUNCTION, static_cast(write_callback)); + curl_easy_setopt(curl.get(), CURLOPT_WRITEDATA, &res_str); +#if defined(_WIN32) + curl_easy_setopt(curl.get(), CURLOPT_SSL_OPTIONS, CURLSSLOPT_NATIVE_CA); +#endif + if (!hf_token.empty()) { + std::string auth_header = "Authorization: Bearer " + hf_token; + http_headers.ptr = curl_slist_append(http_headers.ptr, auth_header.c_str()); + } + // Important: the User-Agent must be "llama-cpp" to get the "ggufFile" field in the response + http_headers.ptr = curl_slist_append(http_headers.ptr, "User-Agent: llama-cpp"); + http_headers.ptr = curl_slist_append(http_headers.ptr, "Accept: application/json"); + curl_easy_setopt(curl.get(), CURLOPT_HTTPHEADER, http_headers.ptr); + + CURLcode res = curl_easy_perform(curl.get()); + + if (res != CURLE_OK) { + throw std::runtime_error("error: cannot make GET request to HF API"); + } + + long res_code; + curl_easy_getinfo(curl.get(), CURLINFO_RESPONSE_CODE, &res_code); + if (res_code == 200) { + model_info = json::parse(res_str); + } else if (res_code == 401) { + throw std::runtime_error("error: model is private or does not exist; if you are accessing a gated model, please provide a valid HF token"); + } else { + throw std::runtime_error(string_format("error from HF API, response code: %ld, data: %s", res_code, res_str.c_str())); + } + + // check response + if (!model_info.contains("ggufFile")) { + throw std::runtime_error("error: model does not have ggufFile"); + } + json & gguf_file = model_info.at("ggufFile"); + if (!gguf_file.contains("rfilename")) { + throw std::runtime_error("error: ggufFile does not have rfilename"); + } + + return std::make_pair(hf_repo, gguf_file.at("rfilename")); } #else struct llama_model * common_load_model_from_url( - const char * /*model_url*/, - const char * /*path_model*/, - const char * /*hf_token*/, + const std::string & /*model_url*/, + const std::string & /*local_path*/, + const std::string & /*hf_token*/, const struct llama_model_params & /*params*/) { LOG_WRN("%s: llama.cpp built without libcurl, downloading from an url not supported.\n", __func__); return nullptr; } struct llama_model * common_load_model_from_hf( - const char * /*repo*/, - const char * /*model*/, - const char * /*path_model*/, - const char * /*hf_token*/, + const std::string & /*repo*/, + const std::string & /*remote_path*/, + const std::string & /*local_path*/, + const std::string & /*hf_token*/, const struct llama_model_params & /*params*/) { LOG_WRN("%s: llama.cpp built without libcurl, downloading from Hugging Face not supported.\n", __func__); return nullptr; } +std::pair common_get_hf_file(const std::string &, const std::string &) { + LOG_WRN("%s: llama.cpp built without libcurl, downloading from Hugging Face not supported.\n", __func__); + return std::make_pair("", ""); +} + #endif // LLAMA_USE_CURL // @@ -1484,6 +1588,66 @@ void common_batch_add( batch.n_tokens++; } +// +// Token utils +// + +size_t common_lcp(const llama_tokens & a, const llama_tokens & b) { + size_t i; + for (i = 0; i < a.size() && i < b.size() && a[i] == b[i]; i++) {} + + return i; +} + +size_t common_lcs(const llama_tokens & a, const llama_tokens & b) { + // check for empty sequences + if (a.empty() || b.empty()) { + return 0; + } + + // get the lengths of the input sequences + size_t a_len = a.size(); + size_t b_len = b.size(); + + // initialize the maximum length of the longest common subsequence (LCS) + size_t max_length = 0; + + // use two rows instead of a 2D matrix to optimize space + std::vector prev_row(b_len + 1, 0); + std::vector curr_row(b_len + 1, 0); + + // iterate through the elements of a + for (size_t i = 1; i <= a_len; i++) { + // iterate through the elements of b + for (size_t j = 1; j <= b_len; j++) { + // if elements at the current positions match + if (a[i - 1] == b[j - 1]) { + // if it's the first element of either sequences, set LCS length to 1 + if (i == 1 || j == 1) { + curr_row[j] = 1; + } else { + // increment LCS length by 1 compared to the previous element + curr_row[j] = prev_row[j - 1] + 1; + } + + // update max_length if necessary + if (curr_row[j] > max_length) { + max_length = curr_row[j]; + } + } else { + // reset LCS length if elements don't match + curr_row[j] = 0; + } + } + + // update the previous row for the next iteration + prev_row = curr_row; + } + + // return the maximum length of the LCS + return max_length; +} + // // Vocab utils // @@ -1493,21 +1657,23 @@ std::vector common_tokenize( const std::string & text, bool add_special, bool parse_special) { - return common_tokenize(llama_get_model(ctx), text, add_special, parse_special); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + return common_tokenize(vocab, text, add_special, parse_special); } std::vector common_tokenize( - const struct llama_model * model, + const struct llama_vocab * vocab, const std::string & text, bool add_special, bool parse_special) { // upper limit for the number of tokens int n_tokens = text.length() + 2 * add_special; std::vector result(n_tokens); - n_tokens = llama_tokenize(model, text.data(), text.length(), result.data(), result.size(), add_special, parse_special); + n_tokens = llama_tokenize(vocab, text.data(), text.length(), result.data(), result.size(), add_special, parse_special); if (n_tokens < 0) { result.resize(-n_tokens); - int check = llama_tokenize(model, text.data(), text.length(), result.data(), result.size(), add_special, parse_special); + int check = llama_tokenize(vocab, text.data(), text.length(), result.data(), result.size(), add_special, parse_special); GGML_ASSERT(check == -n_tokens); } else { result.resize(n_tokens); @@ -1516,12 +1682,18 @@ std::vector common_tokenize( } std::string common_token_to_piece(const struct llama_context * ctx, llama_token token, bool special) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + return common_token_to_piece(vocab, token, special); +} + +std::string common_token_to_piece(const struct llama_vocab * vocab, llama_token token, bool special) { std::string piece; piece.resize(piece.capacity()); // using string internal cache, 15 bytes + '\n' - const int n_chars = llama_token_to_piece(llama_get_model(ctx), token, &piece[0], piece.size(), 0, special); + const int n_chars = llama_token_to_piece(vocab, token, &piece[0], piece.size(), 0, special); if (n_chars < 0) { piece.resize(-n_chars); - int check = llama_token_to_piece(llama_get_model(ctx), token, &piece[0], piece.size(), 0, special); + int check = llama_token_to_piece(vocab, token, &piece[0], piece.size(), 0, special); GGML_ASSERT(check == -n_chars); } else { @@ -1531,13 +1703,19 @@ std::string common_token_to_piece(const struct llama_context * ctx, llama_token return piece; } -std::string common_detokenize(llama_context * ctx, const std::vector & tokens, bool special) { +std::string common_detokenize(const struct llama_context * ctx, const std::vector & tokens, bool special) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + return common_detokenize(vocab, tokens, special); +} + +std::string common_detokenize(const struct llama_vocab * vocab, const std::vector & tokens, bool special) { std::string text; text.resize(std::max(text.capacity(), tokens.size())); - int32_t n_chars = llama_detokenize(llama_get_model(ctx), tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); + int32_t n_chars = llama_detokenize(vocab, tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); if (n_chars < 0) { text.resize(-n_chars); - n_chars = llama_detokenize(llama_get_model(ctx), tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); + n_chars = llama_detokenize(vocab, tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); GGML_ASSERT(n_chars <= (int32_t)text.size()); // whitespace trimming is performed after per-token detokenization } @@ -1551,9 +1729,14 @@ std::string common_detokenize(llama_context * ctx, const std::vector= 0; } @@ -1564,16 +1747,16 @@ std::string common_chat_apply_template(const struct llama_model * model, int alloc_size = 0; bool fallback = false; // indicate if we must fallback to default chatml std::vector chat; - for (auto & msg : msgs) { + for (const auto & msg : msgs) { chat.push_back({msg.role.c_str(), msg.content.c_str()}); alloc_size += (msg.role.size() + msg.content.size()) * 1.25; } - const char * ptr_tmpl = tmpl.empty() ? nullptr : tmpl.c_str(); + const char * ptr_tmpl = tmpl.empty() ? llama_model_chat_template(model) : tmpl.c_str(); std::vector buf(alloc_size); // run the first time to get the total output length - int32_t res = llama_chat_apply_template(model, ptr_tmpl, chat.data(), chat.size(), add_ass, buf.data(), buf.size()); + int32_t res = llama_chat_apply_template(ptr_tmpl, chat.data(), chat.size(), add_ass, buf.data(), buf.size()); // error: chat template is not supported if (res < 0) { @@ -1581,18 +1764,17 @@ std::string common_chat_apply_template(const struct llama_model * model, // if the custom "tmpl" is not supported, we throw an error // this is a bit redundant (for good), since we're not sure if user validated the custom template with llama_chat_verify_template() throw std::runtime_error("this custom template is not supported"); - } else { - // If the built-in template is not supported, we default to chatml - res = llama_chat_apply_template(nullptr, "chatml", chat.data(), chat.size(), add_ass, buf.data(), buf.size()); - fallback = true; } + + // If the built-in template is not supported, we default to chatml + res = llama_chat_apply_template("chatml", chat.data(), chat.size(), add_ass, buf.data(), buf.size()); + fallback = true; } // if it turns out that our buffer is too small, we resize it if ((size_t) res > buf.size()) { buf.resize(res); res = llama_chat_apply_template( - fallback ? nullptr : model, fallback ? "chatml" : ptr_tmpl, chat.data(), chat.size(), add_ass, buf.data(), buf.size()); } @@ -1720,7 +1902,9 @@ void common_embd_normalize(const float * inp, float * out, int n, int embd_norm) break; case 0: // max absolute for (int i = 0; i < n; i++) { - if (sum < std::abs(inp[i])) sum = std::abs(inp[i]); + if (sum < std::abs(inp[i])) { + sum = std::abs(inp[i]); + } } sum /= 32760.0; // make an int16 range break; @@ -1890,218 +2074,3 @@ common_control_vector_data common_control_vector_load(const std::vector & data) { - if (data.empty()) { - fprintf(stream, "%s:\n", prop_name); - return; - } - - fprintf(stream, "%s: [", prop_name); - for (size_t i = 0; i < data.size() - 1; ++i) { - fprintf(stream, "%e, ", data[i]); - } - fprintf(stream, "%e]\n", data.back()); -} - -void yaml_dump_vector_int(FILE * stream, const char * prop_name, const std::vector & data) { - if (data.empty()) { - fprintf(stream, "%s:\n", prop_name); - return; - } - - fprintf(stream, "%s: [", prop_name); - for (size_t i = 0; i < data.size() - 1; ++i) { - fprintf(stream, "%d, ", data[i]); - } - fprintf(stream, "%d]\n", data.back()); -} - -void yaml_dump_string_multiline(FILE * stream, const char * prop_name, const char * data) { - std::string data_str(data == NULL ? "" : data); - - if (data_str.empty()) { - fprintf(stream, "%s:\n", prop_name); - return; - } - - size_t pos_start = 0; - size_t pos_found = 0; - - if (std::isspace(data_str[0]) || std::isspace(data_str.back())) { - data_str = std::regex_replace(data_str, std::regex("\n"), "\\n"); - data_str = std::regex_replace(data_str, std::regex("\""), "\\\""); - data_str = std::regex_replace(data_str, std::regex(R"(\\[^n"])"), R"(\$&)"); - data_str = "\"" + data_str + "\""; - fprintf(stream, "%s: %s\n", prop_name, data_str.c_str()); - return; - } - - if (data_str.find('\n') == std::string::npos) { - fprintf(stream, "%s: %s\n", prop_name, data_str.c_str()); - return; - } - - fprintf(stream, "%s: |\n", prop_name); - while ((pos_found = data_str.find('\n', pos_start)) != std::string::npos) { - fprintf(stream, " %s\n", data_str.substr(pos_start, pos_found-pos_start).c_str()); - pos_start = pos_found + 1; - } -} - -void yaml_dump_non_result_info(FILE * stream, const common_params & params, const llama_context * lctx, - const std::string & timestamp, const std::vector & prompt_tokens, const char * model_desc) { - ggml_cpu_init(); // some ARM features are detected at runtime - - const auto & sparams = params.sparams; - - fprintf(stream, "build_commit: %s\n", LLAMA_COMMIT); - fprintf(stream, "build_number: %d\n", LLAMA_BUILD_NUMBER); - fprintf(stream, "cpu_has_arm_fma: %s\n", ggml_cpu_has_arm_fma() ? "true" : "false"); - fprintf(stream, "cpu_has_avx: %s\n", ggml_cpu_has_avx() ? "true" : "false"); - fprintf(stream, "cpu_has_avx_vnni: %s\n", ggml_cpu_has_avx_vnni() ? "true" : "false"); - fprintf(stream, "cpu_has_avx2: %s\n", ggml_cpu_has_avx2() ? "true" : "false"); - fprintf(stream, "cpu_has_avx512: %s\n", ggml_cpu_has_avx512() ? "true" : "false"); - fprintf(stream, "cpu_has_avx512_vbmi: %s\n", ggml_cpu_has_avx512_vbmi() ? "true" : "false"); - fprintf(stream, "cpu_has_avx512_vnni: %s\n", ggml_cpu_has_avx512_vnni() ? "true" : "false"); - fprintf(stream, "cpu_has_cuda: %s\n", ggml_cpu_has_cuda() ? "true" : "false"); - fprintf(stream, "cpu_has_vulkan: %s\n", ggml_cpu_has_vulkan() ? "true" : "false"); - fprintf(stream, "cpu_has_kompute: %s\n", ggml_cpu_has_kompute() ? "true" : "false"); - fprintf(stream, "cpu_has_fma: %s\n", ggml_cpu_has_fma() ? "true" : "false"); - fprintf(stream, "cpu_has_gpublas: %s\n", ggml_cpu_has_gpublas() ? "true" : "false"); - fprintf(stream, "cpu_has_neon: %s\n", ggml_cpu_has_neon() ? "true" : "false"); - fprintf(stream, "cpu_has_sve: %s\n", ggml_cpu_has_sve() ? "true" : "false"); - fprintf(stream, "cpu_has_f16c: %s\n", ggml_cpu_has_f16c() ? "true" : "false"); - fprintf(stream, "cpu_has_fp16_va: %s\n", ggml_cpu_has_fp16_va() ? "true" : "false"); - fprintf(stream, "cpu_has_riscv_v: %s\n", ggml_cpu_has_riscv_v() ? "true" : "false"); - fprintf(stream, "cpu_has_wasm_simd: %s\n", ggml_cpu_has_wasm_simd() ? "true" : "false"); - fprintf(stream, "cpu_has_blas: %s\n", ggml_cpu_has_blas() ? "true" : "false"); - fprintf(stream, "cpu_has_sse3: %s\n", ggml_cpu_has_sse3() ? "true" : "false"); - fprintf(stream, "cpu_has_vsx: %s\n", ggml_cpu_has_vsx() ? "true" : "false"); - fprintf(stream, "cpu_has_matmul_int8: %s\n", ggml_cpu_has_matmul_int8() ? "true" : "false"); - -#ifdef NDEBUG - fprintf(stream, "debug: false\n"); -#else - fprintf(stream, "debug: true\n"); -#endif // NDEBUG - - fprintf(stream, "model_desc: %s\n", model_desc); - fprintf(stream, "n_vocab: %d # output size of the final layer, 32001 for some models\n", llama_n_vocab(llama_get_model(lctx))); - -#ifdef __OPTIMIZE__ - fprintf(stream, "optimize: true\n"); -#else - fprintf(stream, "optimize: false\n"); -#endif // __OPTIMIZE__ - - fprintf(stream, "time: %s\n", timestamp.c_str()); - - fprintf(stream, "\n"); - fprintf(stream, "###############\n"); - fprintf(stream, "# User Inputs #\n"); - fprintf(stream, "###############\n"); - fprintf(stream, "\n"); - - fprintf(stream, "alias: %s # default: unknown\n", params.model_alias.c_str()); - fprintf(stream, "batch_size: %d # default: 512\n", params.n_batch); - fprintf(stream, "chunks: %d # default: -1 (unlimited)\n", params.n_chunks); - fprintf(stream, "color: %s # default: false\n", params.use_color ? "true" : "false"); - fprintf(stream, "ctx_size: %d # default: 512\n", params.n_ctx); - fprintf(stream, "dry_allowed_length: %d # default: 2\n", sparams.dry_allowed_length); - fprintf(stream, "dry_base: %.2f # default: 1.75\n", sparams.dry_base); - fprintf(stream, "dry_multiplier: %.1f # default: 0.0\n", sparams.dry_multiplier); - fprintf(stream, "dry_penalty_last_n: %d # default: -1 (0 = disable, -1 = context size)\n", sparams.dry_penalty_last_n); - fprintf(stream, "escape: %s # default: false\n", params.escape ? "true" : "false"); - fprintf(stream, "file: # never logged, see prompt instead. Can still be specified for input.\n"); - fprintf(stream, "frequency_penalty: %f # default: 0.0 \n", sparams.penalty_freq); - yaml_dump_string_multiline(stream, "grammar", sparams.grammar.c_str()); - fprintf(stream, "grammar-file: # never logged, see grammar instead. Can still be specified for input.\n"); - fprintf(stream, "hellaswag: %s # default: false\n", params.hellaswag ? "true" : "false"); - fprintf(stream, "hellaswag_tasks: %zu # default: 400\n", params.hellaswag_tasks); - fprintf(stream, "ignore_eos: %s # default: false\n", sparams.ignore_eos ? "true" : "false"); - - yaml_dump_string_multiline(stream, "in_prefix", params.input_prefix.c_str()); - fprintf(stream, "in_prefix_bos: %s # default: false\n", params.input_prefix_bos ? "true" : "false"); - yaml_dump_string_multiline(stream, "in_suffix", params.input_prefix.c_str()); - fprintf(stream, "interactive: %s # default: false\n", params.interactive ? "true" : "false"); - fprintf(stream, "interactive_first: %s # default: false\n", params.interactive_first ? "true" : "false"); - fprintf(stream, "keep: %d # default: 0\n", params.n_keep); - fprintf(stream, "logdir: %s # default: unset (no logging)\n", params.logdir.c_str()); - - fprintf(stream, "logit_bias:\n"); - for (const auto & logit_bias : sparams.logit_bias) { - fprintf(stream, " %d: %f", logit_bias.token, logit_bias.bias); - } - - fprintf(stream, "lora:\n"); - for (auto & la : params.lora_adapters) { - if (la.scale == 1.0f) { - fprintf(stream, " - %s\n", la.path.c_str()); - } - } - fprintf(stream, "lora_scaled:\n"); - for (auto & la : params.lora_adapters) { - if (la.scale != 1.0f) { - fprintf(stream, " - %s: %f\n", la.path.c_str(), la.scale); - } - } - fprintf(stream, "lora_init_without_apply: %s # default: false\n", params.lora_init_without_apply ? "true" : "false"); - fprintf(stream, "main_gpu: %d # default: 0\n", params.main_gpu); - fprintf(stream, "min_keep: %d # default: 0 (disabled)\n", sparams.min_keep); - fprintf(stream, "mirostat: %d # default: 0 (disabled)\n", sparams.mirostat); - fprintf(stream, "mirostat_ent: %f # default: 5.0\n", sparams.mirostat_tau); - fprintf(stream, "mirostat_lr: %f # default: 0.1\n", sparams.mirostat_eta); - fprintf(stream, "mlock: %s # default: false\n", params.use_mlock ? "true" : "false"); - fprintf(stream, "model: %s # default: %s\n", params.model.c_str(), DEFAULT_MODEL_PATH); - fprintf(stream, "model_draft: %s # default:\n", params.model_draft.c_str()); - fprintf(stream, "multiline_input: %s # default: false\n", params.multiline_input ? "true" : "false"); - fprintf(stream, "n_gpu_layers: %d # default: -1\n", params.n_gpu_layers); - fprintf(stream, "n_predict: %d # default: -1 (unlimited)\n", params.n_predict); - fprintf(stream, "n_probs: %d # only used by server binary, default: 0\n", sparams.n_probs); - fprintf(stream, "no_mmap: %s # default: false\n", !params.use_mmap ? "true" : "false"); - fprintf(stream, "penalize_nl: %s # default: false\n", sparams.penalize_nl ? "true" : "false"); - fprintf(stream, "ppl_output_type: %d # default: 0\n", params.ppl_output_type); - fprintf(stream, "ppl_stride: %d # default: 0\n", params.ppl_stride); - fprintf(stream, "presence_penalty: %f # default: 0.0\n", sparams.penalty_present); - yaml_dump_string_multiline(stream, "prompt", params.prompt.c_str()); - fprintf(stream, "prompt_cache: %s\n", params.path_prompt_cache.c_str()); - fprintf(stream, "prompt_cache_all: %s # default: false\n", params.prompt_cache_all ? "true" : "false"); - fprintf(stream, "prompt_cache_ro: %s # default: false\n", params.prompt_cache_ro ? "true" : "false"); - yaml_dump_vector_int(stream, "prompt_tokens", prompt_tokens); - fprintf(stream, "repeat_penalty: %f # default: 1.1\n", sparams.penalty_repeat); - - fprintf(stream, "reverse_prompt:\n"); - for (std::string ap : params.antiprompt) { - size_t pos = 0; - while ((pos = ap.find('\n', pos)) != std::string::npos) { - ap.replace(pos, 1, "\\n"); - pos += 1; - } - - fprintf(stream, " - %s\n", ap.c_str()); - } - - fprintf(stream, "rope_freq_base: %f # default: 10000.0\n", params.rope_freq_base); - fprintf(stream, "rope_freq_scale: %f # default: 1.0\n", params.rope_freq_scale); - fprintf(stream, "simple_io: %s # default: false\n", params.simple_io ? "true" : "false"); - fprintf(stream, "cont_batching: %s # default: false\n", params.cont_batching ? "true" : "false"); - fprintf(stream, "flash_attn: %s # default: false\n", params.flash_attn ? "true" : "false"); - fprintf(stream, "temp: %f # default: 0.8\n", sparams.temp); - - const std::vector tensor_split_vector(params.tensor_split, params.tensor_split + llama_max_devices()); - yaml_dump_vector_float(stream, "tensor_split", tensor_split_vector); - - fprintf(stream, "threads: %d # default: %u\n", params.cpuparams.n_threads, std::thread::hardware_concurrency()); - fprintf(stream, "top_k: %d # default: 40\n", sparams.top_k); - fprintf(stream, "top_p: %f # default: 0.95\n", sparams.top_p); - fprintf(stream, "min_p: %f # default: 0.0\n", sparams.min_p); - fprintf(stream, "xtc_probability: %f # default: 0.0\n", sparams.xtc_probability); - fprintf(stream, "xtc_threshold: %f # default: 0.1\n", sparams.xtc_threshold); - fprintf(stream, "typ_p: %f # default: 1.0\n", sparams.typ_p); - fprintf(stream, "verbose_prompt: %s # default: false\n", params.verbose_prompt ? "true" : "false"); - fprintf(stream, "display_prompt: %s # default: true\n", params.display_prompt ? "true" : "false"); -} diff --git a/common/common.h b/common/common.h index 727f85baa..4fab1319a 100644 --- a/common/common.h +++ b/common/common.h @@ -2,7 +2,7 @@ #pragma once -#include "llama.h" +#include "llama-cpp.h" #include #include @@ -24,20 +24,20 @@ #define DEFAULT_MODEL_PATH "models/7B/ggml-model-f16.gguf" -struct common_lora_adapter_info { +struct common_adapter_lora_info { std::string path; float scale; + + struct llama_adapter_lora * ptr; }; -struct common_lora_adapter_container : common_lora_adapter_info { - struct llama_lora_adapter * adapter; -}; +using llama_tokens = std::vector; // build info extern int LLAMA_BUILD_NUMBER; -extern char const * LLAMA_COMMIT; -extern char const * LLAMA_COMPILER; -extern char const * LLAMA_BUILD_TARGET; +extern const char * LLAMA_COMMIT; +extern const char * LLAMA_COMPILER; +extern const char * LLAMA_BUILD_TARGET; struct common_control_vector_load_info; @@ -78,6 +78,7 @@ enum llama_example { LLAMA_EXAMPLE_LLAVA, LLAMA_EXAMPLE_LOOKUP, LLAMA_EXAMPLE_PARALLEL, + LLAMA_EXAMPLE_TTS, LLAMA_EXAMPLE_COUNT, }; @@ -93,6 +94,7 @@ enum common_sampler_type { COMMON_SAMPLER_TYPE_TEMPERATURE = 7, COMMON_SAMPLER_TYPE_XTC = 8, COMMON_SAMPLER_TYPE_INFILL = 9, + COMMON_SAMPLER_TYPE_PENALTIES = 10, }; // dimensionality reduction methods, used by cvector-generator @@ -101,8 +103,14 @@ enum dimre_method { DIMRE_METHOD_MEAN, }; -// sampler parameters -struct common_sampler_params { +enum common_conversation_mode { + COMMON_CONVERSATION_MODE_DISABLED = 0, + COMMON_CONVERSATION_MODE_ENABLED = 1, + COMMON_CONVERSATION_MODE_AUTO = 2, +}; + +// sampling parameters +struct common_params_sampling { uint32_t seed = LLAMA_DEFAULT_SEED; // the seed used to initialize llama_sampler int32_t n_prev = 64; // number of previous tokens to remember @@ -128,14 +136,15 @@ struct common_sampler_params { int32_t mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0 float mirostat_tau = 5.00f; // target entropy float mirostat_eta = 0.10f; // learning rate - bool penalize_nl = false; // consider newlines as a repeatable token bool ignore_eos = false; bool no_perf = false; // disable performance metrics + bool timing_per_token = false; std::vector dry_sequence_breakers = {"\n", ":", "\"", "*"}; // default sequence breakers for DRY std::vector samplers = { + COMMON_SAMPLER_TYPE_PENALTIES, COMMON_SAMPLER_TYPE_DRY, COMMON_SAMPLER_TYPE_TOP_K, COMMON_SAMPLER_TYPE_TYPICAL_P, @@ -153,21 +162,39 @@ struct common_sampler_params { std::string print() const; }; +struct common_params_speculative { + std::vector devices; // devices to use for offloading + + int32_t n_ctx = 0; // draft context size + int32_t n_max = 16; // maximum number of tokens to draft during speculative decoding + int32_t n_min = 5; // minimum number of draft tokens to use for speculative decoding + int32_t n_gpu_layers = -1; // number of layers to store in VRAM for the draft model (-1 - use default) + float p_split = 0.1f; // speculative decoding split probability + float p_min = 0.9f; // minimum speculative decoding probability (greedy) + + struct cpu_params cpuparams; + struct cpu_params cpuparams_batch; + + std::string model = ""; // draft model for speculative decoding // NOLINT +}; + +struct common_params_vocoder { + std::string hf_repo = ""; // HF repo // NOLINT + std::string hf_file = ""; // HF file // NOLINT + + std::string model = ""; // model path // NOLINT + std::string model_url = ""; // model url to download // NOLINT +}; + struct common_params { int32_t n_predict = -1; // new tokens to predict int32_t n_ctx = 4096; // context size int32_t n_batch = 2048; // logical batch size for prompt processing (must be >=32 to use BLAS) int32_t n_ubatch = 512; // physical batch size for prompt processing (must be >=32 to use BLAS) int32_t n_keep = 0; // number of tokens to keep from initial prompt - int32_t n_draft = 5; // number of tokens to draft during speculative decoding int32_t n_chunks = -1; // max number of chunks to process (-1 = unlimited) int32_t n_parallel = 1; // number of parallel sequences to decode int32_t n_sequences = 1; // number of sequences to decode - float p_split = 0.1f; // speculative decoding split probability - int32_t n_gpu_layers = -1; // number of layers to store in VRAM (-1 - use default) - int32_t n_gpu_layers_draft = -1; // number of layers to store in VRAM for the draft model (-1 - use default) - int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors - float tensor_split[128] = {0}; // how split tensors should be distributed across GPUs int32_t grp_attn_n = 1; // group-attention factor int32_t grp_attn_w = 512; // group-attention width int32_t n_print = -1; // print token count every n tokens (-1 = disabled) @@ -178,28 +205,35 @@ struct common_params { float yarn_beta_fast = 32.0f; // YaRN low correction dim float yarn_beta_slow = 1.0f; // YaRN high correction dim int32_t yarn_orig_ctx = 0; // YaRN original context length - float defrag_thold = -1.0f; // KV cache defragmentation threshold + float defrag_thold = 0.1f; // KV cache defragmentation threshold + + // offload params + std::vector devices; // devices to use for offloading + + int32_t n_gpu_layers = -1; // number of layers to store in VRAM (-1 - use default) + int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors + float tensor_split[128] = {0}; // how split tensors should be distributed across GPUs + + enum llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs struct cpu_params cpuparams; struct cpu_params cpuparams_batch; - struct cpu_params draft_cpuparams; - struct cpu_params draft_cpuparams_batch; ggml_backend_sched_eval_callback cb_eval = nullptr; void * cb_eval_user_data = nullptr; ggml_numa_strategy numa = GGML_NUMA_STRATEGY_DISABLED; - enum llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs enum llama_rope_scaling_type rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED; enum llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_UNSPECIFIED; // pooling type for embeddings enum llama_attention_type attention_type = LLAMA_ATTENTION_TYPE_UNSPECIFIED; // attention type for embeddings - struct common_sampler_params sparams; + struct common_params_sampling sampling; + struct common_params_speculative speculative; + struct common_params_vocoder vocoder; std::string model = ""; // model path // NOLINT - std::string model_draft = ""; // draft model for speculative decoding // NOLINT - std::string model_alias = "unknown"; // model alias // NOLINT + std::string model_alias = ""; // model alias // NOLINT std::string model_url = ""; // model url to download // NOLINT std::string hf_token = ""; // HF token // NOLINT std::string hf_repo = ""; // HF repo // NOLINT @@ -209,7 +243,6 @@ struct common_params { std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state // NOLINT std::string input_prefix = ""; // string to prefix user inputs with // NOLINT std::string input_suffix = ""; // string to suffix user inputs with // NOLINT - std::string logdir = ""; // directory in which to save YAML log files // NOLINT std::string lookup_cache_static = ""; // path of static ngram cache file for lookup decoding // NOLINT std::string lookup_cache_dynamic = ""; // path of dynamic ngram cache file for lookup decoding // NOLINT std::string logits_file = ""; // file for saving *all* logits // NOLINT @@ -219,8 +252,8 @@ struct common_params { std::vector antiprompt; // strings upon which more user input is prompted (a.k.a. reverse prompts) std::vector kv_overrides; - bool lora_init_without_apply = false; // only load lora to memory, but do not apply it to ctx (user can manually apply lora later using llama_lora_adapter_apply) - std::vector lora_adapters; // lora adapter path with user defined scale + bool lora_init_without_apply = false; // only load lora to memory, but do not apply it to ctx (user can manually apply lora later using llama_adapter_lora_apply) + std::vector lora_adapters; // lora adapter path with user defined scale std::vector control_vectors; // control vector with user defined scale @@ -248,7 +281,6 @@ struct common_params { bool special = false; // enable special token output bool interactive = false; // interactive mode bool interactive_first = false; // wait for user input immediately - bool conversation = false; // conversation mode (does not print special tokens and suffix/prefix) bool prompt_cache_all = false; // save user input and generations to prompt cache bool prompt_cache_ro = false; // open the prompt cache read-only and do not update it @@ -271,8 +303,10 @@ struct common_params { bool warmup = true; // warmup run bool check_tensors = false; // validate tensor data - std::string cache_type_k = "f16"; // KV cache data type for the K - std::string cache_type_v = "f16"; // KV cache data type for the V + ggml_type cache_type_k = GGML_TYPE_F16; // KV cache data type for the K + ggml_type cache_type_v = GGML_TYPE_F16; // KV cache data type for the V + + common_conversation_mode conversation_mode = COMMON_CONVERSATION_MODE_AUTO; // multimodal models (see examples/llava) std::string mmproj = ""; // path to multimodal projector // NOLINT @@ -422,6 +456,16 @@ std::vector string_split(const std::string & input, ch return parts; } +static bool string_starts_with(const std::string & str, + const std::string & prefix) { // While we wait for C++20's std::string::starts_with... + return str.rfind(prefix, 0) == 0; +} + +static bool string_ends_with(const std::string & str, + const std::string & suffix) { // While we wait for C++20's std::string::ends_with... + return str.size() >= suffix.size() && str.compare(str.size()-suffix.size(), suffix.size(), suffix) == 0; +} + bool string_parse_kv_override(const char * data, std::vector & overrides); void string_process_escapes(std::string & input); @@ -444,25 +488,41 @@ std::string fs_get_cache_file(const std::string & filename); // Model utils // +// note: defines object's lifetime struct common_init_result { - struct llama_model * model = nullptr; - struct llama_context * context = nullptr; - std::vector lora_adapters; + llama_model_ptr model; + llama_context_ptr context; + + std::vector lora; }; struct common_init_result common_init_from_params(common_params & params); -struct llama_model_params common_model_params_to_llama (const common_params & params); +struct llama_model_params common_model_params_to_llama ( common_params & params); struct llama_context_params common_context_params_to_llama(const common_params & params); struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const cpu_params & params); -struct llama_model * common_load_model_from_url(const char * model_url, const char * path_model, const char * hf_token, const struct llama_model_params & params); -struct llama_model * common_load_model_from_hf(const char * repo, const char * file, const char * path_model, const char * hf_token, const struct llama_model_params & params); +struct llama_model * common_load_model_from_url( + const std::string & model_url, + const std::string & local_path, + const std::string & hf_token, + const struct llama_model_params & params); +struct llama_model * common_load_model_from_hf( + const std::string & repo, + const std::string & remote_path, + const std::string & local_path, + const std::string & hf_token, + const struct llama_model_params & params); +std::pair common_get_hf_file( + const std::string & hf_repo_with_tag, + const std::string & hf_token); // clear LoRA adapters from context, then apply new list of adapters -void common_lora_adapters_apply(struct llama_context * ctx, std::vector & lora_adapters); +void common_set_adapter_lora(struct llama_context * ctx, std::vector & lora); +// // Batch utils +// void common_batch_clear(struct llama_batch & batch); @@ -473,6 +533,16 @@ void common_batch_add( const std::vector & seq_ids, bool logits); +// +// Token utils +// + +// longest common prefix +size_t common_lcp(const llama_tokens & a, const llama_tokens & b); + +// longet common subsequence +size_t common_lcs(const llama_tokens & a, const llama_tokens & b); + // // Vocab utils // @@ -486,7 +556,7 @@ std::vector common_tokenize( bool parse_special = false); std::vector common_tokenize( - const struct llama_model * model, + const struct llama_vocab * vocab, const std::string & text, bool add_special, bool parse_special = false); @@ -498,11 +568,21 @@ std::string common_token_to_piece( llama_token token, bool special = true); +std::string common_token_to_piece( + const struct llama_vocab * vocab, + llama_token token, + bool special = true); + // detokenizes a vector of tokens into a string // should work similar to Python's `tokenizer.decode` // optionally renders special/control tokens std::string common_detokenize( - llama_context * ctx, + const struct llama_context * ctx, + const std::vector & tokens, + bool special = true); + +std::string common_detokenize( + const struct llama_vocab * vocab, const std::vector & tokens, bool special = true); @@ -516,6 +596,9 @@ struct common_chat_msg { std::string content; }; +// Get the built-in chat template for the model. Return empty string if not present. +std::string common_get_builtin_chat_template(const struct llama_model * model); + // Check if the template supplied via "--chat-template" is supported or not. Returns true if it's valid bool common_chat_verify_template(const std::string & tmpl); @@ -552,7 +635,8 @@ void common_kv_cache_dump_view_seqs(const llama_kv_cache_view & view, int row_si // Embedding utils // -void common_embd_normalize(const float * inp, float * out, int n, int embd_norm = 2); +// TODO: repace embd_norm with an enum +void common_embd_normalize(const float * inp, float * out, int n, int embd_norm); float common_embd_similarity_cos(const float * embd1, const float * embd2, int n); @@ -581,18 +665,10 @@ common_control_vector_data common_control_vector_load(const std::vector & data); -void yaml_dump_vector_int (FILE * stream, const char * prop_name, const std::vector & data); -void yaml_dump_string_multiline(FILE * stream, const char * prop_name, const char * data); - -void yaml_dump_non_result_info( - FILE * stream, const common_params & params, const llama_context * lctx, - const std::string & timestamp, const std::vector & prompt_tokens, const char * model_desc); +} diff --git a/common/ngram-cache.cpp b/common/ngram-cache.cpp index a9dfb6714..a057ae45f 100644 --- a/common/ngram-cache.cpp +++ b/common/ngram-cache.cpp @@ -65,13 +65,13 @@ constexpr int draft_min_percent_strict[LLAMA_NGRAM_MAX] = {75, 66, 66, 66}; static llama_token try_draft(common_ngram_cache & nc_static, const common_ngram ngram_static) { common_ngram_cache::iterator part_static_it = nc_static.find(ngram_static); if (part_static_it == nc_static.end()) { - return -1; + return LLAMA_TOKEN_NULL; } const common_ngram_cache_part part_static = part_static_it->second; int max_count_static = 0; int sum_count_static = 0; - llama_token max_token = -1; + llama_token max_token = LLAMA_TOKEN_NULL; for (std::pair token_count_static : part_static) { const llama_token token = token_count_static.first; @@ -85,10 +85,10 @@ static llama_token try_draft(common_ngram_cache & nc_static, const common_ngram } if (sum_count_static < draft_min_sample_size_lax[LLAMA_NGRAM_STATIC-1]) { - return -1; + return LLAMA_TOKEN_NULL; } if (100*max_count_static < draft_min_percent_lax[LLAMA_NGRAM_STATIC-1]*sum_count_static) { - return -1; + return LLAMA_TOKEN_NULL; } return max_token; } @@ -98,9 +98,9 @@ static llama_token try_draft( common_ngram_cache & nc_primary, const std::vector & ngrams_primary, common_ngram_cache_part & part_static, const int * min_sample_size, const int * min_percent) { - llama_token drafted_token = -1; + llama_token drafted_token = LLAMA_TOKEN_NULL; - for (int i = ngrams_primary.size()-1; i >= 0 && drafted_token == -1; --i) { + for (int i = ngrams_primary.size()-1; i >= 0 && drafted_token == LLAMA_TOKEN_NULL; --i) { const common_ngram ngram_primary = ngrams_primary[i]; common_ngram_cache::iterator part_primary_it = nc_primary.find(ngram_primary); @@ -112,7 +112,7 @@ static llama_token try_draft( int max_count_primary = 0; int max_count_static = 0; int sum_count_primary = 0; - llama_token max_token = -1; + llama_token max_token = LLAMA_TOKEN_NULL; for (std::pair token_count_primary : part_primary) { const llama_token token = token_count_primary.first; @@ -154,7 +154,7 @@ void common_ngram_cache_draft( } while ((int) draft.size()-1 < n_draft) { - llama_token drafted_token = -1; + llama_token drafted_token = LLAMA_TOKEN_NULL; const int ngram_start_static = inp_size-LLAMA_NGRAM_STATIC + draft.size()-1; common_ngram ngram_static; @@ -177,17 +177,17 @@ void common_ngram_cache_draft( } ngrams_cd.push_back(ngram_cd); } - if (drafted_token == -1) { + if (drafted_token == LLAMA_TOKEN_NULL) { drafted_token = try_draft(nc_context, ngrams_cd, part_static, draft_min_sample_size_lax, draft_min_percent_lax); } - if (drafted_token == -1) { + if (drafted_token == LLAMA_TOKEN_NULL) { drafted_token = try_draft(nc_dynamic, ngrams_cd, part_static, draft_min_sample_size_strict, draft_min_percent_strict); } - if (drafted_token == -1) { + if (drafted_token == LLAMA_TOKEN_NULL) { drafted_token = try_draft(nc_static, ngram_static); } - if (drafted_token == -1) { + if (drafted_token == LLAMA_TOKEN_NULL) { break; } diff --git a/common/ngram-cache.h b/common/ngram-cache.h index 09c2b0319..dfe012abe 100644 --- a/common/ngram-cache.h +++ b/common/ngram-cache.h @@ -17,13 +17,13 @@ struct common_ngram { common_ngram() { for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) { - tokens[i] = -1; + tokens[i] = LLAMA_TOKEN_NULL; } } common_ngram(const llama_token * input, const int ngram_size) { for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) { - tokens[i] = i < ngram_size ? input[i] : -1; + tokens[i] = i < ngram_size ? input[i] : LLAMA_TOKEN_NULL; } } diff --git a/common/sampling.cpp b/common/sampling.cpp index 7922fde47..7241ac321 100644 --- a/common/sampling.cpp +++ b/common/sampling.cpp @@ -99,7 +99,7 @@ struct ring_buffer { }; struct common_sampler { - common_sampler_params params; + common_params_sampling params; struct llama_sampler * grmr; struct llama_sampler * chain; @@ -113,7 +113,10 @@ struct common_sampler { void set_logits(struct llama_context * ctx, int idx) { const auto * logits = llama_get_logits_ith(ctx, idx); - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_vocab = llama_vocab_n_tokens(vocab); cur.resize(n_vocab); @@ -125,7 +128,7 @@ struct common_sampler { } }; -std::string common_sampler_params::print() const { +std::string common_params_sampling::print() const { char result[1024]; snprintf(result, sizeof(result), @@ -141,14 +144,16 @@ std::string common_sampler_params::print() const { return std::string(result); } -struct common_sampler * common_sampler_init(const struct llama_model * model, const struct common_sampler_params & params) { +struct common_sampler * common_sampler_init(const struct llama_model * model, const struct common_params_sampling & params) { + const llama_vocab * vocab = llama_model_get_vocab(model); + llama_sampler_chain_params lparams = llama_sampler_chain_default_params(); lparams.no_perf = params.no_perf; auto * result = new common_sampler { /* .params = */ params, - /* .grmr = */ llama_sampler_init_grammar(model, params.grammar.c_str(), "root"), + /* .grmr = */ llama_sampler_init_grammar(vocab, params.grammar.c_str(), "root"), /* .chain = */ llama_sampler_chain_init(lparams), /* .prev = */ ring_buffer(std::max(32, params.n_prev)), /* .cur = */ {}, @@ -157,36 +162,24 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, co llama_sampler_chain_add(result->chain, llama_sampler_init_logit_bias( - llama_n_vocab(model), + llama_vocab_n_tokens(vocab), params.logit_bias.size(), params.logit_bias.data())); - llama_sampler_chain_add(result->chain, - llama_sampler_init_penalties( - llama_n_vocab (model), - llama_token_eos(model), - llama_token_nl (model), - params.penalty_last_n, - params.penalty_repeat, - params.penalty_freq, - params.penalty_present, - params.penalize_nl, - params.ignore_eos)); - if (params.mirostat == 0) { for (const auto & cnstr : params.samplers) { switch (cnstr) { - case COMMON_SAMPLER_TYPE_DRY: + case COMMON_SAMPLER_TYPE_DRY: { - std::vector c_breakers; + std::vector c_breakers; c_breakers.reserve(params.dry_sequence_breakers.size()); - for (const auto& str : params.dry_sequence_breakers) { + for (const auto & str : params.dry_sequence_breakers) { c_breakers.push_back(str.c_str()); } - llama_sampler_chain_add(result->chain, llama_sampler_init_dry (model, params.dry_multiplier, params.dry_base, params.dry_allowed_length, params.dry_penalty_last_n, c_breakers.data(), c_breakers.size())); + llama_sampler_chain_add(result->chain, llama_sampler_init_dry (vocab, llama_model_n_ctx_train(model), params.dry_multiplier, params.dry_base, params.dry_allowed_length, params.dry_penalty_last_n, c_breakers.data(), c_breakers.size())); } - break; + break; case COMMON_SAMPLER_TYPE_TOP_K: llama_sampler_chain_add(result->chain, llama_sampler_init_top_k (params.top_k)); break; @@ -206,7 +199,10 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, co llama_sampler_chain_add(result->chain, llama_sampler_init_temp_ext (params.temp, params.dynatemp_range, params.dynatemp_exponent)); break; case COMMON_SAMPLER_TYPE_INFILL: - llama_sampler_chain_add(result->chain, llama_sampler_init_infill (model)); + llama_sampler_chain_add(result->chain, llama_sampler_init_infill (vocab)); + break; + case COMMON_SAMPLER_TYPE_PENALTIES: + llama_sampler_chain_add(result->chain, llama_sampler_init_penalties(params.penalty_last_n, params.penalty_repeat, params.penalty_freq, params.penalty_present)); break; default: GGML_ASSERT(false && "unknown sampler type"); @@ -215,7 +211,7 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, co llama_sampler_chain_add(result->chain, llama_sampler_init_dist(params.seed)); } else if (params.mirostat == 1) { llama_sampler_chain_add(result->chain, llama_sampler_init_temp(params.temp)); - llama_sampler_chain_add(result->chain, llama_sampler_init_mirostat(llama_n_vocab(model), params.seed, params.mirostat_tau, params.mirostat_eta, 100)); + llama_sampler_chain_add(result->chain, llama_sampler_init_mirostat(llama_vocab_n_tokens(vocab), params.seed, params.mirostat_tau, params.mirostat_eta, 100)); } else if (params.mirostat == 2) { llama_sampler_chain_add(result->chain, llama_sampler_init_temp(params.temp)); llama_sampler_chain_add(result->chain, llama_sampler_init_mirostat_v2(params.seed, params.mirostat_tau, params.mirostat_eta)); @@ -320,6 +316,45 @@ llama_token common_sampler_sample(struct common_sampler * gsmpl, struct llama_co return cur_p.data[cur_p.selected].id; } +std::vector common_sampler_sample_and_accept_n(struct common_sampler * gsmpl, struct llama_context * ctx, const std::vector & idxs, const llama_tokens & draft, bool grammar_first) { + GGML_ASSERT(idxs.size() == draft.size() + 1 && "idxs.size() must be draft.size() + 1"); + + std::vector result; + result.reserve(idxs.size()); + + size_t i = 0; + for (; i < draft.size(); i++) { + const llama_token id = common_sampler_sample(gsmpl, ctx, idxs[i], grammar_first); + + common_sampler_accept(gsmpl, id, true); + + result.push_back(id); + + if (draft[i] != id) { + break; + } + } + + if (i == draft.size()) { + const llama_token id = common_sampler_sample(gsmpl, ctx, idxs[i], grammar_first); + + common_sampler_accept(gsmpl, id, true); + + result.push_back(id); + } + + return result; +} + +std::vector common_sampler_sample_and_accept_n(struct common_sampler * gsmpl, struct llama_context * ctx, const llama_tokens & draft, bool grammar_first) { + std::vector idxs(draft.size() + 1); + for (size_t i = 0; i < idxs.size(); ++i) { + idxs[i] = i; + } + + return common_sampler_sample_and_accept_n(gsmpl, ctx, idxs, draft, grammar_first); +} + uint32_t common_sampler_get_seed(const struct common_sampler * gsmpl) { return llama_sampler_get_seed(gsmpl->chain); } @@ -376,6 +411,7 @@ char common_sampler_type_to_chr(enum common_sampler_type cnstr) { case COMMON_SAMPLER_TYPE_TEMPERATURE: return 't'; case COMMON_SAMPLER_TYPE_XTC: return 'x'; case COMMON_SAMPLER_TYPE_INFILL: return 'i'; + case COMMON_SAMPLER_TYPE_PENALTIES: return 'e'; default : return '?'; } } @@ -390,6 +426,7 @@ std::string common_sampler_type_to_str(enum common_sampler_type cnstr) { case COMMON_SAMPLER_TYPE_TEMPERATURE: return "temperature"; case COMMON_SAMPLER_TYPE_XTC: return "xtc"; case COMMON_SAMPLER_TYPE_INFILL: return "infill"; + case COMMON_SAMPLER_TYPE_PENALTIES: return "penalties"; default : return ""; } } @@ -404,6 +441,7 @@ std::vector common_sampler_types_from_names(const std::vect { "temperature", COMMON_SAMPLER_TYPE_TEMPERATURE }, { "xtc", COMMON_SAMPLER_TYPE_XTC }, { "infill", COMMON_SAMPLER_TYPE_INFILL }, + { "penalties", COMMON_SAMPLER_TYPE_PENALTIES }, }; // since samplers names are written multiple ways @@ -450,6 +488,7 @@ std::vector common_sampler_types_from_chars(const std::stri { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_TEMPERATURE), COMMON_SAMPLER_TYPE_TEMPERATURE }, { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_XTC), COMMON_SAMPLER_TYPE_XTC }, { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_INFILL), COMMON_SAMPLER_TYPE_INFILL }, + { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_PENALTIES), COMMON_SAMPLER_TYPE_PENALTIES }, }; std::vector samplers; diff --git a/common/sampling.h b/common/sampling.h index d37f25ad3..348911b18 100644 --- a/common/sampling.h +++ b/common/sampling.h @@ -36,7 +36,7 @@ struct common_sampler; // llama_sampler API overloads -struct common_sampler * common_sampler_init(const struct llama_model * model, const struct common_sampler_params & params); +struct common_sampler * common_sampler_init(const struct llama_model * model, const struct common_params_sampling & params); void common_sampler_free(struct common_sampler * gsmpl); @@ -60,6 +60,27 @@ void common_perf_print(const struct llama_context * ctx, const struct common_sam // llama_token common_sampler_sample(struct common_sampler * gsmpl, struct llama_context * ctx, int idx, bool grammar_first = false); +// generalized version of common_sampler_sample +// +// will cross-reference the sampled tokens with a batch of draft tokens and accept those that match +// if the sampler disagrees at some point, we stop and return the accepted tokens up to now +// +// common_sampler_sample_n(gsmpl, ctx, { idx }, {}); +// +// is equivalent to +// +// common_sampler_sample(gsmpl, ctx, idx); +// common_sampler_accept(gsmpl, token, true); +// +// requires: idxs.size() == draft.size() + 1 +// +// returns at least 1 token, up to idxs.size() +// +std::vector common_sampler_sample_and_accept_n(struct common_sampler * gsmpl, struct llama_context * ctx, const std::vector & idxs, const llama_tokens & draft, bool grammar_first = false); + +// assume idxs == [ 0, 1, 2, ..., draft.size() ] +std::vector common_sampler_sample_and_accept_n(struct common_sampler * gsmpl, struct llama_context * ctx, const llama_tokens & draft, bool grammar_first = false); + uint32_t common_sampler_get_seed(const struct common_sampler * gsmpl); // helpers diff --git a/common/speculative.cpp b/common/speculative.cpp new file mode 100644 index 000000000..318e96ea3 --- /dev/null +++ b/common/speculative.cpp @@ -0,0 +1,277 @@ +#include "speculative.h" + +#include "log.h" +#include "common.h" +#include "sampling.h" + +#include + +#define SPEC_VOCAB_MAX_SIZE_DIFFERENCE 128 +#define SPEC_VOCAB_CHECK_START_TOKEN_ID 5 + +struct common_speculative { + struct llama_context * ctx; + struct common_sampler * smpl; + + llama_batch batch; + llama_tokens prompt; +}; + +struct common_speculative * common_speculative_init( + struct llama_context * ctx_dft) { + auto * result = new common_speculative { + /* .ctx = */ ctx_dft, + /* .smpl = */ nullptr, + /* .batch = */ llama_batch_init(llama_n_batch(ctx_dft), 0, 1), + /* .prompt = */ {}, + }; + + // TODO: optimize or pass from outside? +#if 0 + { + common_params_sampling params; + params.no_perf = false; + + params.top_k = 40; + params.top_p = 0.9; + + params.samplers = { + COMMON_SAMPLER_TYPE_TOP_K, + COMMON_SAMPLER_TYPE_TOP_P, + COMMON_SAMPLER_TYPE_INFILL, + }; + + result->smpl = common_sampler_init(llama_get_model(ctx_dft), params); + } +#else + { + common_params_sampling params; + params.no_perf = false; + + params.top_k = 10; + + params.samplers = { + COMMON_SAMPLER_TYPE_TOP_K, + }; + + result->smpl = common_sampler_init(llama_get_model(ctx_dft), params); + } +#endif + + return result; +} + +void common_speculative_free(struct common_speculative * spec) { + if (spec == nullptr) { + return; + } + + common_sampler_free(spec->smpl); + + llama_batch_free(spec->batch); + + delete spec; +} + +bool common_speculative_are_compatible( + const struct llama_context * ctx_tgt, + const struct llama_context * ctx_dft) { + const struct llama_model * model_tgt = llama_get_model(ctx_tgt); + const struct llama_model * model_dft = llama_get_model(ctx_dft); + + const struct llama_vocab * vocab_tgt = llama_model_get_vocab(model_tgt); + const struct llama_vocab * vocab_dft = llama_model_get_vocab(model_dft); + + const bool vocab_type_tgt = llama_vocab_type(vocab_tgt); + LOG_DBG("%s: vocab_type tgt: %d\n", __func__, vocab_type_tgt); + + const bool vocab_type_dft = llama_vocab_type(vocab_dft); + LOG_DBG("%s: vocab_type dft: %d\n", __func__, vocab_type_dft); + + if (vocab_type_tgt != vocab_type_dft) { + LOG_ERR("%s: draft model vocab type must match target model to use speculation but " + "vocab_type_dft = %d while vocab_type_tgt = %d\n", __func__, vocab_type_dft, vocab_type_tgt); + return false; + } + + if (llama_vocab_get_add_bos(vocab_tgt) != llama_vocab_get_add_bos(vocab_dft) || + llama_vocab_get_add_eos(vocab_tgt) != llama_vocab_get_add_eos(vocab_dft) || + llama_vocab_bos(vocab_tgt) != llama_vocab_bos(vocab_dft) || + llama_vocab_eos(vocab_tgt) != llama_vocab_eos(vocab_dft)) { + LOG_ERR("%s: draft vocab special tokens must match target vocab to use speculation\n", __func__); + LOG_ERR("%s: tgt: bos = %d (%d), eos = %d (%d)\n", __func__, llama_vocab_bos(vocab_tgt), llama_vocab_get_add_bos(vocab_tgt), llama_vocab_eos(vocab_tgt), llama_vocab_get_add_eos(vocab_tgt)); + LOG_ERR("%s: dft: bos = %d (%d), eos = %d (%d)\n", __func__, llama_vocab_bos(vocab_dft), llama_vocab_get_add_bos(vocab_dft), llama_vocab_eos(vocab_dft), llama_vocab_get_add_eos(vocab_dft)); + return false; + } + + { + const int n_vocab_tgt = llama_vocab_n_tokens(vocab_tgt); + const int n_vocab_dft = llama_vocab_n_tokens(vocab_dft); + + const int vocab_diff = std::abs(n_vocab_tgt - n_vocab_dft); + + if (vocab_diff > SPEC_VOCAB_MAX_SIZE_DIFFERENCE) { + LOG_ERR("%s: draft model vocab must closely match target model to use speculation but " + "target vocab size %d does not match draft vocab size %d - difference %d, max allowed %d\n", + __func__, n_vocab_tgt, llama_vocab_n_tokens(vocab_dft), vocab_diff, SPEC_VOCAB_MAX_SIZE_DIFFERENCE); + return false; + } + + for (int i = SPEC_VOCAB_CHECK_START_TOKEN_ID; i < std::min(n_vocab_tgt, n_vocab_dft); ++i) { + const char * token_text_tgt = llama_vocab_get_text(vocab_tgt, i); + const char * token_text_dft = llama_vocab_get_text(vocab_dft, i); + if (std::strcmp(token_text_tgt, token_text_dft) != 0) { + LOG_ERR("%s: draft vocab vocab must match target vocab to use speculation but " + "token %d content differs - target '%s', draft '%s'\n", __func__, i, + common_token_to_piece(ctx_tgt, i).c_str(), + common_token_to_piece(ctx_dft, i).c_str()); + return false; + } + } + } + + return true; +} + +llama_tokens common_speculative_gen_draft( + struct common_speculative * spec, + struct common_speculative_params params, + const llama_tokens & prompt_tgt, + llama_token id_last) { + auto & batch = spec->batch; + auto & ctx = spec->ctx; + auto & smpl = spec->smpl; + auto & prompt = spec->prompt; + + int reuse_i = 0; + int reuse_n = 0; + + const int n_ctx = llama_n_ctx(ctx) - params.n_draft; + + const int i_start = std::max(0, (int) prompt_tgt.size() - n_ctx); + + // reuse as much as possible from the old draft context + // ideally, the draft context should be as big as the target context and we will always reuse the entire prompt + for (int i = 0; i < (int) prompt.size(); ++i) { + int cur = 0; + while (i_start + cur < (int) prompt_tgt.size() && + i + cur < (int) prompt.size() && + prompt_tgt[i_start + cur] == prompt[i + cur]) { + cur++; + } + + if ((cur >= params.n_reuse || n_ctx >= (int) prompt_tgt.size()) && cur > reuse_n) { + reuse_i = i; + reuse_n = cur; + } + } + + LOG_DBG("%s: reuse_i = %d, reuse_n = %d, prompt = %d\n", __func__, reuse_i, reuse_n, (int) prompt.size()); + + llama_tokens result; + result.reserve(params.n_draft); + + if (reuse_n == 0) { + llama_kv_cache_clear(ctx); + + prompt.clear(); + } else { + // this happens when a previous draft has been discarded (for example, due to being too small), but the + // target model agreed with it. in this case, we simply pass back the previous results to save compute + if (reuse_i + reuse_n < (int) prompt.size() && prompt[reuse_i + reuse_n] == id_last) { + for (int i = reuse_i + reuse_n + 1; i < (int) prompt.size(); ++i) { + result.push_back(prompt[i]); + + if (params.n_draft <= (int) result.size()) { + break; + } + } + + return result; + } + + if (reuse_i > 0) { + llama_kv_cache_seq_rm (ctx, 0, 0, reuse_i); + llama_kv_cache_seq_add(ctx, 0, reuse_i, -1, -reuse_i); + + prompt.erase(prompt.begin(), prompt.begin() + reuse_i); + } + + if (reuse_n < (int) prompt.size()) { + llama_kv_cache_seq_rm (ctx, 0, reuse_n, -1); + + prompt.erase(prompt.begin() + reuse_n, prompt.end()); + } + } + + // prepare a batch to evaluate any new tokens in the prompt + common_batch_clear(batch); + + for (size_t i = i_start + reuse_n; i < prompt_tgt.size(); ++i) { + //LOG_DBG("i = %d, i_start = %d, reuse_n = %d, i - i_start = %d, id = %6d\n", i, i_start, reuse_n, i - i_start, prompt_tgt[i]); + common_batch_add(batch, prompt_tgt[i], i - i_start, { 0 }, false); + + prompt.push_back(prompt_tgt[i]); + } + + // we should rarely end-up here during normal decoding + if (batch.n_tokens > 0) { + //LOG_DBG("%s: draft prompt batch: %s\n", __func__, string_from(ctx, batch).c_str()); + + llama_decode(ctx, batch); + } + + const llama_pos n_past = prompt.size(); + + LOG_DBG("%s: n_past = %d\n", __func__, n_past); + + common_batch_clear(batch); + common_batch_add (batch, id_last, n_past, { 0 }, true); + + prompt.push_back(id_last); + + //LOG_DBG("%s: draft prompt: %s\n", __func__, string_from(ctx, prompt).c_str()); + + llama_decode(ctx, batch); + + common_sampler_reset(smpl); + + // sample n_draft tokens from the draft model + for (int i = 0; i < params.n_draft; ++i) { + common_batch_clear(batch); + + common_sampler_sample(smpl, ctx, 0, true); + + const auto * cur_p = common_sampler_get_candidates(smpl); + + for (int k = 0; k < std::min(3, (int) cur_p->size); ++k) { + LOG_DBG(" - draft candidate %3d, pos %3d: %6d (%8.3f) '%s'\n", + k, i, cur_p->data[k].id, cur_p->data[k].p, common_token_to_piece(ctx, cur_p->data[k].id).c_str()); + } + + // add drafted token for each sequence + const llama_token id = cur_p->data[0].id; + + // only collect very high-confidence draft tokens + if (cur_p->data[0].p < params.p_min) { + break; + } + + common_sampler_accept(smpl, id, true); + + result.push_back(id); + + if (params.n_draft <= (int) result.size()) { + break; + } + + common_batch_add(batch, id, n_past + i + 1, { 0 }, true); + + // evaluate the drafted tokens on the draft model + llama_decode(ctx, batch); + + prompt.push_back(id); + } + + return result; +} diff --git a/common/speculative.h b/common/speculative.h new file mode 100644 index 000000000..50ec03446 --- /dev/null +++ b/common/speculative.h @@ -0,0 +1,28 @@ +#pragma once + +#include "llama.h" +#include "common.h" + +struct common_speculative; + +struct common_speculative_params { + int n_draft = 16; // max drafted tokens + int n_reuse = 256; + + float p_min = 0.9f; // min probabiliy required to accept a token in the draft +}; + +struct common_speculative * common_speculative_init(struct llama_context * ctx_dft); + +void common_speculative_free(struct common_speculative * spec); + +bool common_speculative_are_compatible( + const struct llama_context * ctx_tgt, + const struct llama_context * ctx_dft); + +// sample up to n_draft tokens and add them to the batch using the draft model +llama_tokens common_speculative_gen_draft( + struct common_speculative * spec, + struct common_speculative_params params, + const llama_tokens & prompt, + llama_token id_last); diff --git a/convert_hf_to_gguf.py b/convert_hf_to_gguf.py index 39afa5ef4..4dc9837ab 100755 --- a/convert_hf_to_gguf.py +++ b/convert_hf_to_gguf.py @@ -221,17 +221,17 @@ class Model: self.gguf_writer.add_context_length(n_ctx) logger.info(f"gguf: context length = {n_ctx}") - n_embd = self.find_hparam(["hidden_size", "n_embd"]) - self.gguf_writer.add_embedding_length(n_embd) - logger.info(f"gguf: embedding length = {n_embd}") + if (n_embd := self.find_hparam(["hidden_size", "n_embd"], optional=True)) is not None: + self.gguf_writer.add_embedding_length(n_embd) + logger.info(f"gguf: embedding length = {n_embd}") if (n_ff := self.find_hparam(["intermediate_size", "n_inner"], optional=True)) is not None: self.gguf_writer.add_feed_forward_length(n_ff) logger.info(f"gguf: feed forward length = {n_ff}") - n_head = self.find_hparam(["num_attention_heads", "n_head"]) - self.gguf_writer.add_head_count(n_head) - logger.info(f"gguf: head count = {n_head}") + if (n_head := self.find_hparam(["num_attention_heads", "n_head"], optional=True)) is not None: + self.gguf_writer.add_head_count(n_head) + logger.info(f"gguf: head count = {n_head}") if (n_head_kv := self.hparams.get("num_key_value_heads")) is not None: self.gguf_writer.add_head_count_kv(n_head_kv) @@ -296,7 +296,9 @@ class Model: break for new_name, data_torch in (self.modify_tensors(data_torch, name, bid)): - data = data_torch.squeeze().numpy() + # TODO: why do we squeeze here? + # data = data_torch.squeeze().numpy() + data = data_torch.numpy() # if data ends up empty, it means data_torch was a scalar tensor -> restore if len(data.shape) == 0: @@ -324,6 +326,9 @@ class Model: gguf.MODEL_TENSOR.TIME_MIX_W2, gguf.MODEL_TENSOR.TIME_MIX_DECAY_W1, gguf.MODEL_TENSOR.TIME_MIX_DECAY_W2, + gguf.MODEL_TENSOR.TIME_MIX_LERP_FUSED, + gguf.MODEL_TENSOR.POSNET_NORM1, + gguf.MODEL_TENSOR.POSNET_NORM2, ) ) or not new_name.endswith(".weight") @@ -473,6 +478,11 @@ class Model: return modelcls return func + @classmethod + def print_registered_models(cls): + for name in sorted(cls._model_classes.keys()): + logger.error(f"- {name}") + @classmethod def from_model_architecture(cls, arch: str) -> type[Model]: try: @@ -525,9 +535,19 @@ class Model: else: token: str = reverse_vocab[i] if token in added_vocab: + # The tokenizer in llama.cpp assumes the CONTROL and USER_DEFINED tokens are pre-normalized. + # To avoid unexpected issues - we make sure to normalize non-normalized tokens + if not tokenizer.added_tokens_decoder[i].normalized: + previous_token = token + token = tokenizer.decode(tokenizer.encode(token, add_special_tokens=False)) + if previous_token != token: + logger.info(f"{repr(previous_token)} is encoded and decoded back to {repr(token)} using AutoTokenizer") + if tokenizer.added_tokens_decoder[i].special or self.does_token_look_special(token): toktypes.append(gguf.TokenType.CONTROL) else: + # NOTE: this was added for Gemma. + # Encoding and decoding the tokens above isn't sufficient for this case. token = token.replace(b"\xe2\x96\x81".decode("utf-8"), " ") # pre-normalize user-defined spaces toktypes.append(gguf.TokenType.USER_DEFINED) else: @@ -571,6 +591,9 @@ class Model: if chkhsh == "8aeee3860c56296a157a1fe2fad249ec40aa59b1bb5709f4ade11c4e6fe652ed": # ref: https://huggingface.co/tiiuae/falcon-7b res = "falcon" + if chkhsh == "9d032fcbd5501f4a38150912590928bfb36091efb5df11b8e2124b0390e3fb1e": + # ref: https://huggingface.co/tiiuae/Falcon3-7B-Base + res = "falcon3" if chkhsh == "0876d13b50744004aa9aeae05e7b0647eac9d801b5ba4668afc01e709c15e19f": # ref: https://huggingface.co/BAAI/bge-small-en-v1.5 res = "bert-bge" @@ -658,6 +681,21 @@ class Model: if chkhsh == "60824e3c0d9401f89943cbb2fff727f0e2d4c545ba4df2d6e4f09a6db0f5b450": # ref: https://huggingface.co/facebook/chameleon-7b res = "chameleon" + if chkhsh == "1431a23e583c97432bc230bff598d103ddb5a1f89960c8f1d1051aaa944d0b35": + # ref: https://huggingface.co/sapienzanlp/Minerva-7B-base-v1.0 + res = "minerva-7b" + if chkhsh == "8b5a93ed704057481f240da0be7e7dca721d7f8f4755263b6807227a2cbeae65": + # ref: https://huggingface.co/sentence-transformers/stsb-roberta-base + res = "roberta-bpe" + if chkhsh == "ad851be1dba641f2e3711822f816db2c265f788b37c63b4e1aeacb9ee92de8eb": + # ref: https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct + res = "gigachat" + if chkhsh == "d4c8f286ea6b520b3d495c4455483cfa2302c0cfcd4be05d781b6a8a0a7cdaf1": + # ref: https://huggingface.co/Infinigence/Megrez-3B-Instruct + res = "megrez" + if chkhsh == "877081d19cf6996e2c4ff0e1236341e9b7bde288f5311a56a937f0afbbb3aeb5": + # ref: https://huggingface.co/deepseek-ai/DeepSeek-V3 + res = "deepseek-v3" if res is None: logger.warning("\n") @@ -680,6 +718,9 @@ class Model: return res # Marker: End get_vocab_base_pre + def _set_vocab_none(self) -> None: + self.gguf_writer.add_tokenizer_model("none") + def _set_vocab_gpt2(self) -> None: tokens, toktypes, tokpre = self.get_vocab_base() self.gguf_writer.add_tokenizer_model("gpt2") @@ -1663,6 +1704,178 @@ class LlamaModel(Model): raise ValueError(f"Unprocessed experts: {experts}") +@Model.register("DeciLMForCausalLM") +class DeciModel(Model): + model_arch = gguf.MODEL_ARCH.DECI + + @staticmethod + def _ffn_mult_to_intermediate_size(ffn_mult: float, n_embd: int) -> int: + # DeciLM-specific code + intermediate_size = int(2 * ffn_mult * n_embd / 3) + return DeciModel._find_multiple(intermediate_size, 256) + + @staticmethod + def _find_multiple(n: int, k: int) -> int: + # DeciLM-specific code + if n % k == 0: + return n + return n + k - (n % k) + + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + + if "block_configs" in self.hparams: # Llama-3_1-Nemotron-51B + _block_configs: list[dict[str,Any]] = self.hparams["block_configs"] + assert self.block_count == len(_block_configs) + self._num_kv_heads = list() + self._num_heads = list() + _ffn_multipliers = list() + # ***linear attention layer*** + # if n_heads_in_group is None and replace_with_linear is True + # then _num_kv_heads[il] is 0 and _num_heads[il] is num_attention_heads + # ***attention-free layer*** + # if n_heads_in_group is None and replace_with_linear is False + # then _num_kv_heads[il] is 0 and _num_heads[il] is 0 + # ***normal attention-layer*** + # if n_heads_in_group is not None, then + # _num_kv_heads[il] is num_attention_head // n_heads_in_group and + # _num_heads[il] is num_attention_head + for il in range(len(_block_configs)): + if _block_configs[il]["attention"]["n_heads_in_group"] is None: + if _block_configs[il]["attention"]["replace_with_linear"] is True: + self._num_kv_heads.append(0) + self._num_heads.append(self.hparams["num_attention_heads"]) + else: + self._num_kv_heads.append(0) + self._num_heads.append(0) + else: + self._num_kv_heads.append(self.hparams["num_attention_heads"] // _block_configs[il]["attention"]["n_heads_in_group"]) + self._num_heads.append(self.hparams["num_attention_heads"]) + _ffn_multipliers.append(_block_configs[il]["ffn"]["ffn_mult"]) + assert self.block_count == len(self._num_kv_heads) + assert self.block_count == len(self._num_heads) + assert self.block_count == len(_ffn_multipliers) + assert isinstance(self._num_kv_heads, list) and isinstance(self._num_kv_heads[0], int) + assert isinstance(self._num_heads, list) and isinstance(self._num_heads[0], int) + assert isinstance(_ffn_multipliers, list) and isinstance(_ffn_multipliers[0], float) + self._ffn_dims: list[int] = [ + DeciModel._ffn_mult_to_intermediate_size(multiplier, self.hparams["hidden_size"]) + for multiplier in _ffn_multipliers + ] + + def set_vocab(self): + # Please change tokenizer_config.json of Llama-3_1-Nemotron-51B's + # eos_token from '|eot_id|' to '|end_of_text|' + if self.hparams.get("vocab_size", 128256) == 128256: + tokens, toktypes, tokpre = self.get_vocab_base() + self.gguf_writer.add_tokenizer_model("gpt2") + self.gguf_writer.add_tokenizer_pre(tokpre) + self.gguf_writer.add_token_list(tokens) + self.gguf_writer.add_token_types(toktypes) + + special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=True) + special_vocab.add_to_gguf(self.gguf_writer) + else: + # DeciLM-7B + self._set_vocab_llama_hf() + + def set_gguf_parameters(self): + if "block_configs" in self.hparams: # Llama-3_1-Nemotron-51B + assert self.block_count == len(self._num_kv_heads) + assert self.block_count == len(self._num_heads) + assert self.block_count == len(self._ffn_dims) + if (rope_theta := self.hparams.get("rope_theta")) is not None: + self.gguf_writer.add_rope_freq_base(rope_theta) + self.gguf_writer.add_head_count_kv(self._num_kv_heads) + self.gguf_writer.add_head_count(self._num_heads) + self.gguf_writer.add_feed_forward_length(self._ffn_dims) + self.gguf_writer.add_block_count(self.block_count) + self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) + self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) + self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"]) + self.gguf_writer.add_key_length(self.hparams["hidden_size"] // self.hparams["num_attention_heads"]) + self.gguf_writer.add_value_length(self.hparams["hidden_size"] // self.hparams["num_attention_heads"]) + self.gguf_writer.add_file_type(self.ftype) + else: # DeciLM-7B + super().set_gguf_parameters() + if "num_key_value_heads_per_layer" in self.hparams: # DeciLM-7B + self._num_kv_heads: list[int] = self.hparams["num_key_value_heads_per_layer"] + assert self.block_count == len(self._num_kv_heads) + self.gguf_writer.add_head_count_kv(self._num_kv_heads) + hparams = self.hparams + self.gguf_writer.add_vocab_size(hparams["vocab_size"]) + + if "head_dim" in hparams: + rope_dim = hparams["head_dim"] + else: + rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"] + self.gguf_writer.add_rope_dimension_count(rope_dim) + + if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]: + if self.hparams["rope_scaling"].get("type") == "linear": + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR) + self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"]) + + @staticmethod + def permute(weights: Tensor, n_head: int, n_head_kv: int | None): + if n_head_kv is not None and n_head != n_head_kv: + n_head = n_head_kv + return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:]) + .swapaxes(1, 2) + .reshape(weights.shape)) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + n_head = self.hparams["num_attention_heads"] + if bid is not None: + if "num_key_value_heads_per_layer" in self.hparams: + n_kv_head = self.hparams["num_key_value_heads_per_layer"][bid] + elif "block_configs" in self.hparams: + n_kv_head = self._num_kv_heads[bid] + n_head = self._num_heads[bid] + else: + n_kv_head = self.hparams.get("num_key_value_heads") + else: + n_kv_head = self.hparams.get("num_key_value_heads") + + if name.endswith(("q_proj.weight", "q_proj.bias")): + data_torch = DeciModel.permute(data_torch, n_head, n_head) + if name.endswith(("k_proj.weight", "k_proj.bias")): + data_torch = DeciModel.permute(data_torch, n_head, n_kv_head) + return [(self.map_tensor_name(name), data_torch)] + + def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]: + if rope_scaling := self.find_hparam(["rope_scaling"], optional=True): + if rope_scaling.get("rope_type", '').lower() == "llama3": + base = self.hparams.get("rope_theta", 10000.0) + dim = self.hparams.get("head_dim", self.hparams["hidden_size"] // self.hparams["num_attention_heads"]) + freqs = 1.0 / (base ** (torch.arange(0, dim, 2, dtype=torch.float32) / dim)) + + factor = rope_scaling.get("factor", 8.0) + low_freq_factor = rope_scaling.get("low_freq_factor", 1.0) + high_freq_factor = rope_scaling.get("high_freq_factor", 4.0) + old_context_len = self.hparams.get("original_max_position_embeddings", 8192) + + low_freq_wavelen = old_context_len / low_freq_factor + high_freq_wavelen = old_context_len / high_freq_factor + assert low_freq_wavelen != high_freq_wavelen + + rope_factors = [] + for freq in freqs: + wavelen = 2 * math.pi / freq + if wavelen < high_freq_wavelen: + rope_factors.append(1) + elif wavelen > low_freq_wavelen: + rope_factors.append(factor) + else: + smooth = (old_context_len / wavelen - low_freq_factor) / (high_freq_factor - low_freq_factor) + rope_factors.append(1 / ((1 - smooth) / factor + smooth)) + + yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FREQS), torch.tensor(rope_factors, dtype=torch.float32)) + + def prepare_tensors(self): + super().prepare_tensors() + + @Model.register("BitnetForCausalLM") class BitnetModel(Model): model_arch = gguf.MODEL_ARCH.BITNET @@ -1831,29 +2044,40 @@ class MiniCPMModel(Model): model_arch = gguf.MODEL_ARCH.MINICPM def set_gguf_parameters(self): - block_count = self.hparams["num_hidden_layers"] - self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) - self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) - self.gguf_writer.add_block_count(block_count) - self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"]) - self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"]) - self.gguf_writer.add_head_count(self.hparams["num_attention_heads"]) - self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"]) - self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"]) - self.gguf_writer.add_file_type(self.ftype) + super().set_gguf_parameters() + embedding_scale = float(self.hparams["scale_emb"]) + self.gguf_writer.add_embedding_scale(embedding_scale) + logger.info(f"gguf: (minicpm) embedding_scale = {embedding_scale}") + residual_scale = self.hparams["scale_depth"] / self.hparams["num_hidden_layers"] ** 0.5 + self.gguf_writer.add_residual_scale(residual_scale) + logger.info(f"gguf: (minicpm) residual_scale = {residual_scale}") + logit_scale = self.hparams["hidden_size"] / self.hparams["dim_model_base"] + self.gguf_writer.add_logit_scale(logit_scale) + logger.info(f"gguf: (minicpm) logit_scale = {logit_scale}") + if self.hparams.get("rope_scaling") is not None: + if self.hparams["rope_scaling"].get("type") == "longrope": + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LONGROPE) + logger.info(f"gguf: (minicpm) rope_scaling_type = {gguf.RopeScalingType.LONGROPE}") + + def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]: + rope_dims = self.hparams["hidden_size"] // self.hparams["num_attention_heads"] + + rope_scaling = self.find_hparam(['rope_scaling'], True) + if rope_scaling is not None: + long_factors = rope_scaling.get('long_factor', None) + short_factors = rope_scaling.get('short_factor', None) + + if long_factors is None or short_factors is None: + raise KeyError('Missing the required key rope_scaling.long_factor or rope_scaling_short_factor') + + if len(long_factors) != len(short_factors) or len(long_factors) != rope_dims / 2: + raise ValueError(f'The length of rope long and short factors must be {rope_dims / 2}') + + yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_LONG), torch.tensor(long_factors, dtype=torch.float32)) + yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT), torch.tensor(short_factors, dtype=torch.float32)) def set_vocab(self): - self._set_vocab_llama_hf() - - def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor: - if n_kv_head is not None and n_head != n_kv_head: - n_head //= n_kv_head - - return ( - weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:]) - .swapaxes(1, 2) - .reshape(weights.shape) - ) + self._set_vocab_sentencepiece() def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: del bid # unused @@ -1863,9 +2087,9 @@ class MiniCPMModel(Model): # HF models permute some of the tensors, so we need to undo that if name.endswith(("q_proj.weight")): - data_torch = self._reverse_hf_permute(data_torch, n_head, n_head) + data_torch = LlamaModel.permute(data_torch, n_head, n_head) if name.endswith(("k_proj.weight")): - data_torch = self._reverse_hf_permute(data_torch, n_head, n_kv_head) + data_torch = LlamaModel.permute(data_torch, n_head, n_kv_head) return [(self.map_tensor_name(name), data_torch)] @@ -1975,6 +2199,75 @@ class Qwen2Model(Model): except FileNotFoundError: self._set_vocab_gpt2() + def set_gguf_parameters(self): + super().set_gguf_parameters() + if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]: + if self.hparams["rope_scaling"].get("type") == "yarn": + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN) + self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"]) + self.gguf_writer.add_rope_scaling_orig_ctx_len(self.hparams["rope_scaling"]["original_max_position_embeddings"]) + + +@Model.register("Qwen2VLForConditionalGeneration") +class Qwen2VLModel(Model): + model_arch = gguf.MODEL_ARCH.QWEN2VL + + def set_gguf_parameters(self): + super().set_gguf_parameters() + mrope_section = self.hparams["rope_scaling"]["mrope_section"] + mrope_section += [0] * max(0, 4 - len(mrope_section)) + self.gguf_writer.add_rope_dimension_sections(mrope_section) + + def set_vocab(self): + try: + self._set_vocab_sentencepiece() + except FileNotFoundError: + self._set_vocab_gpt2() + + def get_tensors(self) -> Iterator[tuple[str, Tensor]]: + for name, data in super().get_tensors(): + if name.startswith("visual."): + continue + yield name, data + + +@Model.register("WavTokenizerDec") +class WavTokenizerDecModel(Model): + model_arch = gguf.MODEL_ARCH.WAVTOKENIZER_DEC + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + del bid # unused + + if \ + name.endswith("codebook.cluster_size") or \ + name.endswith("codebook.embed_avg") or \ + name.endswith("codebook.inited"): + logger.debug(f"Skipping {name!r}") + return [] + + logger.info(f"{self.map_tensor_name(name)} -> {data_torch.shape}") + + return [(self.map_tensor_name(name), data_torch)] + + def set_vocab(self): + self._set_vocab_none() + + def set_gguf_parameters(self): + super().set_gguf_parameters() + self.gguf_writer.add_vocab_size (self.hparams["vocab_size"]) + self.gguf_writer.add_features_length (self.hparams["n_embd_features"]) + self.gguf_writer.add_feed_forward_length(self.hparams["n_ff"]) + self.gguf_writer.add_group_norm_eps (self.hparams["group_norm_epsilon"]) + self.gguf_writer.add_group_norm_groups (self.hparams["group_norm_groups"]) + + self.gguf_writer.add_posnet_embedding_length(self.hparams["posnet"]["n_embd"]) + self.gguf_writer.add_posnet_block_count (self.hparams["posnet"]["n_layer"]) + + self.gguf_writer.add_convnext_embedding_length(self.hparams["convnext"]["n_embd"]) + self.gguf_writer.add_convnext_block_count (self.hparams["convnext"]["n_layer"]) + + self.gguf_writer.add_causal_attention(False) + @Model.register("Qwen2MoeForCausalLM") class Qwen2MoeModel(Model): @@ -2104,6 +2397,15 @@ class Phi3MiniModel(Model): model_arch = gguf.MODEL_ARCH.PHI3 def set_vocab(self): + # Phi-4 model uses GPT2Tokenizer + tokenizer_config_file = self.dir_model / 'tokenizer_config.json' + if tokenizer_config_file.is_file(): + with open(tokenizer_config_file, "r", encoding="utf-8") as f: + tokenizer_config_json = json.load(f) + tokenizer_class = tokenizer_config_json['tokenizer_class'] + if tokenizer_class == 'GPT2Tokenizer': + return self._set_vocab_gpt2() + from sentencepiece import SentencePieceProcessor tokenizer_path = self.dir_model / 'tokenizer.model' @@ -2220,7 +2522,11 @@ class Phi3MiniModel(Model): self.gguf_writer.add_rope_dimension_count(rope_dims) self.gguf_writer.add_rope_freq_base(self.find_hparam(["rope_theta"])) self.gguf_writer.add_file_type(self.ftype) - self.gguf_writer.add_sliding_window(self.find_hparam(["sliding_window"])) + sliding_window = self.hparams.get("sliding_window") + # use zero value of sliding_window to distinguish Phi-4 from other PHI3 models + if sliding_window is None: + sliding_window = 0 + self.gguf_writer.add_sliding_window(sliding_window) def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]: n_embd = self.find_hparam(["hidden_size", "n_embd"]) @@ -2262,6 +2568,63 @@ class Phi3MiniModel(Model): yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT), torch.tensor(short_factors, dtype=torch.float32)) +@Model.register("PhiMoEForCausalLM") +class PhiMoeModel(Phi3MiniModel): + model_arch = gguf.MODEL_ARCH.PHIMOE + + _experts: list[dict[str, Tensor]] | None = None + + def set_gguf_parameters(self): + super().set_gguf_parameters() + self.gguf_writer.add_expert_used_count(self.hparams["num_experts_per_tok"]) + self.gguf_writer.add_expert_count(self.hparams["num_local_experts"]) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + # process the experts separately + if name.find("block_sparse_moe.experts") != -1: + n_experts = self.hparams["num_local_experts"] + assert bid is not None + + if self._experts is None: + self._experts = [{} for _ in range(self.block_count)] + + self._experts[bid][name] = data_torch + + if len(self._experts[bid]) >= n_experts * 3: + tensors: list[tuple[str, Tensor]] = [] + + # merge the experts into a single 3d tensor + for w_name in ["w1", "w2", "w3"]: + datas: list[Tensor] = [] + + for xid in range(n_experts): + ename = f"model.layers.{bid}.block_sparse_moe.experts.{xid}.{w_name}.weight" + datas.append(self._experts[bid][ename]) + del self._experts[bid][ename] + + data_torch = torch.stack(datas, dim=0) + + merged_name = f"model.layers.{bid}.block_sparse_moe.experts.{w_name}.weight" + + new_name = self.map_tensor_name(merged_name) + + tensors.append((new_name, data_torch)) + return tensors + else: + return [] + + return [(self.map_tensor_name(name), data_torch)] + + def prepare_tensors(self): + super().prepare_tensors() + + if self._experts is not None: + # flatten `list[dict[str, Tensor]]` into `list[str]` + experts = [k for d in self._experts for k in d.keys()] + if len(experts) > 0: + raise ValueError(f"Unprocessed experts: {experts}") + + @Model.register("PlamoForCausalLM") class PlamoModel(Model): model_arch = gguf.MODEL_ARCH.PLAMO @@ -2519,7 +2882,7 @@ class InternLM2Model(Model): return [(self.map_tensor_name(name), data_torch)] -@Model.register("BertModel", "CamembertModel") +@Model.register("BertModel", "BertForMaskedLM", "CamembertModel") class BertModel(Model): model_arch = gguf.MODEL_ARCH.BERT @@ -2560,7 +2923,8 @@ class BertModel(Model): # we need this to validate the size of the token_type embeddings # though currently we are passing all zeros to the token_type embeddings - self.gguf_writer.add_token_type_count(2) # "Sequence A" or "Sequence B" + # "Sequence A" or "Sequence B" + self.gguf_writer.add_token_type_count(self.hparams.get("type_vocab_size", 1)) # convert to phantom space vocab def phantom(tok): @@ -2584,13 +2948,73 @@ class BertModel(Model): def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: del bid # unused + if name.startswith("bert."): + name = name[5:] + + if name.endswith(".gamma"): + name = name[:-6] + ".weight" + + if name.endswith(".beta"): + name = name[:-5] + ".bias" + # we are only using BERT for embeddings so we don't need the pooling layer if name in ("embeddings.position_ids", "pooler.dense.weight", "pooler.dense.bias"): return [] # we don't need these + if name.startswith("cls.predictions"): + return [] + + if name.startswith("cls.seq_relationship"): + return [] + return [(self.map_tensor_name(name), data_torch)] +@Model.register("RobertaModel") +class RobertaModel(BertModel): + model_arch = gguf.MODEL_ARCH.BERT + + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + + # we need the pad_token_id to know how to chop down position_embd matrix + if (pad_token_id := self.hparams.get("pad_token_id")) is not None: + self._position_offset = 1 + pad_token_id + if "max_position_embeddings" in self.hparams: + self.hparams["max_position_embeddings"] -= self._position_offset + else: + self._position_offset = None + + def set_vocab(self): + """Support BPE tokenizers for roberta models""" + bpe_tok_path = self.dir_model / "tokenizer.json" + if bpe_tok_path.exists(): + self._set_vocab_gpt2() + self.gguf_writer.add_add_bos_token(True) + self.gguf_writer.add_add_eos_token(True) + + # we need this to validate the size of the token_type embeddings + # though currently we are passing all zeros to the token_type embeddings + # "Sequence A" or "Sequence B" + self.gguf_writer.add_token_type_count(self.hparams.get("type_vocab_size", 1)) + + else: + return super().set_vocab() + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + # if name starts with "roberta.", remove the prefix + # e.g. https://huggingface.co/BAAI/bge-reranker-v2-m3/tree/main + if name.startswith("roberta."): + name = name[8:] + + # position embeddings start at pad_token_id + 1, so just chop down the weight tensor + if name == "embeddings.position_embeddings.weight": + if self._position_offset is not None: + data_torch = data_torch[self._position_offset:,:] + + return super().modify_tensors(data_torch, name, bid) + + @Model.register("NomicBertModel") class NomicBertModel(BertModel): model_arch = gguf.MODEL_ARCH.NOMIC_BERT @@ -2707,7 +3131,7 @@ class XLMRobertaModel(BertModel): self.gguf_writer.add_token_scores(scores) self.gguf_writer.add_token_types(toktypes) self.gguf_writer.add_add_space_prefix(add_prefix) - self.gguf_writer.add_token_type_count(1) + self.gguf_writer.add_token_type_count(self.hparams.get("type_vocab_size", 1)) self.gguf_writer.add_remove_extra_whitespaces(remove_whitespaces) if precompiled_charsmap: self.gguf_writer.add_precompiled_charsmap(precompiled_charsmap) @@ -2898,6 +3322,8 @@ class Rwkv6Model(Model): # required by llama.cpp, unused self.gguf_writer.add_head_count(0) + lerp_weights: dict[int, dict[str, Tensor]] = {} + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: new_name = self.map_tensor_name(name) @@ -2910,14 +3336,87 @@ class Rwkv6Model(Model): if new_name.endswith("time_mix_w2.weight"): data_torch = data_torch.permute(0, 2, 1) - rescale_every_n_layers = self.hparams["rescale_every"] - if rescale_every_n_layers > 0: - if new_name.endswith("time_mix_output.weight") or new_name.endswith("channel_mix_value.weight"): - data_torch = data_torch.div_(2 ** int(bid // rescale_every_n_layers)) + if new_name.endswith("time_mix_decay.weight") or "lerp" in new_name: + data_torch = data_torch.squeeze() + + try: + rescale_every_n_layers = self.hparams["rescale_every"] + if rescale_every_n_layers > 0: + if new_name.endswith("time_mix_output.weight") or new_name.endswith("channel_mix_value.weight"): + data_torch = data_torch.div_(2 ** int(bid // rescale_every_n_layers)) + except KeyError: + pass + + # concat time_mix_lerp weights to reduce some cpu overhead + # also reduces the number of tensors in the model + if bid is not None and "time_mix_lerp" in new_name and "time_mix_lerp_x" not in new_name: + try: + self.lerp_weights[bid][new_name] = data_torch + except KeyError: + self.lerp_weights[bid] = {new_name: data_torch} + if all(f"blk.{bid}.time_mix_lerp_{i}.weight" in self.lerp_weights[bid].keys() for i in ["w", "k", "v", "r", "g"]): + new_name = f"blk.{bid}.time_mix_lerp_fused.weight" + data = torch.stack([self.lerp_weights[bid][f"blk.{bid}.time_mix_lerp_{i}.weight"].unsqueeze(0) for i in ["w", "k", "v", "r", "g"]], dim=0).unsqueeze(1) + yield (new_name, data) + return yield (new_name, data_torch) +@Model.register("RWKV6Qwen2ForCausalLM") +class RWKV6Qwen2Model(Rwkv6Model): + model_arch = gguf.MODEL_ARCH.RWKV6QWEN2 + + def set_vocab(self): + try: + self._set_vocab_sentencepiece() + except FileNotFoundError: + self._set_vocab_gpt2() + + def set_gguf_parameters(self): + block_count = self.hparams["num_hidden_layers"] + num_attention_heads = self.hparams["num_attention_heads"] + num_key_value_heads = self.hparams["num_key_value_heads"] + hidden_size = self.hparams["hidden_size"] + head_size = hidden_size // num_attention_heads + rms_norm_eps = self.hparams["rms_norm_eps"] + intermediate_size = self.hparams["intermediate_size"] + time_mix_extra_dim = 64 if hidden_size >= 4096 else 32 + time_decay_extra_dim = 128 if hidden_size >= 4096 else 64 + + # RWKV isn't context limited + self.gguf_writer.add_context_length(1048576) + self.gguf_writer.add_embedding_length(hidden_size) + self.gguf_writer.add_block_count(block_count) + self.gguf_writer.add_wkv_head_size(head_size) + self.gguf_writer.add_time_mix_extra_dim(time_mix_extra_dim) + self.gguf_writer.add_time_decay_extra_dim(time_decay_extra_dim) + self.gguf_writer.add_feed_forward_length(intermediate_size) + self.gguf_writer.add_file_type(self.ftype) + + # special parameters for time_mixing in RWKV6QWEN2 + self.gguf_writer.add_layer_norm_rms_eps(rms_norm_eps) + self.gguf_writer.add_token_shift_count(1) + # RWKV6QWEN2 use grouped key/value like GQA + self.gguf_writer.add_head_count_kv(num_key_value_heads) + + # required by llama.cpp, unused + self.gguf_writer.add_head_count(0) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + for new_name, data in super().modify_tensors(data_torch, name, bid): + if "time_mix_w1" in new_name or "time_mix_w2" in new_name: + data = data.view(5, -1, data.shape[-1]) + # rwkv6qwen2 has a different order of rkvwg instead of the original wkvrg + # permute them here to avoid code changes + data = torch.stack([data[3], data[1], data[2], data[0], data[4]], dim=0).view(-1, data.shape[-1]) + if "w2" in new_name: + data = data.view(5, -1, data.shape[-1]) + yield (new_name, data) + continue + yield (new_name, data) + + @Model.register("MambaForCausalLM", "MambaLMHeadModel", "FalconMambaForCausalLM") class MambaModel(Model): model_arch = gguf.MODEL_ARCH.MAMBA @@ -3012,6 +3511,24 @@ class CommandR2Model(Model): self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE) +@Model.register("Cohere2ForCausalLM") +class Cohere2Model(Model): + model_arch = gguf.MODEL_ARCH.COHERE2 + + def set_gguf_parameters(self): + super().set_gguf_parameters() + + self.gguf_writer.add_logit_scale(self.hparams["logit_scale"]) + self.gguf_writer.add_sliding_window(self.hparams["sliding_window"]) + self.gguf_writer.add_vocab_size(self.hparams["vocab_size"]) + + rotary_pct = self.hparams["rotary_pct"] + hidden_size = self.hparams["hidden_size"] + num_attention_heads = self.hparams["num_attention_heads"] + self.gguf_writer.add_rope_dimension_count(int(rotary_pct * (hidden_size // num_attention_heads))) + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE) + + @Model.register("OlmoForCausalLM") @Model.register("OLMoForCausalLM") class OlmoModel(Model): @@ -3040,6 +3557,11 @@ class OlmoModel(Model): return [(self.map_tensor_name(name), data_torch)] +@Model.register("Olmo2ForCausalLM") +class Olmo2Model(Model): + model_arch = gguf.MODEL_ARCH.OLMO2 + + @Model.register("OlmoeForCausalLM") class OlmoeModel(Model): model_arch = gguf.MODEL_ARCH.OLMOE @@ -3373,7 +3895,99 @@ class ArcticModel(Model): raise ValueError(f"Unprocessed experts: {experts}") +@Model.register("DeepseekForCausalLM") +class DeepseekModel(Model): + model_arch = gguf.MODEL_ARCH.DEEPSEEK + + def set_vocab(self): + try: + self._set_vocab_sentencepiece() + except FileNotFoundError: + self._set_vocab_gpt2() + + def set_gguf_parameters(self): + super().set_gguf_parameters() + hparams = self.hparams + if "head_dim" in hparams: + rope_dim = hparams["head_dim"] + else: + rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"] + + self.gguf_writer.add_rope_dimension_count(rope_dim) + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE) + self.gguf_writer.add_leading_dense_block_count(hparams["first_k_dense_replace"]) + self.gguf_writer.add_vocab_size(hparams["vocab_size"]) + self.gguf_writer.add_expert_feed_forward_length(hparams["moe_intermediate_size"]) + self.gguf_writer.add_expert_weights_scale(1.0) + self.gguf_writer.add_expert_count(hparams["n_routed_experts"]) + self.gguf_writer.add_expert_shared_count(hparams["n_shared_experts"]) + + _experts: list[dict[str, Tensor]] | None = None + + @staticmethod + def permute(weights: Tensor, n_head: int, n_head_kv: int | None): + if n_head_kv is not None and n_head != n_head_kv: + n_head = n_head_kv + return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:]) + .swapaxes(1, 2) + .reshape(weights.shape)) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + n_head = self.hparams["num_attention_heads"] + n_kv_head = self.hparams.get("num_key_value_heads") + + if name.endswith(("q_proj.weight", "q_proj.bias")): + data_torch = DeepseekModel.permute(data_torch, n_head, n_head) + if name.endswith(("k_proj.weight", "k_proj.bias")): + data_torch = DeepseekModel.permute(data_torch, n_head, n_kv_head) + + # process the experts separately + if name.find("mlp.experts") != -1: + n_experts = self.hparams["n_routed_experts"] + assert bid is not None + + if self._experts is None: + self._experts = [{} for _ in range(self.block_count)] + + self._experts[bid][name] = data_torch + + if len(self._experts[bid]) >= n_experts * 3: + tensors: list[tuple[str, Tensor]] = [] + + # merge the experts into a single 3d tensor + for w_name in ["down_proj", "gate_proj", "up_proj"]: + datas: list[Tensor] = [] + + for xid in range(n_experts): + ename = f"model.layers.{bid}.mlp.experts.{xid}.{w_name}.weight" + datas.append(self._experts[bid][ename]) + del self._experts[bid][ename] + + data_torch = torch.stack(datas, dim=0) + + merged_name = f"model.layers.{bid}.mlp.experts.{w_name}.weight" + + new_name = self.map_tensor_name(merged_name) + + tensors.append((new_name, data_torch)) + return tensors + else: + return [] + + return [(self.map_tensor_name(name), data_torch)] + + def prepare_tensors(self): + super().prepare_tensors() + + if self._experts is not None: + # flatten `list[dict[str, Tensor]]` into `list[str]` + experts = [k for d in self._experts for k in d.keys()] + if len(experts) > 0: + raise ValueError(f"Unprocessed experts: {experts}") + + @Model.register("DeepseekV2ForCausalLM") +@Model.register("DeepseekV3ForCausalLM") class DeepseekV2Model(Model): model_arch = gguf.MODEL_ARCH.DEEPSEEK2 @@ -3395,6 +4009,15 @@ class DeepseekV2Model(Model): self.gguf_writer.add_expert_count(hparams["n_routed_experts"]) self.gguf_writer.add_expert_shared_count(hparams["n_shared_experts"]) self.gguf_writer.add_expert_weights_scale(hparams["routed_scaling_factor"]) + self.gguf_writer.add_expert_weights_norm(hparams["norm_topk_prob"]) + + if hparams["scoring_func"] == "sigmoid": + self.gguf_writer.add_expert_gating_func(gguf.ExpertGatingFuncType.SIGMOID) + elif hparams["scoring_func"] == "softmax": + self.gguf_writer.add_expert_gating_func(gguf.ExpertGatingFuncType.SOFTMAX) + else: + raise ValueError(f"Unsupported scoring_func value: {hparams['scoring_func']}") + self.gguf_writer.add_rope_dimension_count(hparams["qk_rope_head_dim"]) if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]: @@ -3407,6 +4030,16 @@ class DeepseekV2Model(Model): _experts: list[dict[str, Tensor]] | None = None def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + # rename e_score_correction_bias tensors + if name.endswith("e_score_correction_bias"): + name = name.replace("e_score_correction_bias", "e_score_correction.bias") + + # skip Multi-Token Prediction (MTP) layers + block_count = self.hparams["num_hidden_layers"] + match = re.match(r"model.layers.(\d+)", name) + if match and int(match.group(1)) >= block_count: + return [] + # process the experts separately if name.find("mlp.experts") != -1: n_experts = self.hparams["n_routed_experts"] @@ -4301,6 +4934,7 @@ def parse_args() -> argparse.Namespace: parser.add_argument( "model", type=Path, help="directory containing model file", + nargs="?", ) parser.add_argument( "--use-temp-file", action="store_true", @@ -4338,8 +4972,15 @@ def parse_args() -> argparse.Namespace: "--metadata", type=Path, help="Specify the path for an authorship metadata override file" ) + parser.add_argument( + "--print-supported-models", action="store_true", + help="Print the supported models" + ) - return parser.parse_args() + args = parser.parse_args() + if not args.print_supported_models and args.model is None: + parser.error("the following arguments are required: model") + return args def split_str_to_n_bytes(split_str: str) -> int: @@ -4363,6 +5004,11 @@ def split_str_to_n_bytes(split_str: str) -> int: def main() -> None: args = parse_args() + if args.print_supported_models: + logger.error("Supported models:") + Model.print_registered_models() + sys.exit(0) + if args.verbose: logging.basicConfig(level=logging.DEBUG) else: diff --git a/convert_hf_to_gguf_update.py b/convert_hf_to_gguf_update.py index 28cd02e5a..56edc64a7 100755 --- a/convert_hf_to_gguf_update.py +++ b/convert_hf_to_gguf_update.py @@ -17,7 +17,7 @@ # # python3 convert_hf_to_gguf_update.py # -# - Copy-paste the generated get_vocab_base_pre() function into convert_hf_to_gguf.py +# - The convert_hf_to_gguf.py script will have had its get_vocab_base_pre() function updated # - Update llama.cpp with the new pre-tokenizer if necessary # # TODO: generate tokenizer tests for llama.cpp @@ -72,6 +72,7 @@ models = [ {"name": "deepseek-coder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base", }, {"name": "falcon", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/tiiuae/falcon-7b", }, {"name": "bert-bge", "tokt": TOKENIZER_TYPE.WPM, "repo": "https://huggingface.co/BAAI/bge-small-en-v1.5", }, + {"name": "falcon3", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/tiiuae/Falcon3-7B-Base", }, {"name": "bert-bge-large", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/BAAI/bge-large-zh-v1.5", }, {"name": "mpt", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/mosaicml/mpt-7b", }, {"name": "starcoder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/bigcode/starcoder2-3b", }, @@ -102,6 +103,11 @@ models = [ {"name": "exaone", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct", }, {"name": "phi-2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/microsoft/phi-2", }, {"name": "chameleon", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/facebook/chameleon-7b", }, + {"name": "minerva-7b", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/sapienzanlp/Minerva-7B-base-v1.0", }, + {"name": "roberta-bpe", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/sentence-transformers/stsb-roberta-base"}, + {"name": "gigachat", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct"}, + {"name": "megrez", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/Infinigence/Megrez-3B-Instruct"}, + {"name": "deepseek-v3", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/DeepSeek-V3"}, ] diff --git a/convert_lora_to_gguf.py b/convert_lora_to_gguf.py index ed1014cae..6dea14a23 100755 --- a/convert_lora_to_gguf.py +++ b/convert_lora_to_gguf.py @@ -226,6 +226,9 @@ def get_base_tensor_name(lora_tensor_name: str) -> str: base_name = lora_tensor_name.replace("base_model.model.", "") base_name = base_name.replace(".lora_A.weight", ".weight") base_name = base_name.replace(".lora_B.weight", ".weight") + # models produced by mergekit-extract-lora have token embeddings in the adapter + base_name = base_name.replace(".lora_embedding_A", ".weight") + base_name = base_name.replace(".lora_embedding_B", ".weight") return base_name @@ -260,6 +263,10 @@ def parse_args() -> argparse.Namespace: "--base", type=Path, help="directory containing Hugging Face model config files (config.json, tokenizer.json) for the base model that the adapter is based on - only config is needed, actual model weights are not required. If base model is unspecified, it will be loaded from Hugging Face hub based on the adapter config", ) + parser.add_argument( + "--base-model-id", type=str, + help="the model ID of the base model, if it is not available locally or in the adapter config. If specified, it will ignore --base and load the base model config from the Hugging Face hub (Example: 'meta-llama/Llama-3.2-1B-Instruct')", + ) parser.add_argument( "lora_path", type=Path, help="directory containing Hugging Face PEFT LoRA config (adapter_model.json) and weights (adapter_model.safetensors or adapter_model.bin)", @@ -290,6 +297,7 @@ if __name__ == '__main__': dir_base_model: Path | None = args.base dir_lora: Path = args.lora_path + base_model_id: str | None = args.base_model_id lora_config = dir_lora / "adapter_config.json" input_model = dir_lora / "adapter_model.safetensors" @@ -313,7 +321,10 @@ if __name__ == '__main__': lparams: dict[str, Any] = json.load(f) # load base model - if dir_base_model is None: + if base_model_id is not None: + logger.info(f"Loading base model from Hugging Face: {base_model_id}") + hparams = load_hparams_from_hf(base_model_id) + elif dir_base_model is None: if "base_model_name_or_path" in lparams: model_id = lparams["base_model_name_or_path"] logger.info(f"Loading base model from Hugging Face: {model_id}") @@ -371,11 +382,16 @@ if __name__ == '__main__': if self.lazy: tensor = LazyTorchTensor.from_eager(tensor) base_name = get_base_tensor_name(name) - is_lora_a = ".lora_A.weight" in name - is_lora_b = ".lora_B.weight" in name + # note: mergekit-extract-lora also adds token embeddings to the adapter + is_lora_a = ".lora_A.weight" in name or ".lora_embedding_A" in name + is_lora_b = ".lora_B.weight" in name or ".lora_embedding_B" in name if not is_lora_a and not is_lora_b: if ".base_layer.weight" in name: continue + # mergekit-extract-lora add these layernorm to the adapter, we need to keep them + if "_layernorm" in name or ".norm" in name: + yield (base_name, tensor) + continue logger.error(f"Unexpected name '{name}': Not a lora_A or lora_B tensor") if ".embed_tokens.weight" in name or ".lm_head.weight" in name: logger.error("Embeddings is present in the adapter. This can be due to new tokens added during fine tuning") @@ -407,9 +423,21 @@ if __name__ == '__main__': if name == "lm_head.weight" and len(dest) == 0: raise ValueError("lm_head is present in adapter, but is ignored in base model") for dest_name, dest_data in dest: + # mergekit-extract-lora add these layernorm to the adapter + if "_norm" in dest_name: + assert dest_data.dim() == 1 + yield (dest_name, dest_data) + continue + + # otherwise, we must get the lora_A and lora_B tensors assert isinstance(dest_data, LoraTorchTensor) lora_a, lora_b = dest_data.get_lora_A_B() + # note: mergekit-extract-lora flip and transpose A and B + # here we only need to transpose token_embd.lora_a, see llm_build_inp_embd() + if "token_embd.weight" in dest_name: + lora_a = lora_a.T + yield (dest_name + ".lora_a", lora_a) yield (dest_name + ".lora_b", lora_b) diff --git a/docs/android.md b/docs/android.md index 320b62240..47530c6c1 100644 --- a/docs/android.md +++ b/docs/android.md @@ -23,10 +23,10 @@ $ curl -L {model-url} -o ~/{model}.gguf Then, if you are not already in the repo directory, `cd` into `llama.cpp` and: ``` -$ ./build/bin/llama-simple -m ~/{model}.gguf -c {context-size} -p "{your-prompt}" +$ ./build/bin/llama-cli -m ~/{model}.gguf -c {context-size} -p "{your-prompt}" ``` -Here, we show `llama-simple`, but any of the executables under `examples` should work, in theory. Be sure to set `context-size` to a reasonable number (say, 4096) to start with; otherwise, memory could spike and kill your terminal. +Here, we show `llama-cli`, but any of the executables under `examples` should work, in theory. Be sure to set `context-size` to a reasonable number (say, 4096) to start with; otherwise, memory could spike and kill your terminal. To see what it might look like visually, here's an old demo of an interactive session running on a Pixel 5 phone: diff --git a/docs/backend/BLIS.md b/docs/backend/BLIS.md index 35d06bd0f..904548577 100644 --- a/docs/backend/BLIS.md +++ b/docs/backend/BLIS.md @@ -27,13 +27,6 @@ We recommend using openmp since it's easier to modify the cores being used. ### llama.cpp compilation -Makefile: - -```bash -make GGML_BLIS=1 -j -# make GGML_BLIS=1 llama-benchmark-matmult -``` - CMake: ```bash diff --git a/docs/backend/CANN.md b/docs/backend/CANN.md index 6bdd9d2da..23f10175a 100644 --- a/docs/backend/CANN.md +++ b/docs/backend/CANN.md @@ -23,6 +23,8 @@ The llama.cpp CANN backend is designed to support Ascend NPU. It utilize the abi ## News +- 2024.11 + - Support F16 and F32 data type model for Ascend 310P NPU. - 2024.8 - Support `Q4_0` and `Q8_0` data type for Ascend NPU. - 2024.7 @@ -40,9 +42,11 @@ The llama.cpp CANN backend is designed to support Ascend NPU. It utilize the abi ### Ascend NPU **Verified devices** + | Ascend NPU | Status | |:-----------------------------:|:-------:| | Atlas 300T A2 | Support | +| Atlas 300I Duo | Support | *Notes:* diff --git a/docs/backend/SYCL.md b/docs/backend/SYCL.md index bc8c0f886..8d8312e91 100644 --- a/docs/backend/SYCL.md +++ b/docs/backend/SYCL.md @@ -34,13 +34,16 @@ The SYCL backend would be broken by some PRs due to no online CI. The following release is verified with good quality: -|Commit ID|Tag|Release|Verified Platform| -|-|-|-|-| -|fb76ec31a9914b7761c1727303ab30380fd4f05c|b3038 |[llama-b3038-bin-win-sycl-x64.zip](https://github.com/ggerganov/llama.cpp/releases/download/b3038/llama-b3038-bin-win-sycl-x64.zip) |Arc770/Linux/oneAPI 2024.1
MTL Arc GPU/Windows 11/oneAPI 2024.1| +|Commit ID|Tag|Release|Verified Platform| Update date| +|-|-|-|-|-| +|3bcd40b3c593d14261fb2abfabad3c0fb5b9e318|b4040 |[llama-b4040-bin-win-sycl-x64.zip](https://github.com/ggerganov/llama.cpp/releases/download/b4040/llama-b4040-bin-win-sycl-x64.zip) |Arc770/Linux/oneAPI 2024.1
MTL Arc GPU/Windows 11/oneAPI 2024.1| 2024-11-19| +|fb76ec31a9914b7761c1727303ab30380fd4f05c|b3038 |[llama-b3038-bin-win-sycl-x64.zip](https://github.com/ggerganov/llama.cpp/releases/download/b3038/llama-b3038-bin-win-sycl-x64.zip) |Arc770/Linux/oneAPI 2024.1
MTL Arc GPU/Windows 11/oneAPI 2024.1|| ## News +- 2024.11 + - Use syclcompat to improve the performance on some platforms. This requires to use oneAPI 2025.0 or newer. - 2024.8 - Use oneDNN as the default GEMM library, improve the compatibility for new Intel GPUs. @@ -310,12 +313,14 @@ export CPLUS_INCLUDE_DIR=/path/to/oneMKL/buildWithCublas/include:$CPLUS_INCLUDE_ export CPLUS_INCLUDE_DIR=/path/to/oneMKL/include:$CPLUS_INCLUDE_DIR # Build LLAMA with Nvidia BLAS acceleration through SYCL +# Setting GGML_SYCL_DEVICE_ARCH is optional but can improve performance +GGML_SYCL_DEVICE_ARCH=sm_80 # Example architecture # Option 1: Use FP32 (recommended for better performance in most cases) -cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=NVIDIA -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx +cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=NVIDIA -DGGML_SYCL_DEVICE_ARCH=${GGML_SYCL_DEVICE_ARCH} -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx # Option 2: Use FP16 -cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=NVIDIA -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON +cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=NVIDIA -DGGML_SYCL_DEVICE_ARCH=${GGML_SYCL_DEVICE_ARCH} -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_SYCL_F16=ON # build all binary cmake --build build --config Release -j -v @@ -333,8 +338,9 @@ export CPLUS_INCLUDE_DIR=/path/to/oneMKL/buildWithrocBLAS/include:$CPLUS_INCLUDE ## AMD # Use FP32, FP16 is not supported -# Find your GGML_SYCL_HIP_TARGET with rocminfo, under the key 'Name:' -cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=AMD -DGGML_SYCL_HIP_TARGET=${GGML_SYCL_HIP_TARGET} -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx +# Find your GGML_SYCL_DEVICE_ARCH with rocminfo, under the key 'Name:' +GGML_SYCL_DEVICE_ARCH=gfx90a # Example architecture +cmake -B build -DGGML_SYCL=ON -DGGML_SYCL_TARGET=AMD -DGGML_SYCL_DEVICE_ARCH=${GGML_SYCL_DEVICE_ARCH} -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx # build all binary cmake --build build --config Release -j -v @@ -644,6 +650,7 @@ use 1 SYCL GPUs: [0] with Max compute units:512 |--------------------|---------------------------------------|---------------------------------------------| | GGML_SYCL | ON (mandatory) | Enable build with SYCL code path.
FP32 path - recommended for better perforemance than FP16 on quantized model| | GGML_SYCL_TARGET | INTEL *(default)* \| NVIDIA \| AMD | Set the SYCL target device type. | +| GGML_SYCL_DEVICE_ARCH | Optional (except for AMD) | Set the SYCL device architecture, optional except for AMD. Setting the device architecture can improve the performance. See the table [--offload-arch](https://github.com/intel/llvm/blob/sycl/sycl/doc/design/OffloadDesign.md#--offload-arch) for a list of valid architectures. | | GGML_SYCL_F16 | OFF *(default)* \|ON *(optional)* | Enable FP16 build with SYCL code path. | | CMAKE_C_COMPILER | `icx` *(Linux)*, `icx/cl` *(Windows)* | Set `icx` compiler for SYCL code path. | | CMAKE_CXX_COMPILER | `icpx` *(Linux)*, `icx` *(Windows)* | Set `icpx/icx` compiler for SYCL code path. | diff --git a/docs/build.md b/docs/build.md index 4e362ebc7..3b0d2211d 100644 --- a/docs/build.md +++ b/docs/build.md @@ -7,124 +7,75 @@ git clone https://github.com/ggerganov/llama.cpp cd llama.cpp ``` -In order to build llama.cpp you have four different options. +The following sections describe how to build with different backends and options. -- Using `make`: - - On Linux or MacOS: +## CPU Build - ```bash - make - ``` +Build llama.cpp using `CMake`: - - On Windows (x86/x64 only, arm64 requires cmake): +```bash +cmake -B build +cmake --build build --config Release +``` - 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). - 2. Extract `w64devkit` on your pc. - 3. Run `w64devkit.exe`. - 4. Use the `cd` command to reach the `llama.cpp` folder. - 5. From here you can run: - ```bash - make - ``` +**Notes**: - - Notes: - - For `Q4_0_4_4` quantization type build, add the `GGML_NO_LLAMAFILE=1` flag. For example, use `make GGML_NO_LLAMAFILE=1`. - - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `make -j 8` will run 8 jobs in parallel. - - For faster repeated compilation, install [ccache](https://ccache.dev/). - - For debug builds, run `make LLAMA_DEBUG=1` +- For faster compilation, add the `-j` argument to run multiple jobs in parallel, or use a generator that does this automatically such as Ninja. For example, `cmake --build build --config Release -j 8` will run 8 jobs in parallel. +- For faster repeated compilation, install [ccache](https://ccache.dev/) +- For debug builds, there are two cases: -- Using `CMake`: + 1. Single-config generators (e.g. default = `Unix Makefiles`; note that they just ignore the `--config` flag): - ```bash - cmake -B build + ```bash + cmake -B build -DCMAKE_BUILD_TYPE=Debug + cmake --build build + ``` + + 2. Multi-config generators (`-G` param set to Visual Studio, XCode...): + + ```bash + cmake -B build -G "Xcode" + cmake --build build --config Debug + ``` + + For more details and a list of supported generators, see the [CMake documentation](https://cmake.org/cmake/help/latest/manual/cmake-generators.7.html). +- For static builds, add `-DBUILD_SHARED_LIBS=OFF`: + ``` + cmake -B build -DBUILD_SHARED_LIBS=OFF cmake --build build --config Release ``` - **Notes**: - - - For `Q4_0_4_4` quantization type build, add the `-DGGML_LLAMAFILE=OFF` cmake option. For example, use `cmake -B build -DGGML_LLAMAFILE=OFF`. - - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `cmake --build build --config Release -j 8` will run 8 jobs in parallel. - - For faster repeated compilation, install [ccache](https://ccache.dev/). - - For debug builds, there are two cases: - - 1. Single-config generators (e.g. default = `Unix Makefiles`; note that they just ignore the `--config` flag): +- Building for Windows (x86, x64 and arm64) with MSVC or clang as compilers: + - Install Visual Studio 2022, e.g. via the [Community Edition](https://visualstudio.microsoft.com/de/vs/community/). In the installer, select at least the following options (this also automatically installs the required additional tools like CMake,...): + - Tab Workload: Desktop-development with C++ + - Tab Components (select quickly via search): C++-_CMake_ Tools for Windows, _Git_ for Windows, C++-_Clang_ Compiler for Windows, MS-Build Support for LLVM-Toolset (clang) + - Please remember to always use a Developer Command Prompt / PowerShell for VS2022 for git, build, test + - For Windows on ARM (arm64, WoA) build with: + ```bash + cmake --preset arm64-windows-llvm-release -D GGML_OPENMP=OFF + cmake --build build-arm64-windows-llvm-release + ``` + Building for arm64 can also be done with the MSVC compiler with the build-arm64-windows-MSVC preset, or the standard CMake build instructions. However, note that the MSVC compiler does not support inline ARM assembly code, used e.g. for the accelerated Q4_0_N_M CPU kernels. + For building with ninja generator and clang compiler as default: + -set path:set LIB=C:\Program Files (x86)\Windows Kits\10\Lib\10.0.22621.0\um\x64;C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.41.34120\lib\x64\uwp;C:\Program Files (x86)\Windows Kits\10\Lib\10.0.22621.0\ucrt\x64 ```bash - cmake -B build -DCMAKE_BUILD_TYPE=Debug - cmake --build build + cmake --preset x64-windows-llvm-release + cmake --build build-x64-windows-llvm-release ``` - 2. Multi-config generators (`-G` param set to Visual Studio, XCode...): - - ```bash - cmake -B build -G "Xcode" - cmake --build build --config Debug - ``` - - Building for Windows (x86, x64 and arm64) with MSVC or clang as compilers: - - Install Visual Studio 2022, e.g. via the [Community Edition](https://visualstudio.microsoft.com/de/vs/community/). In the installer, select at least the following options (this also automatically installs the required additional tools like CMake,...): - - Tab Workload: Desktop-development with C++ - - Tab Components (select quickly via search): C++-_CMake_ Tools for Windows, _Git_ for Windows, C++-_Clang_ Compiler for Windows, MS-Build Support for LLVM-Toolset (clang) - - Please remember to always use a Developer Command Prompt / PowerShell for VS2022 for git, build, test - - For Windows on ARM (arm64, WoA) build with: - ```bash - cmake --preset arm64-windows-llvm-release -D GGML_OPENMP=OFF - cmake --build build-arm64-windows-llvm-release - ``` - Note: Building for arm64 could also be done just with MSVC (with the build-arm64-windows-MSVC preset, or the standard CMake build instructions). But MSVC does not support inline ARM assembly-code, used e.g. for the accelerated Q4_0_4_8 CPU kernels. - -- Using `gmake` (FreeBSD): - - 1. Install and activate [DRM in FreeBSD](https://wiki.freebsd.org/Graphics) - 2. Add your user to **video** group - 3. Install compilation dependencies. - - ```bash - sudo pkg install gmake automake autoconf pkgconf llvm15 openblas - - gmake CC=/usr/local/bin/clang15 CXX=/usr/local/bin/clang++15 -j4 - ``` - -## Metal Build - -On MacOS, Metal is enabled by default. Using Metal makes the computation run on the GPU. -To disable the Metal build at compile time use the `GGML_NO_METAL=1` flag or the `GGML_METAL=OFF` cmake option. - -When built with Metal support, you can explicitly disable GPU inference with the `--n-gpu-layers|-ngl 0` command-line -argument. - ## BLAS Build -Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Support with CPU-only BLAS implementations doesn't affect the normal generation performance. We may see generation performance improvements with GPU-involved BLAS implementations, e.g. cuBLAS, hipBLAS. There are currently several different BLAS implementations available for build and use: +Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Using BLAS doesn't affect the generation performance. There are currently several different BLAS implementations available for build and use: -### Accelerate Framework: +### Accelerate Framework This is only available on Mac PCs and it's enabled by default. You can just build using the normal instructions. -### OpenBLAS: +### OpenBLAS This provides BLAS acceleration using only the CPU. Make sure to have OpenBLAS installed on your machine. -- Using `make`: - - On Linux: - ```bash - make GGML_OPENBLAS=1 - ``` - - - On Windows: - - 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). - 2. Download the latest version of [OpenBLAS for Windows](https://github.com/xianyi/OpenBLAS/releases). - 3. Extract `w64devkit` on your pc. - 4. From the OpenBLAS zip that you just downloaded copy `libopenblas.a`, located inside the `lib` folder, inside `w64devkit\x86_64-w64-mingw32\lib`. - 5. From the same OpenBLAS zip copy the content of the `include` folder inside `w64devkit\x86_64-w64-mingw32\include`. - 6. Run `w64devkit.exe`. - 7. Use the `cd` command to reach the `llama.cpp` folder. - 8. From here you can run: - - ```bash - make GGML_OPENBLAS=1 - ``` - - Using `CMake` on Linux: ```bash @@ -136,14 +87,6 @@ This provides BLAS acceleration using only the CPU. Make sure to have OpenBLAS i Check [BLIS.md](./backend/BLIS.md) for more information. -### SYCL - -SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators. - -llama.cpp based on SYCL is used to **support Intel GPU** (Data Center Max series, Flex series, Arc series, Built-in GPU and iGPU). - -For detailed info, please refer to [llama.cpp for SYCL](./backend/SYCL.md). - ### Intel oneMKL Building through oneAPI compilers will make avx_vnni instruction set available for intel processors that do not support avx512 and avx512_vnni. Please note that this build config **does not support Intel GPU**. For Intel GPU support, please refer to [llama.cpp for SYCL](./backend/SYCL.md). @@ -161,16 +104,31 @@ Building through oneAPI compilers will make avx_vnni instruction set available f Check [Optimizing and Running LLaMA2 on Intel® CPU](https://www.intel.com/content/www/us/en/content-details/791610/optimizing-and-running-llama2-on-intel-cpu.html) for more information. -### CUDA +### Other BLAS libraries -This provides GPU acceleration using the CUDA cores of your Nvidia GPU. Make sure to have the CUDA toolkit installed. You can download it from your Linux distro's package manager (e.g. `apt install nvidia-cuda-toolkit`) or from here: [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads). +Any other BLAS library can be used by setting the `GGML_BLAS_VENDOR` option. See the [CMake documentation](https://cmake.org/cmake/help/latest/module/FindBLAS.html#blas-lapack-vendors) for a list of supported vendors. -For Jetson user, if you have Jetson Orin, you can try this: [Offical Support](https://www.jetson-ai-lab.com/tutorial_text-generation.html). If you are using an old model(nano/TX2), need some additional operations before compiling. +## Metal Build + +On MacOS, Metal is enabled by default. Using Metal makes the computation run on the GPU. +To disable the Metal build at compile time use the `-DGGML_METAL=OFF` cmake option. + +When built with Metal support, you can explicitly disable GPU inference with the `--n-gpu-layers 0` command-line argument. + +## SYCL + +SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators. + +llama.cpp based on SYCL is used to **support Intel GPU** (Data Center Max series, Flex series, Arc series, Built-in GPU and iGPU). + +For detailed info, please refer to [llama.cpp for SYCL](./backend/SYCL.md). + +## CUDA + +This provides GPU acceleration using an NVIDIA GPU. Make sure to have the CUDA toolkit installed. You can download it from your Linux distro's package manager (e.g. `apt install nvidia-cuda-toolkit`) or from the [NVIDIA developer site](https://developer.nvidia.com/cuda-downloads). + +If you are using Fedora (using Fedora Workstation, or an 'Atomic' variant such as Silverblue), or would like to set up CUDA in a toolbox, please consider our [Fedora CUDA guide](./cuda-fedora.md). Unfortunately, the process is not as simple as one might expect. -- Using `make`: - ```bash - make GGML_CUDA=1 - ``` - Using `CMake`: ```bash @@ -186,24 +144,16 @@ The following compilation options are also available to tweak performance: | Option | Legal values | Default | Description | |-------------------------------|------------------------|---------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| GGML_CUDA_FORCE_DMMV | Boolean | false | Force the use of dequantization + matrix vector multiplication kernels instead of using kernels that do matrix vector multiplication on quantized data. By default the decision is made based on compute capability (MMVQ for 6.1/Pascal/GTX 1000 or higher). Does not affect k-quants. | -| GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the CUDA dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | -| GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the CUDA mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. | | GGML_CUDA_FORCE_MMQ | Boolean | false | Force the use of custom matrix multiplication kernels for quantized models instead of FP16 cuBLAS even if there is no int8 tensor core implementation available (affects V100, RDNA3). MMQ kernels are enabled by default on GPUs with int8 tensor core support. With MMQ force enabled, speed for large batch sizes will be worse but VRAM consumption will be lower. | | GGML_CUDA_FORCE_CUBLAS | Boolean | false | Force the use of FP16 cuBLAS instead of custom matrix multiplication kernels for quantized models | | GGML_CUDA_F16 | Boolean | false | If enabled, use half-precision floating point arithmetic for the CUDA dequantization + mul mat vec kernels and for the q4_1 and q5_1 matrix matrix multiplication kernels. Can improve performance on relatively recent GPUs. | -| GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per CUDA thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | | GGML_CUDA_PEER_MAX_BATCH_SIZE | Positive integer | 128 | Maximum batch size for which to enable peer access between multiple GPUs. Peer access requires either Linux or NVLink. When using NVLink enabling peer access for larger batch sizes is potentially beneficial. | | GGML_CUDA_FA_ALL_QUANTS | Boolean | false | Compile support for all KV cache quantization type (combinations) for the FlashAttention CUDA kernels. More fine-grained control over KV cache size but compilation takes much longer. | -### MUSA +## MUSA This provides GPU acceleration using the MUSA cores of your Moore Threads MTT GPU. Make sure to have the MUSA SDK installed. You can download it from here: [MUSA SDK](https://developer.mthreads.com/sdk/download/musa). -- Using `make`: - ```bash - make GGML_MUSA=1 - ``` - Using `CMake`: ```bash @@ -217,20 +167,16 @@ The environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` can be used to enab Most of the compilation options available for CUDA should also be available for MUSA, though they haven't been thoroughly tested yet. -### hipBLAS +## HIP -This provides BLAS acceleration on HIP-supported AMD GPUs. +This provides GPU acceleration on HIP-supported AMD GPUs. Make sure to have ROCm installed. You can download it from your Linux distro's package manager or from here: [ROCm Quick Start (Linux)](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/tutorial/quick-start.html#rocm-install-quick). -- Using `make`: - ```bash - make GGML_HIPBLAS=1 - ``` - Using `CMake` for Linux (assuming a gfx1030-compatible AMD GPU): ```bash HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ - cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ + cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ && cmake --build build --config Release -- -j 16 ``` On Linux it is also possible to use unified memory architecture (UMA) to share main memory between the CPU and integrated GPU by setting `-DGGML_HIP_UMA=ON`. @@ -247,19 +193,14 @@ You can download it from your Linux distro's package manager or from here: [ROCm ```bash HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -p)" \ HIP_DEVICE_LIB_PATH= \ - cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ + cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ && cmake --build build -- -j 16 ``` -- Using `make` (example for target gfx1030, build with 16 CPU threads): - ```bash - make -j16 GGML_HIPBLAS=1 GGML_HIP_UMA=1 AMDGPU_TARGETS=gfx1030 - ``` - - Using `CMake` for Windows (using x64 Native Tools Command Prompt for VS, and assuming a gfx1100-compatible AMD GPU): ```bash set PATH=%HIP_PATH%\bin;%PATH% - cmake -S . -B build -G Ninja -DAMDGPU_TARGETS=gfx1100 -DGGML_HIPBLAS=ON -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release + cmake -S . -B build -G Ninja -DAMDGPU_TARGETS=gfx1100 -DGGML_HIP=ON -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release cmake --build build ``` Make sure that `AMDGPU_TARGETS` is set to the GPU arch you want to compile for. The above example uses `gfx1100` that corresponds to Radeon RX 7900XTX/XT/GRE. You can find a list of targets [here](https://llvm.org/docs/AMDGPUUsage.html#processors) @@ -268,23 +209,16 @@ You can download it from your Linux distro's package manager or from here: [ROCm The environment variable [`HIP_VISIBLE_DEVICES`](https://rocm.docs.amd.com/en/latest/understand/gpu_isolation.html#hip-visible-devices) can be used to specify which GPU(s) will be used. If your GPU is not officially supported you can use the environment variable [`HSA_OVERRIDE_GFX_VERSION`] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3. -The following compilation options are also available to tweak performance (yes, they refer to CUDA, not HIP, because it uses the same code as the cuBLAS version above): -| Option | Legal values | Default | Description | -|------------------------|------------------------|---------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| -| GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the HIP dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | -| GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the HIP mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. Does not affect k-quants. | -| GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per HIP thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | - -### Vulkan +## Vulkan **Windows** -#### w64devkit +### w64devkit -Download and extract [w64devkit](https://github.com/skeeto/w64devkit/releases). +Download and extract [`w64devkit`](https://github.com/skeeto/w64devkit/releases). -Download and install the [Vulkan SDK](https://vulkan.lunarg.com/sdk/home#windows). When selecting components, only the Vulkan SDK Core is required. +Download and install the [`Vulkan SDK`](https://vulkan.lunarg.com/sdk/home#windows) with the default settings. Launch `w64devkit.exe` and run the following commands to copy Vulkan dependencies: ```sh @@ -300,18 +234,47 @@ Libs: -lvulkan-1 EOF ``` -Switch into the `llama.cpp` directory and run `make GGML_VULKAN=1`. -#### MSYS2 +Switch into the `llama.cpp` directory and build using CMake. +```sh +cmake -B build -DGGML_VULKAN=ON +cmake --build build --config Release +``` + +### Git Bash MINGW64 + +Download and install [`Git-SCM`](https://git-scm.com/downloads/win) with the default settings + +Download and install [`Visual Studio Community Edition`](https://visualstudio.microsoft.com/) and make sure you select `C++` + +Download and install [`CMake`](https://cmake.org/download/) with the default settings + +Download and install the [`Vulkan SDK`](https://vulkan.lunarg.com/sdk/home#windows) with the default settings. + +Go into your `llama.cpp` directory and right click, select `Open Git Bash Here` and then run the following commands + +``` +cmake -B build -DGGML_VULKAN=ON +cmake --build build --config Release +``` + +Now you can load the model in conversation mode using `Vulkan` + +```sh +build/bin/Release/llama-cli -m "[PATH TO MODEL]" -ngl 100 -c 16384 -t 10 -n -2 -cnv +``` + +### MSYS2 Install [MSYS2](https://www.msys2.org/) and then run the following commands in a UCRT terminal to install dependencies. - ```sh - pacman -S git \ - mingw-w64-ucrt-x86_64-gcc \ - mingw-w64-ucrt-x86_64-cmake \ - mingw-w64-ucrt-x86_64-vulkan-devel \ - mingw-w64-ucrt-x86_64-shaderc - ``` -Switch into `llama.cpp` directory and build using CMake. +```sh +pacman -S git \ + mingw-w64-ucrt-x86_64-gcc \ + mingw-w64-ucrt-x86_64-cmake \ + mingw-w64-ucrt-x86_64-vulkan-devel \ + mingw-w64-ucrt-x86_64-shaderc +``` + +Switch into the `llama.cpp` directory and build using CMake. ```sh cmake -B build -DGGML_VULKAN=ON cmake --build build --config Release @@ -360,7 +323,7 @@ cmake --build build --config Release # ggml_vulkan: Using Intel(R) Graphics (ADL GT2) | uma: 1 | fp16: 1 | warp size: 32 ``` -### CANN +## CANN This provides NPU acceleration using the AI cores of your Ascend NPU. And [CANN](https://www.hiascend.com/en/software/cann) is a hierarchical APIs to help you to quickly build AI applications and service based on Ascend NPU. For more information about Ascend NPU in [Ascend Community](https://www.hiascend.com/en/). @@ -375,22 +338,26 @@ cmake --build build --config release You can test with: -`./build/llama-cli -m PATH_TO_MODEL -p "Building a website can be done in 10 steps:" -ngl 32` - -If the fllowing info is output on screen, you are using `llama.cpp by CANN backend`: ```bash -llm_load_tensors: CANN buffer size = 13313.00 MiB +./build/bin/llama-cli -m PATH_TO_MODEL -p "Building a website can be done in 10 steps:" -ngl 32 +``` + +If the following info is output on screen, you are using `llama.cpp` with the CANN backend: +```bash +llm_load_tensors: CANN model buffer size = 13313.00 MiB llama_new_context_with_model: CANN compute buffer size = 1260.81 MiB ``` For detailed info, such as model/device supports, CANN install, please refer to [llama.cpp for CANN](./backend/CANN.md). -### Android +## Android To read documentation for how to build on Android, [click here](./android.md) -### Arm CPU optimized mulmat kernels +## Notes about GPU-accelerated backends -Llama.cpp includes a set of optimized mulmat kernels for the Arm architecture, leveraging Arm® Neon™, int8mm and SVE instructions. These kernels are enabled at build time through the appropriate compiler cpu-type flags, such as `-DCMAKE_C_FLAGS=-march=armv8.2a+i8mm+sve`. Note that these optimized kernels require the model to be quantized into one of the formats: `Q4_0_4_4` (Arm Neon), `Q4_0_4_8` (int8mm) or `Q4_0_8_8` (SVE). The SVE mulmat kernel specifically requires a vector width of 256 bits. When running on devices with a different vector width, it is recommended to use the `Q4_0_4_8` (int8mm) or `Q4_0_4_4` (Arm Neon) formats for better performance. Refer to [examples/quantize/README.md](../examples/quantize/README.md) for more information on the quantization formats. +The GPU may still be used to accelerate some parts of the computation even when using the `-ngl 0` option. You can fully disable GPU acceleration by using `--device none`. -To support `Q4_0_4_4`, you must build with `GGML_NO_LLAMAFILE=1` (`make`) or `-DGGML_LLAMAFILE=OFF` (`cmake`). +In most cases, it is possible to build and use multiple backends at the same time. For example, you can build llama.cpp with both CUDA and Vulkan support by using the `-DGGML_CUDA=ON -DGGML_VULKAN=ON` options with CMake. At runtime, you can specify which backend devices to use with the `--device` option. To see a list of available devices, use the `--list-devices` option. + +Backends can be built as dynamic libraries that can be loaded dynamically at runtime. This allows you to use the same llama.cpp binary on different machines with different GPUs. To enable this feature, use the `GGML_BACKEND_DL` option when building. diff --git a/docs/cuda-fedora.md b/docs/cuda-fedora.md new file mode 100644 index 000000000..b993386c8 --- /dev/null +++ b/docs/cuda-fedora.md @@ -0,0 +1,317 @@ +# Setting Up CUDA on Fedora + +In this guide we setup [Nvidia CUDA](https://docs.nvidia.com/cuda/) in a toolbox container. This guide is applicable for: +- [Fedora Workstation](https://fedoraproject.org/workstation/) +- [Atomic Desktops for Fedora](https://fedoraproject.org/atomic-desktops/) +- [Fedora Spins](https://fedoraproject.org/spins) +- [Other Distributions](https://containertoolbx.org/distros/), including `Red Hat Enterprise Linux >= 8.`, `Arch Linux`, and `Ubuntu`. + + +## Table of Contents + +- [Prerequisites](#prerequisites) +- [Monitoring NVIDIA CUDA Repositories](#monitoring-nvidia-cuda-repositories) +- [Using the Fedora 39 CUDA Repository](#using-the-fedora-39-cuda-repository) +- [Creating a Fedora Toolbox Environment](#creating-a-fedora-toolbox-environment) +- [Installing Essential Development Tools](#installing-essential-development-tools) +- [Adding the CUDA Repository](#adding-the-cuda-repository) +- [Installing `nvidia-driver-libs`](#installing-nvidia-driver-libs) +- [Manually Resolving Package Conflicts](#manually-resolving-package-conflicts) +- [Finalizing the Installation of `nvidia-driver-libs`](#finalizing-the-installation-of-nvidia-driver-libs) +- [Installing the CUDA Meta-Package](#installing-the-cuda-meta-package) +- [Configuring the Environment](#configuring-the-environment) +- [Verifying the Installation](#verifying-the-installation) +- [Conclusion](#conclusion) +- [Troubleshooting](#troubleshooting) +- [Additional Notes](#additional-notes) +- [References](#references) + +## Prerequisites + +- **Toolbox Installed on the Host System** `Fedora Silverblue` and `Fedora Workstation` both have toolbox by default, other distributions may need to install the [toolbox package](https://containertoolbx.org/install/). +- **NVIDIA Drivers and Graphics Card installed on Host System (optional)** To run CUDA program, such as `llama.cpp`, the host should be setup to access your NVIDIA hardware. Fedora Hosts can use the [RPM Fusion Repository](https://rpmfusion.org/Howto/NVIDIA). +- **Internet connectivity** to download packages. + +### Monitoring NVIDIA CUDA Repositories + +Before proceeding, it is advisable to check if NVIDIA has updated their CUDA repositories for your Fedora version. NVIDIA's repositories can be found at: + +- [Fedora 40 CUDA Repository](https://developer.download.nvidia.com/compute/cuda/repos/fedora40/x86_64/) +- [Fedora 41 CUDA Repository](https://developer.download.nvidia.com/compute/cuda/repos/fedora41/x86_64/) + +As of the latest update, these repositories do not contain the `cuda` meta-package or are missing essential components. + +### Using the Fedora 39 CUDA Repository + +Since the newer repositories are incomplete, we'll use the Fedora 39 repository: + +- [Fedora 39 CUDA Repository](https://developer.download.nvidia.com/compute/cuda/repos/fedora39/x86_64/) + +**Note:** Fedora 39 is no longer maintained, so we recommend using a toolbox environment to prevent system conflicts. + +## Creating a Fedora Toolbox Environment + +This guide focuses on Fedora hosts, but with small adjustments, it can work for other hosts. Using a Fedora 39 toolbox allows us to install the necessary packages without affecting the host system. + +**Note:** Toolbox is available for other systems, and even without Toolbox, it is possible to use Podman or Docker. + +We do not recommend installing on the host system, as Fedora 39 is out-of-maintenance, and instead you should upgrade to a maintained version of Fedora for your host. + +1. **Create a Fedora 39 Toolbox:** + + ```bash + toolbox create --image registry.fedoraproject.org/fedora-toolbox:39 --container fedora-toolbox-39-cuda + ``` + +2. **Enter the Toolbox:** + + ```bash + toolbox enter --container fedora-toolbox-39-cuda + ``` + + Inside the toolbox, you have root privileges and can install packages without affecting the host system. + +## Installing Essential Development Tools + +1. **Synchronize the DNF Package Manager:** + + ```bash + sudo dnf distro-sync + ``` + +2. **Install the Default Text Editor (Optional):** + + ```bash + sudo dnf install vim-default-editor --allowerasing + ``` + + The `--allowerasing` flag resolves any package conflicts. + +3. **Install Development Tools and Libraries:** + + ```bash + sudo dnf install @c-development @development-tools cmake + ``` + + This installs essential packages for compiling software, including `gcc`, `make`, and other development headers. + +## Adding the CUDA Repository + +Add the NVIDIA CUDA repository to your DNF configuration: + +```bash +sudo dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/fedora39/x86_64/cuda-fedora39.repo +``` + +After adding the repository, synchronize the package manager again: + +```bash +sudo dnf distro-sync +``` + +## Installing `nvidia-driver-libs` + +Attempt to install `nvidia-driver-libs`: + +```bash +sudo dnf install nvidia-driver-libs +``` + +**Explanation:** + +- `nvidia-driver-libs` contains necessary NVIDIA driver libraries required by CUDA. +- This step might fail due to conflicts with existing NVIDIA drivers on the host system. + +## Manually Resolving Package Conflicts + +If the installation fails due to conflicts, we'll manually download and install the required packages, excluding conflicting files. + +### 1. Download the `nvidia-driver-libs` RPM + +```bash +sudo dnf download --arch x86_64 nvidia-driver-libs +``` + +You should see a file similar to: + +``` +nvidia-driver-libs-560.35.05-1.fc39.x86_64.rpm +``` + +### 2. Attempt to Install the RPM + +```bash +sudo dnf install nvidia-driver-libs-560.35.05-1.fc39.x86_64.rpm +``` + +**Expected Error:** + +Installation may fail with errors pointing to conflicts with `egl-gbm` and `egl-wayland`. + +**Note: It is important to carefully read the error messages to identify the exact paths that need to be excluded.** + +### 3. Download Dependencies + +```bash +sudo dnf download --arch x86_64 egl-gbm egl-wayland +``` + +### 4. Install `egl-gbm` with Excluded Paths + +Exclude conflicting files during installation: + +```bash +sudo rpm --install --verbose --hash \ + --excludepath=/usr/lib64/libnvidia-egl-gbm.so.1.1.2 \ + --excludepath=/usr/share/egl/egl_external_platform.d/15_nvidia_gbm.json \ + egl-gbm-1.1.2^20240919gitb24587d-3.fc39.x86_64.rpm +``` + +**Explanation:** + +- The `--excludepath` option skips installing files that conflict with existing files. +- Adjust the paths based on the error messages you receive. + +### 5. Install `egl-wayland` with Excluded Paths + +```bash +sudo rpm --install --verbose --hash \ + --excludepath=/usr/share/egl/egl_external_platform.d/10_nvidia_wayland.json \ + egl-wayland-1.1.17^20241118giteeb29e1-5.fc39.x86_64.rpm +``` + +### 6. Install `nvidia-driver-libs` with Excluded Paths + +```bash +sudo rpm --install --verbose --hash \ + --excludepath=/usr/share/glvnd/egl_vendor.d/10_nvidia.json \ + --excludepath=/usr/share/nvidia/nvoptix.bin \ + nvidia-driver-libs-560.35.05-1.fc39.x86_64.rpm +``` + +**Note:** + +- Replace the paths with the ones causing conflicts in your installation if they differ. +- The `--verbose` and `--hash` options provide detailed output during installation. + +## Finalizing the Installation of `nvidia-driver-libs` + +After manually installing the dependencies, run: + +```bash +sudo dnf install nvidia-driver-libs +``` + +You should receive a message indicating the package is already installed: + +``` +Package nvidia-driver-libs-3:560.35.05-1.fc39.x86_64 is already installed. +Dependencies resolved. +Nothing to do. +Complete! +``` + +## Installing the CUDA Meta-Package + +Now that the driver libraries are installed, proceed to install CUDA: + +```bash +sudo dnf install cuda +``` + +This installs the CUDA toolkit and associated packages. + +## Configuring the Environment + +To use CUDA, add its binary directory to your system's `PATH`. + +1. **Create a Profile Script:** + + ```bash + sudo sh -c 'echo "export PATH=\$PATH:/usr/local/cuda/bin" >> /etc/profile.d/cuda.sh' + ``` + + **Explanation:** + + - We add to `/etc/profile.d/` as the `/etc/` folder is unique to this particular container, and is not shared with other containers or the host system. + - The backslash `\` before `$PATH` ensures the variable is correctly written into the script. + +2. **Make the Script Executable:** + + ```bash + sudo chmod +x /etc/profile.d/cuda.sh + ``` + +3. **Source the Script to Update Your Environment:** + + ```bash + source /etc/profile.d/cuda.sh + ``` + + **Note:** This command updates your current shell session with the new `PATH`. The `/etc/profile.d/cuda.sh` script ensures that the CUDA binaries are available in your `PATH` for all future sessions. + +## Verifying the Installation + +To confirm that CUDA is correctly installed and configured, check the version of the NVIDIA CUDA Compiler (`nvcc`): + +```bash +nvcc --version +``` + +You should see output similar to: + +``` +nvcc: NVIDIA (R) Cuda compiler driver +Copyright (c) 2005-2024 NVIDIA Corporation +Built on Tue_Oct_29_23:50:19_PDT_2024 +Cuda compilation tools, release 12.6, V12.6.85 +Build cuda_12.6.r12.6/compiler.35059454_0 +``` + +This output confirms that the CUDA compiler is accessible and indicates the installed version. + +## Conclusion + +You have successfully set up CUDA on Fedora within a toolbox environment using the Fedora 39 CUDA repository. By manually resolving package conflicts and configuring the environment, you can develop CUDA applications without affecting your host system. + +## Troubleshooting + +- **Installation Failures:** + - If you encounter errors during installation, carefully read the error messages. They often indicate conflicting files or missing dependencies. + - Use the `--excludepath` option with `rpm` to exclude conflicting files during manual installations. + +- **Driver Conflicts:** + - Since the host system may already have NVIDIA drivers installed, conflicts can arise. Using the toolbox environment helps isolate these issues. + +- **Environment Variables Not Set:** + - If `nvcc` is not found after installation, ensure that `/usr/local/cuda/bin` is in your `PATH`. + - Run `echo $PATH` to check if the path is included. + - Re-source the profile script or open a new terminal session. + +## Additional Notes + +- **Updating CUDA in the Future:** + - Keep an eye on the official NVIDIA repositories for updates to your Fedora version. + - When an updated repository becomes available, adjust your `dnf` configuration accordingly. + +- **Building `llama.cpp`:** + - With CUDA installed, you can follow these [build instructions for `llama.cpp`](https://github.com/ggerganov/llama.cpp/blob/master/docs/build.md) to compile it with CUDA support. + - Ensure that any CUDA-specific build flags or paths are correctly set in your build configuration. + +- **Using the Toolbox Environment:** + - The toolbox environment is isolated from your host system, which helps prevent conflicts. + - Remember that system files and configurations inside the toolbox are separate from the host. By default the home directory of the user is shared between the host and the toolbox. + +--- + +**Disclaimer:** Manually installing and modifying system packages can lead to instability of the container. The above steps are provided as a guideline and may need adjustments based on your specific system configuration. Always back up important data before making significant system changes, especially as your home folder is writable and shared with he toolbox. + +**Acknowledgments:** Special thanks to the Fedora community and NVIDIA documentation for providing resources that assisted in creating this guide. + +## References + +- [Fedora Toolbox Documentation](https://docs.fedoraproject.org/en-US/fedora-silverblue/toolbox/) +- [NVIDIA CUDA Installation Guide](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html) +- [Podman Documentation](https://podman.io/get-started) + +--- diff --git a/docs/development/HOWTO-add-model.md b/docs/development/HOWTO-add-model.md index 04c5ccbbe..8fcd70811 100644 --- a/docs/development/HOWTO-add-model.md +++ b/docs/development/HOWTO-add-model.md @@ -28,7 +28,7 @@ The required steps to implement for an HF model are: ```python @Model.register("MyModelForCausalLM") class MyModel(Model): - model_arch = gguf.MODEL_ARCH.GROK + model_arch = gguf.MODEL_ARCH.MYMODEL ``` 2. Define the layout of the GGUF tensors in [constants.py](/gguf-py/gguf/constants.py) @@ -79,14 +79,14 @@ Depending on the model configuration, tokenizer, code and tensors layout, you wi - `Model#set_vocab` - `Model#write_tensors` -NOTE: Tensor names must end with `.weight` suffix, that is the convention and several tools like `quantize` expect this to proceed the weights. +NOTE: Tensor names must end with `.weight` or `.bias` suffixes, that is the convention and several tools like `quantize` expect this to proceed the weights. ### 2. Define the model architecture in `llama.cpp` The model params and tensors layout must be defined in `llama.cpp`: 1. Define a new `llm_arch` 2. Define the tensors layout in `LLM_TENSOR_NAMES` -3. Add any non standard metadata in `llm_load_hparams` +3. Add any non-standard metadata in `llm_load_hparams` 4. Create the tensors for inference in `llm_load_tensors` 5. If the model has a RoPE operation, add the rope type in `llama_rope_type` @@ -96,9 +96,9 @@ NOTE: The dimensions in `ggml` are typically in the reverse order of the `pytorc This is the funniest part, you have to provide the inference graph implementation of the new model architecture in `llama_build_graph`. -Have a look at existing implementation like `build_llama`, `build_dbrx` or `build_bert`. +Have a look at existing implementations like `build_llama`, `build_dbrx` or `build_bert`. -When implementing a new graph, please note that the underlying `ggml` backends might not support them all, support for missing backend operations can be added in another PR. +Some `ggml` backends do not support all operations. Backend implementations can be added in a separate PR. Note: to debug the inference graph: you can use [llama-eval-callback](/examples/eval-callback/). diff --git a/examples/CMakeLists.txt b/examples/CMakeLists.txt index d63a96c1c..66cfab2c3 100644 --- a/examples/CMakeLists.txt +++ b/examples/CMakeLists.txt @@ -6,20 +6,26 @@ find_package(Threads REQUIRED) # ... +# flags + +llama_add_compile_flags() + # examples include_directories(${CMAKE_CURRENT_SOURCE_DIR}) if (EMSCRIPTEN) else() - add_subdirectory(cvector-generator) add_subdirectory(batched-bench) add_subdirectory(batched) - add_subdirectory(convert-llama2c-to-ggml) add_subdirectory(embedding) add_subdirectory(eval-callback) - add_subdirectory(export-lora) - add_subdirectory(gbnf-validator) + + if (NOT WIN32) + # disabled on Windows because it uses internal functions not exported with LLAMA_API + add_subdirectory(gbnf-validator) + endif() + add_subdirectory(gguf-hash) add_subdirectory(gguf-split) add_subdirectory(gguf) @@ -27,28 +33,41 @@ else() add_subdirectory(imatrix) add_subdirectory(infill) add_subdirectory(llama-bench) - add_subdirectory(llava) add_subdirectory(lookahead) add_subdirectory(lookup) add_subdirectory(main) add_subdirectory(parallel) add_subdirectory(passkey) add_subdirectory(perplexity) - add_subdirectory(quantize-stats) add_subdirectory(quantize) add_subdirectory(retrieval) - if (GGML_RPC) - add_subdirectory(rpc) - endif() if (LLAMA_BUILD_SERVER) - add_subdirectory(server) - endif() - if (GGML_SYCL) - add_subdirectory(sycl) + add_subdirectory(server) endif() add_subdirectory(save-load-state) + add_subdirectory(run) add_subdirectory(simple) add_subdirectory(simple-chat) add_subdirectory(speculative) + add_subdirectory(speculative-simple) add_subdirectory(tokenize) + add_subdirectory(tts) + add_subdirectory(gen-docs) + if (NOT GGML_BACKEND_DL) + # these examples use the backends directly and cannot be built with dynamic loading + add_subdirectory(convert-llama2c-to-ggml) + add_subdirectory(cvector-generator) + add_subdirectory(export-lora) + if (NOT WIN32) + # disabled on Windows because it uses internal functions not exported with LLAMA_API + add_subdirectory(quantize-stats) + endif() + add_subdirectory(llava) + if (GGML_RPC) + add_subdirectory(rpc) + endif() + if (GGML_SYCL) + add_subdirectory(sycl) + endif() + endif() endif() diff --git a/examples/base-translate.sh b/examples/base-translate.sh deleted file mode 100755 index 103a52f55..000000000 --- a/examples/base-translate.sh +++ /dev/null @@ -1,61 +0,0 @@ -#!/bin/bash -# -# Few-shot translation example. -# Requires a base model (i.e. no fine-tuned or instruct models). -# -# Usage: -# -# cd llama.cpp -# make -j -# -# ./examples/base-translate.sh "" [extra-main-args] -# - -if [ $# -lt 2 ]; then - echo "Usage: ./base-translate.sh \"\" [extra-main-args]" - exit 1 -fi - -eargs="" -if [ $# -gt 2 ]; then - eargs="${@:3}" -fi - -ftmp="__llama.cpp_example_tmp__.txt" -trap "rm -f $ftmp" EXIT - -echo "Translate from English to French: - -=== - -sea otter, peppermint, plush girafe: - -sea otter => loutre de mer -peppermint => menthe poivrée -plush girafe => girafe peluche - -=== - -violin - -violin => violon - -=== - -phone, computer, mouse, keyboard: - -phone => téléphone -computer => ordinateur -mouse => souris -keyboard => clavier - -=== -" > $ftmp - -echo "$2 -" >> $ftmp - -model=$1 - -# generate the most likely continuation until the string "===" is found -./llama-cli -m $model -f $ftmp -n 64 --temp 0 --repeat-penalty 1.0 --no-penalize-nl -r "===" $eargs diff --git a/examples/batched-bench/CMakeLists.txt b/examples/batched-bench/CMakeLists.txt index 959acaeee..68ad707f3 100644 --- a/examples/batched-bench/CMakeLists.txt +++ b/examples/batched-bench/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-batched-bench) add_executable(${TARGET} batched-bench.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/batched-bench/batched-bench.cpp b/examples/batched-bench/batched-bench.cpp index a3b21ad6b..0659ab6f1 100644 --- a/examples/batched-bench/batched-bench.cpp +++ b/examples/batched-bench/batched-bench.cpp @@ -38,7 +38,7 @@ int main(int argc, char ** argv) { llama_model_params model_params = common_model_params_to_llama(params); - llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params); + llama_model * model = llama_model_load_from_file(params.model.c_str(), model_params); if (model == NULL) { fprintf(stderr , "%s: error: unable to load model\n" , __func__); @@ -50,7 +50,7 @@ int main(int argc, char ** argv) { // ensure enough sequences are available ctx_params.n_seq_max = n_pl.empty() ? 1 : *std::max_element(n_pl.begin(), n_pl.end()); - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); if (ctx == NULL) { fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__); @@ -194,7 +194,7 @@ int main(int argc, char ** argv) { llama_batch_free(batch); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); llama_backend_free(); diff --git a/examples/batched.swift/Sources/main.swift b/examples/batched.swift/Sources/main.swift index 10f2e7fd1..371917b2e 100644 --- a/examples/batched.swift/Sources/main.swift +++ b/examples/batched.swift/Sources/main.swift @@ -23,12 +23,12 @@ defer { } let model_params = llama_model_default_params() -guard let model = llama_load_model_from_file(modelPath.cString(using: .utf8), model_params) else { +guard let model = llama_model_load_from_file(modelPath.cString(using: .utf8), model_params) else { print("Failed to load model") exit(1) } defer { - llama_free_model(model) + llama_model_free(model) } var tokens = tokenize(text: prompt, add_bos: true) @@ -141,7 +141,7 @@ while n_cur <= n_len { let new_token_id = llama_sampler_sample(smpl, context, i_batch[i]) // is it an end of stream? -> mark the stream as finished - if llama_token_is_eog(model, new_token_id) || n_cur == n_len { + if llama_vocab_is_eog(model, new_token_id) || n_cur == n_len { i_batch[i] = -1 // print("") if n_parallel > 1 { diff --git a/examples/batched/CMakeLists.txt b/examples/batched/CMakeLists.txt index 77e33343b..0d439f498 100644 --- a/examples/batched/CMakeLists.txt +++ b/examples/batched/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-batched) add_executable(${TARGET} batched.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/batched/batched.cpp b/examples/batched/batched.cpp index 3b554033e..21b95ef5e 100644 --- a/examples/batched/batched.cpp +++ b/examples/batched/batched.cpp @@ -41,17 +41,19 @@ int main(int argc, char ** argv) { llama_model_params model_params = common_model_params_to_llama(params); - llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params); + llama_model * model = llama_model_load_from_file(params.model.c_str(), model_params); if (model == NULL) { LOG_ERR("%s: error: unable to load model\n" , __func__); return 1; } + const llama_vocab * vocab = llama_model_get_vocab(model); + // tokenize the prompt std::vector tokens_list; - tokens_list = common_tokenize(model, params.prompt, true); + tokens_list = common_tokenize(vocab, params.prompt, true); const int n_kv_req = tokens_list.size() + (n_predict - tokens_list.size())*n_parallel; @@ -62,16 +64,17 @@ int main(int argc, char ** argv) { ctx_params.n_ctx = n_kv_req; ctx_params.n_batch = std::max(n_predict, n_parallel); - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); auto sparams = llama_sampler_chain_default_params(); + sparams.no_perf = false; llama_sampler * smpl = llama_sampler_chain_init(sparams); - llama_sampler_chain_add(smpl, llama_sampler_init_top_k(params.sparams.top_k)); - llama_sampler_chain_add(smpl, llama_sampler_init_top_p(params.sparams.top_p, params.sparams.min_keep)); - llama_sampler_chain_add(smpl, llama_sampler_init_temp (params.sparams.temp)); - llama_sampler_chain_add(smpl, llama_sampler_init_dist (params.sparams.seed)); + llama_sampler_chain_add(smpl, llama_sampler_init_top_k(params.sampling.top_k)); + llama_sampler_chain_add(smpl, llama_sampler_init_top_p(params.sampling.top_p, params.sampling.min_keep)); + llama_sampler_chain_add(smpl, llama_sampler_init_temp (params.sampling.temp)); + llama_sampler_chain_add(smpl, llama_sampler_init_dist (params.sampling.seed)); if (ctx == NULL) { LOG_ERR("%s: error: failed to create the llama_context\n" , __func__); @@ -119,8 +122,8 @@ int main(int argc, char ** argv) { } llama_token decoder_start_token_id = llama_model_decoder_start_token(model); - if (decoder_start_token_id == -1) { - decoder_start_token_id = llama_token_bos(model); + if (decoder_start_token_id == LLAMA_TOKEN_NULL) { + decoder_start_token_id = llama_vocab_bos(vocab); } common_batch_clear(batch); @@ -173,7 +176,7 @@ int main(int argc, char ** argv) { const llama_token new_token_id = llama_sampler_sample(smpl, ctx, i_batch[i]); // is it an end of generation? -> mark the stream as finished - if (llama_token_is_eog(model, new_token_id) || n_cur == n_predict) { + if (llama_vocab_is_eog(vocab, new_token_id) || n_cur == n_predict) { i_batch[i] = -1; LOG("\n"); if (n_parallel > 1) { @@ -235,7 +238,7 @@ int main(int argc, char ** argv) { llama_sampler_free(smpl); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); llama_backend_free(); diff --git a/examples/chat-persistent.sh b/examples/chat-persistent.sh index d9cab9836..9d761ebb8 100755 --- a/examples/chat-persistent.sh +++ b/examples/chat-persistent.sh @@ -23,8 +23,9 @@ CUR_PROMPT_CACHE="${CHAT_SAVE_DIR}/current-cache.bin" NEXT_PROMPT_FILE="${CHAT_SAVE_DIR}/next-prompt.txt" NEXT_PROMPT_CACHE="${CHAT_SAVE_DIR}/next-cache.bin" -SESSION_SIZE_MSG_PATTERN='main: session file matches [[:digit:]]+ / [[:digit:]]+' -SAMPLE_TIME_MSG_PATTERN='sample time =[[:space:]]+[[:digit:]]+.[[:digit:]]+ ms /[[:space:]]+[[:digit:]]+' +SESSION_AND_SAMPLE_PATTERN='main: session file matches [[:digit:]]+ / [[:digit:]]+'\ +'|'\ +'sampling time =[[:space:]]+[[:digit:]]+.[[:digit:]]+ ms /[[:space:]]+[[:digit:]]+' SED_DELETE_MESSAGES="/^(${USER_NAME}:|${AI_NAME}:|\\.\\.\\.)/,\$d" CTX_SIZE=2048 @@ -129,15 +130,12 @@ while read -e line; do printf ' ' - # HACK get num tokens from debug message - # TODO get both messages in one go - if ! session_size_msg="$(tail -n30 "$LOG" | grep -oE "$SESSION_SIZE_MSG_PATTERN")" || - ! sample_time_msg="$(tail -n10 "$LOG" | grep -oE "$SAMPLE_TIME_MSG_PATTERN")"; then + if ! session_and_sample_msg=$(tail -n30 "$LOG" | grep -oE "$SESSION_AND_SAMPLE_PATTERN"); then echo >&2 "Couldn't get number of tokens from ./llama-cli output!" exit 1 fi - n_tokens=$(($(cut -d/ -f2 <<<"$session_size_msg") + $(cut -d/ -f2 <<<"$sample_time_msg"))) + n_tokens=$(awk '{sum+=$1} END {print sum}' <<< "$(cut -d/ -f2 <<< "$session_and_sample_msg")") if ((n_tokens > CTX_ROTATE_POINT)); then tail -c+$((n_prompt_len_pre + 1)) "$CUR_PROMPT_FILE" >>"$NEXT_PROMPT_FILE" diff --git a/examples/convert-llama2c-to-ggml/CMakeLists.txt b/examples/convert-llama2c-to-ggml/CMakeLists.txt index a6790e617..44e5f722a 100644 --- a/examples/convert-llama2c-to-ggml/CMakeLists.txt +++ b/examples/convert-llama2c-to-ggml/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-convert-llama2c-to-ggml) add_executable(${TARGET} convert-llama2c-to-ggml.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/convert-llama2c-to-ggml/README.md b/examples/convert-llama2c-to-ggml/README.md index 5774ac83c..46a42da69 100644 --- a/examples/convert-llama2c-to-ggml/README.md +++ b/examples/convert-llama2c-to-ggml/README.md @@ -2,11 +2,8 @@ This example reads weights from project [llama2.c](https://github.com/karpathy/llama2.c) and saves them in ggml compatible format. The vocab that is available in `models/ggml-vocab.bin` is used by default. -To convert the model first download the models from the [llama2.c](https://github.com/karpathy/llama2.c) repository: +To convert the model first download the models from the [llama2.c](https://github.com/karpathy/llama2.c) repository. -`$ make -j` - -After successful compilation, following usage options are available: ``` usage: ./llama-convert-llama2c-to-ggml [options] diff --git a/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp b/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp index 988a584c9..bdf0eed2a 100644 --- a/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp +++ b/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp @@ -1,4 +1,6 @@ #include "ggml.h" +#include "gguf.h" + #include "llama.h" #include "common.h" #include "log.h" @@ -434,12 +436,12 @@ static void print_matrix(struct ggml_tensor * probs) { } } -struct llama_file { +struct my_llama_file { // use FILE * so we don't have to re-open the file to mmap FILE * fp; size_t size; - llama_file(const char * fname, const char * mode) { + my_llama_file(const char * fname, const char * mode) { fp = std::fopen(fname, mode); if (fp == NULL) { size = 0; @@ -500,7 +502,7 @@ struct llama_file { return std::string(chars.data(), len); } - ~llama_file() { + ~my_llama_file() { if (fp) { std::fclose(fp); } @@ -508,7 +510,7 @@ struct llama_file { }; static bool is_ggml_file(const char * filename) { - llama_file file(filename, "rb"); + my_llama_file file(filename, "rb"); if (file.size < 4) { return false; } @@ -576,7 +578,7 @@ static void load_vocab(const char * filename, const Config * config, struct my_l } else { // assume llama2.c vocabulary LOG_INF("%s: Assuming llama2.c vocabulary since %s is not a gguf file\n", __func__, filename); - llama_file file(filename, "rb"); + my_llama_file file(filename, "rb"); if (!file.fp) { die_fmt("%s: %s", strerror(errno), filename); } @@ -689,8 +691,8 @@ static void save_as_llama_model( gguf_set_val_u32(ctx, KV_TOKENIZER_UNK_ID, UNKNOWN_TOKEN_ID); gguf_set_val_u32(ctx, KV_TOKENIZER_BOS_ID, BOS_TOKEN_ID); gguf_set_val_u32(ctx, KV_TOKENIZER_EOS_ID, EOS_TOKEN_ID); - gguf_set_val_u32(ctx, KV_TOKENIZER_SEP_ID, -1); - gguf_set_val_u32(ctx, KV_TOKENIZER_PAD_ID, -1); + gguf_set_val_u32(ctx, KV_TOKENIZER_SEP_ID, LLAMA_TOKEN_NULL); + gguf_set_val_u32(ctx, KV_TOKENIZER_PAD_ID, LLAMA_TOKEN_NULL); gguf_set_val_u32(ctx, KV_CONTEXT_LENGTH, model->hparams.n_ctx); gguf_set_val_u32(ctx, KV_EMBEDDING_LENGTH, model->hparams.n_embd); @@ -909,7 +911,7 @@ int main(int argc, char ** argv) { load_vocab(params.fn_vocab_model, &config, &vocab); struct my_llama_model model; - model.hparams.n_vocab = config.vocab_size; //llama_n_vocab(lctx); + model.hparams.n_vocab = config.vocab_size; //llama_vocab_n_vocab(lctx); model.hparams.n_ctx = params.n_ctx; model.hparams.n_embd = config.dim; //params.n_embd; model.hparams.n_ff = config.hidden_dim; diff --git a/examples/convert_legacy_llama.py b/examples/convert_legacy_llama.py index 9ab9ab06e..c4ec5c524 100755 --- a/examples/convert_legacy_llama.py +++ b/examples/convert_legacy_llama.py @@ -840,6 +840,8 @@ class OutputFile: self.gguf.add_base_model_version(key, base_model_entry["version"]) if "organization" in base_model_entry: self.gguf.add_base_model_organization(key, base_model_entry["organization"]) + if "description" in base_model_entry: + self.gguf.add_base_model_description(key, base_model_entry["description"]) if "url" in base_model_entry: self.gguf.add_base_model_url(key, base_model_entry["url"]) if "doi" in base_model_entry: @@ -849,12 +851,32 @@ class OutputFile: if "repo_url" in base_model_entry: self.gguf.add_base_model_repo_url(key, base_model_entry["repo_url"]) + if metadata.datasets is not None: + self.gguf.add_dataset_count(len(metadata.datasets)) + for key, dataset_entry in enumerate(metadata.datasets): + if "name" in dataset_entry: + self.gguf.add_dataset_name(key, dataset_entry["name"]) + if "author" in dataset_entry: + self.gguf.add_dataset_author(key, dataset_entry["author"]) + if "version" in dataset_entry: + self.gguf.add_dataset_version(key, dataset_entry["version"]) + if "organization" in dataset_entry: + self.gguf.add_dataset_organization(key, dataset_entry["organization"]) + if "description" in dataset_entry: + self.gguf.add_dataset_description(key, dataset_entry["description"]) + if "url" in dataset_entry: + self.gguf.add_dataset_url(key, dataset_entry["url"]) + if "doi" in dataset_entry: + self.gguf.add_dataset_doi(key, dataset_entry["doi"]) + if "uuid" in dataset_entry: + self.gguf.add_dataset_uuid(key, dataset_entry["uuid"]) + if "repo_url" in dataset_entry: + self.gguf.add_dataset_repo_url(key, dataset_entry["repo_url"]) + if metadata.tags is not None: self.gguf.add_tags(metadata.tags) if metadata.languages is not None: self.gguf.add_languages(metadata.languages) - if metadata.datasets is not None: - self.gguf.add_datasets(metadata.datasets) def add_meta_arch(self, params: Params) -> None: # Metadata About The Neural Architecture Itself diff --git a/examples/cvector-generator/CMakeLists.txt b/examples/cvector-generator/CMakeLists.txt index 0a559d60c..49ad9561c 100644 --- a/examples/cvector-generator/CMakeLists.txt +++ b/examples/cvector-generator/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-cvector-generator) add_executable(${TARGET} cvector-generator.cpp pca.hpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/cvector-generator/cvector-generator.cpp b/examples/cvector-generator/cvector-generator.cpp index d1731bba6..413b71d34 100644 --- a/examples/cvector-generator/cvector-generator.cpp +++ b/examples/cvector-generator/cvector-generator.cpp @@ -1,7 +1,9 @@ +#include "ggml.h" +#include "gguf.h" + #include "arg.h" #include "common.h" #include "llama.h" -#include "ggml.h" #include "pca.hpp" #include "mean.hpp" @@ -271,7 +273,9 @@ struct tokenized_prompt { size_t max_seq_len; tokenized_prompt(llama_context * ctx, std::string pos, std::string neg) { - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + const bool add_bos = llama_vocab_get_add_bos(vocab); tokens_pos = common_tokenize(ctx, pos, add_bos, true); tokens_neg = common_tokenize(ctx, neg, add_bos, true); max_seq_len = std::max(tokens_pos.size(), tokens_neg.size()); @@ -415,12 +419,13 @@ int main(int argc, char ** argv) { // load the model to get hparams common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); // int n_ctx = llama_n_ctx(ctx); - int n_layers = llama_n_layer(model); - int n_embd = llama_n_embd(model); + int n_layers = llama_model_n_layer(model); + int n_embd = llama_model_n_embd(model); + // get model hint param (a.k.a model arch name) char model_hint[128]; llama_model_meta_val_str(model, "general.architecture", model_hint, 128); @@ -474,8 +479,6 @@ int main(int argc, char ** argv) { // done with the model, we can now free it to make gain some memory printf("Done evaluate prompts, unload model...\n"); - llama_free(ctx); - llama_free_model(model); bool use_pca = params.cvector_dimre_method == DIMRE_METHOD_PCA; diff --git a/examples/cvector-generator/mean.hpp b/examples/cvector-generator/mean.hpp index 16be5ce3e..4eeac1eeb 100644 --- a/examples/cvector-generator/mean.hpp +++ b/examples/cvector-generator/mean.hpp @@ -15,7 +15,7 @@ static void run( for (size_t il = 0; il < v_input.size(); ++il) { // prepare output vector struct ggml_tensor * ctrl_out = v_output[il]; - ggml_format_name(ctrl_out, "direction.%ld", il+1); + ggml_format_name(ctrl_out, "direction.%zu", il+1); // calculate mean vector struct ggml_tensor * t_layer = v_input[il]; diff --git a/examples/cvector-generator/pca.hpp b/examples/cvector-generator/pca.hpp index f6e307fbc..e88bbdde9 100644 --- a/examples/cvector-generator/pca.hpp +++ b/examples/cvector-generator/pca.hpp @@ -302,7 +302,7 @@ static void run_pca( // prepare output vector struct ggml_tensor * ctrl_out = v_output[il]; - ggml_format_name(ctrl_out, "direction.%ld", il+1); + ggml_format_name(ctrl_out, "direction.%zu", il+1); // run power_iteration params.i_layer = il; diff --git a/examples/deprecation-warning/deprecation-warning.cpp b/examples/deprecation-warning/deprecation-warning.cpp index 11b35d2c2..c2958ea12 100644 --- a/examples/deprecation-warning/deprecation-warning.cpp +++ b/examples/deprecation-warning/deprecation-warning.cpp @@ -12,7 +12,7 @@ int main(int argc, char** argv) { } // Get only the program name from the full path - auto pos = filename.find_last_of('/'); + auto pos = filename.find_last_of("/\\"); if (pos != std::string::npos) { filename = filename.substr(pos+1); } diff --git a/examples/embedding/CMakeLists.txt b/examples/embedding/CMakeLists.txt index 8256e789a..809040307 100644 --- a/examples/embedding/CMakeLists.txt +++ b/examples/embedding/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-embedding) add_executable(${TARGET} embedding.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/embedding/embedding.cpp b/examples/embedding/embedding.cpp index 3f18fc6a7..38d22c90f 100644 --- a/examples/embedding/embedding.cpp +++ b/examples/embedding/embedding.cpp @@ -97,14 +97,17 @@ int main(int argc, char ** argv) { // load the model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + if (model == NULL) { LOG_ERR("%s: unable to load model\n", __func__); return 1; } - const int n_ctx_train = llama_n_ctx_train(model); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_ctx_train = llama_model_n_ctx_train(model); const int n_ctx = llama_n_ctx(ctx); const enum llama_pooling_type pooling_type = llama_pooling_type(ctx); @@ -147,7 +150,7 @@ int main(int argc, char ** argv) { // check if the last token is SEP // it should be automatically added by the tokenizer when 'tokenizer.ggml.add_eos_token' is set to 'true' for (auto & inp : inputs) { - if (inp.empty() || inp.back() != llama_token_sep(model)) { + if (inp.empty() || inp.back() != llama_vocab_sep(vocab)) { LOG_WRN("%s: last token in the prompt is not SEP\n", __func__); LOG_WRN("%s: 'tokenizer.ggml.add_eos_token' should be set to 'true' in the GGUF header\n", __func__); } @@ -180,7 +183,7 @@ int main(int argc, char ** argv) { } // allocate output - const int n_embd = llama_n_embd(model); + const int n_embd = llama_model_n_embd(model); std::vector embeddings(n_embd_count * n_embd, 0); float * emb = embeddings.data(); @@ -316,8 +319,6 @@ int main(int argc, char ** argv) { // clean up llama_batch_free(batch); - llama_free(ctx); - llama_free_model(model); llama_backend_free(); return 0; diff --git a/examples/eval-callback/CMakeLists.txt b/examples/eval-callback/CMakeLists.txt index a48753d38..95915ed91 100644 --- a/examples/eval-callback/CMakeLists.txt +++ b/examples/eval-callback/CMakeLists.txt @@ -2,8 +2,9 @@ set(TARGET llama-eval-callback) add_executable(${TARGET} eval-callback.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TEST_TARGET test-eval-callback) -add_test(NAME ${TEST_TARGET} COMMAND llama-eval-callback --hf-repo ggml-org/models --hf-file tinyllamas/stories260K.gguf --model stories260K.gguf --prompt hello --seed 42 -ngl 0) +add_test(NAME ${TEST_TARGET} + COMMAND llama-eval-callback --hf-repo ggml-org/models --hf-file tinyllamas/stories260K.gguf --model stories260K.gguf --prompt hello --seed 42 -ngl 0) set_property(TEST ${TEST_TARGET} PROPERTY LABELS eval-callback curl) diff --git a/examples/eval-callback/eval-callback.cpp b/examples/eval-callback/eval-callback.cpp index c08e3e5f6..fb188f5a9 100644 --- a/examples/eval-callback/eval-callback.cpp +++ b/examples/eval-callback/eval-callback.cpp @@ -127,7 +127,10 @@ static bool ggml_debug(struct ggml_tensor * t, bool ask, void * user_data) { } static bool run(llama_context * ctx, const common_params & params) { - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const bool add_bos = llama_vocab_get_add_bos(vocab); std::vector tokens = common_tokenize(ctx, params.prompt, add_bos); @@ -162,8 +165,9 @@ int main(int argc, char ** argv) { // init common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + if (model == nullptr || ctx == nullptr) { LOG_ERR("%s : failed to init\n", __func__); return 1; @@ -184,9 +188,6 @@ int main(int argc, char ** argv) { LOG("\n"); llama_perf_context_print(ctx); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); return 0; diff --git a/examples/export-lora/CMakeLists.txt b/examples/export-lora/CMakeLists.txt index 1cef6e716..310455787 100644 --- a/examples/export-lora/CMakeLists.txt +++ b/examples/export-lora/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-export-lora) add_executable(${TARGET} export-lora.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/export-lora/export-lora.cpp b/examples/export-lora/export-lora.cpp index 67662313d..99063b5d5 100644 --- a/examples/export-lora/export-lora.cpp +++ b/examples/export-lora/export-lora.cpp @@ -1,12 +1,13 @@ -#include "arg.h" -#include "common.h" #include "ggml.h" #include "ggml-alloc.h" +#include "gguf.h" + +#include "arg.h" +#include "common.h" #include #include #include -#include #include static bool g_verbose = false; @@ -128,7 +129,7 @@ struct lora_merge_ctx { lora_merge_ctx( std::string & base_fname, - std::vector & lora_files, + std::vector & lora_files, std::string & outfile, int n_threads) : base_model(base_fname, 0), n_threads(n_threads), fout(outfile, std::ios::binary) { fout.exceptions(std::ofstream::failbit); // fail fast on write errors @@ -265,8 +266,8 @@ struct lora_merge_ctx { fout.write((const char *)data.data(), data.size()); } - printf("%s : merged %ld tensors with lora adapters\n", __func__, n_merged); - printf("%s : wrote %ld tensors to output file\n", __func__, trans.size()); + printf("%s : merged %zu tensors with lora adapters\n", __func__, n_merged); + printf("%s : wrote %zu tensors to output file\n", __func__, trans.size()); } void copy_tensor(struct ggml_tensor * base) { @@ -352,7 +353,7 @@ struct lora_merge_ctx { const float scale = alpha ? adapters[i]->scale * alpha / rank : adapters[i]->scale; delta = ggml_scale(ctx0, delta, scale); cur = ggml_add(ctx0, delta, cur); - printf("%s : + merging from adapter[%ld] type=%s\n", __func__, i, ggml_type_name(inp_a[i]->type)); + printf("%s : + merging from adapter[%zu] type=%s\n", __func__, i, ggml_type_name(inp_a[i]->type)); printf("%s : input_scale=%f calculated_scale=%f rank=%d\n", __func__, adapters[i]->scale, scale, (int) inp_b[i]->ne[0]); } cur = ggml_cast(ctx0, cur, out->type); diff --git a/examples/gbnf-validator/CMakeLists.txt b/examples/gbnf-validator/CMakeLists.txt index 4edd6ec73..d2cb524c0 100644 --- a/examples/gbnf-validator/CMakeLists.txt +++ b/examples/gbnf-validator/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-gbnf-validator) add_executable(${TARGET} gbnf-validator.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gbnf-validator/gbnf-validator.cpp b/examples/gbnf-validator/gbnf-validator.cpp index 7493af9d3..17a0e27c4 100644 --- a/examples/gbnf-validator/gbnf-validator.cpp +++ b/examples/gbnf-validator/gbnf-validator.cpp @@ -11,19 +11,15 @@ static bool llama_grammar_validate(struct llama_grammar * grammar, const std::string & input_str, size_t & error_pos, std::string & error_msg) { const auto cpts = unicode_cpts_from_utf8(input_str); - const llama_grammar_rules & rules = llama_grammar_get_rules (grammar); - llama_grammar_stacks & stacks_cur = llama_grammar_get_stacks(grammar); + auto & stacks_cur = llama_grammar_get_stacks(grammar); size_t pos = 0; for (const auto & cpt : cpts) { - const llama_grammar_stacks stacks_prev = llama_grammar_get_stacks(grammar); // copy - - llama_grammar_accept(rules, stacks_prev, cpt, stacks_cur); + llama_grammar_accept(grammar, cpt); if (stacks_cur.empty()) { error_pos = pos; error_msg = "Unexpected character '" + unicode_cpt_to_utf8(cpt) + "'"; - stacks_cur = stacks_prev; return false; } ++pos; @@ -82,7 +78,8 @@ int main(int argc, char** argv) { llama_grammar * grammar = llama_grammar_init_impl(nullptr, grammar_str.c_str(), "root"); if (grammar == nullptr) { - throw std::runtime_error("Failed to initialize llama_grammar"); + fprintf(stdout, "Failed to initialize llama_grammar\n"); + return 1; } // Read the input file std::string input_str; diff --git a/examples/gen-docs/CMakeLists.txt b/examples/gen-docs/CMakeLists.txt index c94cda776..25de0af35 100644 --- a/examples/gen-docs/CMakeLists.txt +++ b/examples/gen-docs/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-gen-docs) add_executable(${TARGET} gen-docs.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gguf-hash/CMakeLists.txt b/examples/gguf-hash/CMakeLists.txt index 633f45535..15c5c68c6 100644 --- a/examples/gguf-hash/CMakeLists.txt +++ b/examples/gguf-hash/CMakeLists.txt @@ -4,12 +4,19 @@ install(TARGETS ${TARGET} RUNTIME) # clibs dependencies include_directories(deps/) + add_library(xxhash OBJECT deps/xxhash/xxhash.c deps/xxhash/xxhash.h) target_link_libraries(${TARGET} PRIVATE xxhash) + add_library(sha1 OBJECT deps/sha1/sha1.c deps/sha1/sha1.h) target_link_libraries(${TARGET} PRIVATE sha1) +if (NOT MSVC) + # disable warnings in 3rd party code + target_compile_options(sha1 PRIVATE -w) +endif() + add_library(sha256 OBJECT deps/sha256/sha256.c deps/sha256/sha256.h) target_link_libraries(${TARGET} PRIVATE sha256) target_link_libraries(${TARGET} PRIVATE ggml ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gguf-hash/gguf-hash.cpp b/examples/gguf-hash/gguf-hash.cpp index e96c75117..9523ec122 100644 --- a/examples/gguf-hash/gguf-hash.cpp +++ b/examples/gguf-hash/gguf-hash.cpp @@ -1,4 +1,5 @@ #include "ggml.h" +#include "gguf.h" #include /* abort() */ #include diff --git a/examples/gguf-split/CMakeLists.txt b/examples/gguf-split/CMakeLists.txt index f63887da7..c407e2f0a 100644 --- a/examples/gguf-split/CMakeLists.txt +++ b/examples/gguf-split/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-gguf-split) add_executable(${TARGET} gguf-split.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gguf-split/gguf-split.cpp b/examples/gguf-split/gguf-split.cpp index 7e62657e1..ef3ceb686 100644 --- a/examples/gguf-split/gguf-split.cpp +++ b/examples/gguf-split/gguf-split.cpp @@ -1,18 +1,19 @@ +#include "ggml.h" +#include "gguf.h" #include "llama.h" #include "common.h" #include -#include +#include +#include +#include #include +#include +#include #include #include #include -#include -#include -#include -#include - #if defined(_WIN32) #include #ifndef PATH_MAX @@ -287,7 +288,7 @@ struct split_strategy { } void print_info() { - printf("n_split: %ld\n", ctx_outs.size()); + printf("n_split: %zu\n", ctx_outs.size()); int i_split = 0; for (auto & ctx_out : ctx_outs) { // re-calculate the real gguf size for each split (= metadata size + total size of all tensors) @@ -297,7 +298,7 @@ struct split_strategy { total_size += ggml_nbytes(t); } total_size = total_size / 1000 / 1000; // convert to megabytes - printf("split %05d: n_tensors = %d, total_size = %ldM\n", i_split + 1, gguf_get_n_tensors(ctx_out), total_size); + printf("split %05d: n_tensors = %" PRIi64 ", total_size = %zuM\n", i_split + 1, gguf_get_n_tensors(ctx_out), total_size); i_split++; } } diff --git a/examples/gguf-split/tests.sh b/examples/gguf-split/tests.sh index d5a92d605..05a932227 100755 --- a/examples/gguf-split/tests.sh +++ b/examples/gguf-split/tests.sh @@ -41,7 +41,7 @@ echo PASS echo # 2b. Test the sharded model is loading properly -$MAIN --model $WORK_PATH/ggml-model-split-00001-of-00006.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-split-00001-of-00006.gguf --n-predict 32 echo PASS echo @@ -51,7 +51,7 @@ echo PASS echo # 3b. Test the merged model is loading properly -$MAIN --model $WORK_PATH/ggml-model-merge.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-merge.gguf --n-predict 32 echo PASS echo @@ -61,7 +61,7 @@ echo PASS echo # 4b. Test the sharded model is loading properly -$MAIN --model $WORK_PATH/ggml-model-split-32-tensors-00001-of-00007.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-split-32-tensors-00001-of-00007.gguf --n-predict 32 echo PASS echo @@ -71,7 +71,7 @@ echo #echo # 5b. Test the merged model is loading properly -#$MAIN --model $WORK_PATH/ggml-model-merge-2.gguf --n-predict 32 +#$MAIN -no-cnv --model $WORK_PATH/ggml-model-merge-2.gguf --n-predict 32 #echo PASS #echo @@ -81,7 +81,7 @@ echo PASS echo # 6b. Test the sharded model is loading properly -$MAIN --model $WORK_PATH/ggml-model-split-2G-00001-of-00002.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-split-2G-00001-of-00002.gguf --n-predict 32 echo PASS echo diff --git a/examples/gguf/CMakeLists.txt b/examples/gguf/CMakeLists.txt index a9569b411..fb04eb83f 100644 --- a/examples/gguf/CMakeLists.txt +++ b/examples/gguf/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-gguf) add_executable(${TARGET} gguf.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE ggml ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gguf/gguf.cpp b/examples/gguf/gguf.cpp index 7498f85ef..f31989c8c 100644 --- a/examples/gguf/gguf.cpp +++ b/examples/gguf/gguf.cpp @@ -1,10 +1,9 @@ #include "ggml.h" +#include "gguf.h" #include -#include #include #include -#include #include #undef MIN @@ -135,9 +134,10 @@ static bool gguf_ex_read_0(const std::string & fname) { for (int i = 0; i < n_tensors; ++i) { const char * name = gguf_get_tensor_name (ctx, i); + const size_t size = gguf_get_tensor_size (ctx, i); const size_t offset = gguf_get_tensor_offset(ctx, i); - printf("%s: tensor[%d]: name = %s, offset = %zu\n", __func__, i, name, offset); + printf("%s: tensor[%d]: name = %s, size = %zu, offset = %zu\n", __func__, i, name, size, offset); } } @@ -182,9 +182,10 @@ static bool gguf_ex_read_1(const std::string & fname, bool check_data) { for (int i = 0; i < n_tensors; ++i) { const char * name = gguf_get_tensor_name (ctx, i); + const size_t size = gguf_get_tensor_size (ctx, i); const size_t offset = gguf_get_tensor_offset(ctx, i); - printf("%s: tensor[%d]: name = %s, offset = %zu\n", __func__, i, name, offset); + printf("%s: tensor[%d]: name = %s, size = %zu, offset = %zu\n", __func__, i, name, size, offset); } } @@ -199,7 +200,8 @@ static bool gguf_ex_read_1(const std::string & fname, bool check_data) { struct ggml_tensor * cur = ggml_get_tensor(ctx_data, name); - printf("%s: tensor[%d]: n_dims = %d, name = %s, data = %p\n", __func__, i, ggml_n_dims(cur), cur->name, cur->data); + printf("%s: tensor[%d]: n_dims = %d, ne = (%d, %d, %d, %d), name = %s, data = %p\n", + __func__, i, ggml_n_dims(cur), int(cur->ne[0]), int(cur->ne[1]), int(cur->ne[2]), int(cur->ne[3]), cur->name, cur->data); // print first 10 elements const float * data = (const float *) cur->data; @@ -215,7 +217,7 @@ static bool gguf_ex_read_1(const std::string & fname, bool check_data) { const float * data = (const float *) cur->data; for (int j = 0; j < ggml_nelements(cur); ++j) { if (data[j] != 100 + i) { - fprintf(stderr, "%s: tensor[%d]: data[%d] = %f\n", __func__, i, j, data[j]); + fprintf(stderr, "%s: tensor[%d], data[%d]: found %f, expected %f\n", __func__, i, j, data[j], float(100 + i)); gguf_free(ctx); return false; } @@ -245,6 +247,8 @@ int main(int argc, char ** argv) { check_data = false; } + srand(123456); + const std::string fname(argv[1]); const std::string mode (argv[2]); diff --git a/examples/gritlm/CMakeLists.txt b/examples/gritlm/CMakeLists.txt index 86dfddca3..fa1b4dc70 100644 --- a/examples/gritlm/CMakeLists.txt +++ b/examples/gritlm/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-gritlm) add_executable(${TARGET} gritlm.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/gritlm/gritlm.cpp b/examples/gritlm/gritlm.cpp index 6e42fa073..72eb46257 100644 --- a/examples/gritlm/gritlm.cpp +++ b/examples/gritlm/gritlm.cpp @@ -11,6 +11,7 @@ static std::vector> encode(llama_context * ctx, const std::ve std::vector> result; const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); llama_batch batch = llama_batch_init(llama_n_batch(ctx), 0, 1); @@ -19,16 +20,16 @@ static std::vector> encode(llama_context * ctx, const std::ve const std::string input_string = instruction + sentences[i]; - std::vector inputs = common_tokenize(model, input_string, true, false); + std::vector inputs = common_tokenize(vocab, input_string, true, false); const int32_t n_toks = inputs.size(); // GritLM seems to have EOS = "" // https://github.com/ContextualAI/gritlm/blob/92025b16534712b31b3c4aaaf069350e222bd5f8/gritlm/gritlm.py#L18 - // inputs.push_back(llama_token_eos(model)); + // inputs.push_back(llama_vocab_eos(vocab)); // we want to ignore instruction tokens for mean pooling - const int32_t n_inst = common_tokenize(model, instruction, true, false).size(); + const int32_t n_inst = common_tokenize(vocab, instruction, true, false).size(); #ifdef GRIT_DEBUG // debug tokens - should be matching as referenced in the GritLM sample @@ -52,7 +53,7 @@ static std::vector> encode(llama_context * ctx, const std::ve llama_decode(ctx, batch); // get embedding dimensions - uint64_t n_embd = llama_n_embd(model); + uint64_t n_embd = llama_model_n_embd(model); // allocate embedding output std::vector emb_unorm(n_embd, 0.0f); @@ -75,7 +76,7 @@ static std::vector> encode(llama_context * ctx, const std::ve } std::vector emb_norm(emb_unorm.size()); - common_embd_normalize(emb_unorm.data(), emb_norm.data(), n_embd); + common_embd_normalize(emb_unorm.data(), emb_norm.data(), n_embd, 2); result.push_back(emb_norm); #ifdef GRIT_DEBUG @@ -97,7 +98,9 @@ static std::string generate(llama_context * ctx, llama_sampler * smpl, const std std::string result; const llama_model * model = llama_get_model(ctx); - llama_token eos_token = llama_token_eos(model); + const llama_vocab * vocab = llama_model_get_vocab(model); + + llama_token eos_token = llama_vocab_eos(vocab); llama_kv_cache_clear(ctx); llama_set_embeddings(ctx, false); @@ -105,7 +108,7 @@ static std::string generate(llama_context * ctx, llama_sampler * smpl, const std llama_batch bat = llama_batch_init(llama_n_batch(ctx), 0, 1); - std::vector inputs = common_tokenize(model, prompt, false, true); + std::vector inputs = common_tokenize(vocab, prompt, false, true); int32_t i_current_token = 0; while (true) { @@ -165,10 +168,10 @@ int main(int argc, char * argv[]) { llama_backend_init(); - llama_model * model = llama_load_model_from_file(params.model.c_str(), mparams); + llama_model * model = llama_model_load_from_file(params.model.c_str(), mparams); // create generation context - llama_context * ctx = llama_new_context_with_model(model, cparams); + llama_context * ctx = llama_init_from_model(model, cparams); auto sparams = llama_sampler_chain_default_params(); @@ -197,7 +200,7 @@ int main(int argc, char * argv[]) { const std::vector> d_rep = encode(ctx, documents, gritlm_instruction("")); const std::vector> q_rep = encode(ctx, queries, gritlm_instruction(instruction)); - const int n_embd = llama_n_embd(model); + const int n_embd = llama_model_n_embd(model); const float cosine_sim_q0_d0 = common_embd_similarity_cos(q_rep[0].data(), d_rep[0].data(), n_embd); const float cosine_sim_q0_d1 = common_embd_similarity_cos(q_rep[0].data(), d_rep[1].data(), n_embd); @@ -219,7 +222,7 @@ int main(int argc, char * argv[]) { llama_sampler_free(smpl); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); llama_backend_free(); return 0; diff --git a/examples/imatrix/CMakeLists.txt b/examples/imatrix/CMakeLists.txt index d4c8265bd..412696c47 100644 --- a/examples/imatrix/CMakeLists.txt +++ b/examples/imatrix/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-imatrix) add_executable(${TARGET} imatrix.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/imatrix/README.md b/examples/imatrix/README.md index bb5faec94..9c056986b 100644 --- a/examples/imatrix/README.md +++ b/examples/imatrix/README.md @@ -25,8 +25,6 @@ For faster computation, make sure to use GPU offloading via the `-ngl` argument ## Example ```bash -GGML_CUDA=1 make -j - # generate importance matrix (imatrix.dat) ./llama-imatrix -m ggml-model-f16.gguf -f train-data.txt -ngl 99 diff --git a/examples/imatrix/imatrix.cpp b/examples/imatrix/imatrix.cpp index 70ff47768..b5f3feb9f 100644 --- a/examples/imatrix/imatrix.cpp +++ b/examples/imatrix/imatrix.cpp @@ -7,7 +7,6 @@ #include #include #include -#include #include #include #include @@ -40,7 +39,7 @@ public: void set_params(common_params params) { m_params = std::move(params); } bool collect_imatrix(struct ggml_tensor * t, bool ask, void * user_data); void save_imatrix(int ncall = -1) const; - bool load_imatrix(const char * file_name); + bool load_imatrix(const char * fname); private: std::unordered_map m_stats; common_params m_params; @@ -429,10 +428,14 @@ static void process_logits( } static bool compute_imatrix(llama_context * ctx, const common_params & params) { - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); - GGML_ASSERT(!llama_add_eos_token(llama_get_model(ctx))); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const bool add_bos = llama_vocab_get_add_bos(vocab); const int n_ctx = llama_n_ctx(ctx); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); + auto tim1 = std::chrono::high_resolution_clock::now(); LOG_INF("%s: tokenizing the input ..\n", __func__); @@ -467,7 +470,7 @@ static bool compute_imatrix(llama_context * ctx, const common_params & params) { const int n_chunk_max = tokens.size() / n_ctx; const int n_chunk = params.n_chunks < 0 ? n_chunk_max : std::min(params.n_chunks, n_chunk_max); - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); const int n_batch = params.n_batch; int count = 0; @@ -507,7 +510,7 @@ static bool compute_imatrix(llama_context * ctx, const common_params & params) { // add BOS token for the first batch of each chunk if (add_bos && j == 0) { - tokens[batch_start] = llama_token_bos(llama_get_model(ctx)); + tokens[batch_start] = llama_vocab_bos(vocab); } common_batch_clear(batch); @@ -618,14 +621,15 @@ int main(int argc, char ** argv) { // init common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + if (model == nullptr || ctx == nullptr) { LOG_ERR("%s : failed to init\n", __func__); return 1; } - const int n_ctx_train = llama_n_ctx_train(model); + const int n_ctx_train = llama_model_n_ctx_train(model); if (params.n_ctx > n_ctx_train) { LOG_WRN("%s: model was trained on only %d context tokens (%d specified)\n", __func__, n_ctx_train, params.n_ctx); @@ -637,18 +641,24 @@ int main(int argc, char ** argv) { LOG_INF("%s\n", common_params_get_system_info(params).c_str()); } - if (!compute_imatrix(ctx, params)) { - return 1; + if (params.prompt.empty()) { + if (params.in_files.empty()) { + LOG_ERR("Error: No prompt provided and no precomputed matrices (--in-file) to combine.\n"); + return 1; + } + LOG_INF("No prompt provided; combining precomputed matrices only.\n"); + } else { + if (!compute_imatrix(ctx, params)) { + return 1; + } } + g_collector.save_imatrix(); LOG("\n"); llama_perf_context_print(ctx); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); return 0; diff --git a/examples/infill/CMakeLists.txt b/examples/infill/CMakeLists.txt index 9b1aa3b63..fb26628d8 100644 --- a/examples/infill/CMakeLists.txt +++ b/examples/infill/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-infill) add_executable(${TARGET} infill.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/infill/README.md b/examples/infill/README.md index 810a0c5e7..df4d976f2 100644 --- a/examples/infill/README.md +++ b/examples/infill/README.md @@ -14,7 +14,7 @@ In this section, we cover the most commonly used options for running the `infill - `-m FNAME, --model FNAME`: Specify the path to the LLaMA model file (e.g., `models/7B/ggml-model.bin`). - `-i, --interactive`: Run the program in interactive mode, allowing you to provide input directly and receive real-time responses. - `-n N, --n-predict N`: Set the number of tokens to predict when generating text. Adjusting this value can influence the length of the generated text. -- `-c N, --ctx-size N`: Set the size of the prompt context. The default is 512, but LLaMA models were built with a context of 2048, which will provide better results for longer input/inference. +- `-c N, --ctx-size N`: Set the size of the prompt context. The default is 4096, but if a LLaMA model was built with a longer context, increasing this value will provide better results for longer input/inference. - `--spm-infill`: Use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. ## Input Prompts diff --git a/examples/infill/infill.cpp b/examples/infill/infill.cpp index f18362c91..489a208b6 100644 --- a/examples/infill/infill.cpp +++ b/examples/infill/infill.cpp @@ -43,50 +43,6 @@ static std::vector * g_output_tokens; static bool is_interacting = false; -static void write_logfile( - const llama_context * ctx, const common_params & params, const llama_model * model, - const std::vector & input_tokens, const std::string & output, - const std::vector & output_tokens -) { - if (params.logdir.empty()) { - return; - } - - const std::string timestamp = string_get_sortable_timestamp(); - - const bool success = fs_create_directory_with_parents(params.logdir); - if (!success) { - LOG_ERR("%s: warning: failed to create logdir %s, cannot write logfile\n", - __func__, params.logdir.c_str()); - return; - } - - const std::string logfile_path = params.logdir + timestamp + ".yml"; - FILE * logfile = fopen(logfile_path.c_str(), "w"); - - if (logfile == NULL) { - LOG_ERR("%s: failed to open logfile %s\n", __func__, logfile_path.c_str()); - return; - } - - fprintf(logfile, "binary: infill\n"); - char model_desc[128]; - llama_model_desc(model, model_desc, sizeof(model_desc)); - yaml_dump_non_result_info(logfile, params, ctx, timestamp, input_tokens, model_desc); - - fprintf(logfile, "\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "# Generation Results #\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "\n"); - - yaml_dump_string_multiline(logfile, "output", output.c_str()); - yaml_dump_vector_int(logfile, "output_tokens", output_tokens); - - llama_perf_dump_yaml(logfile, ctx); - fclose(logfile); -} - #if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32) static void sigint_handler(int signo) { if (signo == SIGINT) { @@ -96,7 +52,6 @@ static void sigint_handler(int signo) { console::cleanup(); LOG("\n"); common_perf_print(*g_ctx, *g_smpl); - write_logfile(*g_ctx, *g_params, *g_model, *g_input_tokens, g_output_ss->str(), *g_output_tokens); // make sure all logs are flushed LOG("Interrupted by user\n"); @@ -118,7 +73,7 @@ int main(int argc, char ** argv) { common_init(); - auto & sparams = params.sparams; + auto & sparams = params.sampling; console::init(params.simple_io, params.use_color); atexit([]() { console::cleanup(); }); @@ -176,15 +131,17 @@ int main(int argc, char ** argv) { LOG_INF("%s: load the model and apply lora adapter, if any\n", __func__); common_init_result llama_init = common_init_from_params(params); - model = llama_init.model; - ctx = llama_init.context; + model = llama_init.model.get(); + ctx = llama_init.context.get(); if (model == NULL) { LOG_ERR("%s: unable to load model\n", __func__); return 1; } - const int n_ctx_train = llama_n_ctx_train(model); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_ctx_train = llama_model_n_ctx_train(model); const int n_ctx = llama_n_ctx(ctx); LOG_DBG("n_ctx: %d\n", n_ctx); @@ -197,28 +154,28 @@ int main(int argc, char ** argv) { LOG_INF("\n"); LOG_INF("%s\n", common_params_get_system_info(params).c_str()); } - const bool add_bos = llama_add_bos_token(model); - GGML_ASSERT(!llama_add_eos_token(model)); + const bool add_bos = llama_vocab_get_add_bos(vocab); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); std::vector embd_inp; std::vector embd_end; std::vector inp_pfx = common_tokenize(ctx, params.input_prefix, false); std::vector inp_sfx = common_tokenize(ctx, params.input_suffix, false); - GGML_ASSERT(llama_token_fim_pre(model) >= 0); - GGML_ASSERT(llama_token_fim_suf(model) >= 0); + GGML_ASSERT(llama_vocab_fim_pre(vocab) >= 0); + GGML_ASSERT(llama_vocab_fim_suf(vocab) >= 0); - inp_pfx.insert(inp_pfx.begin(), llama_token_fim_pre(model)); - inp_sfx.insert(inp_sfx.begin(), llama_token_fim_suf(model)); + inp_pfx.insert(inp_pfx.begin(), llama_vocab_fim_pre(vocab)); + inp_sfx.insert(inp_sfx.begin(), llama_vocab_fim_suf(vocab)); embd_inp = params.spm_infill ? inp_sfx : inp_pfx; embd_end = params.spm_infill ? inp_pfx : inp_sfx; if (add_bos) { - embd_inp.insert(embd_inp.begin(), llama_token_bos(model)); + embd_inp.insert(embd_inp.begin(), llama_vocab_bos(vocab)); } embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end()); - const llama_token middle_token = llama_token_fim_mid(model); + const llama_token middle_token = llama_vocab_fim_mid(vocab); if (middle_token >= 0) { embd_inp.push_back(middle_token); } @@ -230,7 +187,7 @@ int main(int argc, char ** argv) { // Should not run without any tokens if (embd_inp.empty()) { - embd_inp.push_back(llama_token_bos(model)); + embd_inp.push_back(llama_vocab_bos(vocab)); LOG_WRN("embd_inp was considered empty and bos was added: %s\n", string_from(ctx, embd_inp).c_str()); } @@ -465,10 +422,10 @@ int main(int argc, char ** argv) { // if not currently processing queued inputs; if ((int) embd_inp.size() <= n_consumed) { // deal with eot token in infill mode - if ((common_sampler_last(smpl) == llama_token_eot(model) || is_interacting) && params.interactive){ + if ((common_sampler_last(smpl) == llama_vocab_eot(vocab) || is_interacting) && params.interactive){ if (is_interacting && !params.interactive_first) { // print an eot token - LOG("%s", common_token_to_piece(ctx, llama_token_eot(model)).c_str()); + LOG("%s", common_token_to_piece(ctx, llama_vocab_eot(vocab)).c_str()); } LOG("\n"); console::set_display(console::user_input); @@ -508,13 +465,13 @@ int main(int argc, char ** argv) { std::vector inp_pfx = common_tokenize(ctx, params.input_prefix, false); std::vector inp_sfx = common_tokenize(ctx, params.input_suffix, false); - inp_pfx.insert(inp_pfx.begin(), llama_token_fim_pre(model)); - inp_sfx.insert(inp_sfx.begin(), llama_token_fim_suf(model)); + inp_pfx.insert(inp_pfx.begin(), llama_vocab_fim_pre(vocab)); + inp_sfx.insert(inp_sfx.begin(), llama_vocab_fim_suf(vocab)); embd_inp = params.spm_infill ? inp_sfx : inp_pfx; embd_end = params.spm_infill ? inp_pfx : inp_sfx; if (add_bos) { - embd_inp.insert(embd_inp.begin(), llama_token_bos(model)); + embd_inp.insert(embd_inp.begin(), llama_vocab_bos(vocab)); } embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end()); @@ -529,7 +486,7 @@ int main(int argc, char ** argv) { is_interacting = false; } // deal with end of generation tokens in interactive mode - else if (llama_token_is_eog(model, common_sampler_last(smpl))) { + else if (llama_vocab_is_eog(vocab, common_sampler_last(smpl))) { LOG_DBG("found EOS token\n"); if (params.interactive) { @@ -545,7 +502,7 @@ int main(int argc, char ** argv) { if (params.input_prefix_bos) { LOG_DBG("adding input prefix BOS token\n"); - embd_inp.push_back(llama_token_bos(model)); + embd_inp.push_back(llama_vocab_bos(vocab)); } std::string buffer; @@ -608,7 +565,7 @@ int main(int argc, char ** argv) { } // end of generation - if (!embd.empty() && llama_token_is_eog(model, embd.back()) && !params.interactive) { + if (!embd.empty() && llama_vocab_is_eog(vocab, embd.back()) && !params.interactive) { break; } @@ -620,15 +577,11 @@ int main(int argc, char ** argv) { } } if (!params.interactive && n_remain <= 0) { - LOG("%s", common_token_to_piece(ctx, llama_token_eot(model)).c_str()); + LOG("%s", common_token_to_piece(ctx, llama_vocab_eot(vocab)).c_str()); } LOG("\n"); common_perf_print(ctx, smpl); - write_logfile(ctx, params, model, input_tokens, output_ss.str(), output_tokens); - - llama_free(ctx); - llama_free_model(model); common_sampler_free(smpl); llama_backend_free(); diff --git a/examples/llama-bench/CMakeLists.txt b/examples/llama-bench/CMakeLists.txt index 5bdbea4e2..17e3b9b87 100644 --- a/examples/llama-bench/CMakeLists.txt +++ b/examples/llama-bench/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-bench) add_executable(${TARGET} llama-bench.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/llama-bench/llama-bench.cpp b/examples/llama-bench/llama-bench.cpp index e7873a143..a3b4c5ac8 100644 --- a/examples/llama-bench/llama-bench.cpp +++ b/examples/llama-bench/llama-bench.cpp @@ -6,28 +6,28 @@ #include #include #include +#include #include #include -#include #include #include #include #include #include #include -#include #include +#include +#include "common.h" #include "ggml.h" #include "llama.h" -#include "common.h" #ifdef _WIN32 -#define WIN32_LEAN_AND_MEAN -#ifndef NOMINMAX -# define NOMINMAX -#endif -#include +# define WIN32_LEAN_AND_MEAN +# ifndef NOMINMAX +# define NOMINMAX +# endif +# include #endif // utils @@ -36,8 +36,7 @@ static uint64_t get_time_ns() { return std::chrono::nanoseconds(clock::now().time_since_epoch()).count(); } -template -static std::string join(const std::vector & values, const std::string & delim) { +template static std::string join(const std::vector & values, const std::string & delim) { std::ostringstream str; for (size_t i = 0; i < values.size(); i++) { str << values[i]; @@ -48,38 +47,35 @@ static std::string join(const std::vector & values, const std::string & delim return str.str(); } -template -static std::vector transform_to_str(const std::vector & values, F f) { +template static std::vector transform_to_str(const std::vector & values, F f) { std::vector str_values; std::transform(values.begin(), values.end(), std::back_inserter(str_values), f); return str_values; } -template -static T avg(const std::vector & v) { +template static T avg(const std::vector & v) { if (v.empty()) { return 0; } T sum = std::accumulate(v.begin(), v.end(), T(0)); - return sum / (T)v.size(); + return sum / (T) v.size(); } -template -static T stdev(const std::vector & v) { +template static T stdev(const std::vector & v) { if (v.size() <= 1) { return 0; } - T mean = avg(v); + T mean = avg(v); T sq_sum = std::inner_product(v.begin(), v.end(), v.begin(), T(0)); - T stdev = std::sqrt(sq_sum / (T)(v.size() - 1) - mean * mean * (T)v.size() / (T)(v.size() - 1)); + T stdev = std::sqrt(sq_sum / (T) (v.size() - 1) - mean * mean * (T) v.size() / (T) (v.size() - 1)); return stdev; } static std::string get_cpu_info() { std::vector cpu_list; for (size_t i = 0; i < ggml_backend_dev_count(); i++) { - auto * dev = ggml_backend_dev_get(i); - auto dev_type = ggml_backend_dev_type(dev); + auto * dev = ggml_backend_dev_get(i); + auto dev_type = ggml_backend_dev_type(dev); if (dev_type == GGML_BACKEND_DEVICE_TYPE_CPU || dev_type == GGML_BACKEND_DEVICE_TYPE_ACCEL) { cpu_list.push_back(ggml_backend_dev_description(dev)); } @@ -90,8 +86,8 @@ static std::string get_cpu_info() { static std::string get_gpu_info() { std::vector gpu_list; for (size_t i = 0; i < ggml_backend_dev_count(); i++) { - auto * dev = ggml_backend_dev_get(i); - auto dev_type = ggml_backend_dev_type(dev); + auto * dev = ggml_backend_dev_get(i); + auto dev_type = ggml_backend_dev_type(dev); if (dev_type == GGML_BACKEND_DEVICE_TYPE_GPU) { gpu_list.push_back(ggml_backend_dev_description(dev)); } @@ -100,17 +96,24 @@ static std::string get_gpu_info() { } // command line params -enum output_formats {NONE, CSV, JSON, JSONL, MARKDOWN, SQL}; +enum output_formats { NONE, CSV, JSON, JSONL, MARKDOWN, SQL }; static const char * output_format_str(output_formats format) { switch (format) { - case NONE: return "none"; - case CSV: return "csv"; - case JSON: return "json"; - case JSONL: return "jsonl"; - case MARKDOWN: return "md"; - case SQL: return "sql"; - default: GGML_ABORT("invalid output format"); + case NONE: + return "none"; + case CSV: + return "csv"; + case JSON: + return "json"; + case JSONL: + return "jsonl"; + case MARKDOWN: + return "md"; + case SQL: + return "sql"; + default: + GGML_ABORT("invalid output format"); } } @@ -135,10 +138,14 @@ static bool output_format_from_str(const std::string & s, output_formats & forma static const char * split_mode_str(llama_split_mode mode) { switch (mode) { - case LLAMA_SPLIT_MODE_NONE: return "none"; - case LLAMA_SPLIT_MODE_LAYER: return "layer"; - case LLAMA_SPLIT_MODE_ROW: return "row"; - default: GGML_ABORT("invalid split mode"); + case LLAMA_SPLIT_MODE_NONE: + return "none"; + case LLAMA_SPLIT_MODE_LAYER: + return "layer"; + case LLAMA_SPLIT_MODE_ROW: + return "row"; + default: + GGML_ABORT("invalid split mode"); } } @@ -149,59 +156,59 @@ static std::string pair_str(const std::pair & p) { } struct cmd_params { - std::vector model; - std::vector n_prompt; - std::vector n_gen; + std::vector model; + std::vector n_prompt; + std::vector n_gen; std::vector> n_pg; - std::vector n_batch; - std::vector n_ubatch; - std::vector type_k; - std::vector type_v; - std::vector n_threads; - std::vector cpu_mask; - std::vector cpu_strict; - std::vector poll; - std::vector n_gpu_layers; - std::vector rpc_servers; - std::vector split_mode; - std::vector main_gpu; - std::vector no_kv_offload; - std::vector flash_attn; - std::vector> tensor_split; - std::vector use_mmap; - std::vector embeddings; - ggml_numa_strategy numa; - int reps; - ggml_sched_priority prio; - int delay; - bool verbose; - bool progress; - output_formats output_format; - output_formats output_format_stderr; + std::vector n_batch; + std::vector n_ubatch; + std::vector type_k; + std::vector type_v; + std::vector n_threads; + std::vector cpu_mask; + std::vector cpu_strict; + std::vector poll; + std::vector n_gpu_layers; + std::vector rpc_servers; + std::vector split_mode; + std::vector main_gpu; + std::vector no_kv_offload; + std::vector flash_attn; + std::vector> tensor_split; + std::vector use_mmap; + std::vector embeddings; + ggml_numa_strategy numa; + int reps; + ggml_sched_priority prio; + int delay; + bool verbose; + bool progress; + output_formats output_format; + output_formats output_format_stderr; }; static const cmd_params cmd_params_defaults = { - /* model */ {"models/7B/ggml-model-q4_0.gguf"}, - /* n_prompt */ {512}, - /* n_gen */ {128}, + /* model */ { "models/7B/ggml-model-q4_0.gguf" }, + /* n_prompt */ { 512 }, + /* n_gen */ { 128 }, /* n_pg */ {}, - /* n_batch */ {2048}, - /* n_ubatch */ {512}, - /* type_k */ {GGML_TYPE_F16}, - /* type_v */ {GGML_TYPE_F16}, - /* n_threads */ {cpu_get_num_math()}, - /* cpu_mask */ {"0x0"}, - /* cpu_strict */ {false}, - /* poll */ {50}, - /* n_gpu_layers */ {99}, - /* rpc_servers */ {""}, - /* split_mode */ {LLAMA_SPLIT_MODE_LAYER}, - /* main_gpu */ {0}, - /* no_kv_offload */ {false}, - /* flash_attn */ {false}, - /* tensor_split */ {std::vector(llama_max_devices(), 0.0f)}, - /* use_mmap */ {true}, - /* embeddings */ {false}, + /* n_batch */ { 2048 }, + /* n_ubatch */ { 512 }, + /* type_k */ { GGML_TYPE_F16 }, + /* type_v */ { GGML_TYPE_F16 }, + /* n_threads */ { cpu_get_num_math() }, + /* cpu_mask */ { "0x0" }, + /* cpu_strict */ { false }, + /* poll */ { 50 }, + /* n_gpu_layers */ { 99 }, + /* rpc_servers */ { "" }, + /* split_mode */ { LLAMA_SPLIT_MODE_LAYER }, + /* main_gpu */ { 0 }, + /* no_kv_offload */ { false }, + /* flash_attn */ { false }, + /* tensor_split */ { std::vector(llama_max_devices(), 0.0f) }, + /* use_mmap */ { true }, + /* embeddings */ { false }, /* numa */ GGML_NUMA_STRATEGY_DISABLED, /* reps */ 5, /* prio */ GGML_SCHED_PRIO_NORMAL, @@ -218,44 +225,68 @@ static void print_usage(int /* argc */, char ** argv) { printf("options:\n"); printf(" -h, --help\n"); printf(" -m, --model (default: %s)\n", join(cmd_params_defaults.model, ",").c_str()); - printf(" -p, --n-prompt (default: %s)\n", join(cmd_params_defaults.n_prompt, ",").c_str()); + printf(" -p, --n-prompt (default: %s)\n", + join(cmd_params_defaults.n_prompt, ",").c_str()); printf(" -n, --n-gen (default: %s)\n", join(cmd_params_defaults.n_gen, ",").c_str()); - printf(" -pg (default: %s)\n", join(transform_to_str(cmd_params_defaults.n_pg, pair_str), ",").c_str()); - printf(" -b, --batch-size (default: %s)\n", join(cmd_params_defaults.n_batch, ",").c_str()); - printf(" -ub, --ubatch-size (default: %s)\n", join(cmd_params_defaults.n_ubatch, ",").c_str()); - printf(" -ctk, --cache-type-k (default: %s)\n", join(transform_to_str(cmd_params_defaults.type_k, ggml_type_name), ",").c_str()); - printf(" -ctv, --cache-type-v (default: %s)\n", join(transform_to_str(cmd_params_defaults.type_v, ggml_type_name), ",").c_str()); - printf(" -t, --threads (default: %s)\n", join(cmd_params_defaults.n_threads, ",").c_str()); - printf(" -C, --cpu-mask (default: %s)\n", join(cmd_params_defaults.cpu_mask, ",").c_str()); - printf(" --cpu-strict <0|1> (default: %s)\n", join(cmd_params_defaults.cpu_strict, ",").c_str()); + printf(" -pg (default: %s)\n", + join(transform_to_str(cmd_params_defaults.n_pg, pair_str), ",").c_str()); + printf(" -b, --batch-size (default: %s)\n", + join(cmd_params_defaults.n_batch, ",").c_str()); + printf(" -ub, --ubatch-size (default: %s)\n", + join(cmd_params_defaults.n_ubatch, ",").c_str()); + printf(" -ctk, --cache-type-k (default: %s)\n", + join(transform_to_str(cmd_params_defaults.type_k, ggml_type_name), ",").c_str()); + printf(" -ctv, --cache-type-v (default: %s)\n", + join(transform_to_str(cmd_params_defaults.type_v, ggml_type_name), ",").c_str()); + printf(" -t, --threads (default: %s)\n", + join(cmd_params_defaults.n_threads, ",").c_str()); + printf(" -C, --cpu-mask (default: %s)\n", + join(cmd_params_defaults.cpu_mask, ",").c_str()); + printf(" --cpu-strict <0|1> (default: %s)\n", + join(cmd_params_defaults.cpu_strict, ",").c_str()); printf(" --poll <0...100> (default: %s)\n", join(cmd_params_defaults.poll, ",").c_str()); - printf(" -ngl, --n-gpu-layers (default: %s)\n", join(cmd_params_defaults.n_gpu_layers, ",").c_str()); + printf(" -ngl, --n-gpu-layers (default: %s)\n", + join(cmd_params_defaults.n_gpu_layers, ",").c_str()); if (llama_supports_rpc()) { - printf(" -rpc, --rpc (default: %s)\n", join(cmd_params_defaults.rpc_servers, ",").c_str()); + printf(" -rpc, --rpc (default: %s)\n", + join(cmd_params_defaults.rpc_servers, ",").c_str()); } - printf(" -sm, --split-mode (default: %s)\n", join(transform_to_str(cmd_params_defaults.split_mode, split_mode_str), ",").c_str()); - printf(" -mg, --main-gpu (default: %s)\n", join(cmd_params_defaults.main_gpu, ",").c_str()); - printf(" -nkvo, --no-kv-offload <0|1> (default: %s)\n", join(cmd_params_defaults.no_kv_offload, ",").c_str()); - printf(" -fa, --flash-attn <0|1> (default: %s)\n", join(cmd_params_defaults.flash_attn, ",").c_str()); - printf(" -mmp, --mmap <0|1> (default: %s)\n", join(cmd_params_defaults.use_mmap, ",").c_str()); + printf(" -sm, --split-mode (default: %s)\n", + join(transform_to_str(cmd_params_defaults.split_mode, split_mode_str), ",").c_str()); + printf(" -mg, --main-gpu (default: %s)\n", + join(cmd_params_defaults.main_gpu, ",").c_str()); + printf(" -nkvo, --no-kv-offload <0|1> (default: %s)\n", + join(cmd_params_defaults.no_kv_offload, ",").c_str()); + printf(" -fa, --flash-attn <0|1> (default: %s)\n", + join(cmd_params_defaults.flash_attn, ",").c_str()); + printf(" -mmp, --mmap <0|1> (default: %s)\n", + join(cmd_params_defaults.use_mmap, ",").c_str()); printf(" --numa (default: disabled)\n"); - printf(" -embd, --embeddings <0|1> (default: %s)\n", join(cmd_params_defaults.embeddings, ",").c_str()); + printf(" -embd, --embeddings <0|1> (default: %s)\n", + join(cmd_params_defaults.embeddings, ",").c_str()); printf(" -ts, --tensor-split (default: 0)\n"); printf(" -r, --repetitions (default: %d)\n", cmd_params_defaults.reps); printf(" --prio <0|1|2|3> (default: %d)\n", cmd_params_defaults.prio); printf(" --delay <0...N> (seconds) (default: %d)\n", cmd_params_defaults.delay); - printf(" -o, --output (default: %s)\n", output_format_str(cmd_params_defaults.output_format)); - printf(" -oe, --output-err (default: %s)\n", output_format_str(cmd_params_defaults.output_format_stderr)); + printf(" -o, --output (default: %s)\n", + output_format_str(cmd_params_defaults.output_format)); + printf(" -oe, --output-err (default: %s)\n", + output_format_str(cmd_params_defaults.output_format_stderr)); printf(" -v, --verbose (default: %s)\n", cmd_params_defaults.verbose ? "1" : "0"); printf(" --progress (default: %s)\n", cmd_params_defaults.progress ? "1" : "0"); printf("\n"); - printf("Multiple values can be given for each parameter by separating them with ',' or by specifying the parameter multiple times.\n"); + printf( + "Multiple values can be given for each parameter by separating them with ',' or by specifying the parameter " + "multiple times.\n"); } static ggml_type ggml_type_from_name(const std::string & s) { if (s == "f16") { return GGML_TYPE_F16; } + if (s == "bf16") { + return GGML_TYPE_BF16; + } if (s == "q8_0") { return GGML_TYPE_Q8_0; } @@ -278,22 +309,21 @@ static ggml_type ggml_type_from_name(const std::string & s) { return GGML_TYPE_COUNT; } - static cmd_params parse_cmd_params(int argc, char ** argv) { - cmd_params params; - std::string arg; - bool invalid_param = false; - const std::string arg_prefix = "--"; - const char split_delim = ','; + cmd_params params; + std::string arg; + bool invalid_param = false; + const std::string arg_prefix = "--"; + const char split_delim = ','; - params.verbose = cmd_params_defaults.verbose; - params.output_format = cmd_params_defaults.output_format; + params.verbose = cmd_params_defaults.verbose; + params.output_format = cmd_params_defaults.output_format; params.output_format_stderr = cmd_params_defaults.output_format_stderr; - params.reps = cmd_params_defaults.reps; - params.numa = cmd_params_defaults.numa; - params.prio = cmd_params_defaults.prio; - params.delay = cmd_params_defaults.delay; - params.progress = cmd_params_defaults.progress; + params.reps = cmd_params_defaults.reps; + params.numa = cmd_params_defaults.numa; + params.prio = cmd_params_defaults.prio; + params.delay = cmd_params_defaults.delay; + params.progress = cmd_params_defaults.progress; for (int i = 1; i < argc; i++) { arg = argv[i]; @@ -335,7 +365,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { invalid_param = true; break; } - params.n_pg.push_back({std::stoi(p[0]), std::stoi(p[1])}); + params.n_pg.push_back({ std::stoi(p[0]), std::stoi(p[1]) }); } else if (arg == "-b" || arg == "--batch-size") { if (++i >= argc) { invalid_param = true; @@ -355,7 +385,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { invalid_param = true; break; } - auto p = string_split(argv[i], split_delim); + auto p = string_split(argv[i], split_delim); std::vector types; for (const auto & t : p) { ggml_type gt = ggml_type_from_name(t); @@ -374,7 +404,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { invalid_param = true; break; } - auto p = string_split(argv[i], split_delim); + auto p = string_split(argv[i], split_delim); std::vector types; for (const auto & t : p) { ggml_type gt = ggml_type_from_name(t); @@ -434,7 +464,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { invalid_param = true; break; } - auto p = string_split(argv[i], split_delim); + auto p = string_split(argv[i], split_delim); std::vector modes; for (const auto & m : p) { llama_split_mode mode; @@ -473,10 +503,16 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { break; } else { std::string value(argv[i]); - /**/ if (value == "distribute" || value == "" ) { params.numa = GGML_NUMA_STRATEGY_DISTRIBUTE; } - else if (value == "isolate") { params.numa = GGML_NUMA_STRATEGY_ISOLATE; } - else if (value == "numactl") { params.numa = GGML_NUMA_STRATEGY_NUMACTL; } - else { invalid_param = true; break; } + /**/ if (value == "distribute" || value == "") { + params.numa = GGML_NUMA_STRATEGY_DISTRIBUTE; + } else if (value == "isolate") { + params.numa = GGML_NUMA_STRATEGY_ISOLATE; + } else if (value == "numactl") { + params.numa = GGML_NUMA_STRATEGY_NUMACTL; + } else { + invalid_param = true; + break; + } } } else if (arg == "-fa" || arg == "--flash-attn") { if (++i >= argc) { @@ -506,9 +542,9 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { } for (auto ts : string_split(argv[i], split_delim)) { // split string by ; and / - const std::regex regex{R"([;/]+)"}; - std::sregex_token_iterator it{ts.begin(), ts.end(), regex, -1}; - std::vector split_arg{it, {}}; + const std::regex regex{ R"([;/]+)" }; + std::sregex_token_iterator it{ ts.begin(), ts.end(), regex, -1 }; + std::vector split_arg{ it, {} }; GGML_ASSERT(split_arg.size() <= llama_max_devices()); std::vector tensor_split(llama_max_devices()); @@ -567,52 +603,94 @@ static cmd_params parse_cmd_params(int argc, char ** argv) { } // set defaults - if (params.model.empty()) { params.model = cmd_params_defaults.model; } - if (params.n_prompt.empty()) { params.n_prompt = cmd_params_defaults.n_prompt; } - if (params.n_gen.empty()) { params.n_gen = cmd_params_defaults.n_gen; } - if (params.n_pg.empty()) { params.n_pg = cmd_params_defaults.n_pg; } - if (params.n_batch.empty()) { params.n_batch = cmd_params_defaults.n_batch; } - if (params.n_ubatch.empty()) { params.n_ubatch = cmd_params_defaults.n_ubatch; } - if (params.type_k.empty()) { params.type_k = cmd_params_defaults.type_k; } - if (params.type_v.empty()) { params.type_v = cmd_params_defaults.type_v; } - if (params.n_gpu_layers.empty()) { params.n_gpu_layers = cmd_params_defaults.n_gpu_layers; } - if (params.rpc_servers.empty()) { params.rpc_servers = cmd_params_defaults.rpc_servers; } - if (params.split_mode.empty()) { params.split_mode = cmd_params_defaults.split_mode; } - if (params.main_gpu.empty()) { params.main_gpu = cmd_params_defaults.main_gpu; } - if (params.no_kv_offload.empty()){ params.no_kv_offload = cmd_params_defaults.no_kv_offload; } - if (params.flash_attn.empty()) { params.flash_attn = cmd_params_defaults.flash_attn; } - if (params.tensor_split.empty()) { params.tensor_split = cmd_params_defaults.tensor_split; } - if (params.use_mmap.empty()) { params.use_mmap = cmd_params_defaults.use_mmap; } - if (params.embeddings.empty()) { params.embeddings = cmd_params_defaults.embeddings; } - if (params.n_threads.empty()) { params.n_threads = cmd_params_defaults.n_threads; } - if (params.cpu_mask.empty()) { params.cpu_mask = cmd_params_defaults.cpu_mask; } - if (params.cpu_strict.empty()) { params.cpu_strict = cmd_params_defaults.cpu_strict; } - if (params.poll.empty()) { params.poll = cmd_params_defaults.poll; } + if (params.model.empty()) { + params.model = cmd_params_defaults.model; + } + if (params.n_prompt.empty()) { + params.n_prompt = cmd_params_defaults.n_prompt; + } + if (params.n_gen.empty()) { + params.n_gen = cmd_params_defaults.n_gen; + } + if (params.n_pg.empty()) { + params.n_pg = cmd_params_defaults.n_pg; + } + if (params.n_batch.empty()) { + params.n_batch = cmd_params_defaults.n_batch; + } + if (params.n_ubatch.empty()) { + params.n_ubatch = cmd_params_defaults.n_ubatch; + } + if (params.type_k.empty()) { + params.type_k = cmd_params_defaults.type_k; + } + if (params.type_v.empty()) { + params.type_v = cmd_params_defaults.type_v; + } + if (params.n_gpu_layers.empty()) { + params.n_gpu_layers = cmd_params_defaults.n_gpu_layers; + } + if (params.rpc_servers.empty()) { + params.rpc_servers = cmd_params_defaults.rpc_servers; + } + if (params.split_mode.empty()) { + params.split_mode = cmd_params_defaults.split_mode; + } + if (params.main_gpu.empty()) { + params.main_gpu = cmd_params_defaults.main_gpu; + } + if (params.no_kv_offload.empty()) { + params.no_kv_offload = cmd_params_defaults.no_kv_offload; + } + if (params.flash_attn.empty()) { + params.flash_attn = cmd_params_defaults.flash_attn; + } + if (params.tensor_split.empty()) { + params.tensor_split = cmd_params_defaults.tensor_split; + } + if (params.use_mmap.empty()) { + params.use_mmap = cmd_params_defaults.use_mmap; + } + if (params.embeddings.empty()) { + params.embeddings = cmd_params_defaults.embeddings; + } + if (params.n_threads.empty()) { + params.n_threads = cmd_params_defaults.n_threads; + } + if (params.cpu_mask.empty()) { + params.cpu_mask = cmd_params_defaults.cpu_mask; + } + if (params.cpu_strict.empty()) { + params.cpu_strict = cmd_params_defaults.cpu_strict; + } + if (params.poll.empty()) { + params.poll = cmd_params_defaults.poll; + } return params; } struct cmd_params_instance { - std::string model; - int n_prompt; - int n_gen; - int n_batch; - int n_ubatch; - ggml_type type_k; - ggml_type type_v; - int n_threads; - std::string cpu_mask; - bool cpu_strict; - int poll; - int n_gpu_layers; - std::string rpc_servers; - llama_split_mode split_mode; - int main_gpu; - bool no_kv_offload; - bool flash_attn; + std::string model; + int n_prompt; + int n_gen; + int n_batch; + int n_ubatch; + ggml_type type_k; + ggml_type type_v; + int n_threads; + std::string cpu_mask; + bool cpu_strict; + int poll; + int n_gpu_layers; + std::string rpc_servers; + llama_split_mode split_mode; + int main_gpu; + bool no_kv_offload; + bool flash_attn; std::vector tensor_split; - bool use_mmap; - bool embeddings; + bool use_mmap; + bool embeddings; llama_model_params to_llama_mparams() const { llama_model_params mparams = llama_model_default_params(); @@ -621,35 +699,31 @@ struct cmd_params_instance { if (!rpc_servers.empty()) { mparams.rpc_servers = rpc_servers.c_str(); } - mparams.split_mode = split_mode; - mparams.main_gpu = main_gpu; + mparams.split_mode = split_mode; + mparams.main_gpu = main_gpu; mparams.tensor_split = tensor_split.data(); - mparams.use_mmap = use_mmap; + mparams.use_mmap = use_mmap; return mparams; } bool equal_mparams(const cmd_params_instance & other) const { - return model == other.model && - n_gpu_layers == other.n_gpu_layers && - rpc_servers == other.rpc_servers && - split_mode == other.split_mode && - main_gpu == other.main_gpu && - use_mmap == other.use_mmap && + return model == other.model && n_gpu_layers == other.n_gpu_layers && rpc_servers == other.rpc_servers && + split_mode == other.split_mode && main_gpu == other.main_gpu && use_mmap == other.use_mmap && tensor_split == other.tensor_split; } llama_context_params to_llama_cparams() const { llama_context_params cparams = llama_context_default_params(); - cparams.n_ctx = n_prompt + n_gen; - cparams.n_batch = n_batch; - cparams.n_ubatch = n_ubatch; - cparams.type_k = type_k; - cparams.type_v = type_v; + cparams.n_ctx = n_prompt + n_gen; + cparams.n_batch = n_batch; + cparams.n_ubatch = n_ubatch; + cparams.type_k = type_k; + cparams.type_v = type_v; cparams.offload_kqv = !no_kv_offload; - cparams.flash_attn = flash_attn; - cparams.embeddings = embeddings; + cparams.flash_attn = flash_attn; + cparams.embeddings = embeddings; return cparams; } @@ -659,6 +733,7 @@ static std::vector get_cmd_params_instances(const cmd_param std::vector instances; // this ordering minimizes the number of times that each model needs to be reloaded + // clang-format off for (const auto & m : params.model) for (const auto & nl : params.n_gpu_layers) for (const auto & rpc : params.rpc_servers) @@ -764,109 +839,94 @@ static std::vector get_cmd_params_instances(const cmd_param instances.push_back(instance); } } + // clang-format on return instances; } struct test { static const std::string build_commit; - static const int build_number; - static const bool cuda; - static const bool vulkan; - static const bool kompute; - static const bool metal; - static const bool sycl; - static const bool gpu_blas; - static const bool blas; + static const int build_number; static const std::string cpu_info; static const std::string gpu_info; - std::string model_filename; - std::string model_type; - uint64_t model_size; - uint64_t model_n_params; - int n_batch; - int n_ubatch; - int n_threads; - std::string cpu_mask; - bool cpu_strict; - int poll; - bool has_rpc; - ggml_type type_k; - ggml_type type_v; - int n_gpu_layers; - llama_split_mode split_mode; - int main_gpu; - bool no_kv_offload; - bool flash_attn; - std::vector tensor_split; - bool use_mmap; - bool embeddings; - int n_prompt; - int n_gen; - std::string test_time; - std::vector samples_ns; + std::string model_filename; + std::string model_type; + uint64_t model_size; + uint64_t model_n_params; + int n_batch; + int n_ubatch; + int n_threads; + std::string cpu_mask; + bool cpu_strict; + int poll; + ggml_type type_k; + ggml_type type_v; + int n_gpu_layers; + llama_split_mode split_mode; + int main_gpu; + bool no_kv_offload; + bool flash_attn; + std::vector tensor_split; + bool use_mmap; + bool embeddings; + int n_prompt; + int n_gen; + std::string test_time; + std::vector samples_ns; test(const cmd_params_instance & inst, const llama_model * lmodel, const llama_context * ctx) { model_filename = inst.model; char buf[128]; llama_model_desc(lmodel, buf, sizeof(buf)); - model_type = buf; - model_size = llama_model_size(lmodel); + model_type = buf; + model_size = llama_model_size(lmodel); model_n_params = llama_model_n_params(lmodel); - n_batch = inst.n_batch; - n_ubatch = inst.n_ubatch; - n_threads = inst.n_threads; - cpu_mask = inst.cpu_mask; - cpu_strict = inst.cpu_strict; - poll = inst.poll; - has_rpc = !inst.rpc_servers.empty(); - type_k = inst.type_k; - type_v = inst.type_v; - n_gpu_layers = inst.n_gpu_layers; - split_mode = inst.split_mode; - main_gpu = inst.main_gpu; - no_kv_offload = inst.no_kv_offload; - flash_attn = inst.flash_attn; - tensor_split = inst.tensor_split; - use_mmap = inst.use_mmap; - embeddings = inst.embeddings; - n_prompt = inst.n_prompt; - n_gen = inst.n_gen; + n_batch = inst.n_batch; + n_ubatch = inst.n_ubatch; + n_threads = inst.n_threads; + cpu_mask = inst.cpu_mask; + cpu_strict = inst.cpu_strict; + poll = inst.poll; + type_k = inst.type_k; + type_v = inst.type_v; + n_gpu_layers = inst.n_gpu_layers; + split_mode = inst.split_mode; + main_gpu = inst.main_gpu; + no_kv_offload = inst.no_kv_offload; + flash_attn = inst.flash_attn; + tensor_split = inst.tensor_split; + use_mmap = inst.use_mmap; + embeddings = inst.embeddings; + n_prompt = inst.n_prompt; + n_gen = inst.n_gen; // RFC 3339 date-time format - time_t t = time(NULL); + time_t t = time(NULL); std::strftime(buf, sizeof(buf), "%FT%TZ", gmtime(&t)); test_time = buf; (void) ctx; } - uint64_t avg_ns() const { - return ::avg(samples_ns); - } + uint64_t avg_ns() const { return ::avg(samples_ns); } - uint64_t stdev_ns() const { - return ::stdev(samples_ns); - } + uint64_t stdev_ns() const { return ::stdev(samples_ns); } std::vector get_ts() const { - int n_tokens = n_prompt + n_gen; + int n_tokens = n_prompt + n_gen; std::vector ts; - std::transform(samples_ns.begin(), samples_ns.end(), std::back_inserter(ts), [n_tokens](uint64_t t) { return 1e9 * n_tokens / t; }); + std::transform(samples_ns.begin(), samples_ns.end(), std::back_inserter(ts), + [n_tokens](uint64_t t) { return 1e9 * n_tokens / t; }); return ts; } - double avg_ts() const { - return ::avg(get_ts()); - } + double avg_ts() const { return ::avg(get_ts()); } - double stdev_ts() const { - return ::stdev(get_ts()); - } + double stdev_ts() const { return ::stdev(get_ts()); } static std::string get_backend() { std::vector backends; for (size_t i = 0; i < ggml_backend_reg_count(); i++) { - auto * reg = ggml_backend_reg_get(i); + auto * reg = ggml_backend_reg_get(i); std::string name = ggml_backend_reg_name(reg); if (name != "CPU") { backends.push_back(ggml_backend_reg_name(reg)); @@ -877,38 +937,27 @@ struct test { static const std::vector & get_fields() { static const std::vector fields = { - "build_commit", "build_number", - "cuda", "vulkan", "kompute", "metal", "sycl", "rpc", "gpu_blas", "blas", - "cpu_info", "gpu_info", - "model_filename", "model_type", "model_size", "model_n_params", - "n_batch", "n_ubatch", - "n_threads", "cpu_mask", "cpu_strict", "poll", - "type_k", "type_v", - "n_gpu_layers", "split_mode", - "main_gpu", "no_kv_offload", "flash_attn", - "tensor_split", "use_mmap", "embeddings", - "n_prompt", "n_gen", "test_time", - "avg_ns", "stddev_ns", - "avg_ts", "stddev_ts", + "build_commit", "build_number", "cpu_info", "gpu_info", "backends", "model_filename", + "model_type", "model_size", "model_n_params", "n_batch", "n_ubatch", "n_threads", + "cpu_mask", "cpu_strict", "poll", "type_k", "type_v", "n_gpu_layers", + "split_mode", "main_gpu", "no_kv_offload", "flash_attn", "tensor_split", "use_mmap", + "embeddings", "n_prompt", "n_gen", "test_time", "avg_ns", "stddev_ns", + "avg_ts", "stddev_ts", }; return fields; } - enum field_type {STRING, BOOL, INT, FLOAT}; + enum field_type { STRING, BOOL, INT, FLOAT }; static field_type get_field_type(const std::string & field) { - if (field == "build_number" || field == "n_batch" || field == "n_ubatch" || - field == "n_threads" || field == "poll" || - field == "model_size" || field == "model_n_params" || - field == "n_gpu_layers" || field == "main_gpu" || - field == "n_prompt" || field == "n_gen" || - field == "avg_ns" || field == "stddev_ns") { + if (field == "build_number" || field == "n_batch" || field == "n_ubatch" || field == "n_threads" || + field == "poll" || field == "model_size" || field == "model_n_params" || field == "n_gpu_layers" || + field == "main_gpu" || field == "n_prompt" || field == "n_gen" || field == "avg_ns" || + field == "stddev_ns") { return INT; } - if (field == "cuda" || field == "vulkan" || field == "kompute" || field == "metal" || - field == "gpu_blas" || field == "blas" || field == "sycl" ||field == "f16_kv" || field == "no_kv_offload" || - field == "cpu_strict" || - field == "flash_attn" || field == "use_mmap" || field == "embeddings") { + if (field == "f16_kv" || field == "no_kv_offload" || field == "cpu_strict" || field == "flash_attn" || + field == "use_mmap" || field == "embeddings") { return BOOL; } if (field == "avg_ts" || field == "stddev_ts") { @@ -919,7 +968,7 @@ struct test { std::vector get_values() const { std::string tensor_split_str; - int max_nonzero = 0; + int max_nonzero = 0; for (size_t i = 0; i < llama_max_devices(); i++) { if (tensor_split[i] > 0) { max_nonzero = i; @@ -933,44 +982,53 @@ struct test { tensor_split_str += "/"; } } - std::vector values = { - build_commit, std::to_string(build_number), - std::to_string(cuda), std::to_string(vulkan), std::to_string(vulkan), - std::to_string(metal), std::to_string(sycl), std::to_string(has_rpc), std::to_string(gpu_blas), std::to_string(blas), - cpu_info, gpu_info, - model_filename, model_type, std::to_string(model_size), std::to_string(model_n_params), - std::to_string(n_batch), std::to_string(n_ubatch), - std::to_string(n_threads), cpu_mask, std::to_string(cpu_strict), std::to_string(poll), - ggml_type_name(type_k), ggml_type_name(type_v), - std::to_string(n_gpu_layers), split_mode_str(split_mode), - std::to_string(main_gpu), std::to_string(no_kv_offload), std::to_string(flash_attn), - tensor_split_str, std::to_string(use_mmap), std::to_string(embeddings), - std::to_string(n_prompt), std::to_string(n_gen), test_time, - std::to_string(avg_ns()), std::to_string(stdev_ns()), - std::to_string(avg_ts()), std::to_string(stdev_ts()) - }; + std::vector values = { build_commit, + std::to_string(build_number), + cpu_info, + gpu_info, + get_backend(), + model_filename, + model_type, + std::to_string(model_size), + std::to_string(model_n_params), + std::to_string(n_batch), + std::to_string(n_ubatch), + std::to_string(n_threads), + cpu_mask, + std::to_string(cpu_strict), + std::to_string(poll), + ggml_type_name(type_k), + ggml_type_name(type_v), + std::to_string(n_gpu_layers), + split_mode_str(split_mode), + std::to_string(main_gpu), + std::to_string(no_kv_offload), + std::to_string(flash_attn), + tensor_split_str, + std::to_string(use_mmap), + std::to_string(embeddings), + std::to_string(n_prompt), + std::to_string(n_gen), + test_time, + std::to_string(avg_ns()), + std::to_string(stdev_ns()), + std::to_string(avg_ts()), + std::to_string(stdev_ts()) }; return values; } std::map get_map() const { std::map map; - auto fields = get_fields(); - auto values = get_values(); - std::transform(fields.begin(), fields.end(), values.begin(), - std::inserter(map, map.end()), std::make_pair); + auto fields = get_fields(); + auto values = get_values(); + std::transform(fields.begin(), fields.end(), values.begin(), std::inserter(map, map.end()), + std::make_pair); return map; } }; const std::string test::build_commit = LLAMA_COMMIT; const int test::build_number = LLAMA_BUILD_NUMBER; -const bool test::cuda = !!ggml_cpu_has_cuda(); -const bool test::vulkan = !!ggml_cpu_has_vulkan(); -const bool test::kompute = !!ggml_cpu_has_kompute(); -const bool test::metal = !!ggml_cpu_has_metal(); -const bool test::gpu_blas = !!ggml_cpu_has_gpublas(); -const bool test::blas = !!ggml_cpu_has_blas(); -const bool test::sycl = !!ggml_cpu_has_sycl(); const std::string test::cpu_info = get_cpu_info(); const std::string test::gpu_info = get_gpu_info(); @@ -978,9 +1036,12 @@ struct printer { virtual ~printer() {} FILE * fout; + virtual void print_header(const cmd_params & params) { (void) params; } + virtual void print_test(const test & t) = 0; - virtual void print_footer() { } + + virtual void print_footer() {} }; struct csv_printer : public printer { @@ -996,7 +1057,7 @@ struct csv_printer : public printer { return escaped; } - void print_header(const cmd_params & params) override { + void print_header(const cmd_params & params) override { std::vector fields = test::get_fields(); fprintf(fout, "%s\n", join(fields, ",").c_str()); (void) params; @@ -1009,7 +1070,6 @@ struct csv_printer : public printer { } }; - static std::string escape_json(const std::string & value) { std::string escaped; for (auto c : value) { @@ -1017,7 +1077,7 @@ static std::string escape_json(const std::string & value) { escaped += "\\\""; } else if (c == '\\') { escaped += "\\\\"; - } else if (c <= 0x1f) { + } else if (c <= 0x1f) { char buf[8]; snprintf(buf, sizeof(buf), "\\u%04x", c); escaped += buf; @@ -1050,7 +1110,8 @@ struct json_printer : public printer { void print_fields(const std::vector & fields, const std::vector & values) { assert(fields.size() == values.size()); for (size_t i = 0; i < fields.size(); i++) { - fprintf(fout, " \"%s\": %s,\n", fields.at(i).c_str(), format_json_value(fields.at(i), values.at(i)).c_str()); + fprintf(fout, " \"%s\": %s,\n", fields.at(i).c_str(), + format_json_value(fields.at(i), values.at(i)).c_str()); } } @@ -1068,12 +1129,9 @@ struct json_printer : public printer { fflush(fout); } - void print_footer() override { - fprintf(fout, "\n]\n"); - } + void print_footer() override { fprintf(fout, "\n]\n"); } }; - struct jsonl_printer : public printer { void print_fields(const std::vector & fields, const std::vector & values) { assert(fields.size() == values.size()); @@ -1133,7 +1191,7 @@ struct markdown_printer : public printer { return 13; } - int width = std::max((int)field.length(), 10); + int width = std::max((int) field.length(), 10); if (test::get_field_type(field) == test::STRING) { return -width; @@ -1175,7 +1233,8 @@ struct markdown_printer : public printer { fields.emplace_back("size"); fields.emplace_back("params"); fields.emplace_back("backend"); - bool is_cpu_backend = test::get_backend() == "CPU" || test::get_backend() == "BLAS"; + bool is_cpu_backend = test::get_backend().find("CPU") != std::string::npos || + test::get_backend().find("BLAS") != std::string::npos; if (!is_cpu_backend) { fields.emplace_back("n_gpu_layers"); } @@ -1246,18 +1305,18 @@ struct markdown_printer : public printer { fprintf(fout, "|"); for (const auto & field : fields) { std::string value; - char buf[128]; + char buf[128]; if (field == "model") { value = t.model_type; } else if (field == "size") { - if (t.model_size < 1024*1024*1024) { + if (t.model_size < 1024 * 1024 * 1024) { snprintf(buf, sizeof(buf), "%.2f MiB", t.model_size / 1024.0 / 1024.0); } else { snprintf(buf, sizeof(buf), "%.2f GiB", t.model_size / 1024.0 / 1024.0 / 1024.0); } value = buf; } else if (field == "params") { - if (t.model_n_params < 1000*1000*1000) { + if (t.model_n_params < 1000 * 1000 * 1000) { snprintf(buf, sizeof(buf), "%.2f M", t.model_n_params / 1e6); } else { snprintf(buf, sizeof(buf), "%.2f B", t.model_n_params / 1e9); @@ -1265,9 +1324,6 @@ struct markdown_printer : public printer { value = buf; } else if (field == "backend") { value = test::get_backend(); - if (t.has_rpc) { - value += "+RPC"; - } } else if (field == "test") { if (t.n_prompt > 0 && t.n_gen == 0) { snprintf(buf, sizeof(buf), "pp%d", t.n_prompt); @@ -1322,7 +1378,8 @@ struct sql_printer : public printer { std::vector fields = test::get_fields(); fprintf(fout, "CREATE TABLE IF NOT EXISTS test (\n"); for (size_t i = 0; i < fields.size(); i++) { - fprintf(fout, " %s %s%s\n", fields.at(i).c_str(), get_sql_field_type(fields.at(i)).c_str(), i < fields.size() - 1 ? "," : ""); + fprintf(fout, " %s %s%s\n", fields.at(i).c_str(), get_sql_field_type(fields.at(i)).c_str(), + i < fields.size() - 1 ? "," : ""); } fprintf(fout, ");\n"); fprintf(fout, "\n"); @@ -1343,8 +1400,9 @@ struct sql_printer : public printer { static void test_prompt(llama_context * ctx, int n_prompt, int n_batch, int n_threads) { llama_set_n_threads(ctx, n_threads, n_threads); - const llama_model * model = llama_get_model(ctx); - const int32_t n_vocab = llama_n_vocab(model); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + const int32_t n_vocab = llama_vocab_n_tokens(vocab); std::vector tokens(n_batch); @@ -1352,7 +1410,7 @@ static void test_prompt(llama_context * ctx, int n_prompt, int n_batch, int n_th while (n_processed < n_prompt) { int n_tokens = std::min(n_prompt - n_processed, n_batch); - tokens[0] = n_processed == 0 && llama_add_bos_token(model) ? llama_token_bos(model) : std::rand() % n_vocab; + tokens[0] = n_processed == 0 && llama_vocab_get_add_bos(vocab) ? llama_vocab_bos(vocab) : std::rand() % n_vocab; for (int i = 1; i < n_tokens; i++) { tokens[i] = std::rand() % n_vocab; } @@ -1366,10 +1424,11 @@ static void test_prompt(llama_context * ctx, int n_prompt, int n_batch, int n_th static void test_gen(llama_context * ctx, int n_gen, int n_threads) { llama_set_n_threads(ctx, n_threads, n_threads); - const llama_model * model = llama_get_model(ctx); - const int32_t n_vocab = llama_n_vocab(model); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + const int32_t n_vocab = llama_vocab_n_tokens(vocab); - llama_token token = llama_add_bos_token(model) ? llama_token_bos(model) : std::rand() % n_vocab; + llama_token token = llama_vocab_get_add_bos(vocab) ? llama_vocab_bos(vocab) : std::rand() % n_vocab; for (int i = 0; i < n_gen; i++) { llama_decode(ctx, llama_batch_get_one(&token, 1)); @@ -1420,6 +1479,17 @@ int main(int argc, char ** argv) { cmd_params params = parse_cmd_params(argc, argv); + // initialize backends + ggml_backend_load_all(); + auto * cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + if (!cpu_dev) { + fprintf(stderr, "%s: error: CPU backend is not loaded\n", __func__); + return 1; + } + auto * cpu_reg = ggml_backend_dev_backend_reg(cpu_dev); + auto * ggml_threadpool_new_fn = (decltype(ggml_threadpool_new) *) ggml_backend_reg_get_proc_address(cpu_reg, "ggml_threadpool_new"); + auto * ggml_threadpool_free_fn = (decltype(ggml_threadpool_free) *) ggml_backend_reg_get_proc_address(cpu_reg, "ggml_threadpool_free"); + // initialize llama.cpp if (!params.verbose) { llama_log_set(llama_null_log_callback, NULL); @@ -1430,7 +1500,7 @@ int main(int argc, char ** argv) { set_process_priority(params.prio); // initialize printer - std::unique_ptr p = create_printer(params.output_format); + std::unique_ptr p = create_printer(params.output_format); std::unique_ptr p_err = create_printer(params.output_format_stderr); if (p) { @@ -1445,23 +1515,23 @@ int main(int argc, char ** argv) { std::vector params_instances = get_cmd_params_instances(params); - llama_model * lmodel = nullptr; + llama_model * lmodel = nullptr; const cmd_params_instance * prev_inst = nullptr; - int params_idx = 0; + int params_idx = 0; auto params_count = params_instances.size(); for (const auto & inst : params_instances) { - params_idx ++; + params_idx++; if (params.progress) { - fprintf(stderr, "llama-bench: benchmark %d/%ld: starting\n", params_idx, params_count); + fprintf(stderr, "llama-bench: benchmark %d/%zu: starting\n", params_idx, params_count); } // keep the same model between tests when possible if (!lmodel || !prev_inst || !inst.equal_mparams(*prev_inst)) { if (lmodel) { - llama_free_model(lmodel); + llama_model_free(lmodel); } - lmodel = llama_load_model_from_file(inst.model.c_str(), inst.to_llama_mparams()); + lmodel = llama_model_load_from_file(inst.model.c_str(), inst.to_llama_mparams()); if (lmodel == NULL) { fprintf(stderr, "%s: error: failed to load model '%s'\n", __func__, inst.model.c_str()); return 1; @@ -1469,10 +1539,10 @@ int main(int argc, char ** argv) { prev_inst = &inst; } - llama_context * ctx = llama_new_context_with_model(lmodel, inst.to_llama_cparams()); + llama_context * ctx = llama_init_from_model(lmodel, inst.to_llama_cparams()); if (ctx == NULL) { fprintf(stderr, "%s: error: failed to create context with model '%s'\n", __func__, inst.model.c_str()); - llama_free_model(lmodel); + llama_model_free(lmodel); return 1; } @@ -1494,7 +1564,7 @@ int main(int argc, char ** argv) { tpp.poll = t.poll; tpp.prio = params.prio; - struct ggml_threadpool* threadpool = ggml_threadpool_new(&tpp); + struct ggml_threadpool * threadpool = ggml_threadpool_new_fn(&tpp); if (!threadpool) { fprintf(stderr, "%s: threadpool create failed : n_threads %d\n", __func__, tpp.n_threads); exit(1); @@ -1505,14 +1575,14 @@ int main(int argc, char ** argv) { // warmup run if (t.n_prompt > 0) { if (params.progress) { - fprintf(stderr, "llama-bench: benchmark %d/%ld: warmup prompt run\n", params_idx, params_count); + fprintf(stderr, "llama-bench: benchmark %d/%zu: warmup prompt run\n", params_idx, params_count); } //test_prompt(ctx, std::min(t.n_batch, std::min(t.n_prompt, 32)), 0, t.n_batch, t.n_threads); test_prompt(ctx, t.n_prompt, t.n_batch, t.n_threads); } if (t.n_gen > 0) { if (params.progress) { - fprintf(stderr, "llama-bench: benchmark %d/%ld: warmup generation run\n", params_idx, params_count); + fprintf(stderr, "llama-bench: benchmark %d/%zu: warmup generation run\n", params_idx, params_count); } test_gen(ctx, 1, t.n_threads); } @@ -1524,13 +1594,15 @@ int main(int argc, char ** argv) { if (t.n_prompt > 0) { if (params.progress) { - fprintf(stderr, "llama-bench: benchmark %d/%ld: prompt run %d/%d\n", params_idx, params_count, i + 1, params.reps); + fprintf(stderr, "llama-bench: benchmark %d/%zu: prompt run %d/%d\n", params_idx, params_count, + i + 1, params.reps); } test_prompt(ctx, t.n_prompt, t.n_batch, t.n_threads); } if (t.n_gen > 0) { if (params.progress) { - fprintf(stderr, "llama-bench: benchmark %d/%ld: generation run %d/%d\n", params_idx, params_count, i + 1, params.reps); + fprintf(stderr, "llama-bench: benchmark %d/%zu: generation run %d/%d\n", params_idx, params_count, + i + 1, params.reps); } test_gen(ctx, t.n_gen, t.n_threads); } @@ -1553,10 +1625,10 @@ int main(int argc, char ** argv) { llama_free(ctx); - ggml_threadpool_free(threadpool); + ggml_threadpool_free_fn(threadpool); } - llama_free_model(lmodel); + llama_model_free(lmodel); if (p) { p->print_footer(); diff --git a/examples/llama.android/llama/build.gradle.kts b/examples/llama.android/llama/build.gradle.kts index 2d1dfba20..28dbc1904 100644 --- a/examples/llama.android/llama/build.gradle.kts +++ b/examples/llama.android/llama/build.gradle.kts @@ -19,6 +19,7 @@ android { externalNativeBuild { cmake { arguments += "-DLLAMA_BUILD_COMMON=ON" + arguments += "-DGGML_LLAMAFILE=OFF" arguments += "-DCMAKE_BUILD_TYPE=Release" cppFlags += listOf() arguments += listOf() diff --git a/examples/llama.android/llama/src/main/cpp/llama-android.cpp b/examples/llama.android/llama/src/main/cpp/llama-android.cpp index b3858ddfb..99b14961d 100644 --- a/examples/llama.android/llama/src/main/cpp/llama-android.cpp +++ b/examples/llama.android/llama/src/main/cpp/llama-android.cpp @@ -87,7 +87,7 @@ Java_android_llama_cpp_LLamaAndroid_load_1model(JNIEnv *env, jobject, jstring fi auto path_to_model = env->GetStringUTFChars(filename, 0); LOGi("Loading model from %s", path_to_model); - auto model = llama_load_model_from_file(path_to_model, model_params); + auto model = llama_model_load_from_file(path_to_model, model_params); env->ReleaseStringUTFChars(filename, path_to_model); if (!model) { @@ -102,7 +102,7 @@ Java_android_llama_cpp_LLamaAndroid_load_1model(JNIEnv *env, jobject, jstring fi extern "C" JNIEXPORT void JNICALL Java_android_llama_cpp_LLamaAndroid_free_1model(JNIEnv *, jobject, jlong model) { - llama_free_model(reinterpret_cast(model)); + llama_model_free(reinterpret_cast(model)); } extern "C" @@ -305,7 +305,9 @@ Java_android_llama_cpp_LLamaAndroid_new_1batch(JNIEnv *, jobject, jint n_tokens, extern "C" JNIEXPORT void JNICALL Java_android_llama_cpp_LLamaAndroid_free_1batch(JNIEnv *, jobject, jlong batch_pointer) { - llama_batch_free(*reinterpret_cast(batch_pointer)); + //llama_batch_free(*reinterpret_cast(batch_pointer)); + const auto batch = reinterpret_cast(batch_pointer); + delete batch; } extern "C" @@ -403,6 +405,7 @@ Java_android_llama_cpp_LLamaAndroid_completion_1loop( const auto batch = reinterpret_cast(batch_pointer); const auto sampler = reinterpret_cast(sampler_pointer); const auto model = llama_get_model(context); + const auto vocab = llama_model_get_vocab(model); if (!la_int_var) la_int_var = env->GetObjectClass(intvar_ncur); if (!la_int_var_value) la_int_var_value = env->GetMethodID(la_int_var, "getValue", "()I"); @@ -412,7 +415,7 @@ Java_android_llama_cpp_LLamaAndroid_completion_1loop( const auto new_token_id = llama_sampler_sample(sampler, context, -1); const auto n_cur = env->CallIntMethod(intvar_ncur, la_int_var_value); - if (llama_token_is_eog(model, new_token_id) || n_cur == n_len) { + if (llama_vocab_is_eog(vocab, new_token_id) || n_cur == n_len) { return nullptr; } diff --git a/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift b/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift index 65cd4eb51..477c3e6f2 100644 --- a/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift +++ b/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift @@ -52,8 +52,8 @@ actor LlamaContext { deinit { llama_sampler_free(sampling) llama_batch_free(batch) + llama_model_free(model) llama_free(context) - llama_free_model(model) llama_backend_free() } @@ -65,7 +65,7 @@ actor LlamaContext { model_params.n_gpu_layers = 0 print("Running on simulator, force use n_gpu_layers = 0") #endif - let model = llama_load_model_from_file(path, model_params) + let model = llama_model_load_from_file(path, model_params) guard let model else { print("Could not load model at \(path)") throw LlamaError.couldNotInitializeContext @@ -151,7 +151,7 @@ actor LlamaContext { new_token_id = llama_sampler_sample(sampling, context, batch.n_tokens - 1) - if llama_token_is_eog(model, new_token_id) || n_cur == n_len { + if llama_vocab_is_eog(model, new_token_id) || n_cur == n_len { print("\n") is_done = true let new_token_str = String(cString: temporary_invalid_cchars + [0]) @@ -210,20 +210,20 @@ actor LlamaContext { llama_kv_cache_clear(context) - let t_pp_start = ggml_time_us() + let t_pp_start = DispatchTime.now().uptimeNanoseconds / 1000; if llama_decode(context, batch) != 0 { print("llama_decode() failed during prompt") } llama_synchronize(context) - let t_pp_end = ggml_time_us() + let t_pp_end = DispatchTime.now().uptimeNanoseconds / 1000; // bench text generation llama_kv_cache_clear(context) - let t_tg_start = ggml_time_us() + let t_tg_start = DispatchTime.now().uptimeNanoseconds / 1000; for i in 0..) if (BUILD_SHARED_LIBS) @@ -35,11 +35,18 @@ add_executable(${TARGET} llava-cli.cpp) set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME llama-llava-cli) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llava ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TARGET llama-minicpmv-cli) add_executable(${TARGET} minicpmv-cli.cpp) set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME llama-minicpmv-cli) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llava ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) + +set(TARGET llama-qwen2vl-cli) +add_executable(${TARGET} qwen2vl-cli.cpp) +set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME llama-qwen2vl-cli) +install(TARGETS ${TARGET} RUNTIME) +target_link_libraries(${TARGET} PRIVATE common llava ${CMAKE_THREAD_LIBS_INIT}) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp index aae49c965..7a8a3156b 100644 --- a/examples/llava/clip.cpp +++ b/examples/llava/clip.cpp @@ -7,22 +7,27 @@ #include "ggml-cpu.h" #include "ggml-alloc.h" #include "ggml-backend.h" +#include "gguf.h" -#ifdef GGML_USE_CUDA -#include "ggml-cuda.h" -#endif - -#ifdef GGML_USE_METAL -#include "ggml-metal.h" -#endif - -#ifdef GGML_USE_CANN -#include "ggml-cann.h" -#endif - -#ifdef GGML_USE_VULKAN -#include "ggml-vulkan.h" -#endif +//#ifdef GGML_USE_CUDA +//#include "ggml-cuda.h" +//#endif +// +//#ifdef GGML_USE_SYCL +//#include "ggml-sycl.h" +//#endif +// +//#ifdef GGML_USE_METAL +//#include "ggml-metal.h" +//#endif +// +//#ifdef GGML_USE_CANN +//#include "ggml-cann.h" +//#endif +// +//#ifdef GGML_USE_VULKAN +//#include "ggml-vulkan.h" +//#endif #define STB_IMAGE_IMPLEMENTATION #include "stb_image.h" @@ -40,10 +45,17 @@ #include #include -#define LOG_INF(...) do { fprintf(stdout, __VA_ARGS__); } while (0) -#define LOG_WRN(...) do { fprintf(stderr, __VA_ARGS__); } while (0) -#define LOG_ERR(...) do { fprintf(stderr, __VA_ARGS__); } while (0) -#define LOG_DBG(...) do { fprintf(stderr, __VA_ARGS__); } while (0) +#if defined(LLAVA_LOG_OFF) +# define LOG_INF(...) +# define LOG_WRN(...) +# define LOG_ERR(...) +# define LOG_DBG(...) +#else // defined(LLAVA_LOG_OFF) +# define LOG_INF(...) do { fprintf(stdout, __VA_ARGS__); } while (0) +# define LOG_WRN(...) do { fprintf(stderr, __VA_ARGS__); } while (0) +# define LOG_ERR(...) do { fprintf(stderr, __VA_ARGS__); } while (0) +# define LOG_DBG(...) do { fprintf(stdout, __VA_ARGS__); } while (0) +#endif // defined(LLAVA_LOG_OFF) //#define CLIP_DEBUG_FUNCTIONS @@ -91,7 +103,9 @@ static std::string format(const char * fmt, ...) { #define KEY_HAS_LLAVA_PROJ "clip.has_llava_projector" #define KEY_HAS_MINICPMV_PROJ "clip.has_minicpmv_projector" #define KEY_MINICPMV_VERSION "clip.minicpmv_version" +#define KEY_HAS_QWEN2VL_MERGER "clip.has_qwen2vl_merger" #define KEY_USE_GELU "clip.use_gelu" +#define KEY_USE_SILU "clip.use_silu" #define KEY_N_EMBD "clip.%s.embedding_length" #define KEY_N_FF "clip.%s.feed_forward_length" #define KEY_N_BLOCK "clip.%s.block_count" @@ -118,7 +132,8 @@ static std::string format(const char * fmt, ...) { #define TN_TOKEN_EMBD "%s.token_embd.weight" #define TN_POS_EMBD "%s.position_embd.weight" #define TN_CLASS_EMBD "v.class_embd" -#define TN_PATCH_EMBD "v.patch_embd.weight" +#define TN_PATCH_EMBD "v.patch_embd.weight" // not rename tensor with ".0" postfix for backwrad compat +#define TN_PATCH_EMBD_1 "v.patch_embd.weight.1" #define TN_PATCH_BIAS "v.patch_embd.bias" #define TN_ATTN_K "%s.blk.%d.attn_k.%s" #define TN_ATTN_Q "%s.blk.%d.attn_q.%s" @@ -152,6 +167,7 @@ enum projector_type { PROJECTOR_TYPE_LDP, PROJECTOR_TYPE_LDPV2, PROJECTOR_TYPE_RESAMPLER, + PROJECTOR_TYPE_MERGER, PROJECTOR_TYPE_UNKNOWN, }; @@ -160,6 +176,7 @@ static std::map PROJECTOR_TYPE_NAMES = { { PROJECTOR_TYPE_LDP, "ldp" }, { PROJECTOR_TYPE_LDPV2, "ldpv2"}, { PROJECTOR_TYPE_RESAMPLER, "resampler"}, + { PROJECTOR_TYPE_MERGER, "qwen2vl_merger"}, }; @@ -246,7 +263,7 @@ static std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i) { { const enum gguf_type arr_type = gguf_get_arr_type(ctx_gguf, i); int arr_n = gguf_get_arr_n(ctx_gguf, i); - const void * data = gguf_get_arr_data(ctx_gguf, i); + const void * data = arr_type == GGUF_TYPE_STRING ? nullptr : gguf_get_arr_data(ctx_gguf, i); std::stringstream ss; ss << "["; for (int j = 0; j < arr_n; j++) { @@ -452,7 +469,8 @@ struct clip_vision_model { // embeddings struct ggml_tensor * class_embedding; - struct ggml_tensor * patch_embeddings; + struct ggml_tensor * patch_embeddings_0; + struct ggml_tensor * patch_embeddings_1; // second Conv2D kernel when we decouple Conv3D along temproal dimension (Qwen2VL) struct ggml_tensor * patch_bias; struct ggml_tensor * position_embeddings; @@ -542,6 +560,7 @@ struct clip_ctx { bool has_vision_encoder = false; bool has_llava_projector = false; bool has_minicpmv_projector = false; + bool has_qwen2vl_merger = false; int minicpmv_version = 2; struct clip_vision_model vision_model; @@ -550,6 +569,7 @@ struct clip_ctx { float image_mean[3]; float image_std[3]; bool use_gelu = false; + bool use_silu = false; int32_t ftype = 1; bool has_class_embedding = true; @@ -595,14 +615,26 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 image_size_height = imgs->data->ny; } } + else if (ctx->has_qwen2vl_merger) { + // use the image's native resolution when image is avaible + if (is_inf) { + // if (imgs->data->nx && imgs->data->ny) { + image_size_width = imgs->data->nx; + image_size_height = imgs->data->ny; + } + } const int patch_size = hparams.patch_size; const int num_patches = ((image_size_width / patch_size) * (image_size_height / patch_size)); + const int patches_w = image_size_width / patch_size; + const int patches_h = image_size_height / patch_size; const int num_positions = num_patches + (ctx->has_class_embedding ? 1 : 0); + const int num_position_ids = ctx->has_qwen2vl_merger ? num_positions * 4 : num_positions; const int hidden_size = hparams.hidden_size; const int n_head = hparams.n_head; const int d_head = hidden_size / n_head; int n_layer = hparams.n_layer; const float eps = hparams.eps; + int mrope_sections[4] = {d_head/4, d_head/4, d_head/4, d_head/4}; const int batch_size = imgs->size; @@ -623,10 +655,30 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 ggml_set_name(inp_raw, "inp_raw"); ggml_set_input(inp_raw); - struct ggml_tensor * inp = ggml_conv_2d(ctx0, model.patch_embeddings, inp_raw, patch_size, patch_size, 0, 0, 1, 1); + struct ggml_tensor * inp = ggml_conv_2d(ctx0, model.patch_embeddings_0, inp_raw, patch_size, patch_size, 0, 0, 1, 1); - inp = ggml_reshape_3d(ctx0, inp, num_patches, hidden_size, batch_size); - inp = ggml_cont(ctx0, ggml_permute(ctx0, inp, 1, 0, 2, 3)); + if (ctx->has_qwen2vl_merger) { + GGML_ASSERT(image_size_width % (patch_size * 2) == 0); + GGML_ASSERT(image_size_height % (patch_size * 2) == 0); + + auto inp_1 = ggml_conv_2d(ctx0, model.patch_embeddings_1, inp_raw, patch_size, patch_size, 0, 0, 1, 1); + inp = ggml_add(ctx0, inp, inp_1); + inp = ggml_cont(ctx0, ggml_permute(ctx0, inp, 1, 2, 0, 3)); // [w, h, c, b] -> [c, w, h, b] + inp = ggml_reshape_4d( + ctx0, inp, + hidden_size * 2, patches_w / 2, patches_h, batch_size); + inp = ggml_reshape_4d( + ctx0, inp, + hidden_size * 2, patches_w / 2, 2, batch_size * (patches_h / 2)); + inp = ggml_cont(ctx0, ggml_permute(ctx0, inp, 0, 2, 1, 3)); + inp = ggml_reshape_3d( + ctx0, inp, + hidden_size, patches_w * patches_h, batch_size); + } + else { + inp = ggml_reshape_3d(ctx0, inp, num_patches, hidden_size, batch_size); + inp = ggml_cont(ctx0, ggml_permute(ctx0, inp, 1, 0, 2, 3)); + } if (ctx->has_patch_bias) { // inp = ggml_add(ctx0, inp, ggml_repeat(ctx0, model.patch_bias, inp)); @@ -648,12 +700,14 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 } } - struct ggml_tensor * positions = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, num_positions); + struct ggml_tensor * positions = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, num_position_ids); ggml_set_name(positions, "positions"); ggml_set_input(positions); - embeddings = - ggml_add(ctx0, embeddings, ggml_get_rows(ctx0, model.position_embeddings, positions)); + if (!ctx->has_qwen2vl_merger) { // qwen2vl use rope position embedding + embeddings = + ggml_add(ctx0, embeddings, ggml_get_rows(ctx0, model.position_embeddings, positions)); + } if (ctx->has_minicpmv_projector) { int pos_w = image_size_width/patch_size; @@ -677,7 +731,8 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 } // loop over layers - if (ctx->has_minicpmv_projector) { + if (ctx->has_minicpmv_projector || ctx->has_qwen2vl_merger) { + // TODO: figure out why we doing thing in this way ??? n_layer += 1; } for (int il = 0; il < n_layer - 1; il++) { @@ -699,8 +754,13 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 struct ggml_tensor * Q = ggml_add(ctx0, ggml_mul_mat(ctx0, model.layers[il].q_w, cur), model.layers[il].q_b); - Q = ggml_scale_inplace(ctx0, Q, 1.0f / sqrt((float)d_head)); Q = ggml_reshape_4d(ctx0, Q, d_head, n_head, num_positions, batch_size); + if (ctx->has_qwen2vl_merger) { + Q = ggml_rope_multi( + ctx0, Q, positions, nullptr, + d_head/2, mrope_sections, GGML_ROPE_TYPE_VISION, 32768, 10000, 1, 0, 1, 32, 1); + } + Q = ggml_scale_inplace(ctx0, Q, 1.0f / sqrt((float)d_head)); Q = ggml_cont(ctx0, ggml_permute(ctx0, Q, 0, 2, 1, 3)); Q = ggml_reshape_3d(ctx0, Q, d_head, num_positions, n_head * batch_size); @@ -708,6 +768,11 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 ggml_add(ctx0, ggml_mul_mat(ctx0, model.layers[il].k_w, cur), model.layers[il].k_b); K = ggml_reshape_4d(ctx0, K, d_head, n_head, num_positions, batch_size); + if (ctx->has_qwen2vl_merger) { + K = ggml_rope_multi( + ctx0, K, positions, nullptr, + d_head/2, mrope_sections, GGML_ROPE_TYPE_VISION, 32768, 10000, 1, 0, 1, 32, 1); + } K = ggml_cont(ctx0, ggml_permute(ctx0, K, 0, 2, 1, 3)); K = ggml_reshape_3d(ctx0, K, d_head, num_positions, n_head * batch_size); @@ -747,6 +812,8 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 if (ctx->use_gelu) { cur = ggml_gelu_inplace(ctx0, cur); + } else if (ctx->use_silu) { + cur = ggml_silu_inplace(ctx0, cur); } else { cur = ggml_gelu_quick_inplace(ctx0, cur); } @@ -758,6 +825,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 cur = ggml_add(ctx0, embeddings, cur); embeddings = cur; + } // post-layernorm @@ -829,7 +897,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 mlp_3 = ggml_cont(ctx0, ggml_permute(ctx0, mlp_3, 1, 0, 2, 3)); mlp_3 = ggml_reshape_4d(ctx0, mlp_3, n_patch, n_patch, mlp_3->ne[1], mlp_3->ne[2]); // stride = 1, padding = 1, bias is nullptr - block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_1_block_0_0_w, mlp_3, 1, 1, 1, 1, 1, 1); + block_1 = ggml_conv_2d_dw(ctx0, model.mm_model_block_1_block_0_0_w, mlp_3, 1, 1, 1, 1, 1, 1); // layer norm // // block_1 shape = [1, 2048, 24, 24], ne = [24, 24, 2048, 1] @@ -877,7 +945,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 // block_2 { // stride = 2 - block_1 = ggml_conv_depthwise_2d(ctx0, model.mm_model_block_2_block_0_0_w, block_1, 2, 2, 1, 1, 1, 1); + block_1 = ggml_conv_2d_dw(ctx0, model.mm_model_block_2_block_0_0_w, block_1, 2, 2, 1, 1, 1, 1); // block_1 shape = [1, 2048, 12, 12], ne = [12, 12, 2048, 1] // layer norm @@ -938,7 +1006,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 // mlp_2 ne [24, 24, 2048, 1] mlp_2 = ggml_pool_2d(ctx0, mlp_2, GGML_OP_POOL_AVG, 2, 2, 2, 2, 0, 0); // weight ne = [3, 3, 2048, 1] - struct ggml_tensor * peg_0 = ggml_conv_depthwise_2d(ctx0, model.mm_model_peg_0_w, mlp_2, 1, 1, 1, 1, 1, 1); + struct ggml_tensor * peg_0 = ggml_conv_2d_dw(ctx0, model.mm_model_peg_0_w, mlp_2, 1, 1, 1, 1, 1, 1); peg_0 = ggml_cont(ctx0, ggml_permute(ctx0, peg_0, 1, 2, 0, 3)); peg_0 = ggml_add(ctx0, peg_0, model.mm_model_peg_0_b); mlp_2 = ggml_cont(ctx0, ggml_permute(ctx0, mlp_2, 1, 2, 0, 3)); @@ -1019,6 +1087,19 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 GGML_ASSERT(false); } } + else if (ctx->proj_type == PROJECTOR_TYPE_MERGER) { + embeddings = ggml_reshape_3d(ctx0, embeddings, hidden_size * 4, num_positions / 4, batch_size); + + embeddings = ggml_mul_mat(ctx0, model.mm_0_w, embeddings); + embeddings = ggml_add(ctx0, embeddings, model.mm_0_b); + + // GELU activation + embeddings = ggml_gelu(ctx0, embeddings); + + // Second linear layer + embeddings = ggml_mul_mat(ctx0, model.mm_1_w, embeddings); + embeddings = ggml_add(ctx0, embeddings, model.mm_1_b); + } // build the graph ggml_build_forward_expand(gf, embeddings); @@ -1142,25 +1223,30 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { } } -#ifdef GGML_USE_CUDA - new_clip->backend = ggml_backend_cuda_init(0); - LOG_INF("%s: CLIP using CUDA backend\n", __func__); -#endif - -#ifdef GGML_USE_METAL - new_clip->backend = ggml_backend_metal_init(); - LOG_INF("%s: CLIP using Metal backend\n", __func__); -#endif - -#ifdef GGML_USE_CANN - new_clip->backend = ggml_backend_cann_init(0); - LOG_INF("%s: CLIP using CANN backend\n", __func__); -#endif - -#ifdef GGML_USE_VULKAN - new_clip->backend = ggml_backend_vk_init(0); - LOG_INF("%s: CLIP using Vulkan backend\n", __func__); -#endif +//#ifdef GGML_USE_CUDA +// new_clip->backend = ggml_backend_cuda_init(0); +// LOG_INF("%s: CLIP using CUDA backend\n", __func__); +//#endif +// +//#ifdef GGML_USE_METAL +// new_clip->backend = ggml_backend_metal_init(); +// LOG_INF("%s: CLIP using Metal backend\n", __func__); +//#endif +// +//#ifdef GGML_USE_CANN +// new_clip->backend = ggml_backend_cann_init(0); +// LOG_INF("%s: CLIP using CANN backend\n", __func__); +//#endif +// +//#ifdef GGML_USE_VULKAN +// new_clip->backend = ggml_backend_vk_init(0); +// LOG_INF("%s: CLIP using Vulkan backend\n", __func__); +//#endif +// +//#ifdef GGML_USE_SYCL +// new_clip->backend = ggml_backend_sycl_init(0); +// LOG_INF("%s: CLIP using SYCL backend\n", __func__); +//#endif if (!new_clip->backend) { new_clip->backend = ggml_backend_cpu_init(); @@ -1190,6 +1276,10 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { new_clip->minicpmv_version = gguf_get_val_i32(ctx, idx); } + idx = gguf_find_key(ctx, KEY_HAS_QWEN2VL_MERGER); + if (idx != -1) { + new_clip->has_qwen2vl_merger = gguf_get_val_bool(ctx, idx); + } // GGML_ASSERT(new_clip->has_llava_projector); // see monatis/clip.cpp for image and/or text encoding for semantic search GGML_ASSERT(new_clip->has_vision_encoder); @@ -1198,6 +1288,13 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { idx = get_key_idx(ctx, KEY_USE_GELU); new_clip->use_gelu = gguf_get_val_bool(ctx, idx); + try { + idx = get_key_idx(ctx, KEY_USE_SILU); + new_clip->use_silu = gguf_get_val_bool(ctx, idx); + } catch (std::runtime_error & /*e*/) { + new_clip->use_silu = false; + } + if (verbosity >= 1) { LOG_INF("%s: text_encoder: %d\n", __func__, new_clip->has_text_encoder); LOG_INF("%s: vision_encoder: %d\n", __func__, new_clip->has_vision_encoder); @@ -1373,11 +1470,16 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { } try { - vision_model.patch_embeddings = get_tensor(new_clip->ctx_data, TN_PATCH_EMBD); + vision_model.patch_embeddings_0 = get_tensor(new_clip->ctx_data, TN_PATCH_EMBD); vision_model.position_embeddings = get_tensor(new_clip->ctx_data, format(TN_POS_EMBD, "v")); } catch(const std::exception& /*e*/) { LOG_ERR("%s: failed to load vision model tensors\n", __func__); } + try { + vision_model.patch_embeddings_1 = get_tensor(new_clip->ctx_data, TN_PATCH_EMBD_1); + } catch(const std::exception& /*e*/) { + new_clip->has_qwen2vl_merger = false; + } // LLaVA projection if (new_clip->proj_type == PROJECTOR_TYPE_MLP || new_clip->proj_type == PROJECTOR_TYPE_MLP_NORM) { @@ -1465,6 +1567,12 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { vision_model.mm_model_ln_post_w = get_tensor(new_clip->ctx_data, format(TN_MINICPMV_LN, "post", "weight")); vision_model.mm_model_ln_post_b = get_tensor(new_clip->ctx_data, format(TN_MINICPMV_LN, "post", "bias")); } + else if (new_clip->proj_type == PROJECTOR_TYPE_MERGER) { + vision_model.mm_0_w = get_tensor(new_clip->ctx_data, format(TN_LLAVA_PROJ, 0, "weight")); + vision_model.mm_0_b = get_tensor(new_clip->ctx_data, format(TN_LLAVA_PROJ, 0, "bias")); + vision_model.mm_1_w = get_tensor(new_clip->ctx_data, format(TN_LLAVA_PROJ, 2, "weight")); + vision_model.mm_1_b = get_tensor(new_clip->ctx_data, format(TN_LLAVA_PROJ, 2, "bias")); + } else { std::string proj_type = PROJECTOR_TYPE_NAMES[new_clip->proj_type]; throw std::runtime_error(format("%s: don't support projector with: %s currently\n", __func__, proj_type.c_str())); @@ -1503,6 +1611,7 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { new_clip->compute_alloc = ggml_gallocr_new(ggml_backend_get_default_buffer_type(new_clip->backend)); clip_image_f32_batch batch; batch.size = 1; + batch.data = nullptr; ggml_cgraph * gf = clip_image_build_graph(new_clip, &batch, nullptr, false); ggml_gallocr_reserve(new_clip->compute_alloc, gf); size_t compute_memory_buffer_size = ggml_gallocr_get_buffer_size(new_clip->compute_alloc, 0); @@ -1516,6 +1625,10 @@ void clip_add_load_image_size(struct clip_ctx * ctx_clip, struct clip_image_size ctx_clip->load_image_size = load_image_size; } +struct clip_image_size * clip_get_load_image_size(struct clip_ctx * ctx_clip) { + return ctx_clip->load_image_size; +} + struct clip_image_size * clip_image_size_init() { struct clip_image_size * load_image_size = new struct clip_image_size(); load_image_size->width = 448; @@ -1968,6 +2081,23 @@ bool clip_image_preprocess(struct clip_ctx * ctx, const clip_image_u8 * img, cli } return true; } + else if (ctx->has_qwen2vl_merger) { + clip_image_u8 * resized = clip_image_u8_init(); + auto patch_size = clip_patch_size(ctx) * 2; + int nx = ceil((float)img->nx / patch_size) * patch_size; + int ny = ceil((float)img->ny / patch_size) * patch_size; + bicubic_resize(*img, *resized, nx, ny); + + res_imgs->data = new clip_image_f32[1]; + // clip_image_f32 * res = clip_image_f32_init(); + normalize_image_u8_to_f32(resized, res_imgs->data, ctx->image_mean, ctx->image_std); + // res_imgs->data[0] = *res; + res_imgs->size = 1; + + // clip_image_f32_free(res); + clip_image_u8_free(resized); + return true; + } bool pad_to_square = true; if (!ctx->has_vision_encoder) { @@ -2157,6 +2287,13 @@ size_t clip_embd_nbytes(const struct clip_ctx * ctx) { return clip_n_patches(ctx) * clip_n_mmproj_embd(ctx) * sizeof(float); } +size_t clip_embd_nbytes_by_img(const struct clip_ctx * ctx, int img_h, int img_w) { + clip_image_f32 img; + img.nx = img_w; + img.ny = img_h; + return clip_n_patches_by_img(ctx, &img) * clip_n_mmproj_embd(ctx) * sizeof(float); +} + int32_t clip_image_size(const struct clip_ctx * ctx) { return ctx->vision_model.hparams.image_size; } @@ -2178,6 +2315,13 @@ const int32_t * clip_image_grid(const struct clip_ctx * ctx) { } int clip_n_patches(const struct clip_ctx * ctx) { + clip_image_f32 img; + img.nx = ctx->vision_model.hparams.image_size; + img.ny = ctx->vision_model.hparams.image_size; + return clip_n_patches_by_img(ctx, &img); +} + +int clip_n_patches_by_img(const struct clip_ctx * ctx, struct clip_image_f32 * img) { const auto & params = ctx->vision_model.hparams; int n_patches = (params.image_size / params.patch_size) * (params.image_size / params.patch_size); @@ -2191,6 +2335,11 @@ int clip_n_patches(const struct clip_ctx * ctx) { else if (ctx->minicpmv_version == 3) { n_patches = 64; } + } else if (ctx->proj_type == PROJECTOR_TYPE_MERGER) { + int patch_size = params.patch_size * 2; + int x_patch = img->nx / patch_size + (int)(img->nx % patch_size > 0); + int y_patch = img->ny / patch_size + (int)(img->ny % patch_size > 0); + n_patches = x_patch * y_patch; } return n_patches; @@ -2319,7 +2468,7 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima const int image_size = hparams.image_size; int image_size_width = image_size; int image_size_height = image_size; - if (ctx->has_minicpmv_projector) { + if (ctx->has_minicpmv_projector | ctx->has_qwen2vl_merger) { image_size_width = imgs->data[0].nx; image_size_height = imgs->data[0].ny; } @@ -2339,7 +2488,7 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima for (size_t i = 0; i < imgs->size; i++) { const int nx = imgs->data[i].nx; const int ny = imgs->data[i].ny; - if (!ctx->has_minicpmv_projector) { + if (!(ctx->has_minicpmv_projector | ctx->has_qwen2vl_merger)) { GGML_ASSERT(nx == image_size && ny == image_size); } @@ -2397,9 +2546,9 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima auto pos_embed_t = get_2d_sincos_pos_embed(embed_dim, std::make_pair(pos_w, pos_h)); float * pos_embed_data = (float *)malloc(ggml_nbytes(pos_embed)); - for(int i=0;ihas_qwen2vl_merger) { + struct ggml_tensor * positions = ggml_graph_get_tensor(gf, "positions"); + + const int pw = image_size_width / patch_size; + const int ph = image_size_height / patch_size; + int* positions_data = (int*)malloc(ggml_nbytes(positions)); + + int ptr = 0; + for (int y = 0; y < ph; y+=2) + { + for (int x = 0; x < pw; x+=2) + { + for (int dy = 0; dy < 2; dy++) { + for (int dx = 0; dx < 2; dx++) { + positions_data[ptr] = y + dy; + positions_data[num_patches + ptr] = x + dx; + positions_data[num_patches * 2 + ptr] = y + dy; + positions_data[num_patches * 3 + ptr] = x + dx; + ptr++; + } + } + } + } + + ggml_backend_tensor_set(positions, positions_data, 0, ggml_nbytes(positions)); + free(positions_data); + } + else { struct ggml_tensor * positions = ggml_graph_get_tensor(gf, "positions"); int* positions_data = (int*)malloc(ggml_nbytes(positions)); @@ -2428,16 +2604,16 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima } ggml_backend_tensor_set(positions, positions_data, 0, ggml_nbytes(positions)); free(positions_data); - } - { - struct ggml_tensor * patches = ggml_graph_get_tensor(gf, "patches"); - int* patches_data = (int*)malloc(ggml_nbytes(patches)); - for (int i = 0; i < num_patches; i++) { - patches_data[i] = i + 1; + { + struct ggml_tensor * patches = ggml_graph_get_tensor(gf, "patches"); + int* patches_data = (int*)malloc(ggml_nbytes(patches)); + for (int i = 0; i < num_patches; i++) { + patches_data[i] = i + 1; + } + ggml_backend_tensor_set(patches, patches_data, 0, ggml_nbytes(patches)); + free(patches_data); } - ggml_backend_tensor_set(patches, patches_data, 0, ggml_nbytes(patches)); - free(patches_data); } } @@ -2559,7 +2735,8 @@ bool clip_model_quantize(const char * fname_inp, const char * fname_out, const i total_size_org += orig_size; total_size_new += new_size; gguf_set_tensor_type(ctx_out, name.c_str(), new_type); - gguf_set_tensor_data(ctx_out, name.c_str(), new_data, new_size); + GGML_ASSERT(gguf_get_tensor_size(ctx_out, gguf_find_tensor(ctx_out, name.c_str())) == new_size); + gguf_set_tensor_data(ctx_out, name.c_str(), new_data); fout.write((const char *)new_data, new_size); size_t pad = GGML_PAD(new_size, gguf_get_alignment(ctx_out)) - new_size; for (size_t j = 0; j < pad; ++j) { @@ -2610,6 +2787,9 @@ int clip_n_mmproj_embd(const struct clip_ctx * ctx) { return 3584; } } + if (ctx->proj_type == PROJECTOR_TYPE_MERGER) { + return ctx->vision_model.mm_1_b->ne[0]; + } std::string proj_type = PROJECTOR_TYPE_NAMES[ctx->proj_type]; throw std::runtime_error(format("%s: don't support projector with: %s currently\n", __func__, proj_type.c_str())); @@ -2621,3 +2801,21 @@ int clip_is_minicpmv(const struct clip_ctx * ctx) { } return 0; } + +bool clip_is_qwen2vl(const struct clip_ctx * ctx) { + return ctx->has_qwen2vl_merger; +} + + +bool clip_encode_float_image (struct clip_ctx * ctx, int n_threads, float * img, int h, int w, float * vec) { + clip_image_f32 clip_img; + clip_img.buf.resize(h * w * 3); + for (int i = 0; i < h*w*3; i++) + { + clip_img.buf[i] = img[i]; + } + clip_img.nx = w; + clip_img.ny = h; + clip_image_encode(ctx, n_threads, &clip_img, vec); + return true; +} diff --git a/examples/llava/clip.h b/examples/llava/clip.h index 78588bdf1..1603edd26 100644 --- a/examples/llava/clip.h +++ b/examples/llava/clip.h @@ -45,6 +45,7 @@ CLIP_API struct clip_ctx * clip_model_load_cpu(const char * fname, int verbosity CLIP_API void clip_free(struct clip_ctx * ctx); CLIP_API size_t clip_embd_nbytes(const struct clip_ctx * ctx); +CLIP_API size_t clip_embd_nbytes_by_img(const struct clip_ctx * ctx, int img_h, int img_w); CLIP_API int32_t clip_image_size (const struct clip_ctx * ctx); CLIP_API int32_t clip_patch_size (const struct clip_ctx * ctx); @@ -55,11 +56,13 @@ CLIP_API const char * clip_patch_merge_type(const struct clip_ctx * ctx); CLIP_API const int32_t * clip_image_grid(const struct clip_ctx * ctx); -CLIP_API int clip_n_patches (const struct clip_ctx * ctx); -CLIP_API int clip_n_mmproj_embd(const struct clip_ctx * ctx); +CLIP_API int clip_n_patches (const struct clip_ctx * ctx); +CLIP_API int clip_n_patches_by_img (const struct clip_ctx * ctx, struct clip_image_f32 * img); +CLIP_API int clip_n_mmproj_embd (const struct clip_ctx * ctx); CLIP_API int clip_uhd_num_image_embeds_col(struct clip_ctx * ctx_clip); CLIP_API void clip_add_load_image_size(struct clip_ctx * ctx_clip, struct clip_image_size * load_image_size); +CLIP_API struct clip_image_size * clip_get_load_image_size(struct clip_ctx * ctx_clip); CLIP_API struct clip_image_size * clip_image_size_init(); CLIP_API struct clip_image_u8 * clip_image_u8_init (); @@ -86,6 +89,9 @@ CLIP_API bool clip_image_batch_encode(struct clip_ctx * ctx, int n_threads, cons CLIP_API bool clip_model_quantize(const char * fname_inp, const char * fname_out, int itype); CLIP_API int clip_is_minicpmv(const struct clip_ctx * ctx); +CLIP_API bool clip_is_qwen2vl(const struct clip_ctx * ctx); + +CLIP_API bool clip_encode_float_image (struct clip_ctx * ctx, int n_threads, float * img, int h, int w, float * vec); #ifdef __cplusplus } diff --git a/examples/llava/llava-cli.cpp b/examples/llava/llava-cli.cpp index 161098585..40aa0876f 100644 --- a/examples/llava/llava-cli.cpp +++ b/examples/llava/llava-cli.cpp @@ -47,8 +47,12 @@ static const char * sample(struct common_sampler * smpl, int * n_past) { const llama_token id = common_sampler_sample(smpl, ctx_llama, -1); common_sampler_accept(smpl, id, true); + + const llama_model * model = llama_get_model(ctx_llama); + const llama_vocab * vocab = llama_model_get_vocab(model); + static std::string ret; - if (llama_token_is_eog(llama_get_model(ctx_llama), id)) { + if (llama_vocab_is_eog(vocab, id)) { ret = ""; } else { ret = common_token_to_piece(ctx_llama, id); @@ -191,7 +195,7 @@ static void process_prompt(struct llava_context * ctx_llava, struct llava_image_ LOG("\n"); - struct common_sampler * smpl = common_sampler_init(ctx_llava->model, params->sparams); + struct common_sampler * smpl = common_sampler_init(ctx_llava->model, params->sampling); if (!smpl) { LOG_ERR("%s: failed to initialize sampling subsystem\n", __func__); exit(1); @@ -221,7 +225,7 @@ static struct llama_model * llava_init(common_params * params) { llama_model_params model_params = common_model_params_to_llama(*params); - llama_model * model = llama_load_model_from_file(params->model.c_str(), model_params); + llama_model * model = llama_model_load_from_file(params->model.c_str(), model_params); if (model == NULL) { LOG_ERR("%s: unable to load model\n" , __func__); return NULL; @@ -239,11 +243,10 @@ static struct llava_context * llava_init_context(common_params * params, llama_m auto ctx_clip = clip_model_load(clip_path, /*verbosity=*/ 1); - llama_context_params ctx_params = common_context_params_to_llama(*params); ctx_params.n_ctx = params->n_ctx < 2048 ? 2048 : params->n_ctx; // we need a longer context size to process image embeddings - llama_context * ctx_llama = llama_new_context_with_model(model, ctx_params); + llama_context * ctx_llama = llama_init_from_model(model, ctx_params); if (ctx_llama == NULL) { LOG_ERR("%s: failed to create the llama_context\n" , __func__); @@ -265,7 +268,7 @@ static void llava_free(struct llava_context * ctx_llava) { } llama_free(ctx_llava->ctx_llama); - llama_free_model(ctx_llava->model); + llama_model_free(ctx_llava->model); llama_backend_free(); } @@ -323,7 +326,7 @@ int main(int argc, char ** argv) { } } - llama_free_model(model); + llama_model_free(model); return 0; } diff --git a/examples/llava/llava.cpp b/examples/llava/llava.cpp index be6988540..c598caf3d 100644 --- a/examples/llava/llava.cpp +++ b/examples/llava/llava.cpp @@ -11,13 +11,17 @@ #include #include -#define die(msg) do { fputs("error: " msg "\n", stderr); exit(1); } while (0) -#define die_fmt(fmt, ...) do { fprintf(stderr, "error: " fmt "\n", __VA_ARGS__); exit(1); } while (0) - -#define LOG_INF(...) do { fprintf(stdout, __VA_ARGS__); } while (0) -#define LOG_WRN(...) do { fprintf(stderr, __VA_ARGS__); } while (0) -#define LOG_ERR(...) do { fprintf(stderr, __VA_ARGS__); } while (0) -#define LOG_DBG(...) do { fprintf(stdout, __VA_ARGS__); } while (0) +#if defined(LLAVA_LOG_OFF) +# define LOG_INF(...) +# define LOG_WRN(...) +# define LOG_ERR(...) +# define LOG_DBG(...) +#else // defined(LLAVA_LOG_OFF) +# define LOG_INF(...) do { fprintf(stdout, __VA_ARGS__); } while (0) +# define LOG_WRN(...) do { fprintf(stderr, __VA_ARGS__); } while (0) +# define LOG_ERR(...) do { fprintf(stderr, __VA_ARGS__); } while (0) +# define LOG_DBG(...) do { fprintf(stdout, __VA_ARGS__); } while (0) +#endif // defined(LLAVA_LOG_OFF) // RGB uint8 image struct clip_image_u8 { @@ -255,25 +259,33 @@ static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const cli const char * mm_patch_merge_type = clip_patch_merge_type(ctx_clip); - if (clip_is_minicpmv(ctx_clip)) { + if (clip_is_minicpmv(ctx_clip) || clip_is_qwen2vl(ctx_clip)) { std::vector image_embd_v; image_embd_v.resize(img_res_v.size); struct clip_image_size * load_image_size = clip_image_size_init(); + for (size_t i = 0; i < img_res_v.size; i++) { const int64_t t_img_enc_step_start_us = ggml_time_us(); - image_embd_v[i] = (float *)malloc(clip_embd_nbytes(ctx_clip)); + image_embd_v[i] = (float *)malloc(clip_embd_nbytes_by_img(ctx_clip, img_res_v.data[i].nx, img_res_v.data[i].ny)); int patch_size=14; load_image_size->width = img_res_v.data[i].nx; load_image_size->height = img_res_v.data[i].ny; clip_add_load_image_size(ctx_clip, load_image_size); + bool encoded = false; - int has_minicpmv_projector = clip_is_minicpmv(ctx_clip); - if (has_minicpmv_projector == 2) { - encoded = clip_image_encode(ctx_clip, n_threads, only_v2_5_reshape_by_patch(&img_res_v.data[i], patch_size), image_embd_v[i]); - } - else if (has_minicpmv_projector == 3) { + if (clip_is_qwen2vl(ctx_clip)) { encoded = clip_image_encode(ctx_clip, n_threads, &img_res_v.data[i], image_embd_v[i]); } + else { + int has_minicpmv_projector = clip_is_minicpmv(ctx_clip); + if (has_minicpmv_projector == 2) { + encoded = clip_image_encode(ctx_clip, n_threads, only_v2_5_reshape_by_patch(&img_res_v.data[i], patch_size), image_embd_v[i]); + } + else if (has_minicpmv_projector == 3) { + encoded = clip_image_encode(ctx_clip, n_threads, &img_res_v.data[i], image_embd_v[i]); + } + } + if (!encoded) { LOG_ERR("Unable to encode image - spatial_unpad - subimage %d of %d\n", (int) i+1, (int) img_res_v.size); return false; @@ -286,8 +298,11 @@ static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const cli int n_img_pos_out = 0; for (size_t i = 0; i < image_embd_v.size(); i++) { - std::memcpy(image_embd + n_img_pos_out * clip_n_mmproj_embd(ctx_clip), image_embd_v[i], clip_embd_nbytes(ctx_clip)); - n_img_pos_out += clip_n_patches(ctx_clip); + std::memcpy( + image_embd + n_img_pos_out * clip_n_mmproj_embd(ctx_clip), + image_embd_v[i], + clip_embd_nbytes_by_img(ctx_clip, img_res_v.data[i].nx, img_res_v.data[i].ny)); + n_img_pos_out += clip_n_patches_by_img(ctx_clip, &img_res_v.data[i]); } *n_img_pos = n_img_pos_out; for (size_t i = 0; i < image_embd_v.size(); i++) { @@ -369,7 +384,7 @@ static bool encode_image_with_clip(clip_ctx * ctx_clip, int n_threads, const cli bool llava_validate_embed_size(const llama_context * ctx_llama, const clip_ctx * ctx_clip) { // make sure that the correct mmproj was used, i.e., compare apples to apples - int n_llama_embd = llama_n_embd(llama_get_model(ctx_llama)); + int n_llama_embd = llama_model_n_embd(llama_get_model(ctx_llama)); auto n_image_embd = clip_n_mmproj_embd(ctx_clip); if (n_image_embd != n_llama_embd) { LOG_ERR("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_image_embd, n_llama_embd); @@ -383,7 +398,13 @@ bool llava_image_embed_make_with_clip_img(clip_ctx * ctx_clip, int n_threads, co if (clip_is_minicpmv(ctx_clip)) { num_max_patches = 10; } - float * image_embd = (float *)malloc(clip_embd_nbytes(ctx_clip)*num_max_patches); // TODO: base on gridsize/llava model + float * image_embd; + if (clip_is_qwen2vl(ctx_clip)) { + // qwen2vl don't split image into chunks, so `num_max_patches` is not needed. + image_embd = (float *)malloc(clip_embd_nbytes_by_img(ctx_clip, img->nx, img->ny)); + } else { + image_embd = (float *)malloc(clip_embd_nbytes(ctx_clip)*num_max_patches); // TODO: base on gridsize/llava model + } if (!image_embd) { LOG_ERR("Unable to allocate memory for image embeddings\n"); return false; @@ -435,7 +456,7 @@ struct llava_embd_batch { }; bool llava_eval_image_embed(llama_context * ctx_llama, const struct llava_image_embed * image_embed, int n_batch, int * n_past) { - int n_embd = llama_n_embd(llama_get_model(ctx_llama)); + int n_embd = llama_model_n_embd(llama_get_model(ctx_llama)); for (int i = 0; i < image_embed->n_image_pos; i += n_batch) { int n_eval = image_embed->n_image_pos - i; @@ -498,10 +519,16 @@ static bool load_file_to_bytes(const char* path, unsigned char** bytesOut, long errno = 0; size_t ret = fread(buffer, 1, fileSize, file); // Read the file into the buffer if (ferror(file)) { - die_fmt("read error: %s", strerror(errno)); + LOG_ERR("read error: %s", strerror(errno)); + free(buffer); + fclose(file); + return false; } if (ret != (size_t) fileSize) { - die("unexpectedly reached end of file"); + LOG_ERR("unexpectedly reached end of file"); + free(buffer); + fclose(file); + return false; } fclose(file); // Close the file diff --git a/examples/llava/minicpmv-cli.cpp b/examples/llava/minicpmv-cli.cpp index cbecec343..38c44e130 100644 --- a/examples/llava/minicpmv-cli.cpp +++ b/examples/llava/minicpmv-cli.cpp @@ -31,7 +31,7 @@ static struct llama_model * llava_init(common_params * params) { llama_model_params model_params = common_model_params_to_llama(*params); - llama_model * model = llama_load_model_from_file(params->model.c_str(), model_params); + llama_model * model = llama_model_load_from_file(params->model.c_str(), model_params); if (model == NULL) { LOG_ERR("%s: unable to load model\n" , __func__); return NULL; @@ -54,7 +54,7 @@ static struct llava_context * llava_init_context(common_params * params, llama_m ctx_params.n_ctx = params->n_ctx; } - llama_context * ctx_llama = llama_new_context_with_model(model, ctx_params); + llama_context * ctx_llama = llama_init_from_model(model, ctx_params); if (ctx_llama == NULL) { LOG_ERR("%s: failed to create the llama_context\n" , __func__); @@ -75,7 +75,7 @@ static void llava_free(struct llava_context * ctx_llava) { } llama_free(ctx_llava->ctx_llama); - llama_free_model(ctx_llava->model); + llama_model_free(ctx_llava->model); llama_backend_free(); } @@ -167,8 +167,12 @@ static const char * sample(struct common_sampler * smpl, int * n_past) { const llama_token id = common_sampler_sample(smpl, ctx_llama, -1); common_sampler_accept(smpl, id, true); + + const llama_model * model = llama_get_model(ctx_llama); + const llama_vocab * vocab = llama_model_get_vocab(model); + static std::string ret; - if (llama_token_is_eog(llama_get_model(ctx_llama), id)) { + if (llama_vocab_is_eog(vocab, id)) { ret = ""; } else { ret = common_token_to_piece(ctx_llama, id); @@ -237,7 +241,7 @@ static struct common_sampler * llama_init(struct llava_context * ctx_llava, comm LOG_INF("\n"); - struct common_sampler * smpl = common_sampler_init(ctx_llava->model, params->sparams); + struct common_sampler * smpl = common_sampler_init(ctx_llava->model, params->sampling); return smpl; } diff --git a/examples/llava/qwen2_vl_surgery.py b/examples/llava/qwen2_vl_surgery.py new file mode 100644 index 000000000..c87606b4f --- /dev/null +++ b/examples/llava/qwen2_vl_surgery.py @@ -0,0 +1,165 @@ +import argparse +from typing import Dict + +import torch +import numpy as np +from gguf import * +from transformers import ( + Qwen2VLForConditionalGeneration, + Qwen2VLProcessor, + AutoProcessor, + Qwen2VLConfig +) + + +VISION = "clip.vision" + + +def k(raw_key: str, arch: str) -> str: + return raw_key.format(arch=arch) + + +def to_gguf_name(name: str) -> str: + og = name + name = name.replace("text_model", "t").replace("vision_model", "v") + name = name.replace("blocks", "blk").replace("embeddings.", "") + name = name.replace("attn.", "attn_") + name = name.replace("mlp.fc1", "ffn_down").replace("mlp.fc2", "ffn_up").replace("proj.", "out.") + # name = name.replace("layrnorm", "ln").replace("layer_norm", "ln").replace("layernorm", "ln") + name = name.replace("norm1", "ln1").replace("norm2", "ln2") + name = name.replace("merger.mlp", 'mm') + print(f"[to_gguf_name] {og} --> {name}") + return name + + +def find_vision_tensors(qwen2vl, dtype) -> Dict[str, np.ndarray]: + vision_model = qwen2vl.visual + tensor_map = {} + for name, ten in vision_model.state_dict().items(): + ten = ten.numpy() + if 'qkv' in name: + if ten.ndim == 2: # weight + c3, _ = ten.shape + else: # bias + c3 = ten.shape[0] + assert c3 % 3 == 0 + c = c3 // 3 + wq = ten[:c] + wk = ten[c: c * 2] + wv = ten[c * 2:] + tensor_map[to_gguf_name(f"vision_model.{name}").replace("qkv", "q")] = wq + tensor_map[to_gguf_name(f"vision_model.{name}").replace("qkv", "k")] = wk + tensor_map[to_gguf_name(f"vision_model.{name}").replace("qkv", "v")] = wv + elif 'merger' in name: + if name.endswith("ln_q.weight"): + tensor_map['v.post_ln.weight'] = ten + elif name.endswith("ln_q.bias"): + tensor_map['v.post_ln.bias'] = ten + else: + # "merger.mlp.%d.weight/bias" --> "mm.%d.weight/bias" + tensor_map[to_gguf_name(name)] = ten + elif 'patch_embed.proj.weight' in name: + # NOTE: split Conv3D into Conv2Ds + c1, c2, kt, kh, kw = ten.shape + assert kt == 2, "Current implmentation only support temporal_patch_size of 2" + tensor_map["v.patch_embd.weight"] = ten[:, :, 0, ...] + tensor_map["v.patch_embd.weight.1"] = ten[:, :, 1, ...] + else: + tensor_map[to_gguf_name(f"vision_model.{name}")] = ten + + for new_name, ten in tensor_map.items(): + if ten.ndim <= 1 or new_name.endswith("_norm.weight"): + tensor_map[new_name] = ten.astype(np.float32) + else: + tensor_map[new_name] = ten.astype(dtype) + tensor_map["v.position_embd.weight"] = np.zeros([10, 10], dtype=np.float32) # dummy tensor, just here as a placeholder + return tensor_map + + +def main(args): + if args.data_type == 'fp32': + dtype = torch.float32 + np_dtype = np.float32 + ftype = 0 + elif args.data_type == 'fp16': + dtype = torch.float32 + np_dtype = np.float16 + ftype = 1 + else: + raise ValueError() + + local_model = False + model_path = "" + model_name = args.model_name + print("model_name: ", model_name) + qwen2vl = Qwen2VLForConditionalGeneration.from_pretrained( + model_name, torch_dtype=dtype, device_map="cpu" + ) + cfg: Qwen2VLConfig = qwen2vl.config # type: ignore[reportAssignmentType] + vcfg = cfg.vision_config + + if os.path.isdir(model_name): + local_model = True + if model_name.endswith(os.sep): + model_name = model_name[:-1] + model_path = model_name + model_name = os.path.basename(model_name) + fname_out = f"{model_name.replace('/', '-').lower()}-vision.gguf" + + fout = GGUFWriter(path=fname_out, arch="clip") + fout.add_description("image encoder for Qwen2VL") + + fout.add_file_type(ftype) + fout.add_bool("clip.has_text_encoder", False) + fout.add_bool("clip.has_vision_encoder", True) + fout.add_bool("clip.has_qwen2vl_merger", True) + fout.add_string("clip.projector_type", "qwen2vl_merger") + + print(cfg.vision_config) + if 'silu' in cfg.vision_config.hidden_act.lower(): + fout.add_bool("clip.use_silu", True) + fout.add_bool("clip.use_gelu", False) + elif 'gelu' in cfg.vision_config.hidden_act.lower(): + fout.add_bool("clip.use_silu", False) + fout.add_bool("clip.use_gelu", 'quick' not in cfg.vision_config.hidden_act.lower()) + else: + raise ValueError() + + tensor_map = find_vision_tensors(qwen2vl, np_dtype) + for name, data in tensor_map.items(): + fout.add_tensor(name, data) + + fout.add_uint32("clip.vision.patch_size", vcfg.patch_size) + fout.add_uint32("clip.vision.image_size", 14 * 40) # some reasonable size that is divable by (14*2) + fout.add_uint32(k(KEY_EMBEDDING_LENGTH, VISION), vcfg.embed_dim) + fout.add_uint32("clip.vision.projection_dim", vcfg.hidden_size) + fout.add_uint32(k(KEY_ATTENTION_HEAD_COUNT, VISION), vcfg.num_heads) + fout.add_float32(k(KEY_ATTENTION_LAYERNORM_EPS, VISION), 1e-6) + fout.add_uint32(k(KEY_BLOCK_COUNT, VISION), vcfg.depth) + fout.add_uint32(k(KEY_FEED_FORWARD_LENGTH, VISION), 0) # not sure what this does, put 0 here as a placeholder + fout.add_name(model_name) + """ + HACK: Since vision rope related parameter aren't stored in the `Qwen2VLConfig, + it will be hardcoded in the `clip_image_build_graph` from `clip.cpp`. + """ + + if local_model: + processor: Qwen2VLProcessor = AutoProcessor.from_pretrained(model_path) + else: + processor: Qwen2VLProcessor = AutoProcessor.from_pretrained(model_name) + fout.add_array("clip.vision.image_mean", processor.image_processor.image_mean) # type: ignore[reportAttributeAccessIssue] + fout.add_array("clip.vision.image_std", processor.image_processor.image_std) # type: ignore[reportAttributeAccessIssue] + + fout.write_header_to_file() + fout.write_kv_data_to_file() + fout.write_tensors_to_file() + fout.close() + print("save model as: ", fname_out) + + +if __name__ == "__main__": + parser = argparse.ArgumentParser() + parser.add_argument("model_name", nargs='?', default="Qwen/Qwen2-VL-2B-Instruct") + parser.add_argument("--data_type", nargs='?', choices=['fp32', 'fp16'], default="fp32") + args = parser.parse_args() + main(args) diff --git a/examples/llava/qwen2vl-cli.cpp b/examples/llava/qwen2vl-cli.cpp new file mode 100644 index 000000000..132a7da54 --- /dev/null +++ b/examples/llava/qwen2vl-cli.cpp @@ -0,0 +1,584 @@ +#include "arg.h" +#include "base64.hpp" +#include "log.h" +#include "common.h" +#include "sampling.h" +#include "clip.h" +#include "llava.h" +#include "llama.h" +#include "ggml.h" + +#ifdef GGML_USE_CUDA +#include "ggml-cuda.h" +#endif +#ifdef NDEBUG +#include "ggml-alloc.h" +#include "ggml-backend.h" +#endif + +#include +#include +#include +#include +#include +#include +#include + + +static bool qwen2vl_eval_image_embed(llama_context * ctx_llama, const struct llava_image_embed * image_embed, + int n_batch, int * n_past, int * st_pos_id, struct clip_image_size * image_size) { + int n_embd = llama_model_n_embd(llama_get_model(ctx_llama)); + const int patch_size = 14 * 2; + const int ph = image_size->height / patch_size + (image_size->height % patch_size > 0); + const int pw = image_size->width / patch_size + (image_size->width % patch_size > 0); + auto img_tokens = image_embed->n_image_pos; + // llama_pos mrope_pos[img_tokens * 4]; + std::vector mrope_pos; + mrope_pos.resize(img_tokens * 4); + + for (int y = 0; y < ph; y++) + { + for (int x = 0; x < pw; x++) + { + int i = y * pw + x; + mrope_pos[i] = *st_pos_id; + mrope_pos[i + img_tokens] = *st_pos_id + y; + mrope_pos[i + img_tokens * 2] = *st_pos_id + x; + mrope_pos[i + img_tokens * 3] = 0; + } + } + *st_pos_id += std::max(pw, ph); + + int processed = 0; + std::vector batch_mrope_pos; + batch_mrope_pos.resize(img_tokens * 4); + + for (int i = 0; i < img_tokens; i += n_batch) { + int n_eval = img_tokens - i; + if (n_eval > n_batch) { + n_eval = n_batch; + } + + // llama_pos batch_mrope_pos[n_eval * 4]; + std::fill(batch_mrope_pos.begin(), batch_mrope_pos.end(), 0); + memcpy(batch_mrope_pos.data(), &mrope_pos[processed], n_eval * sizeof(llama_pos)); + memcpy(&batch_mrope_pos[n_eval * 1], &mrope_pos[img_tokens * 1 + processed], n_eval * sizeof(llama_pos)); + memcpy(&batch_mrope_pos[n_eval * 2], &mrope_pos[img_tokens * 2 + processed], n_eval * sizeof(llama_pos)); + memcpy(&batch_mrope_pos[n_eval * 3], &mrope_pos[img_tokens * 3 + processed], n_eval * sizeof(llama_pos)); + + llama_batch batch = { + int32_t(n_eval), // n_tokens + nullptr, // token + (image_embed->embed+i*n_embd), // embed + batch_mrope_pos.data(), // pos + nullptr, // n_seq_id + nullptr, // seq_id + nullptr, // logits + }; + + if (llama_decode(ctx_llama, batch)) { + LOG_ERR("%s : failed to eval\n", __func__); + return false; + } + *n_past += n_eval; + processed += n_eval; + } + return true; +} + + +static bool eval_tokens(struct llama_context * ctx_llama, std::vector tokens, int n_batch, int * n_past, int * st_pos_id) { + int N = (int) tokens.size(); + std::vector pos; + for (int i = 0; i < N; i += n_batch) { + int n_eval = (int) tokens.size() - i; + if (n_eval > n_batch) { + n_eval = n_batch; + } + auto batch = llama_batch_get_one(&tokens[i], n_eval); + // TODO: add mrope pos ids somewhere else + pos.resize(batch.n_tokens * 4); + std::fill(pos.begin(), pos.end(), 0); + for (int j = 0; j < batch.n_tokens * 3; j ++) { + pos[j] = *st_pos_id + (j % batch.n_tokens); + } + batch.pos = pos.data(); + + if (llama_decode(ctx_llama, batch)) { + LOG_ERR("%s : failed to eval. token %d/%d (batch size %d, n_past %d)\n", __func__, i, N, n_batch, *n_past); + return false; + } + *n_past += n_eval; + *st_pos_id += n_eval; + } + return true; +} + +static bool eval_id(struct llama_context * ctx_llama, int id, int * n_past, int * st_pos_id) { + std::vector tokens; + tokens.push_back(id); + return eval_tokens(ctx_llama, tokens, 1, n_past, st_pos_id); +} + +static bool eval_string(struct llama_context * ctx_llama, const char* str, int n_batch, int * n_past, int * st_pos_id, bool add_bos){ + std::string str2 = str; + std::vector embd_inp = common_tokenize(ctx_llama, str2, add_bos, true); + eval_tokens(ctx_llama, embd_inp, n_batch, n_past, st_pos_id); + return true; +} + +static const char * sample(struct common_sampler * smpl, + struct llama_context * ctx_llama, + int * n_past, int * st_pos_id) { + const llama_token id = common_sampler_sample(smpl, ctx_llama, -1); + common_sampler_accept(smpl, id, true); + + const llama_model * model = llama_get_model(ctx_llama); + const llama_vocab * vocab = llama_model_get_vocab(model); + + static std::string ret; + if (llama_vocab_is_eog(vocab, id)) { + ret = ""; + } else { + ret = common_token_to_piece(ctx_llama, id); + } + eval_id(ctx_llama, id, n_past, st_pos_id); + return ret.c_str(); +} + +static const char* IMG_BASE64_TAG_BEGIN = ""; + +static void find_image_tag_in_prompt(const std::string& prompt, size_t& begin_out, size_t& end_out) { + begin_out = prompt.find(IMG_BASE64_TAG_BEGIN); + end_out = prompt.find(IMG_BASE64_TAG_END, (begin_out == std::string::npos) ? 0UL : begin_out); +} + +static bool prompt_contains_image(const std::string& prompt) { + size_t begin, end; + find_image_tag_in_prompt(prompt, begin, end); + return (begin != std::string::npos); +} + +// replaces the base64 image tag in the prompt with `replacement` +static llava_image_embed * llava_image_embed_make_with_prompt_base64(struct clip_ctx * ctx_clip, int n_threads, const std::string& prompt) { + size_t img_base64_str_start, img_base64_str_end; + find_image_tag_in_prompt(prompt, img_base64_str_start, img_base64_str_end); + if (img_base64_str_start == std::string::npos || img_base64_str_end == std::string::npos) { + LOG_ERR("%s: invalid base64 image tag. must be %s%s\n", __func__, IMG_BASE64_TAG_BEGIN, IMG_BASE64_TAG_END); + return NULL; + } + + auto base64_bytes_start = img_base64_str_start + strlen(IMG_BASE64_TAG_BEGIN); + auto base64_bytes_count = img_base64_str_end - base64_bytes_start; + auto base64_str = prompt.substr(base64_bytes_start, base64_bytes_count ); + + auto required_bytes = base64::required_encode_size(base64_str.size()); + auto img_bytes = std::vector(required_bytes); + base64::decode(base64_str.begin(), base64_str.end(), img_bytes.begin()); + + auto embed = llava_image_embed_make_with_bytes(ctx_clip, n_threads, img_bytes.data(), img_bytes.size()); + if (!embed) { + LOG_ERR("%s: could not load image from base64 string.\n", __func__); + return NULL; + } + + return embed; +} + +static std::string remove_image_from_prompt(const std::string& prompt, const char * replacement = "") { + size_t begin, end; + find_image_tag_in_prompt(prompt, begin, end); + if (begin == std::string::npos || end == std::string::npos) { + return prompt; + } + auto pre = prompt.substr(0, begin); + auto post = prompt.substr(end + strlen(IMG_BASE64_TAG_END)); + return pre + replacement + post; +} + +struct llava_context { + struct clip_ctx * ctx_clip = NULL; + struct llama_context * ctx_llama = NULL; + struct llama_model * model = NULL; +}; + +static void print_usage(int, char ** argv) { + LOG("\n example usage:\n"); + LOG("\n %s -m --mmproj --image --image [--temp 0.1] [-p \"describe the image in detail.\"]\n", argv[0]); + LOG("\n note: a lower temperature value like 0.1 is recommended for better quality.\n"); +} + +static struct llava_image_embed * load_image(llava_context * ctx_llava, common_params * params, const std::string & fname) { + + // load and preprocess the image + llava_image_embed * embed = NULL; + auto prompt = params->prompt; + if (prompt_contains_image(prompt)) { + if (!params->image.empty()) { + LOG_INF("using base64 encoded image instead of command line image path\n"); + } + embed = llava_image_embed_make_with_prompt_base64(ctx_llava->ctx_clip, params->cpuparams.n_threads, prompt); + if (!embed) { + LOG_ERR("%s: can't load image from prompt\n", __func__); + return NULL; + } + params->prompt = remove_image_from_prompt(prompt); + } else { + embed = llava_image_embed_make_with_filename(ctx_llava->ctx_clip, params->cpuparams.n_threads, fname.c_str()); + if (!embed) { + fprintf(stderr, "%s: is %s really an image file?\n", __func__, fname.c_str()); + return NULL; + } + } + + return embed; +} + +static void process_prompt(struct llava_context * ctx_llava, struct llava_image_embed * image_embed, common_params * params, const std::string & prompt) { + int n_past = 0; + int cur_pos_id = 0; + + const int max_tgt_len = params->n_predict < 0 ? 256 : params->n_predict; + + std::string system_prompt, user_prompt; + size_t image_pos = prompt.find("<|vision_start|>"); + if (image_pos != std::string::npos) { + // new templating mode: Provide the full prompt including system message and use as a placeholder for the image + system_prompt = prompt.substr(0, image_pos); + user_prompt = prompt.substr(image_pos + std::string("<|vision_pad|>").length()); + LOG_INF("system_prompt: %s\n", system_prompt.c_str()); + if (params->verbose_prompt) { + auto tmp = common_tokenize(ctx_llava->ctx_llama, system_prompt, true, true); + for (int i = 0; i < (int) tmp.size(); i++) { + LOG_INF("%6d -> '%s'\n", tmp[i], common_token_to_piece(ctx_llava->ctx_llama, tmp[i]).c_str()); + } + } + LOG_INF("user_prompt: %s\n", user_prompt.c_str()); + if (params->verbose_prompt) { + auto tmp = common_tokenize(ctx_llava->ctx_llama, user_prompt, true, true); + for (int i = 0; i < (int) tmp.size(); i++) { + LOG_INF("%6d -> '%s'\n", tmp[i], common_token_to_piece(ctx_llava->ctx_llama, tmp[i]).c_str()); + } + } + } else { + // llava-1.5 native mode + system_prompt = "<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|>"; + user_prompt = "<|vision_end|>" + prompt + "<|im_end|>\n<|im_start|>assistant\n"; + if (params->verbose_prompt) { + auto tmp = common_tokenize(ctx_llava->ctx_llama, user_prompt, true, true); + for (int i = 0; i < (int) tmp.size(); i++) { + LOG_INF("%6d -> '%s'\n", tmp[i], common_token_to_piece(ctx_llava->ctx_llama, tmp[i]).c_str()); + } + } + } + + eval_string(ctx_llava->ctx_llama, system_prompt.c_str(), params->n_batch, &n_past, &cur_pos_id, true); + if (image_embed != nullptr) { + auto image_size = clip_get_load_image_size(ctx_llava->ctx_clip); + qwen2vl_eval_image_embed(ctx_llava->ctx_llama, image_embed, params->n_batch, &n_past, &cur_pos_id, image_size); + } + eval_string(ctx_llava->ctx_llama, user_prompt.c_str(), params->n_batch, &n_past, &cur_pos_id, false); + + // generate the response + + LOG("\n"); + + struct common_sampler * smpl = common_sampler_init(ctx_llava->model, params->sampling); + if (!smpl) { + LOG_ERR("%s: failed to initialize sampling subsystem\n", __func__); + exit(1); + } + + std::string response = ""; + for (int i = 0; i < max_tgt_len; i++) { + const char * tmp = sample(smpl, ctx_llava->ctx_llama, &n_past, &cur_pos_id); + response += tmp; + if (strcmp(tmp, "") == 0) break; + if (strstr(tmp, "###")) break; // Yi-VL behavior + LOG("%s", tmp); + if (strstr(response.c_str(), "<|im_end|>")) break; // Yi-34B llava-1.6 - for some reason those decode not as the correct token (tokenizer works) + if (strstr(response.c_str(), "<|im_start|>")) break; // Yi-34B llava-1.6 + if (strstr(response.c_str(), "USER:")) break; // mistral llava-1.6 + + fflush(stdout); + } + + common_sampler_free(smpl); + LOG("\n"); +} + +static struct llama_model * llava_init(common_params * params) { + llama_backend_init(); + llama_numa_init(params->numa); + + llama_model_params model_params = common_model_params_to_llama(*params); + + llama_model * model = llama_model_load_from_file(params->model.c_str(), model_params); + if (model == NULL) { + LOG_ERR("%s: unable to load model\n" , __func__); + return NULL; + } + return model; +} + +static struct llava_context * llava_init_context(common_params * params, llama_model * model) { + const char * clip_path = params->mmproj.c_str(); + + auto prompt = params->prompt; + if (prompt.empty()) { + prompt = "describe the image in detail."; + } + + auto ctx_clip = clip_model_load(clip_path, /*verbosity=*/ 1); + + llama_context_params ctx_params = common_context_params_to_llama(*params); + ctx_params.n_ctx = params->n_ctx < 2048 ? 2048 : params->n_ctx; // we need a longer context size to process image embeddings + + llama_context * ctx_llama = llama_init_from_model(model, ctx_params); + + if (ctx_llama == NULL) { + LOG_ERR("%s: failed to create the llama_context\n" , __func__); + return NULL; + } + + auto * ctx_llava = (struct llava_context *)malloc(sizeof(llava_context)); + + ctx_llava->ctx_llama = ctx_llama; + ctx_llava->ctx_clip = ctx_clip; + ctx_llava->model = model; + return ctx_llava; +} + +static void llava_free(struct llava_context * ctx_llava) { + if (ctx_llava->ctx_clip) { + clip_free(ctx_llava->ctx_clip); + ctx_llava->ctx_clip = NULL; + } + + llama_free(ctx_llava->ctx_llama); + llama_model_free(ctx_llava->model); + llama_backend_free(); +} + +#ifndef NDEBUG + +static void debug_test_mrope_2d() { + // 1. Initialize backend + ggml_backend_t backend = NULL; + std::string backend_name = ""; +#ifdef GGML_USE_CUDA + fprintf(stderr, "%s: using CUDA backend\n", __func__); + backend = ggml_backend_cuda_init(0); // init device 0 + backend_name = "cuda"; + if (!backend) { + fprintf(stderr, "%s: ggml_backend_cuda_init() failed\n", __func__); + } +#endif + // if there aren't GPU Backends fallback to CPU backend + if (!backend) { + backend = ggml_backend_cpu_init(); + backend_name = "cpu"; + } + + // Calculate the size needed to allocate + size_t ctx_size = 0; + ctx_size += 2 * ggml_tensor_overhead(); // tensors + // no need to allocate anything else! + + // 2. Allocate `ggml_context` to store tensor data + struct ggml_init_params params = { + /*.mem_size =*/ ctx_size, + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, // the tensors will be allocated later by ggml_backend_alloc_ctx_tensors() + }; + struct ggml_context * ctx = ggml_init(params); + + struct ggml_tensor * inp_raw = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, 128, 12, 30); + ggml_set_name(inp_raw, "inp_raw"); + ggml_set_input(inp_raw); + + struct ggml_tensor * pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 30 * 4); + ggml_set_name(pos, "pos"); + ggml_set_input(pos); + + std::vector dummy_q; + dummy_q.resize(128 * 12 * 30); + std::fill(dummy_q.begin(), dummy_q.end(), 0.1); + // memcpy(inp_raw->data, dummy_q.data(), 128 * 12 * 30 * ggml_element_size(inp_raw)); + + std::vector pos_id; + pos_id.resize(30 * 4); + for (int i = 0; i < 30; i ++) { + pos_id[i] = i; + pos_id[i + 30] = i + 10; + pos_id[i + 60] = i + 20; + pos_id[i + 90] = i + 30; + } + int sections[4] = {32, 32, 0, 0}; + + // 4. Allocate a `ggml_backend_buffer` to store all tensors + ggml_backend_buffer_t buffer = ggml_backend_alloc_ctx_tensors(ctx, backend); + + // 5. Copy tensor data from main memory (RAM) to backend buffer + ggml_backend_tensor_set(inp_raw, dummy_q.data(), 0, ggml_nbytes(inp_raw)); + ggml_backend_tensor_set(pos, pos_id.data(), 0, ggml_nbytes(pos)); + + // 6. Create a `ggml_cgraph` for mul_mat operation + struct ggml_cgraph * gf = NULL; + struct ggml_context * ctx_cgraph = NULL; + + // create a temporally context to build the graph + struct ggml_init_params params0 = { + /*.mem_size =*/ ggml_tensor_overhead()*GGML_DEFAULT_GRAPH_SIZE + ggml_graph_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, // the tensors will be allocated later by ggml_gallocr_alloc_graph() + }; + ctx_cgraph = ggml_init(params0); + gf = ggml_new_graph(ctx_cgraph); + + struct ggml_tensor * result0 = ggml_rope_multi( + ctx_cgraph, inp_raw, pos, nullptr, + 128/2, sections, LLAMA_ROPE_TYPE_VISION, 32768, 1000000, 1, + 0, 1, 32, 1); + + // Add "result" tensor and all of its dependencies to the cgraph + ggml_build_forward_expand(gf, result0); + + // 7. Create a `ggml_gallocr` for cgraph computation + ggml_gallocr_t allocr = ggml_gallocr_new(ggml_backend_get_default_buffer_type(backend)); + ggml_gallocr_alloc_graph(allocr, gf); + + // 9. Run the computation + int n_threads = 1; // Optional: number of threads to perform some operations with multi-threading + if (ggml_backend_is_cpu(backend)) { + ggml_backend_cpu_set_n_threads(backend, n_threads); + } + ggml_backend_graph_compute(backend, gf); + + // 10. Retrieve results (output tensors) + // in this example, output tensor is always the last tensor in the graph + struct ggml_tensor * result = result0; + // struct ggml_tensor * result = gf->nodes[gf->n_nodes - 1]; + float * result_data = (float *)malloc(ggml_nbytes(result)); + // because the tensor data is stored in device buffer, we need to copy it back to RAM + ggml_backend_tensor_get(result, result_data, 0, ggml_nbytes(result)); + const std::string bin_file = "mrope_2d_" + backend_name +".bin"; + std::ofstream outFile(bin_file, std::ios::binary); + + if (outFile.is_open()) { + outFile.write(reinterpret_cast(result_data), ggml_nbytes(result)); + outFile.close(); + std::cout << "Data successfully written to " + bin_file << std::endl; + } else { + std::cerr << "Error opening file!" << std::endl; + } + + free(result_data); + // 11. Free memory and exit + ggml_free(ctx_cgraph); + ggml_gallocr_free(allocr); + ggml_free(ctx); + ggml_backend_buffer_free(buffer); + ggml_backend_free(backend); +} + +static void debug_dump_img_embed(struct llava_context * ctx_llava) { + int n_embd = llama_model_n_embd(llama_get_model(ctx_llava->ctx_llama)); + int ne = n_embd * 4; + float vals[56 * 56 * 3]; + // float embd[ne]; + std::vector embd; + embd.resize(ne); + + for (int i = 0; i < 56*56; i++) + { + for (int c = 0; c < 3; c++) + vals[i * 3 + c] = (float)(i % (56 * 56)) / (56*56); + } + + clip_encode_float_image(ctx_llava->ctx_clip, 16, vals, 56, 56, embd.data()); + + std::ofstream outFile("img_embed.bin", std::ios::binary); + if (outFile.is_open()) { + outFile.write(reinterpret_cast(embd.data()), ne * sizeof(float)); + + outFile.close(); + std::cout << "Data successfully written to mrope.bin" << std::endl; + } else { + std::cerr << "Error opening file!" << std::endl; + } +} + +#endif + + +int main(int argc, char ** argv) { + ggml_time_init(); + + common_params params; + + if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_LLAVA, print_usage)) { + return 1; + } + + common_init(); + + if (params.mmproj.empty() || (params.image.empty() && !prompt_contains_image(params.prompt))) { + print_usage(argc, argv); + return 1; + } + + auto * model = llava_init(¶ms); + if (model == NULL) { + fprintf(stderr, "%s: error: failed to init llava model\n", __func__); + return 1; + } + + if (prompt_contains_image(params.prompt)) { + auto * ctx_llava = llava_init_context(¶ms, model); + + auto * image_embed = load_image(ctx_llava, ¶ms, ""); + + // process the prompt + process_prompt(ctx_llava, image_embed, ¶ms, params.prompt); + + llama_perf_context_print(ctx_llava->ctx_llama); + llava_image_embed_free(image_embed); + ctx_llava->model = NULL; + llava_free(ctx_llava); +#ifndef NDEBUG + } else if (params.image[0].empty()) { + auto ctx_llava = llava_init_context(¶ms, model); + + debug_test_mrope_2d(); + debug_dump_img_embed(ctx_llava); + + llama_perf_context_print(ctx_llava->ctx_llama); + ctx_llava->model = NULL; + llava_free(ctx_llava); +#endif + } else { + for (auto & image : params.image) { + auto * ctx_llava = llava_init_context(¶ms, model); + + auto * image_embed = load_image(ctx_llava, ¶ms, image); + if (!image_embed) { + LOG_ERR("%s: failed to load image %s. Terminating\n\n", __func__, image.c_str()); + return 1; + } + + // process the prompt + process_prompt(ctx_llava, image_embed, ¶ms, params.prompt); + + llama_perf_context_print(ctx_llava->ctx_llama); + llava_image_embed_free(image_embed); + ctx_llava->model = NULL; + llava_free(ctx_llava); + } + } + + llama_model_free(model); + + return 0; +} diff --git a/examples/lookahead/CMakeLists.txt b/examples/lookahead/CMakeLists.txt index f0ae5cd89..346861314 100644 --- a/examples/lookahead/CMakeLists.txt +++ b/examples/lookahead/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-lookahead) add_executable(${TARGET} lookahead.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/lookahead/lookahead.cpp b/examples/lookahead/lookahead.cpp index 3c0ccfea2..2f0898e62 100644 --- a/examples/lookahead/lookahead.cpp +++ b/examples/lookahead/lookahead.cpp @@ -58,8 +58,10 @@ int main(int argc, char ** argv) { // load the target model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + + const llama_vocab * vocab = llama_model_get_vocab(model); // Tokenize the prompt std::vector inp; @@ -115,7 +117,7 @@ int main(int argc, char ** argv) { llama_batch batch = llama_batch_init(params.n_ctx, 0, W + G + 1); // target model sampling context - struct common_sampler * smpl = common_sampler_init(model, params.sparams); + struct common_sampler * smpl = common_sampler_init(model, params.sampling); // verification n-grams std::vector ngrams_cur(G); @@ -147,7 +149,7 @@ int main(int argc, char ** argv) { } // here we keep adding new n-grams as we go - ngram_container ngrams_observed(llama_n_vocab(model), N, G); + ngram_container ngrams_observed(llama_vocab_n_tokens(vocab), N, G); // debug struct llama_kv_cache_view kvc_view = llama_kv_cache_view_init(ctx, W + G + 1); @@ -297,7 +299,7 @@ int main(int argc, char ** argv) { } fflush(stdout); - if (llama_token_is_eog(model, id)) { + if (llama_vocab_is_eog(vocab, id)) { has_eos = true; } @@ -474,9 +476,6 @@ int main(int argc, char ** argv) { llama_batch_free(batch); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); LOG("\n\n"); diff --git a/examples/lookup/CMakeLists.txt b/examples/lookup/CMakeLists.txt index ef19fe25e..fba78ceda 100644 --- a/examples/lookup/CMakeLists.txt +++ b/examples/lookup/CMakeLists.txt @@ -2,22 +2,22 @@ set(TARGET llama-lookup) add_executable(${TARGET} lookup.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TARGET llama-lookup-create) add_executable(${TARGET} lookup-create.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TARGET llama-lookup-merge) add_executable(${TARGET} lookup-merge.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TARGET llama-lookup-stats) add_executable(${TARGET} lookup-stats.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/lookup/lookup-create.cpp b/examples/lookup/lookup-create.cpp index 7ced0aa97..3da45ed9e 100644 --- a/examples/lookup/lookup-create.cpp +++ b/examples/lookup/lookup-create.cpp @@ -1,14 +1,9 @@ #include "arg.h" #include "common.h" #include "ngram-cache.h" -#include "ggml.h" #include "llama.h" -#include -#include -#include #include -#include #include int main(int argc, char ** argv){ @@ -25,16 +20,16 @@ int main(int argc, char ** argv){ // load the model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model_ptr & model = llama_init.model; + llama_context_ptr & ctx = llama_init.context; + GGML_ASSERT(model != nullptr); // tokenize the prompt std::vector inp; - inp = common_tokenize(ctx, params.prompt, true, true); + inp = common_tokenize(ctx.get(), params.prompt, true, true); fprintf(stderr, "%s: tokenization done\n", __func__); - common_ngram_cache ngram_cache; common_ngram_cache_update(ngram_cache, LLAMA_NGRAM_STATIC, LLAMA_NGRAM_STATIC, inp, inp.size(), true); fprintf(stderr, "%s: hashing done, writing file to %s\n", __func__, params.lookup_cache_static.c_str()); diff --git a/examples/lookup/lookup-stats.cpp b/examples/lookup/lookup-stats.cpp index 7faebe7ba..fcb289abe 100644 --- a/examples/lookup/lookup-stats.cpp +++ b/examples/lookup/lookup-stats.cpp @@ -21,7 +21,7 @@ int main(int argc, char ** argv){ common_init(); - const int n_draft = params.n_draft; + const int n_draft = params.speculative.n_max; // init llama.cpp llama_backend_init(); @@ -30,16 +30,16 @@ int main(int argc, char ** argv){ // load the model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_context_ptr & ctx = llama_init.context; // tokenize the prompt std::vector inp; - inp = common_tokenize(ctx, params.prompt, true, true); + inp = common_tokenize(ctx.get(), params.prompt, true, true); common_ngram_cache ngram_cache_context; common_ngram_cache ngram_cache_dynamic; common_ngram_cache ngram_cache_static; + int64_t t_draft_flat_us = 0; int64_t t_draft_us = 0; @@ -65,7 +65,7 @@ int main(int argc, char ** argv){ } const int n_input = inp.size(); - const int n_ctx = llama_n_ctx(ctx); + const int n_ctx = llama_n_ctx(ctx.get()); int n_drafted = 0; int n_accept = 0; @@ -149,9 +149,6 @@ int main(int argc, char ** argv){ LOG_INF("n_accept = %d\n", n_accept); LOG_INF("accept = %.3f%%\n", 100.0f * n_accept / n_drafted); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); LOG("\n\n"); diff --git a/examples/lookup/lookup.cpp b/examples/lookup/lookup.cpp index a04728b18..dbd0444ec 100644 --- a/examples/lookup/lookup.cpp +++ b/examples/lookup/lookup.cpp @@ -22,7 +22,7 @@ int main(int argc, char ** argv){ common_init(); // max. number of additional tokens to draft if match is found - const int n_draft = params.n_draft; + const int n_draft = params.speculative.n_max; const bool dump_kv_cache = params.dump_kv_cache; @@ -33,8 +33,10 @@ int main(int argc, char ** argv){ // load the model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + + const llama_vocab * vocab = llama_model_get_vocab(model); // tokenize the prompt std::vector inp; @@ -102,7 +104,7 @@ int main(int argc, char ** argv){ bool has_eos = false; - struct common_sampler * smpl = common_sampler_init(model, params.sparams); + struct common_sampler * smpl = common_sampler_init(model, params.sampling); std::vector draft; @@ -136,7 +138,7 @@ int main(int argc, char ** argv){ LOG("%s", token_str.c_str()); } - if (llama_token_is_eog(model, id)) { + if (llama_vocab_is_eog(vocab, id)) { has_eos = true; } @@ -243,9 +245,6 @@ int main(int argc, char ** argv){ llama_batch_free(batch_tgt); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); LOG("\n\n"); diff --git a/examples/main-cmake-pkg/CMakeLists.txt b/examples/main-cmake-pkg/CMakeLists.txt index 3b38db292..5563f4de0 100644 --- a/examples/main-cmake-pkg/CMakeLists.txt +++ b/examples/main-cmake-pkg/CMakeLists.txt @@ -29,4 +29,4 @@ add_executable(${TARGET} ${CMAKE_CURRENT_LIST_DIR}/../main/main.cpp) target_include_directories(${TARGET} PRIVATE ${_common_path}) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/main/CMakeLists.txt b/examples/main/CMakeLists.txt index 5f6efaa9a..af3d9150f 100644 --- a/examples/main/CMakeLists.txt +++ b/examples/main/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-cli) add_executable(${TARGET} main.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/main/README.md b/examples/main/README.md index 145216938..17d80a622 100644 --- a/examples/main/README.md +++ b/examples/main/README.md @@ -66,7 +66,7 @@ In this section, we cover the most commonly used options for running the `llama- - `-mu MODEL_URL --model-url MODEL_URL`: Specify a remote http url to download the file (e.g [https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true](https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true)). - `-i, --interactive`: Run the program in interactive mode, allowing you to provide input directly and receive real-time responses. - `-n N, --n-predict N`: Set the number of tokens to predict when generating text. Adjusting this value can influence the length of the generated text. -- `-c N, --ctx-size N`: Set the size of the prompt context. The default is 512, but LLaMA models were built with a context of 2048, which will provide better results for longer input/inference. +- `-c N, --ctx-size N`: Set the size of the prompt context. The default is 4096, but if a LLaMA model was built with a longer context, increasing this value will provide better results for longer input/inference. - `-mli, --multiline-input`: Allows you to write or paste multiple lines without ending each in '\' - `-t N, --threads N`: Set the number of threads to use during generation. For optimal performance, it is recommended to set this value to the number of physical CPU cores your system has. - `-ngl N, --n-gpu-layers N`: When compiled with GPU support, this option allows offloading some layers to the GPU for computation. Generally results in increased performance. @@ -131,7 +131,7 @@ During text generation, LLaMA models have a limited context size, which means th ### Context Size -- `-c N, --ctx-size N`: Set the size of the prompt context (default: 0, 0 = loaded from model). The LLaMA models were built with a context of 2048-8192, which will yield the best results on longer input/inference. +- `-c N, --ctx-size N`: Set the size of the prompt context (default: 4096, 0 = loaded from model). If a LLaMA model was built with a longer context, increasing this value will yield the best results on longer input/inference. ### Extended Context Size @@ -177,16 +177,11 @@ Example usage: `--temp 0` - `--repeat-penalty N`: Control the repetition of token sequences in the generated text default: 1.0, 1.0 = disabled). - `--repeat-last-n N`: Last n tokens to consider for penalizing repetition (default: 64, 0 = disabled, -1 = ctx-size). -- `--no-penalize-nl`: Disable penalization for newline tokens when applying the repeat penalty. The `repeat-penalty` option helps prevent the model from generating repetitive or monotonous text. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. The default value is 1. The `repeat-last-n` option controls the number of tokens in the history to consider for penalizing repetition. A larger value will look further back in the generated text to prevent repetitions, while a smaller value will only consider recent tokens. A value of 0 disables the penalty, and a value of -1 sets the number of tokens considered equal to the context size (`ctx-size`). -Use the `--no-penalize-nl` option to disable newline penalization when applying the repeat penalty. This option is particularly useful for generating chat conversations, dialogues, code, poetry, or any text where newline tokens play a significant role in structure and formatting. Disabling newline penalization helps maintain the natural flow and intended formatting in these specific use cases. - -Example usage: `--repeat-penalty 1.15 --repeat-last-n 128 --no-penalize-nl` - ### DRY Repetition Penalty DRY (Don't Repeat Yourself) sampling is an effective technique for reducing repetition in generated text even across long contexts by penalizing tokens based on their recent usage patterns (original [PR link](https://github.com/oobabooga/text-generation-webui/pull/5677)). @@ -348,6 +343,7 @@ These options provide extra functionality and customization when running the LLa - `-h, --help`: Display a help message showing all available options and their default values. This is particularly useful for checking the latest options and default values, as they can change frequently, and the information in this document may become outdated. - `--verbose-prompt`: Print the prompt before generating text. +- `--no-display-prompt`: Don't print prompt at generation. - `-mg i, --main-gpu i`: When using multiple GPUs this option controls which GPU is used for small tensors for which the overhead of splitting the computation across all GPUs is not worthwhile. The GPU in question will use slightly more VRAM to store a scratch buffer for temporary results. By default GPU 0 is used. - `-ts SPLIT, --tensor-split SPLIT`: When using multiple GPUs this option controls how large tensors should be split across all GPUs. `SPLIT` is a comma-separated list of non-negative values that assigns the proportion of data that each GPU should get in order. For example, "3,2" will assign 60% of the data to GPU 0 and 40% to GPU 1. By default the data is split in proportion to VRAM but this may not be optimal for performance. - `-hfr URL --hf-repo URL`: The url to the Hugging Face model repository. Used in conjunction with `--hf-file` or `-hff`. The model is downloaded and stored in the file provided by `-m` or `--model`. If `-m` is not provided, the model is auto-stored in the path specified by the `LLAMA_CACHE` environment variable or in an OS-specific local cache. diff --git a/examples/main/main.cpp b/examples/main/main.cpp index 374ed47ad..39666a0e8 100644 --- a/examples/main/main.cpp +++ b/examples/main/main.cpp @@ -5,7 +5,6 @@ #include "sampling.h" #include "llama.h" -#include #include #include #include @@ -31,6 +30,8 @@ #pragma warning(disable: 4244 4267) // possible loss of data #endif +static const char * DEFAULT_SYSTEM_MESSAGE = "You are a helpful assistant"; + static llama_context ** g_ctx; static llama_model ** g_model; static common_sampler ** g_smpl; @@ -62,49 +63,6 @@ static bool file_is_empty(const std::string & path) { return f.tellg() == 0; } -static void write_logfile( - const llama_context * ctx, const common_params & params, const llama_model * model, - const std::vector & input_tokens, const std::string & output, - const std::vector & output_tokens -) { - if (params.logdir.empty()) { - return; - } - - const std::string timestamp = string_get_sortable_timestamp(); - - const bool success = fs_create_directory_with_parents(params.logdir); - if (!success) { - LOG_ERR("%s: failed to create logdir %s, cannot write logfile\n", __func__, params.logdir.c_str()); - return; - } - - const std::string logfile_path = params.logdir + timestamp + ".yml"; - FILE * logfile = fopen(logfile_path.c_str(), "w"); - - if (logfile == NULL) { - LOG_ERR("%s: failed to open logfile %s\n", __func__, logfile_path.c_str()); - return; - } - - fprintf(logfile, "binary: main\n"); - char model_desc[128]; - llama_model_desc(model, model_desc, sizeof(model_desc)); - yaml_dump_non_result_info(logfile, params, ctx, timestamp, input_tokens, model_desc); - - fprintf(logfile, "\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "# Generation Results #\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "\n"); - - yaml_dump_string_multiline(logfile, "output", output.c_str()); - yaml_dump_vector_int(logfile, "output_tokens", output_tokens); - - llama_perf_dump_yaml(logfile, ctx); - fclose(logfile); -} - #if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32) static void sigint_handler(int signo) { if (signo == SIGINT) { @@ -115,7 +73,6 @@ static void sigint_handler(int signo) { console::cleanup(); LOG("\n"); common_perf_print(*g_ctx, *g_smpl); - write_logfile(*g_ctx, *g_params, *g_model, *g_input_tokens, g_output_ss->str(), *g_output_tokens); // make sure all logs are flushed LOG("Interrupted by user\n"); @@ -144,7 +101,7 @@ int main(int argc, char ** argv) { common_init(); - auto & sparams = params.sparams; + auto & sparams = params.sampling; // save choice to use color for later // (note for later: this is a slightly awkward choice) @@ -189,26 +146,32 @@ int main(int argc, char ** argv) { llama_context * ctx = nullptr; common_sampler * smpl = nullptr; - std::vector chat_msgs; - g_model = &model; g_ctx = &ctx; g_smpl = &smpl; + std::vector chat_msgs; + // load the model and apply lora adapter, if any LOG_INF("%s: load the model and apply lora adapter, if any\n", __func__); common_init_result llama_init = common_init_from_params(params); - model = llama_init.model; - ctx = llama_init.context; + model = llama_init.model.get(); + ctx = llama_init.context.get(); if (model == NULL) { LOG_ERR("%s: error: unable to load model\n", __func__); return 1; } + const llama_vocab * vocab = llama_model_get_vocab(model); + LOG_INF("%s: llama threadpool init, n_threads = %d\n", __func__, (int) params.cpuparams.n_threads); + auto * reg = ggml_backend_dev_backend_reg(ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU)); + auto * ggml_threadpool_new_fn = (decltype(ggml_threadpool_new) *) ggml_backend_reg_get_proc_address(reg, "ggml_threadpool_new"); + auto * ggml_threadpool_free_fn = (decltype(ggml_threadpool_free) *) ggml_backend_reg_get_proc_address(reg, "ggml_threadpool_free"); + struct ggml_threadpool_params tpp_batch = ggml_threadpool_params_from_cpu_params(params.cpuparams_batch); struct ggml_threadpool_params tpp = @@ -218,7 +181,7 @@ int main(int argc, char ** argv) { struct ggml_threadpool * threadpool_batch = NULL; if (!ggml_threadpool_params_match(&tpp, &tpp_batch)) { - threadpool_batch = ggml_threadpool_new(&tpp_batch); + threadpool_batch = ggml_threadpool_new_fn(&tpp_batch); if (!threadpool_batch) { LOG_ERR("%s: batch threadpool create failed : n_threads %d\n", __func__, tpp_batch.n_threads); return 1; @@ -228,7 +191,7 @@ int main(int argc, char ** argv) { tpp.paused = true; } - struct ggml_threadpool * threadpool = ggml_threadpool_new(&tpp); + struct ggml_threadpool * threadpool = ggml_threadpool_new_fn(&tpp); if (!threadpool) { LOG_ERR("%s: threadpool create failed : n_threads %d\n", __func__, tpp.n_threads); return 1; @@ -236,15 +199,31 @@ int main(int argc, char ** argv) { llama_attach_threadpool(ctx, threadpool, threadpool_batch); - const int n_ctx_train = llama_n_ctx_train(model); + const int n_ctx_train = llama_model_n_ctx_train(model); const int n_ctx = llama_n_ctx(ctx); if (n_ctx > n_ctx_train) { LOG_WRN("%s: model was trained on only %d context tokens (%d specified)\n", __func__, n_ctx_train, n_ctx); } + // auto enable conversation mode if chat template is available + const bool has_chat_template = !common_get_builtin_chat_template(model).empty() || !params.chat_template.empty(); + if (params.conversation_mode == COMMON_CONVERSATION_MODE_AUTO) { + if (has_chat_template) { + LOG_INF("%s: chat template is available, enabling conversation mode (disable it with -no-cnv)\n", __func__); + params.conversation_mode = COMMON_CONVERSATION_MODE_ENABLED; + } else { + params.conversation_mode = COMMON_CONVERSATION_MODE_DISABLED; + } + } + + // in case user force-activate conversation mode (via -cnv) without proper chat template, we show a warning + if (params.conversation_mode && !has_chat_template) { + LOG_WRN("%s: chat template is not available or is not supported. This may cause the model to output suboptimal responses\n", __func__); + } + // print chat template example in conversation mode - if (params.conversation) { + if (params.conversation_mode) { if (params.enable_chat_template) { LOG_INF("%s: chat template example:\n%s\n", __func__, common_chat_format_example(model, params.chat_template).c_str()); } else { @@ -281,9 +260,9 @@ int main(int argc, char ** argv) { } } - const bool add_bos = llama_add_bos_token(model); + const bool add_bos = llama_vocab_get_add_bos(vocab); if (!llama_model_has_encoder(model)) { - GGML_ASSERT(!llama_add_eos_token(model)); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); } LOG_DBG("n_ctx: %d, add_bos: %d\n", n_ctx, add_bos); @@ -291,8 +270,10 @@ int main(int argc, char ** argv) { std::vector embd_inp; { - auto prompt = (params.conversation && params.enable_chat_template && !params.prompt.empty()) - ? chat_add_and_format(model, chat_msgs, "system", params.prompt) // format the system prompt in conversation mode + auto prompt = (params.conversation_mode && params.enable_chat_template) + // format the system prompt in conversation mode (fallback to default if empty) + ? chat_add_and_format(model, chat_msgs, "system", params.prompt.empty() ? DEFAULT_SYSTEM_MESSAGE : params.prompt) + // otherwise use the prompt as is : params.prompt; if (params.interactive_first || !params.prompt.empty() || session_tokens.empty()) { LOG_DBG("tokenize the prompt\n"); @@ -309,7 +290,7 @@ int main(int argc, char ** argv) { // Should not run without any tokens if (embd_inp.empty()) { if (add_bos) { - embd_inp.push_back(llama_token_bos(model)); + embd_inp.push_back(llama_vocab_bos(vocab)); LOG_WRN("embd_inp was considered empty and bos was added: %s\n", string_from(ctx, embd_inp).c_str()); } else { LOG_ERR("input is empty\n"); @@ -366,7 +347,7 @@ int main(int argc, char ** argv) { params.n_keep += add_bos; // always keep the BOS token } - if (params.conversation) { + if (params.conversation_mode) { params.interactive_first = true; } @@ -490,7 +471,11 @@ int main(int argc, char ** argv) { #if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32) LOG_INF( " - Press Ctrl+C to interject at any time.\n"); #endif - LOG_INF( "%s\n", control_message); + LOG_INF( "%s", control_message); + if (params.conversation_mode && params.enable_chat_template && params.prompt.empty()) { + LOG_INF( " - Using default system message. To change it, set a different value via -p PROMPT or -f FILE argument.\n"); + } + LOG_INF("\n"); is_interacting = params.interactive_first; } @@ -534,8 +519,8 @@ int main(int argc, char ** argv) { } llama_token decoder_start_token_id = llama_model_decoder_start_token(model); - if (decoder_start_token_id == -1) { - decoder_start_token_id = llama_token_bos(model); + if (decoder_start_token_id == LLAMA_TOKEN_NULL) { + decoder_start_token_id = llama_vocab_bos(vocab); } embd_inp.clear(); @@ -782,7 +767,7 @@ int main(int argc, char ** argv) { } // deal with end of generation tokens in interactive mode - if (llama_token_is_eog(model, common_sampler_last(smpl))) { + if (llama_vocab_is_eog(vocab, common_sampler_last(smpl))) { LOG_DBG("found an EOG token\n"); if (params.interactive) { @@ -802,7 +787,7 @@ int main(int argc, char ** argv) { } // if current token is not EOG, we add it to current assistant message - if (params.conversation) { + if (params.conversation_mode) { const auto id = common_sampler_last(smpl); assistant_ss << common_token_to_piece(ctx, id, false); } @@ -810,17 +795,17 @@ int main(int argc, char ** argv) { if (n_past > 0 && is_interacting) { LOG_DBG("waiting for user input\n"); - if (params.conversation) { + if (params.conversation_mode) { LOG("\n> "); } if (params.input_prefix_bos) { LOG_DBG("adding input prefix BOS token\n"); - embd_inp.push_back(llama_token_bos(model)); + embd_inp.push_back(llama_vocab_bos(vocab)); } std::string buffer; - if (!params.input_prefix.empty() && !params.conversation) { + if (!params.input_prefix.empty() && !params.conversation_mode) { LOG_DBG("appending input prefix: '%s'\n", params.input_prefix.c_str()); LOG("%s", params.input_prefix.c_str()); } @@ -844,7 +829,7 @@ int main(int argc, char ** argv) { // Entering a empty line lets the user pass control back if (buffer.length() > 1) { // append input suffix if any - if (!params.input_suffix.empty() && !params.conversation) { + if (!params.input_suffix.empty() && !params.conversation_mode) { LOG_DBG("appending input suffix: '%s'\n", params.input_suffix.c_str()); LOG("%s", params.input_suffix.c_str()); } @@ -857,7 +842,7 @@ int main(int argc, char ** argv) { string_process_escapes(buffer); } - bool format_chat = params.conversation && params.enable_chat_template; + bool format_chat = params.conversation_mode && params.enable_chat_template; std::string user_inp = format_chat ? chat_add_and_format(model, chat_msgs, "user", std::move(buffer)) : std::move(buffer); @@ -870,8 +855,8 @@ int main(int argc, char ** argv) { // if user stop generation mid-way, we must add EOT to finish model's last response if (need_insert_eot && format_chat) { - llama_token eot = llama_token_eot(model); - embd_inp.push_back(eot == -1 ? llama_token_eos(model) : eot); + llama_token eot = llama_vocab_eot(vocab); + embd_inp.push_back(eot == LLAMA_TOKEN_NULL ? llama_vocab_eos(vocab) : eot); need_insert_eot = false; } @@ -906,7 +891,7 @@ int main(int argc, char ** argv) { } // end of generation - if (!embd.empty() && llama_token_is_eog(model, embd.back()) && !(params.interactive)) { + if (!embd.empty() && llama_vocab_is_eog(vocab, embd.back()) && !(params.interactive)) { LOG(" [end of text]\n"); break; } @@ -926,17 +911,13 @@ int main(int argc, char ** argv) { LOG("\n\n"); common_perf_print(ctx, smpl); - write_logfile(ctx, params, model, input_tokens, output_ss.str(), output_tokens); common_sampler_free(smpl); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); - ggml_threadpool_free(threadpool); - ggml_threadpool_free(threadpool_batch); + ggml_threadpool_free_fn(threadpool); + ggml_threadpool_free_fn(threadpool_batch); return 0; } diff --git a/examples/parallel/CMakeLists.txt b/examples/parallel/CMakeLists.txt index c13557bac..847e916de 100644 --- a/examples/parallel/CMakeLists.txt +++ b/examples/parallel/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-parallel) add_executable(${TARGET} parallel.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/parallel/parallel.cpp b/examples/parallel/parallel.cpp index 43c8f3ed5..7ef43d5e1 100644 --- a/examples/parallel/parallel.cpp +++ b/examples/parallel/parallel.cpp @@ -132,8 +132,10 @@ int main(int argc, char ** argv) { // load the target model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + + const llama_vocab * vocab = llama_model_get_vocab(model); // load the prompts from an external file if there are any if (params.prompt.empty()) { @@ -160,7 +162,7 @@ int main(int argc, char ** argv) { for (size_t i = 0; i < clients.size(); ++i) { auto & client = clients[i]; client.id = i; - client.smpl = common_sampler_init(model, params.sparams); + client.smpl = common_sampler_init(model, params.sampling); } std::vector tokens_system; @@ -358,7 +360,7 @@ int main(int argc, char ** argv) { // client.id, client.seq_id, id, client.n_decoded, client.i_batch, token_str.c_str()); if (client.n_decoded > 2 && - (llama_token_is_eog(model, id) || + (llama_vocab_is_eog(vocab, id) || (params.n_predict > 0 && client.n_decoded + client.n_prompt >= params.n_predict) || client.response.find("User:") != std::string::npos || client.response.find('\n') != std::string::npos)) { @@ -416,9 +418,6 @@ int main(int argc, char ** argv) { llama_batch_free(batch); - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); LOG("\n\n"); diff --git a/examples/passkey/CMakeLists.txt b/examples/passkey/CMakeLists.txt index dc467a5d3..9bc5110c2 100644 --- a/examples/passkey/CMakeLists.txt +++ b/examples/passkey/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-passkey) add_executable(${TARGET} passkey.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/passkey/passkey.cpp b/examples/passkey/passkey.cpp index 09bba708f..5953928d4 100644 --- a/examples/passkey/passkey.cpp +++ b/examples/passkey/passkey.cpp @@ -63,22 +63,24 @@ int main(int argc, char ** argv) { llama_model_params model_params = common_model_params_to_llama(params); - llama_model * model = llama_load_model_from_file(params.model.c_str(), model_params); + llama_model * model = llama_model_load_from_file(params.model.c_str(), model_params); if (model == NULL) { LOG_ERR("%s: unable to load model\n" , __func__); return 1; } + const llama_vocab * vocab = llama_model_get_vocab(model); + // initialize the context llama_context_params ctx_params = common_context_params_to_llama(params); - ctx_params.n_ctx = llama_n_ctx_train(model)*n_grp + n_keep; + ctx_params.n_ctx = llama_model_n_ctx_train(model)*n_grp + n_keep; GGML_ASSERT(ctx_params.n_batch % n_grp == 0 && "n_batch must be divisible by n_grp"); - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); if (ctx == NULL) { LOG_ERR("%s: failed to create the llama_context\n" , __func__); return 1; @@ -223,7 +225,7 @@ int main(int argc, char ** argv) { const llama_token new_token_id = llama_sampler_sample(smpl, ctx, batch.n_tokens - 1); // is it an end of generation? - if (llama_token_is_eog(model, new_token_id) || n_cur == n_len) { + if (llama_vocab_is_eog(vocab, new_token_id) || n_cur == n_len) { LOG("\n"); break; @@ -266,7 +268,7 @@ int main(int argc, char ** argv) { llama_batch_free(batch); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); llama_backend_free(); diff --git a/examples/perplexity/CMakeLists.txt b/examples/perplexity/CMakeLists.txt index be0f2fd02..3e6864093 100644 --- a/examples/perplexity/CMakeLists.txt +++ b/examples/perplexity/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-perplexity) add_executable(${TARGET} perplexity.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/perplexity/perplexity.cpp b/examples/perplexity/perplexity.cpp index e803ff143..9bf6c5743 100644 --- a/examples/perplexity/perplexity.cpp +++ b/examples/perplexity/perplexity.cpp @@ -34,55 +34,6 @@ struct results_log_softmax { float prob; }; -static void write_logfile( - const llama_context * ctx, const common_params & params, const llama_model * model, - const struct results_perplexity & results -) { - if (params.logdir.empty()) { - return; - } - - if (params.hellaswag) { - LOG_WRN("%s: logging results is not implemented for HellaSwag. No files will be written.\n", __func__); - return; - } - - const std::string timestamp = string_get_sortable_timestamp(); - - const bool success = fs_create_directory_with_parents(params.logdir); - if (!success) { - LOG_WRN("%s: failed to create logdir %s, cannot write logfile\n", - __func__, params.logdir.c_str()); - return; - } - - const std::string logfile_path = params.logdir + timestamp + ".yml"; - FILE * logfile = fopen(logfile_path.c_str(), "w"); - - if (logfile == NULL) { - LOG_ERR("%s: failed to open logfile %s\n", __func__, logfile_path.c_str()); - return; - } - - fprintf(logfile, "binary: main\n"); - char model_desc[128]; - llama_model_desc(model, model_desc, sizeof(model_desc)); - yaml_dump_non_result_info(logfile, params, ctx, timestamp, results.tokens, model_desc); - - fprintf(logfile, "\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "# Perplexity Results #\n"); - fprintf(logfile, "######################\n"); - fprintf(logfile, "\n"); - - yaml_dump_vector_float(logfile, "logits", results.logits); - fprintf(logfile, "ppl_value: %f\n", results.ppl_value); - yaml_dump_vector_float(logfile, "probs", results.probs); - - llama_perf_dump_yaml(logfile, ctx); - fclose(logfile); -} - static std::vector softmax(const std::vector& logits) { std::vector probs(logits.size()); float max_logit = logits[0]; @@ -345,8 +296,11 @@ static results_perplexity perplexity_v2(llama_context * ctx, const common_params // Output: `perplexity: 13.5106 [114/114]` // BOS tokens will be added for each chunk before eval - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); - GGML_ASSERT(!llama_add_eos_token(llama_get_model(ctx))); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const bool add_bos = llama_vocab_get_add_bos(vocab); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); LOG_INF("%s: tokenizing the input ..\n", __func__); @@ -387,7 +341,7 @@ static results_perplexity perplexity_v2(llama_context * ctx, const common_params const int n_chunk = params.n_chunks < 0 ? n_chunk_max : std::min(params.n_chunks, n_chunk_max); const int n_batch = params.n_batch; - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); int count = 0; double nll = 0.0; @@ -431,7 +385,7 @@ static results_perplexity perplexity_v2(llama_context * ctx, const common_params // add BOS token for the first batch of each chunk if (add_bos && j == 0) { - tokens[batch_start] = llama_token_bos(llama_get_model(ctx)); + tokens[batch_start] = llama_vocab_bos(vocab); } const auto * batch_logits = llama_get_logits(ctx); @@ -493,8 +447,11 @@ static results_perplexity perplexity(llama_context * ctx, const common_params & // Output: `perplexity: 13.5106 [114/114]` // BOS tokens will be added for each chunk before eval - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); - GGML_ASSERT(!llama_add_eos_token(llama_get_model(ctx))); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const bool add_bos = llama_vocab_get_add_bos(vocab); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); std::ofstream logits_stream; if (!params.logits_file.empty()) { @@ -534,7 +491,7 @@ static results_perplexity perplexity(llama_context * ctx, const common_params & const int n_chunk = params.n_chunks < 0 ? n_chunk_max : std::min(params.n_chunks, n_chunk_max); const int n_batch = params.n_batch; - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); int count = 0; double nll = 0.0; @@ -606,7 +563,7 @@ static results_perplexity perplexity(llama_context * ctx, const common_params & // add BOS token for the first batch of each chunk if (add_bos && j == 0) { - tokens[seq_start] = llama_token_bos(llama_get_model(ctx)); + tokens[seq_start] = llama_vocab_bos(vocab); } for (int k = 0; k < batch_size; ++k) { @@ -781,6 +738,9 @@ static void compute_logprobs(const float * batch_logits, int n_vocab, std::vecto } static void hellaswag_score(llama_context * ctx, const common_params & params) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + // Calculates hellaswag score (acc_norm) from prompt // // Data extracted from the HellaSwag validation dataset (MIT license) https://github.com/rowanz/hellaswag/blob/master/data/hellaswag_val.jsonl @@ -814,7 +774,7 @@ static void hellaswag_score(llama_context * ctx, const common_params & params) { size_t hs_task_count = prompt_lines.size()/6; LOG_INF("%s : loaded %zu tasks from prompt.\n", __func__, hs_task_count); - const bool is_spm = llama_vocab_type(llama_get_model(ctx)) == LLAMA_VOCAB_TYPE_SPM; + const bool is_spm = llama_vocab_type(vocab) == LLAMA_VOCAB_TYPE_SPM; LOG_INF("================================= is_spm = %d\n", is_spm); // The tasks should be randomized so the score stabilizes quickly. @@ -897,7 +857,7 @@ static void hellaswag_score(llama_context * ctx, const common_params & params) { const int n_ctx = llama_n_ctx(ctx); const int n_batch = params.n_batch; - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); const int max_tasks_per_batch = 32; const int max_seq = std::min(4*max_tasks_per_batch, (int) llama_n_seq_max(ctx)); @@ -1121,6 +1081,8 @@ static std::vector load_winogrande_from_csv(const std::string * */ static void winogrande_score(llama_context * ctx, const common_params & params) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); constexpr int k_min_trailing_ctx = 3; @@ -1179,7 +1141,7 @@ static void winogrande_score(llama_context * ctx, const common_params & params) const int n_ctx = llama_n_ctx(ctx); const int n_batch = params.n_batch; - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); const int max_tasks_per_batch = 128; const int max_seq = std::min(2*max_tasks_per_batch, (int) llama_n_seq_max(ctx)); @@ -1423,6 +1385,8 @@ static bool multiple_choice_prepare_one_task(llama_context * ctx, multiple_choic // https://huggingface.co/datasets/truthful_qa // static void multiple_choice_score(llama_context * ctx, const common_params & params) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); std::istringstream strstream(params.prompt); uint32_t n_task; @@ -1531,7 +1495,7 @@ static void multiple_choice_score(llama_context * ctx, const common_params & par const int n_ctx = llama_n_ctx(ctx); const int n_batch = params.n_batch; - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const int n_vocab = llama_vocab_n_tokens(vocab); const int max_tasks_per_batch = 32; const int max_seq = std::min(4*max_tasks_per_batch, (int) llama_n_seq_max(ctx)); @@ -1704,6 +1668,9 @@ static void multiple_choice_score(llama_context * ctx, const common_params & par } static void kl_divergence(llama_context * ctx, const common_params & params) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + if (params.logits_file.empty()) { LOG_ERR("%s: you must provide a name of a file containing the log probabilities of the base model\n", __func__); return; @@ -1737,8 +1704,8 @@ static void kl_divergence(llama_context * ctx, const common_params & params) { LOG_ERR("%s: failed reading n_vocab, n_chunk from %s\n", __func__, params.logits_file.c_str()); return; } - if (n_vocab != llama_n_vocab(llama_get_model(ctx))) { - LOG_ERR("%s: inconsistent vocabulary (%d vs %d)\n", __func__, n_vocab, llama_n_vocab(llama_get_model(ctx))); + if (n_vocab != llama_vocab_n_tokens(vocab)) { + LOG_ERR("%s: inconsistent vocabulary (%d vs %d)\n", __func__, n_vocab, llama_vocab_n_tokens(vocab)); } std::vector tokens(size_t(n_ctx) * n_chunk); @@ -1750,8 +1717,8 @@ static void kl_divergence(llama_context * ctx, const common_params & params) { const int n_batch = params.n_batch; const int num_batches = (n_ctx + n_batch - 1)/n_batch; const int nv = 2*((n_vocab + 1)/2) + 4; - const bool add_bos = llama_add_bos_token(llama_get_model(ctx)); - GGML_ASSERT(!llama_add_eos_token(llama_get_model(ctx))); + const bool add_bos = llama_vocab_get_add_bos(vocab); + GGML_ASSERT(!llama_vocab_get_add_eos(vocab)); std::vector log_probs_uint16(size_t(n_ctx - 1 - n_ctx/2) * nv); std::vector kld_values(size_t(n_ctx - 1 - n_ctx/2)*n_chunk); @@ -1810,7 +1777,7 @@ static void kl_divergence(llama_context * ctx, const common_params & params) { // add BOS token for the first batch of each chunk if (add_bos && j == 0) { - tokens[batch_start] = llama_token_bos(llama_get_model(ctx)); + tokens[batch_start] = llama_vocab_bos(vocab); } common_batch_clear(batch); @@ -2036,14 +2003,15 @@ int main(int argc, char ** argv) { // load the model and apply lora adapter, if any common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); + if (model == NULL) { LOG_ERR("%s: unable to load model\n", __func__); return 1; } - const int n_ctx_train = llama_n_ctx_train(model); + const int n_ctx_train = llama_model_n_ctx_train(model); if (params.n_ctx > n_ctx_train) { LOG_WRN("%s: model was trained on only %d context tokens (%d specified)\n", @@ -2072,11 +2040,6 @@ int main(int argc, char ** argv) { LOG("\n"); llama_perf_context_print(ctx); - write_logfile(ctx, params, model, results); - - llama_free(ctx); - llama_free_model(model); - llama_backend_free(); return 0; diff --git a/examples/quantize-stats/CMakeLists.txt b/examples/quantize-stats/CMakeLists.txt index bb986a716..9a3a0d3cd 100644 --- a/examples/quantize-stats/CMakeLists.txt +++ b/examples/quantize-stats/CMakeLists.txt @@ -3,4 +3,4 @@ add_executable(${TARGET} quantize-stats.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE llama build_info ${CMAKE_THREAD_LIBS_INIT}) target_include_directories(${TARGET} PRIVATE ../../common) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/quantize-stats/quantize-stats.cpp b/examples/quantize-stats/quantize-stats.cpp index e372856c6..bd2f73467 100644 --- a/examples/quantize-stats/quantize-stats.cpp +++ b/examples/quantize-stats/quantize-stats.cpp @@ -1,7 +1,7 @@ -#include "common.h" #include "ggml.h" #include "llama.h" -#include "llama-impl.h" +#include "llama-context.h" +#include "common.h" #include #include @@ -9,11 +9,9 @@ #include #include #include -#include #include #include #include -#include #include #include #include @@ -142,7 +140,7 @@ static bool tensor_is_contiguous(const struct ggml_tensor * tensor) { } static void test_roundtrip_on_chunk( - const ggml_tensor * layer, int64_t offset, int64_t chunk_size, const ggml_type_traits & qfns, bool use_reference, + const ggml_tensor * layer, int64_t offset, int64_t chunk_size, const ggml_type_traits & qfns, const ggml_type_traits_cpu & qfns_cpu, bool use_reference, float * input_scratch, char * quantized_scratch, float * output_scratch, error_stats & stats ) { if (layer->type == GGML_TYPE_F16) { @@ -156,7 +154,7 @@ static void test_roundtrip_on_chunk( if (use_reference) { qfns.from_float_ref(input_scratch, quantized_scratch, chunk_size); } else { - qfns.from_float(input_scratch, quantized_scratch, chunk_size); + qfns_cpu.from_float(input_scratch, quantized_scratch, chunk_size); } qfns.to_float(quantized_scratch, output_scratch, chunk_size); @@ -166,7 +164,7 @@ static void test_roundtrip_on_chunk( // Run quantization function for a single layer and update error stats static void test_roundtrip_on_layer( - std::string & name, bool print_layer_stats, const ggml_type_traits & qfns, bool use_reference, + std::string & name, bool print_layer_stats, const ggml_type_traits & qfns, const ggml_type_traits_cpu & qfns_cpu, bool use_reference, const ggml_tensor * layer, std::vector & input_scratch, std::vector & quantized_scratch, std::vector & output_scratch, error_stats & total_error, int max_thread = 0 ) { @@ -187,13 +185,13 @@ static void test_roundtrip_on_layer( int num_chunks = (nelements + chunk_size - 1)/chunk_size; if (num_chunks < 2 || max_thread < 2) { - test_roundtrip_on_chunk(layer, 0, nelements, qfns, use_reference, input_scratch_ptr, quantized_scratch.data(), + test_roundtrip_on_chunk(layer, 0, nelements, qfns, qfns_cpu, use_reference, input_scratch_ptr, quantized_scratch.data(), output_scratch.data(), print_layer_stats ? layer_error : total_error); } else { auto & stats = print_layer_stats ? layer_error : total_error; std::mutex mutex; uint64_t counter = 0; - auto compute = [&mutex, &counter, &stats, &qfns, nelements, layer, use_reference, input_scratch_ptr, + auto compute = [&mutex, &counter, &stats, &qfns, &qfns_cpu, nelements, layer, use_reference, input_scratch_ptr, &quantized_scratch, &output_scratch, chunk_size] () { error_stats local_stats {}; while (true) { @@ -205,7 +203,7 @@ static void test_roundtrip_on_layer( } lock.unlock(); uint64_t chunk = offset + chunk_size < nelements ? chunk_size : nelements - offset; - test_roundtrip_on_chunk(layer, offset, chunk, qfns, use_reference, input_scratch_ptr + offset, + test_roundtrip_on_chunk(layer, offset, chunk, qfns, qfns_cpu, use_reference, input_scratch_ptr + offset, quantized_scratch.data() + 4*offset, output_scratch.data() + offset, local_stats); } }; @@ -311,7 +309,7 @@ int main(int argc, char ** argv) { auto mparams = llama_model_default_params(); mparams.use_mlock = false; - model = llama_load_model_from_file(params.model.c_str(), mparams); + model = llama_model_load_from_file(params.model.c_str(), mparams); if (model == NULL) { fprintf(stderr, "%s: error: failed to load model '%s'\n", __func__, params.model.c_str()); @@ -321,22 +319,22 @@ int main(int argc, char ** argv) { auto cparams = llama_context_default_params(); cparams.n_ctx = 256; - ctx = llama_new_context_with_model(model, cparams); + ctx = llama_init_from_model(model, cparams); if (ctx == NULL) { fprintf(stderr, "%s: error: failed to create context with model '%s'\n", __func__, params.model.c_str()); - llama_free_model(model); + llama_model_free(model); return 1; } } - const auto &tensors = llama_internal_get_tensor_map(ctx); + const auto & tensors = llama_internal_get_tensor_map(ctx); // check layer tensors int included_layers = 0; int64_t max_nelements = 0; bool is_f16 = false; - for (const auto& kv_tensor : tensors) { + for (const auto & kv_tensor : tensors) { if (!layer_included(params, kv_tensor.first)) { continue; } @@ -349,7 +347,7 @@ int main(int argc, char ** argv) { fprintf(stderr, "%s: error: Quantization should be tested with a float model, " "this model contains already quantized layers (%s is type %d)\n", __func__, kv_tensor.first.c_str(), kv_tensor.second->type); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); return 1; } included_layers++; @@ -371,8 +369,9 @@ int main(int argc, char ** argv) { if (!params.include_types.empty() && std::find(params.include_types.begin(), params.include_types.end(), i) == params.include_types.end()) { continue; } - const auto * qfns = ggml_get_type_traits(type); - if (qfns->from_float && qfns->to_float) { + const auto * qfns = ggml_get_type_traits(type); + const auto * qfns_cpu = ggml_get_type_traits_cpu(type); + if (qfns_cpu->from_float && qfns->to_float) { if (params.verbose) { printf("testing %s ...\n", ggml_type_name(type)); } @@ -381,7 +380,7 @@ int main(int argc, char ** argv) { error_stats global_stats {}; - for (const auto& kv_tensor : tensors) { + for (const auto & kv_tensor : tensors) { if (!layer_included(params, kv_tensor.first)) { continue; } @@ -393,7 +392,7 @@ int main(int argc, char ** argv) { test_roundtrip_on_layer( layer_name, params.per_layer_stats, - *qfns, + *qfns, *qfns_cpu, params.reference, kv_tensor.second, input_scratch, @@ -410,7 +409,7 @@ int main(int argc, char ** argv) { llama_free(ctx); - llama_free_model(model); + llama_model_free(model); // report timing { const int64_t t_main_end_us = ggml_time_us(); diff --git a/examples/quantize/CMakeLists.txt b/examples/quantize/CMakeLists.txt index 62680cda4..47e5cbe30 100644 --- a/examples/quantize/CMakeLists.txt +++ b/examples/quantize/CMakeLists.txt @@ -3,4 +3,4 @@ add_executable(${TARGET} quantize.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) target_include_directories(${TARGET} PRIVATE ../../common) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/quantize/README.md b/examples/quantize/README.md index 704f0d56b..f9cce7b21 100644 --- a/examples/quantize/README.md +++ b/examples/quantize/README.md @@ -54,8 +54,6 @@ As the models are currently fully loaded into memory, you will need adequate dis Several quantization methods are supported. They differ in the resulting model disk size and inference speed. -The quantization formats `Q4_0_4_4`, `Q4_0_4_8` and `Q4_0_8_8` are block interleaved variants of the `Q4_0` format, providing a data layout that is better suited for specific implementations of optimized mulmat kernels. Since these formats differ only in data layout, they have the same quantized size as the `Q4_0` format. - *(outdated)* | Model | Measure | F16 | Q4_0 | Q4_1 | Q5_0 | Q5_1 | Q8_0 | @@ -83,7 +81,7 @@ The quantization formats `Q4_0_4_4`, `Q4_0_4_8` and `Q4_0_8_8` are block interle - [#4930 - imatrix for all k-quants](https://github.com/ggerganov/llama.cpp/pull/4930) - [#4951 - imatrix on the GPU](https://github.com/ggerganov/llama.cpp/pull/4957) - [#4969 - imatrix for legacy quants](https://github.com/ggerganov/llama.cpp/pull/4969) - - [#4996 - k-qunats tuning](https://github.com/ggerganov/llama.cpp/pull/4996) + - [#4996 - k-quants tuning](https://github.com/ggerganov/llama.cpp/pull/4996) - [#5060 - Q3_K_XS](https://github.com/ggerganov/llama.cpp/pull/5060) - [#5196 - 3-bit i-quants](https://github.com/ggerganov/llama.cpp/pull/5196) - [quantization tuning](https://github.com/ggerganov/llama.cpp/pull/5320), [another one](https://github.com/ggerganov/llama.cpp/pull/5334), and [another one](https://github.com/ggerganov/llama.cpp/pull/5361) diff --git a/examples/quantize/quantize.cpp b/examples/quantize/quantize.cpp index b98993210..8d47b17b6 100644 --- a/examples/quantize/quantize.cpp +++ b/examples/quantize/quantize.cpp @@ -48,9 +48,6 @@ static const std::vector QUANT_OPTIONS = { { "Q5_K_M", LLAMA_FTYPE_MOSTLY_Q5_K_M, " 5.33G, +0.0569 ppl @ Llama-3-8B", }, { "Q6_K", LLAMA_FTYPE_MOSTLY_Q6_K, " 6.14G, +0.0217 ppl @ Llama-3-8B", }, { "Q8_0", LLAMA_FTYPE_MOSTLY_Q8_0, " 7.96G, +0.0026 ppl @ Llama-3-8B", }, - { "Q4_0_4_4", LLAMA_FTYPE_MOSTLY_Q4_0_4_4, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, - { "Q4_0_4_8", LLAMA_FTYPE_MOSTLY_Q4_0_4_8, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, - { "Q4_0_8_8", LLAMA_FTYPE_MOSTLY_Q4_0_8_8, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, { "F16", LLAMA_FTYPE_MOSTLY_F16, "14.00G, +0.0020 ppl @ Mistral-7B", }, { "BF16", LLAMA_FTYPE_MOSTLY_BF16, "14.00G, -0.0050 ppl @ Mistral-7B", }, { "F32", LLAMA_FTYPE_ALL_F32, "26.00G @ 7B", }, diff --git a/examples/quantize/tests.sh b/examples/quantize/tests.sh index 24bc970e8..70f7610f9 100644 --- a/examples/quantize/tests.sh +++ b/examples/quantize/tests.sh @@ -47,7 +47,7 @@ echo PASS echo # 3a. Test the requanted model is loading properly -$MAIN --model $WORK_PATH/ggml-model-requant-00001-of-00006.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-requant-00001-of-00006.gguf --n-predict 32 echo PASS echo @@ -57,7 +57,7 @@ echo PASS echo # 4b. Test the requanted model is loading properly -$MAIN --model $WORK_PATH/ggml-model-requant-merge.gguf --n-predict 32 +$MAIN -no-cnv --model $WORK_PATH/ggml-model-requant-merge.gguf --n-predict 32 echo PASS echo diff --git a/examples/retrieval/CMakeLists.txt b/examples/retrieval/CMakeLists.txt index 66610f311..512a602ec 100644 --- a/examples/retrieval/CMakeLists.txt +++ b/examples/retrieval/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-retrieval) add_executable(${TARGET} retrieval.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/retrieval/retrieval.cpp b/examples/retrieval/retrieval.cpp index 1768aae51..2439022a2 100644 --- a/examples/retrieval/retrieval.cpp +++ b/examples/retrieval/retrieval.cpp @@ -107,7 +107,7 @@ static void batch_decode(llama_context * ctx, llama_batch & batch, float * outpu } float * out = output + batch.seq_id[i][0] * n_embd; - common_embd_normalize(embd, out, n_embd); + common_embd_normalize(embd, out, n_embd, 2); } } @@ -143,7 +143,7 @@ int main(int argc, char ** argv) { std::vector file_chunk = chunk_file(context_file, params.chunk_size, params.chunk_separator); chunks.insert(chunks.end(), file_chunk.begin(), file_chunk.end()); } - LOG_INF("Number of chunks: %ld\n", chunks.size()); + LOG_INF("Number of chunks: %zu\n", chunks.size()); llama_backend_init(); llama_numa_init(params.numa); @@ -151,15 +151,17 @@ int main(int argc, char ** argv) { // load the model common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); if (model == NULL) { LOG_ERR("%s: unable to load model\n", __func__); return 1; } - const int n_ctx_train = llama_n_ctx_train(model); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_ctx_train = llama_model_n_ctx_train(model); const int n_ctx = llama_n_ctx(ctx); const enum llama_pooling_type pooling_type = llama_pooling_type(ctx); @@ -192,8 +194,8 @@ int main(int argc, char ** argv) { return 1; } // add eos if not present - if (llama_token_eos(model) >= 0 && (inp.empty() || inp.back() != llama_token_eos(model))) { - inp.push_back(llama_token_eos(model)); + if (llama_vocab_eos(vocab) >= 0 && (inp.empty() || inp.back() != llama_vocab_eos(vocab))) { + inp.push_back(llama_vocab_eos(vocab)); } chunk.tokens = inp; } @@ -215,7 +217,7 @@ int main(int argc, char ** argv) { struct llama_batch batch = llama_batch_init(n_batch, 0, 1); // allocate output - const int n_embd = llama_n_embd(model); + const int n_embd = llama_model_n_embd(model); std::vector embeddings(n_chunks * n_embd, 0); float * emb = embeddings.data(); @@ -282,8 +284,8 @@ int main(int argc, char ** argv) { return a.second > b.second; }); - LOG("Top %d similar chunks:\n", params.sparams.top_k); - for (int i = 0; i < std::min(params.sparams.top_k, (int) chunks.size()); i++) { + LOG("Top %d similar chunks:\n", params.sampling.top_k); + for (int i = 0; i < std::min(params.sampling.top_k, (int) chunks.size()); i++) { LOG("filename: %s\n", chunks[similarities[i].first].filename.c_str()); LOG("filepos: %lld\n", (long long int) chunks[similarities[i].first].filepos); LOG("similarity: %f\n", similarities[i].second); @@ -298,7 +300,5 @@ int main(int argc, char ** argv) { // clean up llama_batch_free(query_batch); - llama_free(ctx); - llama_free_model(model); llama_backend_free(); } diff --git a/examples/rpc/rpc-server.cpp b/examples/rpc/rpc-server.cpp index 5fe70dac7..8b1b23eda 100644 --- a/examples/rpc/rpc-server.cpp +++ b/examples/rpc/rpc-server.cpp @@ -12,6 +12,10 @@ #include "ggml-vulkan.h" #endif +#ifdef GGML_USE_SYCL +#include "ggml-sycl.h" +#endif + #include "ggml-rpc.h" #ifdef _WIN32 # include @@ -91,6 +95,12 @@ static ggml_backend_t create_backend() { if (!backend) { fprintf(stderr, "%s: ggml_backend_vulkan_init() failed\n", __func__); } +#elif GGML_USE_SYCL + fprintf(stderr, "%s: using SYCL backend\n", __func__); + backend = ggml_backend_sycl_init(0); // init device 0 + if (!backend) { + fprintf(stderr, "%s: ggml_backend_sycl_init() failed\n", __func__); + } #endif // if there aren't GPU Backends fallback to CPU backend @@ -106,6 +116,8 @@ static void get_backend_memory(size_t * free_mem, size_t * total_mem) { ggml_backend_cuda_get_device_memory(0, free_mem, total_mem); #elif GGML_USE_VULKAN ggml_backend_vk_get_device_memory(0, free_mem, total_mem); +#elif GGML_USE_SYCL + ggml_backend_sycl_get_device_memory(0, free_mem, total_mem); #else #ifdef _WIN32 MEMORYSTATUSEX status; diff --git a/examples/run/CMakeLists.txt b/examples/run/CMakeLists.txt new file mode 100644 index 000000000..0686d6305 --- /dev/null +++ b/examples/run/CMakeLists.txt @@ -0,0 +1,5 @@ +set(TARGET llama-run) +add_executable(${TARGET} run.cpp) +install(TARGETS ${TARGET} RUNTIME) +target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/run/README.md b/examples/run/README.md new file mode 100644 index 000000000..a06805441 --- /dev/null +++ b/examples/run/README.md @@ -0,0 +1,51 @@ +# llama.cpp/example/run + +The purpose of this example is to demonstrate a minimal usage of llama.cpp for running models. + +```bash +llama-run granite-code +``` + +```bash +llama-run -h +Description: + Runs a llm + +Usage: + llama-run [options] model [prompt] + +Options: + -c, --context-size + Context size (default: 2048) + -n, --ngl + Number of GPU layers (default: 0) + --temp + Temperature (default: 0.8) + -v, --verbose, --log-verbose + Set verbosity level to infinity (i.e. log all messages, useful for debugging) + -h, --help + Show help message + +Commands: + model + Model is a string with an optional prefix of + huggingface:// (hf://), ollama://, https:// or file://. + If no protocol is specified and a file exists in the specified + path, file:// is assumed, otherwise if a file does not exist in + the specified path, ollama:// is assumed. Models that are being + pulled are downloaded with .partial extension while being + downloaded and then renamed as the file without the .partial + extension when complete. + +Examples: + llama-run llama3 + llama-run ollama://granite-code + llama-run ollama://smollm:135m + llama-run hf://QuantFactory/SmolLM-135M-GGUF/SmolLM-135M.Q2_K.gguf + llama-run huggingface://bartowski/SmolLM-1.7B-Instruct-v0.2-GGUF/SmolLM-1.7B-Instruct-v0.2-IQ3_M.gguf + llama-run https://example.com/some-file1.gguf + llama-run some-file2.gguf + llama-run file://some-file3.gguf + llama-run --ngl 999 some-file4.gguf + llama-run --ngl 999 some-file5.gguf Hello World +``` diff --git a/examples/run/run.cpp b/examples/run/run.cpp new file mode 100644 index 000000000..0ad8bb15b --- /dev/null +++ b/examples/run/run.cpp @@ -0,0 +1,1007 @@ +#if defined(_WIN32) +# include +# include +#else +# include +# include +# include +#endif + +#if defined(LLAMA_USE_CURL) +# include +#endif + +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "common.h" +#include "json.hpp" +#include "llama-cpp.h" + +#if defined(__unix__) || (defined(__APPLE__) && defined(__MACH__)) || defined(_WIN32) +[[noreturn]] static void sigint_handler(int) { + printf("\n\033[0m"); + exit(0); // not ideal, but it's the only way to guarantee exit in all cases +} +#endif + +GGML_ATTRIBUTE_FORMAT(1, 2) +static std::string fmt(const char * fmt, ...) { + va_list ap; + va_list ap2; + va_start(ap, fmt); + va_copy(ap2, ap); + const int size = vsnprintf(NULL, 0, fmt, ap); + GGML_ASSERT(size >= 0 && size < INT_MAX); // NOLINT + std::string buf; + buf.resize(size); + const int size2 = vsnprintf(const_cast(buf.data()), buf.size() + 1, fmt, ap2); + GGML_ASSERT(size2 == size); + va_end(ap2); + va_end(ap); + + return buf; +} + +GGML_ATTRIBUTE_FORMAT(1, 2) +static int printe(const char * fmt, ...) { + va_list args; + va_start(args, fmt); + const int ret = vfprintf(stderr, fmt, args); + va_end(args); + + return ret; +} + +class Opt { + public: + int init(int argc, const char ** argv) { + ctx_params = llama_context_default_params(); + model_params = llama_model_default_params(); + context_size_default = ctx_params.n_batch; + ngl_default = model_params.n_gpu_layers; + common_params_sampling sampling; + temperature_default = sampling.temp; + + if (argc < 2) { + printe("Error: No arguments provided.\n"); + print_help(); + return 1; + } + + // Parse arguments + if (parse(argc, argv)) { + printe("Error: Failed to parse arguments.\n"); + print_help(); + return 1; + } + + // If help is requested, show help and exit + if (help) { + print_help(); + return 2; + } + + ctx_params.n_batch = context_size >= 0 ? context_size : context_size_default; + ctx_params.n_ctx = ctx_params.n_batch; + model_params.n_gpu_layers = ngl >= 0 ? ngl : ngl_default; + temperature = temperature >= 0 ? temperature : temperature_default; + + return 0; // Success + } + + llama_context_params ctx_params; + llama_model_params model_params; + std::string model_; + std::string user; + int context_size = -1, ngl = -1; + float temperature = -1; + bool verbose = false; + + private: + int context_size_default = -1, ngl_default = -1; + float temperature_default = -1; + bool help = false; + + bool parse_flag(const char ** argv, int i, const char * short_opt, const char * long_opt) { + return strcmp(argv[i], short_opt) == 0 || strcmp(argv[i], long_opt) == 0; + } + + int handle_option_with_value(int argc, const char ** argv, int & i, int & option_value) { + if (i + 1 >= argc) { + return 1; + } + + option_value = std::atoi(argv[++i]); + + return 0; + } + + int handle_option_with_value(int argc, const char ** argv, int & i, float & option_value) { + if (i + 1 >= argc) { + return 1; + } + + option_value = std::atof(argv[++i]); + + return 0; + } + + int parse(int argc, const char ** argv) { + bool options_parsing = true; + for (int i = 1, positional_args_i = 0; i < argc; ++i) { + if (options_parsing && (strcmp(argv[i], "-c") == 0 || strcmp(argv[i], "--context-size") == 0)) { + if (handle_option_with_value(argc, argv, i, context_size) == 1) { + return 1; + } + } else if (options_parsing && (strcmp(argv[i], "-n") == 0 || strcmp(argv[i], "--ngl") == 0)) { + if (handle_option_with_value(argc, argv, i, ngl) == 1) { + return 1; + } + } else if (options_parsing && strcmp(argv[i], "--temp") == 0) { + if (handle_option_with_value(argc, argv, i, temperature) == 1) { + return 1; + } + } else if (options_parsing && + (parse_flag(argv, i, "-v", "--verbose") || parse_flag(argv, i, "-v", "--log-verbose"))) { + verbose = true; + } else if (options_parsing && parse_flag(argv, i, "-h", "--help")) { + help = true; + return 0; + } else if (options_parsing && strcmp(argv[i], "--") == 0) { + options_parsing = false; + } else if (positional_args_i == 0) { + if (!argv[i][0] || argv[i][0] == '-') { + return 1; + } + + ++positional_args_i; + model_ = argv[i]; + } else if (positional_args_i == 1) { + ++positional_args_i; + user = argv[i]; + } else { + user += " " + std::string(argv[i]); + } + } + + return 0; + } + + void print_help() const { + printf( + "Description:\n" + " Runs a llm\n" + "\n" + "Usage:\n" + " llama-run [options] model [prompt]\n" + "\n" + "Options:\n" + " -c, --context-size \n" + " Context size (default: %d)\n" + " -n, --ngl \n" + " Number of GPU layers (default: %d)\n" + " --temp \n" + " Temperature (default: %.1f)\n" + " -v, --verbose, --log-verbose\n" + " Set verbosity level to infinity (i.e. log all messages, useful for debugging)\n" + " -h, --help\n" + " Show help message\n" + "\n" + "Commands:\n" + " model\n" + " Model is a string with an optional prefix of \n" + " huggingface:// (hf://), ollama://, https:// or file://.\n" + " If no protocol is specified and a file exists in the specified\n" + " path, file:// is assumed, otherwise if a file does not exist in\n" + " the specified path, ollama:// is assumed. Models that are being\n" + " pulled are downloaded with .partial extension while being\n" + " downloaded and then renamed as the file without the .partial\n" + " extension when complete.\n" + "\n" + "Examples:\n" + " llama-run llama3\n" + " llama-run ollama://granite-code\n" + " llama-run ollama://smollm:135m\n" + " llama-run hf://QuantFactory/SmolLM-135M-GGUF/SmolLM-135M.Q2_K.gguf\n" + " llama-run " + "huggingface://bartowski/SmolLM-1.7B-Instruct-v0.2-GGUF/SmolLM-1.7B-Instruct-v0.2-IQ3_M.gguf\n" + " llama-run https://example.com/some-file1.gguf\n" + " llama-run some-file2.gguf\n" + " llama-run file://some-file3.gguf\n" + " llama-run --ngl 999 some-file4.gguf\n" + " llama-run --ngl 999 some-file5.gguf Hello World\n", + context_size_default, ngl_default, temperature_default); + } +}; + +struct progress_data { + size_t file_size = 0; + std::chrono::steady_clock::time_point start_time = std::chrono::steady_clock::now(); + bool printed = false; +}; + +static int get_terminal_width() { +#if defined(_WIN32) + CONSOLE_SCREEN_BUFFER_INFO csbi; + GetConsoleScreenBufferInfo(GetStdHandle(STD_OUTPUT_HANDLE), &csbi); + return csbi.srWindow.Right - csbi.srWindow.Left + 1; +#else + struct winsize w; + ioctl(STDOUT_FILENO, TIOCGWINSZ, &w); + return w.ws_col; +#endif +} + +#ifdef LLAMA_USE_CURL +class File { + public: + FILE * file = nullptr; + + FILE * open(const std::string & filename, const char * mode) { + file = fopen(filename.c_str(), mode); + + return file; + } + + int lock() { + if (file) { +# ifdef _WIN32 + fd = _fileno(file); + hFile = (HANDLE) _get_osfhandle(fd); + if (hFile == INVALID_HANDLE_VALUE) { + fd = -1; + + return 1; + } + + OVERLAPPED overlapped = {}; + if (!LockFileEx(hFile, LOCKFILE_EXCLUSIVE_LOCK | LOCKFILE_FAIL_IMMEDIATELY, 0, MAXDWORD, MAXDWORD, + &overlapped)) { + fd = -1; + + return 1; + } +# else + fd = fileno(file); + if (flock(fd, LOCK_EX | LOCK_NB) != 0) { + fd = -1; + + return 1; + } +# endif + } + + return 0; + } + + ~File() { + if (fd >= 0) { +# ifdef _WIN32 + if (hFile != INVALID_HANDLE_VALUE) { + OVERLAPPED overlapped = {}; + UnlockFileEx(hFile, 0, MAXDWORD, MAXDWORD, &overlapped); + } +# else + flock(fd, LOCK_UN); +# endif + } + + if (file) { + fclose(file); + } + } + + private: + int fd = -1; +# ifdef _WIN32 + HANDLE hFile = nullptr; +# endif +}; + +class HttpClient { + public: + int init(const std::string & url, const std::vector & headers, const std::string & output_file, + const bool progress, std::string * response_str = nullptr) { + std::string output_file_partial; + curl = curl_easy_init(); + if (!curl) { + return 1; + } + + progress_data data; + File out; + if (!output_file.empty()) { + output_file_partial = output_file + ".partial"; + if (!out.open(output_file_partial, "ab")) { + printe("Failed to open file\n"); + + return 1; + } + + if (out.lock()) { + printe("Failed to exclusively lock file\n"); + + return 1; + } + } + + set_write_options(response_str, out); + data.file_size = set_resume_point(output_file_partial); + set_progress_options(progress, data); + set_headers(headers); + perform(url); + if (!output_file.empty()) { + std::filesystem::rename(output_file_partial, output_file); + } + + return 0; + } + + ~HttpClient() { + if (chunk) { + curl_slist_free_all(chunk); + } + + if (curl) { + curl_easy_cleanup(curl); + } + } + + private: + CURL * curl = nullptr; + struct curl_slist * chunk = nullptr; + + void set_write_options(std::string * response_str, const File & out) { + if (response_str) { + curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, capture_data); + curl_easy_setopt(curl, CURLOPT_WRITEDATA, response_str); + } else { + curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_data); + curl_easy_setopt(curl, CURLOPT_WRITEDATA, out.file); + } + } + + size_t set_resume_point(const std::string & output_file) { + size_t file_size = 0; + if (std::filesystem::exists(output_file)) { + file_size = std::filesystem::file_size(output_file); + curl_easy_setopt(curl, CURLOPT_RESUME_FROM_LARGE, static_cast(file_size)); + } + + return file_size; + } + + void set_progress_options(bool progress, progress_data & data) { + if (progress) { + curl_easy_setopt(curl, CURLOPT_NOPROGRESS, 0L); + curl_easy_setopt(curl, CURLOPT_XFERINFODATA, &data); + curl_easy_setopt(curl, CURLOPT_XFERINFOFUNCTION, update_progress); + } + } + + void set_headers(const std::vector & headers) { + if (!headers.empty()) { + if (chunk) { + curl_slist_free_all(chunk); + chunk = 0; + } + + for (const auto & header : headers) { + chunk = curl_slist_append(chunk, header.c_str()); + } + + curl_easy_setopt(curl, CURLOPT_HTTPHEADER, chunk); + } + } + + void perform(const std::string & url) { + CURLcode res; + curl_easy_setopt(curl, CURLOPT_URL, url.c_str()); + curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L); + curl_easy_setopt(curl, CURLOPT_DEFAULT_PROTOCOL, "https"); + curl_easy_setopt(curl, CURLOPT_FAILONERROR, 1L); + res = curl_easy_perform(curl); + if (res != CURLE_OK) { + printe("curl_easy_perform() failed: %s\n", curl_easy_strerror(res)); + } + } + + static std::string human_readable_time(double seconds) { + int hrs = static_cast(seconds) / 3600; + int mins = (static_cast(seconds) % 3600) / 60; + int secs = static_cast(seconds) % 60; + + if (hrs > 0) { + return fmt("%dh %02dm %02ds", hrs, mins, secs); + } else if (mins > 0) { + return fmt("%dm %02ds", mins, secs); + } else { + return fmt("%ds", secs); + } + } + + static std::string human_readable_size(curl_off_t size) { + static const char * suffix[] = { "B", "KB", "MB", "GB", "TB" }; + char length = sizeof(suffix) / sizeof(suffix[0]); + int i = 0; + double dbl_size = size; + if (size > 1024) { + for (i = 0; (size / 1024) > 0 && i < length - 1; i++, size /= 1024) { + dbl_size = size / 1024.0; + } + } + + return fmt("%.2f %s", dbl_size, suffix[i]); + } + + static int update_progress(void * ptr, curl_off_t total_to_download, curl_off_t now_downloaded, curl_off_t, + curl_off_t) { + progress_data * data = static_cast(ptr); + if (total_to_download <= 0) { + return 0; + } + + total_to_download += data->file_size; + const curl_off_t now_downloaded_plus_file_size = now_downloaded + data->file_size; + const curl_off_t percentage = calculate_percentage(now_downloaded_plus_file_size, total_to_download); + std::string progress_prefix = generate_progress_prefix(percentage); + + const double speed = calculate_speed(now_downloaded, data->start_time); + const double tim = (total_to_download - now_downloaded) / speed; + std::string progress_suffix = + generate_progress_suffix(now_downloaded_plus_file_size, total_to_download, speed, tim); + + int progress_bar_width = calculate_progress_bar_width(progress_prefix, progress_suffix); + std::string progress_bar; + generate_progress_bar(progress_bar_width, percentage, progress_bar); + + print_progress(progress_prefix, progress_bar, progress_suffix); + data->printed = true; + + return 0; + } + + static curl_off_t calculate_percentage(curl_off_t now_downloaded_plus_file_size, curl_off_t total_to_download) { + return (now_downloaded_plus_file_size * 100) / total_to_download; + } + + static std::string generate_progress_prefix(curl_off_t percentage) { return fmt("%3ld%% |", static_cast(percentage)); } + + static double calculate_speed(curl_off_t now_downloaded, const std::chrono::steady_clock::time_point & start_time) { + const auto now = std::chrono::steady_clock::now(); + const std::chrono::duration elapsed_seconds = now - start_time; + return now_downloaded / elapsed_seconds.count(); + } + + static std::string generate_progress_suffix(curl_off_t now_downloaded_plus_file_size, curl_off_t total_to_download, + double speed, double estimated_time) { + const int width = 10; + return fmt("%*s/%*s%*s/s%*s", width, human_readable_size(now_downloaded_plus_file_size).c_str(), width, + human_readable_size(total_to_download).c_str(), width, human_readable_size(speed).c_str(), width, + human_readable_time(estimated_time).c_str()); + } + + static int calculate_progress_bar_width(const std::string & progress_prefix, const std::string & progress_suffix) { + int progress_bar_width = get_terminal_width() - progress_prefix.size() - progress_suffix.size() - 3; + if (progress_bar_width < 1) { + progress_bar_width = 1; + } + + return progress_bar_width; + } + + static std::string generate_progress_bar(int progress_bar_width, curl_off_t percentage, + std::string & progress_bar) { + const curl_off_t pos = (percentage * progress_bar_width) / 100; + for (int i = 0; i < progress_bar_width; ++i) { + progress_bar.append((i < pos) ? "█" : " "); + } + + return progress_bar; + } + + static void print_progress(const std::string & progress_prefix, const std::string & progress_bar, + const std::string & progress_suffix) { + printe("\r%*s\r%s%s| %s", get_terminal_width(), " ", progress_prefix.c_str(), progress_bar.c_str(), + progress_suffix.c_str()); + } + // Function to write data to a file + static size_t write_data(void * ptr, size_t size, size_t nmemb, void * stream) { + FILE * out = static_cast(stream); + return fwrite(ptr, size, nmemb, out); + } + + // Function to capture data into a string + static size_t capture_data(void * ptr, size_t size, size_t nmemb, void * stream) { + std::string * str = static_cast(stream); + str->append(static_cast(ptr), size * nmemb); + return size * nmemb; + } +}; +#endif + +class LlamaData { + public: + llama_model_ptr model; + llama_sampler_ptr sampler; + llama_context_ptr context; + std::vector messages; + std::vector msg_strs; + std::vector fmtted; + + int init(Opt & opt) { + model = initialize_model(opt); + if (!model) { + return 1; + } + + context = initialize_context(model, opt); + if (!context) { + return 1; + } + + sampler = initialize_sampler(opt); + return 0; + } + + private: +#ifdef LLAMA_USE_CURL + int download(const std::string & url, const std::vector & headers, const std::string & output_file, + const bool progress, std::string * response_str = nullptr) { + HttpClient http; + if (http.init(url, headers, output_file, progress, response_str)) { + return 1; + } + + return 0; + } +#else + int download(const std::string &, const std::vector &, const std::string &, const bool, + std::string * = nullptr) { + printe("%s: llama.cpp built without libcurl, downloading from an url not supported.\n", __func__); + return 1; + } +#endif + + int huggingface_dl(const std::string & model, const std::vector headers, const std::string & bn) { + // Find the second occurrence of '/' after protocol string + size_t pos = model.find('/'); + pos = model.find('/', pos + 1); + if (pos == std::string::npos) { + return 1; + } + + const std::string hfr = model.substr(0, pos); + const std::string hff = model.substr(pos + 1); + const std::string url = "https://huggingface.co/" + hfr + "/resolve/main/" + hff; + return download(url, headers, bn, true); + } + + int ollama_dl(std::string & model, const std::vector headers, const std::string & bn) { + if (model.find('/') == std::string::npos) { + model = "library/" + model; + } + + std::string model_tag = "latest"; + size_t colon_pos = model.find(':'); + if (colon_pos != std::string::npos) { + model_tag = model.substr(colon_pos + 1); + model = model.substr(0, colon_pos); + } + + std::string manifest_url = "https://registry.ollama.ai/v2/" + model + "/manifests/" + model_tag; + std::string manifest_str; + const int ret = download(manifest_url, headers, "", false, &manifest_str); + if (ret) { + return ret; + } + + nlohmann::json manifest = nlohmann::json::parse(manifest_str); + std::string layer; + for (const auto & l : manifest["layers"]) { + if (l["mediaType"] == "application/vnd.ollama.image.model") { + layer = l["digest"]; + break; + } + } + + std::string blob_url = "https://registry.ollama.ai/v2/" + model + "/blobs/" + layer; + return download(blob_url, headers, bn, true); + } + + std::string basename(const std::string & path) { + const size_t pos = path.find_last_of("/\\"); + if (pos == std::string::npos) { + return path; + } + + return path.substr(pos + 1); + } + + int remove_proto(std::string & model_) { + const std::string::size_type pos = model_.find("://"); + if (pos == std::string::npos) { + return 1; + } + + model_ = model_.substr(pos + 3); // Skip past "://" + return 0; + } + + int resolve_model(std::string & model_) { + int ret = 0; + if (string_starts_with(model_, "file://") || std::filesystem::exists(model_)) { + remove_proto(model_); + + return ret; + } + + const std::string bn = basename(model_); + const std::vector headers = { "--header", + "Accept: application/vnd.docker.distribution.manifest.v2+json" }; + if (string_starts_with(model_, "hf://") || string_starts_with(model_, "huggingface://")) { + remove_proto(model_); + ret = huggingface_dl(model_, headers, bn); + } else if (string_starts_with(model_, "ollama://")) { + remove_proto(model_); + ret = ollama_dl(model_, headers, bn); + } else if (string_starts_with(model_, "https://")) { + download(model_, headers, bn, true); + } else { + ret = ollama_dl(model_, headers, bn); + } + + model_ = bn; + + return ret; + } + + // Initializes the model and returns a unique pointer to it + llama_model_ptr initialize_model(Opt & opt) { + ggml_backend_load_all(); + resolve_model(opt.model_); + printe( + "\r%*s" + "\rLoading model", + get_terminal_width(), " "); + llama_model_ptr model(llama_model_load_from_file(opt.model_.c_str(), opt.model_params)); + if (!model) { + printe("%s: error: unable to load model from file: %s\n", __func__, opt.model_.c_str()); + } + + printe("\r%*s\r", static_cast(sizeof("Loading model")), " "); + return model; + } + + // Initializes the context with the specified parameters + llama_context_ptr initialize_context(const llama_model_ptr & model, const Opt & opt) { + llama_context_ptr context(llama_init_from_model(model.get(), opt.ctx_params)); + if (!context) { + printe("%s: error: failed to create the llama_context\n", __func__); + } + + return context; + } + + // Initializes and configures the sampler + llama_sampler_ptr initialize_sampler(const Opt & opt) { + llama_sampler_ptr sampler(llama_sampler_chain_init(llama_sampler_chain_default_params())); + llama_sampler_chain_add(sampler.get(), llama_sampler_init_min_p(0.05f, 1)); + llama_sampler_chain_add(sampler.get(), llama_sampler_init_temp(opt.temperature)); + llama_sampler_chain_add(sampler.get(), llama_sampler_init_dist(LLAMA_DEFAULT_SEED)); + + return sampler; + } +}; + +// Add a message to `messages` and store its content in `msg_strs` +static void add_message(const char * role, const std::string & text, LlamaData & llama_data) { + llama_data.msg_strs.push_back(std::move(text)); + llama_data.messages.push_back({ role, llama_data.msg_strs.back().c_str() }); +} + +// Function to apply the chat template and resize `formatted` if needed +static int apply_chat_template(LlamaData & llama_data, const bool append) { + int result = llama_chat_apply_template( + llama_model_chat_template(llama_data.model.get()), llama_data.messages.data(), llama_data.messages.size(), append, + append ? llama_data.fmtted.data() : nullptr, append ? llama_data.fmtted.size() : 0); + if (append && result > static_cast(llama_data.fmtted.size())) { + llama_data.fmtted.resize(result); + result = llama_chat_apply_template(llama_model_chat_template(llama_data.model.get()), llama_data.messages.data(), + llama_data.messages.size(), append, llama_data.fmtted.data(), + llama_data.fmtted.size()); + } + + return result; +} + +// Function to tokenize the prompt +static int tokenize_prompt(const llama_vocab * vocab, const std::string & prompt, + std::vector & prompt_tokens) { + const int n_prompt_tokens = -llama_tokenize(vocab, prompt.c_str(), prompt.size(), NULL, 0, true, true); + prompt_tokens.resize(n_prompt_tokens); + if (llama_tokenize(vocab, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), true, + true) < 0) { + printe("failed to tokenize the prompt\n"); + return -1; + } + + return n_prompt_tokens; +} + +// Check if we have enough space in the context to evaluate this batch +static int check_context_size(const llama_context_ptr & ctx, const llama_batch & batch) { + const int n_ctx = llama_n_ctx(ctx.get()); + const int n_ctx_used = llama_get_kv_cache_used_cells(ctx.get()); + if (n_ctx_used + batch.n_tokens > n_ctx) { + printf("\033[0m\n"); + printe("context size exceeded\n"); + return 1; + } + + return 0; +} + +// convert the token to a string +static int convert_token_to_string(const llama_vocab * vocab, const llama_token token_id, std::string & piece) { + char buf[256]; + int n = llama_token_to_piece(vocab, token_id, buf, sizeof(buf), 0, true); + if (n < 0) { + printe("failed to convert token to piece\n"); + return 1; + } + + piece = std::string(buf, n); + return 0; +} + +static void print_word_and_concatenate_to_response(const std::string & piece, std::string & response) { + printf("%s", piece.c_str()); + fflush(stdout); + response += piece; +} + +// helper function to evaluate a prompt and generate a response +static int generate(LlamaData & llama_data, const std::string & prompt, std::string & response) { + const llama_vocab * vocab = llama_model_get_vocab(llama_data.model.get()); + + std::vector tokens; + if (tokenize_prompt(vocab, prompt, tokens) < 0) { + return 1; + } + + // prepare a batch for the prompt + llama_batch batch = llama_batch_get_one(tokens.data(), tokens.size()); + llama_token new_token_id; + while (true) { + check_context_size(llama_data.context, batch); + if (llama_decode(llama_data.context.get(), batch)) { + printe("failed to decode\n"); + return 1; + } + + // sample the next token, check is it an end of generation? + new_token_id = llama_sampler_sample(llama_data.sampler.get(), llama_data.context.get(), -1); + if (llama_vocab_is_eog(vocab, new_token_id)) { + break; + } + + std::string piece; + if (convert_token_to_string(vocab, new_token_id, piece)) { + return 1; + } + + print_word_and_concatenate_to_response(piece, response); + + // prepare the next batch with the sampled token + batch = llama_batch_get_one(&new_token_id, 1); + } + + return 0; +} + +static int read_user_input(std::string & user) { + std::getline(std::cin, user); + if (std::cin.eof()) { + printf("\n"); + return 1; + } + + if (user == "/bye") { + return 1; + } + + if (user.empty()) { + return 2; + } + + return 0; // Should have data in happy path +} + +// Function to generate a response based on the prompt +static int generate_response(LlamaData & llama_data, const std::string & prompt, std::string & response, + const bool stdout_a_terminal) { + // Set response color + if (stdout_a_terminal) { + printf("\033[33m"); + } + + if (generate(llama_data, prompt, response)) { + printe("failed to generate response\n"); + return 1; + } + + // End response with color reset and newline + printf("\n%s", stdout_a_terminal ? "\033[0m" : ""); + return 0; +} + +// Helper function to apply the chat template and handle errors +static int apply_chat_template_with_error_handling(LlamaData & llama_data, const bool append, int & output_length) { + const int new_len = apply_chat_template(llama_data, append); + if (new_len < 0) { + printe("failed to apply the chat template\n"); + return -1; + } + + output_length = new_len; + return 0; +} + +// Helper function to handle user input +static int handle_user_input(std::string & user_input, const std::string & user) { + if (!user.empty()) { + user_input = user; + return 0; // No need for interactive input + } + + printf( + "\r%*s" + "\r\033[32m> \033[0m", + get_terminal_width(), " "); + return read_user_input(user_input); // Returns true if input ends the loop +} + +static bool is_stdin_a_terminal() { +#if defined(_WIN32) + HANDLE hStdin = GetStdHandle(STD_INPUT_HANDLE); + DWORD mode; + return GetConsoleMode(hStdin, &mode); +#else + return isatty(STDIN_FILENO); +#endif +} + +static bool is_stdout_a_terminal() { +#if defined(_WIN32) + HANDLE hStdout = GetStdHandle(STD_OUTPUT_HANDLE); + DWORD mode; + return GetConsoleMode(hStdout, &mode); +#else + return isatty(STDOUT_FILENO); +#endif +} + +// Function to handle user input +static int get_user_input(std::string & user_input, const std::string & user) { + while (true) { + const int ret = handle_user_input(user_input, user); + if (ret == 1) { + return 1; + } + + if (ret == 2) { + continue; + } + + break; + } + + return 0; +} + +// Main chat loop function +static int chat_loop(LlamaData & llama_data, const std::string & user) { + int prev_len = 0; + llama_data.fmtted.resize(llama_n_ctx(llama_data.context.get())); + static const bool stdout_a_terminal = is_stdout_a_terminal(); + while (true) { + // Get user input + std::string user_input; + if (get_user_input(user_input, user) == 1) { + return 0; + } + + add_message("user", user.empty() ? user_input : user, llama_data); + int new_len; + if (apply_chat_template_with_error_handling(llama_data, true, new_len) < 0) { + return 1; + } + + std::string prompt(llama_data.fmtted.begin() + prev_len, llama_data.fmtted.begin() + new_len); + std::string response; + if (generate_response(llama_data, prompt, response, stdout_a_terminal)) { + return 1; + } + + if (!user.empty()) { + break; + } + + add_message("assistant", response, llama_data); + if (apply_chat_template_with_error_handling(llama_data, false, prev_len) < 0) { + return 1; + } + } + + return 0; +} + +static void log_callback(const enum ggml_log_level level, const char * text, void * p) { + const Opt * opt = static_cast(p); + if (opt->verbose || level == GGML_LOG_LEVEL_ERROR) { + printe("%s", text); + } +} + +static std::string read_pipe_data() { + std::ostringstream result; + result << std::cin.rdbuf(); // Read all data from std::cin + return result.str(); +} + +static void ctrl_c_handling() { +#if defined(__unix__) || (defined(__APPLE__) && defined(__MACH__)) + struct sigaction sigint_action; + sigint_action.sa_handler = sigint_handler; + sigemptyset(&sigint_action.sa_mask); + sigint_action.sa_flags = 0; + sigaction(SIGINT, &sigint_action, NULL); +#elif defined(_WIN32) + auto console_ctrl_handler = +[](DWORD ctrl_type) -> BOOL { + return (ctrl_type == CTRL_C_EVENT) ? (sigint_handler(SIGINT), true) : false; + }; + SetConsoleCtrlHandler(reinterpret_cast(console_ctrl_handler), true); +#endif +} + +int main(int argc, const char ** argv) { + ctrl_c_handling(); + Opt opt; + const int ret = opt.init(argc, argv); + if (ret == 2) { + return 0; + } else if (ret) { + return 1; + } + + if (!is_stdin_a_terminal()) { + if (!opt.user.empty()) { + opt.user += "\n\n"; + } + + opt.user += read_pipe_data(); + } + + llama_log_set(log_callback, &opt); + LlamaData llama_data; + if (llama_data.init(opt)) { + return 1; + } + + if (chat_loop(llama_data, opt.user)) { + return 1; + } + + return 0; +} diff --git a/examples/save-load-state/CMakeLists.txt b/examples/save-load-state/CMakeLists.txt index 0fb5e359b..0f50e50de 100644 --- a/examples/save-load-state/CMakeLists.txt +++ b/examples/save-load-state/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-save-load-state) add_executable(${TARGET} save-load-state.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/save-load-state/save-load-state.cpp b/examples/save-load-state/save-load-state.cpp index 8c49a52a6..cf7cbd815 100644 --- a/examples/save-load-state/save-load-state.cpp +++ b/examples/save-load-state/save-load-state.cpp @@ -9,7 +9,7 @@ int main(int argc, char ** argv) { common_params params; params.prompt = "The quick brown fox"; - params.sparams.seed = 1234; + params.sampling.seed = 1234; if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_COMMON)) { return 1; @@ -30,8 +30,8 @@ int main(int argc, char ** argv) { // init common_init_result llama_init = common_init_from_params(params); - llama_model * model = llama_init.model; - llama_context * ctx = llama_init.context; + llama_model * model = llama_init.model.get(); + llama_context * ctx = llama_init.context.get(); if (model == nullptr || ctx == nullptr) { fprintf(stderr, "%s : failed to init\n", __func__); @@ -42,7 +42,7 @@ int main(int argc, char ** argv) { llama_sampler * smpl = llama_sampler_chain_init(sparams); - llama_sampler_chain_add(smpl, llama_sampler_init_dist(params.sparams.seed)); + llama_sampler_chain_add(smpl, llama_sampler_init_dist(params.sampling.seed)); // tokenize prompt auto tokens = common_tokenize(ctx, params.prompt, true); @@ -89,8 +89,6 @@ int main(int argc, char ** argv) { if (llama_decode(ctx, batch)) { fprintf(stderr, "\n%s : failed to evaluate\n", __func__); llama_batch_free(batch); - llama_free(ctx); - llama_free_model(model); return 1; } n_past += 1; @@ -98,15 +96,12 @@ int main(int argc, char ** argv) { printf("\n\n"); - // free old context - llama_free(ctx); - // make new context - auto * ctx2 = llama_new_context_with_model(model, common_context_params_to_llama(params)); + llama_context * ctx2 = llama_init_from_model(model, common_context_params_to_llama(params)); llama_sampler * smpl2 = llama_sampler_chain_init(sparams); - llama_sampler_chain_add(smpl2, llama_sampler_init_dist(params.sparams.seed)); + llama_sampler_chain_add(smpl2, llama_sampler_init_dist(params.sampling.seed)); printf("\nsecond run: %s", params.prompt.c_str()); @@ -123,8 +118,6 @@ int main(int argc, char ** argv) { if (read != llama_state_set_data(ctx2, state_mem.data(), state_mem.size())) { fprintf(stderr, "\n%s : failed to read state\n", __func__); - llama_free(ctx2); - llama_free_model(model); return 1; } @@ -148,8 +141,6 @@ int main(int argc, char ** argv) { if (llama_decode(ctx2, batch)) { fprintf(stderr, "\n%s : failed to evaluate\n", __func__); llama_batch_free(batch); - llama_free(ctx2); - llama_free_model(model); return 1; } n_past += 1; @@ -157,19 +148,17 @@ int main(int argc, char ** argv) { printf("\n\n"); - llama_free(ctx2); - if (result0 != result1) { fprintf(stderr, "\n%s : error : the 2 generations are different\n", __func__); return 1; } // make new context - auto * ctx3 = llama_new_context_with_model(model, common_context_params_to_llama(params)); + llama_context * ctx3 = llama_init_from_model(model, common_context_params_to_llama(params)); llama_sampler * smpl3 = llama_sampler_chain_init(sparams); - llama_sampler_chain_add(smpl3, llama_sampler_init_dist(params.sparams.seed)); + llama_sampler_chain_add(smpl3, llama_sampler_init_dist(params.sampling.seed)); printf("\nsingle seq run: %s", params.prompt.c_str()); @@ -186,8 +175,6 @@ int main(int argc, char ** argv) { if (read != llama_state_set_data(ctx3, state_mem.data(), state_mem.size())) { fprintf(stderr, "\n%s : failed to read state\n", __func__); - llama_free(ctx3); - llama_free_model(model); return 1; } @@ -204,8 +191,6 @@ int main(int argc, char ** argv) { const size_t ncopy = llama_state_seq_get_data(ctx3, seq_store.data(), seq_store.size(), 0); if (ncopy != seq_store.size()) { fprintf(stderr, "\n%s : seq copy data length %zd does not match expected length %zd\n", __func__, ncopy, seq_store.size()); - llama_free(ctx3); - llama_free_model(model); return 1; } fprintf(stderr, "%s : seq 0 copied, %zd bytes\n", __func__, ncopy); @@ -218,8 +203,6 @@ int main(int argc, char ** argv) { const size_t nset = llama_state_seq_set_data(ctx3, seq_store.data(), seq_store.size(), 1); if (nset != seq_store.size()) { fprintf(stderr, "\n%s : seq set data length %zd does not match expected length %zd\n", __func__, nset, seq_store.size()); - llama_free(ctx3); - llama_free_model(model); return 1; } fprintf(stderr, "%s : seq 1 restored, %zd bytes\n", __func__, nset); @@ -239,8 +222,6 @@ int main(int argc, char ** argv) { if (llama_decode(ctx3, batch)) { fprintf(stderr, "\n%s : failed to evaluate\n", __func__); llama_batch_free(batch); - llama_free(ctx3); - llama_free_model(model); return 1; } n_past += 1; @@ -253,8 +234,6 @@ int main(int argc, char ** argv) { llama_sampler_free(smpl3); llama_batch_free(batch); - llama_free(ctx3); - llama_free_model(model); if (result0 != result2) { fprintf(stderr, "\n%s : error : the seq restore generation is different\n", __func__); diff --git a/examples/server/CMakeLists.txt b/examples/server/CMakeLists.txt index 93e876f5a..1b7cc8c13 100644 --- a/examples/server/CMakeLists.txt +++ b/examples/server/CMakeLists.txt @@ -15,13 +15,8 @@ set(TARGET_SRCS httplib.h ) set(PUBLIC_ASSETS - index.html - completion.js + index.html.gz loading.html - deps_daisyui.min.css - deps_markdown-it.js - deps_tailwindcss.js - deps_vue.esm-browser.js ) foreach(asset ${PUBLIC_ASSETS}) @@ -33,11 +28,13 @@ foreach(asset ${PUBLIC_ASSETS}) OUTPUT "${output}" COMMAND "${CMAKE_COMMAND}" "-DINPUT=${input}" "-DOUTPUT=${output}" -P "${PROJECT_SOURCE_DIR}/scripts/xxd.cmake" ) + set_source_files_properties(${output} PROPERTIES GENERATED TRUE) endforeach() add_executable(${TARGET} ${TARGET_SRCS}) install(TARGETS ${TARGET} RUNTIME) +target_include_directories(${TARGET} PRIVATE ${CMAKE_SOURCE_DIR}) target_link_libraries(${TARGET} PRIVATE common ${CMAKE_THREAD_LIBS_INIT}) if (LLAMA_SERVER_SSL) @@ -50,4 +47,4 @@ if (WIN32) TARGET_LINK_LIBRARIES(${TARGET} PRIVATE ws2_32) endif() -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/server/README.md b/examples/server/README.md index 562494077..1f0a27d96 100644 --- a/examples/server/README.md +++ b/examples/server/README.md @@ -39,16 +39,13 @@ The project is under active development, and we are [looking for feedback and co | `--cpu-strict-batch <0\|1>` | use strict CPU placement (default: same as --cpu-strict) | | `--prio-batch N` | set process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: 0)
| | `--poll-batch <0\|1>` | use polling to wait for work (default: same as --poll) | -| `-c, --ctx-size N` | size of the prompt context (default: 0, 0 = loaded from model)
(env: LLAMA_ARG_CTX_SIZE) | +| `-c, --ctx-size N` | size of the prompt context (default: 4096, 0 = loaded from model)
(env: LLAMA_ARG_CTX_SIZE) | | `-n, --predict, --n-predict N` | number of tokens to predict (default: -1, -1 = infinity, -2 = until context filled)
(env: LLAMA_ARG_N_PREDICT) | | `-b, --batch-size N` | logical maximum batch size (default: 2048)
(env: LLAMA_ARG_BATCH) | | `-ub, --ubatch-size N` | physical maximum batch size (default: 512)
(env: LLAMA_ARG_UBATCH) | | `--keep N` | number of tokens to keep from the initial prompt (default: 0, -1 = all) | | `-fa, --flash-attn` | enable Flash Attention (default: disabled)
(env: LLAMA_ARG_FLASH_ATTN) | -| `-p, --prompt PROMPT` | prompt to start generation with | | `--no-perf` | disable internal libllama performance timings (default: false)
(env: LLAMA_ARG_NO_PERF) | -| `-f, --file FNAME` | a file containing the prompt (default: none) | -| `-bf, --binary-file FNAME` | binary file containing the prompt (default: none) | | `-e, --escape` | process escapes sequences (\n, \r, \t, \', \", \\) (default: true) | | `--no-escape` | do not process escape sequences | | `--rope-scaling {none,linear,yarn}` | RoPE frequency scaling method, defaults to linear unless specified by the model
(env: LLAMA_ARG_ROPE_SCALING_TYPE) | @@ -62,13 +59,15 @@ The project is under active development, and we are [looking for feedback and co | `--yarn-beta-fast N` | YaRN: low correction dim or beta (default: 32.0)
(env: LLAMA_ARG_YARN_BETA_FAST) | | `-dkvc, --dump-kv-cache` | verbose print of the KV cache | | `-nkvo, --no-kv-offload` | disable KV offload
(env: LLAMA_ARG_NO_KV_OFFLOAD) | -| `-ctk, --cache-type-k TYPE` | KV cache data type for K (default: f16)
(env: LLAMA_ARG_CACHE_TYPE_K) | -| `-ctv, --cache-type-v TYPE` | KV cache data type for V (default: f16)
(env: LLAMA_ARG_CACHE_TYPE_V) | -| `-dt, --defrag-thold N` | KV cache defragmentation threshold (default: -1.0, < 0 - disabled)
(env: LLAMA_ARG_DEFRAG_THOLD) | +| `-ctk, --cache-type-k TYPE` | KV cache data type for K
allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1
(default: f16)
(env: LLAMA_ARG_CACHE_TYPE_K) | +| `-ctv, --cache-type-v TYPE` | KV cache data type for V
allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1
(default: f16)
(env: LLAMA_ARG_CACHE_TYPE_V) | +| `-dt, --defrag-thold N` | KV cache defragmentation threshold (default: 0.1, < 0 - disabled)
(env: LLAMA_ARG_DEFRAG_THOLD) | | `-np, --parallel N` | number of parallel sequences to decode (default: 1)
(env: LLAMA_ARG_N_PARALLEL) | | `--mlock` | force system to keep model in RAM rather than swapping or compressing
(env: LLAMA_ARG_MLOCK) | | `--no-mmap` | do not memory-map model (slower load but may reduce pageouts if not using mlock)
(env: LLAMA_ARG_NO_MMAP) | | `--numa TYPE` | attempt optimizations that help on some NUMA systems
- distribute: spread execution evenly over all nodes
- isolate: only spawn threads on CPUs on the node that execution started on
- numactl: use the CPU map provided by numactl
if run without this previously, it is recommended to drop the system page cache before using this
see https://github.com/ggerganov/llama.cpp/issues/1437
(env: LLAMA_ARG_NUMA) | +| `-dev, --device ` | comma-separated list of devices to use for offloading (none = don't offload)
use --list-devices to see a list of available devices
(env: LLAMA_ARG_DEVICE) | +| `--list-devices` | print list of available devices and exit | | `-ngl, --gpu-layers, --n-gpu-layers N` | number of layers to store in VRAM
(env: LLAMA_ARG_N_GPU_LAYERS) | | `-sm, --split-mode {none,layer,row}` | how to split the model across multiple GPUs, one of:
- none: use one GPU only
- layer (default): split layers and KV across GPUs
- row: split rows across GPUs
(env: LLAMA_ARG_SPLIT_MODE) | | `-ts, --tensor-split N0,N1,N2,...` | fraction of the model to offload to each GPU, comma-separated list of proportions, e.g. 3,1
(env: LLAMA_ARG_TENSOR_SPLIT) | @@ -85,7 +84,6 @@ The project is under active development, and we are [looking for feedback and co | `-hfr, --hf-repo REPO` | Hugging Face model repository (default: unused)
(env: LLAMA_ARG_HF_REPO) | | `-hff, --hf-file FILE` | Hugging Face model file (default: unused)
(env: LLAMA_ARG_HF_FILE) | | `-hft, --hf-token TOKEN` | Hugging Face access token (default: value from HF_TOKEN environment variable)
(env: HF_TOKEN) | -| `-ld, --logdir LOGDIR` | path under which to save YAML logs (no logging if unset) | | `--log-disable` | Log disable | | `--log-file FNAME` | Log to file | | `--log-colors` | Enable colored logging
(env: LLAMA_LOG_COLORS) | @@ -99,25 +97,26 @@ The project is under active development, and we are [looking for feedback and co | Argument | Explanation | | -------- | ----------- | -| `--samplers SAMPLERS` | samplers that will be used for generation in the order, separated by ';'
(default: top_k;typ_p;top_p;min_p;temperature) | +| `--samplers SAMPLERS` | samplers that will be used for generation in the order, separated by ';'
(default: dry;top_k;typ_p;top_p;min_p;xtc;temperature) | | `-s, --seed SEED` | RNG seed (default: -1, use random seed for -1) | -| `--sampling-seq SEQUENCE` | simplified sequence for samplers that will be used (default: kfypmt) | +| `--sampling-seq SEQUENCE` | simplified sequence for samplers that will be used (default: dkypmxt) | | `--ignore-eos` | ignore end of stream token and continue generating (implies --logit-bias EOS-inf) | -| `--penalize-nl` | penalize newline tokens (default: false) | | `--temp N` | temperature (default: 0.8) | | `--top-k N` | top-k sampling (default: 40, 0 = disabled) | | `--top-p N` | top-p sampling (default: 0.9, 1.0 = disabled) | | `--min-p N` | min-p sampling (default: 0.1, 0.0 = disabled) | +| `--xtc-probability N` | xtc probability (default: 0.0, 0.0 = disabled) | +| `--xtc-threshold N` | xtc threshold (default: 0.1, 1.0 = disabled) | | `--typical N` | locally typical sampling, parameter p (default: 1.0, 1.0 = disabled) | | `--repeat-last-n N` | last n tokens to consider for penalize (default: 64, 0 = disabled, -1 = ctx_size) | | `--repeat-penalty N` | penalize repeat sequence of tokens (default: 1.0, 1.0 = disabled) | | `--presence-penalty N` | repeat alpha presence penalty (default: 0.0, 0.0 = disabled) | | `--frequency-penalty N` | repeat alpha frequency penalty (default: 0.0, 0.0 = disabled) | -| `--dry-multiplier N` | DRY sampling multiplier (default: 0.0, 0.0 = disabled) | -| `--dry-base N` | DRY sampling base value (default: 1.75) | -| `--dry-allowed-length N` | allowed length for DRY sampling (default: 2) | -| `--dry-penalty-last-n N` | DRY penalty for the last n tokens (default: -1, 0 = disable, -1 = context size) | -| `--dry-sequence-breaker STRING` | add sequence breaker for DRY sampling, clearing out default breakers (`['\n', ':', '"', '*']`) in the process; use `"none"` to not use any sequence breakers +| `--dry-multiplier N` | set DRY sampling multiplier (default: 0.0, 0.0 = disabled) | +| `--dry-base N` | set DRY sampling base value (default: 1.75) | +| `--dry-allowed-length N` | set allowed length for DRY sampling (default: 2) | +| `--dry-penalty-last-n N` | set DRY penalty for the last n tokens (default: -1, 0 = disable, -1 = context size) | +| `--dry-sequence-breaker STRING` | add sequence breaker for DRY sampling, clearing out default breakers ('\n', ':', '"', '*') in the process; use "none" to not use any sequence breakers
| | `--dynatemp-range N` | dynamic temperature range (default: 0.0, 0.0 = disabled) | | `--dynatemp-exp N` | dynamic temperature exponent (default: 1.0) | | `--mirostat N` | use Mirostat sampling.
Top K, Nucleus and Locally Typical samplers are ignored if used.
(default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0) | @@ -135,6 +134,7 @@ The project is under active development, and we are [looking for feedback and co | -------- | ----------- | | `--no-context-shift` | disables context shift on inifinite text generation (default: disabled)
(env: LLAMA_ARG_NO_CONTEXT_SHIFT) | | `-sp, --special` | special tokens output enabled (default: false) | +| `--no-warmup` | skip warming up the model with an empty run | | `--spm-infill` | use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: disabled) | | `--pooling {none,mean,cls,last,rank}` | pooling type for embeddings, use model default if unspecified
(env: LLAMA_ARG_POOLING) | | `-cb, --cont-batching` | enable continuous batching (a.k.a dynamic batching) (default: enabled)
(env: LLAMA_ARG_CONT_BATCHING) | @@ -143,6 +143,7 @@ The project is under active development, and we are [looking for feedback and co | `--host HOST` | ip address to listen (default: 127.0.0.1)
(env: LLAMA_ARG_HOST) | | `--port PORT` | port to listen (default: 8080)
(env: LLAMA_ARG_PORT) | | `--path PATH` | path to serve static files from (default: )
(env: LLAMA_ARG_STATIC_PATH) | +| `--no-webui` | Disable the Web UI (default: enabled)
(env: LLAMA_ARG_NO_WEBUI) | | `--embedding, --embeddings` | restrict to only support embedding use case; use only with dedicated embedding models (default: disabled)
(env: LLAMA_ARG_EMBEDDINGS) | | `--reranking, --rerank` | enable reranking endpoint on server (default: disabled)
(env: LLAMA_ARG_RERANKING) | | `--api-key KEY` | API key to use for authentication (default: none)
(env: LLAMA_API_KEY) | @@ -157,9 +158,16 @@ The project is under active development, and we are [looking for feedback and co | `--props` | enable changing global properties via POST /props (default: disabled)
(env: LLAMA_ARG_ENDPOINT_PROPS) | | `--no-slots` | disables slots monitoring endpoint
(env: LLAMA_ARG_NO_ENDPOINT_SLOTS) | | `--slot-save-path PATH` | path to save slot kv cache (default: disabled) | -| `--chat-template JINJA_TEMPLATE` | set custom jinja chat template (default: template taken from model's metadata)
if suffix/prefix are specified, template will be disabled
only commonly used templates are accepted:
https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template
(env: LLAMA_ARG_CHAT_TEMPLATE) | +| `--chat-template JINJA_TEMPLATE` | set custom jinja chat template (default: template taken from model's metadata)
if suffix/prefix are specified, template will be disabled
list of built-in templates:
chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, exaone3, gemma, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, monarch, openchat, orion, phi3, rwkv-world, vicuna, vicuna-orca, zephyr
(env: LLAMA_ARG_CHAT_TEMPLATE) | | `-sps, --slot-prompt-similarity SIMILARITY` | how much the prompt of a request must match the prompt of a slot in order to use that slot (default: 0.50, 0.0 = disabled)
| | `--lora-init-without-apply` | load LoRA adapters without applying them (apply later via POST /lora-adapters) (default: disabled) | +| `--draft-max, --draft, --draft-n N` | number of tokens to draft for speculative decoding (default: 16)
(env: LLAMA_ARG_DRAFT_MAX) | +| `--draft-min, --draft-n-min N` | minimum number of draft tokens to use for speculative decoding (default: 5)
(env: LLAMA_ARG_DRAFT_MIN) | +| `--draft-p-min P` | minimum speculative decoding probability (greedy) (default: 0.9)
(env: LLAMA_ARG_DRAFT_P_MIN) | +| `-cd, --ctx-size-draft N` | size of the prompt context for the draft model (default: 0, 0 = loaded from model)
(env: LLAMA_ARG_CTX_SIZE_DRAFT) | +| `-devd, --device-draft ` | comma-separated list of devices to use for offloading the draft model (none = don't offload)
use --list-devices to see a list of available devices | +| `-ngld, --gpu-layers-draft, --n-gpu-layers-draft N` | number of layers to store in VRAM for the draft model
(env: LLAMA_ARG_N_GPU_LAYERS_DRAFT) | +| `-md, --model-draft FNAME` | draft model for speculative decoding (default: unused)
(env: LLAMA_ARG_MODEL_DRAFT) | Note: If both command line argument and environment variable are both set for the same param, the argument will take precedence over env var. @@ -187,12 +195,6 @@ services: `llama-server` is built alongside everything else from the root of the project -- Using `make`: - - ```bash - make llama-server - ``` - - Using `CMake`: ```bash @@ -206,15 +208,6 @@ services: `llama-server` can also be built with SSL support using OpenSSL 3 -- Using `make`: - - ```bash - # NOTE: For non-system openssl, use the following: - # CXXFLAGS="-I /path/to/openssl/include" - # LDFLAGS="-L /path/to/openssl/lib" - make LLAMA_SERVER_SSL=true llama-server - ``` - - Using `CMake`: ```bash @@ -222,6 +215,37 @@ services: cmake --build build --config Release -t llama-server ``` +## Web UI + +The project includes a web-based user interface that enables interaction with the model through the `/chat/completions` endpoint. + +The web UI is developed using: +- `vue` framework for frontend development +- `tailwindcss` and `daisyui` for styling +- `vite` for build tooling + +A pre-built version is available as a single HTML file under `/public` directory. + +To build or to run the dev server (with hot reload): + +```sh +# make sure you have nodejs installed +cd examples/server/webui +npm i + +# to run the dev server +npm run dev + +# to build the public/index.html +npm run build +``` + +NOTE: if you are using the vite dev server, you can change the API base URL to llama.cpp. To do that, run this code snippet in browser's console: + +```js +localStorage.setItem('base', 'http://localhost:8080') +``` + ## Quick Start To get started right away, run the following command, making sure to use the correct path for the model you have: @@ -276,23 +300,23 @@ mkdir llama-client cd llama-client ``` -Create a index.js file and put this inside: +Create an index.js file and put this inside: ```javascript -const prompt = `Building a website can be done in 10 simple steps:`; +const prompt = "Building a website can be done in 10 simple steps:" -async function Test() { +async function test() { let response = await fetch("http://127.0.0.1:8080/completion", { - method: 'POST', + method: "POST", body: JSON.stringify({ prompt, - n_predict: 512, + n_predict: 64, }) }) console.log((await response.json()).content) } -Test() +test() ``` And run it: @@ -316,149 +340,198 @@ node index.js ### POST `/completion`: Given a `prompt`, it returns the predicted completion. - *Options:* +> [!IMPORTANT] +> +> This endpoint is **not** OAI-compatible. For OAI-compatible client, use `/v1/completions` instead. - `prompt`: Provide the prompt for this completion as a string or as an array of strings or numbers representing tokens. Internally, if `cache_prompt` is `true`, the prompt is compared to the previous completion and only the "unseen" suffix is evaluated. A `BOS` token is inserted at the start, if all of the following conditions are true: +*Options:* - - The prompt is a string or an array with the first element given as a string - - The model's `tokenizer.ggml.add_bos_token` metadata is `true` +`prompt`: Provide the prompt for this completion as a string or as an array of strings or numbers representing tokens. Internally, if `cache_prompt` is `true`, the prompt is compared to the previous completion and only the "unseen" suffix is evaluated. A `BOS` token is inserted at the start, if all of the following conditions are true: - These input shapes and data type are allowed for `prompt`: + - The prompt is a string or an array with the first element given as a string + - The model's `tokenizer.ggml.add_bos_token` metadata is `true` - - Single string: `"string"` - - Single sequence of tokens: `[12, 34, 56]` - - Mixed tokens and strings: `[12, 34, "string", 56, 78]` +These input shapes and data type are allowed for `prompt`: - Multiple prompts are also supported. In this case, the completion result will be an array. + - Single string: `"string"` + - Single sequence of tokens: `[12, 34, 56]` + - Mixed tokens and strings: `[12, 34, "string", 56, 78]` - - Only strings: `["string1", "string2"]` - - Strings and sequences of tokens: `["string1", [12, 34, 56]]` - - Mixed types: `[[12, 34, "string", 56, 78], [12, 34, 56], "string"]` +Multiple prompts are also supported. In this case, the completion result will be an array. - `temperature`: Adjust the randomness of the generated text. Default: `0.8` + - Only strings: `["string1", "string2"]` + - Strings and sequences of tokens: `["string1", [12, 34, 56]]` + - Mixed types: `[[12, 34, "string", 56, 78], [12, 34, 56], "string"]` - `dynatemp_range`: Dynamic temperature range. The final temperature will be in the range of `[temperature - dynatemp_range; temperature + dynatemp_range]` Default: `0.0`, which is disabled. +`temperature`: Adjust the randomness of the generated text. Default: `0.8` - `dynatemp_exponent`: Dynamic temperature exponent. Default: `1.0` +`dynatemp_range`: Dynamic temperature range. The final temperature will be in the range of `[temperature - dynatemp_range; temperature + dynatemp_range]` Default: `0.0`, which is disabled. - `top_k`: Limit the next token selection to the K most probable tokens. Default: `40` +`dynatemp_exponent`: Dynamic temperature exponent. Default: `1.0` - `top_p`: Limit the next token selection to a subset of tokens with a cumulative probability above a threshold P. Default: `0.95` +`top_k`: Limit the next token selection to the K most probable tokens. Default: `40` - `min_p`: The minimum probability for a token to be considered, relative to the probability of the most likely token. Default: `0.05` +`top_p`: Limit the next token selection to a subset of tokens with a cumulative probability above a threshold P. Default: `0.95` - `n_predict`: Set the maximum number of tokens to predict when generating text. **Note:** May exceed the set limit slightly if the last token is a partial multibyte character. When 0, no tokens will be generated but the prompt is evaluated into the cache. Default: `-1`, where `-1` is infinity. +`min_p`: The minimum probability for a token to be considered, relative to the probability of the most likely token. Default: `0.05` - `n_indent`: Specify the minimum line indentation for the generated text in number of whitespace characters. Useful for code completion tasks. Default: `0` +`n_predict`: Set the maximum number of tokens to predict when generating text. **Note:** May exceed the set limit slightly if the last token is a partial multibyte character. When 0, no tokens will be generated but the prompt is evaluated into the cache. Default: `-1`, where `-1` is infinity. - `n_keep`: Specify the number of tokens from the prompt to retain when the context size is exceeded and tokens need to be discarded. The number excludes the BOS token. - By default, this value is set to `0`, meaning no tokens are kept. Use `-1` to retain all tokens from the prompt. +`n_indent`: Specify the minimum line indentation for the generated text in number of whitespace characters. Useful for code completion tasks. Default: `0` - `stream`: It allows receiving each predicted token in real-time instead of waiting for the completion to finish. To enable this, set to `true`. +`n_keep`: Specify the number of tokens from the prompt to retain when the context size is exceeded and tokens need to be discarded. The number excludes the BOS token. +By default, this value is set to `0`, meaning no tokens are kept. Use `-1` to retain all tokens from the prompt. - `stop`: Specify a JSON array of stopping strings. - These words will not be included in the completion, so make sure to add them to the prompt for the next iteration. Default: `[]` +`stream`: Allows receiving each predicted token in real-time instead of waiting for the completion to finish (uses a different response format). To enable this, set to `true`. - `typical_p`: Enable locally typical sampling with parameter p. Default: `1.0`, which is disabled. +`stop`: Specify a JSON array of stopping strings. +These words will not be included in the completion, so make sure to add them to the prompt for the next iteration. Default: `[]` - `repeat_penalty`: Control the repetition of token sequences in the generated text. Default: `1.1` +`typical_p`: Enable locally typical sampling with parameter p. Default: `1.0`, which is disabled. - `repeat_last_n`: Last n tokens to consider for penalizing repetition. Default: `64`, where `0` is disabled and `-1` is ctx-size. +`repeat_penalty`: Control the repetition of token sequences in the generated text. Default: `1.1` - `penalize_nl`: Penalize newline tokens when applying the repeat penalty. Default: `true` +`repeat_last_n`: Last n tokens to consider for penalizing repetition. Default: `64`, where `0` is disabled and `-1` is ctx-size. - `presence_penalty`: Repeat alpha presence penalty. Default: `0.0`, which is disabled. +`presence_penalty`: Repeat alpha presence penalty. Default: `0.0`, which is disabled. - `frequency_penalty`: Repeat alpha frequency penalty. Default: `0.0`, which is disabled. +`frequency_penalty`: Repeat alpha frequency penalty. Default: `0.0`, which is disabled. - `dry_multiplier`: Set the DRY (Don't Repeat Yourself) repetition penalty multiplier. Default: `0.0`, which is disabled. +`dry_multiplier`: Set the DRY (Don't Repeat Yourself) repetition penalty multiplier. Default: `0.0`, which is disabled. - `dry_base`: Set the DRY repetition penalty base value. Default: `1.75` +`dry_base`: Set the DRY repetition penalty base value. Default: `1.75` - `dry_allowed_length`: Tokens that extend repetition beyond this receive exponentially increasing penalty: multiplier * base ^ (length of repeating sequence before token - allowed length). Default: `2` +`dry_allowed_length`: Tokens that extend repetition beyond this receive exponentially increasing penalty: multiplier * base ^ (length of repeating sequence before token - allowed length). Default: `2` - `dry_penalty_last_n`: How many tokens to scan for repetitions. Default: `-1`, where `0` is disabled and `-1` is context size. +`dry_penalty_last_n`: How many tokens to scan for repetitions. Default: `-1`, where `0` is disabled and `-1` is context size. - `dry_sequence_breakers`: Specify an array of sequence breakers for DRY sampling. Only a JSON array of strings is accepted. Default: `['\n', ':', '"', '*']` +`dry_sequence_breakers`: Specify an array of sequence breakers for DRY sampling. Only a JSON array of strings is accepted. Default: `['\n', ':', '"', '*']` - `mirostat`: Enable Mirostat sampling, controlling perplexity during text generation. Default: `0`, where `0` is disabled, `1` is Mirostat, and `2` is Mirostat 2.0. +`xtc_probability`: Set the chance for token removal via XTC sampler. Default: `0.0`, which is disabled. - `mirostat_tau`: Set the Mirostat target entropy, parameter tau. Default: `5.0` +`xtc_threshold`: Set a minimum probability threshold for tokens to be removed via XTC sampler. Default: `0.1` (> `0.5` disables XTC) - `mirostat_eta`: Set the Mirostat learning rate, parameter eta. Default: `0.1` +`mirostat`: Enable Mirostat sampling, controlling perplexity during text generation. Default: `0`, where `0` is disabled, `1` is Mirostat, and `2` is Mirostat 2.0. - `grammar`: Set grammar for grammar-based sampling. Default: no grammar +`mirostat_tau`: Set the Mirostat target entropy, parameter tau. Default: `5.0` - `json_schema`: Set a JSON schema for grammar-based sampling (e.g. `{"items": {"type": "string"}, "minItems": 10, "maxItems": 100}` of a list of strings, or `{}` for any JSON). See [tests](../../tests/test-json-schema-to-grammar.cpp) for supported features. Default: no JSON schema. +`mirostat_eta`: Set the Mirostat learning rate, parameter eta. Default: `0.1` - `seed`: Set the random number generator (RNG) seed. Default: `-1`, which is a random seed. +`grammar`: Set grammar for grammar-based sampling. Default: no grammar - `ignore_eos`: Ignore end of stream token and continue generating. Default: `false` +`json_schema`: Set a JSON schema for grammar-based sampling (e.g. `{"items": {"type": "string"}, "minItems": 10, "maxItems": 100}` of a list of strings, or `{}` for any JSON). See [tests](../../tests/test-json-schema-to-grammar.cpp) for supported features. Default: no JSON schema. - `logit_bias`: Modify the likelihood of a token appearing in the generated text completion. For example, use `"logit_bias": [[15043,1.0]]` to increase the likelihood of the token 'Hello', or `"logit_bias": [[15043,-1.0]]` to decrease its likelihood. Setting the value to false, `"logit_bias": [[15043,false]]` ensures that the token `Hello` is never produced. The tokens can also be represented as strings, e.g. `[["Hello, World!",-0.5]]` will reduce the likelihood of all the individual tokens that represent the string `Hello, World!`, just like the `presence_penalty` does. Default: `[]` +`seed`: Set the random number generator (RNG) seed. Default: `-1`, which is a random seed. - `n_probs`: If greater than 0, the response also contains the probabilities of top N tokens for each generated token given the sampling settings. Note that for temperature < 0 the tokens are sampled greedily but token probabilities are still being calculated via a simple softmax of the logits without considering any other sampler settings. Default: `0` +`ignore_eos`: Ignore end of stream token and continue generating. Default: `false` - `min_keep`: If greater than 0, force samplers to return N possible tokens at minimum. Default: `0` +`logit_bias`: Modify the likelihood of a token appearing in the generated text completion. For example, use `"logit_bias": [[15043,1.0]]` to increase the likelihood of the token 'Hello', or `"logit_bias": [[15043,-1.0]]` to decrease its likelihood. Setting the value to false, `"logit_bias": [[15043,false]]` ensures that the token `Hello` is never produced. The tokens can also be represented as strings, e.g. `[["Hello, World!",-0.5]]` will reduce the likelihood of all the individual tokens that represent the string `Hello, World!`, just like the `presence_penalty` does. Default: `[]` - `t_max_predict_ms`: Set a time limit in milliseconds for the prediction (a.k.a. text-generation) phase. The timeout will trigger if the generation takes more than the specified time (measured since the first token was generated) and if a new-line character has already been generated. Useful for FIM applications. Default: `0`, which is disabled. +`n_probs`: If greater than 0, the response also contains the probabilities of top N tokens for each generated token given the sampling settings. Note that for temperature < 0 the tokens are sampled greedily but token probabilities are still being calculated via a simple softmax of the logits without considering any other sampler settings. Default: `0` - `image_data`: An array of objects to hold base64-encoded image `data` and its `id`s to be reference in `prompt`. You can determine the place of the image in the prompt as in the following: `USER:[img-12]Describe the image in detail.\nASSISTANT:`. In this case, `[img-12]` will be replaced by the embeddings of the image with id `12` in the following `image_data` array: `{..., "image_data": [{"data": "", "id": 12}]}`. Use `image_data` only with multimodal models, e.g., LLaVA. +`min_keep`: If greater than 0, force samplers to return N possible tokens at minimum. Default: `0` - `id_slot`: Assign the completion task to an specific slot. If is -1 the task will be assigned to a Idle slot. Default: `-1` +`t_max_predict_ms`: Set a time limit in milliseconds for the prediction (a.k.a. text-generation) phase. The timeout will trigger if the generation takes more than the specified time (measured since the first token was generated) and if a new-line character has already been generated. Useful for FIM applications. Default: `0`, which is disabled. - `cache_prompt`: Re-use KV cache from a previous request if possible. This way the common prefix does not have to be re-processed, only the suffix that differs between the requests. Because (depending on the backend) the logits are **not** guaranteed to be bit-for-bit identical for different batch sizes (prompt processing vs. token generation) enabling this option can cause nondeterministic results. Default: `false` +`image_data`: An array of objects to hold base64-encoded image `data` and its `id`s to be reference in `prompt`. You can determine the place of the image in the prompt as in the following: `USER:[img-12]Describe the image in detail.\nASSISTANT:`. In this case, `[img-12]` will be replaced by the embeddings of the image with id `12` in the following `image_data` array: `{..., "image_data": [{"data": "", "id": 12}]}`. Use `image_data` only with multimodal models, e.g., LLaVA. - `samplers`: The order the samplers should be applied in. An array of strings representing sampler type names. If a sampler is not set, it will not be used. If a sampler is specified more than once, it will be applied multiple times. Default: `["top_k", "typical_p", "top_p", "min_p", "temperature"]` - these are all the available values. +`id_slot`: Assign the completion task to an specific slot. If is -1 the task will be assigned to a Idle slot. Default: `-1` + +`cache_prompt`: Re-use KV cache from a previous request if possible. This way the common prefix does not have to be re-processed, only the suffix that differs between the requests. Because (depending on the backend) the logits are **not** guaranteed to be bit-for-bit identical for different batch sizes (prompt processing vs. token generation) enabling this option can cause nondeterministic results. Default: `true` + +`return_tokens`: Return the raw generated token ids in the `tokens` field. Otherwise `tokens` remains empty. Default: `false` + +`samplers`: The order the samplers should be applied in. An array of strings representing sampler type names. If a sampler is not set, it will not be used. If a sampler is specified more than once, it will be applied multiple times. Default: `["dry", "top_k", "typ_p", "top_p", "min_p", "xtc", "temperature"]` - these are all the available values. + +`timings_per_token`: Include prompt processing and text generation speed information in each response. Default: `false` + +`post_sampling_probs`: Returns the probabilities of top `n_probs` tokens after applying sampling chain. + +`response_fields`: A list of response fields, for example: `"response_fields": ["content", "generation_settings/n_predict"]`. If the specified field is missing, it will simply be omitted from the response without triggering an error. Note that fields with a slash will be unnested; for example, `generation_settings/n_predict` will move the field `n_predict` from the `generation_settings` object to the root of the response and give it a new name. + +`lora`: A list of LoRA adapters to be applied to this specific request. Each object in the list must contain `id` and `scale` fields. For example: `[{"id": 0, "scale": 0.5}, {"id": 1, "scale": 1.1}]`. If a LoRA adapter is not specified in the list, its scale will default to `0.0`. Please note that requests with different LoRA configurations will not be batched together, which may result in performance degradation. **Response format** -- Note: When using streaming mode (`stream`), only `content` and `stop` will be returned until end of completion. +- Note: In streaming mode (`stream`), only `content`, `tokens` and `stop` will be returned until end of completion. Responses are sent using the [Server-sent events](https://html.spec.whatwg.org/multipage/server-sent-events.html) standard. Note: the browser's `EventSource` interface cannot be used due to its lack of `POST` request support. -- `completion_probabilities`: An array of token probabilities for each completion. The array's length is `n_predict`. Each item in the array has the following structure: - -```json -{ - "content": "", - "probs": [ - { - "prob": float, - "tok_str": "" - }, - { - "prob": float, - "tok_str": "" - }, +- `completion_probabilities`: An array of token probabilities for each completion. The array's length is `n_predict`. Each item in the array has a nested array `top_logprobs`. It contains at **maximum** `n_probs` elements: + ```json + { + "content": "", + "tokens": [ generated token ids if requested ], ... - ] -}, -``` - -Notice that each `probs` is an array of length `n_probs`. + "probs": [ + { + "id": , + "logprob": float, + "token": "", + "bytes": [int, int, ...], + "top_logprobs": [ + { + "id": , + "logprob": float, + "token": "", + "bytes": [int, int, ...], + }, + { + "id": , + "logprob": float, + "token": "", + "bytes": [int, int, ...], + }, + ... + ] + }, + { + "id": , + "logprob": float, + "token": "", + "bytes": [int, int, ...], + "top_logprobs": [ + ... + ] + }, + ... + ] + }, + ``` + Please note that if `post_sampling_probs` is set to `true`: + - `logprob` will be replaced with `prob`, with the value between 0.0 and 1.0 + - `top_logprobs` will be replaced with `top_probs`. Each element contains: + - `id`: token ID + - `token`: token in string + - `bytes`: token in bytes + - `prob`: token probability, with the value between 0.0 and 1.0 + - Number of elements in `top_probs` may be less than `n_probs` - `content`: Completion result as a string (excluding `stopping_word` if any). In case of streaming mode, will contain the next token as a string. +- `tokens`: Same as `content` but represented as raw token ids. Only populated if `"return_tokens": true` or `"stream": true` in the request. - `stop`: Boolean for use with `stream` to check whether the generation has stopped (Note: This is not related to stopping words array `stop` from input options) - `generation_settings`: The provided options above excluding `prompt` but including `n_ctx`, `model`. These options may differ from the original ones in some way (e.g. bad values filtered out, strings converted to tokens, etc.). -- `model`: The path to the model loaded with `-m` -- `prompt`: The provided `prompt` -- `stopped_eos`: Indicating whether the completion has stopped because it encountered the EOS token -- `stopped_limit`: Indicating whether the completion stopped because `n_predict` tokens were generated before stop words or EOS was encountered -- `stopped_word`: Indicating whether the completion stopped due to encountering a stopping word from `stop` JSON array provided +- `model`: The model alias (for model path, please use `/props` endpoint) +- `prompt`: The processed `prompt` (special tokens may be added) +- `stop_type`: Indicating whether the completion has stopped. Possible values are: + - `none`: Generating (not stopped) + - `eos`: Stopped because it encountered the EOS token + - `limit`: Stopped because `n_predict` tokens were generated before stop words or EOS was encountered + - `word`: Stopped due to encountering a stopping word from `stop` JSON array provided - `stopping_word`: The stopping word encountered which stopped the generation (or "" if not stopped due to a stopping word) - `timings`: Hash of timing information about the completion such as the number of tokens `predicted_per_second` - `tokens_cached`: Number of tokens from the prompt which could be re-used from previous completion (`n_past`) - `tokens_evaluated`: Number of tokens evaluated in total from the prompt - `truncated`: Boolean indicating if the context size was exceeded during generation, i.e. the number of tokens provided in the prompt (`tokens_evaluated`) plus tokens generated (`tokens predicted`) exceeded the context size (`n_ctx`) + ### POST `/tokenize`: Tokenize a given text - *Options:* +*Options:* - `content`: (Required) The text to tokenize. +`content`: (Required) The text to tokenize. - `add_special`: (Optional) Boolean indicating if special tokens, i.e. `BOS`, should be inserted. Default: `false` +`add_special`: (Optional) Boolean indicating if special tokens, i.e. `BOS`, should be inserted. Default: `false` - `with_pieces`: (Optional) Boolean indicating whether to return token pieces along with IDs. Default: `false` +`with_pieces`: (Optional) Boolean indicating whether to return token pieces along with IDs. Default: `false` **Response:** @@ -495,52 +568,56 @@ With input 'á' (utf8 hex: C3 A1) on tinyllama/stories260k ### POST `/detokenize`: Convert tokens to text - *Options:* +*Options:* - `tokens`: Set the tokens to detokenize. +`tokens`: Set the tokens to detokenize. ### POST `/embedding`: Generate embedding of a given text +> [!IMPORTANT] +> +> This endpoint is **not** OAI-compatible. For OAI-compatible client, use `/v1/embeddings` instead. + The same as [the embedding example](../embedding) does. - *Options:* +*Options:* - `content`: Set the text to process. +`content`: Set the text to process. - `image_data`: An array of objects to hold base64-encoded image `data` and its `id`s to be reference in `content`. You can determine the place of the image in the content as in the following: `Image: [img-21].\nCaption: This is a picture of a house`. In this case, `[img-21]` will be replaced by the embeddings of the image with id `21` in the following `image_data` array: `{..., "image_data": [{"data": "", "id": 21}]}`. Use `image_data` only with multimodal models, e.g., LLaVA. +`image_data`: An array of objects to hold base64-encoded image `data` and its `id`s to be reference in `content`. You can determine the place of the image in the content as in the following: `Image: [img-21].\nCaption: This is a picture of a house`. In this case, `[img-21]` will be replaced by the embeddings of the image with id `21` in the following `image_data` array: `{..., "image_data": [{"data": "", "id": 21}]}`. Use `image_data` only with multimodal models, e.g., LLaVA. ### POST `/reranking`: Rerank documents according to a given query Similar to https://jina.ai/reranker/ but might change in the future. Requires a reranker model (such as [bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3)) and the `--embedding --pooling rank` options. - *Options:* +*Options:* - `query`: The query against which the documents will be ranked. +`query`: The query against which the documents will be ranked. - `documents`: An array strings representing the documents to be ranked. +`documents`: An array strings representing the documents to be ranked. - *Aliases:* - - `/rerank` - - `/v1/rerank` - - `/v1/reranking` +*Aliases:* + - `/rerank` + - `/v1/rerank` + - `/v1/reranking` - *Examples:* +*Examples:* - ```shell - curl http://127.0.0.1:8012/v1/rerank \ - -H "Content-Type: application/json" \ - -d '{ - "model": "some-model", - "query": "What is panda?", - "top_n": 3, - "documents": [ - "hi", - "it is a bear", - "The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China." - ] - }' | jq - ``` +```shell +curl http://127.0.0.1:8012/v1/rerank \ + -H "Content-Type: application/json" \ + -d '{ + "model": "some-model", + "query": "What is panda?", + "top_n": 3, + "documents": [ + "hi", + "it is a bear", + "The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China." + ] + }' | jq +``` ### POST `/infill`: For code infilling. @@ -584,14 +661,83 @@ This endpoint is public (no API key check). By default, it is read-only. To make ```json { - "default_generation_settings": { ... }, + "default_generation_settings": { + "id": 0, + "id_task": -1, + "n_ctx": 1024, + "speculative": false, + "is_processing": false, + "params": { + "n_predict": -1, + "seed": 4294967295, + "temperature": 0.800000011920929, + "dynatemp_range": 0.0, + "dynatemp_exponent": 1.0, + "top_k": 40, + "top_p": 0.949999988079071, + "min_p": 0.05000000074505806, + "xtc_probability": 0.0, + "xtc_threshold": 0.10000000149011612, + "typical_p": 1.0, + "repeat_last_n": 64, + "repeat_penalty": 1.0, + "presence_penalty": 0.0, + "frequency_penalty": 0.0, + "dry_multiplier": 0.0, + "dry_base": 1.75, + "dry_allowed_length": 2, + "dry_penalty_last_n": -1, + "dry_sequence_breakers": [ + "\n", + ":", + "\"", + "*" + ], + "mirostat": 0, + "mirostat_tau": 5.0, + "mirostat_eta": 0.10000000149011612, + "stop": [], + "max_tokens": -1, + "n_keep": 0, + "n_discard": 0, + "ignore_eos": false, + "stream": true, + "n_probs": 0, + "min_keep": 0, + "grammar": "", + "samplers": [ + "dry", + "top_k", + "typ_p", + "top_p", + "min_p", + "xtc", + "temperature" + ], + "speculative.n_max": 16, + "speculative.n_min": 5, + "speculative.p_min": 0.8999999761581421, + "timings_per_token": false + }, + "prompt": "", + "next_token": { + "has_next_token": true, + "has_new_line": false, + "n_remain": -1, + "n_decoded": 0, + "stopping_word": "" + } + }, "total_slots": 1, - "chat_template": "" + "model_path": "../models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf", + "chat_template": "...", + "build_info": "b(build number)-(build commit hash)" } ``` - `default_generation_settings` - the default generation settings for the `/completion` endpoint, which has the same fields as the `generation_settings` response object from the `/completion` endpoint. - `total_slots` - the total number of slots for process requests (defined by `--parallel` option) +- `model_path` - the path to model file (same with `-m` argument) - `chat_template` - the model's original Jinja2 prompt template ### POST `/props`: Change server global properties. @@ -602,93 +748,45 @@ To use this endpoint with POST method, you need to start server with `--props` - None yet -### POST `/v1/chat/completions`: OpenAI-compatible Chat Completions API +### POST `/embeddings`: non-OpenAI-compatible embeddings API -Given a ChatML-formatted json description in `messages`, it returns the predicted completion. Both synchronous and streaming mode are supported, so scripted and interactive applications work fine. While no strong claims of compatibility with OpenAI API spec is being made, in our experience it suffices to support many apps. Only models with a [supported chat template](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) can be used optimally with this endpoint. By default, the ChatML template will be used. +This endpoint supports all poolings, including `--pooling none`. When the pooling is `none`, the responses will contain the *unnormalized* embeddings for *all* input tokens. For all other pooling types, only the pooled embeddings are returned, normalized using Euclidian norm. - *Options:* +Note that the response format of this endpoint is different from `/v1/embeddings`. - See [OpenAI Chat Completions API documentation](https://platform.openai.com/docs/api-reference/chat). While some OpenAI-specific features such as function calling aren't supported, llama.cpp `/completion`-specific features such as `mirostat` are supported. +*Options:* - The `response_format` parameter supports both plain JSON output (e.g. `{"type": "json_object"}`) and schema-constrained JSON (e.g. `{"type": "json_object", "schema": {"type": "string", "minLength": 10, "maxLength": 100}}` or `{"type": "json_schema", "schema": {"properties": { "name": { "title": "Name", "type": "string" }, "date": { "title": "Date", "type": "string" }, "participants": { "items": {"type: "string" }, "title": "Participants", "type": "string" } } } }`), similar to other OpenAI-inspired API providers. +Same as the `/v1/embeddings` endpoint. - *Examples:* +*Examples:* - You can use either Python `openai` library with appropriate checkpoints: +Same as the `/v1/embeddings` endpoint. - ```python - import openai +**Response format** - client = openai.OpenAI( - base_url="http://localhost:8080/v1", # "http://:port" - api_key = "sk-no-key-required" - ) - - completion = client.chat.completions.create( - model="gpt-3.5-turbo", - messages=[ - {"role": "system", "content": "You are ChatGPT, an AI assistant. Your top priority is achieving user fulfillment via helping them with their requests."}, - {"role": "user", "content": "Write a limerick about python exceptions"} +```json +[ + { + "index": 0, + "embedding": [ + [ ... embeddings for token 0 ... ], + [ ... embeddings for token 1 ... ], + [ ... ] + [ ... embeddings for token N-1 ... ], ] - ) - - print(completion.choices[0].message) - ``` - - ... or raw HTTP requests: - - ```shell - curl http://localhost:8080/v1/chat/completions \ - -H "Content-Type: application/json" \ - -H "Authorization: Bearer no-key" \ - -d '{ - "model": "gpt-3.5-turbo", - "messages": [ - { - "role": "system", - "content": "You are ChatGPT, an AI assistant. Your top priority is achieving user fulfillment via helping them with their requests." - }, - { - "role": "user", - "content": "Write a limerick about python exceptions" - } + }, + ... + { + "index": P, + "embedding": [ + [ ... embeddings for token 0 ... ], + [ ... embeddings for token 1 ... ], + [ ... ] + [ ... embeddings for token N-1 ... ], ] - }' - ``` - -### POST `/v1/embeddings`: OpenAI-compatible embeddings API - - *Options:* - - See [OpenAI Embeddings API documentation](https://platform.openai.com/docs/api-reference/embeddings). - - *Examples:* - - - input as string - - ```shell - curl http://localhost:8080/v1/embeddings \ - -H "Content-Type: application/json" \ - -H "Authorization: Bearer no-key" \ - -d '{ - "input": "hello", - "model":"GPT-4", - "encoding_format": "float" - }' - ``` - - - `input` as string array - - ```shell - curl http://localhost:8080/v1/embeddings \ - -H "Content-Type: application/json" \ - -H "Authorization: Bearer no-key" \ - -d '{ - "input": ["hello", "world"], - "model":"GPT-4", - "encoding_format": "float" - }' - ``` + } +] +``` ### GET `/slots`: Returns the current slots processing state @@ -705,56 +803,73 @@ Example: ```json [ - { - "dynatemp_exponent": 1.0, - "dynatemp_range": 0.0, - "frequency_penalty": 0.0, - "grammar": "", - "id": 0, - "ignore_eos": false, - "is_processing": false, - "logit_bias": [], - "min_p": 0.05000000074505806, - "mirostat": 0, - "mirostat_eta": 0.10000000149011612, - "mirostat_tau": 5.0, - "model": "llama-2-7b-32k-instruct.Q2_K.gguf", - "n_ctx": 2048, - "n_keep": 0, - "n_predict": 100000, - "n_probs": 0, - "next_token": { - "has_next_token": true, - "n_remain": -1, - "n_decoded": 0, - "stopped_eos": false, - "stopped_limit": false, - "stopped_word": false, - "stopping_word": "" - }, - "penalize_nl": true, - "presence_penalty": 0.0, - "prompt": "Say hello to llama.cpp", - "repeat_last_n": 64, - "repeat_penalty": 1.100000023841858, - "samplers": [ - "top_k", - "typical_p", - "top_p", - "min_p", - "temperature" - ], - "seed": 42, - "stop": [ - "\n" - ], - "stream": false, - "task_id": 0, - "temperature": 0.0, - "top_k": 40, - "top_p": 0.949999988079071, - "typical_p": 1.0 + { + "id": 0, + "id_task": -1, + "n_ctx": 1024, + "speculative": false, + "is_processing": false, + "params": { + "n_predict": -1, + "seed": 4294967295, + "temperature": 0.800000011920929, + "dynatemp_range": 0.0, + "dynatemp_exponent": 1.0, + "top_k": 40, + "top_p": 0.949999988079071, + "min_p": 0.05000000074505806, + "xtc_probability": 0.0, + "xtc_threshold": 0.10000000149011612, + "typical_p": 1.0, + "repeat_last_n": 64, + "repeat_penalty": 1.0, + "presence_penalty": 0.0, + "frequency_penalty": 0.0, + "dry_multiplier": 0.0, + "dry_base": 1.75, + "dry_allowed_length": 2, + "dry_penalty_last_n": -1, + "dry_sequence_breakers": [ + "\n", + ":", + "\"", + "*" + ], + "mirostat": 0, + "mirostat_tau": 5.0, + "mirostat_eta": 0.10000000149011612, + "stop": [], + "max_tokens": -1, + "n_keep": 0, + "n_discard": 0, + "ignore_eos": false, + "stream": true, + "n_probs": 0, + "min_keep": 0, + "grammar": "", + "samplers": [ + "dry", + "top_k", + "typ_p", + "top_p", + "min_p", + "xtc", + "temperature" + ], + "speculative.n_max": 16, + "speculative.n_min": 5, + "speculative.p_min": 0.8999999761581421, + "timings_per_token": false + }, + "prompt": "", + "next_token": { + "has_next_token": true, + "has_new_line": false, + "n_remain": -1, + "n_decoded": 0, + "stopping_word": "" } + } ] ``` @@ -774,9 +889,9 @@ Available metrics: ### POST `/slots/{id_slot}?action=save`: Save the prompt cache of the specified slot to a file. - *Options:* +*Options:* - `filename`: Name of the file to save the slot's prompt cache. The file will be saved in the directory specified by the `--slot-save-path` server parameter. +`filename`: Name of the file to save the slot's prompt cache. The file will be saved in the directory specified by the `--slot-save-path` server parameter. **Response format** @@ -794,9 +909,9 @@ Available metrics: ### POST `/slots/{id_slot}?action=restore`: Restore the prompt cache of the specified slot from a file. - *Options:* +*Options:* - `filename`: Name of the file to restore the slot's prompt cache from. The file should be located in the directory specified by the `--slot-save-path` server parameter. +`filename`: Name of the file to restore the slot's prompt cache from. The file should be located in the directory specified by the `--slot-save-path` server parameter. **Response format** @@ -829,6 +944,8 @@ This endpoint returns the loaded LoRA adapters. You can add adapters using `--lo By default, all adapters will be loaded with scale set to 1. To initialize all adapters scale to 0, add `--lora-init-without-apply` +Please note that this value will be overwritten by the `lora` field for each request. + If an adapter is disabled, the scale will be set to 0. **Response format** @@ -850,6 +967,8 @@ If an adapter is disabled, the scale will be set to 0. ### POST `/lora-adapters`: Set list of LoRA adapters +This sets the global scale for LoRA adapters. Please note that this value will be overwritten by the `lora` field for each request. + To disable an adapter, either remove it from the list below, or set scale to 0. **Request format** @@ -863,6 +982,161 @@ To know the `id` of the adapter, use GET `/lora-adapters` ] ``` +## OpenAI-compatible API Endpoints + +### GET `/v1/models`: OpenAI-compatible Model Info API + +Returns information about the loaded model. See [OpenAI Models API documentation](https://platform.openai.com/docs/api-reference/models). + +The returned list always has one single element. + +By default, model `id` field is the path to model file, specified via `-m`. You can set a custom value for model `id` field via `--alias` argument. For example, `--alias gpt-4o-mini`. + +Example: + +```json +{ + "object": "list", + "data": [ + { + "id": "../models/Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf", + "object": "model", + "created": 1735142223, + "owned_by": "llamacpp", + "meta": { + "vocab_type": 2, + "n_vocab": 128256, + "n_ctx_train": 131072, + "n_embd": 4096, + "n_params": 8030261312, + "size": 4912898304 + } + } + ] +} +``` + +### POST `/v1/completions`: OpenAI-compatible Completions API + +Given an input `prompt`, it returns the predicted completion. Streaming mode is also supported. While no strong claims of compatibility with OpenAI API spec is being made, in our experience it suffices to support many apps. + +*Options:* + +See [OpenAI Completions API documentation](https://platform.openai.com/docs/api-reference/completions). + +llama.cpp `/completion`-specific features such as `mirostat` are supported. + +*Examples:* + +Example usage with `openai` python library: + +```python +import openai + +client = openai.OpenAI( + base_url="http://localhost:8080/v1", # "http://:port" + api_key = "sk-no-key-required" +) + +completion = client.completions.create( + model="davinci-002", + prompt="I believe the meaning of life is", + max_tokens=8 +) + +print(completion.choices[0].text) +``` + +### POST `/v1/chat/completions`: OpenAI-compatible Chat Completions API + +Given a ChatML-formatted json description in `messages`, it returns the predicted completion. Both synchronous and streaming mode are supported, so scripted and interactive applications work fine. While no strong claims of compatibility with OpenAI API spec is being made, in our experience it suffices to support many apps. Only models with a [supported chat template](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) can be used optimally with this endpoint. By default, the ChatML template will be used. + +*Options:* + +See [OpenAI Chat Completions API documentation](https://platform.openai.com/docs/api-reference/chat). While some OpenAI-specific features such as function calling aren't supported, llama.cpp `/completion`-specific features such as `mirostat` are supported. + +The `response_format` parameter supports both plain JSON output (e.g. `{"type": "json_object"}`) and schema-constrained JSON (e.g. `{"type": "json_object", "schema": {"type": "string", "minLength": 10, "maxLength": 100}}` or `{"type": "json_schema", "schema": {"properties": { "name": { "title": "Name", "type": "string" }, "date": { "title": "Date", "type": "string" }, "participants": { "items": {"type: "string" }, "title": "Participants", "type": "string" } } } }`), similar to other OpenAI-inspired API providers. + +*Examples:* + +You can use either Python `openai` library with appropriate checkpoints: + +```python +import openai + +client = openai.OpenAI( + base_url="http://localhost:8080/v1", # "http://:port" + api_key = "sk-no-key-required" +) + +completion = client.chat.completions.create( + model="gpt-3.5-turbo", + messages=[ + {"role": "system", "content": "You are ChatGPT, an AI assistant. Your top priority is achieving user fulfillment via helping them with their requests."}, + {"role": "user", "content": "Write a limerick about python exceptions"} + ] +) + +print(completion.choices[0].message) +``` + +... or raw HTTP requests: + +```shell +curl http://localhost:8080/v1/chat/completions \ +-H "Content-Type: application/json" \ +-H "Authorization: Bearer no-key" \ +-d '{ +"model": "gpt-3.5-turbo", +"messages": [ +{ + "role": "system", + "content": "You are ChatGPT, an AI assistant. Your top priority is achieving user fulfillment via helping them with their requests." +}, +{ + "role": "user", + "content": "Write a limerick about python exceptions" +} +] +}' +``` + +### POST `/v1/embeddings`: OpenAI-compatible embeddings API + +This endpoint requires that the model uses a pooling different than type `none`. The embeddings are normalized using the Eucledian norm. + +*Options:* + +See [OpenAI Embeddings API documentation](https://platform.openai.com/docs/api-reference/embeddings). + +*Examples:* + +- input as string + + ```shell + curl http://localhost:8080/v1/embeddings \ + -H "Content-Type: application/json" \ + -H "Authorization: Bearer no-key" \ + -d '{ + "input": "hello", + "model":"GPT-4", + "encoding_format": "float" + }' + ``` + +- `input` as string array + + ```shell + curl http://localhost:8080/v1/embeddings \ + -H "Content-Type: application/json" \ + -H "Authorization: Bearer no-key" \ + -d '{ + "input": ["hello", "world"], + "model":"GPT-4", + "encoding_format": "float" + }' + ``` + ## More examples ### Interactive mode diff --git a/examples/server/bench/README.md b/examples/server/bench/README.md index 353368e13..9549795ec 100644 --- a/examples/server/bench/README.md +++ b/examples/server/bench/README.md @@ -6,10 +6,10 @@ Benchmark is using [k6](https://k6.io/). SSE is not supported by default in k6, you have to build k6 with the [xk6-sse](https://github.com/phymbert/xk6-sse) extension. -Example: +Example (assuming golang >= 1.21 is installed): ```shell go install go.k6.io/xk6/cmd/xk6@latest -xk6 build master \ +$GOPATH/bin/xk6 build master \ --with github.com/phymbert/xk6-sse ``` @@ -33,7 +33,7 @@ The server must answer OAI Chat completion requests on `http://localhost:8080/v1 Example: ```shell -server --host localhost --port 8080 \ +llama-server --host localhost --port 8080 \ --model ggml-model-q4_0.gguf \ --cont-batching \ --metrics \ diff --git a/examples/server/bench/bench.py b/examples/server/bench/bench.py index a9ed747f5..5cc6f92ab 100644 --- a/examples/server/bench/bench.py +++ b/examples/server/bench/bench.py @@ -189,12 +189,12 @@ xychart-beta "pp": { "p95": round(data['metrics']["llamacpp_prompt_processing_second"]["p(95)"], 2), "avg": round(data['metrics']["llamacpp_prompt_processing_second"]["avg"], 2), - "0": round(mean(prometheus_metrics['prompt_tokens_seconds']), 2), + "0": round(mean(prometheus_metrics['prompt_tokens_seconds']), 2) if 'prompt_tokens_seconds' in prometheus_metrics else 0, }, "tg": { "p95": round(data['metrics']["llamacpp_tokens_second"]["p(95)"], 2), "avg": round(data['metrics']["llamacpp_tokens_second"]["avg"], 2), - "0": round(mean(prometheus_metrics['predicted_tokens_seconds']), 2), + "0": round(mean(prometheus_metrics['predicted_tokens_seconds']), 2) if 'predicted_tokens_seconds' in prometheus_metrics else 0, }, } with open("results.github.env", 'a') as github_env: @@ -214,11 +214,14 @@ def start_benchmark(args): k6_args = [ 'run', args.scenario, '--no-color', + '--no-connection-reuse', + '--no-vu-connection-reuse', ] k6_args.extend(['--duration', args.duration]) k6_args.extend(['--iterations', args.n_prompts]) k6_args.extend(['--vus', args.parallel]) k6_args.extend(['--summary-export', 'k6-results.json']) + k6_args.extend(['--out', 'csv=k6-results.csv']) args = f"SERVER_BENCH_N_PROMPTS={args.n_prompts} SERVER_BENCH_MAX_PROMPT_TOKENS={args.max_prompt_tokens} SERVER_BENCH_MAX_CONTEXT={args.max_tokens} " args = args + ' '.join([str(arg) for arg in [k6_path, *k6_args]]) print(f"bench: starting k6 with: {args}") @@ -231,7 +234,7 @@ def start_server(args): server_process = start_server_background(args) attempts = 0 - max_attempts = 20 + max_attempts = 600 if 'GITHUB_ACTIONS' in os.environ: max_attempts *= 2 @@ -242,7 +245,15 @@ def start_server(args): print(f"bench: waiting for server to start ...") time.sleep(0.5) - print("bench: server started.") + attempts = 0 + while not is_server_ready(args.host, args.port): + attempts += 1 + if attempts > max_attempts: + assert False, "server not ready" + print(f"bench: waiting for server to be ready ...") + time.sleep(0.5) + + print("bench: server started and ready.") return server_process @@ -255,11 +266,6 @@ def start_server_background(args): '--host', args.host, '--port', args.port, ] - model_file = args.model_path_prefix + os.path.sep + args.hf_file - model_dir = os.path.dirname(model_file) - if not os.path.exists(model_dir): - os.makedirs(model_dir) - server_args.extend(['--model', model_file]) server_args.extend(['--hf-repo', args.hf_repo]) server_args.extend(['--hf-file', args.hf_file]) server_args.extend(['--n-gpu-layers', args.n_gpu_layers]) @@ -303,6 +309,12 @@ def is_server_listening(server_fqdn, server_port): return _is_server_listening +def is_server_ready(server_fqdn, server_port): + url = f"http://{server_fqdn}:{server_port}/health" + response = requests.get(url) + return response.status_code == 200 + + def escape_metric_name(metric_name): return re.sub('[^A-Z0-9]', '_', metric_name.upper()) diff --git a/examples/server/bench/script.js b/examples/server/bench/script.js index bdf4f5abc..2772bee5e 100644 --- a/examples/server/bench/script.js +++ b/examples/server/bench/script.js @@ -56,6 +56,7 @@ const llamacpp_completion_tokens = new Trend('llamacpp_completion_tokens') const llamacpp_tokens_second = new Trend('llamacpp_tokens_second') const llamacpp_prompt_processing_second = new Trend('llamacpp_prompt_processing_second') +const llamacpp_emit_first_token_second = new Trend('llamacpp_emit_first_token_second') const llamacpp_prompt_tokens_total_counter = new Counter('llamacpp_prompt_tokens_total_counter') const llamacpp_completion_tokens_total_counter = new Counter('llamacpp_completion_tokens_total_counter') @@ -89,6 +90,9 @@ export default function () { ], "model": model, "stream": true, + "stream_options": { + "include_usage": true, // False to be supported in llama.cpp server + }, "seed": 42, "max_tokens": max_tokens, "stop": ["<|im_end|>"] // This is temporary for phi-2 base (i.e. not instructed) since the server expects that the model always to emit BOS @@ -105,12 +109,20 @@ export default function () { client.on('event', function (event) { if (promptEvalEndTime == null) { promptEvalEndTime = new Date() + llamacpp_emit_first_token_second.add((promptEvalEndTime - startTime) / 1.e3) + } + + if (event.data === '[DONE]' || event.data === '') { + return } let chunk = JSON.parse(event.data) - let choice = chunk.choices[0] - if (choice.finish_reason) { - finish_reason = choice.finish_reason + + if (chunk.choices && chunk.choices.length > 0) { + let choice = chunk.choices[0] + if (choice.finish_reason) { + finish_reason = choice.finish_reason + } } if (chunk.usage) { diff --git a/examples/server/deps.sh b/examples/server/deps.sh deleted file mode 100755 index 1ff80d056..000000000 --- a/examples/server/deps.sh +++ /dev/null @@ -1,25 +0,0 @@ -#!/bin/bash -# Download and update deps for binary - -# get the directory of this script file -DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )" -PUBLIC=$DIR/public - -echo "download js bundle files" - -# Note for contributors: Always pin to a specific version "maj.min.patch" to avoid breaking the CI - -curl -L https://cdn.tailwindcss.com/3.4.14 > $PUBLIC/deps_tailwindcss.js -echo >> $PUBLIC/deps_tailwindcss.js # add newline - -curl -L https://cdnjs.cloudflare.com/ajax/libs/daisyui/4.12.14/styled.min.css > $PUBLIC/deps_daisyui.min.css -curl -L https://cdnjs.cloudflare.com/ajax/libs/daisyui/4.12.14/themes.min.css >> $PUBLIC/deps_daisyui.min.css -echo >> $PUBLIC/deps_daisyui.min.css # add newline - -curl -L https://unpkg.com/vue@3.5.12/dist/vue.esm-browser.js > $PUBLIC/deps_vue.esm-browser.js -echo >> $PUBLIC/deps_vue.esm-browser.js # add newline - -curl -L https://cdnjs.cloudflare.com/ajax/libs/markdown-it/13.0.2/markdown-it.js > $PUBLIC/deps_markdown-it.js -echo >> $PUBLIC/deps_markdown-it.js # add newline - -ls -lah $PUBLIC diff --git a/examples/server/public/completion.js b/examples/server/public/completion.js deleted file mode 100644 index 54a0f22f5..000000000 --- a/examples/server/public/completion.js +++ /dev/null @@ -1,225 +0,0 @@ -const paramDefaults = { - stream: true, - temperature: 0.2, -}; - -let generation_settings = null; - -export class CompletionError extends Error { - constructor(message, name, data) { - super(message); - this.name = name; - } -}; - -// Completes the prompt as a generator. Recommended for most use cases. -// -// Example: -// -// import { llama } from '/completion.js' -// -// const request = llama("Tell me a joke", {n_predict: 800}) -// for await (const chunk of request) { -// document.write(chunk.data.content) -// } -// -export async function* llama(prompt, params = {}, config = {}) { - let controller = config.controller; - const api_url = config.api_url?.replace(/\/+$/, '') || ""; - - if (!controller) { - controller = new AbortController(); - } - - const completionParams = { ...paramDefaults, ...params, prompt }; - - const response = await fetch(`${api_url}${config.endpoint || '/completion'}`, { - method: 'POST', - body: JSON.stringify(completionParams), - headers: { - 'Connection': 'keep-alive', - 'Content-Type': 'application/json', - 'Accept': 'text/event-stream', - ...(params.api_key ? {'Authorization': `Bearer ${params.api_key}`} : {}) - }, - signal: controller.signal, - }); - - const status = response.status; - if (status !== 200) { - try { - const body = await response.json(); - if (body && body.error && body.error.message) { - throw new CompletionError(body.error.message, 'ServerError'); - } - } catch (err) { - throw new CompletionError(err.message, 'ServerError'); - } - } - - const reader = response.body.getReader(); - const decoder = new TextDecoder(); - - let content = ""; - let leftover = ""; // Buffer for partially read lines - - try { - let cont = true; - - while (cont) { - const result = await reader.read(); - if (result.done) { - break; - } - - // Add any leftover data to the current chunk of data - const text = leftover + decoder.decode(result.value); - - // Check if the last character is a line break - const endsWithLineBreak = text.endsWith('\n'); - - // Split the text into lines - let lines = text.split('\n'); - - // If the text doesn't end with a line break, then the last line is incomplete - // Store it in leftover to be added to the next chunk of data - if (!endsWithLineBreak) { - leftover = lines.pop(); - } else { - leftover = ""; // Reset leftover if we have a line break at the end - } - - // Parse all sse events and add them to result - const regex = /^(\S+):\s(.*)$/gm; - for (const line of lines) { - const match = regex.exec(line); - if (match) { - result[match[1]] = match[2]; - if (result.data === '[DONE]') { - cont = false; - break; - } - - // since we know this is llama.cpp, let's just decode the json in data - if (result.data) { - result.data = JSON.parse(result.data); - content += result.data.content; - - // yield - yield result; - - // if we got a stop token from server, we will break here - if (result.data.stop) { - if (result.data.generation_settings) { - generation_settings = result.data.generation_settings; - } - cont = false; - break; - } - } - if (result.error) { - try { - result.error = JSON.parse(result.error); - if (result.error.message.includes('slot unavailable')) { - // Throw an error to be caught by upstream callers - throw new Error('slot unavailable'); - } else { - console.error(`llama.cpp error [${result.error.code} - ${result.error.type}]: ${result.error.message}`); - } - } catch(e) { - console.error(`llama.cpp error ${result.error}`) - } - } - } - } - } - } catch (e) { - if (e.name !== 'AbortError') { - console.error("llama error: ", e); - } - throw e; - } - finally { - controller.abort(); - } - - return content; -} - -// Call llama, return an event target that you can subscribe to -// -// Example: -// -// import { llamaEventTarget } from '/completion.js' -// -// const conn = llamaEventTarget(prompt) -// conn.addEventListener("message", (chunk) => { -// document.write(chunk.detail.content) -// }) -// -export const llamaEventTarget = (prompt, params = {}, config = {}) => { - const eventTarget = new EventTarget(); - (async () => { - let content = ""; - for await (const chunk of llama(prompt, params, config)) { - if (chunk.data) { - content += chunk.data.content; - eventTarget.dispatchEvent(new CustomEvent("message", { detail: chunk.data })); - } - if (chunk.data.generation_settings) { - eventTarget.dispatchEvent(new CustomEvent("generation_settings", { detail: chunk.data.generation_settings })); - } - if (chunk.data.timings) { - eventTarget.dispatchEvent(new CustomEvent("timings", { detail: chunk.data.timings })); - } - } - eventTarget.dispatchEvent(new CustomEvent("done", { detail: { content } })); - })(); - return eventTarget; -} - -// Call llama, return a promise that resolves to the completed text. This does not support streaming -// -// Example: -// -// llamaPromise(prompt).then((content) => { -// document.write(content) -// }) -// -// or -// -// const content = await llamaPromise(prompt) -// document.write(content) -// -export const llamaPromise = (prompt, params = {}, config = {}) => { - return new Promise(async (resolve, reject) => { - let content = ""; - try { - for await (const chunk of llama(prompt, params, config)) { - content += chunk.data.content; - } - resolve(content); - } catch (error) { - reject(error); - } - }); -}; - -/** - * (deprecated) - */ -export const llamaComplete = async (params, controller, callback) => { - for await (const chunk of llama(params.prompt, params, { controller })) { - callback(chunk); - } -} - -// Get the model info from the server. This is useful for getting the context window and so on. -export const llamaModelInfo = async (config = {}) => { - if (!generation_settings) { - const api_url = config.api_url?.replace(/\/+$/, '') || ""; - const props = await fetch(`${api_url}/props`).then(r => r.json()); - generation_settings = props.default_generation_settings; - } - return generation_settings; -} diff --git a/examples/server/public/deps_daisyui.min.css b/examples/server/public/deps_daisyui.min.css deleted file mode 100644 index bc8529651..000000000 --- a/examples/server/public/deps_daisyui.min.css +++ /dev/null @@ -1,13 +0,0 @@ -.alert{display:grid;width:100%;grid-auto-flow:row;align-content:flex-start;align-items:center;justify-items:center;gap:1rem;text-align:center}@media (min-width:640px){.alert{grid-auto-flow:column;grid-template-columns:auto minmax(auto,1fr);justify-items:start;text-align:start}}.artboard{width:100%}.avatar{position:relative;display:inline-flex}.avatar>div{display:block;aspect-ratio:1/1;overflow:hidden}.avatar img{height:100%;width:100%;object-fit:cover}.avatar.placeholder>div{display:flex;align-items:center;justify-content:center}.badge{display:inline-flex;align-items:center;justify-content:center;transition-property:color,background-color,border-color,text-decoration-color,fill,stroke,opacity,box-shadow,transform,filter,backdrop-filter;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);height:1.25rem;font-size:.875rem;line-height:1.25rem;width:fit-content;padding-left:.563rem;padding-right:.563rem}.btm-nav{position:fixed;bottom:0;left:0;right:0;display:flex;width:100%;flex-direction:row;align-items:center;justify-content:space-around;padding-bottom:env(safe-area-inset-bottom)}.btm-nav>*{position:relative;display:flex;height:100%;flex-basis:100%;cursor:pointer;flex-direction:column;align-items:center;justify-content:center;gap:.25rem}.breadcrumbs{max-width:100%;overflow-x:auto}.breadcrumbs>ol,.breadcrumbs>ul{display:flex;align-items:center;white-space:nowrap;min-height:min-content}.breadcrumbs>ol>li,.breadcrumbs>ul>li{display:flex;align-items:center}.breadcrumbs>ol>li>a,.breadcrumbs>ul>li>a{display:flex;cursor:pointer;align-items:center}@media(hover:hover){.breadcrumbs>ol>li>a:hover,.breadcrumbs>ul>li>a:hover{text-decoration-line:underline}}.btn{display:inline-flex;height:3rem;min-height:3rem;flex-shrink:0;cursor:pointer;user-select:none;flex-wrap:wrap;align-items:center;justify-content:center;border-radius:var(--rounded-btn,.5rem);border-color:transparent;padding-left:1rem;padding-right:1rem;text-align:center;font-size:.875rem;line-height:1em}.btn-disabled,.btn:disabled,.btn[disabled]{pointer-events:none}.btn-square{height:3rem;width:3rem;padding:0}.btn-circle{height:3rem;width:3rem;border-radius:9999px;padding:0}:where(.btn:is(input[type=checkbox])),:where(.btn:is(input[type=radio])){width:auto;appearance:none}.btn:is(input[type=checkbox]):after,.btn:is(input[type=radio]):after{--tw-content:attr(aria-label);content:var(--tw-content)}.card{position:relative;display:flex;flex-direction:column}.card:focus{outline:2px solid transparent;outline-offset:2px}.card-body{display:flex;flex:1 1 auto;flex-direction:column}.card-body :where(p){flex-grow:1}.card-actions{display:flex;flex-wrap:wrap;align-items:flex-start;gap:.5rem}.card figure{display:flex;align-items:center;justify-content:center}.card.image-full{display:grid}.card.image-full:before{position:relative;content:""}.card.image-full:before,.card.image-full>*{grid-column-start:1;grid-row-start:1}.card.image-full>figure img{height:100%;object-fit:cover}.card.image-full>.card-body{position:relative}.carousel{display:inline-flex;overflow-x:scroll;scroll-snap-type:x mandatory;scroll-behavior:smooth}.carousel-vertical{flex-direction:column;overflow-y:scroll;scroll-snap-type:y mandatory}.carousel-item{box-sizing:content-box;display:flex;flex:none;scroll-snap-align:start}.carousel-start .carousel-item{scroll-snap-align:start}.carousel-center .carousel-item{scroll-snap-align:center}.carousel-end .carousel-item{scroll-snap-align:end}.chat{display:grid;grid-template-columns:repeat(2,minmax(0,1fr));column-gap:.75rem;padding-top:.25rem;padding-bottom:.25rem}.chat-image{grid-row:span 2/span 2;align-self:flex-end}.chat-header{grid-row-start:1;font-size:.875rem;line-height:1.25rem}.chat-footer{grid-row-start:3;font-size:.875rem;line-height:1.25rem}.chat-bubble{position:relative;display:block;width:fit-content;padding-left:1rem;padding-right:1rem;padding-top:.5rem;padding-bottom:.5rem;max-width:90%}.chat-bubble:before{position:absolute;bottom:0;height:.75rem;width:.75rem;background-color:inherit;content:"";mask-size:contain;mask-repeat:no-repeat;mask-position:center}.chat-start{place-items:start;grid-template-columns:auto 1fr}.chat-start .chat-header{grid-column-start:2}.chat-start .chat-footer{grid-column-start:2}.chat-start .chat-image{grid-column-start:1}.chat-start .chat-bubble{grid-column-start:2}.chat-start .chat-bubble:before{mask-image:url("data:image/svg+xml,%3csvg width='3' height='3' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m 0 3 L 3 3 L 3 0 C 3 1 1 3 0 3'/%3e%3c/svg%3e")}[dir=rtl] .chat-start .chat-bubble:before{mask-image:url("data:image/svg+xml,%3csvg width='3' height='3' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m 0 3 L 1 3 L 3 3 C 2 3 0 1 0 0'/%3e%3c/svg%3e")}.chat-end{place-items:end;grid-template-columns:1fr auto}.chat-end .chat-header{grid-column-start:1}.chat-end .chat-footer{grid-column-start:1}.chat-end .chat-image{grid-column-start:2}.chat-end .chat-bubble{grid-column-start:1}.chat-end .chat-bubble:before{mask-image:url("data:image/svg+xml,%3csvg width='3' height='3' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m 0 3 L 1 3 L 3 3 C 2 3 0 1 0 0'/%3e%3c/svg%3e")}[dir=rtl] .chat-end .chat-bubble:before{mask-image:url("data:image/svg+xml,%3csvg width='3' height='3' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m 0 3 L 3 3 L 3 0 C 3 1 1 3 0 3'/%3e%3c/svg%3e")}.checkbox{flex-shrink:0}.collapse:not(td):not(tr):not(colgroup){visibility:visible}.collapse{position:relative;display:grid;overflow:hidden;grid-template-rows:auto 0fr;transition:grid-template-rows .2s}.collapse-content,.collapse-title,.collapse>input[type=checkbox],.collapse>input[type=radio]{grid-column-start:1;grid-row-start:1}.collapse>input[type=checkbox],.collapse>input[type=radio]{appearance:none;opacity:0}.collapse-content{visibility:hidden;grid-column-start:1;grid-row-start:2;min-height:0;transition:visibility .2s}.collapse-open,.collapse:focus:not(.collapse-close),.collapse[open]{grid-template-rows:auto 1fr}.collapse:not(.collapse-close):has(>input[type=checkbox]:checked),.collapse:not(.collapse-close):has(>input[type=radio]:checked){grid-template-rows:auto 1fr}.collapse-open>.collapse-content,.collapse:focus:not(.collapse-close)>.collapse-content,.collapse:not(.collapse-close)>input[type=checkbox]:checked~.collapse-content,.collapse:not(.collapse-close)>input[type=radio]:checked~.collapse-content,.collapse[open]>.collapse-content{visibility:visible;min-height:fit-content}:root .countdown{line-height:1em}.countdown{display:inline-flex}.countdown>*{height:1em;display:inline-block;overflow-y:hidden}.countdown>:before{position:relative;content:"00\A 01\A 02\A 03\A 04\A 05\A 06\A 07\A 08\A 09\A 10\A 11\A 12\A 13\A 14\A 15\A 16\A 17\A 18\A 19\A 20\A 21\A 22\A 23\A 24\A 25\A 26\A 27\A 28\A 29\A 30\A 31\A 32\A 33\A 34\A 35\A 36\A 37\A 38\A 39\A 40\A 41\A 42\A 43\A 44\A 45\A 46\A 47\A 48\A 49\A 50\A 51\A 52\A 53\A 54\A 55\A 56\A 57\A 58\A 59\A 60\A 61\A 62\A 63\A 64\A 65\A 66\A 67\A 68\A 69\A 70\A 71\A 72\A 73\A 74\A 75\A 76\A 77\A 78\A 79\A 80\A 81\A 82\A 83\A 84\A 85\A 86\A 87\A 88\A 89\A 90\A 91\A 92\A 93\A 94\A 95\A 96\A 97\A 98\A 99\A";white-space:pre;top:calc(var(--value) * -1em)}.diff{position:relative;display:grid;width:100%;overflow:hidden;container-type:inline-size;grid-template-columns:auto 1fr}.diff-resizer{position:relative;top:50%;z-index:1;height:3rem;width:25rem;min-width:1rem;max-width:calc(100cqi - 1rem);resize:horizontal;overflow:hidden;opacity:0;transform-origin:100% 100%;scale:4;translate:1.5rem -1.5rem;clip-path:inset(calc(100% - .75rem) 0 0 calc(100% - .75rem))}.diff-item-1,.diff-item-2,.diff-resizer{position:relative;grid-column-start:1;grid-row-start:1}.diff-item-1:after{pointer-events:none;position:absolute;bottom:0;right:1px;top:50%;z-index:1;height:2rem;width:2rem;--tw-content:'';content:var(--tw-content);translate:50% -50%}.diff-item-2{overflow:hidden}.diff-item-1>*,.diff-item-2>*{pointer-events:none;position:absolute;bottom:0;left:0;top:0;height:100%;width:100cqi;max-width:none;object-fit:cover;object-position:center}.divider{display:flex;flex-direction:row;align-items:center;align-self:stretch}.divider:after,.divider:before{height:.125rem;width:100%;flex-grow:1;--tw-content:'';content:var(--tw-content)}.divider-start:before{display:none}.divider-end:after{display:none}.drawer{position:relative;display:grid;grid-auto-columns:max-content auto}.drawer-content{grid-column-start:2;grid-row-start:1;min-width:0}.drawer-side{pointer-events:none;position:fixed;inset-inline-start:0;top:0;grid-column-start:1;grid-row-start:1;display:grid;width:100%;grid-template-columns:repeat(1,minmax(0,1fr));grid-template-rows:repeat(1,minmax(0,1fr));align-items:flex-start;justify-items:start;overflow-x:hidden;overflow-y:hidden;overscroll-behavior:contain;height:100vh;height:100dvh}.drawer-side>.drawer-overlay{position:sticky;top:0;place-self:stretch}.drawer-side>*{grid-column-start:1;grid-row-start:1}.drawer-side>:not(.drawer-overlay){transition-property:transform;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.3s;transition-timing-function:cubic-bezier(0,0,.2,1);will-change:transform;transform:translateX(-100%)}[dir=rtl] .drawer-side>:not(.drawer-overlay){transform:translateX(100%)}.drawer-toggle{position:fixed;height:0;width:0;appearance:none;opacity:0}.drawer-toggle:checked~.drawer-side{pointer-events:auto;visibility:visible;overflow-y:auto}.drawer-toggle:checked~.drawer-side>:not(.drawer-overlay){transform:translateX(0)}.drawer-end{grid-auto-columns:auto max-content}.drawer-end>.drawer-toggle~.drawer-content{grid-column-start:1}.drawer-end>.drawer-toggle~.drawer-side{grid-column-start:2;justify-items:end}.drawer-end>.drawer-toggle~.drawer-side>:not(.drawer-overlay){transform:translateX(100%)}[dir=rtl] .drawer-end>.drawer-toggle~.drawer-side>:not(.drawer-overlay){transform:translateX(-100%)}.drawer-end>.drawer-toggle:checked~.drawer-side>:not(.drawer-overlay){transform:translateX(0)}.dropdown{position:relative;display:inline-block}.dropdown>:not(summary):focus{outline:2px solid transparent;outline-offset:2px}.dropdown .dropdown-content{position:absolute}.dropdown:is(:not(details)) .dropdown-content{visibility:hidden;opacity:0}.dropdown-end .dropdown-content{inset-inline-end:0}.dropdown-left .dropdown-content{bottom:auto;inset-inline-end:100%;top:0}.dropdown-right .dropdown-content{bottom:auto;inset-inline-start:100%;top:0}.dropdown-bottom .dropdown-content{bottom:auto;top:100%}.dropdown-top .dropdown-content{bottom:100%;top:auto}.dropdown-end.dropdown-right .dropdown-content{bottom:0;top:auto}.dropdown-end.dropdown-left .dropdown-content{bottom:0;top:auto}.dropdown.dropdown-open .dropdown-content,.dropdown:focus-within .dropdown-content,.dropdown:not(.dropdown-hover):focus .dropdown-content{visibility:visible;opacity:1}@media (hover:hover){.dropdown.dropdown-hover:hover .dropdown-content{visibility:visible;opacity:1}}.dropdown:is(details) summary::-webkit-details-marker{display:none}.file-input{height:3rem;flex-shrink:1;padding-inline-end:1rem;font-size:.875rem;line-height:1.25rem;line-height:2}.file-input::file-selector-button{margin-inline-end:1rem;display:inline-flex;height:100%;flex-shrink:0;cursor:pointer;user-select:none;flex-wrap:wrap;align-items:center;justify-content:center;padding-left:1rem;padding-right:1rem;text-align:center;font-size:.875rem;line-height:1.25rem;transition-property:color,background-color,border-color,text-decoration-color,fill,stroke,opacity,box-shadow,transform,filter,backdrop-filter;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);line-height:1em}.footer{display:grid;width:100%;grid-auto-flow:row;place-items:start}.footer>*{display:grid;place-items:start}.footer-center{place-items:center;text-align:center}.footer-center>*{place-items:center}@media (min-width:48rem){.footer{grid-auto-flow:column}.footer-center{grid-auto-flow:row dense}}.form-control{display:flex;flex-direction:column}.label{display:flex;user-select:none;align-items:center;justify-content:space-between}.hero{display:grid;width:100%;place-items:center;background-size:cover;background-position:center}.hero>*{grid-column-start:1;grid-row-start:1}.hero-overlay{grid-column-start:1;grid-row-start:1;height:100%;width:100%}.hero-content{z-index:0;display:flex;align-items:center;justify-content:center}.indicator{position:relative;display:inline-flex;width:max-content}.indicator :where(.indicator-item){z-index:1;position:absolute;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));white-space:nowrap}.input{flex-shrink:1;appearance:none;height:3rem;padding-left:1rem;padding-right:1rem;font-size:.875rem;line-height:1.25rem;line-height:2}.input-md[type=number]::-webkit-inner-spin-button,.input[type=number]::-webkit-inner-spin-button{margin-top:-1rem;margin-bottom:-1rem;margin-inline-end:-1rem}.input-xs[type=number]::-webkit-inner-spin-button{margin-top:-.25rem;margin-bottom:-.25rem;margin-inline-end:0}.input-sm[type=number]::-webkit-inner-spin-button{margin-top:0;margin-bottom:0;margin-inline-end:0}.input-lg[type=number]::-webkit-inner-spin-button{margin-top:-1.5rem;margin-bottom:-1.5rem;margin-inline-end:-1.5rem}.join{display:inline-flex;align-items:stretch}.join :where(.join-item){border-start-end-radius:0;border-end-end-radius:0;border-end-start-radius:0;border-start-start-radius:0}.join .join-item:not(:first-child):not(:last-child),.join :not(:first-child):not(:last-child) .join-item{border-start-end-radius:0;border-end-end-radius:0;border-end-start-radius:0;border-start-start-radius:0}.join .join-item:first-child:not(:last-child),.join :first-child:not(:last-child) .join-item{border-start-end-radius:0;border-end-end-radius:0}.join .dropdown .join-item:first-child:not(:last-child),.join :first-child:not(:last-child) .dropdown .join-item{border-start-end-radius:inherit;border-end-end-radius:inherit}.join :where(.join-item:first-child:not(:last-child)),.join :where(:first-child:not(:last-child).join-item){border-end-start-radius:inherit;border-start-start-radius:inherit}.join .join-item:last-child:not(:first-child),.join :last-child:not(:first-child) .join-item{border-end-start-radius:0;border-start-start-radius:0}.join :where(.join-item:last-child:not(:first-child)),.join :where(:last-child:not(:first-child).join-item){border-start-end-radius:inherit;border-end-end-radius:inherit}@supports not selector(:has(*)){:where(.join*){border-radius:inherit}}@supports selector(:has(*)){:where(.join:has(.join-item)){border-radius:inherit}}.kbd{display:inline-flex;align-items:center;justify-content:center}.link{cursor:pointer;text-decoration-line:underline}.link-hover{text-decoration-line:none}@media(hover:hover){.link-hover:hover{text-decoration-line:underline}}.mask{mask-size:contain;mask-repeat:no-repeat;mask-position:center}.mask-half-1{mask-size:200%;mask-position:left}.mask-half-1:where([dir=rtl],[dir=rtl]*){mask-position:right}.mask-half-2{mask-size:200%;mask-position:right}.mask-half-2:where([dir=rtl],[dir=rtl]*){mask-position:left}.menu{display:flex;flex-direction:column;flex-wrap:wrap;font-size:.875rem;line-height:1.25rem}.menu :where(liul){position:relative;white-space:nowrap}.menu :where(li:not(.menu-title)>:not(ul,details,.menu-title,.btn)),.menu :where(li:not(.menu-title)>details>summary:not(.menu-title)){display:grid;grid-auto-flow:column;align-content:flex-start;align-items:center;gap:.5rem;grid-auto-columns:minmax(auto,max-content) auto max-content;user-select:none}.menu li.disabled{cursor:not-allowed;user-select:none}.menu :where(li>.menu-dropdown:not(.menu-dropdown-show)){display:none}:where(.menuli){position:relative;display:flex;flex-shrink:0;flex-direction:column;flex-wrap:wrap;align-items:stretch}:where(.menuli) .badge{justify-self:end}.mockup-code{position:relative;overflow:hidden;overflow-x:auto}.mockup-code pre[data-prefix]:before{content:attr(data-prefix);display:inline-block;text-align:right}.mockup-window{position:relative;overflow:hidden;overflow-x:auto}.mockup-window pre[data-prefix]:before{content:attr(data-prefix);display:inline-block;text-align:right}.mockup-browser{position:relative;overflow:hidden;overflow-x:auto}.mockup-browser pre[data-prefix]:before{content:attr(data-prefix);display:inline-block;text-align:right}.modal{pointer-events:none;position:fixed;inset:0;margin:0;display:grid;height:100%;max-height:none;width:100%;max-width:none;justify-items:center;padding:0;opacity:0;overscroll-behavior:contain;z-index:999}.modal-scroll{overscroll-behavior:auto}:where(.modal){align-items:center}.modal-box{max-height:calc(100vh - 5em)}.modal-open,.modal-toggle:checked+.modal,.modal:target,.modal[open]{pointer-events:auto;visibility:visible;opacity:1}.modal-action{display:flex}.modal-toggle{position:fixed;height:0;width:0;appearance:none;opacity:0}:root:has(:is(.modal-open,.modal:target,.modal-toggle:checked+.modal,.modal[open])){overflow:hidden;scrollbar-gutter:stable}.navbar{display:flex;align-items:center}:where(.navbar>:not(script,style)){display:inline-flex;align-items:center}.navbar-start{width:50%;justify-content:flex-start}.navbar-center{flex-shrink:0}.navbar-end{width:50%;justify-content:flex-end}.progress{position:relative;width:100%;appearance:none;overflow:hidden}.radial-progress{position:relative;display:inline-grid;height:var(--size);width:var(--size);place-content:center;border-radius:9999px;background-color:transparent;vertical-align:middle;box-sizing:content-box}.radial-progress::-moz-progress-bar{appearance:none;background-color:transparent}.radial-progress::-webkit-progress-value{appearance:none;background-color:transparent}.radial-progress::-webkit-progress-bar{appearance:none;background-color:transparent}.radial-progress:after,.radial-progress:before{position:absolute;border-radius:9999px;content:""}.radial-progress:before{inset:0;background:radial-gradient(farthest-side,currentColor 98%,#0000) top/var(--thickness) var(--thickness) no-repeat,conic-gradient(currentColor calc(var(--value) * 1%),#0000 0);-webkit-mask:radial-gradient(farthest-side,#0000 calc(99% - var(--thickness)),#000 calc(100% - var(--thickness)));mask:radial-gradient(farthest-side,#0000 calc(99% - var(--thickness)),#000 calc(100% - var(--thickness)))}.radial-progress:after{inset:calc(50% - var(--thickness)/ 2);transform:rotate(calc(var(--value) * 3.6deg - 90deg)) translate(calc(var(--size)/ 2 - 50%))}.radio{flex-shrink:0}.range{height:1.5rem;width:100%;cursor:pointer}.range:focus{outline:0}.rating{position:relative;display:inline-flex}.rating :where(input){cursor:pointer;border-radius:0}.select{display:inline-flex;cursor:pointer;user-select:none;appearance:none;height:3rem;min-height:3rem;padding-inline-start:1rem;padding-inline-end:2.5rem;font-size:.875rem;line-height:1.25rem;line-height:2}.select[multiple]{height:auto}.stack{display:inline-grid}.stack>*{grid-column-start:1;grid-row-start:1;transform:translateY(10%) scale(.9);z-index:1}.stack>:nth-child(2){transform:translateY(5%) scale(.95);z-index:2}.stack>:nth-child(1){transform:translateY(0) scale(1);z-index:3}.stats{display:inline-grid}:where(.stats){grid-auto-flow:column}.stat{display:inline-grid;width:100%;grid-template-columns:repeat(1,1fr)}.stat-figure{grid-column-start:2;grid-row:span 3/span 3;grid-row-start:1;place-self:center;justify-self:end}.stat-title{grid-column-start:1;white-space:nowrap}.stat-value{grid-column-start:1;white-space:nowrap}.stat-desc{grid-column-start:1;white-space:nowrap}.stat-actions{grid-column-start:1;white-space:nowrap}.steps{display:inline-grid;grid-auto-flow:column;overflow:hidden;overflow-x:auto;counter-reset:step;grid-auto-columns:1fr}.steps .step{display:grid;grid-template-columns:repeat(1,minmax(0,1fr));grid-template-rows:repeat(2,minmax(0,1fr));place-items:center;text-align:center}.swap{position:relative;display:inline-grid;user-select:none;place-content:center}.swap>*{grid-column-start:1;grid-row-start:1}.swap input{appearance:none}.swap .swap-indeterminate,.swap .swap-on,.swap input:indeterminate~.swap-on{opacity:0}.swap input:checked~.swap-off,.swap input:indeterminate~.swap-off,.swap-active .swap-off{opacity:0}.swap input:checked~.swap-on,.swap input:indeterminate~.swap-indeterminate,.swap-active .swap-on{opacity:1}.tabs{display:grid;align-items:flex-end}.tabs-lifted:has(.tab-content[class*=" rounded-"]) .tab:first-child:not(:is(.tab-active,[aria-selected=true])),.tabs-lifted:has(.tab-content[class^=rounded-]) .tab:first-child:not(:is(.tab-active,[aria-selected=true])){border-bottom-color:transparent}.tab{position:relative;grid-row-start:1;display:inline-flex;height:2rem;cursor:pointer;user-select:none;appearance:none;flex-wrap:wrap;align-items:center;justify-content:center;text-align:center;font-size:.875rem;line-height:1.25rem;line-height:2;--tab-padding:1rem}.tab:is(input[type=radio]){width:auto;border-bottom-right-radius:0;border-bottom-left-radius:0}.tab:is(input[type=radio]):after{--tw-content:attr(aria-label);content:var(--tw-content)}.tab:not(input):empty{cursor:default;grid-column-start:span 9999}.tab-content{grid-column-start:1;grid-column-end:span 9999;grid-row-start:2;margin-top:calc(var(--tab-border) * -1);display:none;border-color:transparent;border-width:var(--tab-border,0)}:checked+.tab-content:nth-child(2),:is(.tab-active,[aria-selected=true])+.tab-content:nth-child(2){border-start-start-radius:0}:is(.tab-active,[aria-selected=true])+.tab-content,input.tab:checked+.tab-content{display:block}.table{position:relative;width:100%}.table :where(.table-pin-rowstheadtr){position:sticky;top:0;z-index:1;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)))}.table :where(.table-pin-rowstfoottr){position:sticky;bottom:0;z-index:1;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)))}.table :where(.table-pin-colstrth){position:sticky;left:0;right:0;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)))}.table-zebra tbody tr:nth-child(even) :where(.table-pin-colstrth){--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)))}.textarea{min-height:3rem;flex-shrink:1;padding-left:1rem;padding-right:1rem;padding-top:.5rem;padding-bottom:.5rem;font-size:.875rem;line-height:1.25rem;line-height:2}.timeline{position:relative;display:flex}:where(.timeline>li){position:relative;display:grid;flex-shrink:0;align-items:center;grid-template-rows:var(--timeline-row-start,minmax(0,1fr)) auto var(--timeline-row-end,minmax(0,1fr));grid-template-columns:var(--timeline-col-start,minmax(0,1fr)) auto var(--timeline-col-end,minmax(0,1fr))}.timeline>li>hr{width:100%;border-width:0}:where(.timeline>li>hr):first-child{grid-column-start:1;grid-row-start:2}:where(.timeline>li>hr):last-child{grid-column-start:3;grid-column-end:none;grid-row-start:2;grid-row-end:auto}.timeline-start{grid-column-start:1;grid-column-end:4;grid-row-start:1;grid-row-end:2;margin:.25rem;align-self:flex-end;justify-self:center}.timeline-middle{grid-column-start:2;grid-row-start:2}.timeline-end{grid-column-start:1;grid-column-end:4;grid-row-start:3;grid-row-end:4;margin:.25rem;align-self:flex-start;justify-self:center}.toast{position:fixed;display:flex;min-width:fit-content;flex-direction:column;white-space:nowrap}.toggle{flex-shrink:0}.alert{border-radius:var(--rounded-box,1rem);border-width:1px;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));padding:1rem;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--alert-bg:var(--fallback-b2,oklch(var(--b2)/1));--alert-bg-mix:var(--fallback-b1,oklch(var(--b1)/1));background-color:var(--alert-bg)}.alert-info{border-color:var(--fallback-in,oklch(var(--in)/.2));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)));--alert-bg:var(--fallback-in,oklch(var(--in)/1));--alert-bg-mix:var(--fallback-b1,oklch(var(--b1)/1))}.alert-success{border-color:var(--fallback-su,oklch(var(--su)/.2));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)));--alert-bg:var(--fallback-su,oklch(var(--su)/1));--alert-bg-mix:var(--fallback-b1,oklch(var(--b1)/1))}.alert-warning{border-color:var(--fallback-wa,oklch(var(--wa)/.2));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)));--alert-bg:var(--fallback-wa,oklch(var(--wa)/1));--alert-bg-mix:var(--fallback-b1,oklch(var(--b1)/1))}.alert-error{border-color:var(--fallback-er,oklch(var(--er)/.2));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)));--alert-bg:var(--fallback-er,oklch(var(--er)/1));--alert-bg-mix:var(--fallback-b1,oklch(var(--b1)/1))}.avatar-group{display:flex;overflow:hidden}.avatar-group :where(.avatar){overflow:hidden;border-radius:9999px;border-width:4px;--tw-border-opacity:1;border-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-border-opacity)))}.badge{border-radius:var(--rounded-badge,1.9rem);border-width:1px;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}.badge-neutral{--tw-border-opacity:1;border-color:var(--fallback-n,oklch(var(--n)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}.badge-primary{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.badge-secondary{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.badge-accent{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.badge-info{border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.badge-success{border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.badge-warning{border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.badge-error{border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.badge-ghost{--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}.badge-outline{border-color:currentColor;--tw-border-opacity:0.5;background-color:transparent;color:currentColor}.badge-outline.badge-neutral{--tw-text-opacity:1;color:var(--fallback-n,oklch(var(--n)/var(--tw-text-opacity)))}.badge-outline.badge-primary{--tw-text-opacity:1;color:var(--fallback-p,oklch(var(--p)/var(--tw-text-opacity)))}.badge-outline.badge-secondary{--tw-text-opacity:1;color:var(--fallback-s,oklch(var(--s)/var(--tw-text-opacity)))}.badge-outline.badge-accent{--tw-text-opacity:1;color:var(--fallback-a,oklch(var(--a)/var(--tw-text-opacity)))}.badge-outline.badge-info{--tw-text-opacity:1;color:var(--fallback-in,oklch(var(--in)/var(--tw-text-opacity)))}.badge-outline.badge-success{--tw-text-opacity:1;color:var(--fallback-su,oklch(var(--su)/var(--tw-text-opacity)))}.badge-outline.badge-warning{--tw-text-opacity:1;color:var(--fallback-wa,oklch(var(--wa)/var(--tw-text-opacity)))}.badge-outline.badge-error{--tw-text-opacity:1;color:var(--fallback-er,oklch(var(--er)/var(--tw-text-opacity)))}.btm-nav{height:4rem;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));color:currentColor}.btm-nav>*{border-color:currentColor}.btm-nav>:not(.active){padding-top:.125rem}.btm-nav>:where(.active){border-top-width:2px;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)))}.btm-nav>.disabled,.btm-nav>[disabled]{pointer-events:none;--tw-border-opacity:0;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}@media (hover:hover){.btm-nav>.disabled:hover,.btm-nav>[disabled]:hover{pointer-events:none;--tw-border-opacity:0;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}}.btm-nav>* .label{font-size:1rem;line-height:1.5rem}.breadcrumbs{padding-top:.5rem;padding-bottom:.5rem}.breadcrumbs>ol>li>a:focus,.breadcrumbs>ul>li>a:focus{outline:2px solid transparent;outline-offset:2px}.breadcrumbs>ol>li>a:focus-visible,.breadcrumbs>ul>li>a:focus-visible{outline:2px solid currentColor;outline-offset:2px}.breadcrumbs>ol>li+:before,.breadcrumbs>ul>li+:before{content:"";margin-left:.5rem;margin-right:.75rem;display:block;height:.375rem;width:.375rem;--tw-rotate:45deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));opacity:.4;border-top:1px solid;border-right:1px solid;background-color:transparent}[dir=rtl] .breadcrumbs>ol>li+:before,[dir=rtl] .breadcrumbs>ul>li+:before{--tw-rotate:-135deg}.btn{gap:.5rem;font-weight:600;text-decoration-line:none;transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);border-width:var(--border-btn,1px);transition-property:color,background-color,border-color,opacity,box-shadow,transform}@media (prefers-reduced-motion:no-preference){.btn{animation:button-pop var(--animation-btn,.25s) ease-out}}.btn:active:focus,.btn:active:hover{animation:button-pop 0s ease-out;transform:scale(var(--btn-focus-scale,.97))}.btn{--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));text-decoration-line:none;--tw-shadow:0 1px 2px 0 rgb(0 0 0 / 0.05);--tw-shadow-colored:0 1px 2px 0 var(--tw-shadow-color);box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow);outline-color:var(--fallback-bc,oklch(var(--bc)/1));background-color:oklch(var(--btn-color,var(--b2)) / var(--tw-bg-opacity));--tw-bg-opacity:1;border-color:oklch(var(--btn-color,var(--b2)) / var(--tw-border-opacity));--tw-border-opacity:1}@supports not (color:oklch(0% 0 0)){.btn{background-color:var(--btn-color,var(--fallback-b2));border-color:var(--btn-color,var(--fallback-b2))}}@media (hover:hover){.btn:hover{--tw-border-opacity:1;border-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn:hover{background-color:color-mix(in oklab,oklch(var(--btn-color,var(--b2)) / var(--tw-bg-opacity,1)) 90%,#000);border-color:color-mix(in oklab,oklch(var(--btn-color,var(--b2)) / var(--tw-border-opacity,1)) 90%,#000)}}@supports not (color:oklch(0% 0 0)){.btn:hover{background-color:var(--btn-color,var(--fallback-b2));border-color:var(--btn-color,var(--fallback-b2))}}}@supports (color:color-mix(in oklab,black,black)){.btn-active{background-color:color-mix(in oklab,oklch(var(--btn-color,var(--b3)) / var(--tw-bg-opacity,1)) 90%,#000);border-color:color-mix(in oklab,oklch(var(--btn-color,var(--b3)) / var(--tw-border-opacity,1)) 90%,#000)}}.btn:focus-visible{outline-style:solid;outline-width:2px;outline-offset:2px}.btn-primary{--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)));outline-color:var(--fallback-p,oklch(var(--p)/1))}@supports (color:oklch(0% 0 0)){.btn-primary{--btn-color:var(--p)}}@supports not (color:oklch(0% 0 0)){.btn-primary{--btn-color:var(--fallback-p)}}.btn-secondary{--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)));outline-color:var(--fallback-s,oklch(var(--s)/1))}@supports (color:oklch(0% 0 0)){.btn-secondary{--btn-color:var(--s)}}@supports not (color:oklch(0% 0 0)){.btn-secondary{--btn-color:var(--fallback-s)}}.btn-accent{--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)));outline-color:var(--fallback-a,oklch(var(--a)/1))}@supports (color:oklch(0% 0 0)){.btn-accent{--btn-color:var(--a)}}@supports not (color:oklch(0% 0 0)){.btn-accent{--btn-color:var(--fallback-a)}}.btn-neutral{--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)));outline-color:var(--fallback-n,oklch(var(--n)/1))}@supports (color:oklch(0% 0 0)){.btn-neutral{--btn-color:var(--n)}}@supports not (color:oklch(0% 0 0)){.btn-neutral{--btn-color:var(--fallback-n)}}.btn-info{--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)));outline-color:var(--fallback-in,oklch(var(--in)/1))}@supports (color:oklch(0% 0 0)){.btn-info{--btn-color:var(--in)}}@supports not (color:oklch(0% 0 0)){.btn-info{--btn-color:var(--fallback-in)}}.btn-success{--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)));outline-color:var(--fallback-su,oklch(var(--su)/1))}@supports (color:oklch(0% 0 0)){.btn-success{--btn-color:var(--su)}}@supports not (color:oklch(0% 0 0)){.btn-success{--btn-color:var(--fallback-su)}}.btn-warning{--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)));outline-color:var(--fallback-wa,oklch(var(--wa)/1))}@supports (color:oklch(0% 0 0)){.btn-warning{--btn-color:var(--wa)}}@supports not (color:oklch(0% 0 0)){.btn-warning{--btn-color:var(--fallback-wa)}}.btn-error{--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)));outline-color:var(--fallback-er,oklch(var(--er)/1))}@supports (color:oklch(0% 0 0)){.btn-error{--btn-color:var(--er)}}@supports not (color:oklch(0% 0 0)){.btn-error{--btn-color:var(--fallback-er)}}.btn.glass{--tw-shadow:0 0 #0000;--tw-shadow-colored:0 0 #0000;box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow);outline-color:currentColor}@media (hover:hover){.btn.glass:hover{--glass-opacity:25%;--glass-border-opacity:15%}}.btn.glass.btn-active{--glass-opacity:25%;--glass-border-opacity:15%}.btn-ghost{border-width:1px;border-color:transparent;background-color:transparent;color:currentColor;--tw-shadow:0 0 #0000;--tw-shadow-colored:0 0 #0000;box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow);outline-color:currentColor}@media (hover:hover){.btn-ghost:hover{border-color:transparent}@supports (color:oklch(0% 0 0)){.btn-ghost:hover{background-color:var(--fallback-bc,oklch(var(--bc)/.2))}}}.btn-ghost.btn-active{border-color:transparent;background-color:var(--fallback-bc,oklch(var(--bc)/.2))}.btn-link{border-color:transparent;background-color:transparent;--tw-text-opacity:1;color:var(--fallback-p,oklch(var(--p)/var(--tw-text-opacity)));text-decoration-line:underline;--tw-shadow:0 0 #0000;--tw-shadow-colored:0 0 #0000;box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow);outline-color:currentColor}@media (hover:hover){.btn-link:hover{border-color:transparent;background-color:transparent;text-decoration-line:underline}}.btn-link.btn-active{border-color:transparent;background-color:transparent;text-decoration-line:underline}.btn-outline{border-color:currentColor;background-color:transparent;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-shadow:0 0 #0000;--tw-shadow-colored:0 0 #0000;box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow)}@media (hover:hover){.btn-outline:hover{--tw-border-opacity:1;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-b1,oklch(var(--b1)/var(--tw-text-opacity)))}}.btn-outline.btn-active{--tw-border-opacity:1;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-b1,oklch(var(--b1)/var(--tw-text-opacity)))}.btn-outline.btn-primary{--tw-text-opacity:1;color:var(--fallback-p,oklch(var(--p)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-primary:hover{--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-primary:hover{background-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000)}}}.btn-outline.btn-primary.btn-active{--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-primary.btn-active{background-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000)}}.btn-outline.btn-secondary{--tw-text-opacity:1;color:var(--fallback-s,oklch(var(--s)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-secondary:hover{--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-secondary:hover{background-color:color-mix(in oklab,var(--fallback-s,oklch(var(--s)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-s,oklch(var(--s)/1)) 90%,#000)}}}.btn-outline.btn-secondary.btn-active{--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-secondary.btn-active{background-color:color-mix(in oklab,var(--fallback-s,oklch(var(--s)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-s,oklch(var(--s)/1)) 90%,#000)}}.btn-outline.btn-accent{--tw-text-opacity:1;color:var(--fallback-a,oklch(var(--a)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-accent:hover{--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-accent:hover{background-color:color-mix(in oklab,var(--fallback-a,oklch(var(--a)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-a,oklch(var(--a)/1)) 90%,#000)}}}.btn-outline.btn-accent.btn-active{--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-accent.btn-active{background-color:color-mix(in oklab,var(--fallback-a,oklch(var(--a)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-a,oklch(var(--a)/1)) 90%,#000)}}.btn-outline.btn-success{--tw-text-opacity:1;color:var(--fallback-su,oklch(var(--su)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-success:hover{--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-success:hover{background-color:color-mix(in oklab,var(--fallback-su,oklch(var(--su)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-su,oklch(var(--su)/1)) 90%,#000)}}}.btn-outline.btn-success.btn-active{--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-success.btn-active{background-color:color-mix(in oklab,var(--fallback-su,oklch(var(--su)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-su,oklch(var(--su)/1)) 90%,#000)}}.btn-outline.btn-info{--tw-text-opacity:1;color:var(--fallback-in,oklch(var(--in)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-info:hover{--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-info:hover{background-color:color-mix(in oklab,var(--fallback-in,oklch(var(--in)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-in,oklch(var(--in)/1)) 90%,#000)}}}.btn-outline.btn-info.btn-active{--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-info.btn-active{background-color:color-mix(in oklab,var(--fallback-in,oklch(var(--in)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-in,oklch(var(--in)/1)) 90%,#000)}}.btn-outline.btn-warning{--tw-text-opacity:1;color:var(--fallback-wa,oklch(var(--wa)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-warning:hover{--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-warning:hover{background-color:color-mix(in oklab,var(--fallback-wa,oklch(var(--wa)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-wa,oklch(var(--wa)/1)) 90%,#000)}}}.btn-outline.btn-warning.btn-active{--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-warning.btn-active{background-color:color-mix(in oklab,var(--fallback-wa,oklch(var(--wa)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-wa,oklch(var(--wa)/1)) 90%,#000)}}.btn-outline.btn-error{--tw-text-opacity:1;color:var(--fallback-er,oklch(var(--er)/var(--tw-text-opacity)))}@media (hover:hover){.btn-outline.btn-error:hover{--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-error:hover{background-color:color-mix(in oklab,var(--fallback-er,oklch(var(--er)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-er,oklch(var(--er)/1)) 90%,#000)}}}.btn-outline.btn-error.btn-active{--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}@supports (color:color-mix(in oklab,black,black)){.btn-outline.btn-error.btn-active{background-color:color-mix(in oklab,var(--fallback-er,oklch(var(--er)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-er,oklch(var(--er)/1)) 90%,#000)}}.btn.btn-disabled,.btn:disabled,.btn[disabled]{--tw-border-opacity:0;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.2;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}@media (hover:hover){.btn-disabled:hover,.btn:disabled:hover,.btn[disabled]:hover{--tw-border-opacity:0;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.2;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}}.btn:is(input[type=checkbox]:checked),.btn:is(input[type=radio]:checked){--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}@media (hover:hover){@supports (color:color-mix(in oklab,black,black)){.btn:is(input[type=checkbox]:checked):hover,.btn:is(input[type=radio]:checked):hover{background-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000);border-color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 90%,#000)}}}.btn:is(input[type=checkbox]:checked):focus-visible,.btn:is(input[type=radio]:checked):focus-visible{outline-color:var(--fallback-p,oklch(var(--p)/1))}@keyframes button-pop{0%{transform:scale(var(--btn-focus-scale,.98))}40%{transform:scale(1.02)}100%{transform:scale(1)}}.card{border-radius:var(--rounded-box,1rem)}.card :where(figure:first-child){overflow:hidden;border-start-start-radius:inherit;border-start-end-radius:inherit;border-end-start-radius:unset;border-end-end-radius:unset}.card :where(figure:last-child){overflow:hidden;border-start-start-radius:unset;border-start-end-radius:unset;border-end-start-radius:inherit;border-end-end-radius:inherit}.card:focus-visible{outline:2px solid currentColor;outline-offset:2px}.card.bordered{border-width:1px;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)))}.card-bordered{border-width:1px;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)))}.card.compact .card-body{padding:1rem;font-size:.875rem;line-height:1.25rem}.card-body{padding:var(--padding-card,2rem);display:flex;flex-direction:column;gap:.5rem}.card-title{display:flex;align-items:center;gap:.5rem;font-size:1.25rem;line-height:1.75rem;font-weight:600}.card.image-full:before{z-index:10;border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));opacity:.75}.card.image-full>.card-body{z-index:20;--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}.card.image-full :where(figure){overflow:hidden;border-radius:inherit}.carousel{-ms-overflow-style:none;scrollbar-width:none}.carousel::-webkit-scrollbar{display:none}.chat-bubble{border-radius:var(--rounded-box,1rem);min-height:2.75rem;min-width:2.75rem}.chat-bubble{--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}.chat-bubble-primary{--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.chat-bubble-secondary{--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.chat-bubble-accent{--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.chat-bubble-info{--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.chat-bubble-success{--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.chat-bubble-warning{--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.chat-bubble-error{--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.chat-start .chat-bubble{border-end-start-radius:0}.chat-start .chat-bubble:before{inset-inline-start:-.749rem}.chat-end .chat-bubble{border-end-end-radius:0}.chat-end .chat-bubble:before{inset-inline-start:99.9%}.checkbox{--chkbg:var(--fallback-bc,oklch(var(--bc)/1));--chkfg:var(--fallback-b1,oklch(var(--b1)/1));height:1.5rem;width:1.5rem;cursor:pointer;appearance:none;border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0.2}.checkbox:focus{box-shadow:none}.checkbox:focus-visible{outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/1))}.checkbox:disabled{border-width:0}.checkbox:checked,.checkbox[aria-checked=true]{background-repeat:no-repeat;animation:checkmark var(--animation-input,.2s) ease-out;background-color:var(--chkbg);background-image:linear-gradient(-45deg,transparent 65%,var(--chkbg) 65.99%),linear-gradient(45deg,transparent 75%,var(--chkbg) 75.99%),linear-gradient(-45deg,var(--chkbg) 40%,transparent 40.99%),linear-gradient(45deg,var(--chkbg) 30%,var(--chkfg) 30.99%,var(--chkfg) 40%,transparent 40.99%),linear-gradient(-45deg,var(--chkfg) 50%,var(--chkbg) 50.99%)}.checkbox:indeterminate{--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));background-repeat:no-repeat;animation:checkmark var(--animation-input,.2s) ease-out;background-image:linear-gradient(90deg,transparent 80%,var(--chkbg) 80%),linear-gradient(-90deg,transparent 80%,var(--chkbg) 80%),linear-gradient(0deg,var(--chkbg) 43%,var(--chkfg) 43%,var(--chkfg) 57%,var(--chkbg) 57%)}.checkbox-primary{--chkbg:var(--fallback-p,oklch(var(--p)/1));--chkfg:var(--fallback-pc,oklch(var(--pc)/1));--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-primary:hover{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}}.checkbox-primary:focus-visible{outline-color:var(--fallback-p,oklch(var(--p)/1))}.checkbox-primary:checked,.checkbox-primary[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.checkbox-secondary{--chkbg:var(--fallback-s,oklch(var(--s)/1));--chkfg:var(--fallback-sc,oklch(var(--sc)/1));--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-secondary:hover{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}}.checkbox-secondary:focus-visible{outline-color:var(--fallback-s,oklch(var(--s)/1))}.checkbox-secondary:checked,.checkbox-secondary[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.checkbox-accent{--chkbg:var(--fallback-a,oklch(var(--a)/1));--chkfg:var(--fallback-ac,oklch(var(--ac)/1));--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-accent:hover{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}}.checkbox-accent:focus-visible{outline-color:var(--fallback-a,oklch(var(--a)/1))}.checkbox-accent:checked,.checkbox-accent[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.checkbox-success{--chkbg:var(--fallback-su,oklch(var(--su)/1));--chkfg:var(--fallback-suc,oklch(var(--suc)/1));--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-success:hover{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}}.checkbox-success:focus-visible{outline-color:var(--fallback-su,oklch(var(--su)/1))}.checkbox-success:checked,.checkbox-success[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.checkbox-warning{--chkbg:var(--fallback-wa,oklch(var(--wa)/1));--chkfg:var(--fallback-wac,oklch(var(--wac)/1));--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-warning:hover{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}}.checkbox-warning:focus-visible{outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.checkbox-warning:checked,.checkbox-warning[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.checkbox-info{--chkbg:var(--fallback-in,oklch(var(--in)/1));--chkfg:var(--fallback-inc,oklch(var(--inc)/1));--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-info:hover{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}}.checkbox-info:focus-visible{outline-color:var(--fallback-in,oklch(var(--in)/1))}.checkbox-info:checked,.checkbox-info[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.checkbox-error{--chkbg:var(--fallback-er,oklch(var(--er)/1));--chkfg:var(--fallback-erc,oklch(var(--erc)/1));--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}@media(hover:hover){.checkbox-error:hover{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}}.checkbox-error:focus-visible{outline-color:var(--fallback-er,oklch(var(--er)/1))}.checkbox-error:checked,.checkbox-error[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.checkbox:disabled{cursor:not-allowed;border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));opacity:.2}@keyframes checkmark{0%{background-position-y:5px}50%{background-position-y:-2px}100%{background-position-y:0}}.checkbox-mark{display:none}.collapse{width:100%;border-radius:var(--rounded-box,1rem)}details.collapse{width:100%}details.collapse summary{position:relative;display:block}details.collapse summary::-webkit-details-marker{display:none}.collapse:focus-visible{outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/1))}details.collapse summary{outline:2px solid transparent;outline-offset:2px}.collapse:has(.collapse-title:focus-visible),.collapse:has(>input[type=checkbox]:focus-visible),.collapse:has(>input[type=radio]:focus-visible){outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/1))}.collapse-arrow>.collapse-title:after{position:absolute;display:block;height:.5rem;width:.5rem;--tw-translate-y:-100%;--tw-rotate:45deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));transition-property:all;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:150ms;transition-timing-function:cubic-bezier(0,0,.2,1);transition-duration:.2s;top:1.9rem;inset-inline-end:1.4rem;content:"";transform-origin:75% 75%;box-shadow:2px 2px;pointer-events:none}.collapse-plus>.collapse-title:after{position:absolute;display:block;height:.5rem;width:.5rem;transition-property:all;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.3s;transition-timing-function:cubic-bezier(0,0,.2,1);top:.9rem;inset-inline-end:1.4rem;content:"+";pointer-events:none}.collapse:not(.collapse-open):not(.collapse-close)>.collapse-title,.collapse:not(.collapse-open):not(.collapse-close)>input[type=checkbox],.collapse:not(.collapse-open):not(.collapse-close)>input[type=radio]:not(:checked){cursor:pointer}.collapse:focus:not(.collapse-open):not(.collapse-close):not(.collapse[open])>.collapse-title{cursor:unset}.collapse-title{position:relative}:where(.collapse>input[type=checkbox]),:where(.collapse>input[type=radio]){z-index:1}.collapse-title,:where(.collapse>input[type=checkbox]),:where(.collapse>input[type=radio]){width:100%;padding:1rem;padding-inline-end:3rem;min-height:3.75rem;transition:background-color .2s ease-out}.collapse-content{padding-left:1rem;padding-right:1rem;cursor:unset;transition:padding .2s ease-out,background-color .2s ease-out}.collapse-open>:where(.collapse-content),.collapse:focus:not(.collapse-close)>:where(.collapse-content),.collapse:not(.collapse-close)>:where(input[type=checkbox]:checked~.collapse-content),.collapse:not(.collapse-close)>:where(input[type=radio]:checked~.collapse-content),.collapse[open]>:where(.collapse-content){padding-bottom:1rem;transition:padding .2s ease-out,background-color .2s ease-out}.collapse-arrow:focus:not(.collapse-close)>.collapse-title:after,.collapse-arrow:not(.collapse-close)>input[type=checkbox]:checked~.collapse-title:after,.collapse-arrow:not(.collapse-close)>input[type=radio]:checked~.collapse-title:after,.collapse-open.collapse-arrow>.collapse-title:after,.collapse[open].collapse-arrow>.collapse-title:after{--tw-translate-y:-50%;--tw-rotate:225deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}.collapse-open.collapse-plus>.collapse-title:after,.collapse-plus:focus:not(.collapse-close)>.collapse-title:after,.collapse-plus:not(.collapse-close)>input[type=checkbox]:checked~.collapse-title:after,.collapse-plus:not(.collapse-close)>input[type=radio]:checked~.collapse-title:after,.collapse[open].collapse-plus>.collapse-title:after{content:"−"}.countdown>:before{text-align:center;transition:all 1s cubic-bezier(1,0,0,1)}.diff-item-1:after{border-radius:9999px;border-width:2px;--tw-border-opacity:1;border-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-border-opacity)));background-color:var(--fallback-b1,oklch(var(--b1)/.5));--tw-shadow:0 1px 2px 0 rgb(0 0 0 / 0.05);--tw-shadow-colored:0 1px 2px 0 var(--tw-shadow-color);box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow);outline-style:solid;outline-offset:-3px;outline-color:var(--fallback-bc,oklch(var(--bc)/.05));--tw-backdrop-blur:blur(8px);backdrop-filter:var(--tw-backdrop-blur) var(--tw-backdrop-brightness) var(--tw-backdrop-contrast) var(--tw-backdrop-grayscale) var(--tw-backdrop-hue-rotate) var(--tw-backdrop-invert) var(--tw-backdrop-opacity) var(--tw-backdrop-saturate) var(--tw-backdrop-sepia);translate:50% -50%}.diff-item-2{border-right-width:2px;--tw-border-opacity:1;border-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-border-opacity)))}.divider{margin-top:1rem;margin-bottom:1rem;height:1rem;white-space:nowrap}.divider:after,.divider:before{background-color:var(--fallback-bc,oklch(var(--bc)/.1))}.divider:not(:empty){gap:1rem}.divider-neutral:after,.divider-neutral:before{--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)))}.divider-primary:after,.divider-primary:before{--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)))}.divider-secondary:after,.divider-secondary:before{--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)))}.divider-accent:after,.divider-accent:before{--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)))}.divider-success:after,.divider-success:before{--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)))}.divider-warning:after,.divider-warning:before{--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)))}.divider-info:after,.divider-info:before{--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)))}.divider-error:after,.divider-error:before{--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)))}.drawer{width:100%}.drawer-side>.drawer-overlay{cursor:pointer;background-color:transparent;transition-property:color,background-color,border-color,text-decoration-color,fill,stroke;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1)}.drawer-toggle:checked~.drawer-side>.drawer-overlay{background-color:#0006}.drawer-toggle:focus-visible~.drawer-content label.drawer-button{outline-style:solid;outline-width:2px;outline-offset:2px}.dropdown:is(:not(details)) .dropdown-content{transform-origin:top;--tw-scale-x:.95;--tw-scale-y:.95;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));transition-property:color,background-color,border-color,text-decoration-color,fill,stroke,opacity,box-shadow,transform,filter,backdrop-filter;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1)}.dropdown-bottom .dropdown-content{transform-origin:top}.dropdown-top .dropdown-content{transform-origin:bottom}.dropdown-left .dropdown-content{transform-origin:right}.dropdown-right .dropdown-content{transform-origin:left}.dropdown.dropdown-open .dropdown-content,.dropdown:focus .dropdown-content,.dropdown:focus-within .dropdown-content{--tw-scale-x:1;--tw-scale-y:1;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}@media (hover:hover){.dropdown.dropdown-hover:hover .dropdown-content{--tw-scale-x:1;--tw-scale-y:1;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}}.file-input{overflow:hidden;border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));font-size:1rem;line-height:1.5rem}.file-input::file-selector-button{border-style:solid;--tw-border-opacity:1;border-color:var(--fallback-n,oklch(var(--n)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));font-weight:600;text-transform:uppercase;--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)));text-decoration-line:none;border-width:var(--border-btn,1px);animation:button-pop var(--animation-btn,.25s) ease-out}.file-input-bordered{--tw-border-opacity:0.2}.file-input:focus{outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/.2))}.file-input-ghost{--tw-bg-opacity:0.05}.file-input-ghost:focus{--tw-bg-opacity:1;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));box-shadow:none}.file-input-ghost::file-selector-button{border-width:1px;border-color:transparent;background-color:transparent;color:currentColor}.file-input-primary{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}.file-input-primary:focus{outline-color:var(--fallback-p,oklch(var(--p)/1))}.file-input-primary::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.file-input-secondary{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}.file-input-secondary:focus{outline-color:var(--fallback-s,oklch(var(--s)/1))}.file-input-secondary::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.file-input-accent{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}.file-input-accent:focus{outline-color:var(--fallback-a,oklch(var(--a)/1))}.file-input-accent::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.file-input-info{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}.file-input-info:focus{outline-color:var(--fallback-in,oklch(var(--in)/1))}.file-input-info::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.file-input-success{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}.file-input-success:focus{outline-color:var(--fallback-su,oklch(var(--su)/1))}.file-input-success::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.file-input-warning{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}.file-input-warning:focus{outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.file-input-warning::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.file-input-error{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}.file-input-error:focus{outline-color:var(--fallback-er,oklch(var(--er)/1))}.file-input-error::file-selector-button{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.file-input-disabled,.file-input[disabled]{cursor:not-allowed;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));--tw-text-opacity:0.2}.file-input-disabled::placeholder,.file-input[disabled]::placeholder{color:var(--fallback-bc,oklch(var(--bc)/var(--tw-placeholder-opacity)));--tw-placeholder-opacity:0.2}.file-input-disabled::file-selector-button,.file-input[disabled]::file-selector-button{--tw-border-opacity:0;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.2;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}.footer{column-gap:1rem;row-gap:2.5rem;font-size:.875rem;line-height:1.25rem}.footer>*{gap:.5rem}.footer-title{margin-bottom:.5rem;font-weight:700;text-transform:uppercase;opacity:.6}.label{padding-left:.25rem;padding-right:.25rem;padding-top:.5rem;padding-bottom:.5rem}.label-text{font-size:.875rem;line-height:1.25rem;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}.label-text-alt{font-size:.75rem;line-height:1rem;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}@media(hover:hover){.label a:hover{--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}}.hero-overlay{background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-bg-opacity:0.5}.hero-content{max-width:80rem;gap:1rem;padding:1rem}.input{border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));font-size:1rem;line-height:1.5rem}.input input{--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));background-color:transparent}.input input:focus{outline:2px solid transparent;outline-offset:2px}.input[list]::-webkit-calendar-picker-indicator{line-height:1em}.input-bordered{border-color:var(--fallback-bc,oklch(var(--bc)/.2))}.input:focus,.input:focus-within{box-shadow:none;border-color:var(--fallback-bc,oklch(var(--bc)/.2));outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/.2))}.input-ghost{--tw-bg-opacity:0.05}.input-ghost:focus,.input-ghost:focus-within{--tw-bg-opacity:1;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));box-shadow:none}.input-primary{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}.input-primary:focus,.input-primary:focus-within{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));outline-color:var(--fallback-p,oklch(var(--p)/1))}.input-secondary{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}.input-secondary:focus,.input-secondary:focus-within{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));outline-color:var(--fallback-s,oklch(var(--s)/1))}.input-accent{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}.input-accent:focus,.input-accent:focus-within{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));outline-color:var(--fallback-a,oklch(var(--a)/1))}.input-info{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}.input-info:focus,.input-info:focus-within{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));outline-color:var(--fallback-in,oklch(var(--in)/1))}.input-success{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}.input-success:focus,.input-success:focus-within{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));outline-color:var(--fallback-su,oklch(var(--su)/1))}.input-warning{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}.input-warning:focus,.input-warning:focus-within{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.input-error{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}.input-error:focus,.input-error:focus-within{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));outline-color:var(--fallback-er,oklch(var(--er)/1))}.input-disabled,.input:disabled,.input:has(>input[disabled]),.input[disabled]{cursor:not-allowed;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));color:var(--fallback-bc,oklch(var(--bc)/.4))}.input-disabled::placeholder,.input:disabled::placeholder,.input:has(>input[disabled])::placeholder,.input[disabled]::placeholder{color:var(--fallback-bc,oklch(var(--bc)/var(--tw-placeholder-opacity)));--tw-placeholder-opacity:0.2}.input:has(>input[disabled])>input[disabled]{cursor:not-allowed}.input::-webkit-date-and-time-value{text-align:inherit}.join{border-radius:var(--rounded-btn,.5rem)}.join>:where(:not(:first-child)){margin-top:0;margin-bottom:0;margin-inline-start:-1px}.join>:where(:not(:first-child)):is(.btn){margin-inline-start:calc(var(--border-btn) * -1)}.join-item:focus{isolation:isolate}.kbd{border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0.2;--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));padding-left:.5rem;padding-right:.5rem;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));border-bottom-width:2px;min-height:2.2em;min-width:2.2em}.link-primary{--tw-text-opacity:1;color:var(--fallback-p,oklch(var(--p)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-primary:hover{color:color-mix(in oklab,var(--fallback-p,oklch(var(--p)/1)) 80%,#000)}}}.link-secondary{--tw-text-opacity:1;color:var(--fallback-s,oklch(var(--s)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-secondary:hover{color:color-mix(in oklab,var(--fallback-s,oklch(var(--s)/1)) 80%,#000)}}}.link-accent{--tw-text-opacity:1;color:var(--fallback-a,oklch(var(--a)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-accent:hover{color:color-mix(in oklab,var(--fallback-a,oklch(var(--a)/1)) 80%,#000)}}}.link-neutral{--tw-text-opacity:1;color:var(--fallback-n,oklch(var(--n)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-neutral:hover{color:color-mix(in oklab,var(--fallback-n,oklch(var(--n)/1)) 80%,#000)}}}.link-success{--tw-text-opacity:1;color:var(--fallback-su,oklch(var(--su)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-success:hover{color:color-mix(in oklab,var(--fallback-su,oklch(var(--su)/1)) 80%,#000)}}}.link-info{--tw-text-opacity:1;color:var(--fallback-in,oklch(var(--in)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-info:hover{color:color-mix(in oklab,var(--fallback-in,oklch(var(--in)/1)) 80%,#000)}}}.link-warning{--tw-text-opacity:1;color:var(--fallback-wa,oklch(var(--wa)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-warning:hover{color:color-mix(in oklab,var(--fallback-wa,oklch(var(--wa)/1)) 80%,#000)}}}.link-error{--tw-text-opacity:1;color:var(--fallback-er,oklch(var(--er)/var(--tw-text-opacity)))}@supports(color:color-mix(in oklab,black,black)){@media(hover:hover){.link-error:hover{color:color-mix(in oklab,var(--fallback-er,oklch(var(--er)/1)) 80%,#000)}}}.link:focus{outline:2px solid transparent;outline-offset:2px}.link:focus-visible{outline:2px solid currentColor;outline-offset:2px}.loading{pointer-events:none;display:inline-block;aspect-ratio:1/1;width:1.5rem;background-color:currentColor;mask-size:100%;mask-repeat:no-repeat;mask-position:center;mask-image:url("data:image/svg+xml,%3Csvg width='24' height='24' stroke='%23000' viewBox='0 0 24 24' xmlns='http://www.w3.org/2000/svg'%3E%3Cstyle%3E.spinner_V8m1%7Btransform-origin:center;animation:spinner_zKoa 2s linear infinite%7D.spinner_V8m1 circle%7Bstroke-linecap:round;animation:spinner_YpZS 1.5s ease-out infinite%7D%40keyframes spinner_zKoa%7B100%25%7Btransform:rotate(360deg)%7D%7D%40keyframes spinner_YpZS%7B0%25%7Bstroke-dasharray:0 150;stroke-dashoffset:0%7D47.5%25%7Bstroke-dasharray:42 150;stroke-dashoffset:-16%7D95%25%2C100%25%7Bstroke-dasharray:42 150;stroke-dashoffset:-59%7D%7D%3C%2Fstyle%3E%3Cg class='spinner_V8m1'%3E%3Ccircle cx='12' cy='12' r='9.5' fill='none' stroke-width='3'%3E%3C%2Fcircle%3E%3C%2Fg%3E%3C%2Fsvg%3E")}.loading-spinner{mask-image:url("data:image/svg+xml,%3Csvg width='24' height='24' stroke='%23000' viewBox='0 0 24 24' xmlns='http://www.w3.org/2000/svg'%3E%3Cstyle%3E.spinner_V8m1%7Btransform-origin:center;animation:spinner_zKoa 2s linear infinite%7D.spinner_V8m1 circle%7Bstroke-linecap:round;animation:spinner_YpZS 1.5s ease-out infinite%7D%40keyframes spinner_zKoa%7B100%25%7Btransform:rotate(360deg)%7D%7D%40keyframes spinner_YpZS%7B0%25%7Bstroke-dasharray:0 150;stroke-dashoffset:0%7D47.5%25%7Bstroke-dasharray:42 150;stroke-dashoffset:-16%7D95%25%2C100%25%7Bstroke-dasharray:42 150;stroke-dashoffset:-59%7D%7D%3C%2Fstyle%3E%3Cg class='spinner_V8m1'%3E%3Ccircle cx='12' cy='12' r='9.5' fill='none' stroke-width='3'%3E%3C%2Fcircle%3E%3C%2Fg%3E%3C%2Fsvg%3E")}.loading-dots{mask-image:url("data:image/svg+xml,%3Csvg width='24' height='24' viewBox='0 0 24 24' xmlns='http://www.w3.org/2000/svg'%3E%3Cstyle%3E.spinner_qM83%7Banimation:spinner_8HQG 1.05s infinite%7D.spinner_oXPr%7Banimation-delay:.1s%7D.spinner_ZTLf%7Banimation-delay:.2s%7D@keyframes spinner_8HQG%7B0%25,57.14%25%7Banimation-timing-function:cubic-bezier(0.33,.66,.66,1);transform:translate(0)%7D28.57%25%7Banimation-timing-function:cubic-bezier(0.33,0,.66,.33);transform:translateY(-6px)%7D100%25%7Btransform:translate(0)%7D%7D%3C/style%3E%3Ccircle class='spinner_qM83' cx='4' cy='12' r='3'/%3E%3Ccircle class='spinner_qM83 spinner_oXPr' cx='12' cy='12' r='3'/%3E%3Ccircle class='spinner_qM83 spinner_ZTLf' cx='20' cy='12' r='3'/%3E%3C/svg%3E")}.loading-ring{mask-image:url("data:image/svg+xml,%3Csvg width='44' height='44' viewBox='0 0 44 44' xmlns='http://www.w3.org/2000/svg' stroke='%23fff'%3E%3Cg fill='none' fill-rule='evenodd' stroke-width='2'%3E%3Ccircle cx='22' cy='22' r='1'%3E%3Canimate attributeName='r' begin='0s' dur='1.8s' values='1; 20' calcMode='spline' keyTimes='0; 1' keySplines='0.165, 0.84, 0.44, 1' repeatCount='indefinite' /%3E%3Canimate attributeName='stroke-opacity' begin='0s' dur='1.8s' values='1; 0' calcMode='spline' keyTimes='0; 1' keySplines='0.3, 0.61, 0.355, 1' repeatCount='indefinite' /%3E%3C/circle%3E%3Ccircle cx='22' cy='22' r='1'%3E%3Canimate attributeName='r' begin='-0.9s' dur='1.8s' values='1; 20' calcMode='spline' keyTimes='0; 1' keySplines='0.165, 0.84, 0.44, 1' repeatCount='indefinite' /%3E%3Canimate attributeName='stroke-opacity' begin='-0.9s' dur='1.8s' values='1; 0' calcMode='spline' keyTimes='0; 1' keySplines='0.3, 0.61, 0.355, 1' repeatCount='indefinite' /%3E%3C/circle%3E%3C/g%3E%3C/svg%3E")}.loading-ball{mask-image:url("data:image/svg+xml,%0A%3Csvg width='24' height='24' viewBox='0 0 24 24' xmlns='http://www.w3.org/2000/svg'%3E%3Cstyle%3E.spinner_rXNP%7Banimation:spinner_YeBj .8s infinite%7D@keyframes spinner_YeBj%7B0%25%7Banimation-timing-function:cubic-bezier(0.33,0,.66,.33);cy:5px%7D46.875%25%7Bcy:20px;rx:4px;ry:4px%7D50%25%7Banimation-timing-function:cubic-bezier(0.33,.66,.66,1);cy:20.5px;rx:4.8px;ry:3px%7D53.125%25%7Brx:4px;ry:4px%7D100%25%7Bcy:5px%7D%7D%3C/style%3E%3Cellipse class='spinner_rXNP' cx='12' cy='5' rx='4' ry='4'/%3E%3C/svg%3E")}.loading-bars{mask-image:url("data:image/svg+xml,%0A%3Csvg width='24' height='24' viewBox='0 0 24 24' xmlns='http://www.w3.org/2000/svg'%3E%3Cstyle%3E.spinner_hzlK%7Banimation:spinner_vc4H .8s linear infinite;animation-delay:-.8s%7D.spinner_koGT%7Banimation-delay:-.65s%7D.spinner_YF1u%7Banimation-delay:-.5s%7D@keyframes spinner_vc4H%7B0%25%7By:1px;height:22px%7D93.75%25%7By:5px;height:14px;opacity:.2%7D%7D%3C/style%3E%3Crect class='spinner_hzlK' x='1' y='1' width='6' height='22'/%3E%3Crect class='spinner_hzlK spinner_koGT' x='9' y='1' width='6' height='22'/%3E%3Crect class='spinner_hzlK spinner_YF1u' x='17' y='1' width='6' height='22'/%3E%3C/svg%3E")}.loading-infinity{mask-image:url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' xmlns:xlink='http://www.w3.org/1999/xlink' style='shape-rendering: auto;' width='200px' height='200px' viewBox='0 0 100 100' preserveAspectRatio='xMidYMid'%3E%3Cpath fill='none' stroke='%230a0a0a' stroke-width='10' stroke-dasharray='205.271142578125 51.317785644531256' d='M24.3 30C11.4 30 5 43.3 5 50s6.4 20 19.3 20c19.3 0 32.1-40 51.4-40 C88.6 30 95 43.3 95 50s-6.4 20-19.3 20C56.4 70 43.6 30 24.3 30z' stroke-linecap='round' style='transform:scale(0.8);transform-origin:50px 50px'%3E%3Canimate attributeName='stroke-dashoffset' repeatCount='indefinite' dur='2s' keyTimes='0;1' values='0;256.58892822265625'%3E%3C/animate%3E%3C/path%3E%3C/svg%3E")}.loading-xs{width:1rem}.loading-sm{width:1.25rem}.loading-md{width:1.5rem}.loading-lg{width:2.5rem}.mask-squircle{mask-image:url("data:image/svg+xml,%3csvg width='200' height='200' xmlns='http://www.w3.org/2000/svg'%3e%3cpath d='M100 0C20 0 0 20 0 100s20 100 100 100 100-20 100-100S180 0 100 0Z'/%3e%3c/svg%3e")}.mask-decagon{mask-image:url("data:image/svg+xml,%3csvg width='192' height='200' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m96 0 58.779 19.098 36.327 50v61.804l-36.327 50L96 200l-58.779-19.098-36.327-50V69.098l36.327-50z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-diamond{mask-image:url("data:image/svg+xml,%3csvg width='200' height='200' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m100 0 100 100-100 100L0 100z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-heart{mask-image:url("data:image/svg+xml,%3csvg width='200' height='185' xmlns='http://www.w3.org/2000/svg'%3e%3cpath d='M100 184.606a15.384 15.384 0 0 1-8.653-2.678C53.565 156.28 37.205 138.695 28.182 127.7 8.952 104.264-.254 80.202.005 54.146.308 24.287 24.264 0 53.406 0c21.192 0 35.869 11.937 44.416 21.879a2.884 2.884 0 0 0 4.356 0C110.725 11.927 125.402 0 146.594 0c29.142 0 53.098 24.287 53.4 54.151.26 26.061-8.956 50.122-28.176 73.554-9.023 10.994-25.383 28.58-63.165 54.228a15.384 15.384 0 0 1-8.653 2.673Z' fill='black' fill-rule='nonzero'/%3e%3c/svg%3e")}.mask-hexagon{mask-image:url("data:image/svg+xml,%3csvg width='182' height='201' xmlns='http://www.w3.org/2000/svg'%3e%3cpath d='M.3 65.486c0-9.196 6.687-20.063 14.211-25.078l61.86-35.946c8.36-5.016 20.899-5.016 29.258 0l61.86 35.946c8.36 5.015 14.211 15.882 14.211 25.078v71.055c0 9.196-6.687 20.063-14.211 25.079l-61.86 35.945c-8.36 4.18-20.899 4.18-29.258 0L14.51 161.62C6.151 157.44.3 145.737.3 136.54V65.486Z' fill='black' fill-rule='nonzero'/%3e%3c/svg%3e")}.mask-hexagon-2{mask-image:url("data:image/svg+xml,%3csvg width='200' height='182' xmlns='http://www.w3.org/2000/svg'%3e%3cpath d='M64.786 181.4c-9.196 0-20.063-6.687-25.079-14.21L3.762 105.33c-5.016-8.36-5.016-20.9 0-29.259l35.945-61.86C44.723 5.851 55.59 0 64.786 0h71.055c9.196 0 20.063 6.688 25.079 14.211l35.945 61.86c4.18 8.36 4.18 20.899 0 29.258l-35.945 61.86c-4.18 8.36-15.883 14.211-25.079 14.211H64.786Z' fill='black' fill-rule='nonzero'/%3e%3c/svg%3e")}.mask-circle{mask-image:url("data:image/svg+xml,%3csvg width='200' height='200' xmlns='http://www.w3.org/2000/svg'%3e%3ccircle fill='black' cx='100' cy='100' r='100' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-parallelogram{mask-image:url("data:image/svg+xml,%3csvg width='200' height='154' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M46.154 0H200l-46.154 153.846H0z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-parallelogram-2{mask-image:url("data:image/svg+xml,%3csvg width='200' height='154' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M153.846 0H0l46.154 153.846H200z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-parallelogram-3{mask-image:url("data:image/svg+xml,%3csvg width='154' height='201' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M.077 47.077v153.846l153.846-46.154V.923z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-parallelogram-4{mask-image:url("data:image/svg+xml,%3csvg width='154' height='201' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M153.923 47.077v153.846L.077 154.77V.923z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-pentagon{mask-image:url("data:image/svg+xml,%3csvg width='192' height='181' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m96 0 95.106 69.098-36.327 111.804H37.22L.894 69.098z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-square{mask-image:url("data:image/svg+xml,%3csvg width='200' height='200' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M0 0h200v200H0z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-star{mask-image:url("data:image/svg+xml,%3csvg width='192' height='180' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m96 137.263-58.779 42.024 22.163-68.389L.894 68.481l72.476-.243L96 0l22.63 68.238 72.476.243-58.49 42.417 22.163 68.389z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-star-2{mask-image:url("data:image/svg+xml,%3csvg width='192' height='180' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m96 153.044-58.779 26.243 7.02-63.513L.894 68.481l63.117-13.01L96 0l31.989 55.472 63.117 13.01-43.347 47.292 7.02 63.513z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-triangle{mask-image:url("data:image/svg+xml,%3csvg width='174' height='149' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m87 148.476-86.603.185L43.86 74.423 87 0l43.14 74.423 43.463 74.238z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-triangle-2{mask-image:url("data:image/svg+xml,%3csvg width='174' height='150' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m87 .738 86.603-.184-43.463 74.238L87 149.214 43.86 74.792.397.554z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-triangle-3{mask-image:url("data:image/svg+xml,%3csvg width='150' height='174' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='m149.369 87.107.185 86.603-74.239-43.463L.893 87.107l74.422-43.14L149.554.505z' fill-rule='evenodd'/%3e%3c/svg%3e")}.mask-triangle-4{mask-image:url("data:image/svg+xml,%3csvg width='150' height='174' xmlns='http://www.w3.org/2000/svg'%3e%3cpath fill='black' d='M.631 87.107.446.505l74.239 43.462 74.422 43.14-74.422 43.14L.446 173.71z' fill-rule='evenodd'/%3e%3c/svg%3e")}.menu{padding:.5rem}:where(.menuli:empty){--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));opacity:.1;margin:.5rem 1rem;height:1px}.menu :where(liul){margin-inline-start:1rem;padding-inline-start:.5rem}.menu :where(liul):before{position:absolute;bottom:.75rem;inset-inline-start:0;top:.75rem;width:1px;--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));opacity:.1;content:""}.menu :where(li:not(.menu-title)>:not(ul,details,.menu-title,.btn)),.menu :where(li:not(.menu-title)>details>summary:not(.menu-title)){border-radius:var(--rounded-btn,.5rem);padding-left:1rem;padding-right:1rem;padding-top:.5rem;padding-bottom:.5rem;text-align:start;transition-property:color,background-color,border-color,text-decoration-color,fill,stroke,opacity,box-shadow,transform,filter,backdrop-filter;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);text-wrap:balance}:where(.menuli:not(.menu-title,.disabled)>:not(ul,details,.menu-title)):is(summary):not(.active,.btn):focus-visible,:where(.menuli:not(.menu-title,.disabled)>:not(ul,details,.menu-title)):not(summary,.active,.btn).focus,:where(.menuli:not(.menu-title,.disabled)>:not(ul,details,.menu-title)):not(summary,.active,.btn):focus,:where(.menuli:not(.menu-title,.disabled)>details>summary:not(.menu-title)):is(summary):not(.active,.btn):focus-visible,:where(.menuli:not(.menu-title,.disabled)>details>summary:not(.menu-title)):not(summary,.active,.btn).focus,:where(.menuli:not(.menu-title,.disabled)>details>summary:not(.menu-title)):not(summary,.active,.btn):focus{cursor:pointer;background-color:var(--fallback-bc,oklch(var(--bc)/.1));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));outline:2px solid transparent;outline-offset:2px}@media (hover:hover){:where(.menuli:not(.menu-title,.disabled)>:not(ul,details,.menu-title)):not(.active,.btn):hover,:where(.menuli:not(.menu-title,.disabled)>details>summary:not(.menu-title)):not(.active,.btn):hover{cursor:pointer;outline:2px solid transparent;outline-offset:2px}@supports (color:oklch(0% 0 0)){:where(.menuli:not(.menu-title,.disabled)>:not(ul,details,.menu-title)):not(.active,.btn):hover,:where(.menuli:not(.menu-title,.disabled)>details>summary:not(.menu-title)):not(.active,.btn):hover{background-color:var(--fallback-bc,oklch(var(--bc)/.1))}}}.menu li>:not(ul,.menu-title,details,.btn).active,.menu li>:not(ul,.menu-title,details,.btn):active,.menu li>details>summary:active{--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}@media(hover:hover){.menu li>:not(ul,.menu-title,details,.btn).active,.menu li>:not(ul,.menu-title,details,.btn):active,.menu li>details>summary:active{--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}}.menu li.disabled{color:var(--fallback-bc,oklch(var(--bc)/.3))}.menu :where(li>details>summary)::-webkit-details-marker{display:none}.menu :where(li>.menu-dropdown-toggle):after,.menu :where(li>details>summary):after{justify-self:end;display:block;margin-top:-.5rem;height:.5rem;width:.5rem;transform:rotate(45deg);transition-property:transform,margin-top;transition-duration:.3s;transition-timing-function:cubic-bezier(.4,0,.2,1);content:"";transform-origin:75% 75%;box-shadow:2px 2px;pointer-events:none}.menu :where(li>.menu-dropdown-toggle.menu-dropdown-show):after,.menu :where(li>details[open]>summary):after{transform:rotate(225deg);margin-top:0}.menu-title{padding-left:1rem;padding-right:1rem;padding-top:.5rem;padding-bottom:.5rem;font-size:.875rem;line-height:1.25rem;font-weight:700;color:var(--fallback-bc,oklch(var(--bc)/.4))}.mockup-code{min-width:18rem;border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));padding-top:1.25rem;padding-bottom:1.25rem;--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)));direction:ltr}.mockup-code:before{content:"";margin-bottom:1rem;display:block;height:.75rem;width:.75rem;border-radius:9999px;opacity:.3;box-shadow:1.4em 0,2.8em 0,4.2em 0}.mockup-code pre{padding-right:1.25rem}.mockup-code pre:before{content:"";margin-right:2ch}.mockup-code pre[data-prefix]:before{content:attr(data-prefix);width:2rem;opacity:.5}.mockup-window{display:flex;flex-direction:column;border-radius:var(--rounded-box,1rem);padding-top:1.25rem}.mockup-window:before{content:"";margin-bottom:1rem;display:block;aspect-ratio:1/1;height:.75rem;flex-shrink:0;align-self:flex-start;border-radius:9999px;opacity:.3}.mockup-window:where([dir=rtl],[dir=rtl]*):before{align-self:flex-end}.mockup-window:before{box-shadow:1.4em 0,2.8em 0,4.2em 0}.mockup-phone{display:inline-block;border:4px solid #444;border-radius:50px;background-color:#000;padding:10px;margin:0 auto;overflow:hidden}.mockup-phone .camera{position:relative;top:0;left:0;background:#000;height:25px;width:150px;margin:0 auto;border-bottom-left-radius:17px;border-bottom-right-radius:17px;z-index:11}.mockup-phone .camera:before{content:"";position:absolute;top:35%;left:50%;width:50px;height:4px;border-radius:5px;background-color:#0c0b0e;transform:translate(-50%,-50%)}.mockup-phone .camera:after{content:"";position:absolute;top:20%;left:70%;width:8px;height:8px;border-radius:5px;background-color:#0f0b25}.mockup-phone .display{overflow:hidden;border-radius:40px;margin-top:-25px}.mockup-browser{border-radius:var(--rounded-box,1rem)}.mockup-browser .mockup-browser-toolbar{margin-top:.75rem;margin-bottom:.75rem;display:inline-flex;width:100%;align-items:center;padding-right:1.4em}.mockup-browser .mockup-browser-toolbar:where([dir=rtl],[dir=rtl]*){flex-direction:row-reverse}.mockup-browser .mockup-browser-toolbar:before{content:"";margin-right:4.8rem;display:inline-block;aspect-ratio:1/1;height:.75rem;border-radius:9999px;opacity:.3;box-shadow:1.4em 0,2.8em 0,4.2em 0}.mockup-browser .mockup-browser-toolbar .input{position:relative;margin-left:auto;margin-right:auto;display:block;height:1.75rem;width:24rem;overflow:hidden;text-overflow:ellipsis;white-space:nowrap;--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));padding-left:2rem;direction:ltr}.mockup-browser .mockup-browser-toolbar .input:before{content:"";position:absolute;left:.5rem;top:50%;aspect-ratio:1/1;height:.75rem;--tw-translate-y:-50%;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));border-radius:9999px;border-width:2px;border-color:currentColor;opacity:.6}.mockup-browser .mockup-browser-toolbar .input:after{content:"";position:absolute;left:1.25rem;top:50%;height:.5rem;--tw-translate-y:25%;--tw-rotate:-45deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));border-radius:9999px;border-width:1px;border-color:currentColor;opacity:.6}.modal{background-color:transparent;color:inherit;transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);transition-property:transform,opacity,visibility;overflow-y:hidden;overscroll-behavior:contain}.modal::backdrop,.modal:not(dialog:not(.modal-open)){background-color:#0006;animation:modal-pop .2s ease-out}.modal-backdrop{z-index:-1;grid-column-start:1;grid-row-start:1;display:grid;align-self:stretch;justify-self:stretch;color:transparent}.modal-box{grid-column-start:1;grid-row-start:1;width:91.666667%;max-width:32rem;--tw-scale-x:.9;--tw-scale-y:.9;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));border-bottom-right-radius:var(--rounded-box,1rem);border-bottom-left-radius:var(--rounded-box,1rem);border-top-left-radius:var(--rounded-box,1rem);border-top-right-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));padding:1.5rem;transition-property:color,background-color,border-color,text-decoration-color,fill,stroke,opacity,box-shadow,transform,filter,backdrop-filter;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.2s;transition-timing-function:cubic-bezier(0,0,.2,1);box-shadow:rgba(0,0,0,.25) 0 25px 50px -12px;overflow-y:auto;overscroll-behavior:contain}.modal-open .modal-box,.modal-toggle:checked+.modal .modal-box,.modal:target .modal-box,.modal[open] .modal-box{--tw-translate-y:0px;--tw-scale-x:1;--tw-scale-y:1;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}.modal-action{margin-top:1.5rem;justify-content:flex-end}.modal-action>:not([hidden])~:not([hidden]){--tw-space-x-reverse:0;margin-right:calc(.5rem * var(--tw-space-x-reverse));margin-left:calc(.5rem * calc(1 - var(--tw-space-x-reverse)))}@keyframes modal-pop{0%{opacity:0}}.navbar{padding:var(--navbar-padding,.5rem);min-height:4rem;width:100%}.progress{height:.5rem;border-radius:var(--rounded-box,1rem);background-color:var(--fallback-bc,oklch(var(--bc)/.2))}.progress::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)))}.progress-primary::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)))}.progress-secondary::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)))}.progress-accent::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)))}.progress-info::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)))}.progress-success::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)))}.progress-warning::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)))}.progress-error::-moz-progress-bar{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)))}.progress:indeterminate{--progress-color:var(--fallback-bc,oklch(var(--bc)/1))}.progress-primary:indeterminate{--progress-color:var(--fallback-p,oklch(var(--p)/1))}.progress-secondary:indeterminate{--progress-color:var(--fallback-s,oklch(var(--s)/1))}.progress-accent:indeterminate{--progress-color:var(--fallback-a,oklch(var(--a)/1))}.progress-info:indeterminate{--progress-color:var(--fallback-in,oklch(var(--in)/1))}.progress-success:indeterminate{--progress-color:var(--fallback-su,oklch(var(--su)/1))}.progress-warning:indeterminate{--progress-color:var(--fallback-wa,oklch(var(--wa)/1))}.progress-error:indeterminate{--progress-color:var(--fallback-er,oklch(var(--er)/1))}.progress::-webkit-progress-bar{border-radius:var(--rounded-box,1rem);background-color:transparent}.progress::-webkit-progress-value{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)))}.progress-primary::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)))}.progress-secondary::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)))}.progress-accent::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)))}.progress-info::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)))}.progress-success::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)))}.progress-warning::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)))}.progress-error::-webkit-progress-value{--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)))}.progress:indeterminate{background-image:repeating-linear-gradient(90deg,var(--progress-color) -1%,var(--progress-color) 10%,transparent 10%,transparent 90%);background-size:200%;background-position-x:15%;animation:progress-loading 5s ease-in-out infinite}.progress:indeterminate::-moz-progress-bar{background-color:transparent;background-image:repeating-linear-gradient(90deg,var(--progress-color) -1%,var(--progress-color) 10%,transparent 10%,transparent 90%);background-size:200%;background-position-x:15%;animation:progress-loading 5s ease-in-out infinite}@keyframes progress-loading{50%{background-position-x:-115%}}.radial-progress{--value:0;--size:5rem;--thickness:calc(var(--size) / 10)}.radial-progress:after{background-color:currentColor}.radio{--chkbg:var(--bc);height:1.5rem;width:1.5rem;cursor:pointer;appearance:none;border-radius:9999px;border-width:1px;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0.2}.radio:focus{box-shadow:none}.radio:focus-visible{outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/1))}.radio:checked,.radio[aria-checked=true]{--tw-bg-opacity:1;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));background-image:none;animation:radiomark var(--animation-input,.2s) ease-out;box-shadow:0 0 0 4px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 4px var(--fallback-b1,oklch(var(--b1)/1)) inset}.radio-primary{--chkbg:var(--p);--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}@media(hover:hover){.radio-primary:hover{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}}.radio-primary:focus-visible{outline-color:var(--fallback-p,oklch(var(--p)/1))}.radio-primary:checked,.radio-primary[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.radio-secondary{--chkbg:var(--s);--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}@media(hover:hover){.radio-secondary:hover{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}}.radio-secondary:focus-visible{outline-color:var(--fallback-s,oklch(var(--s)/1))}.radio-secondary:checked,.radio-secondary[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.radio-accent{--chkbg:var(--a);--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}@media(hover:hover){.radio-accent:hover{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}}.radio-accent:focus-visible{outline-color:var(--fallback-a,oklch(var(--a)/1))}.radio-accent:checked,.radio-accent[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.radio-success{--chkbg:var(--su);--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}@media(hover:hover){.radio-success:hover{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}}.radio-success:focus-visible{outline-color:var(--fallback-su,oklch(var(--su)/1))}.radio-success:checked,.radio-success[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.radio-warning{--chkbg:var(--wa);--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}@media(hover:hover){.radio-warning:hover{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}}.radio-warning:focus-visible{outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.radio-warning:checked,.radio-warning[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.radio-info{--chkbg:var(--in);--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}@media(hover:hover){.radio-info:hover{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}}.radio-info:focus-visible{outline-color:var(--fallback-in,oklch(var(--in)/1))}.radio-info:checked,.radio-info[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.radio-error{--chkbg:var(--er);--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}@media(hover:hover){.radio-error:hover{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}}.radio-error:focus-visible{outline-color:var(--fallback-er,oklch(var(--er)/1))}.radio-error:checked,.radio-error[aria-checked=true]{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.radio:disabled{cursor:not-allowed;opacity:.2}@keyframes radiomark{0%{box-shadow:0 0 0 12px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 12px var(--fallback-b1,oklch(var(--b1)/1)) inset}50%{box-shadow:0 0 0 3px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 3px var(--fallback-b1,oklch(var(--b1)/1)) inset}100%{box-shadow:0 0 0 4px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 4px var(--fallback-b1,oklch(var(--b1)/1)) inset}}.radio-mark{display:none}.range{appearance:none;-webkit-appearance:none;--range-shdw:var(--fallback-bc,oklch(var(--bc)/1));overflow:hidden;border-radius:var(--rounded-box,1rem);background-color:transparent}.range:focus-visible::-webkit-slider-thumb{--focus-shadow:0 0 0 6px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 2rem var(--range-shdw) inset}.range:focus-visible::-moz-range-thumb{--focus-shadow:0 0 0 6px var(--fallback-b1,oklch(var(--b1)/1)) inset,0 0 0 2rem var(--range-shdw) inset}.range::-webkit-slider-runnable-track{height:.5rem;width:100%;border-radius:var(--rounded-box,1rem);background-color:var(--fallback-bc,oklch(var(--bc)/.1))}.range::-moz-range-track{height:.5rem;width:100%;border-radius:var(--rounded-box,1rem);background-color:var(--fallback-bc,oklch(var(--bc)/.1))}.range::-webkit-slider-thumb{position:relative;height:1.5rem;width:1.5rem;border-radius:var(--rounded-box,1rem);border-style:none;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));appearance:none;-webkit-appearance:none;top:50%;color:var(--range-shdw);transform:translateY(-50%);--filler-size:100rem;--filler-offset:0.6rem;box-shadow:0 0 0 3px var(--range-shdw) inset,var(--focus-shadow,0 0),calc(var(--filler-size) * -1 - var(--filler-offset)) 0 0 var(--filler-size)}.range::-moz-range-thumb{position:relative;height:1.5rem;width:1.5rem;border-radius:var(--rounded-box,1rem);border-style:none;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));top:50%;color:var(--range-shdw);--filler-size:100rem;--filler-offset:0.5rem;box-shadow:0 0 0 3px var(--range-shdw) inset,var(--focus-shadow,0 0),calc(var(--filler-size) * -1 - var(--filler-offset)) 0 0 var(--filler-size)}.range-primary{--range-shdw:var(--fallback-p,oklch(var(--p)/1))}.range-secondary{--range-shdw:var(--fallback-s,oklch(var(--s)/1))}.range-accent{--range-shdw:var(--fallback-a,oklch(var(--a)/1))}.range-success{--range-shdw:var(--fallback-su,oklch(var(--su)/1))}.range-warning{--range-shdw:var(--fallback-wa,oklch(var(--wa)/1))}.range-info{--range-shdw:var(--fallback-in,oklch(var(--in)/1))}.range-error{--range-shdw:var(--fallback-er,oklch(var(--er)/1))}.rating input{appearance:none;-webkit-appearance:none}.rating :where(input){animation:rating-pop var(--animation-input,.25s) ease-out;height:1.5rem;width:1.5rem;background-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-bg-opacity)));--tw-bg-opacity:1}.rating .rating-hidden{width:.5rem;background-color:transparent}.rating input[type=radio]:checked{background-image:none}.rating input:checked~input,.rating input[aria-checked=true]~input{--tw-bg-opacity:0.2}.rating input:focus-visible{transition-property:transform;transition-timing-function:cubic-bezier(.4,0,.2,1);transition-duration:.3s;transition-timing-function:cubic-bezier(0,0,.2,1);transform:translateY(-.125em)}.rating input:active:focus{animation:none;transform:translateY(-.125em)}.rating-half :where(input:not(.rating-hidden)){width:.75rem}@keyframes rating-pop{0%{transform:translateY(-.125em)}40%{transform:translateY(-.125em)}100%{transform:translateY(0)}}.select{border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));padding-inline-end:2.5rem}.select-bordered{border-color:var(--fallback-bc,oklch(var(--bc)/.2))}.select{background-image:linear-gradient(45deg,transparent 50%,currentColor 50%),linear-gradient(135deg,currentColor 50%,transparent 50%);background-position:calc(100% - 20px) calc(1px + 50%),calc(100% - 16.1px) calc(1px + 50%);background-size:4px 4px,4px 4px;background-repeat:no-repeat}.select:focus{box-shadow:none;border-color:var(--fallback-bc,oklch(var(--bc)/.2));outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/.2))}.select-ghost{--tw-bg-opacity:0.05}.select-ghost:focus{--tw-bg-opacity:1;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}.select-primary{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}.select-primary:focus{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));outline-color:var(--fallback-p,oklch(var(--p)/1))}.select-secondary{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}.select-secondary:focus{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));outline-color:var(--fallback-s,oklch(var(--s)/1))}.select-accent{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}.select-accent:focus{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));outline-color:var(--fallback-a,oklch(var(--a)/1))}.select-info{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}.select-info:focus{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));outline-color:var(--fallback-in,oklch(var(--in)/1))}.select-success{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}.select-success:focus{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));outline-color:var(--fallback-su,oklch(var(--su)/1))}.select-warning{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}.select-warning:focus{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.select-error{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}.select-error:focus{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));outline-color:var(--fallback-er,oklch(var(--er)/1))}.select-disabled,.select:disabled,.select[disabled]{cursor:not-allowed;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));color:var(--fallback-bc,oklch(var(--bc)/.4))}.select-disabled::placeholder,.select:disabled::placeholder,.select[disabled]::placeholder{color:var(--fallback-bc,oklch(var(--bc)/var(--tw-placeholder-opacity)));--tw-placeholder-opacity:0.2}.select-multiple,.select[multiple],.select[size].select:not([size="1"]){background-image:none;padding-right:1rem}[dir=rtl] .select{background-position:calc(0% + 12px) calc(1px + 50%),calc(0% + 16px) calc(1px + 50%)}.skeleton{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)));will-change:background-position;animation:skeleton 1.8s ease-in-out infinite;background-image:linear-gradient(105deg,transparent 0,transparent 40%,var(--fallback-b1,oklch(var(--b1)/1)) 50%,transparent 60%,transparent 100%);background-size:200% auto;background-repeat:no-repeat;background-position-x:-50%}@media (prefers-reduced-motion){.skeleton{animation-duration:15s}}@keyframes skeleton{from{background-position:150%}to{background-position:-50%}}.stack{place-items:center;align-items:flex-end}.stack>*{width:100%;opacity:.6}.stack>:nth-child(2){opacity:.8}.stack>:nth-child(1){opacity:1}.stats{border-radius:var(--rounded-box,1rem);--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}:where(.stats)>:not([hidden])~:not([hidden]){--tw-divide-x-reverse:0;border-right-width:calc(1px * var(--tw-divide-x-reverse));border-left-width:calc(1px * calc(1 - var(--tw-divide-x-reverse)));--tw-divide-y-reverse:0;border-top-width:calc(0px * calc(1 - var(--tw-divide-y-reverse)));border-bottom-width:calc(0px * var(--tw-divide-y-reverse))}:where(.stats){overflow-x:auto}[dir=rtl] .stats>:not([hidden])~:not([hidden]){--tw-divide-x-reverse:1}.stat{column-gap:1rem;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0.1;padding-left:1.5rem;padding-right:1.5rem;padding-top:1rem;padding-bottom:1rem}.stat-title{color:var(--fallback-bc,oklch(var(--bc)/.6))}.stat-value{font-size:2.25rem;line-height:2.5rem;font-weight:800}.stat-desc{font-size:.75rem;line-height:1rem;color:var(--fallback-bc,oklch(var(--bc)/.6))}.stat-actions{margin-top:1rem}.steps .step{grid-template-rows:40px 1fr;grid-template-columns:auto;min-width:4rem}.steps .step:before{top:0;grid-column-start:1;grid-row-start:1;height:.5rem;width:100%;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y));--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));content:"";margin-inline-start:-100%}.steps .step:after{content:counter(step);counter-increment:step;z-index:1;position:relative;grid-column-start:1;grid-row-start:1;display:grid;height:2rem;width:2rem;place-items:center;place-self:center;border-radius:9999px;--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}.steps .step:first-child:before{content:none}.steps .step[data-content]:after{content:attr(data-content)}.steps .step-neutral+.step-neutral:before,.steps .step-neutral:after{--tw-bg-opacity:1;background-color:var(--fallback-n,oklch(var(--n)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-nc,oklch(var(--nc)/var(--tw-text-opacity)))}.steps .step-primary+.step-primary:before,.steps .step-primary:after{--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.steps .step-secondary+.step-secondary:before,.steps .step-secondary:after{--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.steps .step-accent+.step-accent:before,.steps .step-accent:after{--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.steps .step-info+.step-info:before{--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)))}.steps .step-info:after{--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.steps .step-success+.step-success:before{--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)))}.steps .step-success:after{--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.steps .step-warning+.step-warning:before{--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)))}.steps .step-warning:after{--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.steps .step-error+.step-error:before{--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)))}.steps .step-error:after{--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.swap{cursor:pointer}.swap>*{transition-duration:.3s;transition-timing-function:cubic-bezier(0,0,.2,1);transition-property:transform,opacity}.swap-rotate .swap-indeterminate,.swap-rotate .swap-on,.swap-rotate input:indeterminate~.swap-on{--tw-rotate:45deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}.swap-active:where(.swap-rotate) .swap-off,.swap-rotate input:checked~.swap-off,.swap-rotate input:indeterminate~.swap-off{--tw-rotate:-45deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}.swap-active:where(.swap-rotate) .swap-on,.swap-rotate input:checked~.swap-on,.swap-rotate input:indeterminate~.swap-indeterminate{--tw-rotate:0deg;transform:translate(var(--tw-translate-x),var(--tw-translate-y)) rotate(var(--tw-rotate)) skewX(var(--tw-skew-x)) skewY(var(--tw-skew-y)) scaleX(var(--tw-scale-x)) scaleY(var(--tw-scale-y))}.swap-flip{transform-style:preserve-3d;perspective:16em}.swap-flip .swap-indeterminate,.swap-flip .swap-on,.swap-flip input:indeterminate~.swap-on{transform:rotateY(180deg);backface-visibility:hidden;opacity:1}.swap-active:where(.swap-flip) .swap-off,.swap-flip input:checked~.swap-off,.swap-flip input:indeterminate~.swap-off{transform:rotateY(-180deg);backface-visibility:hidden;opacity:1}.swap-active:where(.swap-flip) .swap-on,.swap-flip input:checked~.swap-on,.swap-flip input:indeterminate~.swap-indeterminate{transform:rotateY(0)}.tabs-lifted>.tab:focus-visible{border-end-end-radius:0;border-end-start-radius:0}.tab{--tw-text-opacity:0.5}@media(hover:hover){.tab:hover{--tw-text-opacity:1}}.tab{--tab-color:var(--fallback-bc,oklch(var(--bc)/1));--tab-bg:var(--fallback-b1,oklch(var(--b1)/1));--tab-border-color:var(--fallback-b3,oklch(var(--b3)/1));color:var(--tab-color);padding-inline-start:var(--tab-padding,1rem);padding-inline-end:var(--tab-padding,1rem)}.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]),.tab:is(input:checked){border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:1;--tw-text-opacity:1}.tab:focus{outline:2px solid transparent;outline-offset:2px}.tab:focus-visible{outline:2px solid currentColor;outline-offset:-5px}.tab-disabled,.tab[disabled]{cursor:not-allowed;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}@media (hover:hover){.tab[disabled],.tab[disabled]:hover{cursor:not-allowed;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));--tw-text-opacity:0.2}}.tabs-bordered>.tab{border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));--tw-border-opacity:0.2;border-style:solid;border-bottom-width:calc(var(--tab-border,1px) + 1px)}.tabs-lifted>.tab{border:var(--tab-border,1px) solid transparent;border-width:0 0 var(--tab-border,1px) 0;border-start-start-radius:var(--tab-radius,.5rem);border-start-end-radius:var(--tab-radius,.5rem);border-bottom-color:var(--tab-border-color);padding-inline-start:var(--tab-padding,1rem);padding-inline-end:var(--tab-padding,1rem);padding-top:var(--tab-border,1px)}.tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]),.tabs-lifted>.tab:is(input:checked){background-color:var(--tab-bg);border-width:var(--tab-border,1px) var(--tab-border,1px) 0 var(--tab-border,1px);border-inline-start-color:var(--tab-border-color);border-inline-end-color:var(--tab-border-color);border-top-color:var(--tab-border-color);padding-inline-start:calc(var(--tab-padding,1rem) - var(--tab-border,1px));padding-inline-end:calc(var(--tab-padding,1rem) - var(--tab-border,1px));padding-bottom:var(--tab-border,1px);padding-top:0}.tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):before,.tabs-lifted>.tab:is(input:checked):before{z-index:1;content:"";display:block;position:absolute;width:calc(100% + var(--tab-radius,.5rem) * 2);height:var(--tab-radius,.5rem);bottom:0;background-size:var(--tab-radius,.5rem);background-position:top left,top right;background-repeat:no-repeat;--tab-grad:calc(69% - var(--tab-border, 1px));--radius-start:radial-gradient( - circle at top left, - transparent var(--tab-grad), - var(--tab-border-color) calc(var(--tab-grad) + 0.25px), - var(--tab-border-color) calc(var(--tab-grad) + var(--tab-border, 1px)), - var(--tab-bg) calc(var(--tab-grad) + var(--tab-border, 1px) + 0.25px) - );--radius-end:radial-gradient( - circle at top right, - transparent var(--tab-grad), - var(--tab-border-color) calc(var(--tab-grad) + 0.25px), - var(--tab-border-color) calc(var(--tab-grad) + var(--tab-border, 1px)), - var(--tab-bg) calc(var(--tab-grad) + var(--tab-border, 1px) + 0.25px) - );background-image:var(--radius-start),var(--radius-end)}.tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):first-child:before,.tabs-lifted>.tab:is(input:checked):first-child:before{background-image:var(--radius-end);background-position:top right}[dir=rtl] .tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):first-child:before,[dir=rtl] .tabs-lifted>.tab:is(input:checked):first-child:before{background-image:var(--radius-start);background-position:top left}.tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):last-child:before,.tabs-lifted>.tab:is(input:checked):last-child:before{background-image:var(--radius-start);background-position:top left}[dir=rtl] .tabs-lifted>.tab:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):last-child:before,[dir=rtl] .tabs-lifted>.tab:is(input:checked):last-child:before{background-image:var(--radius-end);background-position:top right}.tabs-lifted>.tab:is(input:checked)+.tabs-lifted .tab:is(input:checked):before,.tabs-lifted>:is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled])+.tabs-lifted :is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):before{background-image:var(--radius-end);background-position:top right}.tabs-boxed{border-radius:var(--rounded-btn,.5rem);--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));padding:.25rem}.tabs-boxed .tab{border-radius:var(--rounded-btn,.5rem)}.tabs-boxed :is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]),.tabs-boxed :is(input:checked){--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}@media(hover:hover){.tabs-boxed :is(.tab-active,[aria-selected=true]):not(.tab-disabled):not([disabled]):hover,.tabs-boxed :is(input:checked):hover{--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}}.table{border-radius:var(--rounded-box,1rem);text-align:left;font-size:.875rem;line-height:1.25rem}.table:where([dir=rtl],[dir=rtl]*){text-align:right}.table :where(th,td){padding-left:1rem;padding-right:1rem;padding-top:.75rem;padding-bottom:.75rem;vertical-align:middle}.table tr.active,.table tr.active:nth-child(even),.table-zebra tbody tr:nth-child(even){--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)))}@media(hover:hover){.table tr.hover:hover,.table tr.hover:nth-child(even):hover{--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)))}}.table-zebra tr.active,.table-zebra tr.active:nth-child(even),.table-zebra-zebra tbody tr:nth-child(even){--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)))}@media(hover:hover){.table-zebra tr.hover:hover,.table-zebra tr.hover:nth-child(even):hover{--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)))}}.table :where(theadtr,tbodytr:not(:last-child),tbodytr:first-child:last-child){border-bottom-width:1px;--tw-border-opacity:1;border-bottom-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)))}.table :where(thead,tfoot){white-space:nowrap;font-size:.75rem;line-height:1rem;font-weight:700;color:var(--fallback-bc,oklch(var(--bc)/.6))}.table :where(tfoot){border-top-width:1px;--tw-border-opacity:1;border-top-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)))}.textarea{border-radius:var(--rounded-btn,.5rem);border-width:1px;border-color:transparent;--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)))}.textarea-bordered{border-color:var(--fallback-bc,oklch(var(--bc)/.2))}.textarea:focus{box-shadow:none;border-color:var(--fallback-bc,oklch(var(--bc)/.2));outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/.2))}.textarea-ghost{--tw-bg-opacity:0.05}.textarea-ghost:focus{--tw-bg-opacity:1;--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));box-shadow:none}.textarea-primary{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)))}.textarea-primary:focus{--tw-border-opacity:1;border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));outline-color:var(--fallback-p,oklch(var(--p)/1))}.textarea-secondary{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)))}.textarea-secondary:focus{--tw-border-opacity:1;border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));outline-color:var(--fallback-s,oklch(var(--s)/1))}.textarea-accent{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)))}.textarea-accent:focus{--tw-border-opacity:1;border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));outline-color:var(--fallback-a,oklch(var(--a)/1))}.textarea-info{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)))}.textarea-info:focus{--tw-border-opacity:1;border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));outline-color:var(--fallback-in,oklch(var(--in)/1))}.textarea-success{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)))}.textarea-success:focus{--tw-border-opacity:1;border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));outline-color:var(--fallback-su,oklch(var(--su)/1))}.textarea-warning{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)))}.textarea-warning:focus{--tw-border-opacity:1;border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.textarea-error{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)))}.textarea-error:focus{--tw-border-opacity:1;border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));outline-color:var(--fallback-er,oklch(var(--er)/1))}.textarea-disabled,.textarea:disabled,.textarea[disabled]{cursor:not-allowed;--tw-border-opacity:1;border-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b2,oklch(var(--b2)/var(--tw-bg-opacity)));color:var(--fallback-bc,oklch(var(--bc)/.4))}.textarea-disabled::placeholder,.textarea:disabled::placeholder,.textarea[disabled]::placeholder{color:var(--fallback-bc,oklch(var(--bc)/var(--tw-placeholder-opacity)));--tw-placeholder-opacity:0.2}.timeline hr{height:.25rem}:where(.timelinehr){--tw-bg-opacity:1;background-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-bg-opacity)))}:where(.timeline:has(.timeline-middle)hr):first-child{border-start-end-radius:var(--rounded-badge,1.9rem);border-end-end-radius:var(--rounded-badge,1.9rem);border-start-start-radius:0;border-end-start-radius:0}:where(.timeline:has(.timeline-middle)hr):last-child{border-start-start-radius:var(--rounded-badge,1.9rem);border-end-start-radius:var(--rounded-badge,1.9rem);border-start-end-radius:0;border-end-end-radius:0}:where(.timeline:not(:has(.timeline-middle)):first-childhr:last-child){border-start-start-radius:var(--rounded-badge,1.9rem);border-end-start-radius:var(--rounded-badge,1.9rem);border-start-end-radius:0;border-end-end-radius:0}:where(.timeline:not(:has(.timeline-middle)):last-childhr:first-child){border-start-end-radius:var(--rounded-badge,1.9rem);border-end-end-radius:var(--rounded-badge,1.9rem);border-start-start-radius:0;border-end-start-radius:0}.timeline-box{border-radius:var(--rounded-box,1rem);border-width:1px;--tw-border-opacity:1;border-color:var(--fallback-b3,oklch(var(--b3)/var(--tw-border-opacity)));--tw-bg-opacity:1;background-color:var(--fallback-b1,oklch(var(--b1)/var(--tw-bg-opacity)));padding-left:1rem;padding-right:1rem;padding-top:.5rem;padding-bottom:.5rem;--tw-shadow:0 1px 2px 0 rgb(0 0 0 / 0.05);--tw-shadow-colored:0 1px 2px 0 var(--tw-shadow-color);box-shadow:var(--tw-ring-offset-shadow,0 0 #0000),var(--tw-ring-shadow,0 0 #0000),var(--tw-shadow)}.toast{gap:.5rem;padding:1rem}.toast>*{animation:toast-pop .25s ease-out}@keyframes toast-pop{0%{transform:scale(.9);opacity:0}100%{transform:scale(1);opacity:1}}.toggle{--tglbg:var(--fallback-b1,oklch(var(--b1)/1));--handleoffset:1.5rem;--handleoffsetcalculator:calc(var(--handleoffset) * -1);--togglehandleborder:0 0;height:1.5rem;width:3rem;cursor:pointer;appearance:none;border-radius:var(--rounded-badge,1.9rem);border-width:1px;border-color:currentColor;background-color:currentColor;color:var(--fallback-bc,oklch(var(--bc)/.5));transition:background,box-shadow var(--animation-input,.2s) ease-out;box-shadow:var(--handleoffsetcalculator) 0 0 2px var(--tglbg) inset,0 0 0 2px var(--tglbg) inset,var(--togglehandleborder)}[dir=rtl] .toggle{--handleoffsetcalculator:calc(var(--handleoffset) * 1)}.toggle:focus-visible{outline-style:solid;outline-width:2px;outline-offset:2px;outline-color:var(--fallback-bc,oklch(var(--bc)/.2))}.toggle:hover{background-color:currentColor}.toggle:checked,.toggle[aria-checked=true]{background-image:none;--handleoffsetcalculator:var(--handleoffset);--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)))}[dir=rtl] .toggle:checked,[dir=rtl] .toggle[aria-checked=true]{--handleoffsetcalculator:calc(var(--handleoffset) * -1)}.toggle:indeterminate{--tw-text-opacity:1;color:var(--fallback-bc,oklch(var(--bc)/var(--tw-text-opacity)));box-shadow:calc(var(--handleoffset)/ 2) 0 0 2px var(--tglbg) inset,calc(var(--handleoffset)/ -2) 0 0 2px var(--tglbg) inset,0 0 0 2px var(--tglbg) inset}[dir=rtl] .toggle:indeterminate{box-shadow:calc(var(--handleoffset)/ 2) 0 0 2px var(--tglbg) inset,calc(var(--handleoffset)/ -2) 0 0 2px var(--tglbg) inset,0 0 0 2px var(--tglbg) inset}.toggle-primary:focus-visible{outline-color:var(--fallback-p,oklch(var(--p)/1))}.toggle-primary:checked,.toggle-primary[aria-checked=true]{border-color:var(--fallback-p,oklch(var(--p)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-p,oklch(var(--p)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-pc,oklch(var(--pc)/var(--tw-text-opacity)))}.toggle-secondary:focus-visible{outline-color:var(--fallback-s,oklch(var(--s)/1))}.toggle-secondary:checked,.toggle-secondary[aria-checked=true]{border-color:var(--fallback-s,oklch(var(--s)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-s,oklch(var(--s)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-sc,oklch(var(--sc)/var(--tw-text-opacity)))}.toggle-accent:focus-visible{outline-color:var(--fallback-a,oklch(var(--a)/1))}.toggle-accent:checked,.toggle-accent[aria-checked=true]{border-color:var(--fallback-a,oklch(var(--a)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-a,oklch(var(--a)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-ac,oklch(var(--ac)/var(--tw-text-opacity)))}.toggle-success:focus-visible{outline-color:var(--fallback-su,oklch(var(--su)/1))}.toggle-success:checked,.toggle-success[aria-checked=true]{border-color:var(--fallback-su,oklch(var(--su)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-su,oklch(var(--su)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-suc,oklch(var(--suc)/var(--tw-text-opacity)))}.toggle-warning:focus-visible{outline-color:var(--fallback-wa,oklch(var(--wa)/1))}.toggle-warning:checked,.toggle-warning[aria-checked=true]{border-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-wa,oklch(var(--wa)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-wac,oklch(var(--wac)/var(--tw-text-opacity)))}.toggle-info:focus-visible{outline-color:var(--fallback-in,oklch(var(--in)/1))}.toggle-info:checked,.toggle-info[aria-checked=true]{border-color:var(--fallback-in,oklch(var(--in)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-in,oklch(var(--in)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-inc,oklch(var(--inc)/var(--tw-text-opacity)))}.toggle-error:focus-visible{outline-color:var(--fallback-er,oklch(var(--er)/1))}.toggle-error:checked,.toggle-error[aria-checked=true]{border-color:var(--fallback-er,oklch(var(--er)/var(--tw-border-opacity)));--tw-border-opacity:0.1;--tw-bg-opacity:1;background-color:var(--fallback-er,oklch(var(--er)/var(--tw-bg-opacity)));--tw-text-opacity:1;color:var(--fallback-erc,oklch(var(--erc)/var(--tw-text-opacity)))}.toggle:disabled{cursor:not-allowed;--tw-border-opacity:1;border-color:var(--fallback-bc,oklch(var(--bc)/var(--tw-border-opacity)));background-color:transparent;opacity:.3;--togglehandleborder:0 0 0 3px var(--fallback-bc,oklch(var(--bc)/1)) inset,var(--handleoffsetcalculator) 0 0 3px var(--fallback-bc,oklch(var(--bc)/1)) inset}.toggle-mark{display:none}:root .prose{--tw-prose-body:var(--fallback-bc,oklch(var(--bc)/0.8));--tw-prose-headings:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-lead:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-links:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-bold:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-counters:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-bullets:var(--fallback-bc,oklch(var(--bc)/0.5));--tw-prose-hr:var(--fallback-bc,oklch(var(--bc)/0.2));--tw-prose-quotes:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-quote-borders:var(--fallback-bc,oklch(var(--bc)/0.2));--tw-prose-captions:var(--fallback-bc,oklch(var(--bc)/0.5));--tw-prose-code:var(--fallback-bc,oklch(var(--bc)/1));--tw-prose-pre-code:var(--fallback-nc,oklch(var(--nc)/1));--tw-prose-pre-bg:var(--fallback-n,oklch(var(--n)/1));--tw-prose-th-borders:var(--fallback-bc,oklch(var(--bc)/0.5));--tw-prose-td-borders:var(--fallback-bc,oklch(var(--bc)/0.2))}.prose :where(code):not(:where([class~=not-prose]*,pre*)){padding:1px 8px;border-radius:var(--rounded-badge);font-weight:initial;background-color:var(--fallback-bc,oklch(var(--bc)/.1))}@supports not (color:oklch(0% 0 0)){.prose :where(code):not(:where([class~=not-prose]*,pre*)){background-color:var(--fallback-b3,oklch(var(--b3)/1))}}.prose :where(code):not(:where([class~=not-prose],[class~=not-prose]*))::after,.prose :where(code):not(:where([class~=not-prose],[class~=not-prose]*))::before{display:none}.prose pre code{border-radius:0;padding:0}.prose :where(tbodytr,thead):not(:where([class~=not-prose]*)){border-bottom-color:var(--fallback-bc,oklch(var(--bc)/.2))}:root{color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:89.824% 0.06192 275.75;--ac:15.352% 0.0368 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:49.12% 0.3096 275.75;--s:69.71% 0.329 342.55;--sc:98.71% 0.0106 342.55;--a:76.76% 0.184 183.61;--n:32.1785% 0.02476 255.701624;--nc:89.4994% 0.011585 252.096176;--b1:100% 0 0;--b2:96.1151% 0 0;--b3:92.4169% 0.00108 197.137559;--bc:27.8078% 0.029596 256.847952}@media (prefers-color-scheme:dark){:root{color-scheme:dark;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:13.138% 0.0392 275.75;--sc:14.96% 0.052 342.55;--ac:14.902% 0.0334 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:65.69% 0.196 275.75;--s:74.8% 0.26 342.55;--a:74.51% 0.167 183.61;--n:31.3815% 0.021108 254.139175;--nc:74.6477% 0.0216 264.435964;--b1:25.3267% 0.015896 252.417568;--b2:23.2607% 0.013807 253.100675;--b3:21.1484% 0.01165 254.087939;--bc:74.6477% 0.0216 264.435964}}[data-theme=light]{color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:89.824% 0.06192 275.75;--ac:15.352% 0.0368 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:49.12% 0.3096 275.75;--s:69.71% 0.329 342.55;--sc:98.71% 0.0106 342.55;--a:76.76% 0.184 183.61;--n:32.1785% 0.02476 255.701624;--nc:89.4994% 0.011585 252.096176;--b1:100% 0 0;--b2:96.1151% 0 0;--b3:92.4169% 0.00108 197.137559;--bc:27.8078% 0.029596 256.847952}:root:has(input.theme-controller[value=light]:checked){color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:89.824% 0.06192 275.75;--ac:15.352% 0.0368 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:49.12% 0.3096 275.75;--s:69.71% 0.329 342.55;--sc:98.71% 0.0106 342.55;--a:76.76% 0.184 183.61;--n:32.1785% 0.02476 255.701624;--nc:89.4994% 0.011585 252.096176;--b1:100% 0 0;--b2:96.1151% 0 0;--b3:92.4169% 0.00108 197.137559;--bc:27.8078% 0.029596 256.847952}[data-theme=dark]{color-scheme:dark;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:13.138% 0.0392 275.75;--sc:14.96% 0.052 342.55;--ac:14.902% 0.0334 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:65.69% 0.196 275.75;--s:74.8% 0.26 342.55;--a:74.51% 0.167 183.61;--n:31.3815% 0.021108 254.139175;--nc:74.6477% 0.0216 264.435964;--b1:25.3267% 0.015896 252.417568;--b2:23.2607% 0.013807 253.100675;--b3:21.1484% 0.01165 254.087939;--bc:74.6477% 0.0216 264.435964}:root:has(input.theme-controller[value=dark]:checked){color-scheme:dark;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:13.138% 0.0392 275.75;--sc:14.96% 0.052 342.55;--ac:14.902% 0.0334 183.61;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:65.69% 0.196 275.75;--s:74.8% 0.26 342.55;--a:74.51% 0.167 183.61;--n:31.3815% 0.021108 254.139175;--nc:74.6477% 0.0216 264.435964;--b1:25.3267% 0.015896 252.417568;--b2:23.2607% 0.013807 253.100675;--b3:21.1484% 0.01165 254.087939;--bc:74.6477% 0.0216 264.435964}[data-theme=cupcake]{color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:15.2344% 0.017892 200.026556;--sc:15.787% 0.020249 356.29965;--ac:15.8762% 0.029206 78.618794;--nc:84.7148% 0.013247 313.189598;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--p:76.172% 0.089459 200.026556;--s:78.9351% 0.101246 356.29965;--a:79.3811% 0.146032 78.618794;--n:23.5742% 0.066235 313.189598;--b1:97.7882% 0.00418 56.375637;--b2:93.9822% 0.007638 61.449292;--b3:91.5861% 0.006811 53.440502;--bc:23.5742% 0.066235 313.189598;--rounded-btn:1.9rem;--tab-border:2px;--tab-radius:0.7rem}:root:has(input.theme-controller[value=cupcake]:checked){color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:15.2344% 0.017892 200.026556;--sc:15.787% 0.020249 356.29965;--ac:15.8762% 0.029206 78.618794;--nc:84.7148% 0.013247 313.189598;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--p:76.172% 0.089459 200.026556;--s:78.9351% 0.101246 356.29965;--a:79.3811% 0.146032 78.618794;--n:23.5742% 0.066235 313.189598;--b1:97.7882% 0.00418 56.375637;--b2:93.9822% 0.007638 61.449292;--b3:91.5861% 0.006811 53.440502;--bc:23.5742% 0.066235 313.189598;--rounded-btn:1.9rem;--tab-border:2px;--tab-radius:0.7rem}[data-theme=bumblebee]{color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:20% 0 0;--ac:16.254% 0.0314 56.52;--nc:82.55% 0.015 281.99;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:89.51% 0.2132 96.61;--pc:38.92% 0.046 96.61;--s:80.39% 0.194 70.76;--sc:39.38% 0.068 70.76;--a:81.27% 0.157 56.52;--n:12.75% 0.075 281.99;--b1:100% 0 0}:root:has(input.theme-controller[value=bumblebee]:checked){color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:20% 0 0;--ac:16.254% 0.0314 56.52;--nc:82.55% 0.015 281.99;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:89.51% 0.2132 96.61;--pc:38.92% 0.046 96.61;--s:80.39% 0.194 70.76;--sc:39.38% 0.068 70.76;--a:81.27% 0.157 56.52;--n:12.75% 0.075 281.99;--b1:100% 0 0}[data-theme=emerald]{color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:76.6626% 0.135433 153.450024;--pc:33.3872% 0.040618 162.240129;--s:61.3028% 0.202368 261.294233;--sc:100% 0 0;--a:72.7725% 0.149783 33.200363;--ac:0% 0 0;--n:35.5192% 0.032071 262.988584;--nc:98.4625% 0.001706 247.838921;--b1:100% 0 0;--bc:35.5192% 0.032071 262.988584;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}:root:has(input.theme-controller[value=emerald]:checked){color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:76.6626% 0.135433 153.450024;--pc:33.3872% 0.040618 162.240129;--s:61.3028% 0.202368 261.294233;--sc:100% 0 0;--a:72.7725% 0.149783 33.200363;--ac:0% 0 0;--n:35.5192% 0.032071 262.988584;--nc:98.4625% 0.001706 247.838921;--b1:100% 0 0;--bc:35.5192% 0.032071 262.988584;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}[data-theme=corporate]{color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:12.078% 0.0456 269.1;--sc:13.0739% 0.010951 256.688055;--ac:15.3934% 0.022799 163.57888;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--border-btn:1px;--tab-border:1px;--p:60.39% 0.228 269.1;--s:65.3694% 0.054756 256.688055;--a:76.9669% 0.113994 163.57888;--n:22.3899% 0.031305 278.07229;--nc:95.8796% 0.008588 247.915135;--b1:100% 0 0;--bc:22.3899% 0.031305 278.07229;--rounded-box:0.25rem;--rounded-btn:.125rem;--rounded-badge:.125rem;--tab-radius:0.25rem;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}:root:has(input.theme-controller[value=corporate]:checked){color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:12.078% 0.0456 269.1;--sc:13.0739% 0.010951 256.688055;--ac:15.3934% 0.022799 163.57888;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--border-btn:1px;--tab-border:1px;--p:60.39% 0.228 269.1;--s:65.3694% 0.054756 256.688055;--a:76.9669% 0.113994 163.57888;--n:22.3899% 0.031305 278.07229;--nc:95.8796% 0.008588 247.915135;--b1:100% 0 0;--bc:22.3899% 0.031305 278.07229;--rounded-box:0.25rem;--rounded-btn:.125rem;--rounded-badge:.125rem;--tab-radius:0.25rem;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}[data-theme=synthwave]{color-scheme:dark;--b2:20.2941% 0.076211 287.835609;--b3:18.7665% 0.070475 287.835609;--pc:14.4421% 0.031903 342.009383;--sc:15.6543% 0.02362 227.382405;--ac:17.608% 0.0412 93.72;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:72.2105% 0.159514 342.009383;--s:78.2714% 0.118101 227.382405;--a:88.04% 0.206 93.72;--n:25.5554% 0.103537 286.507967;--nc:97.9365% 0.00819 301.358346;--b1:21.8216% 0.081948 287.835609;--bc:97.9365% 0.00819 301.358346;--in:76.5197% 0.12273 231.831603;--inc:23.5017% 0.096418 290.329844;--su:86.0572% 0.115038 178.624677;--suc:23.5017% 0.096418 290.329844;--wa:85.531% 0.122117 93.722227;--wac:23.5017% 0.096418 290.329844;--er:73.7005% 0.121339 32.639257;--erc:23.5017% 0.096418 290.329844}:root:has(input.theme-controller[value=synthwave]:checked){color-scheme:dark;--b2:20.2941% 0.076211 287.835609;--b3:18.7665% 0.070475 287.835609;--pc:14.4421% 0.031903 342.009383;--sc:15.6543% 0.02362 227.382405;--ac:17.608% 0.0412 93.72;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:72.2105% 0.159514 342.009383;--s:78.2714% 0.118101 227.382405;--a:88.04% 0.206 93.72;--n:25.5554% 0.103537 286.507967;--nc:97.9365% 0.00819 301.358346;--b1:21.8216% 0.081948 287.835609;--bc:97.9365% 0.00819 301.358346;--in:76.5197% 0.12273 231.831603;--inc:23.5017% 0.096418 290.329844;--su:86.0572% 0.115038 178.624677;--suc:23.5017% 0.096418 290.329844;--wa:85.531% 0.122117 93.722227;--wac:23.5017% 0.096418 290.329844;--er:73.7005% 0.121339 32.639257;--erc:23.5017% 0.096418 290.329844}[data-theme=retro]{color-scheme:light;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:13.144% 0.0398 27.33;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:76.8664% 0.104092 22.664655;--pc:26.5104% 0.006243 0.522862;--s:80.7415% 0.052534 159.094608;--sc:26.5104% 0.006243 0.522862;--a:70.3919% 0.125455 52.953428;--ac:26.5104% 0.006243 0.522862;--n:28.4181% 0.009519 355.534017;--nc:92.5604% 0.025113 89.217311;--b1:91.6374% 0.034554 90.51575;--b2:88.2722% 0.049418 91.774344;--b3:84.133% 0.065952 90.856665;--bc:26.5104% 0.006243 0.522862;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:65.72% 0.199 27.33;--rounded-box:0.4rem;--rounded-btn:0.4rem;--rounded-badge:0.4rem;--tab-radius:0.4rem}:root:has(input.theme-controller[value=retro]:checked){color-scheme:light;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:13.144% 0.0398 27.33;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:76.8664% 0.104092 22.664655;--pc:26.5104% 0.006243 0.522862;--s:80.7415% 0.052534 159.094608;--sc:26.5104% 0.006243 0.522862;--a:70.3919% 0.125455 52.953428;--ac:26.5104% 0.006243 0.522862;--n:28.4181% 0.009519 355.534017;--nc:92.5604% 0.025113 89.217311;--b1:91.6374% 0.034554 90.51575;--b2:88.2722% 0.049418 91.774344;--b3:84.133% 0.065952 90.856665;--bc:26.5104% 0.006243 0.522862;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:65.72% 0.199 27.33;--rounded-box:0.4rem;--rounded-btn:0.4rem;--rounded-badge:0.4rem;--tab-radius:0.4rem}[data-theme=cyberpunk]{color-scheme:light;--b2:87.8943% 0.16647 104.32;--b3:81.2786% 0.15394 104.32;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:18.902% 0.0358 104.32;--pc:14.844% 0.0418 6.35;--sc:16.666% 0.0368 204.72;--ac:14.372% 0.04352 310.43;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;font-family:ui-monospace,SFMono-Regular,Menlo,Monaco,Consolas,Liberation Mono,Courier New,monospace;--p:74.22% 0.209 6.35;--s:83.33% 0.184 204.72;--a:71.86% 0.2176 310.43;--n:23.04% 0.065 269.31;--nc:94.51% 0.179 104.32;--b1:94.51% 0.179 104.32;--rounded-box:0;--rounded-btn:0;--rounded-badge:0;--tab-radius:0}:root:has(input.theme-controller[value=cyberpunk]:checked){color-scheme:light;--b2:87.8943% 0.16647 104.32;--b3:81.2786% 0.15394 104.32;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:18.902% 0.0358 104.32;--pc:14.844% 0.0418 6.35;--sc:16.666% 0.0368 204.72;--ac:14.372% 0.04352 310.43;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;font-family:ui-monospace,SFMono-Regular,Menlo,Monaco,Consolas,Liberation Mono,Courier New,monospace;--p:74.22% 0.209 6.35;--s:83.33% 0.184 204.72;--a:71.86% 0.2176 310.43;--n:23.04% 0.065 269.31;--nc:94.51% 0.179 104.32;--b1:94.51% 0.179 104.32;--rounded-box:0;--rounded-btn:0;--rounded-badge:0;--tab-radius:0}[data-theme=valentine]{color-scheme:light;--b2:88.0567% 0.024834 337.06289;--b3:81.4288% 0.022964 337.06289;--pc:13.7239% 0.030755 15.066527;--sc:14.3942% 0.029258 293.189609;--ac:14.2537% 0.014961 197.828857;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:14.614% 0.0414 27.33;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:68.6197% 0.153774 15.066527;--s:71.971% 0.14629 293.189609;--a:71.2685% 0.074804 197.828857;--n:54.6053% 0.143342 358.004839;--nc:90.2701% 0.037202 336.955191;--b1:94.6846% 0.026703 337.06289;--bc:37.3085% 0.081131 4.606426;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:73.07% 0.207 27.33;--rounded-btn:1.9rem;--tab-radius:0.7rem}:root:has(input.theme-controller[value=valentine]:checked){color-scheme:light;--b2:88.0567% 0.024834 337.06289;--b3:81.4288% 0.022964 337.06289;--pc:13.7239% 0.030755 15.066527;--sc:14.3942% 0.029258 293.189609;--ac:14.2537% 0.014961 197.828857;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:14.614% 0.0414 27.33;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:68.6197% 0.153774 15.066527;--s:71.971% 0.14629 293.189609;--a:71.2685% 0.074804 197.828857;--n:54.6053% 0.143342 358.004839;--nc:90.2701% 0.037202 336.955191;--b1:94.6846% 0.026703 337.06289;--bc:37.3085% 0.081131 4.606426;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:73.07% 0.207 27.33;--rounded-btn:1.9rem;--tab-radius:0.7rem}[data-theme=halloween]{color-scheme:dark;--b2:23.0416% 0 0;--b3:21.3072% 0 0;--bc:84.9552% 0 0;--sc:89.196% 0.0496 305.03;--nc:84.8742% 0.009322 65.681484;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:13.144% 0.0398 27.33;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:77.48% 0.204 60.62;--pc:19.6935% 0.004671 196.779412;--s:45.98% 0.248 305.03;--a:64.8% 0.223 136.073479;--ac:0% 0 0;--n:24.371% 0.046608 65.681484;--b1:24.7759% 0 0;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:65.72% 0.199 27.33}:root:has(input.theme-controller[value=halloween]:checked){color-scheme:dark;--b2:23.0416% 0 0;--b3:21.3072% 0 0;--bc:84.9552% 0 0;--sc:89.196% 0.0496 305.03;--nc:84.8742% 0.009322 65.681484;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:13.144% 0.0398 27.33;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:77.48% 0.204 60.62;--pc:19.6935% 0.004671 196.779412;--s:45.98% 0.248 305.03;--a:64.8% 0.223 136.073479;--ac:0% 0 0;--n:24.371% 0.046608 65.681484;--b1:24.7759% 0 0;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:65.72% 0.199 27.33}[data-theme=garden]{color-scheme:light;--b2:86.4453% 0.002011 17.197414;--b3:79.9386% 0.00186 17.197414;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--sc:89.699% 0.022197 355.095988;--ac:11.2547% 0.010859 154.390187;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:62.45% 0.278 3.83636;--pc:100% 0 0;--s:48.4952% 0.110985 355.095988;--a:56.2735% 0.054297 154.390187;--n:24.1559% 0.049362 89.070594;--nc:92.9519% 0.002163 17.197414;--b1:92.9519% 0.002163 17.197414;--bc:16.9617% 0.001664 17.32068}:root:has(input.theme-controller[value=garden]:checked){color-scheme:light;--b2:86.4453% 0.002011 17.197414;--b3:79.9386% 0.00186 17.197414;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--sc:89.699% 0.022197 355.095988;--ac:11.2547% 0.010859 154.390187;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:62.45% 0.278 3.83636;--pc:100% 0 0;--s:48.4952% 0.110985 355.095988;--a:56.2735% 0.054297 154.390187;--n:24.1559% 0.049362 89.070594;--nc:92.9519% 0.002163 17.197414;--b1:92.9519% 0.002163 17.197414;--bc:16.9617% 0.001664 17.32068}[data-theme=forest]{color-scheme:dark;--b2:17.522% 0.007709 17.911578;--b3:16.2032% 0.007129 17.911578;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:83.7682% 0.001658 17.911578;--sc:13.9553% 0.027077 168.327128;--ac:14.1257% 0.02389 185.713193;--nc:86.1397% 0.007806 171.364646;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:68.6283% 0.185567 148.958922;--pc:0% 0 0;--s:69.7764% 0.135385 168.327128;--a:70.6285% 0.119451 185.713193;--n:30.6985% 0.039032 171.364646;--b1:18.8409% 0.00829 17.911578;--rounded-btn:1.9rem}:root:has(input.theme-controller[value=forest]:checked){color-scheme:dark;--b2:17.522% 0.007709 17.911578;--b3:16.2032% 0.007129 17.911578;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:83.7682% 0.001658 17.911578;--sc:13.9553% 0.027077 168.327128;--ac:14.1257% 0.02389 185.713193;--nc:86.1397% 0.007806 171.364646;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:68.6283% 0.185567 148.958922;--pc:0% 0 0;--s:69.7764% 0.135385 168.327128;--a:70.6285% 0.119451 185.713193;--n:30.6985% 0.039032 171.364646;--b1:18.8409% 0.00829 17.911578;--rounded-btn:1.9rem}[data-theme=aqua]{color-scheme:dark;--b2:45.3464% 0.118611 261.181672;--b3:41.9333% 0.109683 261.181672;--bc:89.7519% 0.025508 261.181672;--sc:12.1365% 0.02175 309.782946;--ac:18.6854% 0.020445 94.555431;--nc:12.2124% 0.023402 243.760661;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:14.79% 0.038 27.33;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:85.6617% 0.14498 198.6458;--pc:40.1249% 0.068266 197.603872;--s:60.6827% 0.108752 309.782946;--a:93.4269% 0.102225 94.555431;--n:61.0622% 0.117009 243.760661;--b1:48.7596% 0.127539 261.181672;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:73.95% 0.19 27.33}:root:has(input.theme-controller[value=aqua]:checked){color-scheme:dark;--b2:45.3464% 0.118611 261.181672;--b3:41.9333% 0.109683 261.181672;--bc:89.7519% 0.025508 261.181672;--sc:12.1365% 0.02175 309.782946;--ac:18.6854% 0.020445 94.555431;--nc:12.2124% 0.023402 243.760661;--inc:90.923% 0.043042 262.880917;--suc:12.541% 0.033982 149.213788;--wac:13.3168% 0.031484 58.31834;--erc:14.79% 0.038 27.33;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:85.6617% 0.14498 198.6458;--pc:40.1249% 0.068266 197.603872;--s:60.6827% 0.108752 309.782946;--a:93.4269% 0.102225 94.555431;--n:61.0622% 0.117009 243.760661;--b1:48.7596% 0.127539 261.181672;--in:54.615% 0.215208 262.880917;--su:62.7052% 0.169912 149.213788;--wa:66.584% 0.157422 58.31834;--er:73.95% 0.19 27.33}[data-theme=lofi]{color-scheme:light;--inc:15.908% 0.0206 205.9;--suc:18.026% 0.0306 164.14;--wac:17.674% 0.027 79.94;--erc:15.732% 0.03 28.47;--border-btn:1px;--tab-border:1px;--p:15.9066% 0 0;--pc:100% 0 0;--s:21.455% 0.001566 17.278957;--sc:100% 0 0;--a:26.8618% 0 0;--ac:100% 0 0;--n:0% 0 0;--nc:100% 0 0;--b1:100% 0 0;--b2:96.1151% 0 0;--b3:92.268% 0.001082 17.17934;--bc:0% 0 0;--in:79.54% 0.103 205.9;--su:90.13% 0.153 164.14;--wa:88.37% 0.135 79.94;--er:78.66% 0.15 28.47;--rounded-box:0.25rem;--rounded-btn:0.125rem;--rounded-badge:0.125rem;--tab-radius:0.125rem;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}:root:has(input.theme-controller[value=lofi]:checked){color-scheme:light;--inc:15.908% 0.0206 205.9;--suc:18.026% 0.0306 164.14;--wac:17.674% 0.027 79.94;--erc:15.732% 0.03 28.47;--border-btn:1px;--tab-border:1px;--p:15.9066% 0 0;--pc:100% 0 0;--s:21.455% 0.001566 17.278957;--sc:100% 0 0;--a:26.8618% 0 0;--ac:100% 0 0;--n:0% 0 0;--nc:100% 0 0;--b1:100% 0 0;--b2:96.1151% 0 0;--b3:92.268% 0.001082 17.17934;--bc:0% 0 0;--in:79.54% 0.103 205.9;--su:90.13% 0.153 164.14;--wa:88.37% 0.135 79.94;--er:78.66% 0.15 28.47;--rounded-box:0.25rem;--rounded-btn:0.125rem;--rounded-badge:0.125rem;--tab-radius:0.125rem;--animation-btn:0;--animation-input:0;--btn-focus-scale:1}[data-theme=pastel]{color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:20% 0 0;--pc:16.6166% 0.006979 316.8737;--sc:17.6153% 0.009839 8.688364;--ac:17.8419% 0.012056 170.923263;--nc:14.2681% 0.014702 228.183906;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:83.0828% 0.034896 316.8737;--s:88.0763% 0.049197 8.688364;--a:89.2096% 0.06028 170.923263;--n:71.3406% 0.07351 228.183906;--b1:100% 0 0;--b2:98.4625% 0.001706 247.838921;--b3:87.1681% 0.009339 258.338227;--rounded-btn:1.9rem;--tab-radius:0.7rem}:root:has(input.theme-controller[value=pastel]:checked){color-scheme:light;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--bc:20% 0 0;--pc:16.6166% 0.006979 316.8737;--sc:17.6153% 0.009839 8.688364;--ac:17.8419% 0.012056 170.923263;--nc:14.2681% 0.014702 228.183906;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:83.0828% 0.034896 316.8737;--s:88.0763% 0.049197 8.688364;--a:89.2096% 0.06028 170.923263;--n:71.3406% 0.07351 228.183906;--b1:100% 0 0;--b2:98.4625% 0.001706 247.838921;--b3:87.1681% 0.009339 258.338227;--rounded-btn:1.9rem;--tab-radius:0.7rem}[data-theme=fantasy]{color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:87.49% 0.0378 325.02;--sc:90.784% 0.0324 241.36;--ac:15.196% 0.0408 56.72;--nc:85.5616% 0.005919 256.847952;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:37.45% 0.189 325.02;--s:53.92% 0.162 241.36;--a:75.98% 0.204 56.72;--n:27.8078% 0.029596 256.847952;--b1:100% 0 0;--bc:27.8078% 0.029596 256.847952}:root:has(input.theme-controller[value=fantasy]:checked){color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--in:72.06% 0.191 231.6;--su:64.8% 0.150 160;--wa:84.71% 0.199 83.87;--er:71.76% 0.221 22.18;--pc:87.49% 0.0378 325.02;--sc:90.784% 0.0324 241.36;--ac:15.196% 0.0408 56.72;--nc:85.5616% 0.005919 256.847952;--inc:0% 0 0;--suc:0% 0 0;--wac:0% 0 0;--erc:0% 0 0;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:37.45% 0.189 325.02;--s:53.92% 0.162 241.36;--a:75.98% 0.204 56.72;--n:27.8078% 0.029596 256.847952;--b1:100% 0 0;--bc:27.8078% 0.029596 256.847952}[data-theme=wireframe]{color-scheme:light;--bc:20% 0 0;--pc:15.6521% 0 0;--sc:15.6521% 0 0;--ac:15.6521% 0 0;--nc:18.8014% 0 0;--inc:89.0403% 0.062643 264.052021;--suc:90.395% 0.035372 142.495339;--wac:14.1626% 0.019994 108.702381;--erc:12.5591% 0.051537 29.233885;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;font-family:Chalkboard,comic sans ms,sans-serif;--p:78.2604% 0 0;--s:78.2604% 0 0;--a:78.2604% 0 0;--n:94.007% 0 0;--b1:100% 0 0;--b2:94.9119% 0 0;--b3:89.7547% 0 0;--in:45.2014% 0.313214 264.052021;--su:51.9752% 0.176858 142.495339;--wa:70.8131% 0.099969 108.702381;--er:62.7955% 0.257683 29.233885;--rounded-box:0.2rem;--rounded-btn:0.2rem;--rounded-badge:0.2rem;--tab-radius:0.2rem}:root:has(input.theme-controller[value=wireframe]:checked){color-scheme:light;--bc:20% 0 0;--pc:15.6521% 0 0;--sc:15.6521% 0 0;--ac:15.6521% 0 0;--nc:18.8014% 0 0;--inc:89.0403% 0.062643 264.052021;--suc:90.395% 0.035372 142.495339;--wac:14.1626% 0.019994 108.702381;--erc:12.5591% 0.051537 29.233885;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;font-family:Chalkboard,comic sans ms,sans-serif;--p:78.2604% 0 0;--s:78.2604% 0 0;--a:78.2604% 0 0;--n:94.007% 0 0;--b1:100% 0 0;--b2:94.9119% 0 0;--b3:89.7547% 0 0;--in:45.2014% 0.313214 264.052021;--su:51.9752% 0.176858 142.495339;--wa:70.8131% 0.099969 108.702381;--er:62.7955% 0.257683 29.233885;--rounded-box:0.2rem;--rounded-btn:0.2rem;--rounded-badge:0.2rem;--tab-radius:0.2rem}[data-theme=black]{color-scheme:dark;--pc:86.736% 0 0;--sc:86.736% 0 0;--ac:86.736% 0 0;--nc:86.736% 0 0;--inc:89.0403% 0.062643 264.052021;--suc:90.395% 0.035372 142.495339;--wac:19.3597% 0.042201 109.769232;--erc:12.5591% 0.051537 29.233885;--border-btn:1px;--tab-border:1px;--p:33.6799% 0 0;--s:33.6799% 0 0;--a:33.6799% 0 0;--b1:0% 0 0;--b2:19.1251% 0 0;--b3:26.8618% 0 0;--bc:87.6096% 0 0;--n:33.6799% 0 0;--in:45.2014% 0.313214 264.052021;--su:51.9752% 0.176858 142.495339;--wa:96.7983% 0.211006 109.769232;--er:62.7955% 0.257683 29.233885;--rounded-box:0;--rounded-btn:0;--rounded-badge:0;--animation-btn:0;--animation-input:0;--btn-focus-scale:1;--tab-radius:0}:root:has(input.theme-controller[value=black]:checked){color-scheme:dark;--pc:86.736% 0 0;--sc:86.736% 0 0;--ac:86.736% 0 0;--nc:86.736% 0 0;--inc:89.0403% 0.062643 264.052021;--suc:90.395% 0.035372 142.495339;--wac:19.3597% 0.042201 109.769232;--erc:12.5591% 0.051537 29.233885;--border-btn:1px;--tab-border:1px;--p:33.6799% 0 0;--s:33.6799% 0 0;--a:33.6799% 0 0;--b1:0% 0 0;--b2:19.1251% 0 0;--b3:26.8618% 0 0;--bc:87.6096% 0 0;--n:33.6799% 0 0;--in:45.2014% 0.313214 264.052021;--su:51.9752% 0.176858 142.495339;--wa:96.7983% 0.211006 109.769232;--er:62.7955% 0.257683 29.233885;--rounded-box:0;--rounded-btn:0;--rounded-badge:0;--animation-btn:0;--animation-input:0;--btn-focus-scale:1;--tab-radius:0}[data-theme=luxury]{color-scheme:dark;--pc:20% 0 0;--sc:85.5163% 0.012821 261.069149;--ac:87.3349% 0.010348 338.82597;--inc:15.8122% 0.024356 237.133883;--suc:15.6239% 0.038579 132.154381;--wac:17.2255% 0.027305 102.89115;--erc:14.3506% 0.035271 22.568916;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:100% 0 0;--s:27.5815% 0.064106 261.069149;--a:36.6744% 0.051741 338.82597;--n:24.27% 0.057015 59.825019;--nc:93.2033% 0.089631 90.861683;--b1:14.0765% 0.004386 285.822869;--b2:20.2191% 0.004211 308.22937;--b3:29.8961% 0.003818 308.318612;--bc:75.6879% 0.123666 76.890484;--in:79.0612% 0.121778 237.133883;--su:78.1197% 0.192894 132.154381;--wa:86.1274% 0.136524 102.89115;--er:71.7531% 0.176357 22.568916}:root:has(input.theme-controller[value=luxury]:checked){color-scheme:dark;--pc:20% 0 0;--sc:85.5163% 0.012821 261.069149;--ac:87.3349% 0.010348 338.82597;--inc:15.8122% 0.024356 237.133883;--suc:15.6239% 0.038579 132.154381;--wac:17.2255% 0.027305 102.89115;--erc:14.3506% 0.035271 22.568916;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:100% 0 0;--s:27.5815% 0.064106 261.069149;--a:36.6744% 0.051741 338.82597;--n:24.27% 0.057015 59.825019;--nc:93.2033% 0.089631 90.861683;--b1:14.0765% 0.004386 285.822869;--b2:20.2191% 0.004211 308.22937;--b3:29.8961% 0.003818 308.318612;--bc:75.6879% 0.123666 76.890484;--in:79.0612% 0.121778 237.133883;--su:78.1197% 0.192894 132.154381;--wa:86.1274% 0.136524 102.89115;--er:71.7531% 0.176357 22.568916}[data-theme=dracula]{color-scheme:dark;--b2:26.8053% 0.020556 277.508664;--b3:24.7877% 0.019009 277.508664;--pc:15.0922% 0.036614 346.812432;--sc:14.8405% 0.029709 301.883095;--ac:16.6785% 0.024826 66.558491;--nc:87.8891% 0.006515 275.524078;--inc:17.6526% 0.018676 212.846491;--suc:17.4199% 0.043903 148.024881;--wac:19.1068% 0.026849 112.757109;--erc:13.6441% 0.041266 24.430965;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:75.4611% 0.18307 346.812432;--s:74.2023% 0.148546 301.883095;--a:83.3927% 0.124132 66.558491;--n:39.4456% 0.032576 275.524078;--b1:28.8229% 0.022103 277.508664;--bc:97.7477% 0.007913 106.545019;--in:88.263% 0.09338 212.846491;--su:87.0995% 0.219516 148.024881;--wa:95.5338% 0.134246 112.757109;--er:68.2204% 0.206328 24.430965}:root:has(input.theme-controller[value=dracula]:checked){color-scheme:dark;--b2:26.8053% 0.020556 277.508664;--b3:24.7877% 0.019009 277.508664;--pc:15.0922% 0.036614 346.812432;--sc:14.8405% 0.029709 301.883095;--ac:16.6785% 0.024826 66.558491;--nc:87.8891% 0.006515 275.524078;--inc:17.6526% 0.018676 212.846491;--suc:17.4199% 0.043903 148.024881;--wac:19.1068% 0.026849 112.757109;--erc:13.6441% 0.041266 24.430965;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:75.4611% 0.18307 346.812432;--s:74.2023% 0.148546 301.883095;--a:83.3927% 0.124132 66.558491;--n:39.4456% 0.032576 275.524078;--b1:28.8229% 0.022103 277.508664;--bc:97.7477% 0.007913 106.545019;--in:88.263% 0.09338 212.846491;--su:87.0995% 0.219516 148.024881;--wa:95.5338% 0.134246 112.757109;--er:68.2204% 0.206328 24.430965}[data-theme=cmyk]{color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--bc:20% 0 0;--pc:14.3544% 0.02666 239.443325;--sc:12.8953% 0.040552 359.339283;--ac:18.8458% 0.037948 105.306968;--nc:84.3557% 0 0;--inc:13.6952% 0.0189 217.284104;--suc:89.3898% 0.032505 321.406278;--wac:14.2473% 0.031969 52.023412;--erc:12.4027% 0.041677 28.717543;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:71.7722% 0.133298 239.443325;--s:64.4766% 0.202758 359.339283;--a:94.2289% 0.189741 105.306968;--n:21.7787% 0 0;--b1:100% 0 0;--in:68.4759% 0.094499 217.284104;--su:46.949% 0.162524 321.406278;--wa:71.2364% 0.159843 52.023412;--er:62.0133% 0.208385 28.717543}:root:has(input.theme-controller[value=cmyk]:checked){color-scheme:light;--b2:93% 0 0;--b3:86% 0 0;--bc:20% 0 0;--pc:14.3544% 0.02666 239.443325;--sc:12.8953% 0.040552 359.339283;--ac:18.8458% 0.037948 105.306968;--nc:84.3557% 0 0;--inc:13.6952% 0.0189 217.284104;--suc:89.3898% 0.032505 321.406278;--wac:14.2473% 0.031969 52.023412;--erc:12.4027% 0.041677 28.717543;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:71.7722% 0.133298 239.443325;--s:64.4766% 0.202758 359.339283;--a:94.2289% 0.189741 105.306968;--n:21.7787% 0 0;--b1:100% 0 0;--in:68.4759% 0.094499 217.284104;--su:46.949% 0.162524 321.406278;--wa:71.2364% 0.159843 52.023412;--er:62.0133% 0.208385 28.717543}[data-theme=autumn]{color-scheme:light;--b2:89.1077% 0 0;--b3:82.4006% 0 0;--bc:19.1629% 0 0;--pc:88.1446% 0.032232 17.530175;--sc:12.3353% 0.033821 23.865865;--ac:14.6851% 0.018999 60.729616;--nc:90.8734% 0.007475 51.902819;--inc:13.8449% 0.019596 207.284192;--suc:12.199% 0.016032 174.616213;--wac:14.0163% 0.032982 56.844303;--erc:90.614% 0.0482 24.16;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:40.7232% 0.16116 17.530175;--s:61.6763% 0.169105 23.865865;--a:73.4253% 0.094994 60.729616;--n:54.3672% 0.037374 51.902819;--b1:95.8147% 0 0;--in:69.2245% 0.097979 207.284192;--su:60.9951% 0.080159 174.616213;--wa:70.0817% 0.164909 56.844303;--er:53.07% 0.241 24.16}:root:has(input.theme-controller[value=autumn]:checked){color-scheme:light;--b2:89.1077% 0 0;--b3:82.4006% 0 0;--bc:19.1629% 0 0;--pc:88.1446% 0.032232 17.530175;--sc:12.3353% 0.033821 23.865865;--ac:14.6851% 0.018999 60.729616;--nc:90.8734% 0.007475 51.902819;--inc:13.8449% 0.019596 207.284192;--suc:12.199% 0.016032 174.616213;--wac:14.0163% 0.032982 56.844303;--erc:90.614% 0.0482 24.16;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:40.7232% 0.16116 17.530175;--s:61.6763% 0.169105 23.865865;--a:73.4253% 0.094994 60.729616;--n:54.3672% 0.037374 51.902819;--b1:95.8147% 0 0;--in:69.2245% 0.097979 207.284192;--su:60.9951% 0.080159 174.616213;--wa:70.0817% 0.164909 56.844303;--er:53.07% 0.241 24.16}[data-theme=business]{color-scheme:dark;--b2:22.6487% 0 0;--b3:20.944% 0 0;--bc:84.8707% 0 0;--pc:88.3407% 0.019811 251.473931;--sc:12.8185% 0.005481 229.389418;--ac:13.4542% 0.033545 35.791525;--nc:85.4882% 0.00265 253.041249;--inc:12.5233% 0.028702 240.033697;--suc:14.0454% 0.018919 156.59611;--wac:15.4965% 0.023141 81.519177;--erc:90.3221% 0.029356 29.674507;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:41.7036% 0.099057 251.473931;--s:64.0924% 0.027405 229.389418;--a:67.271% 0.167726 35.791525;--n:27.441% 0.01325 253.041249;--b1:24.3535% 0 0;--in:62.6163% 0.143511 240.033697;--su:70.2268% 0.094594 156.59611;--wa:77.4824% 0.115704 81.519177;--er:51.6105% 0.14678 29.674507;--rounded-box:0.25rem;--rounded-btn:.125rem;--rounded-badge:.125rem}:root:has(input.theme-controller[value=business]:checked){color-scheme:dark;--b2:22.6487% 0 0;--b3:20.944% 0 0;--bc:84.8707% 0 0;--pc:88.3407% 0.019811 251.473931;--sc:12.8185% 0.005481 229.389418;--ac:13.4542% 0.033545 35.791525;--nc:85.4882% 0.00265 253.041249;--inc:12.5233% 0.028702 240.033697;--suc:14.0454% 0.018919 156.59611;--wac:15.4965% 0.023141 81.519177;--erc:90.3221% 0.029356 29.674507;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:41.7036% 0.099057 251.473931;--s:64.0924% 0.027405 229.389418;--a:67.271% 0.167726 35.791525;--n:27.441% 0.01325 253.041249;--b1:24.3535% 0 0;--in:62.6163% 0.143511 240.033697;--su:70.2268% 0.094594 156.59611;--wa:77.4824% 0.115704 81.519177;--er:51.6105% 0.14678 29.674507;--rounded-box:0.25rem;--rounded-btn:.125rem;--rounded-badge:.125rem}[data-theme=acid]{color-scheme:light;--b2:91.6146% 0 0;--b3:84.7189% 0 0;--bc:19.7021% 0 0;--pc:14.38% 0.0714 330.759573;--sc:14.674% 0.0448 48.250878;--ac:18.556% 0.0528 122.962951;--nc:84.262% 0.0256 278.68;--inc:12.144% 0.0454 252.05;--suc:17.144% 0.0532 158.53;--wac:18.202% 0.0424 100.5;--erc:12.968% 0.0586 29.349188;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:71.9% 0.357 330.759573;--s:73.37% 0.224 48.250878;--a:92.78% 0.264 122.962951;--n:21.31% 0.128 278.68;--b1:98.5104% 0 0;--in:60.72% 0.227 252.05;--su:85.72% 0.266 158.53;--wa:91.01% 0.212 100.5;--er:64.84% 0.293 29.349188;--rounded-box:1.25rem;--rounded-btn:1rem;--rounded-badge:1rem;--tab-radius:0.7rem}:root:has(input.theme-controller[value=acid]:checked){color-scheme:light;--b2:91.6146% 0 0;--b3:84.7189% 0 0;--bc:19.7021% 0 0;--pc:14.38% 0.0714 330.759573;--sc:14.674% 0.0448 48.250878;--ac:18.556% 0.0528 122.962951;--nc:84.262% 0.0256 278.68;--inc:12.144% 0.0454 252.05;--suc:17.144% 0.0532 158.53;--wac:18.202% 0.0424 100.5;--erc:12.968% 0.0586 29.349188;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:71.9% 0.357 330.759573;--s:73.37% 0.224 48.250878;--a:92.78% 0.264 122.962951;--n:21.31% 0.128 278.68;--b1:98.5104% 0 0;--in:60.72% 0.227 252.05;--su:85.72% 0.266 158.53;--wa:91.01% 0.212 100.5;--er:64.84% 0.293 29.349188;--rounded-box:1.25rem;--rounded-btn:1rem;--rounded-badge:1rem;--tab-radius:0.7rem}[data-theme=lemonade]{color-scheme:light;--b2:91.8003% 0.0186 123.72;--b3:84.8906% 0.0172 123.72;--bc:19.742% 0.004 123.72;--pc:11.784% 0.0398 134.6;--sc:15.55% 0.0392 111.09;--ac:17.078% 0.0402 100.73;--nc:86.196% 0.015 108.6;--inc:17.238% 0.0094 224.14;--suc:17.238% 0.0094 157.85;--wac:17.238% 0.0094 102.15;--erc:17.238% 0.0094 25.85;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:58.92% 0.199 134.6;--s:77.75% 0.196 111.09;--a:85.39% 0.201 100.73;--n:30.98% 0.075 108.6;--b1:98.71% 0.02 123.72;--in:86.19% 0.047 224.14;--su:86.19% 0.047 157.85;--wa:86.19% 0.047 102.15;--er:86.19% 0.047 25.85}:root:has(input.theme-controller[value=lemonade]:checked){color-scheme:light;--b2:91.8003% 0.0186 123.72;--b3:84.8906% 0.0172 123.72;--bc:19.742% 0.004 123.72;--pc:11.784% 0.0398 134.6;--sc:15.55% 0.0392 111.09;--ac:17.078% 0.0402 100.73;--nc:86.196% 0.015 108.6;--inc:17.238% 0.0094 224.14;--suc:17.238% 0.0094 157.85;--wac:17.238% 0.0094 102.15;--erc:17.238% 0.0094 25.85;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:58.92% 0.199 134.6;--s:77.75% 0.196 111.09;--a:85.39% 0.201 100.73;--n:30.98% 0.075 108.6;--b1:98.71% 0.02 123.72;--in:86.19% 0.047 224.14;--su:86.19% 0.047 157.85;--wa:86.19% 0.047 102.15;--er:86.19% 0.047 25.85}[data-theme=night]{color-scheme:dark;--b2:19.3144% 0.037037 265.754874;--b3:17.8606% 0.034249 265.754874;--bc:84.1536% 0.007965 265.754874;--pc:15.0703% 0.027798 232.66148;--sc:13.6023% 0.031661 276.934902;--ac:14.4721% 0.035244 350.048739;--nc:85.5899% 0.00737 260.030984;--suc:15.6904% 0.026506 181.911977;--wac:16.6486% 0.027912 82.95003;--erc:14.3572% 0.034051 13.11834;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:75.3513% 0.138989 232.66148;--s:68.0113% 0.158303 276.934902;--a:72.3603% 0.176218 350.048739;--n:27.9495% 0.036848 260.030984;--b1:20.7682% 0.039824 265.754874;--in:68.4553% 0.148062 237.25135;--inc:0% 0 0;--su:78.452% 0.132529 181.911977;--wa:83.2428% 0.139558 82.95003;--er:71.7858% 0.170255 13.11834}:root:has(input.theme-controller[value=night]:checked){color-scheme:dark;--b2:19.3144% 0.037037 265.754874;--b3:17.8606% 0.034249 265.754874;--bc:84.1536% 0.007965 265.754874;--pc:15.0703% 0.027798 232.66148;--sc:13.6023% 0.031661 276.934902;--ac:14.4721% 0.035244 350.048739;--nc:85.5899% 0.00737 260.030984;--suc:15.6904% 0.026506 181.911977;--wac:16.6486% 0.027912 82.95003;--erc:14.3572% 0.034051 13.11834;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:75.3513% 0.138989 232.66148;--s:68.0113% 0.158303 276.934902;--a:72.3603% 0.176218 350.048739;--n:27.9495% 0.036848 260.030984;--b1:20.7682% 0.039824 265.754874;--in:68.4553% 0.148062 237.25135;--inc:0% 0 0;--su:78.452% 0.132529 181.911977;--wa:83.2428% 0.139558 82.95003;--er:71.7858% 0.170255 13.11834}[data-theme=coffee]{color-scheme:dark;--b2:20.1585% 0.021457 329.708637;--b3:18.6412% 0.019842 329.708637;--pc:14.3993% 0.024765 62.756393;--sc:86.893% 0.00597 199.19444;--ac:88.5243% 0.014881 224.389184;--nc:83.3022% 0.003149 326.261446;--inc:15.898% 0.012774 184.558367;--suc:14.9445% 0.014491 131.116276;--wac:17.6301% 0.028162 87.722413;--erc:15.4637% 0.025644 31.871922;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:71.9967% 0.123825 62.756393;--s:34.465% 0.029849 199.19444;--a:42.6213% 0.074405 224.389184;--n:16.5109% 0.015743 326.261446;--b1:21.6758% 0.023072 329.708637;--bc:72.3547% 0.092794 79.129387;--in:79.4902% 0.063869 184.558367;--su:74.7224% 0.072456 131.116276;--wa:88.1503% 0.140812 87.722413;--er:77.3187% 0.12822 31.871922}:root:has(input.theme-controller[value=coffee]:checked){color-scheme:dark;--b2:20.1585% 0.021457 329.708637;--b3:18.6412% 0.019842 329.708637;--pc:14.3993% 0.024765 62.756393;--sc:86.893% 0.00597 199.19444;--ac:88.5243% 0.014881 224.389184;--nc:83.3022% 0.003149 326.261446;--inc:15.898% 0.012774 184.558367;--suc:14.9445% 0.014491 131.116276;--wac:17.6301% 0.028162 87.722413;--erc:15.4637% 0.025644 31.871922;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:71.9967% 0.123825 62.756393;--s:34.465% 0.029849 199.19444;--a:42.6213% 0.074405 224.389184;--n:16.5109% 0.015743 326.261446;--b1:21.6758% 0.023072 329.708637;--bc:72.3547% 0.092794 79.129387;--in:79.4902% 0.063869 184.558367;--su:74.7224% 0.072456 131.116276;--wa:88.1503% 0.140812 87.722413;--er:77.3187% 0.12822 31.871922}[data-theme=winter]{color-scheme:light;--pc:91.372% 0.051 257.57;--sc:88.5103% 0.03222 282.339433;--ac:11.988% 0.038303 335.171434;--nc:83.9233% 0.012704 257.651965;--inc:17.6255% 0.017178 214.515264;--suc:16.0988% 0.015404 197.823719;--wac:17.8345% 0.009167 71.47031;--erc:14.6185% 0.022037 20.076293;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:56.86% 0.255 257.57;--s:42.5516% 0.161098 282.339433;--a:59.9398% 0.191515 335.171434;--n:19.6166% 0.063518 257.651965;--b1:100% 0 0;--b2:97.4663% 0.011947 259.822565;--b3:93.2686% 0.016223 262.751375;--bc:41.8869% 0.053885 255.824911;--in:88.1275% 0.085888 214.515264;--su:80.4941% 0.077019 197.823719;--wa:89.1725% 0.045833 71.47031;--er:73.0926% 0.110185 20.076293}:root:has(input.theme-controller[value=winter]:checked){color-scheme:light;--pc:91.372% 0.051 257.57;--sc:88.5103% 0.03222 282.339433;--ac:11.988% 0.038303 335.171434;--nc:83.9233% 0.012704 257.651965;--inc:17.6255% 0.017178 214.515264;--suc:16.0988% 0.015404 197.823719;--wac:17.8345% 0.009167 71.47031;--erc:14.6185% 0.022037 20.076293;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:56.86% 0.255 257.57;--s:42.5516% 0.161098 282.339433;--a:59.9398% 0.191515 335.171434;--n:19.6166% 0.063518 257.651965;--b1:100% 0 0;--b2:97.4663% 0.011947 259.822565;--b3:93.2686% 0.016223 262.751375;--bc:41.8869% 0.053885 255.824911;--in:88.1275% 0.085888 214.515264;--su:80.4941% 0.077019 197.823719;--wa:89.1725% 0.045833 71.47031;--er:73.0926% 0.110185 20.076293}[data-theme=dim]{color-scheme:dark;--pc:17.2267% 0.028331 139.549991;--sc:14.6752% 0.033181 35.353059;--ac:14.8459% 0.026728 311.37924;--inc:17.2157% 0.028409 206.182959;--suc:17.2343% 0.028437 166.534048;--wac:17.2327% 0.028447 94.818679;--erc:16.4838% 0.019914 33.756357;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:86.1335% 0.141656 139.549991;--s:73.3759% 0.165904 35.353059;--a:74.2296% 0.133641 311.37924;--n:24.7311% 0.020483 264.094728;--nc:82.9011% 0.031335 222.959324;--b1:30.8577% 0.023243 264.149498;--b2:28.0368% 0.01983 264.182074;--b3:26.3469% 0.018403 262.177739;--bc:82.9011% 0.031335 222.959324;--in:86.0785% 0.142046 206.182959;--su:86.1717% 0.142187 166.534048;--wa:86.1634% 0.142236 94.818679;--er:82.4189% 0.09957 33.756357}:root:has(input.theme-controller[value=dim]:checked){color-scheme:dark;--pc:17.2267% 0.028331 139.549991;--sc:14.6752% 0.033181 35.353059;--ac:14.8459% 0.026728 311.37924;--inc:17.2157% 0.028409 206.182959;--suc:17.2343% 0.028437 166.534048;--wac:17.2327% 0.028447 94.818679;--erc:16.4838% 0.019914 33.756357;--rounded-box:1rem;--rounded-btn:0.5rem;--rounded-badge:1.9rem;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--tab-radius:0.5rem;--p:86.1335% 0.141656 139.549991;--s:73.3759% 0.165904 35.353059;--a:74.2296% 0.133641 311.37924;--n:24.7311% 0.020483 264.094728;--nc:82.9011% 0.031335 222.959324;--b1:30.8577% 0.023243 264.149498;--b2:28.0368% 0.01983 264.182074;--b3:26.3469% 0.018403 262.177739;--bc:82.9011% 0.031335 222.959324;--in:86.0785% 0.142046 206.182959;--su:86.1717% 0.142187 166.534048;--wa:86.1634% 0.142236 94.818679;--er:82.4189% 0.09957 33.756357}[data-theme=nord]{color-scheme:light;--pc:11.8872% 0.015449 254.027774;--sc:13.9303% 0.011822 248.687186;--ac:15.4929% 0.01245 217.469017;--inc:13.8414% 0.012499 332.664922;--suc:15.3654% 0.01498 131.063061;--wac:17.0972% 0.017847 84.093335;--erc:12.122% 0.024119 15.341883;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:59.4359% 0.077246 254.027774;--s:69.6516% 0.059108 248.687186;--a:77.4643% 0.062249 217.469017;--n:45.229% 0.035214 264.1312;--nc:89.9258% 0.016374 262.749256;--b1:95.1276% 0.007445 260.731539;--b2:93.2996% 0.010389 261.788485;--b3:89.9258% 0.016374 262.749256;--bc:32.4374% 0.022945 264.182036;--in:69.2072% 0.062496 332.664922;--su:76.827% 0.074899 131.063061;--wa:85.4862% 0.089234 84.093335;--er:60.61% 0.120594 15.341883;--rounded-box:0.4rem;--rounded-btn:0.2rem;--rounded-badge:0.4rem;--tab-radius:0.2rem}:root:has(input.theme-controller[value=nord]:checked){color-scheme:light;--pc:11.8872% 0.015449 254.027774;--sc:13.9303% 0.011822 248.687186;--ac:15.4929% 0.01245 217.469017;--inc:13.8414% 0.012499 332.664922;--suc:15.3654% 0.01498 131.063061;--wac:17.0972% 0.017847 84.093335;--erc:12.122% 0.024119 15.341883;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:59.4359% 0.077246 254.027774;--s:69.6516% 0.059108 248.687186;--a:77.4643% 0.062249 217.469017;--n:45.229% 0.035214 264.1312;--nc:89.9258% 0.016374 262.749256;--b1:95.1276% 0.007445 260.731539;--b2:93.2996% 0.010389 261.788485;--b3:89.9258% 0.016374 262.749256;--bc:32.4374% 0.022945 264.182036;--in:69.2072% 0.062496 332.664922;--su:76.827% 0.074899 131.063061;--wa:85.4862% 0.089234 84.093335;--er:60.61% 0.120594 15.341883;--rounded-box:0.4rem;--rounded-btn:0.2rem;--rounded-badge:0.4rem;--tab-radius:0.2rem}[data-theme=sunset]{color-scheme:dark;--pc:14.9408% 0.031656 39.94703;--sc:14.5075% 0.035531 2.72034;--ac:14.2589% 0.033336 299.844533;--inc:17.1119% 0.017054 206.015183;--suc:17.1122% 0.017172 144.77874;--wac:17.1139% 0.016961 74.427797;--erc:17.1023% 0.015778 16.886379;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:74.7039% 0.158278 39.94703;--s:72.5375% 0.177654 2.72034;--a:71.2947% 0.166678 299.844533;--n:26% 0.019 237.69;--nc:70% 0.019 237.69;--b1:22% 0.019 237.69;--b2:20% 0.019 237.69;--b3:18% 0.019 237.69;--bc:77.3835% 0.043586 245.096534;--in:85.5596% 0.085271 206.015183;--su:85.5609% 0.08586 144.77874;--wa:85.5695% 0.084806 74.427797;--er:85.5116% 0.07889 16.886379;--rounded-box:1.2rem;--rounded-btn:0.8rem;--rounded-badge:0.4rem;--tab-radius:0.7rem}:root:has(input.theme-controller[value=sunset]:checked){color-scheme:dark;--pc:14.9408% 0.031656 39.94703;--sc:14.5075% 0.035531 2.72034;--ac:14.2589% 0.033336 299.844533;--inc:17.1119% 0.017054 206.015183;--suc:17.1122% 0.017172 144.77874;--wac:17.1139% 0.016961 74.427797;--erc:17.1023% 0.015778 16.886379;--animation-btn:0.25s;--animation-input:.2s;--btn-focus-scale:0.95;--border-btn:1px;--tab-border:1px;--p:74.7039% 0.158278 39.94703;--s:72.5375% 0.177654 2.72034;--a:71.2947% 0.166678 299.844533;--n:26% 0.019 237.69;--nc:70% 0.019 237.69;--b1:22% 0.019 237.69;--b2:20% 0.019 237.69;--b3:18% 0.019 237.69;--bc:77.3835% 0.043586 245.096534;--in:85.5596% 0.085271 206.015183;--su:85.5609% 0.08586 144.77874;--wa:85.5695% 0.084806 74.427797;--er:85.5116% 0.07889 16.886379;--rounded-box:1.2rem;--rounded-btn:0.8rem;--rounded-badge:0.4rem;--tab-radius:0.7rem} diff --git a/examples/server/public/deps_markdown-it.js b/examples/server/public/deps_markdown-it.js deleted file mode 100644 index 1be0cebe6..000000000 --- a/examples/server/public/deps_markdown-it.js +++ /dev/null @@ -1,8442 +0,0 @@ -/*! markdown-it 13.0.2 https://github.com/markdown-it/markdown-it @license MIT */ -(function(global, factory) { - typeof exports === "object" && typeof module !== "undefined" ? module.exports = factory() : typeof define === "function" && define.amd ? define(factory) : (global = typeof globalThis !== "undefined" ? globalThis : global || self, - global.markdownit = factory()); -})(this, (function() { - "use strict"; - function createCommonjsModule(fn, basedir, module) { - return module = { - path: basedir, - exports: {}, - require: function(path, base) { - return commonjsRequire(path, base === undefined || base === null ? module.path : base); - } - }, fn(module, module.exports), module.exports; - } - function getAugmentedNamespace(n) { - if (n.__esModule) return n; - var a = Object.defineProperty({}, "__esModule", { - value: true - }); - Object.keys(n).forEach((function(k) { - var d = Object.getOwnPropertyDescriptor(n, k); - Object.defineProperty(a, k, d.get ? d : { - enumerable: true, - get: function() { - return n[k]; - } - }); - })); - return a; - } - function commonjsRequire() { - throw new Error("Dynamic requires are not currently supported by @rollup/plugin-commonjs"); - } - var require$$0 = { - Aacute: "\xc1", - aacute: "\xe1", - Abreve: "\u0102", - abreve: "\u0103", - ac: "\u223e", - acd: "\u223f", - acE: "\u223e\u0333", - Acirc: "\xc2", - acirc: "\xe2", - acute: "\xb4", - Acy: "\u0410", - acy: "\u0430", - AElig: "\xc6", - aelig: "\xe6", - af: "\u2061", - Afr: "\ud835\udd04", - afr: "\ud835\udd1e", - Agrave: "\xc0", - agrave: "\xe0", - alefsym: "\u2135", - aleph: "\u2135", - Alpha: "\u0391", - alpha: "\u03b1", - Amacr: "\u0100", - amacr: "\u0101", - amalg: "\u2a3f", - amp: "&", - AMP: "&", - andand: "\u2a55", - And: "\u2a53", - and: "\u2227", - andd: "\u2a5c", - andslope: "\u2a58", - andv: "\u2a5a", - ang: "\u2220", - ange: "\u29a4", - angle: "\u2220", - angmsdaa: "\u29a8", - angmsdab: "\u29a9", - angmsdac: "\u29aa", - angmsdad: "\u29ab", - angmsdae: "\u29ac", - angmsdaf: "\u29ad", - angmsdag: "\u29ae", - angmsdah: "\u29af", - angmsd: "\u2221", - angrt: "\u221f", - angrtvb: "\u22be", - angrtvbd: "\u299d", - angsph: "\u2222", - angst: "\xc5", - angzarr: "\u237c", - Aogon: "\u0104", - aogon: "\u0105", - Aopf: "\ud835\udd38", - aopf: "\ud835\udd52", - apacir: "\u2a6f", - ap: "\u2248", - apE: "\u2a70", - ape: "\u224a", - apid: "\u224b", - apos: "'", - ApplyFunction: "\u2061", - approx: "\u2248", - approxeq: "\u224a", - Aring: "\xc5", - aring: "\xe5", - Ascr: "\ud835\udc9c", - ascr: "\ud835\udcb6", - Assign: "\u2254", - ast: "*", - asymp: "\u2248", - asympeq: "\u224d", - Atilde: "\xc3", - atilde: "\xe3", - Auml: "\xc4", - auml: "\xe4", - awconint: "\u2233", - awint: "\u2a11", - backcong: "\u224c", - backepsilon: "\u03f6", - backprime: "\u2035", - backsim: "\u223d", - backsimeq: "\u22cd", - Backslash: "\u2216", - Barv: "\u2ae7", - barvee: "\u22bd", - barwed: "\u2305", - Barwed: "\u2306", - barwedge: "\u2305", - bbrk: "\u23b5", - bbrktbrk: "\u23b6", - bcong: "\u224c", - Bcy: "\u0411", - bcy: "\u0431", - bdquo: "\u201e", - becaus: "\u2235", - because: "\u2235", - Because: "\u2235", - bemptyv: "\u29b0", - bepsi: "\u03f6", - bernou: "\u212c", - Bernoullis: "\u212c", - Beta: "\u0392", - beta: "\u03b2", - beth: "\u2136", - between: "\u226c", - Bfr: "\ud835\udd05", - bfr: "\ud835\udd1f", - bigcap: "\u22c2", - bigcirc: "\u25ef", - bigcup: "\u22c3", - bigodot: "\u2a00", - bigoplus: "\u2a01", - bigotimes: "\u2a02", - bigsqcup: "\u2a06", - bigstar: "\u2605", - bigtriangledown: "\u25bd", - bigtriangleup: "\u25b3", - biguplus: "\u2a04", - bigvee: "\u22c1", - bigwedge: "\u22c0", - bkarow: "\u290d", - blacklozenge: "\u29eb", - blacksquare: "\u25aa", - blacktriangle: "\u25b4", - blacktriangledown: "\u25be", - blacktriangleleft: "\u25c2", - blacktriangleright: "\u25b8", - blank: "\u2423", - blk12: "\u2592", - blk14: "\u2591", - blk34: "\u2593", - block: "\u2588", - bne: "=\u20e5", - bnequiv: "\u2261\u20e5", - bNot: "\u2aed", - bnot: "\u2310", - Bopf: "\ud835\udd39", - bopf: "\ud835\udd53", - bot: "\u22a5", - bottom: "\u22a5", - bowtie: "\u22c8", - boxbox: "\u29c9", - boxdl: "\u2510", - boxdL: "\u2555", - boxDl: "\u2556", - boxDL: "\u2557", - boxdr: "\u250c", - boxdR: "\u2552", - boxDr: "\u2553", - boxDR: "\u2554", - boxh: "\u2500", - boxH: "\u2550", - boxhd: "\u252c", - boxHd: "\u2564", - boxhD: "\u2565", - boxHD: "\u2566", - boxhu: "\u2534", - boxHu: "\u2567", - boxhU: "\u2568", - boxHU: "\u2569", - boxminus: "\u229f", - boxplus: "\u229e", - boxtimes: "\u22a0", - boxul: "\u2518", - boxuL: "\u255b", - boxUl: "\u255c", - boxUL: "\u255d", - boxur: "\u2514", - boxuR: "\u2558", - boxUr: "\u2559", - boxUR: "\u255a", - boxv: "\u2502", - boxV: "\u2551", - boxvh: "\u253c", - boxvH: "\u256a", - boxVh: "\u256b", - boxVH: "\u256c", - boxvl: "\u2524", - boxvL: "\u2561", - boxVl: "\u2562", - boxVL: "\u2563", - boxvr: "\u251c", - boxvR: "\u255e", - boxVr: "\u255f", - boxVR: "\u2560", - bprime: "\u2035", - breve: "\u02d8", - Breve: "\u02d8", - brvbar: "\xa6", - bscr: "\ud835\udcb7", - Bscr: "\u212c", - bsemi: "\u204f", - bsim: "\u223d", - bsime: "\u22cd", - bsolb: "\u29c5", - bsol: "\\", - bsolhsub: "\u27c8", - bull: "\u2022", - bullet: "\u2022", - bump: "\u224e", - bumpE: "\u2aae", - bumpe: "\u224f", - Bumpeq: "\u224e", - bumpeq: "\u224f", - Cacute: "\u0106", - cacute: "\u0107", - capand: "\u2a44", - capbrcup: "\u2a49", - capcap: "\u2a4b", - cap: "\u2229", - Cap: "\u22d2", - capcup: "\u2a47", - capdot: "\u2a40", - CapitalDifferentialD: "\u2145", - caps: "\u2229\ufe00", - caret: "\u2041", - caron: "\u02c7", - Cayleys: "\u212d", - ccaps: "\u2a4d", - Ccaron: "\u010c", - ccaron: "\u010d", - Ccedil: "\xc7", - ccedil: "\xe7", - Ccirc: "\u0108", - ccirc: "\u0109", - Cconint: "\u2230", - ccups: "\u2a4c", - ccupssm: "\u2a50", - Cdot: "\u010a", - cdot: "\u010b", - cedil: "\xb8", - Cedilla: "\xb8", - cemptyv: "\u29b2", - cent: "\xa2", - centerdot: "\xb7", - CenterDot: "\xb7", - cfr: "\ud835\udd20", - Cfr: "\u212d", - CHcy: "\u0427", - chcy: "\u0447", - check: "\u2713", - checkmark: "\u2713", - Chi: "\u03a7", - chi: "\u03c7", - circ: "\u02c6", - circeq: "\u2257", - circlearrowleft: "\u21ba", - circlearrowright: "\u21bb", - circledast: "\u229b", - circledcirc: "\u229a", - circleddash: "\u229d", - CircleDot: "\u2299", - circledR: "\xae", - circledS: "\u24c8", - CircleMinus: "\u2296", - CirclePlus: "\u2295", - CircleTimes: "\u2297", - cir: "\u25cb", - cirE: "\u29c3", - cire: "\u2257", - cirfnint: "\u2a10", - cirmid: "\u2aef", - cirscir: "\u29c2", - ClockwiseContourIntegral: "\u2232", - CloseCurlyDoubleQuote: "\u201d", - CloseCurlyQuote: "\u2019", - clubs: "\u2663", - clubsuit: "\u2663", - colon: ":", - Colon: "\u2237", - Colone: "\u2a74", - colone: "\u2254", - coloneq: "\u2254", - comma: ",", - commat: "@", - comp: "\u2201", - compfn: "\u2218", - complement: "\u2201", - complexes: "\u2102", - cong: "\u2245", - congdot: "\u2a6d", - Congruent: "\u2261", - conint: "\u222e", - Conint: "\u222f", - ContourIntegral: "\u222e", - copf: "\ud835\udd54", - Copf: "\u2102", - coprod: "\u2210", - Coproduct: "\u2210", - copy: "\xa9", - COPY: "\xa9", - copysr: "\u2117", - CounterClockwiseContourIntegral: "\u2233", - crarr: "\u21b5", - cross: "\u2717", - Cross: "\u2a2f", - Cscr: "\ud835\udc9e", - cscr: "\ud835\udcb8", - csub: "\u2acf", - csube: "\u2ad1", - csup: "\u2ad0", - csupe: "\u2ad2", - ctdot: "\u22ef", - cudarrl: "\u2938", - cudarrr: "\u2935", - cuepr: "\u22de", - cuesc: "\u22df", - cularr: "\u21b6", - cularrp: "\u293d", - cupbrcap: "\u2a48", - cupcap: "\u2a46", - CupCap: "\u224d", - cup: "\u222a", - Cup: "\u22d3", - cupcup: "\u2a4a", - cupdot: "\u228d", - cupor: "\u2a45", - cups: "\u222a\ufe00", - curarr: "\u21b7", - curarrm: "\u293c", - curlyeqprec: "\u22de", - curlyeqsucc: "\u22df", - curlyvee: "\u22ce", - curlywedge: "\u22cf", - curren: "\xa4", - curvearrowleft: "\u21b6", - curvearrowright: "\u21b7", - cuvee: "\u22ce", - cuwed: "\u22cf", - cwconint: "\u2232", - cwint: "\u2231", - cylcty: "\u232d", - dagger: "\u2020", - Dagger: "\u2021", - daleth: "\u2138", - darr: "\u2193", - Darr: "\u21a1", - dArr: "\u21d3", - dash: "\u2010", - Dashv: "\u2ae4", - dashv: "\u22a3", - dbkarow: "\u290f", - dblac: "\u02dd", - Dcaron: "\u010e", - dcaron: "\u010f", - Dcy: "\u0414", - dcy: "\u0434", - ddagger: "\u2021", - ddarr: "\u21ca", - DD: "\u2145", - dd: "\u2146", - DDotrahd: "\u2911", - ddotseq: "\u2a77", - deg: "\xb0", - Del: "\u2207", - Delta: "\u0394", - delta: "\u03b4", - demptyv: "\u29b1", - dfisht: "\u297f", - Dfr: "\ud835\udd07", - dfr: "\ud835\udd21", - dHar: "\u2965", - dharl: "\u21c3", - dharr: "\u21c2", - DiacriticalAcute: "\xb4", - DiacriticalDot: "\u02d9", - DiacriticalDoubleAcute: "\u02dd", - DiacriticalGrave: "`", - DiacriticalTilde: "\u02dc", - diam: "\u22c4", - diamond: "\u22c4", - Diamond: "\u22c4", - diamondsuit: "\u2666", - diams: "\u2666", - die: "\xa8", - DifferentialD: "\u2146", - digamma: "\u03dd", - disin: "\u22f2", - div: "\xf7", - divide: "\xf7", - divideontimes: "\u22c7", - divonx: "\u22c7", - DJcy: "\u0402", - djcy: "\u0452", - dlcorn: "\u231e", - dlcrop: "\u230d", - dollar: "$", - Dopf: "\ud835\udd3b", - dopf: "\ud835\udd55", - Dot: "\xa8", - dot: "\u02d9", - DotDot: "\u20dc", - doteq: "\u2250", - doteqdot: "\u2251", - DotEqual: "\u2250", - dotminus: "\u2238", - dotplus: "\u2214", - dotsquare: "\u22a1", - doublebarwedge: "\u2306", - DoubleContourIntegral: "\u222f", - DoubleDot: "\xa8", - DoubleDownArrow: "\u21d3", - DoubleLeftArrow: "\u21d0", - DoubleLeftRightArrow: "\u21d4", - DoubleLeftTee: "\u2ae4", - DoubleLongLeftArrow: "\u27f8", - DoubleLongLeftRightArrow: "\u27fa", - DoubleLongRightArrow: "\u27f9", - DoubleRightArrow: "\u21d2", - DoubleRightTee: "\u22a8", - DoubleUpArrow: "\u21d1", - DoubleUpDownArrow: "\u21d5", - DoubleVerticalBar: "\u2225", - DownArrowBar: "\u2913", - downarrow: "\u2193", - DownArrow: "\u2193", - Downarrow: "\u21d3", - DownArrowUpArrow: "\u21f5", - DownBreve: "\u0311", - downdownarrows: "\u21ca", - downharpoonleft: "\u21c3", - downharpoonright: "\u21c2", - DownLeftRightVector: "\u2950", - DownLeftTeeVector: "\u295e", - DownLeftVectorBar: "\u2956", - DownLeftVector: "\u21bd", - DownRightTeeVector: "\u295f", - DownRightVectorBar: "\u2957", - DownRightVector: "\u21c1", - DownTeeArrow: "\u21a7", - DownTee: "\u22a4", - drbkarow: "\u2910", - drcorn: "\u231f", - drcrop: "\u230c", - Dscr: "\ud835\udc9f", - dscr: "\ud835\udcb9", - DScy: "\u0405", - dscy: "\u0455", - dsol: "\u29f6", - Dstrok: "\u0110", - dstrok: "\u0111", - dtdot: "\u22f1", - dtri: "\u25bf", - dtrif: "\u25be", - duarr: "\u21f5", - duhar: "\u296f", - dwangle: "\u29a6", - DZcy: "\u040f", - dzcy: "\u045f", - dzigrarr: "\u27ff", - Eacute: "\xc9", - eacute: "\xe9", - easter: "\u2a6e", - Ecaron: "\u011a", - ecaron: "\u011b", - Ecirc: "\xca", - ecirc: "\xea", - ecir: "\u2256", - ecolon: "\u2255", - Ecy: "\u042d", - ecy: "\u044d", - eDDot: "\u2a77", - Edot: "\u0116", - edot: "\u0117", - eDot: "\u2251", - ee: "\u2147", - efDot: "\u2252", - Efr: "\ud835\udd08", - efr: "\ud835\udd22", - eg: "\u2a9a", - Egrave: "\xc8", - egrave: "\xe8", - egs: "\u2a96", - egsdot: "\u2a98", - el: "\u2a99", - Element: "\u2208", - elinters: "\u23e7", - ell: "\u2113", - els: "\u2a95", - elsdot: "\u2a97", - Emacr: "\u0112", - emacr: "\u0113", - empty: "\u2205", - emptyset: "\u2205", - EmptySmallSquare: "\u25fb", - emptyv: "\u2205", - EmptyVerySmallSquare: "\u25ab", - emsp13: "\u2004", - emsp14: "\u2005", - emsp: "\u2003", - ENG: "\u014a", - eng: "\u014b", - ensp: "\u2002", - Eogon: "\u0118", - eogon: "\u0119", - Eopf: "\ud835\udd3c", - eopf: "\ud835\udd56", - epar: "\u22d5", - eparsl: "\u29e3", - eplus: "\u2a71", - epsi: "\u03b5", - Epsilon: "\u0395", - epsilon: "\u03b5", - epsiv: "\u03f5", - eqcirc: "\u2256", - eqcolon: "\u2255", - eqsim: "\u2242", - eqslantgtr: "\u2a96", - eqslantless: "\u2a95", - Equal: "\u2a75", - equals: "=", - EqualTilde: "\u2242", - equest: "\u225f", - Equilibrium: "\u21cc", - equiv: "\u2261", - equivDD: "\u2a78", - eqvparsl: "\u29e5", - erarr: "\u2971", - erDot: "\u2253", - escr: "\u212f", - Escr: "\u2130", - esdot: "\u2250", - Esim: "\u2a73", - esim: "\u2242", - Eta: "\u0397", - eta: "\u03b7", - ETH: "\xd0", - eth: "\xf0", - Euml: "\xcb", - euml: "\xeb", - euro: "\u20ac", - excl: "!", - exist: "\u2203", - Exists: "\u2203", - expectation: "\u2130", - exponentiale: "\u2147", - ExponentialE: "\u2147", - fallingdotseq: "\u2252", - Fcy: "\u0424", - fcy: "\u0444", - female: "\u2640", - ffilig: "\ufb03", - fflig: "\ufb00", - ffllig: "\ufb04", - Ffr: "\ud835\udd09", - ffr: "\ud835\udd23", - filig: "\ufb01", - FilledSmallSquare: "\u25fc", - FilledVerySmallSquare: "\u25aa", - fjlig: "fj", - flat: "\u266d", - fllig: "\ufb02", - fltns: "\u25b1", - fnof: "\u0192", - Fopf: "\ud835\udd3d", - fopf: "\ud835\udd57", - forall: "\u2200", - ForAll: "\u2200", - fork: "\u22d4", - forkv: "\u2ad9", - Fouriertrf: "\u2131", - fpartint: "\u2a0d", - frac12: "\xbd", - frac13: "\u2153", - frac14: "\xbc", - frac15: "\u2155", - frac16: "\u2159", - frac18: "\u215b", - frac23: "\u2154", - frac25: "\u2156", - frac34: "\xbe", - frac35: "\u2157", - frac38: "\u215c", - frac45: "\u2158", - frac56: "\u215a", - frac58: "\u215d", - frac78: "\u215e", - frasl: "\u2044", - frown: "\u2322", - fscr: "\ud835\udcbb", - Fscr: "\u2131", - gacute: "\u01f5", - Gamma: "\u0393", - gamma: "\u03b3", - Gammad: "\u03dc", - gammad: "\u03dd", - gap: "\u2a86", - Gbreve: "\u011e", - gbreve: "\u011f", - Gcedil: "\u0122", - Gcirc: "\u011c", - gcirc: "\u011d", - Gcy: "\u0413", - gcy: "\u0433", - Gdot: "\u0120", - gdot: "\u0121", - ge: "\u2265", - gE: "\u2267", - gEl: "\u2a8c", - gel: "\u22db", - geq: "\u2265", - geqq: "\u2267", - geqslant: "\u2a7e", - gescc: "\u2aa9", - ges: "\u2a7e", - gesdot: "\u2a80", - gesdoto: "\u2a82", - gesdotol: "\u2a84", - gesl: "\u22db\ufe00", - gesles: "\u2a94", - Gfr: "\ud835\udd0a", - gfr: "\ud835\udd24", - gg: "\u226b", - Gg: "\u22d9", - ggg: "\u22d9", - gimel: "\u2137", - GJcy: "\u0403", - gjcy: "\u0453", - gla: "\u2aa5", - gl: "\u2277", - glE: "\u2a92", - glj: "\u2aa4", - gnap: "\u2a8a", - gnapprox: "\u2a8a", - gne: "\u2a88", - gnE: "\u2269", - gneq: "\u2a88", - gneqq: "\u2269", - gnsim: "\u22e7", - Gopf: "\ud835\udd3e", - gopf: "\ud835\udd58", - grave: "`", - GreaterEqual: "\u2265", - GreaterEqualLess: "\u22db", - GreaterFullEqual: "\u2267", - GreaterGreater: "\u2aa2", - GreaterLess: "\u2277", - GreaterSlantEqual: "\u2a7e", - GreaterTilde: "\u2273", - Gscr: "\ud835\udca2", - gscr: "\u210a", - gsim: "\u2273", - gsime: "\u2a8e", - gsiml: "\u2a90", - gtcc: "\u2aa7", - gtcir: "\u2a7a", - gt: ">", - GT: ">", - Gt: "\u226b", - gtdot: "\u22d7", - gtlPar: "\u2995", - gtquest: "\u2a7c", - gtrapprox: "\u2a86", - gtrarr: "\u2978", - gtrdot: "\u22d7", - gtreqless: "\u22db", - gtreqqless: "\u2a8c", - gtrless: "\u2277", - gtrsim: "\u2273", - gvertneqq: "\u2269\ufe00", - gvnE: "\u2269\ufe00", - Hacek: "\u02c7", - hairsp: "\u200a", - half: "\xbd", - hamilt: "\u210b", - HARDcy: "\u042a", - hardcy: "\u044a", - harrcir: "\u2948", - harr: "\u2194", - hArr: "\u21d4", - harrw: "\u21ad", - Hat: "^", - hbar: "\u210f", - Hcirc: "\u0124", - hcirc: "\u0125", - hearts: "\u2665", - heartsuit: "\u2665", - hellip: "\u2026", - hercon: "\u22b9", - hfr: "\ud835\udd25", - Hfr: "\u210c", - HilbertSpace: "\u210b", - hksearow: "\u2925", - hkswarow: "\u2926", - hoarr: "\u21ff", - homtht: "\u223b", - hookleftarrow: "\u21a9", - hookrightarrow: "\u21aa", - hopf: "\ud835\udd59", - Hopf: "\u210d", - horbar: "\u2015", - HorizontalLine: "\u2500", - hscr: "\ud835\udcbd", - Hscr: "\u210b", - hslash: "\u210f", - Hstrok: "\u0126", - hstrok: "\u0127", - HumpDownHump: "\u224e", - HumpEqual: "\u224f", - hybull: "\u2043", - hyphen: "\u2010", - Iacute: "\xcd", - iacute: "\xed", - ic: "\u2063", - Icirc: "\xce", - icirc: "\xee", - Icy: "\u0418", - icy: "\u0438", - Idot: "\u0130", - IEcy: "\u0415", - iecy: "\u0435", - iexcl: "\xa1", - iff: "\u21d4", - ifr: "\ud835\udd26", - Ifr: "\u2111", - Igrave: "\xcc", - igrave: "\xec", - ii: "\u2148", - iiiint: "\u2a0c", - iiint: "\u222d", - iinfin: "\u29dc", - iiota: "\u2129", - IJlig: "\u0132", - ijlig: "\u0133", - Imacr: "\u012a", - imacr: "\u012b", - image: "\u2111", - ImaginaryI: "\u2148", - imagline: "\u2110", - imagpart: "\u2111", - imath: "\u0131", - Im: "\u2111", - imof: "\u22b7", - imped: "\u01b5", - Implies: "\u21d2", - incare: "\u2105", - in: "\u2208", - infin: "\u221e", - infintie: "\u29dd", - inodot: "\u0131", - intcal: "\u22ba", - int: "\u222b", - Int: "\u222c", - integers: "\u2124", - Integral: "\u222b", - intercal: "\u22ba", - Intersection: "\u22c2", - intlarhk: "\u2a17", - intprod: "\u2a3c", - InvisibleComma: "\u2063", - InvisibleTimes: "\u2062", - IOcy: "\u0401", - iocy: "\u0451", - Iogon: "\u012e", - iogon: "\u012f", - Iopf: "\ud835\udd40", - iopf: "\ud835\udd5a", - Iota: "\u0399", - iota: "\u03b9", - iprod: "\u2a3c", - iquest: "\xbf", - iscr: "\ud835\udcbe", - Iscr: "\u2110", - isin: "\u2208", - isindot: "\u22f5", - isinE: "\u22f9", - isins: "\u22f4", - isinsv: "\u22f3", - isinv: "\u2208", - it: "\u2062", - Itilde: "\u0128", - itilde: "\u0129", - Iukcy: "\u0406", - iukcy: "\u0456", - Iuml: "\xcf", - iuml: "\xef", - Jcirc: "\u0134", - jcirc: "\u0135", - Jcy: "\u0419", - jcy: "\u0439", - Jfr: "\ud835\udd0d", - jfr: "\ud835\udd27", - jmath: "\u0237", - Jopf: "\ud835\udd41", - jopf: "\ud835\udd5b", - Jscr: "\ud835\udca5", - jscr: "\ud835\udcbf", - Jsercy: "\u0408", - jsercy: "\u0458", - Jukcy: "\u0404", - jukcy: "\u0454", - Kappa: "\u039a", - kappa: "\u03ba", - kappav: "\u03f0", - Kcedil: "\u0136", - kcedil: "\u0137", - Kcy: "\u041a", - kcy: "\u043a", - Kfr: "\ud835\udd0e", - kfr: "\ud835\udd28", - kgreen: "\u0138", - KHcy: "\u0425", - khcy: "\u0445", - KJcy: "\u040c", - kjcy: "\u045c", - Kopf: "\ud835\udd42", - kopf: "\ud835\udd5c", - Kscr: "\ud835\udca6", - kscr: "\ud835\udcc0", - lAarr: "\u21da", - Lacute: "\u0139", - lacute: "\u013a", - laemptyv: "\u29b4", - lagran: "\u2112", - Lambda: "\u039b", - lambda: "\u03bb", - lang: "\u27e8", - Lang: "\u27ea", - langd: "\u2991", - langle: "\u27e8", - lap: "\u2a85", - Laplacetrf: "\u2112", - laquo: "\xab", - larrb: "\u21e4", - larrbfs: "\u291f", - larr: "\u2190", - Larr: "\u219e", - lArr: "\u21d0", - larrfs: "\u291d", - larrhk: "\u21a9", - larrlp: "\u21ab", - larrpl: "\u2939", - larrsim: "\u2973", - larrtl: "\u21a2", - latail: "\u2919", - lAtail: "\u291b", - lat: "\u2aab", - late: "\u2aad", - lates: "\u2aad\ufe00", - lbarr: "\u290c", - lBarr: "\u290e", - lbbrk: "\u2772", - lbrace: "{", - lbrack: "[", - lbrke: "\u298b", - lbrksld: "\u298f", - lbrkslu: "\u298d", - Lcaron: "\u013d", - lcaron: "\u013e", - Lcedil: "\u013b", - lcedil: "\u013c", - lceil: "\u2308", - lcub: "{", - Lcy: "\u041b", - lcy: "\u043b", - ldca: "\u2936", - ldquo: "\u201c", - ldquor: "\u201e", - ldrdhar: "\u2967", - ldrushar: "\u294b", - ldsh: "\u21b2", - le: "\u2264", - lE: "\u2266", - LeftAngleBracket: "\u27e8", - LeftArrowBar: "\u21e4", - leftarrow: "\u2190", - LeftArrow: "\u2190", - Leftarrow: "\u21d0", - LeftArrowRightArrow: "\u21c6", - leftarrowtail: "\u21a2", - LeftCeiling: "\u2308", - LeftDoubleBracket: "\u27e6", - LeftDownTeeVector: "\u2961", - LeftDownVectorBar: "\u2959", - LeftDownVector: "\u21c3", - LeftFloor: "\u230a", - leftharpoondown: "\u21bd", - leftharpoonup: "\u21bc", - leftleftarrows: "\u21c7", - leftrightarrow: "\u2194", - LeftRightArrow: "\u2194", - Leftrightarrow: "\u21d4", - leftrightarrows: "\u21c6", - leftrightharpoons: "\u21cb", - leftrightsquigarrow: "\u21ad", - LeftRightVector: "\u294e", - LeftTeeArrow: "\u21a4", - LeftTee: "\u22a3", - LeftTeeVector: "\u295a", - leftthreetimes: "\u22cb", - LeftTriangleBar: "\u29cf", - LeftTriangle: "\u22b2", - LeftTriangleEqual: "\u22b4", - LeftUpDownVector: "\u2951", - LeftUpTeeVector: "\u2960", - LeftUpVectorBar: "\u2958", - LeftUpVector: "\u21bf", - LeftVectorBar: "\u2952", - LeftVector: "\u21bc", - lEg: "\u2a8b", - leg: "\u22da", - leq: "\u2264", - leqq: "\u2266", - leqslant: "\u2a7d", - lescc: "\u2aa8", - les: "\u2a7d", - lesdot: "\u2a7f", - lesdoto: "\u2a81", - lesdotor: "\u2a83", - lesg: "\u22da\ufe00", - lesges: "\u2a93", - lessapprox: "\u2a85", - lessdot: "\u22d6", - lesseqgtr: "\u22da", - lesseqqgtr: "\u2a8b", - LessEqualGreater: "\u22da", - LessFullEqual: "\u2266", - LessGreater: "\u2276", - lessgtr: "\u2276", - LessLess: "\u2aa1", - lesssim: "\u2272", - LessSlantEqual: "\u2a7d", - LessTilde: "\u2272", - lfisht: "\u297c", - lfloor: "\u230a", - Lfr: "\ud835\udd0f", - lfr: "\ud835\udd29", - lg: "\u2276", - lgE: "\u2a91", - lHar: "\u2962", - lhard: "\u21bd", - lharu: "\u21bc", - lharul: "\u296a", - lhblk: "\u2584", - LJcy: "\u0409", - ljcy: "\u0459", - llarr: "\u21c7", - ll: "\u226a", - Ll: "\u22d8", - llcorner: "\u231e", - Lleftarrow: "\u21da", - llhard: "\u296b", - lltri: "\u25fa", - Lmidot: "\u013f", - lmidot: "\u0140", - lmoustache: "\u23b0", - lmoust: "\u23b0", - lnap: "\u2a89", - lnapprox: "\u2a89", - lne: "\u2a87", - lnE: "\u2268", - lneq: "\u2a87", - lneqq: "\u2268", - lnsim: "\u22e6", - loang: "\u27ec", - loarr: "\u21fd", - lobrk: "\u27e6", - longleftarrow: "\u27f5", - LongLeftArrow: "\u27f5", - Longleftarrow: "\u27f8", - longleftrightarrow: "\u27f7", - LongLeftRightArrow: "\u27f7", - Longleftrightarrow: "\u27fa", - longmapsto: "\u27fc", - longrightarrow: "\u27f6", - LongRightArrow: "\u27f6", - Longrightarrow: "\u27f9", - looparrowleft: "\u21ab", - looparrowright: "\u21ac", - lopar: "\u2985", - Lopf: "\ud835\udd43", - lopf: "\ud835\udd5d", - loplus: "\u2a2d", - lotimes: "\u2a34", - lowast: "\u2217", - lowbar: "_", - LowerLeftArrow: "\u2199", - LowerRightArrow: "\u2198", - loz: "\u25ca", - lozenge: "\u25ca", - lozf: "\u29eb", - lpar: "(", - lparlt: "\u2993", - lrarr: "\u21c6", - lrcorner: "\u231f", - lrhar: "\u21cb", - lrhard: "\u296d", - lrm: "\u200e", - lrtri: "\u22bf", - lsaquo: "\u2039", - lscr: "\ud835\udcc1", - Lscr: "\u2112", - lsh: "\u21b0", - Lsh: "\u21b0", - lsim: "\u2272", - lsime: "\u2a8d", - lsimg: "\u2a8f", - lsqb: "[", - lsquo: "\u2018", - lsquor: "\u201a", - Lstrok: "\u0141", - lstrok: "\u0142", - ltcc: "\u2aa6", - ltcir: "\u2a79", - lt: "<", - LT: "<", - Lt: "\u226a", - ltdot: "\u22d6", - lthree: "\u22cb", - ltimes: "\u22c9", - ltlarr: "\u2976", - ltquest: "\u2a7b", - ltri: "\u25c3", - ltrie: "\u22b4", - ltrif: "\u25c2", - ltrPar: "\u2996", - lurdshar: "\u294a", - luruhar: "\u2966", - lvertneqq: "\u2268\ufe00", - lvnE: "\u2268\ufe00", - macr: "\xaf", - male: "\u2642", - malt: "\u2720", - maltese: "\u2720", - Map: "\u2905", - map: "\u21a6", - mapsto: "\u21a6", - mapstodown: "\u21a7", - mapstoleft: "\u21a4", - mapstoup: "\u21a5", - marker: "\u25ae", - mcomma: "\u2a29", - Mcy: "\u041c", - mcy: "\u043c", - mdash: "\u2014", - mDDot: "\u223a", - measuredangle: "\u2221", - MediumSpace: "\u205f", - Mellintrf: "\u2133", - Mfr: "\ud835\udd10", - mfr: "\ud835\udd2a", - mho: "\u2127", - micro: "\xb5", - midast: "*", - midcir: "\u2af0", - mid: "\u2223", - middot: "\xb7", - minusb: "\u229f", - minus: "\u2212", - minusd: "\u2238", - minusdu: "\u2a2a", - MinusPlus: "\u2213", - mlcp: "\u2adb", - mldr: "\u2026", - mnplus: "\u2213", - models: "\u22a7", - Mopf: "\ud835\udd44", - mopf: "\ud835\udd5e", - mp: "\u2213", - mscr: "\ud835\udcc2", - Mscr: "\u2133", - mstpos: "\u223e", - Mu: "\u039c", - mu: "\u03bc", - multimap: "\u22b8", - mumap: "\u22b8", - nabla: "\u2207", - Nacute: "\u0143", - nacute: "\u0144", - nang: "\u2220\u20d2", - nap: "\u2249", - napE: "\u2a70\u0338", - napid: "\u224b\u0338", - napos: "\u0149", - napprox: "\u2249", - natural: "\u266e", - naturals: "\u2115", - natur: "\u266e", - nbsp: "\xa0", - nbump: "\u224e\u0338", - nbumpe: "\u224f\u0338", - ncap: "\u2a43", - Ncaron: "\u0147", - ncaron: "\u0148", - Ncedil: "\u0145", - ncedil: "\u0146", - ncong: "\u2247", - ncongdot: "\u2a6d\u0338", - ncup: "\u2a42", - Ncy: "\u041d", - ncy: "\u043d", - ndash: "\u2013", - nearhk: "\u2924", - nearr: "\u2197", - neArr: "\u21d7", - nearrow: "\u2197", - ne: "\u2260", - nedot: "\u2250\u0338", - NegativeMediumSpace: "\u200b", - NegativeThickSpace: "\u200b", - NegativeThinSpace: "\u200b", - NegativeVeryThinSpace: "\u200b", - nequiv: "\u2262", - nesear: "\u2928", - nesim: "\u2242\u0338", - NestedGreaterGreater: "\u226b", - NestedLessLess: "\u226a", - NewLine: "\n", - nexist: "\u2204", - nexists: "\u2204", - Nfr: "\ud835\udd11", - nfr: "\ud835\udd2b", - ngE: "\u2267\u0338", - nge: "\u2271", - ngeq: "\u2271", - ngeqq: "\u2267\u0338", - ngeqslant: "\u2a7e\u0338", - nges: "\u2a7e\u0338", - nGg: "\u22d9\u0338", - ngsim: "\u2275", - nGt: "\u226b\u20d2", - ngt: "\u226f", - ngtr: "\u226f", - nGtv: "\u226b\u0338", - nharr: "\u21ae", - nhArr: "\u21ce", - nhpar: "\u2af2", - ni: "\u220b", - nis: "\u22fc", - nisd: "\u22fa", - niv: "\u220b", - NJcy: "\u040a", - njcy: "\u045a", - nlarr: "\u219a", - nlArr: "\u21cd", - nldr: "\u2025", - nlE: "\u2266\u0338", - nle: "\u2270", - nleftarrow: "\u219a", - nLeftarrow: "\u21cd", - nleftrightarrow: "\u21ae", - nLeftrightarrow: "\u21ce", - nleq: "\u2270", - nleqq: "\u2266\u0338", - nleqslant: "\u2a7d\u0338", - nles: "\u2a7d\u0338", - nless: "\u226e", - nLl: "\u22d8\u0338", - nlsim: "\u2274", - nLt: "\u226a\u20d2", - nlt: "\u226e", - nltri: "\u22ea", - nltrie: "\u22ec", - nLtv: "\u226a\u0338", - nmid: "\u2224", - NoBreak: "\u2060", - NonBreakingSpace: "\xa0", - nopf: "\ud835\udd5f", - Nopf: "\u2115", - Not: "\u2aec", - not: "\xac", - NotCongruent: "\u2262", - NotCupCap: "\u226d", - NotDoubleVerticalBar: "\u2226", - NotElement: "\u2209", - NotEqual: "\u2260", - NotEqualTilde: "\u2242\u0338", - NotExists: "\u2204", - NotGreater: "\u226f", - NotGreaterEqual: "\u2271", - NotGreaterFullEqual: "\u2267\u0338", - NotGreaterGreater: "\u226b\u0338", - NotGreaterLess: "\u2279", - NotGreaterSlantEqual: "\u2a7e\u0338", - NotGreaterTilde: "\u2275", - NotHumpDownHump: "\u224e\u0338", - NotHumpEqual: "\u224f\u0338", - notin: "\u2209", - notindot: "\u22f5\u0338", - notinE: "\u22f9\u0338", - notinva: "\u2209", - notinvb: "\u22f7", - notinvc: "\u22f6", - NotLeftTriangleBar: "\u29cf\u0338", - NotLeftTriangle: "\u22ea", - NotLeftTriangleEqual: "\u22ec", - NotLess: "\u226e", - NotLessEqual: "\u2270", - NotLessGreater: "\u2278", - NotLessLess: "\u226a\u0338", - NotLessSlantEqual: "\u2a7d\u0338", - NotLessTilde: "\u2274", - NotNestedGreaterGreater: "\u2aa2\u0338", - NotNestedLessLess: "\u2aa1\u0338", - notni: "\u220c", - notniva: "\u220c", - notnivb: "\u22fe", - notnivc: "\u22fd", - NotPrecedes: "\u2280", - NotPrecedesEqual: "\u2aaf\u0338", - NotPrecedesSlantEqual: "\u22e0", - NotReverseElement: "\u220c", - NotRightTriangleBar: "\u29d0\u0338", - NotRightTriangle: "\u22eb", - NotRightTriangleEqual: "\u22ed", - NotSquareSubset: "\u228f\u0338", - NotSquareSubsetEqual: "\u22e2", - NotSquareSuperset: "\u2290\u0338", - NotSquareSupersetEqual: "\u22e3", - NotSubset: "\u2282\u20d2", - NotSubsetEqual: "\u2288", - NotSucceeds: "\u2281", - NotSucceedsEqual: "\u2ab0\u0338", - NotSucceedsSlantEqual: "\u22e1", - NotSucceedsTilde: "\u227f\u0338", - NotSuperset: "\u2283\u20d2", - NotSupersetEqual: "\u2289", - NotTilde: "\u2241", - NotTildeEqual: "\u2244", - NotTildeFullEqual: "\u2247", - NotTildeTilde: "\u2249", - NotVerticalBar: "\u2224", - nparallel: "\u2226", - npar: "\u2226", - nparsl: "\u2afd\u20e5", - npart: "\u2202\u0338", - npolint: "\u2a14", - npr: "\u2280", - nprcue: "\u22e0", - nprec: "\u2280", - npreceq: "\u2aaf\u0338", - npre: "\u2aaf\u0338", - nrarrc: "\u2933\u0338", - nrarr: "\u219b", - nrArr: "\u21cf", - nrarrw: "\u219d\u0338", - nrightarrow: "\u219b", - nRightarrow: "\u21cf", - nrtri: "\u22eb", - nrtrie: "\u22ed", - nsc: "\u2281", - nsccue: "\u22e1", - nsce: "\u2ab0\u0338", - Nscr: "\ud835\udca9", - nscr: "\ud835\udcc3", - nshortmid: "\u2224", - nshortparallel: "\u2226", - nsim: "\u2241", - nsime: "\u2244", - nsimeq: "\u2244", - nsmid: "\u2224", - nspar: "\u2226", - nsqsube: "\u22e2", - nsqsupe: "\u22e3", - nsub: "\u2284", - nsubE: "\u2ac5\u0338", - nsube: "\u2288", - nsubset: "\u2282\u20d2", - nsubseteq: "\u2288", - nsubseteqq: "\u2ac5\u0338", - nsucc: "\u2281", - nsucceq: "\u2ab0\u0338", - nsup: "\u2285", - nsupE: "\u2ac6\u0338", - nsupe: "\u2289", - nsupset: "\u2283\u20d2", - nsupseteq: "\u2289", - nsupseteqq: "\u2ac6\u0338", - ntgl: "\u2279", - Ntilde: "\xd1", - ntilde: "\xf1", - ntlg: "\u2278", - ntriangleleft: "\u22ea", - ntrianglelefteq: "\u22ec", - ntriangleright: "\u22eb", - ntrianglerighteq: "\u22ed", - Nu: "\u039d", - nu: "\u03bd", - num: "#", - numero: "\u2116", - numsp: "\u2007", - nvap: "\u224d\u20d2", - nvdash: "\u22ac", - nvDash: "\u22ad", - nVdash: "\u22ae", - nVDash: "\u22af", - nvge: "\u2265\u20d2", - nvgt: ">\u20d2", - nvHarr: "\u2904", - nvinfin: "\u29de", - nvlArr: "\u2902", - nvle: "\u2264\u20d2", - nvlt: "<\u20d2", - nvltrie: "\u22b4\u20d2", - nvrArr: "\u2903", - nvrtrie: "\u22b5\u20d2", - nvsim: "\u223c\u20d2", - nwarhk: "\u2923", - nwarr: "\u2196", - nwArr: "\u21d6", - nwarrow: "\u2196", - nwnear: "\u2927", - Oacute: "\xd3", - oacute: "\xf3", - oast: "\u229b", - Ocirc: "\xd4", - ocirc: "\xf4", - ocir: "\u229a", - Ocy: "\u041e", - ocy: "\u043e", - odash: "\u229d", - Odblac: "\u0150", - odblac: "\u0151", - odiv: "\u2a38", - odot: "\u2299", - odsold: "\u29bc", - OElig: "\u0152", - oelig: "\u0153", - ofcir: "\u29bf", - Ofr: "\ud835\udd12", - ofr: "\ud835\udd2c", - ogon: "\u02db", - Ograve: "\xd2", - ograve: "\xf2", - ogt: "\u29c1", - ohbar: "\u29b5", - ohm: "\u03a9", - oint: "\u222e", - olarr: "\u21ba", - olcir: "\u29be", - olcross: "\u29bb", - oline: "\u203e", - olt: "\u29c0", - Omacr: "\u014c", - omacr: "\u014d", - Omega: "\u03a9", - omega: "\u03c9", - Omicron: "\u039f", - omicron: "\u03bf", - omid: "\u29b6", - ominus: "\u2296", - Oopf: "\ud835\udd46", - oopf: "\ud835\udd60", - opar: "\u29b7", - OpenCurlyDoubleQuote: "\u201c", - OpenCurlyQuote: "\u2018", - operp: "\u29b9", - oplus: "\u2295", - orarr: "\u21bb", - Or: "\u2a54", - or: "\u2228", - ord: "\u2a5d", - order: "\u2134", - orderof: "\u2134", - ordf: "\xaa", - ordm: "\xba", - origof: "\u22b6", - oror: "\u2a56", - orslope: "\u2a57", - orv: "\u2a5b", - oS: "\u24c8", - Oscr: "\ud835\udcaa", - oscr: "\u2134", - Oslash: "\xd8", - oslash: "\xf8", - osol: "\u2298", - Otilde: "\xd5", - otilde: "\xf5", - otimesas: "\u2a36", - Otimes: "\u2a37", - otimes: "\u2297", - Ouml: "\xd6", - ouml: "\xf6", - ovbar: "\u233d", - OverBar: "\u203e", - OverBrace: "\u23de", - OverBracket: "\u23b4", - OverParenthesis: "\u23dc", - para: "\xb6", - parallel: "\u2225", - par: "\u2225", - parsim: "\u2af3", - parsl: "\u2afd", - part: "\u2202", - PartialD: "\u2202", - Pcy: "\u041f", - pcy: "\u043f", - percnt: "%", - period: ".", - permil: "\u2030", - perp: "\u22a5", - pertenk: "\u2031", - Pfr: "\ud835\udd13", - pfr: "\ud835\udd2d", - Phi: "\u03a6", - phi: "\u03c6", - phiv: "\u03d5", - phmmat: "\u2133", - phone: "\u260e", - Pi: "\u03a0", - pi: "\u03c0", - pitchfork: "\u22d4", - piv: "\u03d6", - planck: "\u210f", - planckh: "\u210e", - plankv: "\u210f", - plusacir: "\u2a23", - plusb: "\u229e", - pluscir: "\u2a22", - plus: "+", - plusdo: "\u2214", - plusdu: "\u2a25", - pluse: "\u2a72", - PlusMinus: "\xb1", - plusmn: "\xb1", - plussim: "\u2a26", - plustwo: "\u2a27", - pm: "\xb1", - Poincareplane: "\u210c", - pointint: "\u2a15", - popf: "\ud835\udd61", - Popf: "\u2119", - pound: "\xa3", - prap: "\u2ab7", - Pr: "\u2abb", - pr: "\u227a", - prcue: "\u227c", - precapprox: "\u2ab7", - prec: "\u227a", - preccurlyeq: "\u227c", - Precedes: "\u227a", - PrecedesEqual: "\u2aaf", - PrecedesSlantEqual: "\u227c", - PrecedesTilde: "\u227e", - preceq: "\u2aaf", - precnapprox: "\u2ab9", - precneqq: "\u2ab5", - precnsim: "\u22e8", - pre: "\u2aaf", - prE: "\u2ab3", - precsim: "\u227e", - prime: "\u2032", - Prime: "\u2033", - primes: "\u2119", - prnap: "\u2ab9", - prnE: "\u2ab5", - prnsim: "\u22e8", - prod: "\u220f", - Product: "\u220f", - profalar: "\u232e", - profline: "\u2312", - profsurf: "\u2313", - prop: "\u221d", - Proportional: "\u221d", - Proportion: "\u2237", - propto: "\u221d", - prsim: "\u227e", - prurel: "\u22b0", - Pscr: "\ud835\udcab", - pscr: "\ud835\udcc5", - Psi: "\u03a8", - psi: "\u03c8", - puncsp: "\u2008", - Qfr: "\ud835\udd14", - qfr: "\ud835\udd2e", - qint: "\u2a0c", - qopf: "\ud835\udd62", - Qopf: "\u211a", - qprime: "\u2057", - Qscr: "\ud835\udcac", - qscr: "\ud835\udcc6", - quaternions: "\u210d", - quatint: "\u2a16", - quest: "?", - questeq: "\u225f", - quot: '"', - QUOT: '"', - rAarr: "\u21db", - race: "\u223d\u0331", - Racute: "\u0154", - racute: "\u0155", - radic: "\u221a", - raemptyv: "\u29b3", - rang: "\u27e9", - Rang: "\u27eb", - rangd: "\u2992", - range: "\u29a5", - rangle: "\u27e9", - raquo: "\xbb", - rarrap: "\u2975", - rarrb: "\u21e5", - rarrbfs: "\u2920", - rarrc: "\u2933", - rarr: "\u2192", - Rarr: "\u21a0", - rArr: "\u21d2", - rarrfs: "\u291e", - rarrhk: "\u21aa", - rarrlp: "\u21ac", - rarrpl: "\u2945", - rarrsim: "\u2974", - Rarrtl: "\u2916", - rarrtl: "\u21a3", - rarrw: "\u219d", - ratail: "\u291a", - rAtail: "\u291c", - ratio: "\u2236", - rationals: "\u211a", - rbarr: "\u290d", - rBarr: "\u290f", - RBarr: "\u2910", - rbbrk: "\u2773", - rbrace: "}", - rbrack: "]", - rbrke: "\u298c", - rbrksld: "\u298e", - rbrkslu: "\u2990", - Rcaron: "\u0158", - rcaron: "\u0159", - Rcedil: "\u0156", - rcedil: "\u0157", - rceil: "\u2309", - rcub: "}", - Rcy: "\u0420", - rcy: "\u0440", - rdca: "\u2937", - rdldhar: "\u2969", - rdquo: "\u201d", - rdquor: "\u201d", - rdsh: "\u21b3", - real: "\u211c", - realine: "\u211b", - realpart: "\u211c", - reals: "\u211d", - Re: "\u211c", - rect: "\u25ad", - reg: "\xae", - REG: "\xae", - ReverseElement: "\u220b", - ReverseEquilibrium: "\u21cb", - ReverseUpEquilibrium: "\u296f", - rfisht: "\u297d", - rfloor: "\u230b", - rfr: "\ud835\udd2f", - Rfr: "\u211c", - rHar: "\u2964", - rhard: "\u21c1", - rharu: "\u21c0", - rharul: "\u296c", - Rho: "\u03a1", - rho: "\u03c1", - rhov: "\u03f1", - RightAngleBracket: "\u27e9", - RightArrowBar: "\u21e5", - rightarrow: "\u2192", - RightArrow: "\u2192", - Rightarrow: "\u21d2", - RightArrowLeftArrow: "\u21c4", - rightarrowtail: "\u21a3", - RightCeiling: "\u2309", - RightDoubleBracket: "\u27e7", - RightDownTeeVector: "\u295d", - RightDownVectorBar: "\u2955", - RightDownVector: "\u21c2", - RightFloor: "\u230b", - rightharpoondown: "\u21c1", - rightharpoonup: "\u21c0", - rightleftarrows: "\u21c4", - rightleftharpoons: "\u21cc", - rightrightarrows: "\u21c9", - rightsquigarrow: "\u219d", - RightTeeArrow: "\u21a6", - RightTee: "\u22a2", - RightTeeVector: "\u295b", - rightthreetimes: "\u22cc", - RightTriangleBar: "\u29d0", - RightTriangle: "\u22b3", - RightTriangleEqual: "\u22b5", - RightUpDownVector: "\u294f", - RightUpTeeVector: "\u295c", - RightUpVectorBar: "\u2954", - RightUpVector: "\u21be", - RightVectorBar: "\u2953", - RightVector: "\u21c0", - ring: "\u02da", - risingdotseq: "\u2253", - rlarr: "\u21c4", - rlhar: "\u21cc", - rlm: "\u200f", - rmoustache: "\u23b1", - rmoust: "\u23b1", - rnmid: "\u2aee", - roang: "\u27ed", - roarr: "\u21fe", - robrk: "\u27e7", - ropar: "\u2986", - ropf: "\ud835\udd63", - Ropf: "\u211d", - roplus: "\u2a2e", - rotimes: "\u2a35", - RoundImplies: "\u2970", - rpar: ")", - rpargt: "\u2994", - rppolint: "\u2a12", - rrarr: "\u21c9", - Rrightarrow: "\u21db", - rsaquo: "\u203a", - rscr: "\ud835\udcc7", - Rscr: "\u211b", - rsh: "\u21b1", - Rsh: "\u21b1", - rsqb: "]", - rsquo: "\u2019", - rsquor: "\u2019", - rthree: "\u22cc", - rtimes: "\u22ca", - rtri: "\u25b9", - rtrie: "\u22b5", - rtrif: "\u25b8", - rtriltri: "\u29ce", - RuleDelayed: "\u29f4", - ruluhar: "\u2968", - rx: "\u211e", - Sacute: "\u015a", - sacute: "\u015b", - sbquo: "\u201a", - scap: "\u2ab8", - Scaron: "\u0160", - scaron: "\u0161", - Sc: "\u2abc", - sc: "\u227b", - sccue: "\u227d", - sce: "\u2ab0", - scE: "\u2ab4", - Scedil: "\u015e", - scedil: "\u015f", - Scirc: "\u015c", - scirc: "\u015d", - scnap: "\u2aba", - scnE: "\u2ab6", - scnsim: "\u22e9", - scpolint: "\u2a13", - scsim: "\u227f", - Scy: "\u0421", - scy: "\u0441", - sdotb: "\u22a1", - sdot: "\u22c5", - sdote: "\u2a66", - searhk: "\u2925", - searr: "\u2198", - seArr: "\u21d8", - searrow: "\u2198", - sect: "\xa7", - semi: ";", - seswar: "\u2929", - setminus: "\u2216", - setmn: "\u2216", - sext: "\u2736", - Sfr: "\ud835\udd16", - sfr: "\ud835\udd30", - sfrown: "\u2322", - sharp: "\u266f", - SHCHcy: "\u0429", - shchcy: "\u0449", - SHcy: "\u0428", - shcy: "\u0448", - ShortDownArrow: "\u2193", - ShortLeftArrow: "\u2190", - shortmid: "\u2223", - shortparallel: "\u2225", - ShortRightArrow: "\u2192", - ShortUpArrow: "\u2191", - shy: "\xad", - Sigma: "\u03a3", - sigma: "\u03c3", - sigmaf: "\u03c2", - sigmav: "\u03c2", - sim: "\u223c", - simdot: "\u2a6a", - sime: "\u2243", - simeq: "\u2243", - simg: "\u2a9e", - simgE: "\u2aa0", - siml: "\u2a9d", - simlE: "\u2a9f", - simne: "\u2246", - simplus: "\u2a24", - simrarr: "\u2972", - slarr: "\u2190", - SmallCircle: "\u2218", - smallsetminus: "\u2216", - smashp: "\u2a33", - smeparsl: "\u29e4", - smid: "\u2223", - smile: "\u2323", - smt: "\u2aaa", - smte: "\u2aac", - smtes: "\u2aac\ufe00", - SOFTcy: "\u042c", - softcy: "\u044c", - solbar: "\u233f", - solb: "\u29c4", - sol: "/", - Sopf: "\ud835\udd4a", - sopf: "\ud835\udd64", - spades: "\u2660", - spadesuit: "\u2660", - spar: "\u2225", - sqcap: "\u2293", - sqcaps: "\u2293\ufe00", - sqcup: "\u2294", - sqcups: "\u2294\ufe00", - Sqrt: "\u221a", - sqsub: "\u228f", - sqsube: "\u2291", - sqsubset: "\u228f", - sqsubseteq: "\u2291", - sqsup: "\u2290", - sqsupe: "\u2292", - sqsupset: "\u2290", - sqsupseteq: "\u2292", - square: "\u25a1", - Square: "\u25a1", - SquareIntersection: "\u2293", - SquareSubset: "\u228f", - SquareSubsetEqual: "\u2291", - SquareSuperset: "\u2290", - SquareSupersetEqual: "\u2292", - SquareUnion: "\u2294", - squarf: "\u25aa", - squ: "\u25a1", - squf: "\u25aa", - srarr: "\u2192", - Sscr: "\ud835\udcae", - sscr: "\ud835\udcc8", - ssetmn: "\u2216", - ssmile: "\u2323", - sstarf: "\u22c6", - Star: "\u22c6", - star: "\u2606", - starf: "\u2605", - straightepsilon: "\u03f5", - straightphi: "\u03d5", - strns: "\xaf", - sub: "\u2282", - Sub: "\u22d0", - subdot: "\u2abd", - subE: "\u2ac5", - sube: "\u2286", - subedot: "\u2ac3", - submult: "\u2ac1", - subnE: "\u2acb", - subne: "\u228a", - subplus: "\u2abf", - subrarr: "\u2979", - subset: "\u2282", - Subset: "\u22d0", - subseteq: "\u2286", - subseteqq: "\u2ac5", - SubsetEqual: "\u2286", - subsetneq: "\u228a", - subsetneqq: "\u2acb", - subsim: "\u2ac7", - subsub: "\u2ad5", - subsup: "\u2ad3", - succapprox: "\u2ab8", - succ: "\u227b", - succcurlyeq: "\u227d", - Succeeds: "\u227b", - SucceedsEqual: "\u2ab0", - SucceedsSlantEqual: "\u227d", - SucceedsTilde: "\u227f", - succeq: "\u2ab0", - succnapprox: "\u2aba", - succneqq: "\u2ab6", - succnsim: "\u22e9", - succsim: "\u227f", - SuchThat: "\u220b", - sum: "\u2211", - Sum: "\u2211", - sung: "\u266a", - sup1: "\xb9", - sup2: "\xb2", - sup3: "\xb3", - sup: "\u2283", - Sup: "\u22d1", - supdot: "\u2abe", - supdsub: "\u2ad8", - supE: "\u2ac6", - supe: "\u2287", - supedot: "\u2ac4", - Superset: "\u2283", - SupersetEqual: "\u2287", - suphsol: "\u27c9", - suphsub: "\u2ad7", - suplarr: "\u297b", - supmult: "\u2ac2", - supnE: "\u2acc", - supne: "\u228b", - supplus: "\u2ac0", - supset: "\u2283", - Supset: "\u22d1", - supseteq: "\u2287", - supseteqq: "\u2ac6", - supsetneq: "\u228b", - supsetneqq: "\u2acc", - supsim: "\u2ac8", - supsub: "\u2ad4", - supsup: "\u2ad6", - swarhk: "\u2926", - swarr: "\u2199", - swArr: "\u21d9", - swarrow: "\u2199", - swnwar: "\u292a", - szlig: "\xdf", - Tab: "\t", - target: "\u2316", - Tau: "\u03a4", - tau: "\u03c4", - tbrk: "\u23b4", - Tcaron: "\u0164", - tcaron: "\u0165", - Tcedil: "\u0162", - tcedil: "\u0163", - Tcy: "\u0422", - tcy: "\u0442", - tdot: "\u20db", - telrec: "\u2315", - Tfr: "\ud835\udd17", - tfr: "\ud835\udd31", - there4: "\u2234", - therefore: "\u2234", - Therefore: "\u2234", - Theta: "\u0398", - theta: "\u03b8", - thetasym: "\u03d1", - thetav: "\u03d1", - thickapprox: "\u2248", - thicksim: "\u223c", - ThickSpace: "\u205f\u200a", - ThinSpace: "\u2009", - thinsp: "\u2009", - thkap: "\u2248", - thksim: "\u223c", - THORN: "\xde", - thorn: "\xfe", - tilde: "\u02dc", - Tilde: "\u223c", - TildeEqual: "\u2243", - TildeFullEqual: "\u2245", - TildeTilde: "\u2248", - timesbar: "\u2a31", - timesb: "\u22a0", - times: "\xd7", - timesd: "\u2a30", - tint: "\u222d", - toea: "\u2928", - topbot: "\u2336", - topcir: "\u2af1", - top: "\u22a4", - Topf: "\ud835\udd4b", - topf: "\ud835\udd65", - topfork: "\u2ada", - tosa: "\u2929", - tprime: "\u2034", - trade: "\u2122", - TRADE: "\u2122", - triangle: "\u25b5", - triangledown: "\u25bf", - triangleleft: "\u25c3", - trianglelefteq: "\u22b4", - triangleq: "\u225c", - triangleright: "\u25b9", - trianglerighteq: "\u22b5", - tridot: "\u25ec", - trie: "\u225c", - triminus: "\u2a3a", - TripleDot: "\u20db", - triplus: "\u2a39", - trisb: "\u29cd", - tritime: "\u2a3b", - trpezium: "\u23e2", - Tscr: "\ud835\udcaf", - tscr: "\ud835\udcc9", - TScy: "\u0426", - tscy: "\u0446", - TSHcy: "\u040b", - tshcy: "\u045b", - Tstrok: "\u0166", - tstrok: "\u0167", - twixt: "\u226c", - twoheadleftarrow: "\u219e", - twoheadrightarrow: "\u21a0", - Uacute: "\xda", - uacute: "\xfa", - uarr: "\u2191", - Uarr: "\u219f", - uArr: "\u21d1", - Uarrocir: "\u2949", - Ubrcy: "\u040e", - ubrcy: "\u045e", - Ubreve: "\u016c", - ubreve: "\u016d", - Ucirc: "\xdb", - ucirc: "\xfb", - Ucy: "\u0423", - ucy: "\u0443", - udarr: "\u21c5", - Udblac: "\u0170", - udblac: "\u0171", - udhar: "\u296e", - ufisht: "\u297e", - Ufr: "\ud835\udd18", - ufr: "\ud835\udd32", - Ugrave: "\xd9", - ugrave: "\xf9", - uHar: "\u2963", - uharl: "\u21bf", - uharr: "\u21be", - uhblk: "\u2580", - ulcorn: "\u231c", - ulcorner: "\u231c", - ulcrop: "\u230f", - ultri: "\u25f8", - Umacr: "\u016a", - umacr: "\u016b", - uml: "\xa8", - UnderBar: "_", - UnderBrace: "\u23df", - UnderBracket: "\u23b5", - UnderParenthesis: "\u23dd", - Union: "\u22c3", - UnionPlus: "\u228e", - Uogon: "\u0172", - uogon: "\u0173", - Uopf: "\ud835\udd4c", - uopf: "\ud835\udd66", - UpArrowBar: "\u2912", - uparrow: "\u2191", - UpArrow: "\u2191", - Uparrow: "\u21d1", - UpArrowDownArrow: "\u21c5", - updownarrow: "\u2195", - UpDownArrow: "\u2195", - Updownarrow: "\u21d5", - UpEquilibrium: "\u296e", - upharpoonleft: "\u21bf", - upharpoonright: "\u21be", - uplus: "\u228e", - UpperLeftArrow: "\u2196", - UpperRightArrow: "\u2197", - upsi: "\u03c5", - Upsi: "\u03d2", - upsih: "\u03d2", - Upsilon: "\u03a5", - upsilon: "\u03c5", - UpTeeArrow: "\u21a5", - UpTee: "\u22a5", - upuparrows: "\u21c8", - urcorn: "\u231d", - urcorner: "\u231d", - urcrop: "\u230e", - Uring: "\u016e", - uring: "\u016f", - urtri: "\u25f9", - Uscr: "\ud835\udcb0", - uscr: "\ud835\udcca", - utdot: "\u22f0", - Utilde: "\u0168", - utilde: "\u0169", - utri: "\u25b5", - utrif: "\u25b4", - uuarr: "\u21c8", - Uuml: "\xdc", - uuml: "\xfc", - uwangle: "\u29a7", - vangrt: "\u299c", - varepsilon: "\u03f5", - varkappa: "\u03f0", - varnothing: "\u2205", - varphi: "\u03d5", - varpi: "\u03d6", - varpropto: "\u221d", - varr: "\u2195", - vArr: "\u21d5", - varrho: "\u03f1", - varsigma: "\u03c2", - varsubsetneq: "\u228a\ufe00", - varsubsetneqq: "\u2acb\ufe00", - varsupsetneq: "\u228b\ufe00", - varsupsetneqq: "\u2acc\ufe00", - vartheta: "\u03d1", - vartriangleleft: "\u22b2", - vartriangleright: "\u22b3", - vBar: "\u2ae8", - Vbar: "\u2aeb", - vBarv: "\u2ae9", - Vcy: "\u0412", - vcy: "\u0432", - vdash: "\u22a2", - vDash: "\u22a8", - Vdash: "\u22a9", - VDash: "\u22ab", - Vdashl: "\u2ae6", - veebar: "\u22bb", - vee: "\u2228", - Vee: "\u22c1", - veeeq: "\u225a", - vellip: "\u22ee", - verbar: "|", - Verbar: "\u2016", - vert: "|", - Vert: "\u2016", - VerticalBar: "\u2223", - VerticalLine: "|", - VerticalSeparator: "\u2758", - VerticalTilde: "\u2240", - VeryThinSpace: "\u200a", - Vfr: "\ud835\udd19", - vfr: "\ud835\udd33", - vltri: "\u22b2", - vnsub: "\u2282\u20d2", - vnsup: "\u2283\u20d2", - Vopf: "\ud835\udd4d", - vopf: "\ud835\udd67", - vprop: "\u221d", - vrtri: "\u22b3", - Vscr: "\ud835\udcb1", - vscr: "\ud835\udccb", - vsubnE: "\u2acb\ufe00", - vsubne: "\u228a\ufe00", - vsupnE: "\u2acc\ufe00", - vsupne: "\u228b\ufe00", - Vvdash: "\u22aa", - vzigzag: "\u299a", - Wcirc: "\u0174", - wcirc: "\u0175", - wedbar: "\u2a5f", - wedge: "\u2227", - Wedge: "\u22c0", - wedgeq: "\u2259", - weierp: "\u2118", - Wfr: "\ud835\udd1a", - wfr: "\ud835\udd34", - Wopf: "\ud835\udd4e", - wopf: "\ud835\udd68", - wp: "\u2118", - wr: "\u2240", - wreath: "\u2240", - Wscr: "\ud835\udcb2", - wscr: "\ud835\udccc", - xcap: "\u22c2", - xcirc: "\u25ef", - xcup: "\u22c3", - xdtri: "\u25bd", - Xfr: "\ud835\udd1b", - xfr: "\ud835\udd35", - xharr: "\u27f7", - xhArr: "\u27fa", - Xi: "\u039e", - xi: "\u03be", - xlarr: "\u27f5", - xlArr: "\u27f8", - xmap: "\u27fc", - xnis: "\u22fb", - xodot: "\u2a00", - Xopf: "\ud835\udd4f", - xopf: "\ud835\udd69", - xoplus: "\u2a01", - xotime: "\u2a02", - xrarr: "\u27f6", - xrArr: "\u27f9", - Xscr: "\ud835\udcb3", - xscr: "\ud835\udccd", - xsqcup: "\u2a06", - xuplus: "\u2a04", - xutri: "\u25b3", - xvee: "\u22c1", - xwedge: "\u22c0", - Yacute: "\xdd", - yacute: "\xfd", - YAcy: "\u042f", - yacy: "\u044f", - Ycirc: "\u0176", - ycirc: "\u0177", - Ycy: "\u042b", - ycy: "\u044b", - yen: "\xa5", - Yfr: "\ud835\udd1c", - yfr: "\ud835\udd36", - YIcy: "\u0407", - yicy: "\u0457", - Yopf: "\ud835\udd50", - yopf: "\ud835\udd6a", - Yscr: "\ud835\udcb4", - yscr: "\ud835\udcce", - YUcy: "\u042e", - yucy: "\u044e", - yuml: "\xff", - Yuml: "\u0178", - Zacute: "\u0179", - zacute: "\u017a", - Zcaron: "\u017d", - zcaron: "\u017e", - Zcy: "\u0417", - zcy: "\u0437", - Zdot: "\u017b", - zdot: "\u017c", - zeetrf: "\u2128", - ZeroWidthSpace: "\u200b", - Zeta: "\u0396", - zeta: "\u03b6", - zfr: "\ud835\udd37", - Zfr: "\u2128", - ZHcy: "\u0416", - zhcy: "\u0436", - zigrarr: "\u21dd", - zopf: "\ud835\udd6b", - Zopf: "\u2124", - Zscr: "\ud835\udcb5", - zscr: "\ud835\udccf", - zwj: "\u200d", - zwnj: "\u200c" - }; - /*eslint quotes:0*/ var entities = require$$0; - var regex$4 = /[!-#%-\*,-\/:;\?@\[-\]_\{\}\xA1\xA7\xAB\xB6\xB7\xBB\xBF\u037E\u0387\u055A-\u055F\u0589\u058A\u05BE\u05C0\u05C3\u05C6\u05F3\u05F4\u0609\u060A\u060C\u060D\u061B\u061E\u061F\u066A-\u066D\u06D4\u0700-\u070D\u07F7-\u07F9\u0830-\u083E\u085E\u0964\u0965\u0970\u09FD\u0A76\u0AF0\u0C84\u0DF4\u0E4F\u0E5A\u0E5B\u0F04-\u0F12\u0F14\u0F3A-\u0F3D\u0F85\u0FD0-\u0FD4\u0FD9\u0FDA\u104A-\u104F\u10FB\u1360-\u1368\u1400\u166D\u166E\u169B\u169C\u16EB-\u16ED\u1735\u1736\u17D4-\u17D6\u17D8-\u17DA\u1800-\u180A\u1944\u1945\u1A1E\u1A1F\u1AA0-\u1AA6\u1AA8-\u1AAD\u1B5A-\u1B60\u1BFC-\u1BFF\u1C3B-\u1C3F\u1C7E\u1C7F\u1CC0-\u1CC7\u1CD3\u2010-\u2027\u2030-\u2043\u2045-\u2051\u2053-\u205E\u207D\u207E\u208D\u208E\u2308-\u230B\u2329\u232A\u2768-\u2775\u27C5\u27C6\u27E6-\u27EF\u2983-\u2998\u29D8-\u29DB\u29FC\u29FD\u2CF9-\u2CFC\u2CFE\u2CFF\u2D70\u2E00-\u2E2E\u2E30-\u2E4E\u3001-\u3003\u3008-\u3011\u3014-\u301F\u3030\u303D\u30A0\u30FB\uA4FE\uA4FF\uA60D-\uA60F\uA673\uA67E\uA6F2-\uA6F7\uA874-\uA877\uA8CE\uA8CF\uA8F8-\uA8FA\uA8FC\uA92E\uA92F\uA95F\uA9C1-\uA9CD\uA9DE\uA9DF\uAA5C-\uAA5F\uAADE\uAADF\uAAF0\uAAF1\uABEB\uFD3E\uFD3F\uFE10-\uFE19\uFE30-\uFE52\uFE54-\uFE61\uFE63\uFE68\uFE6A\uFE6B\uFF01-\uFF03\uFF05-\uFF0A\uFF0C-\uFF0F\uFF1A\uFF1B\uFF1F\uFF20\uFF3B-\uFF3D\uFF3F\uFF5B\uFF5D\uFF5F-\uFF65]|\uD800[\uDD00-\uDD02\uDF9F\uDFD0]|\uD801\uDD6F|\uD802[\uDC57\uDD1F\uDD3F\uDE50-\uDE58\uDE7F\uDEF0-\uDEF6\uDF39-\uDF3F\uDF99-\uDF9C]|\uD803[\uDF55-\uDF59]|\uD804[\uDC47-\uDC4D\uDCBB\uDCBC\uDCBE-\uDCC1\uDD40-\uDD43\uDD74\uDD75\uDDC5-\uDDC8\uDDCD\uDDDB\uDDDD-\uDDDF\uDE38-\uDE3D\uDEA9]|\uD805[\uDC4B-\uDC4F\uDC5B\uDC5D\uDCC6\uDDC1-\uDDD7\uDE41-\uDE43\uDE60-\uDE6C\uDF3C-\uDF3E]|\uD806[\uDC3B\uDE3F-\uDE46\uDE9A-\uDE9C\uDE9E-\uDEA2]|\uD807[\uDC41-\uDC45\uDC70\uDC71\uDEF7\uDEF8]|\uD809[\uDC70-\uDC74]|\uD81A[\uDE6E\uDE6F\uDEF5\uDF37-\uDF3B\uDF44]|\uD81B[\uDE97-\uDE9A]|\uD82F\uDC9F|\uD836[\uDE87-\uDE8B]|\uD83A[\uDD5E\uDD5F]/; - var encodeCache = {}; - // Create a lookup array where anything but characters in `chars` string - // and alphanumeric chars is percent-encoded. - - function getEncodeCache(exclude) { - var i, ch, cache = encodeCache[exclude]; - if (cache) { - return cache; - } - cache = encodeCache[exclude] = []; - for (i = 0; i < 128; i++) { - ch = String.fromCharCode(i); - if (/^[0-9a-z]$/i.test(ch)) { - // always allow unencoded alphanumeric characters - cache.push(ch); - } else { - cache.push("%" + ("0" + i.toString(16).toUpperCase()).slice(-2)); - } - } - for (i = 0; i < exclude.length; i++) { - cache[exclude.charCodeAt(i)] = exclude[i]; - } - return cache; - } - // Encode unsafe characters with percent-encoding, skipping already - // encoded sequences. - - // - string - string to encode - // - exclude - list of characters to ignore (in addition to a-zA-Z0-9) - // - keepEscaped - don't encode '%' in a correct escape sequence (default: true) - - function encode$2(string, exclude, keepEscaped) { - var i, l, code, nextCode, cache, result = ""; - if (typeof exclude !== "string") { - // encode(string, keepEscaped) - keepEscaped = exclude; - exclude = encode$2.defaultChars; - } - if (typeof keepEscaped === "undefined") { - keepEscaped = true; - } - cache = getEncodeCache(exclude); - for (i = 0, l = string.length; i < l; i++) { - code = string.charCodeAt(i); - if (keepEscaped && code === 37 /* % */ && i + 2 < l) { - if (/^[0-9a-f]{2}$/i.test(string.slice(i + 1, i + 3))) { - result += string.slice(i, i + 3); - i += 2; - continue; - } - } - if (code < 128) { - result += cache[code]; - continue; - } - if (code >= 55296 && code <= 57343) { - if (code >= 55296 && code <= 56319 && i + 1 < l) { - nextCode = string.charCodeAt(i + 1); - if (nextCode >= 56320 && nextCode <= 57343) { - result += encodeURIComponent(string[i] + string[i + 1]); - i++; - continue; - } - } - result += "%EF%BF%BD"; - continue; - } - result += encodeURIComponent(string[i]); - } - return result; - } - encode$2.defaultChars = ";/?:@&=+$,-_.!~*'()#"; - encode$2.componentChars = "-_.!~*'()"; - var encode_1 = encode$2; - /* eslint-disable no-bitwise */ var decodeCache = {}; - function getDecodeCache(exclude) { - var i, ch, cache = decodeCache[exclude]; - if (cache) { - return cache; - } - cache = decodeCache[exclude] = []; - for (i = 0; i < 128; i++) { - ch = String.fromCharCode(i); - cache.push(ch); - } - for (i = 0; i < exclude.length; i++) { - ch = exclude.charCodeAt(i); - cache[ch] = "%" + ("0" + ch.toString(16).toUpperCase()).slice(-2); - } - return cache; - } - // Decode percent-encoded string. - - function decode$2(string, exclude) { - var cache; - if (typeof exclude !== "string") { - exclude = decode$2.defaultChars; - } - cache = getDecodeCache(exclude); - return string.replace(/(%[a-f0-9]{2})+/gi, (function(seq) { - var i, l, b1, b2, b3, b4, chr, result = ""; - for (i = 0, l = seq.length; i < l; i += 3) { - b1 = parseInt(seq.slice(i + 1, i + 3), 16); - if (b1 < 128) { - result += cache[b1]; - continue; - } - if ((b1 & 224) === 192 && i + 3 < l) { - // 110xxxxx 10xxxxxx - b2 = parseInt(seq.slice(i + 4, i + 6), 16); - if ((b2 & 192) === 128) { - chr = b1 << 6 & 1984 | b2 & 63; - if (chr < 128) { - result += "\ufffd\ufffd"; - } else { - result += String.fromCharCode(chr); - } - i += 3; - continue; - } - } - if ((b1 & 240) === 224 && i + 6 < l) { - // 1110xxxx 10xxxxxx 10xxxxxx - b2 = parseInt(seq.slice(i + 4, i + 6), 16); - b3 = parseInt(seq.slice(i + 7, i + 9), 16); - if ((b2 & 192) === 128 && (b3 & 192) === 128) { - chr = b1 << 12 & 61440 | b2 << 6 & 4032 | b3 & 63; - if (chr < 2048 || chr >= 55296 && chr <= 57343) { - result += "\ufffd\ufffd\ufffd"; - } else { - result += String.fromCharCode(chr); - } - i += 6; - continue; - } - } - if ((b1 & 248) === 240 && i + 9 < l) { - // 111110xx 10xxxxxx 10xxxxxx 10xxxxxx - b2 = parseInt(seq.slice(i + 4, i + 6), 16); - b3 = parseInt(seq.slice(i + 7, i + 9), 16); - b4 = parseInt(seq.slice(i + 10, i + 12), 16); - if ((b2 & 192) === 128 && (b3 & 192) === 128 && (b4 & 192) === 128) { - chr = b1 << 18 & 1835008 | b2 << 12 & 258048 | b3 << 6 & 4032 | b4 & 63; - if (chr < 65536 || chr > 1114111) { - result += "\ufffd\ufffd\ufffd\ufffd"; - } else { - chr -= 65536; - result += String.fromCharCode(55296 + (chr >> 10), 56320 + (chr & 1023)); - } - i += 9; - continue; - } - } - result += "\ufffd"; - } - return result; - })); - } - decode$2.defaultChars = ";/?:@&=+$,#"; - decode$2.componentChars = ""; - var decode_1 = decode$2; - var format$1 = function format(url) { - var result = ""; - result += url.protocol || ""; - result += url.slashes ? "//" : ""; - result += url.auth ? url.auth + "@" : ""; - if (url.hostname && url.hostname.indexOf(":") !== -1) { - // ipv6 address - result += "[" + url.hostname + "]"; - } else { - result += url.hostname || ""; - } - result += url.port ? ":" + url.port : ""; - result += url.pathname || ""; - result += url.search || ""; - result += url.hash || ""; - return result; - }; - // Copyright Joyent, Inc. and other Node contributors. - - // Changes from joyent/node: - - // 1. No leading slash in paths, - // e.g. in `url.parse('http://foo?bar')` pathname is ``, not `/` - - // 2. Backslashes are not replaced with slashes, - // so `http:\\example.org\` is treated like a relative path - - // 3. Trailing colon is treated like a part of the path, - // i.e. in `http://example.org:foo` pathname is `:foo` - - // 4. Nothing is URL-encoded in the resulting object, - // (in joyent/node some chars in auth and paths are encoded) - - // 5. `url.parse()` does not have `parseQueryString` argument - - // 6. Removed extraneous result properties: `host`, `path`, `query`, etc., - // which can be constructed using other parts of the url. - - function Url() { - this.protocol = null; - this.slashes = null; - this.auth = null; - this.port = null; - this.hostname = null; - this.hash = null; - this.search = null; - this.pathname = null; - } - // Reference: RFC 3986, RFC 1808, RFC 2396 - // define these here so at least they only have to be - // compiled once on the first module load. - var protocolPattern = /^([a-z0-9.+-]+:)/i, portPattern = /:[0-9]*$/, - // Special case for a simple path URL - simplePathPattern = /^(\/\/?(?!\/)[^\?\s]*)(\?[^\s]*)?$/, - // RFC 2396: characters reserved for delimiting URLs. - // We actually just auto-escape these. - delims = [ "<", ">", '"', "`", " ", "\r", "\n", "\t" ], - // RFC 2396: characters not allowed for various reasons. - unwise = [ "{", "}", "|", "\\", "^", "`" ].concat(delims), - // Allowed by RFCs, but cause of XSS attacks. Always escape these. - autoEscape = [ "'" ].concat(unwise), - // Characters that are never ever allowed in a hostname. - // Note that any invalid chars are also handled, but these - // are the ones that are *expected* to be seen, so we fast-path - // them. - nonHostChars = [ "%", "/", "?", ";", "#" ].concat(autoEscape), hostEndingChars = [ "/", "?", "#" ], hostnameMaxLen = 255, hostnamePartPattern = /^[+a-z0-9A-Z_-]{0,63}$/, hostnamePartStart = /^([+a-z0-9A-Z_-]{0,63})(.*)$/, - // protocols that can allow "unsafe" and "unwise" chars. - /* eslint-disable no-script-url */ - // protocols that never have a hostname. - hostlessProtocol = { - javascript: true, - "javascript:": true - }, - // protocols that always contain a // bit. - slashedProtocol = { - http: true, - https: true, - ftp: true, - gopher: true, - file: true, - "http:": true, - "https:": true, - "ftp:": true, - "gopher:": true, - "file:": true - }; - /* eslint-enable no-script-url */ function urlParse(url, slashesDenoteHost) { - if (url && url instanceof Url) { - return url; - } - var u = new Url; - u.parse(url, slashesDenoteHost); - return u; - } - Url.prototype.parse = function(url, slashesDenoteHost) { - var i, l, lowerProto, hec, slashes, rest = url; - // trim before proceeding. - // This is to support parse stuff like " http://foo.com \n" - rest = rest.trim(); - if (!slashesDenoteHost && url.split("#").length === 1) { - // Try fast path regexp - var simplePath = simplePathPattern.exec(rest); - if (simplePath) { - this.pathname = simplePath[1]; - if (simplePath[2]) { - this.search = simplePath[2]; - } - return this; - } - } - var proto = protocolPattern.exec(rest); - if (proto) { - proto = proto[0]; - lowerProto = proto.toLowerCase(); - this.protocol = proto; - rest = rest.substr(proto.length); - } - // figure out if it's got a host - // user@server is *always* interpreted as a hostname, and url - // resolution will treat //foo/bar as host=foo,path=bar because that's - // how the browser resolves relative URLs. - if (slashesDenoteHost || proto || rest.match(/^\/\/[^@\/]+@[^@\/]+/)) { - slashes = rest.substr(0, 2) === "//"; - if (slashes && !(proto && hostlessProtocol[proto])) { - rest = rest.substr(2); - this.slashes = true; - } - } - if (!hostlessProtocol[proto] && (slashes || proto && !slashedProtocol[proto])) { - // there's a hostname. - // the first instance of /, ?, ;, or # ends the host. - // If there is an @ in the hostname, then non-host chars *are* allowed - // to the left of the last @ sign, unless some host-ending character - // comes *before* the @-sign. - // URLs are obnoxious. - // ex: - // http://a@b@c/ => user:a@b host:c - // http://a@b?@c => user:a host:c path:/?@c - // v0.12 TODO(isaacs): This is not quite how Chrome does things. - // Review our test case against browsers more comprehensively. - // find the first instance of any hostEndingChars - var hostEnd = -1; - for (i = 0; i < hostEndingChars.length; i++) { - hec = rest.indexOf(hostEndingChars[i]); - if (hec !== -1 && (hostEnd === -1 || hec < hostEnd)) { - hostEnd = hec; - } - } - // at this point, either we have an explicit point where the - // auth portion cannot go past, or the last @ char is the decider. - var auth, atSign; - if (hostEnd === -1) { - // atSign can be anywhere. - atSign = rest.lastIndexOf("@"); - } else { - // atSign must be in auth portion. - // http://a@b/c@d => host:b auth:a path:/c@d - atSign = rest.lastIndexOf("@", hostEnd); - } - // Now we have a portion which is definitely the auth. - // Pull that off. - if (atSign !== -1) { - auth = rest.slice(0, atSign); - rest = rest.slice(atSign + 1); - this.auth = auth; - } - // the host is the remaining to the left of the first non-host char - hostEnd = -1; - for (i = 0; i < nonHostChars.length; i++) { - hec = rest.indexOf(nonHostChars[i]); - if (hec !== -1 && (hostEnd === -1 || hec < hostEnd)) { - hostEnd = hec; - } - } - // if we still have not hit it, then the entire thing is a host. - if (hostEnd === -1) { - hostEnd = rest.length; - } - if (rest[hostEnd - 1] === ":") { - hostEnd--; - } - var host = rest.slice(0, hostEnd); - rest = rest.slice(hostEnd); - // pull out port. - this.parseHost(host); - // we've indicated that there is a hostname, - // so even if it's empty, it has to be present. - this.hostname = this.hostname || ""; - // if hostname begins with [ and ends with ] - // assume that it's an IPv6 address. - var ipv6Hostname = this.hostname[0] === "[" && this.hostname[this.hostname.length - 1] === "]"; - // validate a little. - if (!ipv6Hostname) { - var hostparts = this.hostname.split(/\./); - for (i = 0, l = hostparts.length; i < l; i++) { - var part = hostparts[i]; - if (!part) { - continue; - } - if (!part.match(hostnamePartPattern)) { - var newpart = ""; - for (var j = 0, k = part.length; j < k; j++) { - if (part.charCodeAt(j) > 127) { - // we replace non-ASCII char with a temporary placeholder - // we need this to make sure size of hostname is not - // broken by replacing non-ASCII by nothing - newpart += "x"; - } else { - newpart += part[j]; - } - } - // we test again with ASCII char only - if (!newpart.match(hostnamePartPattern)) { - var validParts = hostparts.slice(0, i); - var notHost = hostparts.slice(i + 1); - var bit = part.match(hostnamePartStart); - if (bit) { - validParts.push(bit[1]); - notHost.unshift(bit[2]); - } - if (notHost.length) { - rest = notHost.join(".") + rest; - } - this.hostname = validParts.join("."); - break; - } - } - } - } - if (this.hostname.length > hostnameMaxLen) { - this.hostname = ""; - } - // strip [ and ] from the hostname - // the host field still retains them, though - if (ipv6Hostname) { - this.hostname = this.hostname.substr(1, this.hostname.length - 2); - } - } - // chop off from the tail first. - var hash = rest.indexOf("#"); - if (hash !== -1) { - // got a fragment string. - this.hash = rest.substr(hash); - rest = rest.slice(0, hash); - } - var qm = rest.indexOf("?"); - if (qm !== -1) { - this.search = rest.substr(qm); - rest = rest.slice(0, qm); - } - if (rest) { - this.pathname = rest; - } - if (slashedProtocol[lowerProto] && this.hostname && !this.pathname) { - this.pathname = ""; - } - return this; - }; - Url.prototype.parseHost = function(host) { - var port = portPattern.exec(host); - if (port) { - port = port[0]; - if (port !== ":") { - this.port = port.substr(1); - } - host = host.substr(0, host.length - port.length); - } - if (host) { - this.hostname = host; - } - }; - var parse$1 = urlParse; - var encode$1 = encode_1; - var decode$1 = decode_1; - var format = format$1; - var parse = parse$1; - var mdurl = { - encode: encode$1, - decode: decode$1, - format: format, - parse: parse - }; - var regex$3 = /[\0-\uD7FF\uE000-\uFFFF]|[\uD800-\uDBFF][\uDC00-\uDFFF]|[\uD800-\uDBFF](?![\uDC00-\uDFFF])|(?:[^\uD800-\uDBFF]|^)[\uDC00-\uDFFF]/; - var regex$2 = /[\0-\x1F\x7F-\x9F]/; - var regex$1 = /[\xAD\u0600-\u0605\u061C\u06DD\u070F\u08E2\u180E\u200B-\u200F\u202A-\u202E\u2060-\u2064\u2066-\u206F\uFEFF\uFFF9-\uFFFB]|\uD804[\uDCBD\uDCCD]|\uD82F[\uDCA0-\uDCA3]|\uD834[\uDD73-\uDD7A]|\uDB40[\uDC01\uDC20-\uDC7F]/; - var regex = /[ \xA0\u1680\u2000-\u200A\u2028\u2029\u202F\u205F\u3000]/; - var Any = regex$3; - var Cc = regex$2; - var Cf = regex$1; - var P = regex$4; - var Z = regex; - var uc_micro = { - Any: Any, - Cc: Cc, - Cf: Cf, - P: P, - Z: Z - }; - var utils = createCommonjsModule((function(module, exports) { - function _class(obj) { - return Object.prototype.toString.call(obj); - } - function isString(obj) { - return _class(obj) === "[object String]"; - } - var _hasOwnProperty = Object.prototype.hasOwnProperty; - function has(object, key) { - return _hasOwnProperty.call(object, key); - } - // Merge objects - - function assign(obj /*from1, from2, from3, ...*/) { - var sources = Array.prototype.slice.call(arguments, 1); - sources.forEach((function(source) { - if (!source) { - return; - } - if (typeof source !== "object") { - throw new TypeError(source + "must be object"); - } - Object.keys(source).forEach((function(key) { - obj[key] = source[key]; - })); - })); - return obj; - } - // Remove element from array and put another array at those position. - // Useful for some operations with tokens - function arrayReplaceAt(src, pos, newElements) { - return [].concat(src.slice(0, pos), newElements, src.slice(pos + 1)); - } - //////////////////////////////////////////////////////////////////////////////// - function isValidEntityCode(c) { - /*eslint no-bitwise:0*/ - // broken sequence - if (c >= 55296 && c <= 57343) { - return false; - } - // never used - if (c >= 64976 && c <= 65007) { - return false; - } - if ((c & 65535) === 65535 || (c & 65535) === 65534) { - return false; - } - // control codes - if (c >= 0 && c <= 8) { - return false; - } - if (c === 11) { - return false; - } - if (c >= 14 && c <= 31) { - return false; - } - if (c >= 127 && c <= 159) { - return false; - } - // out of range - if (c > 1114111) { - return false; - } - return true; - } - function fromCodePoint(c) { - /*eslint no-bitwise:0*/ - if (c > 65535) { - c -= 65536; - var surrogate1 = 55296 + (c >> 10), surrogate2 = 56320 + (c & 1023); - return String.fromCharCode(surrogate1, surrogate2); - } - return String.fromCharCode(c); - } - var UNESCAPE_MD_RE = /\\([!"#$%&'()*+,\-.\/:;<=>?@[\\\]^_`{|}~])/g; - var ENTITY_RE = /&([a-z#][a-z0-9]{1,31});/gi; - var UNESCAPE_ALL_RE = new RegExp(UNESCAPE_MD_RE.source + "|" + ENTITY_RE.source, "gi"); - var DIGITAL_ENTITY_TEST_RE = /^#((?:x[a-f0-9]{1,8}|[0-9]{1,8}))$/i; - function replaceEntityPattern(match, name) { - var code; - if (has(entities, name)) { - return entities[name]; - } - if (name.charCodeAt(0) === 35 /* # */ && DIGITAL_ENTITY_TEST_RE.test(name)) { - code = name[1].toLowerCase() === "x" ? parseInt(name.slice(2), 16) : parseInt(name.slice(1), 10); - if (isValidEntityCode(code)) { - return fromCodePoint(code); - } - } - return match; - } - /*function replaceEntities(str) { - if (str.indexOf('&') < 0) { return str; } - - return str.replace(ENTITY_RE, replaceEntityPattern); - }*/ function unescapeMd(str) { - if (str.indexOf("\\") < 0) { - return str; - } - return str.replace(UNESCAPE_MD_RE, "$1"); - } - function unescapeAll(str) { - if (str.indexOf("\\") < 0 && str.indexOf("&") < 0) { - return str; - } - return str.replace(UNESCAPE_ALL_RE, (function(match, escaped, entity) { - if (escaped) { - return escaped; - } - return replaceEntityPattern(match, entity); - })); - } - //////////////////////////////////////////////////////////////////////////////// - var HTML_ESCAPE_TEST_RE = /[&<>"]/; - var HTML_ESCAPE_REPLACE_RE = /[&<>"]/g; - var HTML_REPLACEMENTS = { - "&": "&", - "<": "<", - ">": ">", - '"': """ - }; - function replaceUnsafeChar(ch) { - return HTML_REPLACEMENTS[ch]; - } - function escapeHtml(str) { - if (HTML_ESCAPE_TEST_RE.test(str)) { - return str.replace(HTML_ESCAPE_REPLACE_RE, replaceUnsafeChar); - } - return str; - } - //////////////////////////////////////////////////////////////////////////////// - var REGEXP_ESCAPE_RE = /[.?*+^$[\]\\(){}|-]/g; - function escapeRE(str) { - return str.replace(REGEXP_ESCAPE_RE, "\\$&"); - } - //////////////////////////////////////////////////////////////////////////////// - function isSpace(code) { - switch (code) { - case 9: - case 32: - return true; - } - return false; - } - // Zs (unicode class) || [\t\f\v\r\n] - function isWhiteSpace(code) { - if (code >= 8192 && code <= 8202) { - return true; - } - switch (code) { - case 9: - // \t - case 10: - // \n - case 11: - // \v - case 12: - // \f - case 13: - // \r - case 32: - case 160: - case 5760: - case 8239: - case 8287: - case 12288: - return true; - } - return false; - } - //////////////////////////////////////////////////////////////////////////////// - /*eslint-disable max-len*/ - // Currently without astral characters support. - function isPunctChar(ch) { - return regex$4.test(ch); - } - // Markdown ASCII punctuation characters. - - // !, ", #, $, %, &, ', (, ), *, +, ,, -, ., /, :, ;, <, =, >, ?, @, [, \, ], ^, _, `, {, |, }, or ~ - // http://spec.commonmark.org/0.15/#ascii-punctuation-character - - // Don't confuse with unicode punctuation !!! It lacks some chars in ascii range. - - function isMdAsciiPunct(ch) { - switch (ch) { - case 33 /* ! */ : - case 34 /* " */ : - case 35 /* # */ : - case 36 /* $ */ : - case 37 /* % */ : - case 38 /* & */ : - case 39 /* ' */ : - case 40 /* ( */ : - case 41 /* ) */ : - case 42 /* * */ : - case 43 /* + */ : - case 44 /* , */ : - case 45 /* - */ : - case 46 /* . */ : - case 47 /* / */ : - case 58 /* : */ : - case 59 /* ; */ : - case 60 /* < */ : - case 61 /* = */ : - case 62 /* > */ : - case 63 /* ? */ : - case 64 /* @ */ : - case 91 /* [ */ : - case 92 /* \ */ : - case 93 /* ] */ : - case 94 /* ^ */ : - case 95 /* _ */ : - case 96 /* ` */ : - case 123 /* { */ : - case 124 /* | */ : - case 125 /* } */ : - case 126 /* ~ */ : - return true; - - default: - return false; - } - } - // Hepler to unify [reference labels]. - - function normalizeReference(str) { - // Trim and collapse whitespace - str = str.trim().replace(/\s+/g, " "); - // In node v10 'ẞ'.toLowerCase() === 'Ṿ', which is presumed to be a bug - // fixed in v12 (couldn't find any details). - - // So treat this one as a special case - // (remove this when node v10 is no longer supported). - - if ("\u1e9e".toLowerCase() === "\u1e7e") { - str = str.replace(/\u1e9e/g, "\xdf"); - } - // .toLowerCase().toUpperCase() should get rid of all differences - // between letter variants. - - // Simple .toLowerCase() doesn't normalize 125 code points correctly, - // and .toUpperCase doesn't normalize 6 of them (list of exceptions: - // İ, ϴ, ẞ, Ω, K, Å - those are already uppercased, but have differently - // uppercased versions). - - // Here's an example showing how it happens. Lets take greek letter omega: - // uppercase U+0398 (Θ), U+03f4 (ϴ) and lowercase U+03b8 (θ), U+03d1 (ϑ) - - // Unicode entries: - // 0398;GREEK CAPITAL LETTER THETA;Lu;0;L;;;;;N;;;;03B8; - // 03B8;GREEK SMALL LETTER THETA;Ll;0;L;;;;;N;;;0398;;0398 - // 03D1;GREEK THETA SYMBOL;Ll;0;L; 03B8;;;;N;GREEK SMALL LETTER SCRIPT THETA;;0398;;0398 - // 03F4;GREEK CAPITAL THETA SYMBOL;Lu;0;L; 0398;;;;N;;;;03B8; - - // Case-insensitive comparison should treat all of them as equivalent. - - // But .toLowerCase() doesn't change ϑ (it's already lowercase), - // and .toUpperCase() doesn't change ϴ (already uppercase). - - // Applying first lower then upper case normalizes any character: - // '\u0398\u03f4\u03b8\u03d1'.toLowerCase().toUpperCase() === '\u0398\u0398\u0398\u0398' - - // Note: this is equivalent to unicode case folding; unicode normalization - // is a different step that is not required here. - - // Final result should be uppercased, because it's later stored in an object - // (this avoid a conflict with Object.prototype members, - // most notably, `__proto__`) - - return str.toLowerCase().toUpperCase(); - } - //////////////////////////////////////////////////////////////////////////////// - // Re-export libraries commonly used in both markdown-it and its plugins, - // so plugins won't have to depend on them explicitly, which reduces their - // bundled size (e.g. a browser build). - - exports.lib = {}; - exports.lib.mdurl = mdurl; - exports.lib.ucmicro = uc_micro; - exports.assign = assign; - exports.isString = isString; - exports.has = has; - exports.unescapeMd = unescapeMd; - exports.unescapeAll = unescapeAll; - exports.isValidEntityCode = isValidEntityCode; - exports.fromCodePoint = fromCodePoint; - // exports.replaceEntities = replaceEntities; - exports.escapeHtml = escapeHtml; - exports.arrayReplaceAt = arrayReplaceAt; - exports.isSpace = isSpace; - exports.isWhiteSpace = isWhiteSpace; - exports.isMdAsciiPunct = isMdAsciiPunct; - exports.isPunctChar = isPunctChar; - exports.escapeRE = escapeRE; - exports.normalizeReference = normalizeReference; - })); - // Parse link label - var parse_link_label = function parseLinkLabel(state, start, disableNested) { - var level, found, marker, prevPos, labelEnd = -1, max = state.posMax, oldPos = state.pos; - state.pos = start + 1; - level = 1; - while (state.pos < max) { - marker = state.src.charCodeAt(state.pos); - if (marker === 93 /* ] */) { - level--; - if (level === 0) { - found = true; - break; - } - } - prevPos = state.pos; - state.md.inline.skipToken(state); - if (marker === 91 /* [ */) { - if (prevPos === state.pos - 1) { - // increase level if we find text `[`, which is not a part of any token - level++; - } else if (disableNested) { - state.pos = oldPos; - return -1; - } - } - } - if (found) { - labelEnd = state.pos; - } - // restore old state - state.pos = oldPos; - return labelEnd; - }; - var unescapeAll$2 = utils.unescapeAll; - var parse_link_destination = function parseLinkDestination(str, start, max) { - var code, level, pos = start, result = { - ok: false, - pos: 0, - lines: 0, - str: "" - }; - if (str.charCodeAt(pos) === 60 /* < */) { - pos++; - while (pos < max) { - code = str.charCodeAt(pos); - if (code === 10 /* \n */) { - return result; - } - if (code === 60 /* < */) { - return result; - } - if (code === 62 /* > */) { - result.pos = pos + 1; - result.str = unescapeAll$2(str.slice(start + 1, pos)); - result.ok = true; - return result; - } - if (code === 92 /* \ */ && pos + 1 < max) { - pos += 2; - continue; - } - pos++; - } - // no closing '>' - return result; - } - // this should be ... } else { ... branch - level = 0; - while (pos < max) { - code = str.charCodeAt(pos); - if (code === 32) { - break; - } - // ascii control characters - if (code < 32 || code === 127) { - break; - } - if (code === 92 /* \ */ && pos + 1 < max) { - if (str.charCodeAt(pos + 1) === 32) { - break; - } - pos += 2; - continue; - } - if (code === 40 /* ( */) { - level++; - if (level > 32) { - return result; - } - } - if (code === 41 /* ) */) { - if (level === 0) { - break; - } - level--; - } - pos++; - } - if (start === pos) { - return result; - } - if (level !== 0) { - return result; - } - result.str = unescapeAll$2(str.slice(start, pos)); - result.pos = pos; - result.ok = true; - return result; - }; - var unescapeAll$1 = utils.unescapeAll; - var parse_link_title = function parseLinkTitle(str, start, max) { - var code, marker, lines = 0, pos = start, result = { - ok: false, - pos: 0, - lines: 0, - str: "" - }; - if (pos >= max) { - return result; - } - marker = str.charCodeAt(pos); - if (marker !== 34 /* " */ && marker !== 39 /* ' */ && marker !== 40 /* ( */) { - return result; - } - pos++; - // if opening marker is "(", switch it to closing marker ")" - if (marker === 40) { - marker = 41; - } - while (pos < max) { - code = str.charCodeAt(pos); - if (code === marker) { - result.pos = pos + 1; - result.lines = lines; - result.str = unescapeAll$1(str.slice(start + 1, pos)); - result.ok = true; - return result; - } else if (code === 40 /* ( */ && marker === 41 /* ) */) { - return result; - } else if (code === 10) { - lines++; - } else if (code === 92 /* \ */ && pos + 1 < max) { - pos++; - if (str.charCodeAt(pos) === 10) { - lines++; - } - } - pos++; - } - return result; - }; - var parseLinkLabel = parse_link_label; - var parseLinkDestination = parse_link_destination; - var parseLinkTitle = parse_link_title; - var helpers = { - parseLinkLabel: parseLinkLabel, - parseLinkDestination: parseLinkDestination, - parseLinkTitle: parseLinkTitle - }; - var assign$1 = utils.assign; - var unescapeAll = utils.unescapeAll; - var escapeHtml = utils.escapeHtml; - //////////////////////////////////////////////////////////////////////////////// - var default_rules = {}; - default_rules.code_inline = function(tokens, idx, options, env, slf) { - var token = tokens[idx]; - return "" + escapeHtml(token.content) + ""; - }; - default_rules.code_block = function(tokens, idx, options, env, slf) { - var token = tokens[idx]; - return "" + escapeHtml(tokens[idx].content) + "\n"; - }; - default_rules.fence = function(tokens, idx, options, env, slf) { - var token = tokens[idx], info = token.info ? unescapeAll(token.info).trim() : "", langName = "", langAttrs = "", highlighted, i, arr, tmpAttrs, tmpToken; - if (info) { - arr = info.split(/(\s+)/g); - langName = arr[0]; - langAttrs = arr.slice(2).join(""); - } - if (options.highlight) { - highlighted = options.highlight(token.content, langName, langAttrs) || escapeHtml(token.content); - } else { - highlighted = escapeHtml(token.content); - } - if (highlighted.indexOf("" + highlighted + "\n"; - } - return "
" + highlighted + "
\n"; - }; - default_rules.image = function(tokens, idx, options, env, slf) { - var token = tokens[idx]; - // "alt" attr MUST be set, even if empty. Because it's mandatory and - // should be placed on proper position for tests. - - // Replace content with actual value - token.attrs[token.attrIndex("alt")][1] = slf.renderInlineAsText(token.children, options, env); - return slf.renderToken(tokens, idx, options); - }; - default_rules.hardbreak = function(tokens, idx, options /*, env */) { - return options.xhtmlOut ? "
\n" : "
\n"; - }; - default_rules.softbreak = function(tokens, idx, options /*, env */) { - return options.breaks ? options.xhtmlOut ? "
\n" : "
\n" : "\n"; - }; - default_rules.text = function(tokens, idx /*, options, env */) { - return escapeHtml(tokens[idx].content); - }; - default_rules.html_block = function(tokens, idx /*, options, env */) { - return tokens[idx].content; - }; - default_rules.html_inline = function(tokens, idx /*, options, env */) { - return tokens[idx].content; - }; - /** - * new Renderer() - * - * Creates new [[Renderer]] instance and fill [[Renderer#rules]] with defaults. - **/ function Renderer() { - /** - * Renderer#rules -> Object - * - * Contains render rules for tokens. Can be updated and extended. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')(); - * - * md.renderer.rules.strong_open = function () { return ''; }; - * md.renderer.rules.strong_close = function () { return ''; }; - * - * var result = md.renderInline(...); - * ``` - * - * Each rule is called as independent static function with fixed signature: - * - * ```javascript - * function my_token_render(tokens, idx, options, env, renderer) { - * // ... - * return renderedHTML; - * } - * ``` - * - * See [source code](https://github.com/markdown-it/markdown-it/blob/master/lib/renderer.js) - * for more details and examples. - **/ - this.rules = assign$1({}, default_rules); - } - /** - * Renderer.renderAttrs(token) -> String - * - * Render token attributes to string. - **/ Renderer.prototype.renderAttrs = function renderAttrs(token) { - var i, l, result; - if (!token.attrs) { - return ""; - } - result = ""; - for (i = 0, l = token.attrs.length; i < l; i++) { - result += " " + escapeHtml(token.attrs[i][0]) + '="' + escapeHtml(token.attrs[i][1]) + '"'; - } - return result; - }; - /** - * Renderer.renderToken(tokens, idx, options) -> String - * - tokens (Array): list of tokens - * - idx (Numbed): token index to render - * - options (Object): params of parser instance - * - * Default token renderer. Can be overriden by custom function - * in [[Renderer#rules]]. - **/ Renderer.prototype.renderToken = function renderToken(tokens, idx, options) { - var nextToken, result = "", needLf = false, token = tokens[idx]; - // Tight list paragraphs - if (token.hidden) { - return ""; - } - // Insert a newline between hidden paragraph and subsequent opening - // block-level tag. - - // For example, here we should insert a newline before blockquote: - // - a - // > - - if (token.block && token.nesting !== -1 && idx && tokens[idx - 1].hidden) { - result += "\n"; - } - // Add token name, e.g. ``. - needLf = false; - } - } - } - } - result += needLf ? ">\n" : ">"; - return result; - }; - /** - * Renderer.renderInline(tokens, options, env) -> String - * - tokens (Array): list on block tokens to render - * - options (Object): params of parser instance - * - env (Object): additional data from parsed input (references, for example) - * - * The same as [[Renderer.render]], but for single token of `inline` type. - **/ Renderer.prototype.renderInline = function(tokens, options, env) { - var type, result = "", rules = this.rules; - for (var i = 0, len = tokens.length; i < len; i++) { - type = tokens[i].type; - if (typeof rules[type] !== "undefined") { - result += rules[type](tokens, i, options, env, this); - } else { - result += this.renderToken(tokens, i, options); - } - } - return result; - }; - /** internal - * Renderer.renderInlineAsText(tokens, options, env) -> String - * - tokens (Array): list on block tokens to render - * - options (Object): params of parser instance - * - env (Object): additional data from parsed input (references, for example) - * - * Special kludge for image `alt` attributes to conform CommonMark spec. - * Don't try to use it! Spec requires to show `alt` content with stripped markup, - * instead of simple escaping. - **/ Renderer.prototype.renderInlineAsText = function(tokens, options, env) { - var result = ""; - for (var i = 0, len = tokens.length; i < len; i++) { - if (tokens[i].type === "text") { - result += tokens[i].content; - } else if (tokens[i].type === "image") { - result += this.renderInlineAsText(tokens[i].children, options, env); - } else if (tokens[i].type === "softbreak") { - result += "\n"; - } - } - return result; - }; - /** - * Renderer.render(tokens, options, env) -> String - * - tokens (Array): list on block tokens to render - * - options (Object): params of parser instance - * - env (Object): additional data from parsed input (references, for example) - * - * Takes token stream and generates HTML. Probably, you will never need to call - * this method directly. - **/ Renderer.prototype.render = function(tokens, options, env) { - var i, len, type, result = "", rules = this.rules; - for (i = 0, len = tokens.length; i < len; i++) { - type = tokens[i].type; - if (type === "inline") { - result += this.renderInline(tokens[i].children, options, env); - } else if (typeof rules[type] !== "undefined") { - result += rules[type](tokens, i, options, env, this); - } else { - result += this.renderToken(tokens, i, options, env); - } - } - return result; - }; - var renderer = Renderer; - /** - * class Ruler - * - * Helper class, used by [[MarkdownIt#core]], [[MarkdownIt#block]] and - * [[MarkdownIt#inline]] to manage sequences of functions (rules): - * - * - keep rules in defined order - * - assign the name to each rule - * - enable/disable rules - * - add/replace rules - * - allow assign rules to additional named chains (in the same) - * - cacheing lists of active rules - * - * You will not need use this class directly until write plugins. For simple - * rules control use [[MarkdownIt.disable]], [[MarkdownIt.enable]] and - * [[MarkdownIt.use]]. - **/ - /** - * new Ruler() - **/ function Ruler() { - // List of added rules. Each element is: - // { - // name: XXX, - // enabled: Boolean, - // fn: Function(), - // alt: [ name2, name3 ] - // } - this.__rules__ = []; - // Cached rule chains. - - // First level - chain name, '' for default. - // Second level - diginal anchor for fast filtering by charcodes. - - this.__cache__ = null; - } - //////////////////////////////////////////////////////////////////////////////// - // Helper methods, should not be used directly - // Find rule index by name - - Ruler.prototype.__find__ = function(name) { - for (var i = 0; i < this.__rules__.length; i++) { - if (this.__rules__[i].name === name) { - return i; - } - } - return -1; - }; - // Build rules lookup cache - - Ruler.prototype.__compile__ = function() { - var self = this; - var chains = [ "" ]; - // collect unique names - self.__rules__.forEach((function(rule) { - if (!rule.enabled) { - return; - } - rule.alt.forEach((function(altName) { - if (chains.indexOf(altName) < 0) { - chains.push(altName); - } - })); - })); - self.__cache__ = {}; - chains.forEach((function(chain) { - self.__cache__[chain] = []; - self.__rules__.forEach((function(rule) { - if (!rule.enabled) { - return; - } - if (chain && rule.alt.indexOf(chain) < 0) { - return; - } - self.__cache__[chain].push(rule.fn); - })); - })); - }; - /** - * Ruler.at(name, fn [, options]) - * - name (String): rule name to replace. - * - fn (Function): new rule function. - * - options (Object): new rule options (not mandatory). - * - * Replace rule by name with new function & options. Throws error if name not - * found. - * - * ##### Options: - * - * - __alt__ - array with names of "alternate" chains. - * - * ##### Example - * - * Replace existing typographer replacement rule with new one: - * - * ```javascript - * var md = require('markdown-it')(); - * - * md.core.ruler.at('replacements', function replace(state) { - * //... - * }); - * ``` - **/ Ruler.prototype.at = function(name, fn, options) { - var index = this.__find__(name); - var opt = options || {}; - if (index === -1) { - throw new Error("Parser rule not found: " + name); - } - this.__rules__[index].fn = fn; - this.__rules__[index].alt = opt.alt || []; - this.__cache__ = null; - }; - /** - * Ruler.before(beforeName, ruleName, fn [, options]) - * - beforeName (String): new rule will be added before this one. - * - ruleName (String): name of added rule. - * - fn (Function): rule function. - * - options (Object): rule options (not mandatory). - * - * Add new rule to chain before one with given name. See also - * [[Ruler.after]], [[Ruler.push]]. - * - * ##### Options: - * - * - __alt__ - array with names of "alternate" chains. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')(); - * - * md.block.ruler.before('paragraph', 'my_rule', function replace(state) { - * //... - * }); - * ``` - **/ Ruler.prototype.before = function(beforeName, ruleName, fn, options) { - var index = this.__find__(beforeName); - var opt = options || {}; - if (index === -1) { - throw new Error("Parser rule not found: " + beforeName); - } - this.__rules__.splice(index, 0, { - name: ruleName, - enabled: true, - fn: fn, - alt: opt.alt || [] - }); - this.__cache__ = null; - }; - /** - * Ruler.after(afterName, ruleName, fn [, options]) - * - afterName (String): new rule will be added after this one. - * - ruleName (String): name of added rule. - * - fn (Function): rule function. - * - options (Object): rule options (not mandatory). - * - * Add new rule to chain after one with given name. See also - * [[Ruler.before]], [[Ruler.push]]. - * - * ##### Options: - * - * - __alt__ - array with names of "alternate" chains. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')(); - * - * md.inline.ruler.after('text', 'my_rule', function replace(state) { - * //... - * }); - * ``` - **/ Ruler.prototype.after = function(afterName, ruleName, fn, options) { - var index = this.__find__(afterName); - var opt = options || {}; - if (index === -1) { - throw new Error("Parser rule not found: " + afterName); - } - this.__rules__.splice(index + 1, 0, { - name: ruleName, - enabled: true, - fn: fn, - alt: opt.alt || [] - }); - this.__cache__ = null; - }; - /** - * Ruler.push(ruleName, fn [, options]) - * - ruleName (String): name of added rule. - * - fn (Function): rule function. - * - options (Object): rule options (not mandatory). - * - * Push new rule to the end of chain. See also - * [[Ruler.before]], [[Ruler.after]]. - * - * ##### Options: - * - * - __alt__ - array with names of "alternate" chains. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')(); - * - * md.core.ruler.push('my_rule', function replace(state) { - * //... - * }); - * ``` - **/ Ruler.prototype.push = function(ruleName, fn, options) { - var opt = options || {}; - this.__rules__.push({ - name: ruleName, - enabled: true, - fn: fn, - alt: opt.alt || [] - }); - this.__cache__ = null; - }; - /** - * Ruler.enable(list [, ignoreInvalid]) -> Array - * - list (String|Array): list of rule names to enable. - * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found. - * - * Enable rules with given names. If any rule name not found - throw Error. - * Errors can be disabled by second param. - * - * Returns list of found rule names (if no exception happened). - * - * See also [[Ruler.disable]], [[Ruler.enableOnly]]. - **/ Ruler.prototype.enable = function(list, ignoreInvalid) { - if (!Array.isArray(list)) { - list = [ list ]; - } - var result = []; - // Search by name and enable - list.forEach((function(name) { - var idx = this.__find__(name); - if (idx < 0) { - if (ignoreInvalid) { - return; - } - throw new Error("Rules manager: invalid rule name " + name); - } - this.__rules__[idx].enabled = true; - result.push(name); - }), this); - this.__cache__ = null; - return result; - }; - /** - * Ruler.enableOnly(list [, ignoreInvalid]) - * - list (String|Array): list of rule names to enable (whitelist). - * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found. - * - * Enable rules with given names, and disable everything else. If any rule name - * not found - throw Error. Errors can be disabled by second param. - * - * See also [[Ruler.disable]], [[Ruler.enable]]. - **/ Ruler.prototype.enableOnly = function(list, ignoreInvalid) { - if (!Array.isArray(list)) { - list = [ list ]; - } - this.__rules__.forEach((function(rule) { - rule.enabled = false; - })); - this.enable(list, ignoreInvalid); - }; - /** - * Ruler.disable(list [, ignoreInvalid]) -> Array - * - list (String|Array): list of rule names to disable. - * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found. - * - * Disable rules with given names. If any rule name not found - throw Error. - * Errors can be disabled by second param. - * - * Returns list of found rule names (if no exception happened). - * - * See also [[Ruler.enable]], [[Ruler.enableOnly]]. - **/ Ruler.prototype.disable = function(list, ignoreInvalid) { - if (!Array.isArray(list)) { - list = [ list ]; - } - var result = []; - // Search by name and disable - list.forEach((function(name) { - var idx = this.__find__(name); - if (idx < 0) { - if (ignoreInvalid) { - return; - } - throw new Error("Rules manager: invalid rule name " + name); - } - this.__rules__[idx].enabled = false; - result.push(name); - }), this); - this.__cache__ = null; - return result; - }; - /** - * Ruler.getRules(chainName) -> Array - * - * Return array of active functions (rules) for given chain name. It analyzes - * rules configuration, compiles caches if not exists and returns result. - * - * Default chain name is `''` (empty string). It can't be skipped. That's - * done intentionally, to keep signature monomorphic for high speed. - **/ Ruler.prototype.getRules = function(chainName) { - if (this.__cache__ === null) { - this.__compile__(); - } - // Chain can be empty, if rules disabled. But we still have to return Array. - return this.__cache__[chainName] || []; - }; - var ruler = Ruler; - // Normalize input string - // https://spec.commonmark.org/0.29/#line-ending - var NEWLINES_RE = /\r\n?|\n/g; - var NULL_RE = /\0/g; - var normalize = function normalize(state) { - var str; - // Normalize newlines - str = state.src.replace(NEWLINES_RE, "\n"); - // Replace NULL characters - str = str.replace(NULL_RE, "\ufffd"); - state.src = str; - }; - var block = function block(state) { - var token; - if (state.inlineMode) { - token = new state.Token("inline", "", 0); - token.content = state.src; - token.map = [ 0, 1 ]; - token.children = []; - state.tokens.push(token); - } else { - state.md.block.parse(state.src, state.md, state.env, state.tokens); - } - }; - var inline = function inline(state) { - var tokens = state.tokens, tok, i, l; - // Parse inlines - for (i = 0, l = tokens.length; i < l; i++) { - tok = tokens[i]; - if (tok.type === "inline") { - state.md.inline.parse(tok.content, state.md, state.env, tok.children); - } - } - }; - var arrayReplaceAt = utils.arrayReplaceAt; - function isLinkOpen$1(str) { - return /^\s]/i.test(str); - } - function isLinkClose$1(str) { - return /^<\/a\s*>/i.test(str); - } - var linkify$1 = function linkify(state) { - var i, j, l, tokens, token, currentToken, nodes, ln, text, pos, lastPos, level, htmlLinkLevel, url, fullUrl, urlText, blockTokens = state.tokens, links; - if (!state.md.options.linkify) { - return; - } - for (j = 0, l = blockTokens.length; j < l; j++) { - if (blockTokens[j].type !== "inline" || !state.md.linkify.pretest(blockTokens[j].content)) { - continue; - } - tokens = blockTokens[j].children; - htmlLinkLevel = 0; - // We scan from the end, to keep position when new tags added. - // Use reversed logic in links start/end match - for (i = tokens.length - 1; i >= 0; i--) { - currentToken = tokens[i]; - // Skip content of markdown links - if (currentToken.type === "link_close") { - i--; - while (tokens[i].level !== currentToken.level && tokens[i].type !== "link_open") { - i--; - } - continue; - } - // Skip content of html tag links - if (currentToken.type === "html_inline") { - if (isLinkOpen$1(currentToken.content) && htmlLinkLevel > 0) { - htmlLinkLevel--; - } - if (isLinkClose$1(currentToken.content)) { - htmlLinkLevel++; - } - } - if (htmlLinkLevel > 0) { - continue; - } - if (currentToken.type === "text" && state.md.linkify.test(currentToken.content)) { - text = currentToken.content; - links = state.md.linkify.match(text); - // Now split string to nodes - nodes = []; - level = currentToken.level; - lastPos = 0; - // forbid escape sequence at the start of the string, - // this avoids http\://example.com/ from being linkified as - // http://example.com/ - if (links.length > 0 && links[0].index === 0 && i > 0 && tokens[i - 1].type === "text_special") { - links = links.slice(1); - } - for (ln = 0; ln < links.length; ln++) { - url = links[ln].url; - fullUrl = state.md.normalizeLink(url); - if (!state.md.validateLink(fullUrl)) { - continue; - } - urlText = links[ln].text; - // Linkifier might send raw hostnames like "example.com", where url - // starts with domain name. So we prepend http:// in those cases, - // and remove it afterwards. - - if (!links[ln].schema) { - urlText = state.md.normalizeLinkText("http://" + urlText).replace(/^http:\/\//, ""); - } else if (links[ln].schema === "mailto:" && !/^mailto:/i.test(urlText)) { - urlText = state.md.normalizeLinkText("mailto:" + urlText).replace(/^mailto:/, ""); - } else { - urlText = state.md.normalizeLinkText(urlText); - } - pos = links[ln].index; - if (pos > lastPos) { - token = new state.Token("text", "", 0); - token.content = text.slice(lastPos, pos); - token.level = level; - nodes.push(token); - } - token = new state.Token("link_open", "a", 1); - token.attrs = [ [ "href", fullUrl ] ]; - token.level = level++; - token.markup = "linkify"; - token.info = "auto"; - nodes.push(token); - token = new state.Token("text", "", 0); - token.content = urlText; - token.level = level; - nodes.push(token); - token = new state.Token("link_close", "a", -1); - token.level = --level; - token.markup = "linkify"; - token.info = "auto"; - nodes.push(token); - lastPos = links[ln].lastIndex; - } - if (lastPos < text.length) { - token = new state.Token("text", "", 0); - token.content = text.slice(lastPos); - token.level = level; - nodes.push(token); - } - // replace current node - blockTokens[j].children = tokens = arrayReplaceAt(tokens, i, nodes); - } - } - } - }; - // Simple typographic replacements - // TODO: - // - fractionals 1/2, 1/4, 3/4 -> ½, ¼, ¾ - // - multiplications 2 x 4 -> 2 × 4 - var RARE_RE = /\+-|\.\.|\?\?\?\?|!!!!|,,|--/; - // Workaround for phantomjs - need regex without /g flag, - // or root check will fail every second time - var SCOPED_ABBR_TEST_RE = /\((c|tm|r)\)/i; - var SCOPED_ABBR_RE = /\((c|tm|r)\)/gi; - var SCOPED_ABBR = { - c: "\xa9", - r: "\xae", - tm: "\u2122" - }; - function replaceFn(match, name) { - return SCOPED_ABBR[name.toLowerCase()]; - } - function replace_scoped(inlineTokens) { - var i, token, inside_autolink = 0; - for (i = inlineTokens.length - 1; i >= 0; i--) { - token = inlineTokens[i]; - if (token.type === "text" && !inside_autolink) { - token.content = token.content.replace(SCOPED_ABBR_RE, replaceFn); - } - if (token.type === "link_open" && token.info === "auto") { - inside_autolink--; - } - if (token.type === "link_close" && token.info === "auto") { - inside_autolink++; - } - } - } - function replace_rare(inlineTokens) { - var i, token, inside_autolink = 0; - for (i = inlineTokens.length - 1; i >= 0; i--) { - token = inlineTokens[i]; - if (token.type === "text" && !inside_autolink) { - if (RARE_RE.test(token.content)) { - token.content = token.content.replace(/\+-/g, "\xb1").replace(/\.{2,}/g, "\u2026").replace(/([?!])\u2026/g, "$1..").replace(/([?!]){4,}/g, "$1$1$1").replace(/,{2,}/g, ",").replace(/(^|[^-])---(?=[^-]|$)/gm, "$1\u2014").replace(/(^|\s)--(?=\s|$)/gm, "$1\u2013").replace(/(^|[^-\s])--(?=[^-\s]|$)/gm, "$1\u2013"); - } - } - if (token.type === "link_open" && token.info === "auto") { - inside_autolink--; - } - if (token.type === "link_close" && token.info === "auto") { - inside_autolink++; - } - } - } - var replacements = function replace(state) { - var blkIdx; - if (!state.md.options.typographer) { - return; - } - for (blkIdx = state.tokens.length - 1; blkIdx >= 0; blkIdx--) { - if (state.tokens[blkIdx].type !== "inline") { - continue; - } - if (SCOPED_ABBR_TEST_RE.test(state.tokens[blkIdx].content)) { - replace_scoped(state.tokens[blkIdx].children); - } - if (RARE_RE.test(state.tokens[blkIdx].content)) { - replace_rare(state.tokens[blkIdx].children); - } - } - }; - var isWhiteSpace$1 = utils.isWhiteSpace; - var isPunctChar$1 = utils.isPunctChar; - var isMdAsciiPunct$1 = utils.isMdAsciiPunct; - var QUOTE_TEST_RE = /['"]/; - var QUOTE_RE = /['"]/g; - var APOSTROPHE = "\u2019"; - /* ’ */ function replaceAt(str, index, ch) { - return str.slice(0, index) + ch + str.slice(index + 1); - } - function process_inlines(tokens, state) { - var i, token, text, t, pos, max, thisLevel, item, lastChar, nextChar, isLastPunctChar, isNextPunctChar, isLastWhiteSpace, isNextWhiteSpace, canOpen, canClose, j, isSingle, stack, openQuote, closeQuote; - stack = []; - for (i = 0; i < tokens.length; i++) { - token = tokens[i]; - thisLevel = tokens[i].level; - for (j = stack.length - 1; j >= 0; j--) { - if (stack[j].level <= thisLevel) { - break; - } - } - stack.length = j + 1; - if (token.type !== "text") { - continue; - } - text = token.content; - pos = 0; - max = text.length; - /*eslint no-labels:0,block-scoped-var:0*/ OUTER: while (pos < max) { - QUOTE_RE.lastIndex = pos; - t = QUOTE_RE.exec(text); - if (!t) { - break; - } - canOpen = canClose = true; - pos = t.index + 1; - isSingle = t[0] === "'"; - // Find previous character, - // default to space if it's the beginning of the line - - lastChar = 32; - if (t.index - 1 >= 0) { - lastChar = text.charCodeAt(t.index - 1); - } else { - for (j = i - 1; j >= 0; j--) { - if (tokens[j].type === "softbreak" || tokens[j].type === "hardbreak") break; - // lastChar defaults to 0x20 - if (!tokens[j].content) continue; - // should skip all tokens except 'text', 'html_inline' or 'code_inline' - lastChar = tokens[j].content.charCodeAt(tokens[j].content.length - 1); - break; - } - } - // Find next character, - // default to space if it's the end of the line - - nextChar = 32; - if (pos < max) { - nextChar = text.charCodeAt(pos); - } else { - for (j = i + 1; j < tokens.length; j++) { - if (tokens[j].type === "softbreak" || tokens[j].type === "hardbreak") break; - // nextChar defaults to 0x20 - if (!tokens[j].content) continue; - // should skip all tokens except 'text', 'html_inline' or 'code_inline' - nextChar = tokens[j].content.charCodeAt(0); - break; - } - } - isLastPunctChar = isMdAsciiPunct$1(lastChar) || isPunctChar$1(String.fromCharCode(lastChar)); - isNextPunctChar = isMdAsciiPunct$1(nextChar) || isPunctChar$1(String.fromCharCode(nextChar)); - isLastWhiteSpace = isWhiteSpace$1(lastChar); - isNextWhiteSpace = isWhiteSpace$1(nextChar); - if (isNextWhiteSpace) { - canOpen = false; - } else if (isNextPunctChar) { - if (!(isLastWhiteSpace || isLastPunctChar)) { - canOpen = false; - } - } - if (isLastWhiteSpace) { - canClose = false; - } else if (isLastPunctChar) { - if (!(isNextWhiteSpace || isNextPunctChar)) { - canClose = false; - } - } - if (nextChar === 34 /* " */ && t[0] === '"') { - if (lastChar >= 48 /* 0 */ && lastChar <= 57 /* 9 */) { - // special case: 1"" - count first quote as an inch - canClose = canOpen = false; - } - } - if (canOpen && canClose) { - // Replace quotes in the middle of punctuation sequence, but not - // in the middle of the words, i.e.: - // 1. foo " bar " baz - not replaced - // 2. foo-"-bar-"-baz - replaced - // 3. foo"bar"baz - not replaced - canOpen = isLastPunctChar; - canClose = isNextPunctChar; - } - if (!canOpen && !canClose) { - // middle of word - if (isSingle) { - token.content = replaceAt(token.content, t.index, APOSTROPHE); - } - continue; - } - if (canClose) { - // this could be a closing quote, rewind the stack to get a match - for (j = stack.length - 1; j >= 0; j--) { - item = stack[j]; - if (stack[j].level < thisLevel) { - break; - } - if (item.single === isSingle && stack[j].level === thisLevel) { - item = stack[j]; - if (isSingle) { - openQuote = state.md.options.quotes[2]; - closeQuote = state.md.options.quotes[3]; - } else { - openQuote = state.md.options.quotes[0]; - closeQuote = state.md.options.quotes[1]; - } - // replace token.content *before* tokens[item.token].content, - // because, if they are pointing at the same token, replaceAt - // could mess up indices when quote length != 1 - token.content = replaceAt(token.content, t.index, closeQuote); - tokens[item.token].content = replaceAt(tokens[item.token].content, item.pos, openQuote); - pos += closeQuote.length - 1; - if (item.token === i) { - pos += openQuote.length - 1; - } - text = token.content; - max = text.length; - stack.length = j; - continue OUTER; - } - } - } - if (canOpen) { - stack.push({ - token: i, - pos: t.index, - single: isSingle, - level: thisLevel - }); - } else if (canClose && isSingle) { - token.content = replaceAt(token.content, t.index, APOSTROPHE); - } - } - } - } - var smartquotes = function smartquotes(state) { - /*eslint max-depth:0*/ - var blkIdx; - if (!state.md.options.typographer) { - return; - } - for (blkIdx = state.tokens.length - 1; blkIdx >= 0; blkIdx--) { - if (state.tokens[blkIdx].type !== "inline" || !QUOTE_TEST_RE.test(state.tokens[blkIdx].content)) { - continue; - } - process_inlines(state.tokens[blkIdx].children, state); - } - }; - // Join raw text tokens with the rest of the text - var text_join = function text_join(state) { - var j, l, tokens, curr, max, last, blockTokens = state.tokens; - for (j = 0, l = blockTokens.length; j < l; j++) { - if (blockTokens[j].type !== "inline") continue; - tokens = blockTokens[j].children; - max = tokens.length; - for (curr = 0; curr < max; curr++) { - if (tokens[curr].type === "text_special") { - tokens[curr].type = "text"; - } - } - for (curr = last = 0; curr < max; curr++) { - if (tokens[curr].type === "text" && curr + 1 < max && tokens[curr + 1].type === "text") { - // collapse two adjacent text nodes - tokens[curr + 1].content = tokens[curr].content + tokens[curr + 1].content; - } else { - if (curr !== last) { - tokens[last] = tokens[curr]; - } - last++; - } - } - if (curr !== last) { - tokens.length = last; - } - } - }; - // Token class - /** - * class Token - **/ - /** - * new Token(type, tag, nesting) - * - * Create new token and fill passed properties. - **/ function Token(type, tag, nesting) { - /** - * Token#type -> String - * - * Type of the token (string, e.g. "paragraph_open") - **/ - this.type = type; - /** - * Token#tag -> String - * - * html tag name, e.g. "p" - **/ this.tag = tag; - /** - * Token#attrs -> Array - * - * Html attributes. Format: `[ [ name1, value1 ], [ name2, value2 ] ]` - **/ this.attrs = null; - /** - * Token#map -> Array - * - * Source map info. Format: `[ line_begin, line_end ]` - **/ this.map = null; - /** - * Token#nesting -> Number - * - * Level change (number in {-1, 0, 1} set), where: - * - * - `1` means the tag is opening - * - `0` means the tag is self-closing - * - `-1` means the tag is closing - **/ this.nesting = nesting; - /** - * Token#level -> Number - * - * nesting level, the same as `state.level` - **/ this.level = 0; - /** - * Token#children -> Array - * - * An array of child nodes (inline and img tokens) - **/ this.children = null; - /** - * Token#content -> String - * - * In a case of self-closing tag (code, html, fence, etc.), - * it has contents of this tag. - **/ this.content = ""; - /** - * Token#markup -> String - * - * '*' or '_' for emphasis, fence string for fence, etc. - **/ this.markup = ""; - /** - * Token#info -> String - * - * Additional information: - * - * - Info string for "fence" tokens - * - The value "auto" for autolink "link_open" and "link_close" tokens - * - The string value of the item marker for ordered-list "list_item_open" tokens - **/ this.info = ""; - /** - * Token#meta -> Object - * - * A place for plugins to store an arbitrary data - **/ this.meta = null; - /** - * Token#block -> Boolean - * - * True for block-level tokens, false for inline tokens. - * Used in renderer to calculate line breaks - **/ this.block = false; - /** - * Token#hidden -> Boolean - * - * If it's true, ignore this element when rendering. Used for tight lists - * to hide paragraphs. - **/ this.hidden = false; - } - /** - * Token.attrIndex(name) -> Number - * - * Search attribute index by name. - **/ Token.prototype.attrIndex = function attrIndex(name) { - var attrs, i, len; - if (!this.attrs) { - return -1; - } - attrs = this.attrs; - for (i = 0, len = attrs.length; i < len; i++) { - if (attrs[i][0] === name) { - return i; - } - } - return -1; - }; - /** - * Token.attrPush(attrData) - * - * Add `[ name, value ]` attribute to list. Init attrs if necessary - **/ Token.prototype.attrPush = function attrPush(attrData) { - if (this.attrs) { - this.attrs.push(attrData); - } else { - this.attrs = [ attrData ]; - } - }; - /** - * Token.attrSet(name, value) - * - * Set `name` attribute to `value`. Override old value if exists. - **/ Token.prototype.attrSet = function attrSet(name, value) { - var idx = this.attrIndex(name), attrData = [ name, value ]; - if (idx < 0) { - this.attrPush(attrData); - } else { - this.attrs[idx] = attrData; - } - }; - /** - * Token.attrGet(name) - * - * Get the value of attribute `name`, or null if it does not exist. - **/ Token.prototype.attrGet = function attrGet(name) { - var idx = this.attrIndex(name), value = null; - if (idx >= 0) { - value = this.attrs[idx][1]; - } - return value; - }; - /** - * Token.attrJoin(name, value) - * - * Join value to existing attribute via space. Or create new attribute if not - * exists. Useful to operate with token classes. - **/ Token.prototype.attrJoin = function attrJoin(name, value) { - var idx = this.attrIndex(name); - if (idx < 0) { - this.attrPush([ name, value ]); - } else { - this.attrs[idx][1] = this.attrs[idx][1] + " " + value; - } - }; - var token = Token; - function StateCore(src, md, env) { - this.src = src; - this.env = env; - this.tokens = []; - this.inlineMode = false; - this.md = md; - // link to parser instance - } - // re-export Token class to use in core rules - StateCore.prototype.Token = token; - var state_core = StateCore; - var _rules$2 = [ [ "normalize", normalize ], [ "block", block ], [ "inline", inline ], [ "linkify", linkify$1 ], [ "replacements", replacements ], [ "smartquotes", smartquotes ], - // `text_join` finds `text_special` tokens (for escape sequences) - // and joins them with the rest of the text - [ "text_join", text_join ] ]; - /** - * new Core() - **/ function Core() { - /** - * Core#ruler -> Ruler - * - * [[Ruler]] instance. Keep configuration of core rules. - **/ - this.ruler = new ruler; - for (var i = 0; i < _rules$2.length; i++) { - this.ruler.push(_rules$2[i][0], _rules$2[i][1]); - } - } - /** - * Core.process(state) - * - * Executes core chain rules. - **/ Core.prototype.process = function(state) { - var i, l, rules; - rules = this.ruler.getRules(""); - for (i = 0, l = rules.length; i < l; i++) { - rules[i](state); - } - }; - Core.prototype.State = state_core; - var parser_core = Core; - var isSpace$a = utils.isSpace; - function getLine(state, line) { - var pos = state.bMarks[line] + state.tShift[line], max = state.eMarks[line]; - return state.src.slice(pos, max); - } - function escapedSplit(str) { - var result = [], pos = 0, max = str.length, ch, isEscaped = false, lastPos = 0, current = ""; - ch = str.charCodeAt(pos); - while (pos < max) { - if (ch === 124 /* | */) { - if (!isEscaped) { - // pipe separating cells, '|' - result.push(current + str.substring(lastPos, pos)); - current = ""; - lastPos = pos + 1; - } else { - // escaped pipe, '\|' - current += str.substring(lastPos, pos - 1); - lastPos = pos; - } - } - isEscaped = ch === 92 /* \ */; - pos++; - ch = str.charCodeAt(pos); - } - result.push(current + str.substring(lastPos)); - return result; - } - var table = function table(state, startLine, endLine, silent) { - var ch, lineText, pos, i, l, nextLine, columns, columnCount, token, aligns, t, tableLines, tbodyLines, oldParentType, terminate, terminatorRules, firstCh, secondCh; - // should have at least two lines - if (startLine + 2 > endLine) { - return false; - } - nextLine = startLine + 1; - if (state.sCount[nextLine] < state.blkIndent) { - return false; - } - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[nextLine] - state.blkIndent >= 4) { - return false; - } - // first character of the second line should be '|', '-', ':', - // and no other characters are allowed but spaces; - // basically, this is the equivalent of /^[-:|][-:|\s]*$/ regexp - pos = state.bMarks[nextLine] + state.tShift[nextLine]; - if (pos >= state.eMarks[nextLine]) { - return false; - } - firstCh = state.src.charCodeAt(pos++); - if (firstCh !== 124 /* | */ && firstCh !== 45 /* - */ && firstCh !== 58 /* : */) { - return false; - } - if (pos >= state.eMarks[nextLine]) { - return false; - } - secondCh = state.src.charCodeAt(pos++); - if (secondCh !== 124 /* | */ && secondCh !== 45 /* - */ && secondCh !== 58 /* : */ && !isSpace$a(secondCh)) { - return false; - } - // if first character is '-', then second character must not be a space - // (due to parsing ambiguity with list) - if (firstCh === 45 /* - */ && isSpace$a(secondCh)) { - return false; - } - while (pos < state.eMarks[nextLine]) { - ch = state.src.charCodeAt(pos); - if (ch !== 124 /* | */ && ch !== 45 /* - */ && ch !== 58 /* : */ && !isSpace$a(ch)) { - return false; - } - pos++; - } - lineText = getLine(state, startLine + 1); - columns = lineText.split("|"); - aligns = []; - for (i = 0; i < columns.length; i++) { - t = columns[i].trim(); - if (!t) { - // allow empty columns before and after table, but not in between columns; - // e.g. allow ` |---| `, disallow ` ---||--- ` - if (i === 0 || i === columns.length - 1) { - continue; - } else { - return false; - } - } - if (!/^:?-+:?$/.test(t)) { - return false; - } - if (t.charCodeAt(t.length - 1) === 58 /* : */) { - aligns.push(t.charCodeAt(0) === 58 /* : */ ? "center" : "right"); - } else if (t.charCodeAt(0) === 58 /* : */) { - aligns.push("left"); - } else { - aligns.push(""); - } - } - lineText = getLine(state, startLine).trim(); - if (lineText.indexOf("|") === -1) { - return false; - } - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - columns = escapedSplit(lineText); - if (columns.length && columns[0] === "") columns.shift(); - if (columns.length && columns[columns.length - 1] === "") columns.pop(); - // header row will define an amount of columns in the entire table, - // and align row should be exactly the same (the rest of the rows can differ) - columnCount = columns.length; - if (columnCount === 0 || columnCount !== aligns.length) { - return false; - } - if (silent) { - return true; - } - oldParentType = state.parentType; - state.parentType = "table"; - // use 'blockquote' lists for termination because it's - // the most similar to tables - terminatorRules = state.md.block.ruler.getRules("blockquote"); - token = state.push("table_open", "table", 1); - token.map = tableLines = [ startLine, 0 ]; - token = state.push("thead_open", "thead", 1); - token.map = [ startLine, startLine + 1 ]; - token = state.push("tr_open", "tr", 1); - token.map = [ startLine, startLine + 1 ]; - for (i = 0; i < columns.length; i++) { - token = state.push("th_open", "th", 1); - if (aligns[i]) { - token.attrs = [ [ "style", "text-align:" + aligns[i] ] ]; - } - token = state.push("inline", "", 0); - token.content = columns[i].trim(); - token.children = []; - token = state.push("th_close", "th", -1); - } - token = state.push("tr_close", "tr", -1); - token = state.push("thead_close", "thead", -1); - for (nextLine = startLine + 2; nextLine < endLine; nextLine++) { - if (state.sCount[nextLine] < state.blkIndent) { - break; - } - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - break; - } - lineText = getLine(state, nextLine).trim(); - if (!lineText) { - break; - } - if (state.sCount[nextLine] - state.blkIndent >= 4) { - break; - } - columns = escapedSplit(lineText); - if (columns.length && columns[0] === "") columns.shift(); - if (columns.length && columns[columns.length - 1] === "") columns.pop(); - if (nextLine === startLine + 2) { - token = state.push("tbody_open", "tbody", 1); - token.map = tbodyLines = [ startLine + 2, 0 ]; - } - token = state.push("tr_open", "tr", 1); - token.map = [ nextLine, nextLine + 1 ]; - for (i = 0; i < columnCount; i++) { - token = state.push("td_open", "td", 1); - if (aligns[i]) { - token.attrs = [ [ "style", "text-align:" + aligns[i] ] ]; - } - token = state.push("inline", "", 0); - token.content = columns[i] ? columns[i].trim() : ""; - token.children = []; - token = state.push("td_close", "td", -1); - } - token = state.push("tr_close", "tr", -1); - } - if (tbodyLines) { - token = state.push("tbody_close", "tbody", -1); - tbodyLines[1] = nextLine; - } - token = state.push("table_close", "table", -1); - tableLines[1] = nextLine; - state.parentType = oldParentType; - state.line = nextLine; - return true; - }; - // Code block (4 spaces padded) - var code = function code(state, startLine, endLine /*, silent*/) { - var nextLine, last, token; - if (state.sCount[startLine] - state.blkIndent < 4) { - return false; - } - last = nextLine = startLine + 1; - while (nextLine < endLine) { - if (state.isEmpty(nextLine)) { - nextLine++; - continue; - } - if (state.sCount[nextLine] - state.blkIndent >= 4) { - nextLine++; - last = nextLine; - continue; - } - break; - } - state.line = last; - token = state.push("code_block", "code", 0); - token.content = state.getLines(startLine, last, 4 + state.blkIndent, false) + "\n"; - token.map = [ startLine, state.line ]; - return true; - }; - // fences (``` lang, ~~~ lang) - var fence = function fence(state, startLine, endLine, silent) { - var marker, len, params, nextLine, mem, token, markup, haveEndMarker = false, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine]; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - if (pos + 3 > max) { - return false; - } - marker = state.src.charCodeAt(pos); - if (marker !== 126 /* ~ */ && marker !== 96 /* ` */) { - return false; - } - // scan marker length - mem = pos; - pos = state.skipChars(pos, marker); - len = pos - mem; - if (len < 3) { - return false; - } - markup = state.src.slice(mem, pos); - params = state.src.slice(pos, max); - if (marker === 96 /* ` */) { - if (params.indexOf(String.fromCharCode(marker)) >= 0) { - return false; - } - } - // Since start is found, we can report success here in validation mode - if (silent) { - return true; - } - // search end of block - nextLine = startLine; - for (;;) { - nextLine++; - if (nextLine >= endLine) { - // unclosed block should be autoclosed by end of document. - // also block seems to be autoclosed by end of parent - break; - } - pos = mem = state.bMarks[nextLine] + state.tShift[nextLine]; - max = state.eMarks[nextLine]; - if (pos < max && state.sCount[nextLine] < state.blkIndent) { - // non-empty line with negative indent should stop the list: - // - ``` - // test - break; - } - if (state.src.charCodeAt(pos) !== marker) { - continue; - } - if (state.sCount[nextLine] - state.blkIndent >= 4) { - // closing fence should be indented less than 4 spaces - continue; - } - pos = state.skipChars(pos, marker); - // closing code fence must be at least as long as the opening one - if (pos - mem < len) { - continue; - } - // make sure tail has spaces only - pos = state.skipSpaces(pos); - if (pos < max) { - continue; - } - haveEndMarker = true; - // found! - break; - } - // If a fence has heading spaces, they should be removed from its inner block - len = state.sCount[startLine]; - state.line = nextLine + (haveEndMarker ? 1 : 0); - token = state.push("fence", "code", 0); - token.info = params; - token.content = state.getLines(startLine + 1, nextLine, len, true); - token.markup = markup; - token.map = [ startLine, state.line ]; - return true; - }; - var isSpace$9 = utils.isSpace; - var blockquote = function blockquote(state, startLine, endLine, silent) { - var adjustTab, ch, i, initial, l, lastLineEmpty, lines, nextLine, offset, oldBMarks, oldBSCount, oldIndent, oldParentType, oldSCount, oldTShift, spaceAfterMarker, terminate, terminatorRules, token, isOutdented, oldLineMax = state.lineMax, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine]; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - // check the block quote marker - if (state.src.charCodeAt(pos) !== 62 /* > */) { - return false; - } - // we know that it's going to be a valid blockquote, - // so no point trying to find the end of it in silent mode - if (silent) { - return true; - } - oldBMarks = []; - oldBSCount = []; - oldSCount = []; - oldTShift = []; - terminatorRules = state.md.block.ruler.getRules("blockquote"); - oldParentType = state.parentType; - state.parentType = "blockquote"; - // Search the end of the block - - // Block ends with either: - // 1. an empty line outside: - // ``` - // > test - - // ``` - // 2. an empty line inside: - // ``` - // > - // test - // ``` - // 3. another tag: - // ``` - // > test - // - - - - // ``` - for (nextLine = startLine; nextLine < endLine; nextLine++) { - // check if it's outdented, i.e. it's inside list item and indented - // less than said list item: - // ``` - // 1. anything - // > current blockquote - // 2. checking this line - // ``` - isOutdented = state.sCount[nextLine] < state.blkIndent; - pos = state.bMarks[nextLine] + state.tShift[nextLine]; - max = state.eMarks[nextLine]; - if (pos >= max) { - // Case 1: line is not inside the blockquote, and this line is empty. - break; - } - if (state.src.charCodeAt(pos++) === 62 /* > */ && !isOutdented) { - // This line is inside the blockquote. - // set offset past spaces and ">" - initial = state.sCount[nextLine] + 1; - // skip one optional space after '>' - if (state.src.charCodeAt(pos) === 32 /* space */) { - // ' > test ' - // ^ -- position start of line here: - pos++; - initial++; - adjustTab = false; - spaceAfterMarker = true; - } else if (state.src.charCodeAt(pos) === 9 /* tab */) { - spaceAfterMarker = true; - if ((state.bsCount[nextLine] + initial) % 4 === 3) { - // ' >\t test ' - // ^ -- position start of line here (tab has width===1) - pos++; - initial++; - adjustTab = false; - } else { - // ' >\t test ' - // ^ -- position start of line here + shift bsCount slightly - // to make extra space appear - adjustTab = true; - } - } else { - spaceAfterMarker = false; - } - offset = initial; - oldBMarks.push(state.bMarks[nextLine]); - state.bMarks[nextLine] = pos; - while (pos < max) { - ch = state.src.charCodeAt(pos); - if (isSpace$9(ch)) { - if (ch === 9) { - offset += 4 - (offset + state.bsCount[nextLine] + (adjustTab ? 1 : 0)) % 4; - } else { - offset++; - } - } else { - break; - } - pos++; - } - lastLineEmpty = pos >= max; - oldBSCount.push(state.bsCount[nextLine]); - state.bsCount[nextLine] = state.sCount[nextLine] + 1 + (spaceAfterMarker ? 1 : 0); - oldSCount.push(state.sCount[nextLine]); - state.sCount[nextLine] = offset - initial; - oldTShift.push(state.tShift[nextLine]); - state.tShift[nextLine] = pos - state.bMarks[nextLine]; - continue; - } - // Case 2: line is not inside the blockquote, and the last line was empty. - if (lastLineEmpty) { - break; - } - // Case 3: another tag found. - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - // Quirk to enforce "hard termination mode" for paragraphs; - // normally if you call `tokenize(state, startLine, nextLine)`, - // paragraphs will look below nextLine for paragraph continuation, - // but if blockquote is terminated by another tag, they shouldn't - state.lineMax = nextLine; - if (state.blkIndent !== 0) { - // state.blkIndent was non-zero, we now set it to zero, - // so we need to re-calculate all offsets to appear as - // if indent wasn't changed - oldBMarks.push(state.bMarks[nextLine]); - oldBSCount.push(state.bsCount[nextLine]); - oldTShift.push(state.tShift[nextLine]); - oldSCount.push(state.sCount[nextLine]); - state.sCount[nextLine] -= state.blkIndent; - } - break; - } - oldBMarks.push(state.bMarks[nextLine]); - oldBSCount.push(state.bsCount[nextLine]); - oldTShift.push(state.tShift[nextLine]); - oldSCount.push(state.sCount[nextLine]); - // A negative indentation means that this is a paragraph continuation - - state.sCount[nextLine] = -1; - } - oldIndent = state.blkIndent; - state.blkIndent = 0; - token = state.push("blockquote_open", "blockquote", 1); - token.markup = ">"; - token.map = lines = [ startLine, 0 ]; - state.md.block.tokenize(state, startLine, nextLine); - token = state.push("blockquote_close", "blockquote", -1); - token.markup = ">"; - state.lineMax = oldLineMax; - state.parentType = oldParentType; - lines[1] = state.line; - // Restore original tShift; this might not be necessary since the parser - // has already been here, but just to make sure we can do that. - for (i = 0; i < oldTShift.length; i++) { - state.bMarks[i + startLine] = oldBMarks[i]; - state.tShift[i + startLine] = oldTShift[i]; - state.sCount[i + startLine] = oldSCount[i]; - state.bsCount[i + startLine] = oldBSCount[i]; - } - state.blkIndent = oldIndent; - return true; - }; - var isSpace$8 = utils.isSpace; - var hr = function hr(state, startLine, endLine, silent) { - var marker, cnt, ch, token, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine]; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - marker = state.src.charCodeAt(pos++); - // Check hr marker - if (marker !== 42 /* * */ && marker !== 45 /* - */ && marker !== 95 /* _ */) { - return false; - } - // markers can be mixed with spaces, but there should be at least 3 of them - cnt = 1; - while (pos < max) { - ch = state.src.charCodeAt(pos++); - if (ch !== marker && !isSpace$8(ch)) { - return false; - } - if (ch === marker) { - cnt++; - } - } - if (cnt < 3) { - return false; - } - if (silent) { - return true; - } - state.line = startLine + 1; - token = state.push("hr", "hr", 0); - token.map = [ startLine, state.line ]; - token.markup = Array(cnt + 1).join(String.fromCharCode(marker)); - return true; - }; - var isSpace$7 = utils.isSpace; - // Search `[-+*][\n ]`, returns next pos after marker on success - // or -1 on fail. - function skipBulletListMarker(state, startLine) { - var marker, pos, max, ch; - pos = state.bMarks[startLine] + state.tShift[startLine]; - max = state.eMarks[startLine]; - marker = state.src.charCodeAt(pos++); - // Check bullet - if (marker !== 42 /* * */ && marker !== 45 /* - */ && marker !== 43 /* + */) { - return -1; - } - if (pos < max) { - ch = state.src.charCodeAt(pos); - if (!isSpace$7(ch)) { - // " -test " - is not a list item - return -1; - } - } - return pos; - } - // Search `\d+[.)][\n ]`, returns next pos after marker on success - // or -1 on fail. - function skipOrderedListMarker(state, startLine) { - var ch, start = state.bMarks[startLine] + state.tShift[startLine], pos = start, max = state.eMarks[startLine]; - // List marker should have at least 2 chars (digit + dot) - if (pos + 1 >= max) { - return -1; - } - ch = state.src.charCodeAt(pos++); - if (ch < 48 /* 0 */ || ch > 57 /* 9 */) { - return -1; - } - for (;;) { - // EOL -> fail - if (pos >= max) { - return -1; - } - ch = state.src.charCodeAt(pos++); - if (ch >= 48 /* 0 */ && ch <= 57 /* 9 */) { - // List marker should have no more than 9 digits - // (prevents integer overflow in browsers) - if (pos - start >= 10) { - return -1; - } - continue; - } - // found valid marker - if (ch === 41 /* ) */ || ch === 46 /* . */) { - break; - } - return -1; - } - if (pos < max) { - ch = state.src.charCodeAt(pos); - if (!isSpace$7(ch)) { - // " 1.test " - is not a list item - return -1; - } - } - return pos; - } - function markTightParagraphs(state, idx) { - var i, l, level = state.level + 2; - for (i = idx + 2, l = state.tokens.length - 2; i < l; i++) { - if (state.tokens[i].level === level && state.tokens[i].type === "paragraph_open") { - state.tokens[i + 2].hidden = true; - state.tokens[i].hidden = true; - i += 2; - } - } - } - var list = function list(state, startLine, endLine, silent) { - var ch, contentStart, i, indent, indentAfterMarker, initial, isOrdered, itemLines, l, listLines, listTokIdx, markerCharCode, markerValue, max, offset, oldListIndent, oldParentType, oldSCount, oldTShift, oldTight, pos, posAfterMarker, prevEmptyEnd, start, terminate, terminatorRules, token, nextLine = startLine, isTerminatingParagraph = false, tight = true; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[nextLine] - state.blkIndent >= 4) { - return false; - } - // Special case: - // - item 1 - // - item 2 - // - item 3 - // - item 4 - // - this one is a paragraph continuation - if (state.listIndent >= 0 && state.sCount[nextLine] - state.listIndent >= 4 && state.sCount[nextLine] < state.blkIndent) { - return false; - } - // limit conditions when list can interrupt - // a paragraph (validation mode only) - if (silent && state.parentType === "paragraph") { - // Next list item should still terminate previous list item; - // This code can fail if plugins use blkIndent as well as lists, - // but I hope the spec gets fixed long before that happens. - if (state.sCount[nextLine] >= state.blkIndent) { - isTerminatingParagraph = true; - } - } - // Detect list type and position after marker - if ((posAfterMarker = skipOrderedListMarker(state, nextLine)) >= 0) { - isOrdered = true; - start = state.bMarks[nextLine] + state.tShift[nextLine]; - markerValue = Number(state.src.slice(start, posAfterMarker - 1)); - // If we're starting a new ordered list right after - // a paragraph, it should start with 1. - if (isTerminatingParagraph && markerValue !== 1) return false; - } else if ((posAfterMarker = skipBulletListMarker(state, nextLine)) >= 0) { - isOrdered = false; - } else { - return false; - } - // If we're starting a new unordered list right after - // a paragraph, first line should not be empty. - if (isTerminatingParagraph) { - if (state.skipSpaces(posAfterMarker) >= state.eMarks[nextLine]) return false; - } - // For validation mode we can terminate immediately - if (silent) { - return true; - } - // We should terminate list on style change. Remember first one to compare. - markerCharCode = state.src.charCodeAt(posAfterMarker - 1); - // Start list - listTokIdx = state.tokens.length; - if (isOrdered) { - token = state.push("ordered_list_open", "ol", 1); - if (markerValue !== 1) { - token.attrs = [ [ "start", markerValue ] ]; - } - } else { - token = state.push("bullet_list_open", "ul", 1); - } - token.map = listLines = [ nextLine, 0 ]; - token.markup = String.fromCharCode(markerCharCode); - - // Iterate list items - - prevEmptyEnd = false; - terminatorRules = state.md.block.ruler.getRules("list"); - oldParentType = state.parentType; - state.parentType = "list"; - while (nextLine < endLine) { - pos = posAfterMarker; - max = state.eMarks[nextLine]; - initial = offset = state.sCount[nextLine] + posAfterMarker - (state.bMarks[nextLine] + state.tShift[nextLine]); - while (pos < max) { - ch = state.src.charCodeAt(pos); - if (ch === 9) { - offset += 4 - (offset + state.bsCount[nextLine]) % 4; - } else if (ch === 32) { - offset++; - } else { - break; - } - pos++; - } - contentStart = pos; - if (contentStart >= max) { - // trimming space in "- \n 3" case, indent is 1 here - indentAfterMarker = 1; - } else { - indentAfterMarker = offset - initial; - } - // If we have more than 4 spaces, the indent is 1 - // (the rest is just indented code block) - if (indentAfterMarker > 4) { - indentAfterMarker = 1; - } - // " - test" - // ^^^^^ - calculating total length of this thing - indent = initial + indentAfterMarker; - // Run subparser & write tokens - token = state.push("list_item_open", "li", 1); - token.markup = String.fromCharCode(markerCharCode); - token.map = itemLines = [ nextLine, 0 ]; - if (isOrdered) { - token.info = state.src.slice(start, posAfterMarker - 1); - } - // change current state, then restore it after parser subcall - oldTight = state.tight; - oldTShift = state.tShift[nextLine]; - oldSCount = state.sCount[nextLine]; - // - example list - // ^ listIndent position will be here - // ^ blkIndent position will be here - - oldListIndent = state.listIndent; - state.listIndent = state.blkIndent; - state.blkIndent = indent; - state.tight = true; - state.tShift[nextLine] = contentStart - state.bMarks[nextLine]; - state.sCount[nextLine] = offset; - if (contentStart >= max && state.isEmpty(nextLine + 1)) { - // workaround for this case - // (list item is empty, list terminates before "foo"): - // ~~~~~~~~ - // - - // foo - // ~~~~~~~~ - state.line = Math.min(state.line + 2, endLine); - } else { - state.md.block.tokenize(state, nextLine, endLine, true); - } - // If any of list item is tight, mark list as tight - if (!state.tight || prevEmptyEnd) { - tight = false; - } - // Item become loose if finish with empty line, - // but we should filter last element, because it means list finish - prevEmptyEnd = state.line - nextLine > 1 && state.isEmpty(state.line - 1); - state.blkIndent = state.listIndent; - state.listIndent = oldListIndent; - state.tShift[nextLine] = oldTShift; - state.sCount[nextLine] = oldSCount; - state.tight = oldTight; - token = state.push("list_item_close", "li", -1); - token.markup = String.fromCharCode(markerCharCode); - nextLine = state.line; - itemLines[1] = nextLine; - if (nextLine >= endLine) { - break; - } - - // Try to check if list is terminated or continued. - - if (state.sCount[nextLine] < state.blkIndent) { - break; - } - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[nextLine] - state.blkIndent >= 4) { - break; - } - // fail if terminating block found - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - break; - } - // fail if list has another type - if (isOrdered) { - posAfterMarker = skipOrderedListMarker(state, nextLine); - if (posAfterMarker < 0) { - break; - } - start = state.bMarks[nextLine] + state.tShift[nextLine]; - } else { - posAfterMarker = skipBulletListMarker(state, nextLine); - if (posAfterMarker < 0) { - break; - } - } - if (markerCharCode !== state.src.charCodeAt(posAfterMarker - 1)) { - break; - } - } - // Finalize list - if (isOrdered) { - token = state.push("ordered_list_close", "ol", -1); - } else { - token = state.push("bullet_list_close", "ul", -1); - } - token.markup = String.fromCharCode(markerCharCode); - listLines[1] = nextLine; - state.line = nextLine; - state.parentType = oldParentType; - // mark paragraphs tight if needed - if (tight) { - markTightParagraphs(state, listTokIdx); - } - return true; - }; - var normalizeReference$2 = utils.normalizeReference; - var isSpace$6 = utils.isSpace; - var reference = function reference(state, startLine, _endLine, silent) { - var ch, destEndPos, destEndLineNo, endLine, href, i, l, label, labelEnd, oldParentType, res, start, str, terminate, terminatorRules, title, lines = 0, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine], nextLine = startLine + 1; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - if (state.src.charCodeAt(pos) !== 91 /* [ */) { - return false; - } - // Simple check to quickly interrupt scan on [link](url) at the start of line. - // Can be useful on practice: https://github.com/markdown-it/markdown-it/issues/54 - while (++pos < max) { - if (state.src.charCodeAt(pos) === 93 /* ] */ && state.src.charCodeAt(pos - 1) !== 92 /* \ */) { - if (pos + 1 === max) { - return false; - } - if (state.src.charCodeAt(pos + 1) !== 58 /* : */) { - return false; - } - break; - } - } - endLine = state.lineMax; - // jump line-by-line until empty one or EOF - terminatorRules = state.md.block.ruler.getRules("reference"); - oldParentType = state.parentType; - state.parentType = "reference"; - for (;nextLine < endLine && !state.isEmpty(nextLine); nextLine++) { - // this would be a code block normally, but after paragraph - // it's considered a lazy continuation regardless of what's there - if (state.sCount[nextLine] - state.blkIndent > 3) { - continue; - } - // quirk for blockquotes, this line should already be checked by that rule - if (state.sCount[nextLine] < 0) { - continue; - } - // Some tags can terminate paragraph without empty line. - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - break; - } - } - str = state.getLines(startLine, nextLine, state.blkIndent, false).trim(); - max = str.length; - for (pos = 1; pos < max; pos++) { - ch = str.charCodeAt(pos); - if (ch === 91 /* [ */) { - return false; - } else if (ch === 93 /* ] */) { - labelEnd = pos; - break; - } else if (ch === 10 /* \n */) { - lines++; - } else if (ch === 92 /* \ */) { - pos++; - if (pos < max && str.charCodeAt(pos) === 10) { - lines++; - } - } - } - if (labelEnd < 0 || str.charCodeAt(labelEnd + 1) !== 58 /* : */) { - return false; - } - // [label]: destination 'title' - // ^^^ skip optional whitespace here - for (pos = labelEnd + 2; pos < max; pos++) { - ch = str.charCodeAt(pos); - if (ch === 10) { - lines++; - } else if (isSpace$6(ch)) ; else { - break; - } - } - // [label]: destination 'title' - // ^^^^^^^^^^^ parse this - res = state.md.helpers.parseLinkDestination(str, pos, max); - if (!res.ok) { - return false; - } - href = state.md.normalizeLink(res.str); - if (!state.md.validateLink(href)) { - return false; - } - pos = res.pos; - lines += res.lines; - // save cursor state, we could require to rollback later - destEndPos = pos; - destEndLineNo = lines; - // [label]: destination 'title' - // ^^^ skipping those spaces - start = pos; - for (;pos < max; pos++) { - ch = str.charCodeAt(pos); - if (ch === 10) { - lines++; - } else if (isSpace$6(ch)) ; else { - break; - } - } - // [label]: destination 'title' - // ^^^^^^^ parse this - res = state.md.helpers.parseLinkTitle(str, pos, max); - if (pos < max && start !== pos && res.ok) { - title = res.str; - pos = res.pos; - lines += res.lines; - } else { - title = ""; - pos = destEndPos; - lines = destEndLineNo; - } - // skip trailing spaces until the rest of the line - while (pos < max) { - ch = str.charCodeAt(pos); - if (!isSpace$6(ch)) { - break; - } - pos++; - } - if (pos < max && str.charCodeAt(pos) !== 10) { - if (title) { - // garbage at the end of the line after title, - // but it could still be a valid reference if we roll back - title = ""; - pos = destEndPos; - lines = destEndLineNo; - while (pos < max) { - ch = str.charCodeAt(pos); - if (!isSpace$6(ch)) { - break; - } - pos++; - } - } - } - if (pos < max && str.charCodeAt(pos) !== 10) { - // garbage at the end of the line - return false; - } - label = normalizeReference$2(str.slice(1, labelEnd)); - if (!label) { - // CommonMark 0.20 disallows empty labels - return false; - } - // Reference can not terminate anything. This check is for safety only. - /*istanbul ignore if*/ if (silent) { - return true; - } - if (typeof state.env.references === "undefined") { - state.env.references = {}; - } - if (typeof state.env.references[label] === "undefined") { - state.env.references[label] = { - title: title, - href: href - }; - } - state.parentType = oldParentType; - state.line = startLine + lines + 1; - return true; - }; - // List of valid html blocks names, accorting to commonmark spec - var html_blocks = [ "address", "article", "aside", "base", "basefont", "blockquote", "body", "caption", "center", "col", "colgroup", "dd", "details", "dialog", "dir", "div", "dl", "dt", "fieldset", "figcaption", "figure", "footer", "form", "frame", "frameset", "h1", "h2", "h3", "h4", "h5", "h6", "head", "header", "hr", "html", "iframe", "legend", "li", "link", "main", "menu", "menuitem", "nav", "noframes", "ol", "optgroup", "option", "p", "param", "section", "source", "summary", "table", "tbody", "td", "tfoot", "th", "thead", "title", "tr", "track", "ul" ]; - // Regexps to match html elements - var attr_name = "[a-zA-Z_:][a-zA-Z0-9:._-]*"; - var unquoted = "[^\"'=<>`\\x00-\\x20]+"; - var single_quoted = "'[^']*'"; - var double_quoted = '"[^"]*"'; - var attr_value = "(?:" + unquoted + "|" + single_quoted + "|" + double_quoted + ")"; - var attribute = "(?:\\s+" + attr_name + "(?:\\s*=\\s*" + attr_value + ")?)"; - var open_tag = "<[A-Za-z][A-Za-z0-9\\-]*" + attribute + "*\\s*\\/?>"; - var close_tag = "<\\/[A-Za-z][A-Za-z0-9\\-]*\\s*>"; - var comment = "\x3c!----\x3e|\x3c!--(?:-?[^>-])(?:-?[^-])*--\x3e"; - var processing = "<[?][\\s\\S]*?[?]>"; - var declaration = "]*>"; - var cdata = ""; - var HTML_TAG_RE$1 = new RegExp("^(?:" + open_tag + "|" + close_tag + "|" + comment + "|" + processing + "|" + declaration + "|" + cdata + ")"); - var HTML_OPEN_CLOSE_TAG_RE$1 = new RegExp("^(?:" + open_tag + "|" + close_tag + ")"); - var HTML_TAG_RE_1 = HTML_TAG_RE$1; - var HTML_OPEN_CLOSE_TAG_RE_1 = HTML_OPEN_CLOSE_TAG_RE$1; - var html_re = { - HTML_TAG_RE: HTML_TAG_RE_1, - HTML_OPEN_CLOSE_TAG_RE: HTML_OPEN_CLOSE_TAG_RE_1 - }; - var HTML_OPEN_CLOSE_TAG_RE = html_re.HTML_OPEN_CLOSE_TAG_RE; - // An array of opening and corresponding closing sequences for html tags, - // last argument defines whether it can terminate a paragraph or not - - var HTML_SEQUENCES = [ [ /^<(script|pre|style|textarea)(?=(\s|>|$))/i, /<\/(script|pre|style|textarea)>/i, true ], [ /^/, true ], [ /^<\?/, /\?>/, true ], [ /^/, true ], [ /^/, true ], [ new RegExp("^|$))", "i"), /^$/, true ], [ new RegExp(HTML_OPEN_CLOSE_TAG_RE.source + "\\s*$"), /^$/, false ] ]; - var html_block = function html_block(state, startLine, endLine, silent) { - var i, nextLine, token, lineText, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine]; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - if (!state.md.options.html) { - return false; - } - if (state.src.charCodeAt(pos) !== 60 /* < */) { - return false; - } - lineText = state.src.slice(pos, max); - for (i = 0; i < HTML_SEQUENCES.length; i++) { - if (HTML_SEQUENCES[i][0].test(lineText)) { - break; - } - } - if (i === HTML_SEQUENCES.length) { - return false; - } - if (silent) { - // true if this sequence can be a terminator, false otherwise - return HTML_SEQUENCES[i][2]; - } - nextLine = startLine + 1; - // If we are here - we detected HTML block. - // Let's roll down till block end. - if (!HTML_SEQUENCES[i][1].test(lineText)) { - for (;nextLine < endLine; nextLine++) { - if (state.sCount[nextLine] < state.blkIndent) { - break; - } - pos = state.bMarks[nextLine] + state.tShift[nextLine]; - max = state.eMarks[nextLine]; - lineText = state.src.slice(pos, max); - if (HTML_SEQUENCES[i][1].test(lineText)) { - if (lineText.length !== 0) { - nextLine++; - } - break; - } - } - } - state.line = nextLine; - token = state.push("html_block", "", 0); - token.map = [ startLine, nextLine ]; - token.content = state.getLines(startLine, nextLine, state.blkIndent, true); - return true; - }; - var isSpace$5 = utils.isSpace; - var heading = function heading(state, startLine, endLine, silent) { - var ch, level, tmp, token, pos = state.bMarks[startLine] + state.tShift[startLine], max = state.eMarks[startLine]; - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - ch = state.src.charCodeAt(pos); - if (ch !== 35 /* # */ || pos >= max) { - return false; - } - // count heading level - level = 1; - ch = state.src.charCodeAt(++pos); - while (ch === 35 /* # */ && pos < max && level <= 6) { - level++; - ch = state.src.charCodeAt(++pos); - } - if (level > 6 || pos < max && !isSpace$5(ch)) { - return false; - } - if (silent) { - return true; - } - // Let's cut tails like ' ### ' from the end of string - max = state.skipSpacesBack(max, pos); - tmp = state.skipCharsBack(max, 35, pos); - // # - if (tmp > pos && isSpace$5(state.src.charCodeAt(tmp - 1))) { - max = tmp; - } - state.line = startLine + 1; - token = state.push("heading_open", "h" + String(level), 1); - token.markup = "########".slice(0, level); - token.map = [ startLine, state.line ]; - token = state.push("inline", "", 0); - token.content = state.src.slice(pos, max).trim(); - token.map = [ startLine, state.line ]; - token.children = []; - token = state.push("heading_close", "h" + String(level), -1); - token.markup = "########".slice(0, level); - return true; - }; - // lheading (---, ===) - var lheading = function lheading(state, startLine, endLine /*, silent*/) { - var content, terminate, i, l, token, pos, max, level, marker, nextLine = startLine + 1, oldParentType, terminatorRules = state.md.block.ruler.getRules("paragraph"); - // if it's indented more than 3 spaces, it should be a code block - if (state.sCount[startLine] - state.blkIndent >= 4) { - return false; - } - oldParentType = state.parentType; - state.parentType = "paragraph"; - // use paragraph to match terminatorRules - // jump line-by-line until empty one or EOF - for (;nextLine < endLine && !state.isEmpty(nextLine); nextLine++) { - // this would be a code block normally, but after paragraph - // it's considered a lazy continuation regardless of what's there - if (state.sCount[nextLine] - state.blkIndent > 3) { - continue; - } - - // Check for underline in setext header - - if (state.sCount[nextLine] >= state.blkIndent) { - pos = state.bMarks[nextLine] + state.tShift[nextLine]; - max = state.eMarks[nextLine]; - if (pos < max) { - marker = state.src.charCodeAt(pos); - if (marker === 45 /* - */ || marker === 61 /* = */) { - pos = state.skipChars(pos, marker); - pos = state.skipSpaces(pos); - if (pos >= max) { - level = marker === 61 /* = */ ? 1 : 2; - break; - } - } - } - } - // quirk for blockquotes, this line should already be checked by that rule - if (state.sCount[nextLine] < 0) { - continue; - } - // Some tags can terminate paragraph without empty line. - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - break; - } - } - if (!level) { - // Didn't find valid underline - return false; - } - content = state.getLines(startLine, nextLine, state.blkIndent, false).trim(); - state.line = nextLine + 1; - token = state.push("heading_open", "h" + String(level), 1); - token.markup = String.fromCharCode(marker); - token.map = [ startLine, state.line ]; - token = state.push("inline", "", 0); - token.content = content; - token.map = [ startLine, state.line - 1 ]; - token.children = []; - token = state.push("heading_close", "h" + String(level), -1); - token.markup = String.fromCharCode(marker); - state.parentType = oldParentType; - return true; - }; - // Paragraph - var paragraph = function paragraph(state, startLine, endLine) { - var content, terminate, i, l, token, oldParentType, nextLine = startLine + 1, terminatorRules = state.md.block.ruler.getRules("paragraph"); - oldParentType = state.parentType; - state.parentType = "paragraph"; - // jump line-by-line until empty one or EOF - for (;nextLine < endLine && !state.isEmpty(nextLine); nextLine++) { - // this would be a code block normally, but after paragraph - // it's considered a lazy continuation regardless of what's there - if (state.sCount[nextLine] - state.blkIndent > 3) { - continue; - } - // quirk for blockquotes, this line should already be checked by that rule - if (state.sCount[nextLine] < 0) { - continue; - } - // Some tags can terminate paragraph without empty line. - terminate = false; - for (i = 0, l = terminatorRules.length; i < l; i++) { - if (terminatorRules[i](state, nextLine, endLine, true)) { - terminate = true; - break; - } - } - if (terminate) { - break; - } - } - content = state.getLines(startLine, nextLine, state.blkIndent, false).trim(); - state.line = nextLine; - token = state.push("paragraph_open", "p", 1); - token.map = [ startLine, state.line ]; - token = state.push("inline", "", 0); - token.content = content; - token.map = [ startLine, state.line ]; - token.children = []; - token = state.push("paragraph_close", "p", -1); - state.parentType = oldParentType; - return true; - }; - var isSpace$4 = utils.isSpace; - function StateBlock(src, md, env, tokens) { - var ch, s, start, pos, len, indent, offset, indent_found; - this.src = src; - // link to parser instance - this.md = md; - this.env = env; - - // Internal state vartiables - - this.tokens = tokens; - this.bMarks = []; - // line begin offsets for fast jumps - this.eMarks = []; - // line end offsets for fast jumps - this.tShift = []; - // offsets of the first non-space characters (tabs not expanded) - this.sCount = []; - // indents for each line (tabs expanded) - // An amount of virtual spaces (tabs expanded) between beginning - // of each line (bMarks) and real beginning of that line. - - // It exists only as a hack because blockquotes override bMarks - // losing information in the process. - - // It's used only when expanding tabs, you can think about it as - // an initial tab length, e.g. bsCount=21 applied to string `\t123` - // means first tab should be expanded to 4-21%4 === 3 spaces. - - this.bsCount = []; - // block parser variables - this.blkIndent = 0; - // required block content indent (for example, if we are - // inside a list, it would be positioned after list marker) - this.line = 0; - // line index in src - this.lineMax = 0; - // lines count - this.tight = false; - // loose/tight mode for lists - this.ddIndent = -1; - // indent of the current dd block (-1 if there isn't any) - this.listIndent = -1; - // indent of the current list block (-1 if there isn't any) - // can be 'blockquote', 'list', 'root', 'paragraph' or 'reference' - // used in lists to determine if they interrupt a paragraph - this.parentType = "root"; - this.level = 0; - // renderer - this.result = ""; - // Create caches - // Generate markers. - s = this.src; - indent_found = false; - for (start = pos = indent = offset = 0, len = s.length; pos < len; pos++) { - ch = s.charCodeAt(pos); - if (!indent_found) { - if (isSpace$4(ch)) { - indent++; - if (ch === 9) { - offset += 4 - offset % 4; - } else { - offset++; - } - continue; - } else { - indent_found = true; - } - } - if (ch === 10 || pos === len - 1) { - if (ch !== 10) { - pos++; - } - this.bMarks.push(start); - this.eMarks.push(pos); - this.tShift.push(indent); - this.sCount.push(offset); - this.bsCount.push(0); - indent_found = false; - indent = 0; - offset = 0; - start = pos + 1; - } - } - // Push fake entry to simplify cache bounds checks - this.bMarks.push(s.length); - this.eMarks.push(s.length); - this.tShift.push(0); - this.sCount.push(0); - this.bsCount.push(0); - this.lineMax = this.bMarks.length - 1; - // don't count last fake line - } - // Push new token to "stream". - - StateBlock.prototype.push = function(type, tag, nesting) { - var token$1 = new token(type, tag, nesting); - token$1.block = true; - if (nesting < 0) this.level--; - // closing tag - token$1.level = this.level; - if (nesting > 0) this.level++; - // opening tag - this.tokens.push(token$1); - return token$1; - }; - StateBlock.prototype.isEmpty = function isEmpty(line) { - return this.bMarks[line] + this.tShift[line] >= this.eMarks[line]; - }; - StateBlock.prototype.skipEmptyLines = function skipEmptyLines(from) { - for (var max = this.lineMax; from < max; from++) { - if (this.bMarks[from] + this.tShift[from] < this.eMarks[from]) { - break; - } - } - return from; - }; - // Skip spaces from given position. - StateBlock.prototype.skipSpaces = function skipSpaces(pos) { - var ch; - for (var max = this.src.length; pos < max; pos++) { - ch = this.src.charCodeAt(pos); - if (!isSpace$4(ch)) { - break; - } - } - return pos; - }; - // Skip spaces from given position in reverse. - StateBlock.prototype.skipSpacesBack = function skipSpacesBack(pos, min) { - if (pos <= min) { - return pos; - } - while (pos > min) { - if (!isSpace$4(this.src.charCodeAt(--pos))) { - return pos + 1; - } - } - return pos; - }; - // Skip char codes from given position - StateBlock.prototype.skipChars = function skipChars(pos, code) { - for (var max = this.src.length; pos < max; pos++) { - if (this.src.charCodeAt(pos) !== code) { - break; - } - } - return pos; - }; - // Skip char codes reverse from given position - 1 - StateBlock.prototype.skipCharsBack = function skipCharsBack(pos, code, min) { - if (pos <= min) { - return pos; - } - while (pos > min) { - if (code !== this.src.charCodeAt(--pos)) { - return pos + 1; - } - } - return pos; - }; - // cut lines range from source. - StateBlock.prototype.getLines = function getLines(begin, end, indent, keepLastLF) { - var i, lineIndent, ch, first, last, queue, lineStart, line = begin; - if (begin >= end) { - return ""; - } - queue = new Array(end - begin); - for (i = 0; line < end; line++, i++) { - lineIndent = 0; - lineStart = first = this.bMarks[line]; - if (line + 1 < end || keepLastLF) { - // No need for bounds check because we have fake entry on tail. - last = this.eMarks[line] + 1; - } else { - last = this.eMarks[line]; - } - while (first < last && lineIndent < indent) { - ch = this.src.charCodeAt(first); - if (isSpace$4(ch)) { - if (ch === 9) { - lineIndent += 4 - (lineIndent + this.bsCount[line]) % 4; - } else { - lineIndent++; - } - } else if (first - lineStart < this.tShift[line]) { - // patched tShift masked characters to look like spaces (blockquotes, list markers) - lineIndent++; - } else { - break; - } - first++; - } - if (lineIndent > indent) { - // partially expanding tabs in code blocks, e.g '\t\tfoobar' - // with indent=2 becomes ' \tfoobar' - queue[i] = new Array(lineIndent - indent + 1).join(" ") + this.src.slice(first, last); - } else { - queue[i] = this.src.slice(first, last); - } - } - return queue.join(""); - }; - // re-export Token class to use in block rules - StateBlock.prototype.Token = token; - var state_block = StateBlock; - var _rules$1 = [ - // First 2 params - rule name & source. Secondary array - list of rules, - // which can be terminated by this one. - [ "table", table, [ "paragraph", "reference" ] ], [ "code", code ], [ "fence", fence, [ "paragraph", "reference", "blockquote", "list" ] ], [ "blockquote", blockquote, [ "paragraph", "reference", "blockquote", "list" ] ], [ "hr", hr, [ "paragraph", "reference", "blockquote", "list" ] ], [ "list", list, [ "paragraph", "reference", "blockquote" ] ], [ "reference", reference ], [ "html_block", html_block, [ "paragraph", "reference", "blockquote" ] ], [ "heading", heading, [ "paragraph", "reference", "blockquote" ] ], [ "lheading", lheading ], [ "paragraph", paragraph ] ]; - /** - * new ParserBlock() - **/ function ParserBlock() { - /** - * ParserBlock#ruler -> Ruler - * - * [[Ruler]] instance. Keep configuration of block rules. - **/ - this.ruler = new ruler; - for (var i = 0; i < _rules$1.length; i++) { - this.ruler.push(_rules$1[i][0], _rules$1[i][1], { - alt: (_rules$1[i][2] || []).slice() - }); - } - } - // Generate tokens for input range - - ParserBlock.prototype.tokenize = function(state, startLine, endLine) { - var ok, i, prevLine, rules = this.ruler.getRules(""), len = rules.length, line = startLine, hasEmptyLines = false, maxNesting = state.md.options.maxNesting; - while (line < endLine) { - state.line = line = state.skipEmptyLines(line); - if (line >= endLine) { - break; - } - // Termination condition for nested calls. - // Nested calls currently used for blockquotes & lists - if (state.sCount[line] < state.blkIndent) { - break; - } - // If nesting level exceeded - skip tail to the end. That's not ordinary - // situation and we should not care about content. - if (state.level >= maxNesting) { - state.line = endLine; - break; - } - // Try all possible rules. - // On success, rule should: - - // - update `state.line` - // - update `state.tokens` - // - return true - prevLine = state.line; - for (i = 0; i < len; i++) { - ok = rules[i](state, line, endLine, false); - if (ok) { - if (prevLine >= state.line) { - throw new Error("block rule didn't increment state.line"); - } - break; - } - } - // this can only happen if user disables paragraph rule - if (!ok) throw new Error("none of the block rules matched"); - // set state.tight if we had an empty line before current tag - // i.e. latest empty line should not count - state.tight = !hasEmptyLines; - // paragraph might "eat" one newline after it in nested lists - if (state.isEmpty(state.line - 1)) { - hasEmptyLines = true; - } - line = state.line; - if (line < endLine && state.isEmpty(line)) { - hasEmptyLines = true; - line++; - state.line = line; - } - } - }; - /** - * ParserBlock.parse(str, md, env, outTokens) - * - * Process input string and push block tokens into `outTokens` - **/ ParserBlock.prototype.parse = function(src, md, env, outTokens) { - var state; - if (!src) { - return; - } - state = new this.State(src, md, env, outTokens); - this.tokenize(state, state.line, state.lineMax); - }; - ParserBlock.prototype.State = state_block; - var parser_block = ParserBlock; - // Skip text characters for text token, place those to pending buffer - // Rule to skip pure text - // '{}$%@~+=:' reserved for extentions - // !, ", #, $, %, &, ', (, ), *, +, ,, -, ., /, :, ;, <, =, >, ?, @, [, \, ], ^, _, `, {, |, }, or ~ - // !!!! Don't confuse with "Markdown ASCII Punctuation" chars - // http://spec.commonmark.org/0.15/#ascii-punctuation-character - function isTerminatorChar(ch) { - switch (ch) { - case 10 /* \n */ : - case 33 /* ! */ : - case 35 /* # */ : - case 36 /* $ */ : - case 37 /* % */ : - case 38 /* & */ : - case 42 /* * */ : - case 43 /* + */ : - case 45 /* - */ : - case 58 /* : */ : - case 60 /* < */ : - case 61 /* = */ : - case 62 /* > */ : - case 64 /* @ */ : - case 91 /* [ */ : - case 92 /* \ */ : - case 93 /* ] */ : - case 94 /* ^ */ : - case 95 /* _ */ : - case 96 /* ` */ : - case 123 /* { */ : - case 125 /* } */ : - case 126 /* ~ */ : - return true; - - default: - return false; - } - } - var text = function text(state, silent) { - var pos = state.pos; - while (pos < state.posMax && !isTerminatorChar(state.src.charCodeAt(pos))) { - pos++; - } - if (pos === state.pos) { - return false; - } - if (!silent) { - state.pending += state.src.slice(state.pos, pos); - } - state.pos = pos; - return true; - }; - // Process links like https://example.org/ - // RFC3986: scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) - var SCHEME_RE = /(?:^|[^a-z0-9.+-])([a-z][a-z0-9.+-]*)$/i; - var linkify = function linkify(state, silent) { - var pos, max, match, proto, link, url, fullUrl, token; - if (!state.md.options.linkify) return false; - if (state.linkLevel > 0) return false; - pos = state.pos; - max = state.posMax; - if (pos + 3 > max) return false; - if (state.src.charCodeAt(pos) !== 58 /* : */) return false; - if (state.src.charCodeAt(pos + 1) !== 47 /* / */) return false; - if (state.src.charCodeAt(pos + 2) !== 47 /* / */) return false; - match = state.pending.match(SCHEME_RE); - if (!match) return false; - proto = match[1]; - link = state.md.linkify.matchAtStart(state.src.slice(pos - proto.length)); - if (!link) return false; - url = link.url; - // invalid link, but still detected by linkify somehow; - // need to check to prevent infinite loop below - if (url.length <= proto.length) return false; - // disallow '*' at the end of the link (conflicts with emphasis) - url = url.replace(/\*+$/, ""); - fullUrl = state.md.normalizeLink(url); - if (!state.md.validateLink(fullUrl)) return false; - if (!silent) { - state.pending = state.pending.slice(0, -proto.length); - token = state.push("link_open", "a", 1); - token.attrs = [ [ "href", fullUrl ] ]; - token.markup = "linkify"; - token.info = "auto"; - token = state.push("text", "", 0); - token.content = state.md.normalizeLinkText(url); - token = state.push("link_close", "a", -1); - token.markup = "linkify"; - token.info = "auto"; - } - state.pos += url.length - proto.length; - return true; - }; - var isSpace$3 = utils.isSpace; - var newline = function newline(state, silent) { - var pmax, max, ws, pos = state.pos; - if (state.src.charCodeAt(pos) !== 10 /* \n */) { - return false; - } - pmax = state.pending.length - 1; - max = state.posMax; - // ' \n' -> hardbreak - // Lookup in pending chars is bad practice! Don't copy to other rules! - // Pending string is stored in concat mode, indexed lookups will cause - // convertion to flat mode. - if (!silent) { - if (pmax >= 0 && state.pending.charCodeAt(pmax) === 32) { - if (pmax >= 1 && state.pending.charCodeAt(pmax - 1) === 32) { - // Find whitespaces tail of pending chars. - ws = pmax - 1; - while (ws >= 1 && state.pending.charCodeAt(ws - 1) === 32) ws--; - state.pending = state.pending.slice(0, ws); - state.push("hardbreak", "br", 0); - } else { - state.pending = state.pending.slice(0, -1); - state.push("softbreak", "br", 0); - } - } else { - state.push("softbreak", "br", 0); - } - } - pos++; - // skip heading spaces for next line - while (pos < max && isSpace$3(state.src.charCodeAt(pos))) { - pos++; - } - state.pos = pos; - return true; - }; - var isSpace$2 = utils.isSpace; - var ESCAPED = []; - for (var i = 0; i < 256; i++) { - ESCAPED.push(0); - } - "\\!\"#$%&'()*+,./:;<=>?@[]^_`{|}~-".split("").forEach((function(ch) { - ESCAPED[ch.charCodeAt(0)] = 1; - })); - var _escape = function escape(state, silent) { - var ch1, ch2, origStr, escapedStr, token, pos = state.pos, max = state.posMax; - if (state.src.charCodeAt(pos) !== 92 /* \ */) return false; - pos++; - // '\' at the end of the inline block - if (pos >= max) return false; - ch1 = state.src.charCodeAt(pos); - if (ch1 === 10) { - if (!silent) { - state.push("hardbreak", "br", 0); - } - pos++; - // skip leading whitespaces from next line - while (pos < max) { - ch1 = state.src.charCodeAt(pos); - if (!isSpace$2(ch1)) break; - pos++; - } - state.pos = pos; - return true; - } - escapedStr = state.src[pos]; - if (ch1 >= 55296 && ch1 <= 56319 && pos + 1 < max) { - ch2 = state.src.charCodeAt(pos + 1); - if (ch2 >= 56320 && ch2 <= 57343) { - escapedStr += state.src[pos + 1]; - pos++; - } - } - origStr = "\\" + escapedStr; - if (!silent) { - token = state.push("text_special", "", 0); - if (ch1 < 256 && ESCAPED[ch1] !== 0) { - token.content = escapedStr; - } else { - token.content = origStr; - } - token.markup = origStr; - token.info = "escape"; - } - state.pos = pos + 1; - return true; - }; - // Parse backticks - var backticks = function backtick(state, silent) { - var start, max, marker, token, matchStart, matchEnd, openerLength, closerLength, pos = state.pos, ch = state.src.charCodeAt(pos); - if (ch !== 96 /* ` */) { - return false; - } - start = pos; - pos++; - max = state.posMax; - // scan marker length - while (pos < max && state.src.charCodeAt(pos) === 96 /* ` */) { - pos++; - } - marker = state.src.slice(start, pos); - openerLength = marker.length; - if (state.backticksScanned && (state.backticks[openerLength] || 0) <= start) { - if (!silent) state.pending += marker; - state.pos += openerLength; - return true; - } - matchEnd = pos; - // Nothing found in the cache, scan until the end of the line (or until marker is found) - while ((matchStart = state.src.indexOf("`", matchEnd)) !== -1) { - matchEnd = matchStart + 1; - // scan marker length - while (matchEnd < max && state.src.charCodeAt(matchEnd) === 96 /* ` */) { - matchEnd++; - } - closerLength = matchEnd - matchStart; - if (closerLength === openerLength) { - // Found matching closer length. - if (!silent) { - token = state.push("code_inline", "code", 0); - token.markup = marker; - token.content = state.src.slice(pos, matchStart).replace(/\n/g, " ").replace(/^ (.+) $/, "$1"); - } - state.pos = matchEnd; - return true; - } - // Some different length found, put it in cache as upper limit of where closer can be found - state.backticks[closerLength] = matchStart; - } - // Scanned through the end, didn't find anything - state.backticksScanned = true; - if (!silent) state.pending += marker; - state.pos += openerLength; - return true; - }; - // ~~strike through~~ - // Insert each marker as a separate text token, and add it to delimiter list - - var tokenize$1 = function strikethrough(state, silent) { - var i, scanned, token, len, ch, start = state.pos, marker = state.src.charCodeAt(start); - if (silent) { - return false; - } - if (marker !== 126 /* ~ */) { - return false; - } - scanned = state.scanDelims(state.pos, true); - len = scanned.length; - ch = String.fromCharCode(marker); - if (len < 2) { - return false; - } - if (len % 2) { - token = state.push("text", "", 0); - token.content = ch; - len--; - } - for (i = 0; i < len; i += 2) { - token = state.push("text", "", 0); - token.content = ch + ch; - state.delimiters.push({ - marker: marker, - length: 0, - // disable "rule of 3" length checks meant for emphasis - token: state.tokens.length - 1, - end: -1, - open: scanned.can_open, - close: scanned.can_close - }); - } - state.pos += scanned.length; - return true; - }; - function postProcess$1(state, delimiters) { - var i, j, startDelim, endDelim, token, loneMarkers = [], max = delimiters.length; - for (i = 0; i < max; i++) { - startDelim = delimiters[i]; - if (startDelim.marker !== 126 /* ~ */) { - continue; - } - if (startDelim.end === -1) { - continue; - } - endDelim = delimiters[startDelim.end]; - token = state.tokens[startDelim.token]; - token.type = "s_open"; - token.tag = "s"; - token.nesting = 1; - token.markup = "~~"; - token.content = ""; - token = state.tokens[endDelim.token]; - token.type = "s_close"; - token.tag = "s"; - token.nesting = -1; - token.markup = "~~"; - token.content = ""; - if (state.tokens[endDelim.token - 1].type === "text" && state.tokens[endDelim.token - 1].content === "~") { - loneMarkers.push(endDelim.token - 1); - } - } - // If a marker sequence has an odd number of characters, it's splitted - // like this: `~~~~~` -> `~` + `~~` + `~~`, leaving one marker at the - // start of the sequence. - - // So, we have to move all those markers after subsequent s_close tags. - - while (loneMarkers.length) { - i = loneMarkers.pop(); - j = i + 1; - while (j < state.tokens.length && state.tokens[j].type === "s_close") { - j++; - } - j--; - if (i !== j) { - token = state.tokens[j]; - state.tokens[j] = state.tokens[i]; - state.tokens[i] = token; - } - } - } - // Walk through delimiter list and replace text tokens with tags - - var postProcess_1$1 = function strikethrough(state) { - var curr, tokens_meta = state.tokens_meta, max = state.tokens_meta.length; - postProcess$1(state, state.delimiters); - for (curr = 0; curr < max; curr++) { - if (tokens_meta[curr] && tokens_meta[curr].delimiters) { - postProcess$1(state, tokens_meta[curr].delimiters); - } - } - }; - var strikethrough = { - tokenize: tokenize$1, - postProcess: postProcess_1$1 - }; - // Process *this* and _that_ - // Insert each marker as a separate text token, and add it to delimiter list - - var tokenize = function emphasis(state, silent) { - var i, scanned, token, start = state.pos, marker = state.src.charCodeAt(start); - if (silent) { - return false; - } - if (marker !== 95 /* _ */ && marker !== 42 /* * */) { - return false; - } - scanned = state.scanDelims(state.pos, marker === 42); - for (i = 0; i < scanned.length; i++) { - token = state.push("text", "", 0); - token.content = String.fromCharCode(marker); - state.delimiters.push({ - // Char code of the starting marker (number). - marker: marker, - // Total length of these series of delimiters. - length: scanned.length, - // A position of the token this delimiter corresponds to. - token: state.tokens.length - 1, - // If this delimiter is matched as a valid opener, `end` will be - // equal to its position, otherwise it's `-1`. - end: -1, - // Boolean flags that determine if this delimiter could open or close - // an emphasis. - open: scanned.can_open, - close: scanned.can_close - }); - } - state.pos += scanned.length; - return true; - }; - function postProcess(state, delimiters) { - var i, startDelim, endDelim, token, ch, isStrong, max = delimiters.length; - for (i = max - 1; i >= 0; i--) { - startDelim = delimiters[i]; - if (startDelim.marker !== 95 /* _ */ && startDelim.marker !== 42 /* * */) { - continue; - } - // Process only opening markers - if (startDelim.end === -1) { - continue; - } - endDelim = delimiters[startDelim.end]; - // If the previous delimiter has the same marker and is adjacent to this one, - // merge those into one strong delimiter. - - // `whatever` -> `whatever` - - isStrong = i > 0 && delimiters[i - 1].end === startDelim.end + 1 && - // check that first two markers match and adjacent - delimiters[i - 1].marker === startDelim.marker && delimiters[i - 1].token === startDelim.token - 1 && - // check that last two markers are adjacent (we can safely assume they match) - delimiters[startDelim.end + 1].token === endDelim.token + 1; - ch = String.fromCharCode(startDelim.marker); - token = state.tokens[startDelim.token]; - token.type = isStrong ? "strong_open" : "em_open"; - token.tag = isStrong ? "strong" : "em"; - token.nesting = 1; - token.markup = isStrong ? ch + ch : ch; - token.content = ""; - token = state.tokens[endDelim.token]; - token.type = isStrong ? "strong_close" : "em_close"; - token.tag = isStrong ? "strong" : "em"; - token.nesting = -1; - token.markup = isStrong ? ch + ch : ch; - token.content = ""; - if (isStrong) { - state.tokens[delimiters[i - 1].token].content = ""; - state.tokens[delimiters[startDelim.end + 1].token].content = ""; - i--; - } - } - } - // Walk through delimiter list and replace text tokens with tags - - var postProcess_1 = function emphasis(state) { - var curr, tokens_meta = state.tokens_meta, max = state.tokens_meta.length; - postProcess(state, state.delimiters); - for (curr = 0; curr < max; curr++) { - if (tokens_meta[curr] && tokens_meta[curr].delimiters) { - postProcess(state, tokens_meta[curr].delimiters); - } - } - }; - var emphasis = { - tokenize: tokenize, - postProcess: postProcess_1 - }; - var normalizeReference$1 = utils.normalizeReference; - var isSpace$1 = utils.isSpace; - var link = function link(state, silent) { - var attrs, code, label, labelEnd, labelStart, pos, res, ref, token, href = "", title = "", oldPos = state.pos, max = state.posMax, start = state.pos, parseReference = true; - if (state.src.charCodeAt(state.pos) !== 91 /* [ */) { - return false; - } - labelStart = state.pos + 1; - labelEnd = state.md.helpers.parseLinkLabel(state, state.pos, true); - // parser failed to find ']', so it's not a valid link - if (labelEnd < 0) { - return false; - } - pos = labelEnd + 1; - if (pos < max && state.src.charCodeAt(pos) === 40 /* ( */) { - // Inline link - // might have found a valid shortcut link, disable reference parsing - parseReference = false; - // [link]( "title" ) - // ^^ skipping these spaces - pos++; - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace$1(code) && code !== 10) { - break; - } - } - if (pos >= max) { - return false; - } - // [link]( "title" ) - // ^^^^^^ parsing link destination - start = pos; - res = state.md.helpers.parseLinkDestination(state.src, pos, state.posMax); - if (res.ok) { - href = state.md.normalizeLink(res.str); - if (state.md.validateLink(href)) { - pos = res.pos; - } else { - href = ""; - } - // [link]( "title" ) - // ^^ skipping these spaces - start = pos; - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace$1(code) && code !== 10) { - break; - } - } - // [link]( "title" ) - // ^^^^^^^ parsing link title - res = state.md.helpers.parseLinkTitle(state.src, pos, state.posMax); - if (pos < max && start !== pos && res.ok) { - title = res.str; - pos = res.pos; - // [link]( "title" ) - // ^^ skipping these spaces - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace$1(code) && code !== 10) { - break; - } - } - } - } - if (pos >= max || state.src.charCodeAt(pos) !== 41 /* ) */) { - // parsing a valid shortcut link failed, fallback to reference - parseReference = true; - } - pos++; - } - if (parseReference) { - // Link reference - if (typeof state.env.references === "undefined") { - return false; - } - if (pos < max && state.src.charCodeAt(pos) === 91 /* [ */) { - start = pos + 1; - pos = state.md.helpers.parseLinkLabel(state, pos); - if (pos >= 0) { - label = state.src.slice(start, pos++); - } else { - pos = labelEnd + 1; - } - } else { - pos = labelEnd + 1; - } - // covers label === '' and label === undefined - // (collapsed reference link and shortcut reference link respectively) - if (!label) { - label = state.src.slice(labelStart, labelEnd); - } - ref = state.env.references[normalizeReference$1(label)]; - if (!ref) { - state.pos = oldPos; - return false; - } - href = ref.href; - title = ref.title; - } - - // We found the end of the link, and know for a fact it's a valid link; - // so all that's left to do is to call tokenizer. - - if (!silent) { - state.pos = labelStart; - state.posMax = labelEnd; - token = state.push("link_open", "a", 1); - token.attrs = attrs = [ [ "href", href ] ]; - if (title) { - attrs.push([ "title", title ]); - } - state.linkLevel++; - state.md.inline.tokenize(state); - state.linkLevel--; - token = state.push("link_close", "a", -1); - } - state.pos = pos; - state.posMax = max; - return true; - }; - var normalizeReference = utils.normalizeReference; - var isSpace = utils.isSpace; - var image = function image(state, silent) { - var attrs, code, content, label, labelEnd, labelStart, pos, ref, res, title, token, tokens, start, href = "", oldPos = state.pos, max = state.posMax; - if (state.src.charCodeAt(state.pos) !== 33 /* ! */) { - return false; - } - if (state.src.charCodeAt(state.pos + 1) !== 91 /* [ */) { - return false; - } - labelStart = state.pos + 2; - labelEnd = state.md.helpers.parseLinkLabel(state, state.pos + 1, false); - // parser failed to find ']', so it's not a valid link - if (labelEnd < 0) { - return false; - } - pos = labelEnd + 1; - if (pos < max && state.src.charCodeAt(pos) === 40 /* ( */) { - // Inline link - // [link]( "title" ) - // ^^ skipping these spaces - pos++; - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace(code) && code !== 10) { - break; - } - } - if (pos >= max) { - return false; - } - // [link]( "title" ) - // ^^^^^^ parsing link destination - start = pos; - res = state.md.helpers.parseLinkDestination(state.src, pos, state.posMax); - if (res.ok) { - href = state.md.normalizeLink(res.str); - if (state.md.validateLink(href)) { - pos = res.pos; - } else { - href = ""; - } - } - // [link]( "title" ) - // ^^ skipping these spaces - start = pos; - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace(code) && code !== 10) { - break; - } - } - // [link]( "title" ) - // ^^^^^^^ parsing link title - res = state.md.helpers.parseLinkTitle(state.src, pos, state.posMax); - if (pos < max && start !== pos && res.ok) { - title = res.str; - pos = res.pos; - // [link]( "title" ) - // ^^ skipping these spaces - for (;pos < max; pos++) { - code = state.src.charCodeAt(pos); - if (!isSpace(code) && code !== 10) { - break; - } - } - } else { - title = ""; - } - if (pos >= max || state.src.charCodeAt(pos) !== 41 /* ) */) { - state.pos = oldPos; - return false; - } - pos++; - } else { - // Link reference - if (typeof state.env.references === "undefined") { - return false; - } - if (pos < max && state.src.charCodeAt(pos) === 91 /* [ */) { - start = pos + 1; - pos = state.md.helpers.parseLinkLabel(state, pos); - if (pos >= 0) { - label = state.src.slice(start, pos++); - } else { - pos = labelEnd + 1; - } - } else { - pos = labelEnd + 1; - } - // covers label === '' and label === undefined - // (collapsed reference link and shortcut reference link respectively) - if (!label) { - label = state.src.slice(labelStart, labelEnd); - } - ref = state.env.references[normalizeReference(label)]; - if (!ref) { - state.pos = oldPos; - return false; - } - href = ref.href; - title = ref.title; - } - - // We found the end of the link, and know for a fact it's a valid link; - // so all that's left to do is to call tokenizer. - - if (!silent) { - content = state.src.slice(labelStart, labelEnd); - state.md.inline.parse(content, state.md, state.env, tokens = []); - token = state.push("image", "img", 0); - token.attrs = attrs = [ [ "src", href ], [ "alt", "" ] ]; - token.children = tokens; - token.content = content; - if (title) { - attrs.push([ "title", title ]); - } - } - state.pos = pos; - state.posMax = max; - return true; - }; - // Process autolinks '' - /*eslint max-len:0*/ var EMAIL_RE = /^([a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*)$/; - var AUTOLINK_RE = /^([a-zA-Z][a-zA-Z0-9+.\-]{1,31}):([^<>\x00-\x20]*)$/; - var autolink = function autolink(state, silent) { - var url, fullUrl, token, ch, start, max, pos = state.pos; - if (state.src.charCodeAt(pos) !== 60 /* < */) { - return false; - } - start = state.pos; - max = state.posMax; - for (;;) { - if (++pos >= max) return false; - ch = state.src.charCodeAt(pos); - if (ch === 60 /* < */) return false; - if (ch === 62 /* > */) break; - } - url = state.src.slice(start + 1, pos); - if (AUTOLINK_RE.test(url)) { - fullUrl = state.md.normalizeLink(url); - if (!state.md.validateLink(fullUrl)) { - return false; - } - if (!silent) { - token = state.push("link_open", "a", 1); - token.attrs = [ [ "href", fullUrl ] ]; - token.markup = "autolink"; - token.info = "auto"; - token = state.push("text", "", 0); - token.content = state.md.normalizeLinkText(url); - token = state.push("link_close", "a", -1); - token.markup = "autolink"; - token.info = "auto"; - } - state.pos += url.length + 2; - return true; - } - if (EMAIL_RE.test(url)) { - fullUrl = state.md.normalizeLink("mailto:" + url); - if (!state.md.validateLink(fullUrl)) { - return false; - } - if (!silent) { - token = state.push("link_open", "a", 1); - token.attrs = [ [ "href", fullUrl ] ]; - token.markup = "autolink"; - token.info = "auto"; - token = state.push("text", "", 0); - token.content = state.md.normalizeLinkText(url); - token = state.push("link_close", "a", -1); - token.markup = "autolink"; - token.info = "auto"; - } - state.pos += url.length + 2; - return true; - } - return false; - }; - var HTML_TAG_RE = html_re.HTML_TAG_RE; - function isLinkOpen(str) { - return /^\s]/i.test(str); - } - function isLinkClose(str) { - return /^<\/a\s*>/i.test(str); - } - function isLetter(ch) { - /*eslint no-bitwise:0*/ - var lc = ch | 32; - // to lower case - return lc >= 97 /* a */ && lc <= 122 /* z */; - } - var html_inline = function html_inline(state, silent) { - var ch, match, max, token, pos = state.pos; - if (!state.md.options.html) { - return false; - } - // Check start - max = state.posMax; - if (state.src.charCodeAt(pos) !== 60 /* < */ || pos + 2 >= max) { - return false; - } - // Quick fail on second char - ch = state.src.charCodeAt(pos + 1); - if (ch !== 33 /* ! */ && ch !== 63 /* ? */ && ch !== 47 /* / */ && !isLetter(ch)) { - return false; - } - match = state.src.slice(pos).match(HTML_TAG_RE); - if (!match) { - return false; - } - if (!silent) { - token = state.push("html_inline", "", 0); - token.content = match[0]; - if (isLinkOpen(token.content)) state.linkLevel++; - if (isLinkClose(token.content)) state.linkLevel--; - } - state.pos += match[0].length; - return true; - }; - var has = utils.has; - var isValidEntityCode = utils.isValidEntityCode; - var fromCodePoint = utils.fromCodePoint; - var DIGITAL_RE = /^&#((?:x[a-f0-9]{1,6}|[0-9]{1,7}));/i; - var NAMED_RE = /^&([a-z][a-z0-9]{1,31});/i; - var entity = function entity(state, silent) { - var ch, code, match, token, pos = state.pos, max = state.posMax; - if (state.src.charCodeAt(pos) !== 38 /* & */) return false; - if (pos + 1 >= max) return false; - ch = state.src.charCodeAt(pos + 1); - if (ch === 35 /* # */) { - match = state.src.slice(pos).match(DIGITAL_RE); - if (match) { - if (!silent) { - code = match[1][0].toLowerCase() === "x" ? parseInt(match[1].slice(1), 16) : parseInt(match[1], 10); - token = state.push("text_special", "", 0); - token.content = isValidEntityCode(code) ? fromCodePoint(code) : fromCodePoint(65533); - token.markup = match[0]; - token.info = "entity"; - } - state.pos += match[0].length; - return true; - } - } else { - match = state.src.slice(pos).match(NAMED_RE); - if (match) { - if (has(entities, match[1])) { - if (!silent) { - token = state.push("text_special", "", 0); - token.content = entities[match[1]]; - token.markup = match[0]; - token.info = "entity"; - } - state.pos += match[0].length; - return true; - } - } - } - return false; - }; - // For each opening emphasis-like marker find a matching closing one - function processDelimiters(delimiters) { - var closerIdx, openerIdx, closer, opener, minOpenerIdx, newMinOpenerIdx, isOddMatch, lastJump, openersBottom = {}, max = delimiters.length; - if (!max) return; - // headerIdx is the first delimiter of the current (where closer is) delimiter run - var headerIdx = 0; - var lastTokenIdx = -2; - // needs any value lower than -1 - var jumps = []; - for (closerIdx = 0; closerIdx < max; closerIdx++) { - closer = delimiters[closerIdx]; - jumps.push(0); - // markers belong to same delimiter run if: - // - they have adjacent tokens - // - AND markers are the same - - if (delimiters[headerIdx].marker !== closer.marker || lastTokenIdx !== closer.token - 1) { - headerIdx = closerIdx; - } - lastTokenIdx = closer.token; - // Length is only used for emphasis-specific "rule of 3", - // if it's not defined (in strikethrough or 3rd party plugins), - // we can default it to 0 to disable those checks. - - closer.length = closer.length || 0; - if (!closer.close) continue; - // Previously calculated lower bounds (previous fails) - // for each marker, each delimiter length modulo 3, - // and for whether this closer can be an opener; - // https://github.com/commonmark/cmark/commit/34250e12ccebdc6372b8b49c44fab57c72443460 - if (!openersBottom.hasOwnProperty(closer.marker)) { - openersBottom[closer.marker] = [ -1, -1, -1, -1, -1, -1 ]; - } - minOpenerIdx = openersBottom[closer.marker][(closer.open ? 3 : 0) + closer.length % 3]; - openerIdx = headerIdx - jumps[headerIdx] - 1; - newMinOpenerIdx = openerIdx; - for (;openerIdx > minOpenerIdx; openerIdx -= jumps[openerIdx] + 1) { - opener = delimiters[openerIdx]; - if (opener.marker !== closer.marker) continue; - if (opener.open && opener.end < 0) { - isOddMatch = false; - // from spec: - - // If one of the delimiters can both open and close emphasis, then the - // sum of the lengths of the delimiter runs containing the opening and - // closing delimiters must not be a multiple of 3 unless both lengths - // are multiples of 3. - - if (opener.close || closer.open) { - if ((opener.length + closer.length) % 3 === 0) { - if (opener.length % 3 !== 0 || closer.length % 3 !== 0) { - isOddMatch = true; - } - } - } - if (!isOddMatch) { - // If previous delimiter cannot be an opener, we can safely skip - // the entire sequence in future checks. This is required to make - // sure algorithm has linear complexity (see *_*_*_*_*_... case). - lastJump = openerIdx > 0 && !delimiters[openerIdx - 1].open ? jumps[openerIdx - 1] + 1 : 0; - jumps[closerIdx] = closerIdx - openerIdx + lastJump; - jumps[openerIdx] = lastJump; - closer.open = false; - opener.end = closerIdx; - opener.close = false; - newMinOpenerIdx = -1; - // treat next token as start of run, - // it optimizes skips in **<...>**a**<...>** pathological case - lastTokenIdx = -2; - break; - } - } - } - if (newMinOpenerIdx !== -1) { - // If match for this delimiter run failed, we want to set lower bound for - // future lookups. This is required to make sure algorithm has linear - // complexity. - // See details here: - // https://github.com/commonmark/cmark/issues/178#issuecomment-270417442 - openersBottom[closer.marker][(closer.open ? 3 : 0) + (closer.length || 0) % 3] = newMinOpenerIdx; - } - } - } - var balance_pairs = function link_pairs(state) { - var curr, tokens_meta = state.tokens_meta, max = state.tokens_meta.length; - processDelimiters(state.delimiters); - for (curr = 0; curr < max; curr++) { - if (tokens_meta[curr] && tokens_meta[curr].delimiters) { - processDelimiters(tokens_meta[curr].delimiters); - } - } - }; - // Clean up tokens after emphasis and strikethrough postprocessing: - var fragments_join = function fragments_join(state) { - var curr, last, level = 0, tokens = state.tokens, max = state.tokens.length; - for (curr = last = 0; curr < max; curr++) { - // re-calculate levels after emphasis/strikethrough turns some text nodes - // into opening/closing tags - if (tokens[curr].nesting < 0) level--; - // closing tag - tokens[curr].level = level; - if (tokens[curr].nesting > 0) level++; - // opening tag - if (tokens[curr].type === "text" && curr + 1 < max && tokens[curr + 1].type === "text") { - // collapse two adjacent text nodes - tokens[curr + 1].content = tokens[curr].content + tokens[curr + 1].content; - } else { - if (curr !== last) { - tokens[last] = tokens[curr]; - } - last++; - } - } - if (curr !== last) { - tokens.length = last; - } - }; - var isWhiteSpace = utils.isWhiteSpace; - var isPunctChar = utils.isPunctChar; - var isMdAsciiPunct = utils.isMdAsciiPunct; - function StateInline(src, md, env, outTokens) { - this.src = src; - this.env = env; - this.md = md; - this.tokens = outTokens; - this.tokens_meta = Array(outTokens.length); - this.pos = 0; - this.posMax = this.src.length; - this.level = 0; - this.pending = ""; - this.pendingLevel = 0; - // Stores { start: end } pairs. Useful for backtrack - // optimization of pairs parse (emphasis, strikes). - this.cache = {}; - // List of emphasis-like delimiters for current tag - this.delimiters = []; - // Stack of delimiter lists for upper level tags - this._prev_delimiters = []; - // backtick length => last seen position - this.backticks = {}; - this.backticksScanned = false; - // Counter used to disable inline linkify-it execution - // inside and markdown links - this.linkLevel = 0; - } - // Flush pending text - - StateInline.prototype.pushPending = function() { - var token$1 = new token("text", "", 0); - token$1.content = this.pending; - token$1.level = this.pendingLevel; - this.tokens.push(token$1); - this.pending = ""; - return token$1; - }; - // Push new token to "stream". - // If pending text exists - flush it as text token - - StateInline.prototype.push = function(type, tag, nesting) { - if (this.pending) { - this.pushPending(); - } - var token$1 = new token(type, tag, nesting); - var token_meta = null; - if (nesting < 0) { - // closing tag - this.level--; - this.delimiters = this._prev_delimiters.pop(); - } - token$1.level = this.level; - if (nesting > 0) { - // opening tag - this.level++; - this._prev_delimiters.push(this.delimiters); - this.delimiters = []; - token_meta = { - delimiters: this.delimiters - }; - } - this.pendingLevel = this.level; - this.tokens.push(token$1); - this.tokens_meta.push(token_meta); - return token$1; - }; - // Scan a sequence of emphasis-like markers, and determine whether - // it can start an emphasis sequence or end an emphasis sequence. - - // - start - position to scan from (it should point at a valid marker); - // - canSplitWord - determine if these markers can be found inside a word - - StateInline.prototype.scanDelims = function(start, canSplitWord) { - var pos = start, lastChar, nextChar, count, can_open, can_close, isLastWhiteSpace, isLastPunctChar, isNextWhiteSpace, isNextPunctChar, left_flanking = true, right_flanking = true, max = this.posMax, marker = this.src.charCodeAt(start); - // treat beginning of the line as a whitespace - lastChar = start > 0 ? this.src.charCodeAt(start - 1) : 32; - while (pos < max && this.src.charCodeAt(pos) === marker) { - pos++; - } - count = pos - start; - // treat end of the line as a whitespace - nextChar = pos < max ? this.src.charCodeAt(pos) : 32; - isLastPunctChar = isMdAsciiPunct(lastChar) || isPunctChar(String.fromCharCode(lastChar)); - isNextPunctChar = isMdAsciiPunct(nextChar) || isPunctChar(String.fromCharCode(nextChar)); - isLastWhiteSpace = isWhiteSpace(lastChar); - isNextWhiteSpace = isWhiteSpace(nextChar); - if (isNextWhiteSpace) { - left_flanking = false; - } else if (isNextPunctChar) { - if (!(isLastWhiteSpace || isLastPunctChar)) { - left_flanking = false; - } - } - if (isLastWhiteSpace) { - right_flanking = false; - } else if (isLastPunctChar) { - if (!(isNextWhiteSpace || isNextPunctChar)) { - right_flanking = false; - } - } - if (!canSplitWord) { - can_open = left_flanking && (!right_flanking || isLastPunctChar); - can_close = right_flanking && (!left_flanking || isNextPunctChar); - } else { - can_open = left_flanking; - can_close = right_flanking; - } - return { - can_open: can_open, - can_close: can_close, - length: count - }; - }; - // re-export Token class to use in block rules - StateInline.prototype.Token = token; - var state_inline = StateInline; - //////////////////////////////////////////////////////////////////////////////// - // Parser rules - var _rules = [ [ "text", text ], [ "linkify", linkify ], [ "newline", newline ], [ "escape", _escape ], [ "backticks", backticks ], [ "strikethrough", strikethrough.tokenize ], [ "emphasis", emphasis.tokenize ], [ "link", link ], [ "image", image ], [ "autolink", autolink ], [ "html_inline", html_inline ], [ "entity", entity ] ]; - // `rule2` ruleset was created specifically for emphasis/strikethrough - // post-processing and may be changed in the future. - - // Don't use this for anything except pairs (plugins working with `balance_pairs`). - - var _rules2 = [ [ "balance_pairs", balance_pairs ], [ "strikethrough", strikethrough.postProcess ], [ "emphasis", emphasis.postProcess ], - // rules for pairs separate '**' into its own text tokens, which may be left unused, - // rule below merges unused segments back with the rest of the text - [ "fragments_join", fragments_join ] ]; - /** - * new ParserInline() - **/ function ParserInline() { - var i; - /** - * ParserInline#ruler -> Ruler - * - * [[Ruler]] instance. Keep configuration of inline rules. - **/ this.ruler = new ruler; - for (i = 0; i < _rules.length; i++) { - this.ruler.push(_rules[i][0], _rules[i][1]); - } - /** - * ParserInline#ruler2 -> Ruler - * - * [[Ruler]] instance. Second ruler used for post-processing - * (e.g. in emphasis-like rules). - **/ this.ruler2 = new ruler; - for (i = 0; i < _rules2.length; i++) { - this.ruler2.push(_rules2[i][0], _rules2[i][1]); - } - } - // Skip single token by running all rules in validation mode; - // returns `true` if any rule reported success - - ParserInline.prototype.skipToken = function(state) { - var ok, i, pos = state.pos, rules = this.ruler.getRules(""), len = rules.length, maxNesting = state.md.options.maxNesting, cache = state.cache; - if (typeof cache[pos] !== "undefined") { - state.pos = cache[pos]; - return; - } - if (state.level < maxNesting) { - for (i = 0; i < len; i++) { - // Increment state.level and decrement it later to limit recursion. - // It's harmless to do here, because no tokens are created. But ideally, - // we'd need a separate private state variable for this purpose. - state.level++; - ok = rules[i](state, true); - state.level--; - if (ok) { - if (pos >= state.pos) { - throw new Error("inline rule didn't increment state.pos"); - } - break; - } - } - } else { - // Too much nesting, just skip until the end of the paragraph. - // NOTE: this will cause links to behave incorrectly in the following case, - // when an amount of `[` is exactly equal to `maxNesting + 1`: - // [[[[[[[[[[[[[[[[[[[[[foo]() - // TODO: remove this workaround when CM standard will allow nested links - // (we can replace it by preventing links from being parsed in - // validation mode) - state.pos = state.posMax; - } - if (!ok) { - state.pos++; - } - cache[pos] = state.pos; - }; - // Generate tokens for input range - - ParserInline.prototype.tokenize = function(state) { - var ok, i, prevPos, rules = this.ruler.getRules(""), len = rules.length, end = state.posMax, maxNesting = state.md.options.maxNesting; - while (state.pos < end) { - // Try all possible rules. - // On success, rule should: - // - update `state.pos` - // - update `state.tokens` - // - return true - prevPos = state.pos; - if (state.level < maxNesting) { - for (i = 0; i < len; i++) { - ok = rules[i](state, false); - if (ok) { - if (prevPos >= state.pos) { - throw new Error("inline rule didn't increment state.pos"); - } - break; - } - } - } - if (ok) { - if (state.pos >= end) { - break; - } - continue; - } - state.pending += state.src[state.pos++]; - } - if (state.pending) { - state.pushPending(); - } - }; - /** - * ParserInline.parse(str, md, env, outTokens) - * - * Process input string and push inline tokens into `outTokens` - **/ ParserInline.prototype.parse = function(str, md, env, outTokens) { - var i, rules, len; - var state = new this.State(str, md, env, outTokens); - this.tokenize(state); - rules = this.ruler2.getRules(""); - len = rules.length; - for (i = 0; i < len; i++) { - rules[i](state); - } - }; - ParserInline.prototype.State = state_inline; - var parser_inline = ParserInline; - var re = function(opts) { - var re = {}; - opts = opts || {}; - // Use direct extract instead of `regenerate` to reduse browserified size - re.src_Any = regex$3.source; - re.src_Cc = regex$2.source; - re.src_Z = regex.source; - re.src_P = regex$4.source; - // \p{\Z\P\Cc\CF} (white spaces + control + format + punctuation) - re.src_ZPCc = [ re.src_Z, re.src_P, re.src_Cc ].join("|"); - // \p{\Z\Cc} (white spaces + control) - re.src_ZCc = [ re.src_Z, re.src_Cc ].join("|"); - // Experimental. List of chars, completely prohibited in links - // because can separate it from other part of text - var text_separators = "[><\uff5c]"; - // All possible word characters (everything without punctuation, spaces & controls) - // Defined via punctuation & spaces to save space - // Should be something like \p{\L\N\S\M} (\w but without `_`) - re.src_pseudo_letter = "(?:(?!" + text_separators + "|" + re.src_ZPCc + ")" + re.src_Any + ")"; - // The same as abothe but without [0-9] - // var src_pseudo_letter_non_d = '(?:(?![0-9]|' + src_ZPCc + ')' + src_Any + ')'; - //////////////////////////////////////////////////////////////////////////////// - re.src_ip4 = "(?:(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)"; - // Prohibit any of "@/[]()" in user/pass to avoid wrong domain fetch. - re.src_auth = "(?:(?:(?!" + re.src_ZCc + "|[@/\\[\\]()]).)+@)?"; - re.src_port = "(?::(?:6(?:[0-4]\\d{3}|5(?:[0-4]\\d{2}|5(?:[0-2]\\d|3[0-5])))|[1-5]?\\d{1,4}))?"; - re.src_host_terminator = "(?=$|" + text_separators + "|" + re.src_ZPCc + ")" + "(?!" + (opts["---"] ? "-(?!--)|" : "-|") + "_|:\\d|\\.-|\\.(?!$|" + re.src_ZPCc + "))"; - re.src_path = "(?:" + "[/?#]" + "(?:" + "(?!" + re.src_ZCc + "|" + text_separators + "|[()[\\]{}.,\"'?!\\-;]).|" + "\\[(?:(?!" + re.src_ZCc + "|\\]).)*\\]|" + "\\((?:(?!" + re.src_ZCc + "|[)]).)*\\)|" + "\\{(?:(?!" + re.src_ZCc + "|[}]).)*\\}|" + '\\"(?:(?!' + re.src_ZCc + '|["]).)+\\"|' + "\\'(?:(?!" + re.src_ZCc + "|[']).)+\\'|" + "\\'(?=" + re.src_pseudo_letter + "|[-])|" + // allow `I'm_king` if no pair found - "\\.{2,}[a-zA-Z0-9%/&]|" + // google has many dots in "google search" links (#66, #81). - // github has ... in commit range links, - // Restrict to - // - english - // - percent-encoded - // - parts of file path - // - params separator - // until more examples found. - "\\.(?!" + re.src_ZCc + "|[.]|$)|" + (opts["---"] ? "\\-(?!--(?:[^-]|$))(?:-*)|" : "\\-+|") + ",(?!" + re.src_ZCc + "|$)|" + // allow `,,,` in paths - ";(?!" + re.src_ZCc + "|$)|" + // allow `;` if not followed by space-like char - "\\!+(?!" + re.src_ZCc + "|[!]|$)|" + // allow `!!!` in paths, but not at the end - "\\?(?!" + re.src_ZCc + "|[?]|$)" + ")+" + "|\\/" + ")?"; - // Allow anything in markdown spec, forbid quote (") at the first position - // because emails enclosed in quotes are far more common - re.src_email_name = '[\\-;:&=\\+\\$,\\.a-zA-Z0-9_][\\-;:&=\\+\\$,\\"\\.a-zA-Z0-9_]*'; - re.src_xn = "xn--[a-z0-9\\-]{1,59}"; - // More to read about domain names - // http://serverfault.com/questions/638260/ - re.src_domain_root = - // Allow letters & digits (http://test1) - "(?:" + re.src_xn + "|" + re.src_pseudo_letter + "{1,63}" + ")"; - re.src_domain = "(?:" + re.src_xn + "|" + "(?:" + re.src_pseudo_letter + ")" + "|" + "(?:" + re.src_pseudo_letter + "(?:-|" + re.src_pseudo_letter + "){0,61}" + re.src_pseudo_letter + ")" + ")"; - re.src_host = "(?:" + - // Don't need IP check, because digits are already allowed in normal domain names - // src_ip4 + - // '|' + - "(?:(?:(?:" + re.src_domain + ")\\.)*" + re.src_domain /*_root*/ + ")" + ")"; - re.tpl_host_fuzzy = "(?:" + re.src_ip4 + "|" + "(?:(?:(?:" + re.src_domain + ")\\.)+(?:%TLDS%))" + ")"; - re.tpl_host_no_ip_fuzzy = "(?:(?:(?:" + re.src_domain + ")\\.)+(?:%TLDS%))"; - re.src_host_strict = re.src_host + re.src_host_terminator; - re.tpl_host_fuzzy_strict = re.tpl_host_fuzzy + re.src_host_terminator; - re.src_host_port_strict = re.src_host + re.src_port + re.src_host_terminator; - re.tpl_host_port_fuzzy_strict = re.tpl_host_fuzzy + re.src_port + re.src_host_terminator; - re.tpl_host_port_no_ip_fuzzy_strict = re.tpl_host_no_ip_fuzzy + re.src_port + re.src_host_terminator; - //////////////////////////////////////////////////////////////////////////////// - // Main rules - // Rude test fuzzy links by host, for quick deny - re.tpl_host_fuzzy_test = "localhost|www\\.|\\.\\d{1,3}\\.|(?:\\.(?:%TLDS%)(?:" + re.src_ZPCc + "|>|$))"; - re.tpl_email_fuzzy = "(^|" + text_separators + '|"|\\(|' + re.src_ZCc + ")" + "(" + re.src_email_name + "@" + re.tpl_host_fuzzy_strict + ")"; - re.tpl_link_fuzzy = - // Fuzzy link can't be prepended with .:/\- and non punctuation. - // but can start with > (markdown blockquote) - "(^|(?![.:/\\-_@])(?:[$+<=>^`|\uff5c]|" + re.src_ZPCc + "))" + "((?![$+<=>^`|\uff5c])" + re.tpl_host_port_fuzzy_strict + re.src_path + ")"; - re.tpl_link_no_ip_fuzzy = - // Fuzzy link can't be prepended with .:/\- and non punctuation. - // but can start with > (markdown blockquote) - "(^|(?![.:/\\-_@])(?:[$+<=>^`|\uff5c]|" + re.src_ZPCc + "))" + "((?![$+<=>^`|\uff5c])" + re.tpl_host_port_no_ip_fuzzy_strict + re.src_path + ")"; - return re; - }; - //////////////////////////////////////////////////////////////////////////////// - // Helpers - // Merge objects - - function assign(obj /*from1, from2, from3, ...*/) { - var sources = Array.prototype.slice.call(arguments, 1); - sources.forEach((function(source) { - if (!source) { - return; - } - Object.keys(source).forEach((function(key) { - obj[key] = source[key]; - })); - })); - return obj; - } - function _class(obj) { - return Object.prototype.toString.call(obj); - } - function isString(obj) { - return _class(obj) === "[object String]"; - } - function isObject(obj) { - return _class(obj) === "[object Object]"; - } - function isRegExp(obj) { - return _class(obj) === "[object RegExp]"; - } - function isFunction(obj) { - return _class(obj) === "[object Function]"; - } - function escapeRE(str) { - return str.replace(/[.?*+^$[\]\\(){}|-]/g, "\\$&"); - } - //////////////////////////////////////////////////////////////////////////////// - var defaultOptions = { - fuzzyLink: true, - fuzzyEmail: true, - fuzzyIP: false - }; - function isOptionsObj(obj) { - return Object.keys(obj || {}).reduce((function(acc, k) { - return acc || defaultOptions.hasOwnProperty(k); - }), false); - } - var defaultSchemas = { - "http:": { - validate: function(text, pos, self) { - var tail = text.slice(pos); - if (!self.re.http) { - // compile lazily, because "host"-containing variables can change on tlds update. - self.re.http = new RegExp("^\\/\\/" + self.re.src_auth + self.re.src_host_port_strict + self.re.src_path, "i"); - } - if (self.re.http.test(tail)) { - return tail.match(self.re.http)[0].length; - } - return 0; - } - }, - "https:": "http:", - "ftp:": "http:", - "//": { - validate: function(text, pos, self) { - var tail = text.slice(pos); - if (!self.re.no_http) { - // compile lazily, because "host"-containing variables can change on tlds update. - self.re.no_http = new RegExp("^" + self.re.src_auth + - // Don't allow single-level domains, because of false positives like '//test' - // with code comments - "(?:localhost|(?:(?:" + self.re.src_domain + ")\\.)+" + self.re.src_domain_root + ")" + self.re.src_port + self.re.src_host_terminator + self.re.src_path, "i"); - } - if (self.re.no_http.test(tail)) { - // should not be `://` & `///`, that protects from errors in protocol name - if (pos >= 3 && text[pos - 3] === ":") { - return 0; - } - if (pos >= 3 && text[pos - 3] === "/") { - return 0; - } - return tail.match(self.re.no_http)[0].length; - } - return 0; - } - }, - "mailto:": { - validate: function(text, pos, self) { - var tail = text.slice(pos); - if (!self.re.mailto) { - self.re.mailto = new RegExp("^" + self.re.src_email_name + "@" + self.re.src_host_strict, "i"); - } - if (self.re.mailto.test(tail)) { - return tail.match(self.re.mailto)[0].length; - } - return 0; - } - } - }; - /*eslint-disable max-len*/ - // RE pattern for 2-character tlds (autogenerated by ./support/tlds_2char_gen.js) - var tlds_2ch_src_re = "a[cdefgilmnoqrstuwxz]|b[abdefghijmnorstvwyz]|c[acdfghiklmnoruvwxyz]|d[ejkmoz]|e[cegrstu]|f[ijkmor]|g[abdefghilmnpqrstuwy]|h[kmnrtu]|i[delmnoqrst]|j[emop]|k[eghimnprwyz]|l[abcikrstuvy]|m[acdeghklmnopqrstuvwxyz]|n[acefgilopruz]|om|p[aefghklmnrstwy]|qa|r[eosuw]|s[abcdeghijklmnortuvxyz]|t[cdfghjklmnortvwz]|u[agksyz]|v[aceginu]|w[fs]|y[et]|z[amw]"; - // DON'T try to make PRs with changes. Extend TLDs with LinkifyIt.tlds() instead - var tlds_default = "biz|com|edu|gov|net|org|pro|web|xxx|aero|asia|coop|info|museum|name|shop|\u0440\u0444".split("|"); - /*eslint-enable max-len*/ - //////////////////////////////////////////////////////////////////////////////// - function resetScanCache(self) { - self.__index__ = -1; - self.__text_cache__ = ""; - } - function createValidator(re) { - return function(text, pos) { - var tail = text.slice(pos); - if (re.test(tail)) { - return tail.match(re)[0].length; - } - return 0; - }; - } - function createNormalizer() { - return function(match, self) { - self.normalize(match); - }; - } - // Schemas compiler. Build regexps. - - function compile(self) { - // Load & clone RE patterns. - var re$1 = self.re = re(self.__opts__); - // Define dynamic patterns - var tlds = self.__tlds__.slice(); - self.onCompile(); - if (!self.__tlds_replaced__) { - tlds.push(tlds_2ch_src_re); - } - tlds.push(re$1.src_xn); - re$1.src_tlds = tlds.join("|"); - function untpl(tpl) { - return tpl.replace("%TLDS%", re$1.src_tlds); - } - re$1.email_fuzzy = RegExp(untpl(re$1.tpl_email_fuzzy), "i"); - re$1.link_fuzzy = RegExp(untpl(re$1.tpl_link_fuzzy), "i"); - re$1.link_no_ip_fuzzy = RegExp(untpl(re$1.tpl_link_no_ip_fuzzy), "i"); - re$1.host_fuzzy_test = RegExp(untpl(re$1.tpl_host_fuzzy_test), "i"); - - // Compile each schema - - var aliases = []; - self.__compiled__ = {}; - // Reset compiled data - function schemaError(name, val) { - throw new Error('(LinkifyIt) Invalid schema "' + name + '": ' + val); - } - Object.keys(self.__schemas__).forEach((function(name) { - var val = self.__schemas__[name]; - // skip disabled methods - if (val === null) { - return; - } - var compiled = { - validate: null, - link: null - }; - self.__compiled__[name] = compiled; - if (isObject(val)) { - if (isRegExp(val.validate)) { - compiled.validate = createValidator(val.validate); - } else if (isFunction(val.validate)) { - compiled.validate = val.validate; - } else { - schemaError(name, val); - } - if (isFunction(val.normalize)) { - compiled.normalize = val.normalize; - } else if (!val.normalize) { - compiled.normalize = createNormalizer(); - } else { - schemaError(name, val); - } - return; - } - if (isString(val)) { - aliases.push(name); - return; - } - schemaError(name, val); - })); - - // Compile postponed aliases - - aliases.forEach((function(alias) { - if (!self.__compiled__[self.__schemas__[alias]]) { - // Silently fail on missed schemas to avoid errons on disable. - // schemaError(alias, self.__schemas__[alias]); - return; - } - self.__compiled__[alias].validate = self.__compiled__[self.__schemas__[alias]].validate; - self.__compiled__[alias].normalize = self.__compiled__[self.__schemas__[alias]].normalize; - })); - - // Fake record for guessed links - - self.__compiled__[""] = { - validate: null, - normalize: createNormalizer() - }; - - // Build schema condition - - var slist = Object.keys(self.__compiled__).filter((function(name) { - // Filter disabled & fake schemas - return name.length > 0 && self.__compiled__[name]; - })).map(escapeRE).join("|"); - // (?!_) cause 1.5x slowdown - self.re.schema_test = RegExp("(^|(?!_)(?:[><\uff5c]|" + re$1.src_ZPCc + "))(" + slist + ")", "i"); - self.re.schema_search = RegExp("(^|(?!_)(?:[><\uff5c]|" + re$1.src_ZPCc + "))(" + slist + ")", "ig"); - self.re.schema_at_start = RegExp("^" + self.re.schema_search.source, "i"); - self.re.pretest = RegExp("(" + self.re.schema_test.source + ")|(" + self.re.host_fuzzy_test.source + ")|@", "i"); - - // Cleanup - - resetScanCache(self); - } - /** - * class Match - * - * Match result. Single element of array, returned by [[LinkifyIt#match]] - **/ function Match(self, shift) { - var start = self.__index__, end = self.__last_index__, text = self.__text_cache__.slice(start, end); - /** - * Match#schema -> String - * - * Prefix (protocol) for matched string. - **/ this.schema = self.__schema__.toLowerCase(); - /** - * Match#index -> Number - * - * First position of matched string. - **/ this.index = start + shift; - /** - * Match#lastIndex -> Number - * - * Next position after matched string. - **/ this.lastIndex = end + shift; - /** - * Match#raw -> String - * - * Matched string. - **/ this.raw = text; - /** - * Match#text -> String - * - * Notmalized text of matched string. - **/ this.text = text; - /** - * Match#url -> String - * - * Normalized url of matched string. - **/ this.url = text; - } - function createMatch(self, shift) { - var match = new Match(self, shift); - self.__compiled__[match.schema].normalize(match, self); - return match; - } - /** - * class LinkifyIt - **/ - /** - * new LinkifyIt(schemas, options) - * - schemas (Object): Optional. Additional schemas to validate (prefix/validator) - * - options (Object): { fuzzyLink|fuzzyEmail|fuzzyIP: true|false } - * - * Creates new linkifier instance with optional additional schemas. - * Can be called without `new` keyword for convenience. - * - * By default understands: - * - * - `http(s)://...` , `ftp://...`, `mailto:...` & `//...` links - * - "fuzzy" links and emails (example.com, foo@bar.com). - * - * `schemas` is an object, where each key/value describes protocol/rule: - * - * - __key__ - link prefix (usually, protocol name with `:` at the end, `skype:` - * for example). `linkify-it` makes shure that prefix is not preceeded with - * alphanumeric char and symbols. Only whitespaces and punctuation allowed. - * - __value__ - rule to check tail after link prefix - * - _String_ - just alias to existing rule - * - _Object_ - * - _validate_ - validator function (should return matched length on success), - * or `RegExp`. - * - _normalize_ - optional function to normalize text & url of matched result - * (for example, for @twitter mentions). - * - * `options`: - * - * - __fuzzyLink__ - recognige URL-s without `http(s):` prefix. Default `true`. - * - __fuzzyIP__ - allow IPs in fuzzy links above. Can conflict with some texts - * like version numbers. Default `false`. - * - __fuzzyEmail__ - recognize emails without `mailto:` prefix. - * - **/ function LinkifyIt(schemas, options) { - if (!(this instanceof LinkifyIt)) { - return new LinkifyIt(schemas, options); - } - if (!options) { - if (isOptionsObj(schemas)) { - options = schemas; - schemas = {}; - } - } - this.__opts__ = assign({}, defaultOptions, options); - // Cache last tested result. Used to skip repeating steps on next `match` call. - this.__index__ = -1; - this.__last_index__ = -1; - // Next scan position - this.__schema__ = ""; - this.__text_cache__ = ""; - this.__schemas__ = assign({}, defaultSchemas, schemas); - this.__compiled__ = {}; - this.__tlds__ = tlds_default; - this.__tlds_replaced__ = false; - this.re = {}; - compile(this); - } - /** chainable - * LinkifyIt#add(schema, definition) - * - schema (String): rule name (fixed pattern prefix) - * - definition (String|RegExp|Object): schema definition - * - * Add new rule definition. See constructor description for details. - **/ LinkifyIt.prototype.add = function add(schema, definition) { - this.__schemas__[schema] = definition; - compile(this); - return this; - }; - /** chainable - * LinkifyIt#set(options) - * - options (Object): { fuzzyLink|fuzzyEmail|fuzzyIP: true|false } - * - * Set recognition options for links without schema. - **/ LinkifyIt.prototype.set = function set(options) { - this.__opts__ = assign(this.__opts__, options); - return this; - }; - /** - * LinkifyIt#test(text) -> Boolean - * - * Searches linkifiable pattern and returns `true` on success or `false` on fail. - **/ LinkifyIt.prototype.test = function test(text) { - // Reset scan cache - this.__text_cache__ = text; - this.__index__ = -1; - if (!text.length) { - return false; - } - var m, ml, me, len, shift, next, re, tld_pos, at_pos; - // try to scan for link with schema - that's the most simple rule - if (this.re.schema_test.test(text)) { - re = this.re.schema_search; - re.lastIndex = 0; - while ((m = re.exec(text)) !== null) { - len = this.testSchemaAt(text, m[2], re.lastIndex); - if (len) { - this.__schema__ = m[2]; - this.__index__ = m.index + m[1].length; - this.__last_index__ = m.index + m[0].length + len; - break; - } - } - } - if (this.__opts__.fuzzyLink && this.__compiled__["http:"]) { - // guess schemaless links - tld_pos = text.search(this.re.host_fuzzy_test); - if (tld_pos >= 0) { - // if tld is located after found link - no need to check fuzzy pattern - if (this.__index__ < 0 || tld_pos < this.__index__) { - if ((ml = text.match(this.__opts__.fuzzyIP ? this.re.link_fuzzy : this.re.link_no_ip_fuzzy)) !== null) { - shift = ml.index + ml[1].length; - if (this.__index__ < 0 || shift < this.__index__) { - this.__schema__ = ""; - this.__index__ = shift; - this.__last_index__ = ml.index + ml[0].length; - } - } - } - } - } - if (this.__opts__.fuzzyEmail && this.__compiled__["mailto:"]) { - // guess schemaless emails - at_pos = text.indexOf("@"); - if (at_pos >= 0) { - // We can't skip this check, because this cases are possible: - // 192.168.1.1@gmail.com, my.in@example.com - if ((me = text.match(this.re.email_fuzzy)) !== null) { - shift = me.index + me[1].length; - next = me.index + me[0].length; - if (this.__index__ < 0 || shift < this.__index__ || shift === this.__index__ && next > this.__last_index__) { - this.__schema__ = "mailto:"; - this.__index__ = shift; - this.__last_index__ = next; - } - } - } - } - return this.__index__ >= 0; - }; - /** - * LinkifyIt#pretest(text) -> Boolean - * - * Very quick check, that can give false positives. Returns true if link MAY BE - * can exists. Can be used for speed optimization, when you need to check that - * link NOT exists. - **/ LinkifyIt.prototype.pretest = function pretest(text) { - return this.re.pretest.test(text); - }; - /** - * LinkifyIt#testSchemaAt(text, name, position) -> Number - * - text (String): text to scan - * - name (String): rule (schema) name - * - position (Number): text offset to check from - * - * Similar to [[LinkifyIt#test]] but checks only specific protocol tail exactly - * at given position. Returns length of found pattern (0 on fail). - **/ LinkifyIt.prototype.testSchemaAt = function testSchemaAt(text, schema, pos) { - // If not supported schema check requested - terminate - if (!this.__compiled__[schema.toLowerCase()]) { - return 0; - } - return this.__compiled__[schema.toLowerCase()].validate(text, pos, this); - }; - /** - * LinkifyIt#match(text) -> Array|null - * - * Returns array of found link descriptions or `null` on fail. We strongly - * recommend to use [[LinkifyIt#test]] first, for best speed. - * - * ##### Result match description - * - * - __schema__ - link schema, can be empty for fuzzy links, or `//` for - * protocol-neutral links. - * - __index__ - offset of matched text - * - __lastIndex__ - index of next char after mathch end - * - __raw__ - matched text - * - __text__ - normalized text - * - __url__ - link, generated from matched text - **/ LinkifyIt.prototype.match = function match(text) { - var shift = 0, result = []; - // Try to take previous element from cache, if .test() called before - if (this.__index__ >= 0 && this.__text_cache__ === text) { - result.push(createMatch(this, shift)); - shift = this.__last_index__; - } - // Cut head if cache was used - var tail = shift ? text.slice(shift) : text; - // Scan string until end reached - while (this.test(tail)) { - result.push(createMatch(this, shift)); - tail = tail.slice(this.__last_index__); - shift += this.__last_index__; - } - if (result.length) { - return result; - } - return null; - }; - /** - * LinkifyIt#matchAtStart(text) -> Match|null - * - * Returns fully-formed (not fuzzy) link if it starts at the beginning - * of the string, and null otherwise. - **/ LinkifyIt.prototype.matchAtStart = function matchAtStart(text) { - // Reset scan cache - this.__text_cache__ = text; - this.__index__ = -1; - if (!text.length) return null; - var m = this.re.schema_at_start.exec(text); - if (!m) return null; - var len = this.testSchemaAt(text, m[2], m[0].length); - if (!len) return null; - this.__schema__ = m[2]; - this.__index__ = m.index + m[1].length; - this.__last_index__ = m.index + m[0].length + len; - return createMatch(this, 0); - }; - /** chainable - * LinkifyIt#tlds(list [, keepOld]) -> this - * - list (Array): list of tlds - * - keepOld (Boolean): merge with current list if `true` (`false` by default) - * - * Load (or merge) new tlds list. Those are user for fuzzy links (without prefix) - * to avoid false positives. By default this algorythm used: - * - * - hostname with any 2-letter root zones are ok. - * - biz|com|edu|gov|net|org|pro|web|xxx|aero|asia|coop|info|museum|name|shop|рф - * are ok. - * - encoded (`xn--...`) root zones are ok. - * - * If list is replaced, then exact match for 2-chars root zones will be checked. - **/ LinkifyIt.prototype.tlds = function tlds(list, keepOld) { - list = Array.isArray(list) ? list : [ list ]; - if (!keepOld) { - this.__tlds__ = list.slice(); - this.__tlds_replaced__ = true; - compile(this); - return this; - } - this.__tlds__ = this.__tlds__.concat(list).sort().filter((function(el, idx, arr) { - return el !== arr[idx - 1]; - })).reverse(); - compile(this); - return this; - }; - /** - * LinkifyIt#normalize(match) - * - * Default normalizer (if schema does not define it's own). - **/ LinkifyIt.prototype.normalize = function normalize(match) { - // Do minimal possible changes by default. Need to collect feedback prior - // to move forward https://github.com/markdown-it/linkify-it/issues/1 - if (!match.schema) { - match.url = "http://" + match.url; - } - if (match.schema === "mailto:" && !/^mailto:/i.test(match.url)) { - match.url = "mailto:" + match.url; - } - }; - /** - * LinkifyIt#onCompile() - * - * Override to modify basic RegExp-s. - **/ LinkifyIt.prototype.onCompile = function onCompile() {}; - var linkifyIt = LinkifyIt; - /*! https://mths.be/punycode v1.4.1 by @mathias */ - /** Highest positive signed 32-bit float value */ var maxInt = 2147483647; - // aka. 0x7FFFFFFF or 2^31-1 - /** Bootstring parameters */ var base = 36; - var tMin = 1; - var tMax = 26; - var skew = 38; - var damp = 700; - var initialBias = 72; - var initialN = 128; - // 0x80 - var delimiter = "-"; - // '\x2D' - /** Regular expressions */ var regexPunycode = /^xn--/; - var regexNonASCII = /[^\x20-\x7E]/; - // unprintable ASCII chars + non-ASCII chars - var regexSeparators = /[\x2E\u3002\uFF0E\uFF61]/g; - // RFC 3490 separators - /** Error messages */ var errors = { - overflow: "Overflow: input needs wider integers to process", - "not-basic": "Illegal input >= 0x80 (not a basic code point)", - "invalid-input": "Invalid input" - }; - /** Convenience shortcuts */ var baseMinusTMin = base - tMin; - var floor = Math.floor; - var stringFromCharCode = String.fromCharCode; - /*--------------------------------------------------------------------------*/ - /** - * A generic error utility function. - * @private - * @param {String} type The error type. - * @returns {Error} Throws a `RangeError` with the applicable error message. - */ function error(type) { - throw new RangeError(errors[type]); - } - /** - * A generic `Array#map` utility function. - * @private - * @param {Array} array The array to iterate over. - * @param {Function} callback The function that gets called for every array - * item. - * @returns {Array} A new array of values returned by the callback function. - */ function map(array, fn) { - var length = array.length; - var result = []; - while (length--) { - result[length] = fn(array[length]); - } - return result; - } - /** - * A simple `Array#map`-like wrapper to work with domain name strings or email - * addresses. - * @private - * @param {String} domain The domain name or email address. - * @param {Function} callback The function that gets called for every - * character. - * @returns {Array} A new string of characters returned by the callback - * function. - */ function mapDomain(string, fn) { - var parts = string.split("@"); - var result = ""; - if (parts.length > 1) { - // In email addresses, only the domain name should be punycoded. Leave - // the local part (i.e. everything up to `@`) intact. - result = parts[0] + "@"; - string = parts[1]; - } - // Avoid `split(regex)` for IE8 compatibility. See #17. - string = string.replace(regexSeparators, "."); - var labels = string.split("."); - var encoded = map(labels, fn).join("."); - return result + encoded; - } - /** - * Creates an array containing the numeric code points of each Unicode - * character in the string. While JavaScript uses UCS-2 internally, - * this function will convert a pair of surrogate halves (each of which - * UCS-2 exposes as separate characters) into a single code point, - * matching UTF-16. - * @see `punycode.ucs2.encode` - * @see - * @memberOf punycode.ucs2 - * @name decode - * @param {String} string The Unicode input string (UCS-2). - * @returns {Array} The new array of code points. - */ function ucs2decode(string) { - var output = [], counter = 0, length = string.length, value, extra; - while (counter < length) { - value = string.charCodeAt(counter++); - if (value >= 55296 && value <= 56319 && counter < length) { - // high surrogate, and there is a next character - extra = string.charCodeAt(counter++); - if ((extra & 64512) == 56320) { - // low surrogate - output.push(((value & 1023) << 10) + (extra & 1023) + 65536); - } else { - // unmatched surrogate; only append this code unit, in case the next - // code unit is the high surrogate of a surrogate pair - output.push(value); - counter--; - } - } else { - output.push(value); - } - } - return output; - } - /** - * Creates a string based on an array of numeric code points. - * @see `punycode.ucs2.decode` - * @memberOf punycode.ucs2 - * @name encode - * @param {Array} codePoints The array of numeric code points. - * @returns {String} The new Unicode string (UCS-2). - */ function ucs2encode(array) { - return map(array, (function(value) { - var output = ""; - if (value > 65535) { - value -= 65536; - output += stringFromCharCode(value >>> 10 & 1023 | 55296); - value = 56320 | value & 1023; - } - output += stringFromCharCode(value); - return output; - })).join(""); - } - /** - * Converts a basic code point into a digit/integer. - * @see `digitToBasic()` - * @private - * @param {Number} codePoint The basic numeric code point value. - * @returns {Number} The numeric value of a basic code point (for use in - * representing integers) in the range `0` to `base - 1`, or `base` if - * the code point does not represent a value. - */ function basicToDigit(codePoint) { - if (codePoint - 48 < 10) { - return codePoint - 22; - } - if (codePoint - 65 < 26) { - return codePoint - 65; - } - if (codePoint - 97 < 26) { - return codePoint - 97; - } - return base; - } - /** - * Converts a digit/integer into a basic code point. - * @see `basicToDigit()` - * @private - * @param {Number} digit The numeric value of a basic code point. - * @returns {Number} The basic code point whose value (when used for - * representing integers) is `digit`, which needs to be in the range - * `0` to `base - 1`. If `flag` is non-zero, the uppercase form is - * used; else, the lowercase form is used. The behavior is undefined - * if `flag` is non-zero and `digit` has no uppercase form. - */ function digitToBasic(digit, flag) { - // 0..25 map to ASCII a..z or A..Z - // 26..35 map to ASCII 0..9 - return digit + 22 + 75 * (digit < 26) - ((flag != 0) << 5); - } - /** - * Bias adaptation function as per section 3.4 of RFC 3492. - * https://tools.ietf.org/html/rfc3492#section-3.4 - * @private - */ function adapt(delta, numPoints, firstTime) { - var k = 0; - delta = firstTime ? floor(delta / damp) : delta >> 1; - delta += floor(delta / numPoints); - for (;delta > baseMinusTMin * tMax >> 1; k += base) { - delta = floor(delta / baseMinusTMin); - } - return floor(k + (baseMinusTMin + 1) * delta / (delta + skew)); - } - /** - * Converts a Punycode string of ASCII-only symbols to a string of Unicode - * symbols. - * @memberOf punycode - * @param {String} input The Punycode string of ASCII-only symbols. - * @returns {String} The resulting string of Unicode symbols. - */ function decode(input) { - // Don't use UCS-2 - var output = [], inputLength = input.length, out, i = 0, n = initialN, bias = initialBias, basic, j, index, oldi, w, k, digit, t, - /** Cached calculation results */ - baseMinusT; - // Handle the basic code points: let `basic` be the number of input code - // points before the last delimiter, or `0` if there is none, then copy - // the first basic code points to the output. - basic = input.lastIndexOf(delimiter); - if (basic < 0) { - basic = 0; - } - for (j = 0; j < basic; ++j) { - // if it's not a basic code point - if (input.charCodeAt(j) >= 128) { - error("not-basic"); - } - output.push(input.charCodeAt(j)); - } - // Main decoding loop: start just after the last delimiter if any basic code - // points were copied; start at the beginning otherwise. - for (index = basic > 0 ? basic + 1 : 0; index < inputLength; ) { - // `index` is the index of the next character to be consumed. - // Decode a generalized variable-length integer into `delta`, - // which gets added to `i`. The overflow checking is easier - // if we increase `i` as we go, then subtract off its starting - // value at the end to obtain `delta`. - for (oldi = i, w = 1, k = base; ;k += base) { - if (index >= inputLength) { - error("invalid-input"); - } - digit = basicToDigit(input.charCodeAt(index++)); - if (digit >= base || digit > floor((maxInt - i) / w)) { - error("overflow"); - } - i += digit * w; - t = k <= bias ? tMin : k >= bias + tMax ? tMax : k - bias; - if (digit < t) { - break; - } - baseMinusT = base - t; - if (w > floor(maxInt / baseMinusT)) { - error("overflow"); - } - w *= baseMinusT; - } - out = output.length + 1; - bias = adapt(i - oldi, out, oldi == 0); - // `i` was supposed to wrap around from `out` to `0`, - // incrementing `n` each time, so we'll fix that now: - if (floor(i / out) > maxInt - n) { - error("overflow"); - } - n += floor(i / out); - i %= out; - // Insert `n` at position `i` of the output - output.splice(i++, 0, n); - } - return ucs2encode(output); - } - /** - * Converts a string of Unicode symbols (e.g. a domain name label) to a - * Punycode string of ASCII-only symbols. - * @memberOf punycode - * @param {String} input The string of Unicode symbols. - * @returns {String} The resulting Punycode string of ASCII-only symbols. - */ function encode(input) { - var n, delta, handledCPCount, basicLength, bias, j, m, q, k, t, currentValue, output = [], - /** `inputLength` will hold the number of code points in `input`. */ - inputLength, - /** Cached calculation results */ - handledCPCountPlusOne, baseMinusT, qMinusT; - // Convert the input in UCS-2 to Unicode - input = ucs2decode(input); - // Cache the length - inputLength = input.length; - // Initialize the state - n = initialN; - delta = 0; - bias = initialBias; - // Handle the basic code points - for (j = 0; j < inputLength; ++j) { - currentValue = input[j]; - if (currentValue < 128) { - output.push(stringFromCharCode(currentValue)); - } - } - handledCPCount = basicLength = output.length; - // `handledCPCount` is the number of code points that have been handled; - // `basicLength` is the number of basic code points. - // Finish the basic string - if it is not empty - with a delimiter - if (basicLength) { - output.push(delimiter); - } - // Main encoding loop: - while (handledCPCount < inputLength) { - // All non-basic code points < n have been handled already. Find the next - // larger one: - for (m = maxInt, j = 0; j < inputLength; ++j) { - currentValue = input[j]; - if (currentValue >= n && currentValue < m) { - m = currentValue; - } - } - // Increase `delta` enough to advance the decoder's state to , - // but guard against overflow - handledCPCountPlusOne = handledCPCount + 1; - if (m - n > floor((maxInt - delta) / handledCPCountPlusOne)) { - error("overflow"); - } - delta += (m - n) * handledCPCountPlusOne; - n = m; - for (j = 0; j < inputLength; ++j) { - currentValue = input[j]; - if (currentValue < n && ++delta > maxInt) { - error("overflow"); - } - if (currentValue == n) { - // Represent delta as a generalized variable-length integer - for (q = delta, k = base; ;k += base) { - t = k <= bias ? tMin : k >= bias + tMax ? tMax : k - bias; - if (q < t) { - break; - } - qMinusT = q - t; - baseMinusT = base - t; - output.push(stringFromCharCode(digitToBasic(t + qMinusT % baseMinusT, 0))); - q = floor(qMinusT / baseMinusT); - } - output.push(stringFromCharCode(digitToBasic(q, 0))); - bias = adapt(delta, handledCPCountPlusOne, handledCPCount == basicLength); - delta = 0; - ++handledCPCount; - } - } - ++delta; - ++n; - } - return output.join(""); - } - /** - * Converts a Punycode string representing a domain name or an email address - * to Unicode. Only the Punycoded parts of the input will be converted, i.e. - * it doesn't matter if you call it on a string that has already been - * converted to Unicode. - * @memberOf punycode - * @param {String} input The Punycoded domain name or email address to - * convert to Unicode. - * @returns {String} The Unicode representation of the given Punycode - * string. - */ function toUnicode(input) { - return mapDomain(input, (function(string) { - return regexPunycode.test(string) ? decode(string.slice(4).toLowerCase()) : string; - })); - } - /** - * Converts a Unicode string representing a domain name or an email address to - * Punycode. Only the non-ASCII parts of the domain name will be converted, - * i.e. it doesn't matter if you call it with a domain that's already in - * ASCII. - * @memberOf punycode - * @param {String} input The domain name or email address to convert, as a - * Unicode string. - * @returns {String} The Punycode representation of the given domain name or - * email address. - */ function toASCII(input) { - return mapDomain(input, (function(string) { - return regexNonASCII.test(string) ? "xn--" + encode(string) : string; - })); - } - var version = "1.4.1"; - /** - * An object of methods to convert from JavaScript's internal character - * representation (UCS-2) to Unicode code points, and back. - * @see - * @memberOf punycode - * @type Object - */ var ucs2 = { - decode: ucs2decode, - encode: ucs2encode - }; - var punycode$1 = { - version: version, - ucs2: ucs2, - toASCII: toASCII, - toUnicode: toUnicode, - encode: encode, - decode: decode - }; - var punycode$2 = Object.freeze({ - __proto__: null, - decode: decode, - encode: encode, - toUnicode: toUnicode, - toASCII: toASCII, - version: version, - ucs2: ucs2, - default: punycode$1 - }); - // markdown-it default options - var _default = { - options: { - html: false, - // Enable HTML tags in source - xhtmlOut: false, - // Use '/' to close single tags (
) - breaks: false, - // Convert '\n' in paragraphs into
- langPrefix: "language-", - // CSS language prefix for fenced blocks - linkify: false, - // autoconvert URL-like texts to links - // Enable some language-neutral replacements + quotes beautification - typographer: false, - // Double + single quotes replacement pairs, when typographer enabled, - // and smartquotes on. Could be either a String or an Array. - // For example, you can use '«»„“' for Russian, '„“‚‘' for German, - // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp). - quotes: "\u201c\u201d\u2018\u2019", - /* “”‘’ */ - // Highlighter function. Should return escaped HTML, - // or '' if the source string is not changed and should be escaped externaly. - // If result starts with ) - breaks: false, - // Convert '\n' in paragraphs into
- langPrefix: "language-", - // CSS language prefix for fenced blocks - linkify: false, - // autoconvert URL-like texts to links - // Enable some language-neutral replacements + quotes beautification - typographer: false, - // Double + single quotes replacement pairs, when typographer enabled, - // and smartquotes on. Could be either a String or an Array. - // For example, you can use '«»„“' for Russian, '„“‚‘' for German, - // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp). - quotes: "\u201c\u201d\u2018\u2019", - /* “”‘’ */ - // Highlighter function. Should return escaped HTML, - // or '' if the source string is not changed and should be escaped externaly. - // If result starts with ) - breaks: false, - // Convert '\n' in paragraphs into
- langPrefix: "language-", - // CSS language prefix for fenced blocks - linkify: false, - // autoconvert URL-like texts to links - // Enable some language-neutral replacements + quotes beautification - typographer: false, - // Double + single quotes replacement pairs, when typographer enabled, - // and smartquotes on. Could be either a String or an Array. - // For example, you can use '«»„“' for Russian, '„“‚‘' for German, - // and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp). - quotes: "\u201c\u201d\u2018\u2019", - /* “”‘’ */ - // Highlighter function. Should return escaped HTML, - // or '' if the source string is not changed and should be escaped externaly. - // If result starts with = 0) { - try { - parsed.hostname = punycode.toASCII(parsed.hostname); - } catch (er) {} - } - } - return mdurl.encode(mdurl.format(parsed)); - } - function normalizeLinkText(url) { - var parsed = mdurl.parse(url, true); - if (parsed.hostname) { - // Encode hostnames in urls like: - // `http://host/`, `https://host/`, `mailto:user@host`, `//host/` - // We don't encode unknown schemas, because it's likely that we encode - // something we shouldn't (e.g. `skype:name` treated as `skype:host`) - if (!parsed.protocol || RECODE_HOSTNAME_FOR.indexOf(parsed.protocol) >= 0) { - try { - parsed.hostname = punycode.toUnicode(parsed.hostname); - } catch (er) {} - } - } - // add '%' to exclude list because of https://github.com/markdown-it/markdown-it/issues/720 - return mdurl.decode(mdurl.format(parsed), mdurl.decode.defaultChars + "%"); - } - /** - * class MarkdownIt - * - * Main parser/renderer class. - * - * ##### Usage - * - * ```javascript - * // node.js, "classic" way: - * var MarkdownIt = require('markdown-it'), - * md = new MarkdownIt(); - * var result = md.render('# markdown-it rulezz!'); - * - * // node.js, the same, but with sugar: - * var md = require('markdown-it')(); - * var result = md.render('# markdown-it rulezz!'); - * - * // browser without AMD, added to "window" on script load - * // Note, there are no dash. - * var md = window.markdownit(); - * var result = md.render('# markdown-it rulezz!'); - * ``` - * - * Single line rendering, without paragraph wrap: - * - * ```javascript - * var md = require('markdown-it')(); - * var result = md.renderInline('__markdown-it__ rulezz!'); - * ``` - **/ - /** - * new MarkdownIt([presetName, options]) - * - presetName (String): optional, `commonmark` / `zero` - * - options (Object) - * - * Creates parser instanse with given config. Can be called without `new`. - * - * ##### presetName - * - * MarkdownIt provides named presets as a convenience to quickly - * enable/disable active syntax rules and options for common use cases. - * - * - ["commonmark"](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/commonmark.js) - - * configures parser to strict [CommonMark](http://commonmark.org/) mode. - * - [default](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/default.js) - - * similar to GFM, used when no preset name given. Enables all available rules, - * but still without html, typographer & autolinker. - * - ["zero"](https://github.com/markdown-it/markdown-it/blob/master/lib/presets/zero.js) - - * all rules disabled. Useful to quickly setup your config via `.enable()`. - * For example, when you need only `bold` and `italic` markup and nothing else. - * - * ##### options: - * - * - __html__ - `false`. Set `true` to enable HTML tags in source. Be careful! - * That's not safe! You may need external sanitizer to protect output from XSS. - * It's better to extend features via plugins, instead of enabling HTML. - * - __xhtmlOut__ - `false`. Set `true` to add '/' when closing single tags - * (`
`). This is needed only for full CommonMark compatibility. In real - * world you will need HTML output. - * - __breaks__ - `false`. Set `true` to convert `\n` in paragraphs into `
`. - * - __langPrefix__ - `language-`. CSS language class prefix for fenced blocks. - * Can be useful for external highlighters. - * - __linkify__ - `false`. Set `true` to autoconvert URL-like text to links. - * - __typographer__ - `false`. Set `true` to enable [some language-neutral - * replacement](https://github.com/markdown-it/markdown-it/blob/master/lib/rules_core/replacements.js) + - * quotes beautification (smartquotes). - * - __quotes__ - `“”‘’`, String or Array. Double + single quotes replacement - * pairs, when typographer enabled and smartquotes on. For example, you can - * use `'«»„“'` for Russian, `'„“‚‘'` for German, and - * `['«\xA0', '\xA0»', '‹\xA0', '\xA0›']` for French (including nbsp). - * - __highlight__ - `null`. Highlighter function for fenced code blocks. - * Highlighter `function (str, lang)` should return escaped HTML. It can also - * return empty string if the source was not changed and should be escaped - * externaly. If result starts with `): - * - * ```javascript - * var hljs = require('highlight.js') // https://highlightjs.org/ - * - * // Actual default values - * var md = require('markdown-it')({ - * highlight: function (str, lang) { - * if (lang && hljs.getLanguage(lang)) { - * try { - * return '
' +
-	 *                hljs.highlight(str, { language: lang, ignoreIllegals: true }).value +
-	 *                '
'; - * } catch (__) {} - * } - * - * return '
' + md.utils.escapeHtml(str) + '
'; - * } - * }); - * ``` - * - **/ function MarkdownIt(presetName, options) { - if (!(this instanceof MarkdownIt)) { - return new MarkdownIt(presetName, options); - } - if (!options) { - if (!utils.isString(presetName)) { - options = presetName || {}; - presetName = "default"; - } - } - /** - * MarkdownIt#inline -> ParserInline - * - * Instance of [[ParserInline]]. You may need it to add new rules when - * writing plugins. For simple rules control use [[MarkdownIt.disable]] and - * [[MarkdownIt.enable]]. - **/ this.inline = new parser_inline; - /** - * MarkdownIt#block -> ParserBlock - * - * Instance of [[ParserBlock]]. You may need it to add new rules when - * writing plugins. For simple rules control use [[MarkdownIt.disable]] and - * [[MarkdownIt.enable]]. - **/ this.block = new parser_block; - /** - * MarkdownIt#core -> Core - * - * Instance of [[Core]] chain executor. You may need it to add new rules when - * writing plugins. For simple rules control use [[MarkdownIt.disable]] and - * [[MarkdownIt.enable]]. - **/ this.core = new parser_core; - /** - * MarkdownIt#renderer -> Renderer - * - * Instance of [[Renderer]]. Use it to modify output look. Or to add rendering - * rules for new token types, generated by plugins. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')(); - * - * function myToken(tokens, idx, options, env, self) { - * //... - * return result; - * }; - * - * md.renderer.rules['my_token'] = myToken - * ``` - * - * See [[Renderer]] docs and [source code](https://github.com/markdown-it/markdown-it/blob/master/lib/renderer.js). - **/ this.renderer = new renderer; - /** - * MarkdownIt#linkify -> LinkifyIt - * - * [linkify-it](https://github.com/markdown-it/linkify-it) instance. - * Used by [linkify](https://github.com/markdown-it/markdown-it/blob/master/lib/rules_core/linkify.js) - * rule. - **/ this.linkify = new linkifyIt; - /** - * MarkdownIt#validateLink(url) -> Boolean - * - * Link validation function. CommonMark allows too much in links. By default - * we disable `javascript:`, `vbscript:`, `file:` schemas, and almost all `data:...` schemas - * except some embedded image types. - * - * You can change this behaviour: - * - * ```javascript - * var md = require('markdown-it')(); - * // enable everything - * md.validateLink = function () { return true; } - * ``` - **/ this.validateLink = validateLink; - /** - * MarkdownIt#normalizeLink(url) -> String - * - * Function used to encode link url to a machine-readable format, - * which includes url-encoding, punycode, etc. - **/ this.normalizeLink = normalizeLink; - /** - * MarkdownIt#normalizeLinkText(url) -> String - * - * Function used to decode link url to a human-readable format` - **/ this.normalizeLinkText = normalizeLinkText; - // Expose utils & helpers for easy acces from plugins - /** - * MarkdownIt#utils -> utils - * - * Assorted utility functions, useful to write plugins. See details - * [here](https://github.com/markdown-it/markdown-it/blob/master/lib/common/utils.js). - **/ this.utils = utils; - /** - * MarkdownIt#helpers -> helpers - * - * Link components parser functions, useful to write plugins. See details - * [here](https://github.com/markdown-it/markdown-it/blob/master/lib/helpers). - **/ this.helpers = utils.assign({}, helpers); - this.options = {}; - this.configure(presetName); - if (options) { - this.set(options); - } - } - /** chainable - * MarkdownIt.set(options) - * - * Set parser options (in the same format as in constructor). Probably, you - * will never need it, but you can change options after constructor call. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')() - * .set({ html: true, breaks: true }) - * .set({ typographer, true }); - * ``` - * - * __Note:__ To achieve the best possible performance, don't modify a - * `markdown-it` instance options on the fly. If you need multiple configurations - * it's best to create multiple instances and initialize each with separate - * config. - **/ MarkdownIt.prototype.set = function(options) { - utils.assign(this.options, options); - return this; - }; - /** chainable, internal - * MarkdownIt.configure(presets) - * - * Batch load of all options and compenent settings. This is internal method, - * and you probably will not need it. But if you will - see available presets - * and data structure [here](https://github.com/markdown-it/markdown-it/tree/master/lib/presets) - * - * We strongly recommend to use presets instead of direct config loads. That - * will give better compatibility with next versions. - **/ MarkdownIt.prototype.configure = function(presets) { - var self = this, presetName; - if (utils.isString(presets)) { - presetName = presets; - presets = config[presetName]; - if (!presets) { - throw new Error('Wrong `markdown-it` preset "' + presetName + '", check name'); - } - } - if (!presets) { - throw new Error("Wrong `markdown-it` preset, can't be empty"); - } - if (presets.options) { - self.set(presets.options); - } - if (presets.components) { - Object.keys(presets.components).forEach((function(name) { - if (presets.components[name].rules) { - self[name].ruler.enableOnly(presets.components[name].rules); - } - if (presets.components[name].rules2) { - self[name].ruler2.enableOnly(presets.components[name].rules2); - } - })); - } - return this; - }; - /** chainable - * MarkdownIt.enable(list, ignoreInvalid) - * - list (String|Array): rule name or list of rule names to enable - * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found. - * - * Enable list or rules. It will automatically find appropriate components, - * containing rules with given names. If rule not found, and `ignoreInvalid` - * not set - throws exception. - * - * ##### Example - * - * ```javascript - * var md = require('markdown-it')() - * .enable(['sub', 'sup']) - * .disable('smartquotes'); - * ``` - **/ MarkdownIt.prototype.enable = function(list, ignoreInvalid) { - var result = []; - if (!Array.isArray(list)) { - list = [ list ]; - } - [ "core", "block", "inline" ].forEach((function(chain) { - result = result.concat(this[chain].ruler.enable(list, true)); - }), this); - result = result.concat(this.inline.ruler2.enable(list, true)); - var missed = list.filter((function(name) { - return result.indexOf(name) < 0; - })); - if (missed.length && !ignoreInvalid) { - throw new Error("MarkdownIt. Failed to enable unknown rule(s): " + missed); - } - return this; - }; - /** chainable - * MarkdownIt.disable(list, ignoreInvalid) - * - list (String|Array): rule name or list of rule names to disable. - * - ignoreInvalid (Boolean): set `true` to ignore errors when rule not found. - * - * The same as [[MarkdownIt.enable]], but turn specified rules off. - **/ MarkdownIt.prototype.disable = function(list, ignoreInvalid) { - var result = []; - if (!Array.isArray(list)) { - list = [ list ]; - } - [ "core", "block", "inline" ].forEach((function(chain) { - result = result.concat(this[chain].ruler.disable(list, true)); - }), this); - result = result.concat(this.inline.ruler2.disable(list, true)); - var missed = list.filter((function(name) { - return result.indexOf(name) < 0; - })); - if (missed.length && !ignoreInvalid) { - throw new Error("MarkdownIt. Failed to disable unknown rule(s): " + missed); - } - return this; - }; - /** chainable - * MarkdownIt.use(plugin, params) - * - * Load specified plugin with given params into current parser instance. - * It's just a sugar to call `plugin(md, params)` with curring. - * - * ##### Example - * - * ```javascript - * var iterator = require('markdown-it-for-inline'); - * var md = require('markdown-it')() - * .use(iterator, 'foo_replace', 'text', function (tokens, idx) { - * tokens[idx].content = tokens[idx].content.replace(/foo/g, 'bar'); - * }); - * ``` - **/ MarkdownIt.prototype.use = function(plugin /*, params, ... */) { - var args = [ this ].concat(Array.prototype.slice.call(arguments, 1)); - plugin.apply(plugin, args); - return this; - }; - /** internal - * MarkdownIt.parse(src, env) -> Array - * - src (String): source string - * - env (Object): environment sandbox - * - * Parse input string and return list of block tokens (special token type - * "inline" will contain list of inline tokens). You should not call this - * method directly, until you write custom renderer (for example, to produce - * AST). - * - * `env` is used to pass data between "distributed" rules and return additional - * metadata like reference info, needed for the renderer. It also can be used to - * inject data in specific cases. Usually, you will be ok to pass `{}`, - * and then pass updated object to renderer. - **/ MarkdownIt.prototype.parse = function(src, env) { - if (typeof src !== "string") { - throw new Error("Input data should be a String"); - } - var state = new this.core.State(src, this, env); - this.core.process(state); - return state.tokens; - }; - /** - * MarkdownIt.render(src [, env]) -> String - * - src (String): source string - * - env (Object): environment sandbox - * - * Render markdown string into html. It does all magic for you :). - * - * `env` can be used to inject additional metadata (`{}` by default). - * But you will not need it with high probability. See also comment - * in [[MarkdownIt.parse]]. - **/ MarkdownIt.prototype.render = function(src, env) { - env = env || {}; - return this.renderer.render(this.parse(src, env), this.options, env); - }; - /** internal - * MarkdownIt.parseInline(src, env) -> Array - * - src (String): source string - * - env (Object): environment sandbox - * - * The same as [[MarkdownIt.parse]] but skip all block rules. It returns the - * block tokens list with the single `inline` element, containing parsed inline - * tokens in `children` property. Also updates `env` object. - **/ MarkdownIt.prototype.parseInline = function(src, env) { - var state = new this.core.State(src, this, env); - state.inlineMode = true; - this.core.process(state); - return state.tokens; - }; - /** - * MarkdownIt.renderInline(src [, env]) -> String - * - src (String): source string - * - env (Object): environment sandbox - * - * Similar to [[MarkdownIt.render]] but for single paragraph content. Result - * will NOT be wrapped into `

` tags. - **/ MarkdownIt.prototype.renderInline = function(src, env) { - env = env || {}; - return this.renderer.render(this.parseInline(src, env), this.options, env); - }; - var lib = MarkdownIt; - var markdownIt = lib; - return markdownIt; -})); - diff --git a/examples/server/public/deps_tailwindcss.js b/examples/server/public/deps_tailwindcss.js deleted file mode 100644 index 6736cb8ca..000000000 --- a/examples/server/public/deps_tailwindcss.js +++ /dev/null @@ -1,82 +0,0 @@ -(()=>{var Iv=Object.create;var Ui=Object.defineProperty;var Dv=Object.getOwnPropertyDescriptor;var qv=Object.getOwnPropertyNames;var $v=Object.getPrototypeOf,Lv=Object.prototype.hasOwnProperty;var cf=r=>Ui(r,"__esModule",{value:!0});var pf=r=>{if(typeof require!="undefined")return require(r);throw new Error('Dynamic require of "'+r+'" is not supported')};var R=(r,e)=>()=>(r&&(e=r(r=0)),e);var x=(r,e)=>()=>(e||r((e={exports:{}}).exports,e),e.exports),Ge=(r,e)=>{cf(r);for(var t in e)Ui(r,t,{get:e[t],enumerable:!0})},Mv=(r,e,t)=>{if(e&&typeof e=="object"||typeof e=="function")for(let i of qv(e))!Lv.call(r,i)&&i!=="default"&&Ui(r,i,{get:()=>e[i],enumerable:!(t=Dv(e,i))||t.enumerable});return r},pe=r=>Mv(cf(Ui(r!=null?Iv($v(r)):{},"default",r&&r.__esModule&&"default"in r?{get:()=>r.default,enumerable:!0}:{value:r,enumerable:!0})),r);var m,u=R(()=>{m={platform:"",env:{},versions:{node:"14.17.6"}}});var Nv,be,ft=R(()=>{u();Nv=0,be={readFileSync:r=>self[r]||"",statSync:()=>({mtimeMs:Nv++}),promises:{readFile:r=>Promise.resolve(self[r]||"")}}});var Ns=x((sP,hf)=>{u();"use strict";var df=class{constructor(e={}){if(!(e.maxSize&&e.maxSize>0))throw new TypeError("`maxSize` must be a number greater than 0");if(typeof e.maxAge=="number"&&e.maxAge===0)throw new TypeError("`maxAge` must be a number greater than 0");this.maxSize=e.maxSize,this.maxAge=e.maxAge||1/0,this.onEviction=e.onEviction,this.cache=new Map,this.oldCache=new Map,this._size=0}_emitEvictions(e){if(typeof this.onEviction=="function")for(let[t,i]of e)this.onEviction(t,i.value)}_deleteIfExpired(e,t){return typeof t.expiry=="number"&&t.expiry<=Date.now()?(typeof this.onEviction=="function"&&this.onEviction(e,t.value),this.delete(e)):!1}_getOrDeleteIfExpired(e,t){if(this._deleteIfExpired(e,t)===!1)return t.value}_getItemValue(e,t){return t.expiry?this._getOrDeleteIfExpired(e,t):t.value}_peek(e,t){let i=t.get(e);return this._getItemValue(e,i)}_set(e,t){this.cache.set(e,t),this._size++,this._size>=this.maxSize&&(this._size=0,this._emitEvictions(this.oldCache),this.oldCache=this.cache,this.cache=new Map)}_moveToRecent(e,t){this.oldCache.delete(e),this._set(e,t)}*_entriesAscending(){for(let e of this.oldCache){let[t,i]=e;this.cache.has(t)||this._deleteIfExpired(t,i)===!1&&(yield e)}for(let e of this.cache){let[t,i]=e;this._deleteIfExpired(t,i)===!1&&(yield e)}}get(e){if(this.cache.has(e)){let t=this.cache.get(e);return this._getItemValue(e,t)}if(this.oldCache.has(e)){let t=this.oldCache.get(e);if(this._deleteIfExpired(e,t)===!1)return this._moveToRecent(e,t),t.value}}set(e,t,{maxAge:i=this.maxAge===1/0?void 0:Date.now()+this.maxAge}={}){this.cache.has(e)?this.cache.set(e,{value:t,maxAge:i}):this._set(e,{value:t,expiry:i})}has(e){return this.cache.has(e)?!this._deleteIfExpired(e,this.cache.get(e)):this.oldCache.has(e)?!this._deleteIfExpired(e,this.oldCache.get(e)):!1}peek(e){if(this.cache.has(e))return this._peek(e,this.cache);if(this.oldCache.has(e))return this._peek(e,this.oldCache)}delete(e){let t=this.cache.delete(e);return t&&this._size--,this.oldCache.delete(e)||t}clear(){this.cache.clear(),this.oldCache.clear(),this._size=0}resize(e){if(!(e&&e>0))throw new TypeError("`maxSize` must be a number greater than 0");let t=[...this._entriesAscending()],i=t.length-e;i<0?(this.cache=new Map(t),this.oldCache=new Map,this._size=t.length):(i>0&&this._emitEvictions(t.slice(0,i)),this.oldCache=new Map(t.slice(i)),this.cache=new Map,this._size=0),this.maxSize=e}*keys(){for(let[e]of this)yield e}*values(){for(let[,e]of this)yield e}*[Symbol.iterator](){for(let e of this.cache){let[t,i]=e;this._deleteIfExpired(t,i)===!1&&(yield[t,i.value])}for(let e of this.oldCache){let[t,i]=e;this.cache.has(t)||this._deleteIfExpired(t,i)===!1&&(yield[t,i.value])}}*entriesDescending(){let e=[...this.cache];for(let t=e.length-1;t>=0;--t){let i=e[t],[n,a]=i;this._deleteIfExpired(n,a)===!1&&(yield[n,a.value])}e=[...this.oldCache];for(let t=e.length-1;t>=0;--t){let i=e[t],[n,a]=i;this.cache.has(n)||this._deleteIfExpired(n,a)===!1&&(yield[n,a.value])}}*entriesAscending(){for(let[e,t]of this._entriesAscending())yield[e,t.value]}get size(){if(!this._size)return this.oldCache.size;let e=0;for(let t of this.oldCache.keys())this.cache.has(t)||e++;return Math.min(this._size+e,this.maxSize)}};hf.exports=df});var mf,gf=R(()=>{u();mf=r=>r&&r._hash});function Vi(r){return mf(r,{ignoreUnknown:!0})}var yf=R(()=>{u();gf()});function xt(r){if(r=`${r}`,r==="0")return"0";if(/^[+-]?(\d+|\d*\.\d+)(e[+-]?\d+)?(%|\w+)?$/.test(r))return r.replace(/^[+-]?/,t=>t==="-"?"":"-");let e=["var","calc","min","max","clamp"];for(let t of e)if(r.includes(`${t}(`))return`calc(${r} * -1)`}var Hi=R(()=>{u()});var bf,wf=R(()=>{u();bf=["preflight","container","accessibility","pointerEvents","visibility","position","inset","isolation","zIndex","order","gridColumn","gridColumnStart","gridColumnEnd","gridRow","gridRowStart","gridRowEnd","float","clear","margin","boxSizing","lineClamp","display","aspectRatio","size","height","maxHeight","minHeight","width","minWidth","maxWidth","flex","flexShrink","flexGrow","flexBasis","tableLayout","captionSide","borderCollapse","borderSpacing","transformOrigin","translate","rotate","skew","scale","transform","animation","cursor","touchAction","userSelect","resize","scrollSnapType","scrollSnapAlign","scrollSnapStop","scrollMargin","scrollPadding","listStylePosition","listStyleType","listStyleImage","appearance","columns","breakBefore","breakInside","breakAfter","gridAutoColumns","gridAutoFlow","gridAutoRows","gridTemplateColumns","gridTemplateRows","flexDirection","flexWrap","placeContent","placeItems","alignContent","alignItems","justifyContent","justifyItems","gap","space","divideWidth","divideStyle","divideColor","divideOpacity","placeSelf","alignSelf","justifySelf","overflow","overscrollBehavior","scrollBehavior","textOverflow","hyphens","whitespace","textWrap","wordBreak","borderRadius","borderWidth","borderStyle","borderColor","borderOpacity","backgroundColor","backgroundOpacity","backgroundImage","gradientColorStops","boxDecorationBreak","backgroundSize","backgroundAttachment","backgroundClip","backgroundPosition","backgroundRepeat","backgroundOrigin","fill","stroke","strokeWidth","objectFit","objectPosition","padding","textAlign","textIndent","verticalAlign","fontFamily","fontSize","fontWeight","textTransform","fontStyle","fontVariantNumeric","lineHeight","letterSpacing","textColor","textOpacity","textDecoration","textDecorationColor","textDecorationStyle","textDecorationThickness","textUnderlineOffset","fontSmoothing","placeholderColor","placeholderOpacity","caretColor","accentColor","opacity","backgroundBlendMode","mixBlendMode","boxShadow","boxShadowColor","outlineStyle","outlineWidth","outlineOffset","outlineColor","ringWidth","ringColor","ringOpacity","ringOffsetWidth","ringOffsetColor","blur","brightness","contrast","dropShadow","grayscale","hueRotate","invert","saturate","sepia","filter","backdropBlur","backdropBrightness","backdropContrast","backdropGrayscale","backdropHueRotate","backdropInvert","backdropOpacity","backdropSaturate","backdropSepia","backdropFilter","transitionProperty","transitionDelay","transitionDuration","transitionTimingFunction","willChange","contain","content","forcedColorAdjust"]});function vf(r,e){return r===void 0?e:Array.isArray(r)?r:[...new Set(e.filter(i=>r!==!1&&r[i]!==!1).concat(Object.keys(r).filter(i=>r[i]!==!1)))]}var xf=R(()=>{u()});var kf={};Ge(kf,{default:()=>Qe});var Qe,Wi=R(()=>{u();Qe=new Proxy({},{get:()=>String})});function Bs(r,e,t){typeof m!="undefined"&&m.env.JEST_WORKER_ID||t&&Sf.has(t)||(t&&Sf.add(t),console.warn(""),e.forEach(i=>console.warn(r,"-",i)))}function Fs(r){return Qe.dim(r)}var Sf,G,Be=R(()=>{u();Wi();Sf=new Set;G={info(r,e){Bs(Qe.bold(Qe.cyan("info")),...Array.isArray(r)?[r]:[e,r])},warn(r,e){["content-problems"].includes(r)||Bs(Qe.bold(Qe.yellow("warn")),...Array.isArray(r)?[r]:[e,r])},risk(r,e){Bs(Qe.bold(Qe.magenta("risk")),...Array.isArray(r)?[r]:[e,r])}}});var Af={};Ge(Af,{default:()=>js});function qr({version:r,from:e,to:t}){G.warn(`${e}-color-renamed`,[`As of Tailwind CSS ${r}, \`${e}\` has been renamed to \`${t}\`.`,"Update your configuration file to silence this warning."])}var js,zs=R(()=>{u();Be();js={inherit:"inherit",current:"currentColor",transparent:"transparent",black:"#000",white:"#fff",slate:{50:"#f8fafc",100:"#f1f5f9",200:"#e2e8f0",300:"#cbd5e1",400:"#94a3b8",500:"#64748b",600:"#475569",700:"#334155",800:"#1e293b",900:"#0f172a",950:"#020617"},gray:{50:"#f9fafb",100:"#f3f4f6",200:"#e5e7eb",300:"#d1d5db",400:"#9ca3af",500:"#6b7280",600:"#4b5563",700:"#374151",800:"#1f2937",900:"#111827",950:"#030712"},zinc:{50:"#fafafa",100:"#f4f4f5",200:"#e4e4e7",300:"#d4d4d8",400:"#a1a1aa",500:"#71717a",600:"#52525b",700:"#3f3f46",800:"#27272a",900:"#18181b",950:"#09090b"},neutral:{50:"#fafafa",100:"#f5f5f5",200:"#e5e5e5",300:"#d4d4d4",400:"#a3a3a3",500:"#737373",600:"#525252",700:"#404040",800:"#262626",900:"#171717",950:"#0a0a0a"},stone:{50:"#fafaf9",100:"#f5f5f4",200:"#e7e5e4",300:"#d6d3d1",400:"#a8a29e",500:"#78716c",600:"#57534e",700:"#44403c",800:"#292524",900:"#1c1917",950:"#0c0a09"},red:{50:"#fef2f2",100:"#fee2e2",200:"#fecaca",300:"#fca5a5",400:"#f87171",500:"#ef4444",600:"#dc2626",700:"#b91c1c",800:"#991b1b",900:"#7f1d1d",950:"#450a0a"},orange:{50:"#fff7ed",100:"#ffedd5",200:"#fed7aa",300:"#fdba74",400:"#fb923c",500:"#f97316",600:"#ea580c",700:"#c2410c",800:"#9a3412",900:"#7c2d12",950:"#431407"},amber:{50:"#fffbeb",100:"#fef3c7",200:"#fde68a",300:"#fcd34d",400:"#fbbf24",500:"#f59e0b",600:"#d97706",700:"#b45309",800:"#92400e",900:"#78350f",950:"#451a03"},yellow:{50:"#fefce8",100:"#fef9c3",200:"#fef08a",300:"#fde047",400:"#facc15",500:"#eab308",600:"#ca8a04",700:"#a16207",800:"#854d0e",900:"#713f12",950:"#422006"},lime:{50:"#f7fee7",100:"#ecfccb",200:"#d9f99d",300:"#bef264",400:"#a3e635",500:"#84cc16",600:"#65a30d",700:"#4d7c0f",800:"#3f6212",900:"#365314",950:"#1a2e05"},green:{50:"#f0fdf4",100:"#dcfce7",200:"#bbf7d0",300:"#86efac",400:"#4ade80",500:"#22c55e",600:"#16a34a",700:"#15803d",800:"#166534",900:"#14532d",950:"#052e16"},emerald:{50:"#ecfdf5",100:"#d1fae5",200:"#a7f3d0",300:"#6ee7b7",400:"#34d399",500:"#10b981",600:"#059669",700:"#047857",800:"#065f46",900:"#064e3b",950:"#022c22"},teal:{50:"#f0fdfa",100:"#ccfbf1",200:"#99f6e4",300:"#5eead4",400:"#2dd4bf",500:"#14b8a6",600:"#0d9488",700:"#0f766e",800:"#115e59",900:"#134e4a",950:"#042f2e"},cyan:{50:"#ecfeff",100:"#cffafe",200:"#a5f3fc",300:"#67e8f9",400:"#22d3ee",500:"#06b6d4",600:"#0891b2",700:"#0e7490",800:"#155e75",900:"#164e63",950:"#083344"},sky:{50:"#f0f9ff",100:"#e0f2fe",200:"#bae6fd",300:"#7dd3fc",400:"#38bdf8",500:"#0ea5e9",600:"#0284c7",700:"#0369a1",800:"#075985",900:"#0c4a6e",950:"#082f49"},blue:{50:"#eff6ff",100:"#dbeafe",200:"#bfdbfe",300:"#93c5fd",400:"#60a5fa",500:"#3b82f6",600:"#2563eb",700:"#1d4ed8",800:"#1e40af",900:"#1e3a8a",950:"#172554"},indigo:{50:"#eef2ff",100:"#e0e7ff",200:"#c7d2fe",300:"#a5b4fc",400:"#818cf8",500:"#6366f1",600:"#4f46e5",700:"#4338ca",800:"#3730a3",900:"#312e81",950:"#1e1b4b"},violet:{50:"#f5f3ff",100:"#ede9fe",200:"#ddd6fe",300:"#c4b5fd",400:"#a78bfa",500:"#8b5cf6",600:"#7c3aed",700:"#6d28d9",800:"#5b21b6",900:"#4c1d95",950:"#2e1065"},purple:{50:"#faf5ff",100:"#f3e8ff",200:"#e9d5ff",300:"#d8b4fe",400:"#c084fc",500:"#a855f7",600:"#9333ea",700:"#7e22ce",800:"#6b21a8",900:"#581c87",950:"#3b0764"},fuchsia:{50:"#fdf4ff",100:"#fae8ff",200:"#f5d0fe",300:"#f0abfc",400:"#e879f9",500:"#d946ef",600:"#c026d3",700:"#a21caf",800:"#86198f",900:"#701a75",950:"#4a044e"},pink:{50:"#fdf2f8",100:"#fce7f3",200:"#fbcfe8",300:"#f9a8d4",400:"#f472b6",500:"#ec4899",600:"#db2777",700:"#be185d",800:"#9d174d",900:"#831843",950:"#500724"},rose:{50:"#fff1f2",100:"#ffe4e6",200:"#fecdd3",300:"#fda4af",400:"#fb7185",500:"#f43f5e",600:"#e11d48",700:"#be123c",800:"#9f1239",900:"#881337",950:"#4c0519"},get lightBlue(){return qr({version:"v2.2",from:"lightBlue",to:"sky"}),this.sky},get warmGray(){return qr({version:"v3.0",from:"warmGray",to:"stone"}),this.stone},get trueGray(){return qr({version:"v3.0",from:"trueGray",to:"neutral"}),this.neutral},get coolGray(){return qr({version:"v3.0",from:"coolGray",to:"gray"}),this.gray},get blueGray(){return qr({version:"v3.0",from:"blueGray",to:"slate"}),this.slate}}});function Us(r,...e){for(let t of e){for(let i in t)r?.hasOwnProperty?.(i)||(r[i]=t[i]);for(let i of Object.getOwnPropertySymbols(t))r?.hasOwnProperty?.(i)||(r[i]=t[i])}return r}var Cf=R(()=>{u()});function kt(r){if(Array.isArray(r))return r;let e=r.split("[").length-1,t=r.split("]").length-1;if(e!==t)throw new Error(`Path is invalid. Has unbalanced brackets: ${r}`);return r.split(/\.(?![^\[]*\])|[\[\]]/g).filter(Boolean)}var Gi=R(()=>{u()});function we(r,e){return Qi.future.includes(e)?r.future==="all"||(r?.future?.[e]??_f[e]??!1):Qi.experimental.includes(e)?r.experimental==="all"||(r?.experimental?.[e]??_f[e]??!1):!1}function Ef(r){return r.experimental==="all"?Qi.experimental:Object.keys(r?.experimental??{}).filter(e=>Qi.experimental.includes(e)&&r.experimental[e])}function Of(r){if(m.env.JEST_WORKER_ID===void 0&&Ef(r).length>0){let e=Ef(r).map(t=>Qe.yellow(t)).join(", ");G.warn("experimental-flags-enabled",[`You have enabled experimental features: ${e}`,"Experimental features in Tailwind CSS are not covered by semver, may introduce breaking changes, and can change at any time."])}}var _f,Qi,ct=R(()=>{u();Wi();Be();_f={optimizeUniversalDefaults:!1,generalizedModifiers:!0,disableColorOpacityUtilitiesByDefault:!1,relativeContentPathsByDefault:!1},Qi={future:["hoverOnlyWhenSupported","respectDefaultRingColorOpacity","disableColorOpacityUtilitiesByDefault","relativeContentPathsByDefault"],experimental:["optimizeUniversalDefaults","generalizedModifiers"]}});function Tf(r){(()=>{if(r.purge||!r.content||!Array.isArray(r.content)&&!(typeof r.content=="object"&&r.content!==null))return!1;if(Array.isArray(r.content))return r.content.every(t=>typeof t=="string"?!0:!(typeof t?.raw!="string"||t?.extension&&typeof t?.extension!="string"));if(typeof r.content=="object"&&r.content!==null){if(Object.keys(r.content).some(t=>!["files","relative","extract","transform"].includes(t)))return!1;if(Array.isArray(r.content.files)){if(!r.content.files.every(t=>typeof t=="string"?!0:!(typeof t?.raw!="string"||t?.extension&&typeof t?.extension!="string")))return!1;if(typeof r.content.extract=="object"){for(let t of Object.values(r.content.extract))if(typeof t!="function")return!1}else if(!(r.content.extract===void 0||typeof r.content.extract=="function"))return!1;if(typeof r.content.transform=="object"){for(let t of Object.values(r.content.transform))if(typeof t!="function")return!1}else if(!(r.content.transform===void 0||typeof r.content.transform=="function"))return!1;if(typeof r.content.relative!="boolean"&&typeof r.content.relative!="undefined")return!1}return!0}return!1})()||G.warn("purge-deprecation",["The `purge`/`content` options have changed in Tailwind CSS v3.0.","Update your configuration file to eliminate this warning.","https://tailwindcss.com/docs/upgrade-guide#configure-content-sources"]),r.safelist=(()=>{let{content:t,purge:i,safelist:n}=r;return Array.isArray(n)?n:Array.isArray(t?.safelist)?t.safelist:Array.isArray(i?.safelist)?i.safelist:Array.isArray(i?.options?.safelist)?i.options.safelist:[]})(),r.blocklist=(()=>{let{blocklist:t}=r;if(Array.isArray(t)){if(t.every(i=>typeof i=="string"))return t;G.warn("blocklist-invalid",["The `blocklist` option must be an array of strings.","https://tailwindcss.com/docs/content-configuration#discarding-classes"])}return[]})(),typeof r.prefix=="function"?(G.warn("prefix-function",["As of Tailwind CSS v3.0, `prefix` cannot be a function.","Update `prefix` in your configuration to be a string to eliminate this warning.","https://tailwindcss.com/docs/upgrade-guide#prefix-cannot-be-a-function"]),r.prefix=""):r.prefix=r.prefix??"",r.content={relative:(()=>{let{content:t}=r;return t?.relative?t.relative:we(r,"relativeContentPathsByDefault")})(),files:(()=>{let{content:t,purge:i}=r;return Array.isArray(i)?i:Array.isArray(i?.content)?i.content:Array.isArray(t)?t:Array.isArray(t?.content)?t.content:Array.isArray(t?.files)?t.files:[]})(),extract:(()=>{let t=(()=>r.purge?.extract?r.purge.extract:r.content?.extract?r.content.extract:r.purge?.extract?.DEFAULT?r.purge.extract.DEFAULT:r.content?.extract?.DEFAULT?r.content.extract.DEFAULT:r.purge?.options?.extractors?r.purge.options.extractors:r.content?.options?.extractors?r.content.options.extractors:{})(),i={},n=(()=>{if(r.purge?.options?.defaultExtractor)return r.purge.options.defaultExtractor;if(r.content?.options?.defaultExtractor)return r.content.options.defaultExtractor})();if(n!==void 0&&(i.DEFAULT=n),typeof t=="function")i.DEFAULT=t;else if(Array.isArray(t))for(let{extensions:a,extractor:s}of t??[])for(let o of a)i[o]=s;else typeof t=="object"&&t!==null&&Object.assign(i,t);return i})(),transform:(()=>{let t=(()=>r.purge?.transform?r.purge.transform:r.content?.transform?r.content.transform:r.purge?.transform?.DEFAULT?r.purge.transform.DEFAULT:r.content?.transform?.DEFAULT?r.content.transform.DEFAULT:{})(),i={};return typeof t=="function"?i.DEFAULT=t:typeof t=="object"&&t!==null&&Object.assign(i,t),i})()};for(let t of r.content.files)if(typeof t=="string"&&/{([^,]*?)}/g.test(t)){G.warn("invalid-glob-braces",[`The glob pattern ${Fs(t)} in your Tailwind CSS configuration is invalid.`,`Update it to ${Fs(t.replace(/{([^,]*?)}/g,"$1"))} to silence this warning.`]);break}return r}var Rf=R(()=>{u();ct();Be()});function ke(r){if(Object.prototype.toString.call(r)!=="[object Object]")return!1;let e=Object.getPrototypeOf(r);return e===null||Object.getPrototypeOf(e)===null}var Kt=R(()=>{u()});function St(r){return Array.isArray(r)?r.map(e=>St(e)):typeof r=="object"&&r!==null?Object.fromEntries(Object.entries(r).map(([e,t])=>[e,St(t)])):r}var Yi=R(()=>{u()});function jt(r){return r.replace(/\\,/g,"\\2c ")}var Ki=R(()=>{u()});var Vs,Pf=R(()=>{u();Vs={aliceblue:[240,248,255],antiquewhite:[250,235,215],aqua:[0,255,255],aquamarine:[127,255,212],azure:[240,255,255],beige:[245,245,220],bisque:[255,228,196],black:[0,0,0],blanchedalmond:[255,235,205],blue:[0,0,255],blueviolet:[138,43,226],brown:[165,42,42],burlywood:[222,184,135],cadetblue:[95,158,160],chartreuse:[127,255,0],chocolate:[210,105,30],coral:[255,127,80],cornflowerblue:[100,149,237],cornsilk:[255,248,220],crimson:[220,20,60],cyan:[0,255,255],darkblue:[0,0,139],darkcyan:[0,139,139],darkgoldenrod:[184,134,11],darkgray:[169,169,169],darkgreen:[0,100,0],darkgrey:[169,169,169],darkkhaki:[189,183,107],darkmagenta:[139,0,139],darkolivegreen:[85,107,47],darkorange:[255,140,0],darkorchid:[153,50,204],darkred:[139,0,0],darksalmon:[233,150,122],darkseagreen:[143,188,143],darkslateblue:[72,61,139],darkslategray:[47,79,79],darkslategrey:[47,79,79],darkturquoise:[0,206,209],darkviolet:[148,0,211],deeppink:[255,20,147],deepskyblue:[0,191,255],dimgray:[105,105,105],dimgrey:[105,105,105],dodgerblue:[30,144,255],firebrick:[178,34,34],floralwhite:[255,250,240],forestgreen:[34,139,34],fuchsia:[255,0,255],gainsboro:[220,220,220],ghostwhite:[248,248,255],gold:[255,215,0],goldenrod:[218,165,32],gray:[128,128,128],green:[0,128,0],greenyellow:[173,255,47],grey:[128,128,128],honeydew:[240,255,240],hotpink:[255,105,180],indianred:[205,92,92],indigo:[75,0,130],ivory:[255,255,240],khaki:[240,230,140],lavender:[230,230,250],lavenderblush:[255,240,245],lawngreen:[124,252,0],lemonchiffon:[255,250,205],lightblue:[173,216,230],lightcoral:[240,128,128],lightcyan:[224,255,255],lightgoldenrodyellow:[250,250,210],lightgray:[211,211,211],lightgreen:[144,238,144],lightgrey:[211,211,211],lightpink:[255,182,193],lightsalmon:[255,160,122],lightseagreen:[32,178,170],lightskyblue:[135,206,250],lightslategray:[119,136,153],lightslategrey:[119,136,153],lightsteelblue:[176,196,222],lightyellow:[255,255,224],lime:[0,255,0],limegreen:[50,205,50],linen:[250,240,230],magenta:[255,0,255],maroon:[128,0,0],mediumaquamarine:[102,205,170],mediumblue:[0,0,205],mediumorchid:[186,85,211],mediumpurple:[147,112,219],mediumseagreen:[60,179,113],mediumslateblue:[123,104,238],mediumspringgreen:[0,250,154],mediumturquoise:[72,209,204],mediumvioletred:[199,21,133],midnightblue:[25,25,112],mintcream:[245,255,250],mistyrose:[255,228,225],moccasin:[255,228,181],navajowhite:[255,222,173],navy:[0,0,128],oldlace:[253,245,230],olive:[128,128,0],olivedrab:[107,142,35],orange:[255,165,0],orangered:[255,69,0],orchid:[218,112,214],palegoldenrod:[238,232,170],palegreen:[152,251,152],paleturquoise:[175,238,238],palevioletred:[219,112,147],papayawhip:[255,239,213],peachpuff:[255,218,185],peru:[205,133,63],pink:[255,192,203],plum:[221,160,221],powderblue:[176,224,230],purple:[128,0,128],rebeccapurple:[102,51,153],red:[255,0,0],rosybrown:[188,143,143],royalblue:[65,105,225],saddlebrown:[139,69,19],salmon:[250,128,114],sandybrown:[244,164,96],seagreen:[46,139,87],seashell:[255,245,238],sienna:[160,82,45],silver:[192,192,192],skyblue:[135,206,235],slateblue:[106,90,205],slategray:[112,128,144],slategrey:[112,128,144],snow:[255,250,250],springgreen:[0,255,127],steelblue:[70,130,180],tan:[210,180,140],teal:[0,128,128],thistle:[216,191,216],tomato:[255,99,71],turquoise:[64,224,208],violet:[238,130,238],wheat:[245,222,179],white:[255,255,255],whitesmoke:[245,245,245],yellow:[255,255,0],yellowgreen:[154,205,50]}});function $r(r,{loose:e=!1}={}){if(typeof r!="string")return null;if(r=r.trim(),r==="transparent")return{mode:"rgb",color:["0","0","0"],alpha:"0"};if(r in Vs)return{mode:"rgb",color:Vs[r].map(a=>a.toString())};let t=r.replace(Fv,(a,s,o,l,c)=>["#",s,s,o,o,l,l,c?c+c:""].join("")).match(Bv);if(t!==null)return{mode:"rgb",color:[parseInt(t[1],16),parseInt(t[2],16),parseInt(t[3],16)].map(a=>a.toString()),alpha:t[4]?(parseInt(t[4],16)/255).toString():void 0};let i=r.match(jv)??r.match(zv);if(i===null)return null;let n=[i[2],i[3],i[4]].filter(Boolean).map(a=>a.toString());return n.length===2&&n[0].startsWith("var(")?{mode:i[1],color:[n[0]],alpha:n[1]}:!e&&n.length!==3||n.length<3&&!n.some(a=>/^var\(.*?\)$/.test(a))?null:{mode:i[1],color:n,alpha:i[5]?.toString?.()}}function Hs({mode:r,color:e,alpha:t}){let i=t!==void 0;return r==="rgba"||r==="hsla"?`${r}(${e.join(", ")}${i?`, ${t}`:""})`:`${r}(${e.join(" ")}${i?` / ${t}`:""})`}var Bv,Fv,At,Xi,If,Ct,jv,zv,Ws=R(()=>{u();Pf();Bv=/^#([a-f\d]{2})([a-f\d]{2})([a-f\d]{2})([a-f\d]{2})?$/i,Fv=/^#([a-f\d])([a-f\d])([a-f\d])([a-f\d])?$/i,At=/(?:\d+|\d*\.\d+)%?/,Xi=/(?:\s*,\s*|\s+)/,If=/\s*[,/]\s*/,Ct=/var\(--(?:[^ )]*?)(?:,(?:[^ )]*?|var\(--[^ )]*?\)))?\)/,jv=new RegExp(`^(rgba?)\\(\\s*(${At.source}|${Ct.source})(?:${Xi.source}(${At.source}|${Ct.source}))?(?:${Xi.source}(${At.source}|${Ct.source}))?(?:${If.source}(${At.source}|${Ct.source}))?\\s*\\)$`),zv=new RegExp(`^(hsla?)\\(\\s*((?:${At.source})(?:deg|rad|grad|turn)?|${Ct.source})(?:${Xi.source}(${At.source}|${Ct.source}))?(?:${Xi.source}(${At.source}|${Ct.source}))?(?:${If.source}(${At.source}|${Ct.source}))?\\s*\\)$`)});function Ze(r,e,t){if(typeof r=="function")return r({opacityValue:e});let i=$r(r,{loose:!0});return i===null?t:Hs({...i,alpha:e})}function Ae({color:r,property:e,variable:t}){let i=[].concat(e);if(typeof r=="function")return{[t]:"1",...Object.fromEntries(i.map(a=>[a,r({opacityVariable:t,opacityValue:`var(${t})`})]))};let n=$r(r);return n===null?Object.fromEntries(i.map(a=>[a,r])):n.alpha!==void 0?Object.fromEntries(i.map(a=>[a,r])):{[t]:"1",...Object.fromEntries(i.map(a=>[a,Hs({...n,alpha:`var(${t})`})]))}}var Lr=R(()=>{u();Ws()});function ve(r,e){let t=[],i=[],n=0,a=!1;for(let s=0;s{u()});function Ji(r){return ve(r,",").map(t=>{let i=t.trim(),n={raw:i},a=i.split(Vv),s=new Set;for(let o of a)Df.lastIndex=0,!s.has("KEYWORD")&&Uv.has(o)?(n.keyword=o,s.add("KEYWORD")):Df.test(o)?s.has("X")?s.has("Y")?s.has("BLUR")?s.has("SPREAD")||(n.spread=o,s.add("SPREAD")):(n.blur=o,s.add("BLUR")):(n.y=o,s.add("Y")):(n.x=o,s.add("X")):n.color?(n.unknown||(n.unknown=[]),n.unknown.push(o)):n.color=o;return n.valid=n.x!==void 0&&n.y!==void 0,n})}function qf(r){return r.map(e=>e.valid?[e.keyword,e.x,e.y,e.blur,e.spread,e.color].filter(Boolean).join(" "):e.raw).join(", ")}var Uv,Vv,Df,Gs=R(()=>{u();zt();Uv=new Set(["inset","inherit","initial","revert","unset"]),Vv=/\ +(?![^(]*\))/g,Df=/^-?(\d+|\.\d+)(.*?)$/g});function Qs(r){return Hv.some(e=>new RegExp(`^${e}\\(.*\\)`).test(r))}function K(r,e=null,t=!0){let i=e&&Wv.has(e.property);return r.startsWith("--")&&!i?`var(${r})`:r.includes("url(")?r.split(/(url\(.*?\))/g).filter(Boolean).map(n=>/^url\(.*?\)$/.test(n)?n:K(n,e,!1)).join(""):(r=r.replace(/([^\\])_+/g,(n,a)=>a+" ".repeat(n.length-1)).replace(/^_/g," ").replace(/\\_/g,"_"),t&&(r=r.trim()),r=Gv(r),r)}function Ye(r){return r.includes("=")&&(r=r.replace(/(=.*)/g,(e,t)=>{if(t[1]==="'"||t[1]==='"')return t;if(t.length>2){let i=t[t.length-1];if(t[t.length-2]===" "&&(i==="i"||i==="I"||i==="s"||i==="S"))return`="${t.slice(1,-2)}" ${t[t.length-1]}`}return`="${t.slice(1)}"`})),r}function Gv(r){let e=["theme"],t=["min-content","max-content","fit-content","safe-area-inset-top","safe-area-inset-right","safe-area-inset-bottom","safe-area-inset-left","titlebar-area-x","titlebar-area-y","titlebar-area-width","titlebar-area-height","keyboard-inset-top","keyboard-inset-right","keyboard-inset-bottom","keyboard-inset-left","keyboard-inset-width","keyboard-inset-height","radial-gradient","linear-gradient","conic-gradient","repeating-radial-gradient","repeating-linear-gradient","repeating-conic-gradient","anchor-size"];return r.replace(/(calc|min|max|clamp)\(.+\)/g,i=>{let n="";function a(){let s=n.trimEnd();return s[s.length-1]}for(let s=0;si[s+p]===d)},l=function(f){let d=1/0;for(let h of f){let b=i.indexOf(h,s);b!==-1&&bo(f))){let f=t.find(d=>o(d));n+=f,s+=f.length-1}else e.some(f=>o(f))?n+=l([")"]):o("[")?n+=l(["]"]):["+","-","*","/"].includes(c)&&!["(","+","-","*","/",","].includes(a())?n+=` ${c} `:n+=c}return n.replace(/\s+/g," ")})}function Ys(r){return r.startsWith("url(")}function Ks(r){return!isNaN(Number(r))||Qs(r)}function Mr(r){return r.endsWith("%")&&Ks(r.slice(0,-1))||Qs(r)}function Nr(r){return r==="0"||new RegExp(`^[+-]?[0-9]*.?[0-9]+(?:[eE][+-]?[0-9]+)?${Yv}$`).test(r)||Qs(r)}function $f(r){return Kv.has(r)}function Lf(r){let e=Ji(K(r));for(let t of e)if(!t.valid)return!1;return!0}function Mf(r){let e=0;return ve(r,"_").every(i=>(i=K(i),i.startsWith("var(")?!0:$r(i,{loose:!0})!==null?(e++,!0):!1))?e>0:!1}function Nf(r){let e=0;return ve(r,",").every(i=>(i=K(i),i.startsWith("var(")?!0:Ys(i)||Jv(i)||["element(","image(","cross-fade(","image-set("].some(n=>i.startsWith(n))?(e++,!0):!1))?e>0:!1}function Jv(r){r=K(r);for(let e of Xv)if(r.startsWith(`${e}(`))return!0;return!1}function Bf(r){let e=0;return ve(r,"_").every(i=>(i=K(i),i.startsWith("var(")?!0:Zv.has(i)||Nr(i)||Mr(i)?(e++,!0):!1))?e>0:!1}function Ff(r){let e=0;return ve(r,",").every(i=>(i=K(i),i.startsWith("var(")?!0:i.includes(" ")&&!/(['"])([^"']+)\1/g.test(i)||/^\d/g.test(i)?!1:(e++,!0)))?e>0:!1}function jf(r){return ex.has(r)}function zf(r){return tx.has(r)}function Uf(r){return rx.has(r)}var Hv,Wv,Qv,Yv,Kv,Xv,Zv,ex,tx,rx,Br=R(()=>{u();Ws();Gs();zt();Hv=["min","max","clamp","calc"];Wv=new Set(["scroll-timeline-name","timeline-scope","view-timeline-name","font-palette","anchor-name","anchor-scope","position-anchor","position-try-options","scroll-timeline","animation-timeline","view-timeline","position-try"]);Qv=["cm","mm","Q","in","pc","pt","px","em","ex","ch","rem","lh","rlh","vw","vh","vmin","vmax","vb","vi","svw","svh","lvw","lvh","dvw","dvh","cqw","cqh","cqi","cqb","cqmin","cqmax"],Yv=`(?:${Qv.join("|")})`;Kv=new Set(["thin","medium","thick"]);Xv=new Set(["conic-gradient","linear-gradient","radial-gradient","repeating-conic-gradient","repeating-linear-gradient","repeating-radial-gradient"]);Zv=new Set(["center","top","right","bottom","left"]);ex=new Set(["serif","sans-serif","monospace","cursive","fantasy","system-ui","ui-serif","ui-sans-serif","ui-monospace","ui-rounded","math","emoji","fangsong"]);tx=new Set(["xx-small","x-small","small","medium","large","x-large","xx-large","xxx-large"]);rx=new Set(["larger","smaller"])});function Vf(r){let e=["cover","contain"];return ve(r,",").every(t=>{let i=ve(t,"_").filter(Boolean);return i.length===1&&e.includes(i[0])?!0:i.length!==1&&i.length!==2?!1:i.every(n=>Nr(n)||Mr(n)||n==="auto")})}var Hf=R(()=>{u();Br();zt()});function Wf(r,e){r.walkClasses(t=>{t.value=e(t.value),t.raws&&t.raws.value&&(t.raws.value=jt(t.raws.value))})}function Gf(r,e){if(!_t(r))return;let t=r.slice(1,-1);if(!!e(t))return K(t)}function ix(r,e={},t){let i=e[r];if(i!==void 0)return xt(i);if(_t(r)){let n=Gf(r,t);return n===void 0?void 0:xt(n)}}function Zi(r,e={},{validate:t=()=>!0}={}){let i=e.values?.[r];return i!==void 0?i:e.supportsNegativeValues&&r.startsWith("-")?ix(r.slice(1),e.values,t):Gf(r,t)}function _t(r){return r.startsWith("[")&&r.endsWith("]")}function Qf(r){let e=r.lastIndexOf("/"),t=r.lastIndexOf("[",e),i=r.indexOf("]",e);return r[e-1]==="]"||r[e+1]==="["||t!==-1&&i!==-1&&t")){let e=r;return({opacityValue:t=1})=>e.replace(//g,t)}return r}function Yf(r){return K(r.slice(1,-1))}function nx(r,e={},{tailwindConfig:t={}}={}){if(e.values?.[r]!==void 0)return Xt(e.values?.[r]);let[i,n]=Qf(r);if(n!==void 0){let a=e.values?.[i]??(_t(i)?i.slice(1,-1):void 0);return a===void 0?void 0:(a=Xt(a),_t(n)?Ze(a,Yf(n)):t.theme?.opacity?.[n]===void 0?void 0:Ze(a,t.theme.opacity[n]))}return Zi(r,e,{validate:Mf})}function sx(r,e={}){return e.values?.[r]}function qe(r){return(e,t)=>Zi(e,t,{validate:r})}function ax(r,e){let t=r.indexOf(e);return t===-1?[void 0,r]:[r.slice(0,t),r.slice(t+1)]}function Js(r,e,t,i){if(t.values&&e in t.values)for(let{type:a}of r??[]){let s=Xs[a](e,t,{tailwindConfig:i});if(s!==void 0)return[s,a,null]}if(_t(e)){let a=e.slice(1,-1),[s,o]=ax(a,":");if(!/^[\w-_]+$/g.test(s))o=a;else if(s!==void 0&&!Kf.includes(s))return[];if(o.length>0&&Kf.includes(s))return[Zi(`[${o}]`,t),s,null]}let n=Zs(r,e,t,i);for(let a of n)return a;return[]}function*Zs(r,e,t,i){let n=we(i,"generalizedModifiers"),[a,s]=Qf(e);if(n&&t.modifiers!=null&&(t.modifiers==="any"||typeof t.modifiers=="object"&&(s&&_t(s)||s in t.modifiers))||(a=e,s=void 0),s!==void 0&&a===""&&(a="DEFAULT"),s!==void 0&&typeof t.modifiers=="object"){let l=t.modifiers?.[s]??null;l!==null?s=l:_t(s)&&(s=Yf(s))}for(let{type:l}of r??[]){let c=Xs[l](a,t,{tailwindConfig:i});c!==void 0&&(yield[c,l,s??null])}}var Xs,Kf,Fr=R(()=>{u();Ki();Lr();Br();Hi();Hf();ct();Xs={any:Zi,color:nx,url:qe(Ys),image:qe(Nf),length:qe(Nr),percentage:qe(Mr),position:qe(Bf),lookup:sx,"generic-name":qe(jf),"family-name":qe(Ff),number:qe(Ks),"line-width":qe($f),"absolute-size":qe(zf),"relative-size":qe(Uf),shadow:qe(Lf),size:qe(Vf)},Kf=Object.keys(Xs)});function X(r){return typeof r=="function"?r({}):r}var ea=R(()=>{u()});function Jt(r){return typeof r=="function"}function jr(r,...e){let t=e.pop();for(let i of e)for(let n in i){let a=t(r[n],i[n]);a===void 0?ke(r[n])&&ke(i[n])?r[n]=jr({},r[n],i[n],t):r[n]=i[n]:r[n]=a}return r}function ox(r,...e){return Jt(r)?r(...e):r}function lx(r){return r.reduce((e,{extend:t})=>jr(e,t,(i,n)=>i===void 0?[n]:Array.isArray(i)?[n,...i]:[n,i]),{})}function ux(r){return{...r.reduce((e,t)=>Us(e,t),{}),extend:lx(r)}}function Xf(r,e){if(Array.isArray(r)&&ke(r[0]))return r.concat(e);if(Array.isArray(e)&&ke(e[0])&&ke(r))return[r,...e];if(Array.isArray(e))return e}function fx({extend:r,...e}){return jr(e,r,(t,i)=>!Jt(t)&&!i.some(Jt)?jr({},t,...i,Xf):(n,a)=>jr({},...[t,...i].map(s=>ox(s,n,a)),Xf))}function*cx(r){let e=kt(r);if(e.length===0||(yield e,Array.isArray(r)))return;let t=/^(.*?)\s*\/\s*([^/]+)$/,i=r.match(t);if(i!==null){let[,n,a]=i,s=kt(n);s.alpha=a,yield s}}function px(r){let e=(t,i)=>{for(let n of cx(t)){let a=0,s=r;for(;s!=null&&a(t[i]=Jt(r[i])?r[i](e,ta):r[i],t),{})}function Jf(r){let e=[];return r.forEach(t=>{e=[...e,t];let i=t?.plugins??[];i.length!==0&&i.forEach(n=>{n.__isOptionsFunction&&(n=n()),e=[...e,...Jf([n?.config??{}])]})}),e}function dx(r){return[...r].reduceRight((t,i)=>Jt(i)?i({corePlugins:t}):vf(i,t),bf)}function hx(r){return[...r].reduceRight((t,i)=>[...t,...i],[])}function ra(r){let e=[...Jf(r),{prefix:"",important:!1,separator:":"}];return Tf(Us({theme:px(fx(ux(e.map(t=>t?.theme??{})))),corePlugins:dx(e.map(t=>t.corePlugins)),plugins:hx(r.map(t=>t?.plugins??[]))},...e))}var ta,Zf=R(()=>{u();Hi();wf();xf();zs();Cf();Gi();Rf();Kt();Yi();Fr();Lr();ea();ta={colors:js,negative(r){return Object.keys(r).filter(e=>r[e]!=="0").reduce((e,t)=>{let i=xt(r[t]);return i!==void 0&&(e[`-${t}`]=i),e},{})},breakpoints(r){return Object.keys(r).filter(e=>typeof r[e]=="string").reduce((e,t)=>({...e,[`screen-${t}`]:r[t]}),{})}}});var en=x((l3,ec)=>{u();ec.exports={content:[],presets:[],darkMode:"media",theme:{accentColor:({theme:r})=>({...r("colors"),auto:"auto"}),animation:{none:"none",spin:"spin 1s linear infinite",ping:"ping 1s cubic-bezier(0, 0, 0.2, 1) infinite",pulse:"pulse 2s cubic-bezier(0.4, 0, 0.6, 1) infinite",bounce:"bounce 1s infinite"},aria:{busy:'busy="true"',checked:'checked="true"',disabled:'disabled="true"',expanded:'expanded="true"',hidden:'hidden="true"',pressed:'pressed="true"',readonly:'readonly="true"',required:'required="true"',selected:'selected="true"'},aspectRatio:{auto:"auto",square:"1 / 1",video:"16 / 9"},backdropBlur:({theme:r})=>r("blur"),backdropBrightness:({theme:r})=>r("brightness"),backdropContrast:({theme:r})=>r("contrast"),backdropGrayscale:({theme:r})=>r("grayscale"),backdropHueRotate:({theme:r})=>r("hueRotate"),backdropInvert:({theme:r})=>r("invert"),backdropOpacity:({theme:r})=>r("opacity"),backdropSaturate:({theme:r})=>r("saturate"),backdropSepia:({theme:r})=>r("sepia"),backgroundColor:({theme:r})=>r("colors"),backgroundImage:{none:"none","gradient-to-t":"linear-gradient(to top, var(--tw-gradient-stops))","gradient-to-tr":"linear-gradient(to top right, var(--tw-gradient-stops))","gradient-to-r":"linear-gradient(to right, var(--tw-gradient-stops))","gradient-to-br":"linear-gradient(to bottom right, var(--tw-gradient-stops))","gradient-to-b":"linear-gradient(to bottom, var(--tw-gradient-stops))","gradient-to-bl":"linear-gradient(to bottom left, var(--tw-gradient-stops))","gradient-to-l":"linear-gradient(to left, var(--tw-gradient-stops))","gradient-to-tl":"linear-gradient(to top left, var(--tw-gradient-stops))"},backgroundOpacity:({theme:r})=>r("opacity"),backgroundPosition:{bottom:"bottom",center:"center",left:"left","left-bottom":"left bottom","left-top":"left top",right:"right","right-bottom":"right bottom","right-top":"right top",top:"top"},backgroundSize:{auto:"auto",cover:"cover",contain:"contain"},blur:{0:"0",none:"",sm:"4px",DEFAULT:"8px",md:"12px",lg:"16px",xl:"24px","2xl":"40px","3xl":"64px"},borderColor:({theme:r})=>({...r("colors"),DEFAULT:r("colors.gray.200","currentColor")}),borderOpacity:({theme:r})=>r("opacity"),borderRadius:{none:"0px",sm:"0.125rem",DEFAULT:"0.25rem",md:"0.375rem",lg:"0.5rem",xl:"0.75rem","2xl":"1rem","3xl":"1.5rem",full:"9999px"},borderSpacing:({theme:r})=>({...r("spacing")}),borderWidth:{DEFAULT:"1px",0:"0px",2:"2px",4:"4px",8:"8px"},boxShadow:{sm:"0 1px 2px 0 rgb(0 0 0 / 0.05)",DEFAULT:"0 1px 3px 0 rgb(0 0 0 / 0.1), 0 1px 2px -1px rgb(0 0 0 / 0.1)",md:"0 4px 6px -1px rgb(0 0 0 / 0.1), 0 2px 4px -2px rgb(0 0 0 / 0.1)",lg:"0 10px 15px -3px rgb(0 0 0 / 0.1), 0 4px 6px -4px rgb(0 0 0 / 0.1)",xl:"0 20px 25px -5px rgb(0 0 0 / 0.1), 0 8px 10px -6px rgb(0 0 0 / 0.1)","2xl":"0 25px 50px -12px rgb(0 0 0 / 0.25)",inner:"inset 0 2px 4px 0 rgb(0 0 0 / 0.05)",none:"none"},boxShadowColor:({theme:r})=>r("colors"),brightness:{0:"0",50:".5",75:".75",90:".9",95:".95",100:"1",105:"1.05",110:"1.1",125:"1.25",150:"1.5",200:"2"},caretColor:({theme:r})=>r("colors"),colors:({colors:r})=>({inherit:r.inherit,current:r.current,transparent:r.transparent,black:r.black,white:r.white,slate:r.slate,gray:r.gray,zinc:r.zinc,neutral:r.neutral,stone:r.stone,red:r.red,orange:r.orange,amber:r.amber,yellow:r.yellow,lime:r.lime,green:r.green,emerald:r.emerald,teal:r.teal,cyan:r.cyan,sky:r.sky,blue:r.blue,indigo:r.indigo,violet:r.violet,purple:r.purple,fuchsia:r.fuchsia,pink:r.pink,rose:r.rose}),columns:{auto:"auto",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12","3xs":"16rem","2xs":"18rem",xs:"20rem",sm:"24rem",md:"28rem",lg:"32rem",xl:"36rem","2xl":"42rem","3xl":"48rem","4xl":"56rem","5xl":"64rem","6xl":"72rem","7xl":"80rem"},container:{},content:{none:"none"},contrast:{0:"0",50:".5",75:".75",100:"1",125:"1.25",150:"1.5",200:"2"},cursor:{auto:"auto",default:"default",pointer:"pointer",wait:"wait",text:"text",move:"move",help:"help","not-allowed":"not-allowed",none:"none","context-menu":"context-menu",progress:"progress",cell:"cell",crosshair:"crosshair","vertical-text":"vertical-text",alias:"alias",copy:"copy","no-drop":"no-drop",grab:"grab",grabbing:"grabbing","all-scroll":"all-scroll","col-resize":"col-resize","row-resize":"row-resize","n-resize":"n-resize","e-resize":"e-resize","s-resize":"s-resize","w-resize":"w-resize","ne-resize":"ne-resize","nw-resize":"nw-resize","se-resize":"se-resize","sw-resize":"sw-resize","ew-resize":"ew-resize","ns-resize":"ns-resize","nesw-resize":"nesw-resize","nwse-resize":"nwse-resize","zoom-in":"zoom-in","zoom-out":"zoom-out"},divideColor:({theme:r})=>r("borderColor"),divideOpacity:({theme:r})=>r("borderOpacity"),divideWidth:({theme:r})=>r("borderWidth"),dropShadow:{sm:"0 1px 1px rgb(0 0 0 / 0.05)",DEFAULT:["0 1px 2px rgb(0 0 0 / 0.1)","0 1px 1px rgb(0 0 0 / 0.06)"],md:["0 4px 3px rgb(0 0 0 / 0.07)","0 2px 2px rgb(0 0 0 / 0.06)"],lg:["0 10px 8px rgb(0 0 0 / 0.04)","0 4px 3px rgb(0 0 0 / 0.1)"],xl:["0 20px 13px rgb(0 0 0 / 0.03)","0 8px 5px rgb(0 0 0 / 0.08)"],"2xl":"0 25px 25px rgb(0 0 0 / 0.15)",none:"0 0 #0000"},fill:({theme:r})=>({none:"none",...r("colors")}),flex:{1:"1 1 0%",auto:"1 1 auto",initial:"0 1 auto",none:"none"},flexBasis:({theme:r})=>({auto:"auto",...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%","1/5":"20%","2/5":"40%","3/5":"60%","4/5":"80%","1/6":"16.666667%","2/6":"33.333333%","3/6":"50%","4/6":"66.666667%","5/6":"83.333333%","1/12":"8.333333%","2/12":"16.666667%","3/12":"25%","4/12":"33.333333%","5/12":"41.666667%","6/12":"50%","7/12":"58.333333%","8/12":"66.666667%","9/12":"75%","10/12":"83.333333%","11/12":"91.666667%",full:"100%"}),flexGrow:{0:"0",DEFAULT:"1"},flexShrink:{0:"0",DEFAULT:"1"},fontFamily:{sans:["ui-sans-serif","system-ui","sans-serif",'"Apple Color Emoji"','"Segoe UI Emoji"','"Segoe UI Symbol"','"Noto Color Emoji"'],serif:["ui-serif","Georgia","Cambria",'"Times New Roman"',"Times","serif"],mono:["ui-monospace","SFMono-Regular","Menlo","Monaco","Consolas",'"Liberation Mono"','"Courier New"',"monospace"]},fontSize:{xs:["0.75rem",{lineHeight:"1rem"}],sm:["0.875rem",{lineHeight:"1.25rem"}],base:["1rem",{lineHeight:"1.5rem"}],lg:["1.125rem",{lineHeight:"1.75rem"}],xl:["1.25rem",{lineHeight:"1.75rem"}],"2xl":["1.5rem",{lineHeight:"2rem"}],"3xl":["1.875rem",{lineHeight:"2.25rem"}],"4xl":["2.25rem",{lineHeight:"2.5rem"}],"5xl":["3rem",{lineHeight:"1"}],"6xl":["3.75rem",{lineHeight:"1"}],"7xl":["4.5rem",{lineHeight:"1"}],"8xl":["6rem",{lineHeight:"1"}],"9xl":["8rem",{lineHeight:"1"}]},fontWeight:{thin:"100",extralight:"200",light:"300",normal:"400",medium:"500",semibold:"600",bold:"700",extrabold:"800",black:"900"},gap:({theme:r})=>r("spacing"),gradientColorStops:({theme:r})=>r("colors"),gradientColorStopPositions:{"0%":"0%","5%":"5%","10%":"10%","15%":"15%","20%":"20%","25%":"25%","30%":"30%","35%":"35%","40%":"40%","45%":"45%","50%":"50%","55%":"55%","60%":"60%","65%":"65%","70%":"70%","75%":"75%","80%":"80%","85%":"85%","90%":"90%","95%":"95%","100%":"100%"},grayscale:{0:"0",DEFAULT:"100%"},gridAutoColumns:{auto:"auto",min:"min-content",max:"max-content",fr:"minmax(0, 1fr)"},gridAutoRows:{auto:"auto",min:"min-content",max:"max-content",fr:"minmax(0, 1fr)"},gridColumn:{auto:"auto","span-1":"span 1 / span 1","span-2":"span 2 / span 2","span-3":"span 3 / span 3","span-4":"span 4 / span 4","span-5":"span 5 / span 5","span-6":"span 6 / span 6","span-7":"span 7 / span 7","span-8":"span 8 / span 8","span-9":"span 9 / span 9","span-10":"span 10 / span 10","span-11":"span 11 / span 11","span-12":"span 12 / span 12","span-full":"1 / -1"},gridColumnEnd:{auto:"auto",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12",13:"13"},gridColumnStart:{auto:"auto",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12",13:"13"},gridRow:{auto:"auto","span-1":"span 1 / span 1","span-2":"span 2 / span 2","span-3":"span 3 / span 3","span-4":"span 4 / span 4","span-5":"span 5 / span 5","span-6":"span 6 / span 6","span-7":"span 7 / span 7","span-8":"span 8 / span 8","span-9":"span 9 / span 9","span-10":"span 10 / span 10","span-11":"span 11 / span 11","span-12":"span 12 / span 12","span-full":"1 / -1"},gridRowEnd:{auto:"auto",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12",13:"13"},gridRowStart:{auto:"auto",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12",13:"13"},gridTemplateColumns:{none:"none",subgrid:"subgrid",1:"repeat(1, minmax(0, 1fr))",2:"repeat(2, minmax(0, 1fr))",3:"repeat(3, minmax(0, 1fr))",4:"repeat(4, minmax(0, 1fr))",5:"repeat(5, minmax(0, 1fr))",6:"repeat(6, minmax(0, 1fr))",7:"repeat(7, minmax(0, 1fr))",8:"repeat(8, minmax(0, 1fr))",9:"repeat(9, minmax(0, 1fr))",10:"repeat(10, minmax(0, 1fr))",11:"repeat(11, minmax(0, 1fr))",12:"repeat(12, minmax(0, 1fr))"},gridTemplateRows:{none:"none",subgrid:"subgrid",1:"repeat(1, minmax(0, 1fr))",2:"repeat(2, minmax(0, 1fr))",3:"repeat(3, minmax(0, 1fr))",4:"repeat(4, minmax(0, 1fr))",5:"repeat(5, minmax(0, 1fr))",6:"repeat(6, minmax(0, 1fr))",7:"repeat(7, minmax(0, 1fr))",8:"repeat(8, minmax(0, 1fr))",9:"repeat(9, minmax(0, 1fr))",10:"repeat(10, minmax(0, 1fr))",11:"repeat(11, minmax(0, 1fr))",12:"repeat(12, minmax(0, 1fr))"},height:({theme:r})=>({auto:"auto",...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%","1/5":"20%","2/5":"40%","3/5":"60%","4/5":"80%","1/6":"16.666667%","2/6":"33.333333%","3/6":"50%","4/6":"66.666667%","5/6":"83.333333%",full:"100%",screen:"100vh",svh:"100svh",lvh:"100lvh",dvh:"100dvh",min:"min-content",max:"max-content",fit:"fit-content"}),hueRotate:{0:"0deg",15:"15deg",30:"30deg",60:"60deg",90:"90deg",180:"180deg"},inset:({theme:r})=>({auto:"auto",...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%",full:"100%"}),invert:{0:"0",DEFAULT:"100%"},keyframes:{spin:{to:{transform:"rotate(360deg)"}},ping:{"75%, 100%":{transform:"scale(2)",opacity:"0"}},pulse:{"50%":{opacity:".5"}},bounce:{"0%, 100%":{transform:"translateY(-25%)",animationTimingFunction:"cubic-bezier(0.8,0,1,1)"},"50%":{transform:"none",animationTimingFunction:"cubic-bezier(0,0,0.2,1)"}}},letterSpacing:{tighter:"-0.05em",tight:"-0.025em",normal:"0em",wide:"0.025em",wider:"0.05em",widest:"0.1em"},lineHeight:{none:"1",tight:"1.25",snug:"1.375",normal:"1.5",relaxed:"1.625",loose:"2",3:".75rem",4:"1rem",5:"1.25rem",6:"1.5rem",7:"1.75rem",8:"2rem",9:"2.25rem",10:"2.5rem"},listStyleType:{none:"none",disc:"disc",decimal:"decimal"},listStyleImage:{none:"none"},margin:({theme:r})=>({auto:"auto",...r("spacing")}),lineClamp:{1:"1",2:"2",3:"3",4:"4",5:"5",6:"6"},maxHeight:({theme:r})=>({...r("spacing"),none:"none",full:"100%",screen:"100vh",svh:"100svh",lvh:"100lvh",dvh:"100dvh",min:"min-content",max:"max-content",fit:"fit-content"}),maxWidth:({theme:r,breakpoints:e})=>({...r("spacing"),none:"none",xs:"20rem",sm:"24rem",md:"28rem",lg:"32rem",xl:"36rem","2xl":"42rem","3xl":"48rem","4xl":"56rem","5xl":"64rem","6xl":"72rem","7xl":"80rem",full:"100%",min:"min-content",max:"max-content",fit:"fit-content",prose:"65ch",...e(r("screens"))}),minHeight:({theme:r})=>({...r("spacing"),full:"100%",screen:"100vh",svh:"100svh",lvh:"100lvh",dvh:"100dvh",min:"min-content",max:"max-content",fit:"fit-content"}),minWidth:({theme:r})=>({...r("spacing"),full:"100%",min:"min-content",max:"max-content",fit:"fit-content"}),objectPosition:{bottom:"bottom",center:"center",left:"left","left-bottom":"left bottom","left-top":"left top",right:"right","right-bottom":"right bottom","right-top":"right top",top:"top"},opacity:{0:"0",5:"0.05",10:"0.1",15:"0.15",20:"0.2",25:"0.25",30:"0.3",35:"0.35",40:"0.4",45:"0.45",50:"0.5",55:"0.55",60:"0.6",65:"0.65",70:"0.7",75:"0.75",80:"0.8",85:"0.85",90:"0.9",95:"0.95",100:"1"},order:{first:"-9999",last:"9999",none:"0",1:"1",2:"2",3:"3",4:"4",5:"5",6:"6",7:"7",8:"8",9:"9",10:"10",11:"11",12:"12"},outlineColor:({theme:r})=>r("colors"),outlineOffset:{0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},outlineWidth:{0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},padding:({theme:r})=>r("spacing"),placeholderColor:({theme:r})=>r("colors"),placeholderOpacity:({theme:r})=>r("opacity"),ringColor:({theme:r})=>({DEFAULT:r("colors.blue.500","#3b82f6"),...r("colors")}),ringOffsetColor:({theme:r})=>r("colors"),ringOffsetWidth:{0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},ringOpacity:({theme:r})=>({DEFAULT:"0.5",...r("opacity")}),ringWidth:{DEFAULT:"3px",0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},rotate:{0:"0deg",1:"1deg",2:"2deg",3:"3deg",6:"6deg",12:"12deg",45:"45deg",90:"90deg",180:"180deg"},saturate:{0:"0",50:".5",100:"1",150:"1.5",200:"2"},scale:{0:"0",50:".5",75:".75",90:".9",95:".95",100:"1",105:"1.05",110:"1.1",125:"1.25",150:"1.5"},screens:{sm:"640px",md:"768px",lg:"1024px",xl:"1280px","2xl":"1536px"},scrollMargin:({theme:r})=>({...r("spacing")}),scrollPadding:({theme:r})=>r("spacing"),sepia:{0:"0",DEFAULT:"100%"},skew:{0:"0deg",1:"1deg",2:"2deg",3:"3deg",6:"6deg",12:"12deg"},space:({theme:r})=>({...r("spacing")}),spacing:{px:"1px",0:"0px",.5:"0.125rem",1:"0.25rem",1.5:"0.375rem",2:"0.5rem",2.5:"0.625rem",3:"0.75rem",3.5:"0.875rem",4:"1rem",5:"1.25rem",6:"1.5rem",7:"1.75rem",8:"2rem",9:"2.25rem",10:"2.5rem",11:"2.75rem",12:"3rem",14:"3.5rem",16:"4rem",20:"5rem",24:"6rem",28:"7rem",32:"8rem",36:"9rem",40:"10rem",44:"11rem",48:"12rem",52:"13rem",56:"14rem",60:"15rem",64:"16rem",72:"18rem",80:"20rem",96:"24rem"},stroke:({theme:r})=>({none:"none",...r("colors")}),strokeWidth:{0:"0",1:"1",2:"2"},supports:{},data:{},textColor:({theme:r})=>r("colors"),textDecorationColor:({theme:r})=>r("colors"),textDecorationThickness:{auto:"auto","from-font":"from-font",0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},textIndent:({theme:r})=>({...r("spacing")}),textOpacity:({theme:r})=>r("opacity"),textUnderlineOffset:{auto:"auto",0:"0px",1:"1px",2:"2px",4:"4px",8:"8px"},transformOrigin:{center:"center",top:"top","top-right":"top right",right:"right","bottom-right":"bottom right",bottom:"bottom","bottom-left":"bottom left",left:"left","top-left":"top left"},transitionDelay:{0:"0s",75:"75ms",100:"100ms",150:"150ms",200:"200ms",300:"300ms",500:"500ms",700:"700ms",1e3:"1000ms"},transitionDuration:{DEFAULT:"150ms",0:"0s",75:"75ms",100:"100ms",150:"150ms",200:"200ms",300:"300ms",500:"500ms",700:"700ms",1e3:"1000ms"},transitionProperty:{none:"none",all:"all",DEFAULT:"color, background-color, border-color, text-decoration-color, fill, stroke, opacity, box-shadow, transform, filter, backdrop-filter",colors:"color, background-color, border-color, text-decoration-color, fill, stroke",opacity:"opacity",shadow:"box-shadow",transform:"transform"},transitionTimingFunction:{DEFAULT:"cubic-bezier(0.4, 0, 0.2, 1)",linear:"linear",in:"cubic-bezier(0.4, 0, 1, 1)",out:"cubic-bezier(0, 0, 0.2, 1)","in-out":"cubic-bezier(0.4, 0, 0.2, 1)"},translate:({theme:r})=>({...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%",full:"100%"}),size:({theme:r})=>({auto:"auto",...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%","1/5":"20%","2/5":"40%","3/5":"60%","4/5":"80%","1/6":"16.666667%","2/6":"33.333333%","3/6":"50%","4/6":"66.666667%","5/6":"83.333333%","1/12":"8.333333%","2/12":"16.666667%","3/12":"25%","4/12":"33.333333%","5/12":"41.666667%","6/12":"50%","7/12":"58.333333%","8/12":"66.666667%","9/12":"75%","10/12":"83.333333%","11/12":"91.666667%",full:"100%",min:"min-content",max:"max-content",fit:"fit-content"}),width:({theme:r})=>({auto:"auto",...r("spacing"),"1/2":"50%","1/3":"33.333333%","2/3":"66.666667%","1/4":"25%","2/4":"50%","3/4":"75%","1/5":"20%","2/5":"40%","3/5":"60%","4/5":"80%","1/6":"16.666667%","2/6":"33.333333%","3/6":"50%","4/6":"66.666667%","5/6":"83.333333%","1/12":"8.333333%","2/12":"16.666667%","3/12":"25%","4/12":"33.333333%","5/12":"41.666667%","6/12":"50%","7/12":"58.333333%","8/12":"66.666667%","9/12":"75%","10/12":"83.333333%","11/12":"91.666667%",full:"100%",screen:"100vw",svw:"100svw",lvw:"100lvw",dvw:"100dvw",min:"min-content",max:"max-content",fit:"fit-content"}),willChange:{auto:"auto",scroll:"scroll-position",contents:"contents",transform:"transform"},zIndex:{auto:"auto",0:"0",10:"10",20:"20",30:"30",40:"40",50:"50"}},plugins:[]}});function tn(r){let e=(r?.presets??[tc.default]).slice().reverse().flatMap(n=>tn(n instanceof Function?n():n)),t={respectDefaultRingColorOpacity:{theme:{ringColor:({theme:n})=>({DEFAULT:"#3b82f67f",...n("colors")})}},disableColorOpacityUtilitiesByDefault:{corePlugins:{backgroundOpacity:!1,borderOpacity:!1,divideOpacity:!1,placeholderOpacity:!1,ringOpacity:!1,textOpacity:!1}}},i=Object.keys(t).filter(n=>we(r,n)).map(n=>t[n]);return[r,...i,...e]}var tc,rc=R(()=>{u();tc=pe(en());ct()});var ic={};Ge(ic,{default:()=>zr});function zr(...r){let[,...e]=tn(r[0]);return ra([...r,...e])}var ia=R(()=>{u();Zf();rc()});var Ur={};Ge(Ur,{default:()=>me});var me,et=R(()=>{u();me={resolve:r=>r,extname:r=>"."+r.split(".").pop()}});function rn(r){return typeof r=="object"&&r!==null}function gx(r){return Object.keys(r).length===0}function nc(r){return typeof r=="string"||r instanceof String}function na(r){return rn(r)&&r.config===void 0&&!gx(r)?null:rn(r)&&r.config!==void 0&&nc(r.config)?me.resolve(r.config):rn(r)&&r.config!==void 0&&rn(r.config)?null:nc(r)?me.resolve(r):yx()}function yx(){for(let r of mx)try{let e=me.resolve(r);return be.accessSync(e),e}catch(e){}return null}var mx,sc=R(()=>{u();ft();et();mx=["./tailwind.config.js","./tailwind.config.cjs","./tailwind.config.mjs","./tailwind.config.ts","./tailwind.config.cts","./tailwind.config.mts"]});var ac={};Ge(ac,{default:()=>sa});var sa,aa=R(()=>{u();sa={parse:r=>({href:r})}});var oa=x(()=>{u()});var nn=x((b3,uc)=>{u();"use strict";var oc=(Wi(),kf),lc=oa(),Zt=class extends Error{constructor(e,t,i,n,a,s){super(e);this.name="CssSyntaxError",this.reason=e,a&&(this.file=a),n&&(this.source=n),s&&(this.plugin=s),typeof t!="undefined"&&typeof i!="undefined"&&(typeof t=="number"?(this.line=t,this.column=i):(this.line=t.line,this.column=t.column,this.endLine=i.line,this.endColumn=i.column)),this.setMessage(),Error.captureStackTrace&&Error.captureStackTrace(this,Zt)}setMessage(){this.message=this.plugin?this.plugin+": ":"",this.message+=this.file?this.file:"",typeof this.line!="undefined"&&(this.message+=":"+this.line+":"+this.column),this.message+=": "+this.reason}showSourceCode(e){if(!this.source)return"";let t=this.source;e==null&&(e=oc.isColorSupported),lc&&e&&(t=lc(t));let i=t.split(/\r?\n/),n=Math.max(this.line-3,0),a=Math.min(this.line+2,i.length),s=String(a).length,o,l;if(e){let{bold:c,red:f,gray:d}=oc.createColors(!0);o=p=>c(f(p)),l=p=>d(p)}else o=l=c=>c;return i.slice(n,a).map((c,f)=>{let d=n+1+f,p=" "+(" "+d).slice(-s)+" | ";if(d===this.line){let h=l(p.replace(/\d/g," "))+c.slice(0,this.column-1).replace(/[^\t]/g," ");return o(">")+l(p)+c+` - `+h+o("^")}return" "+l(p)+c}).join(` -`)}toString(){let e=this.showSourceCode();return e&&(e=` - -`+e+` -`),this.name+": "+this.message+e}};uc.exports=Zt;Zt.default=Zt});var sn=x((w3,la)=>{u();"use strict";la.exports.isClean=Symbol("isClean");la.exports.my=Symbol("my")});var ua=x((v3,cc)=>{u();"use strict";var fc={colon:": ",indent:" ",beforeDecl:` -`,beforeRule:` -`,beforeOpen:" ",beforeClose:` -`,beforeComment:` -`,after:` -`,emptyBody:"",commentLeft:" ",commentRight:" ",semicolon:!1};function bx(r){return r[0].toUpperCase()+r.slice(1)}var an=class{constructor(e){this.builder=e}stringify(e,t){if(!this[e.type])throw new Error("Unknown AST node type "+e.type+". Maybe you need to change PostCSS stringifier.");this[e.type](e,t)}document(e){this.body(e)}root(e){this.body(e),e.raws.after&&this.builder(e.raws.after)}comment(e){let t=this.raw(e,"left","commentLeft"),i=this.raw(e,"right","commentRight");this.builder("/*"+t+e.text+i+"*/",e)}decl(e,t){let i=this.raw(e,"between","colon"),n=e.prop+i+this.rawValue(e,"value");e.important&&(n+=e.raws.important||" !important"),t&&(n+=";"),this.builder(n,e)}rule(e){this.block(e,this.rawValue(e,"selector")),e.raws.ownSemicolon&&this.builder(e.raws.ownSemicolon,e,"end")}atrule(e,t){let i="@"+e.name,n=e.params?this.rawValue(e,"params"):"";if(typeof e.raws.afterName!="undefined"?i+=e.raws.afterName:n&&(i+=" "),e.nodes)this.block(e,i+n);else{let a=(e.raws.between||"")+(t?";":"");this.builder(i+n+a,e)}}body(e){let t=e.nodes.length-1;for(;t>0&&e.nodes[t].type==="comment";)t-=1;let i=this.raw(e,"semicolon");for(let n=0;n{if(n=l.raws[t],typeof n!="undefined")return!1})}return typeof n=="undefined"&&(n=fc[i]),s.rawCache[i]=n,n}rawSemicolon(e){let t;return e.walk(i=>{if(i.nodes&&i.nodes.length&&i.last.type==="decl"&&(t=i.raws.semicolon,typeof t!="undefined"))return!1}),t}rawEmptyBody(e){let t;return e.walk(i=>{if(i.nodes&&i.nodes.length===0&&(t=i.raws.after,typeof t!="undefined"))return!1}),t}rawIndent(e){if(e.raws.indent)return e.raws.indent;let t;return e.walk(i=>{let n=i.parent;if(n&&n!==e&&n.parent&&n.parent===e&&typeof i.raws.before!="undefined"){let a=i.raws.before.split(` -`);return t=a[a.length-1],t=t.replace(/\S/g,""),!1}}),t}rawBeforeComment(e,t){let i;return e.walkComments(n=>{if(typeof n.raws.before!="undefined")return i=n.raws.before,i.includes(` -`)&&(i=i.replace(/[^\n]+$/,"")),!1}),typeof i=="undefined"?i=this.raw(t,null,"beforeDecl"):i&&(i=i.replace(/\S/g,"")),i}rawBeforeDecl(e,t){let i;return e.walkDecls(n=>{if(typeof n.raws.before!="undefined")return i=n.raws.before,i.includes(` -`)&&(i=i.replace(/[^\n]+$/,"")),!1}),typeof i=="undefined"?i=this.raw(t,null,"beforeRule"):i&&(i=i.replace(/\S/g,"")),i}rawBeforeRule(e){let t;return e.walk(i=>{if(i.nodes&&(i.parent!==e||e.first!==i)&&typeof i.raws.before!="undefined")return t=i.raws.before,t.includes(` -`)&&(t=t.replace(/[^\n]+$/,"")),!1}),t&&(t=t.replace(/\S/g,"")),t}rawBeforeClose(e){let t;return e.walk(i=>{if(i.nodes&&i.nodes.length>0&&typeof i.raws.after!="undefined")return t=i.raws.after,t.includes(` -`)&&(t=t.replace(/[^\n]+$/,"")),!1}),t&&(t=t.replace(/\S/g,"")),t}rawBeforeOpen(e){let t;return e.walk(i=>{if(i.type!=="decl"&&(t=i.raws.between,typeof t!="undefined"))return!1}),t}rawColon(e){let t;return e.walkDecls(i=>{if(typeof i.raws.between!="undefined")return t=i.raws.between.replace(/[^\s:]/g,""),!1}),t}beforeAfter(e,t){let i;e.type==="decl"?i=this.raw(e,null,"beforeDecl"):e.type==="comment"?i=this.raw(e,null,"beforeComment"):t==="before"?i=this.raw(e,null,"beforeRule"):i=this.raw(e,null,"beforeClose");let n=e.parent,a=0;for(;n&&n.type!=="root";)a+=1,n=n.parent;if(i.includes(` -`)){let s=this.raw(e,null,"indent");if(s.length)for(let o=0;o{u();"use strict";var wx=ua();function fa(r,e){new wx(e).stringify(r)}pc.exports=fa;fa.default=fa});var Hr=x((k3,dc)=>{u();"use strict";var{isClean:on,my:vx}=sn(),xx=nn(),kx=ua(),Sx=Vr();function ca(r,e){let t=new r.constructor;for(let i in r){if(!Object.prototype.hasOwnProperty.call(r,i)||i==="proxyCache")continue;let n=r[i],a=typeof n;i==="parent"&&a==="object"?e&&(t[i]=e):i==="source"?t[i]=n:Array.isArray(n)?t[i]=n.map(s=>ca(s,t)):(a==="object"&&n!==null&&(n=ca(n)),t[i]=n)}return t}var ln=class{constructor(e={}){this.raws={},this[on]=!1,this[vx]=!0;for(let t in e)if(t==="nodes"){this.nodes=[];for(let i of e[t])typeof i.clone=="function"?this.append(i.clone()):this.append(i)}else this[t]=e[t]}error(e,t={}){if(this.source){let{start:i,end:n}=this.rangeBy(t);return this.source.input.error(e,{line:i.line,column:i.column},{line:n.line,column:n.column},t)}return new xx(e)}warn(e,t,i){let n={node:this};for(let a in i)n[a]=i[a];return e.warn(t,n)}remove(){return this.parent&&this.parent.removeChild(this),this.parent=void 0,this}toString(e=Sx){e.stringify&&(e=e.stringify);let t="";return e(this,i=>{t+=i}),t}assign(e={}){for(let t in e)this[t]=e[t];return this}clone(e={}){let t=ca(this);for(let i in e)t[i]=e[i];return t}cloneBefore(e={}){let t=this.clone(e);return this.parent.insertBefore(this,t),t}cloneAfter(e={}){let t=this.clone(e);return this.parent.insertAfter(this,t),t}replaceWith(...e){if(this.parent){let t=this,i=!1;for(let n of e)n===this?i=!0:i?(this.parent.insertAfter(t,n),t=n):this.parent.insertBefore(t,n);i||this.remove()}return this}next(){if(!this.parent)return;let e=this.parent.index(this);return this.parent.nodes[e+1]}prev(){if(!this.parent)return;let e=this.parent.index(this);return this.parent.nodes[e-1]}before(e){return this.parent.insertBefore(this,e),this}after(e){return this.parent.insertAfter(this,e),this}root(){let e=this;for(;e.parent&&e.parent.type!=="document";)e=e.parent;return e}raw(e,t){return new kx().raw(this,e,t)}cleanRaws(e){delete this.raws.before,delete this.raws.after,e||delete this.raws.between}toJSON(e,t){let i={},n=t==null;t=t||new Map;let a=0;for(let s in this){if(!Object.prototype.hasOwnProperty.call(this,s)||s==="parent"||s==="proxyCache")continue;let o=this[s];if(Array.isArray(o))i[s]=o.map(l=>typeof l=="object"&&l.toJSON?l.toJSON(null,t):l);else if(typeof o=="object"&&o.toJSON)i[s]=o.toJSON(null,t);else if(s==="source"){let l=t.get(o.input);l==null&&(l=a,t.set(o.input,a),a++),i[s]={inputId:l,start:o.start,end:o.end}}else i[s]=o}return n&&(i.inputs=[...t.keys()].map(s=>s.toJSON())),i}positionInside(e){let t=this.toString(),i=this.source.start.column,n=this.source.start.line;for(let a=0;ae.root().toProxy():e[t]}}}toProxy(){return this.proxyCache||(this.proxyCache=new Proxy(this,this.getProxyProcessor())),this.proxyCache}addToError(e){if(e.postcssNode=this,e.stack&&this.source&&/\n\s{4}at /.test(e.stack)){let t=this.source;e.stack=e.stack.replace(/\n\s{4}at /,`$&${t.input.from}:${t.start.line}:${t.start.column}$&`)}return e}markDirty(){if(this[on]){this[on]=!1;let e=this;for(;e=e.parent;)e[on]=!1}}get proxyOf(){return this}};dc.exports=ln;ln.default=ln});var Wr=x((S3,hc)=>{u();"use strict";var Ax=Hr(),un=class extends Ax{constructor(e){e&&typeof e.value!="undefined"&&typeof e.value!="string"&&(e={...e,value:String(e.value)});super(e);this.type="decl"}get variable(){return this.prop.startsWith("--")||this.prop[0]==="$"}};hc.exports=un;un.default=un});var pa=x((A3,mc)=>{u();mc.exports=function(r,e){return{generate:()=>{let t="";return r(e,i=>{t+=i}),[t]}}}});var Gr=x((C3,gc)=>{u();"use strict";var Cx=Hr(),fn=class extends Cx{constructor(e){super(e);this.type="comment"}};gc.exports=fn;fn.default=fn});var Et=x((_3,Cc)=>{u();"use strict";var{isClean:yc,my:bc}=sn(),wc=Wr(),vc=Gr(),_x=Hr(),xc,da,ha,kc;function Sc(r){return r.map(e=>(e.nodes&&(e.nodes=Sc(e.nodes)),delete e.source,e))}function Ac(r){if(r[yc]=!1,r.proxyOf.nodes)for(let e of r.proxyOf.nodes)Ac(e)}var Fe=class extends _x{push(e){return e.parent=this,this.proxyOf.nodes.push(e),this}each(e){if(!this.proxyOf.nodes)return;let t=this.getIterator(),i,n;for(;this.indexes[t]{let n;try{n=e(t,i)}catch(a){throw t.addToError(a)}return n!==!1&&t.walk&&(n=t.walk(e)),n})}walkDecls(e,t){return t?e instanceof RegExp?this.walk((i,n)=>{if(i.type==="decl"&&e.test(i.prop))return t(i,n)}):this.walk((i,n)=>{if(i.type==="decl"&&i.prop===e)return t(i,n)}):(t=e,this.walk((i,n)=>{if(i.type==="decl")return t(i,n)}))}walkRules(e,t){return t?e instanceof RegExp?this.walk((i,n)=>{if(i.type==="rule"&&e.test(i.selector))return t(i,n)}):this.walk((i,n)=>{if(i.type==="rule"&&i.selector===e)return t(i,n)}):(t=e,this.walk((i,n)=>{if(i.type==="rule")return t(i,n)}))}walkAtRules(e,t){return t?e instanceof RegExp?this.walk((i,n)=>{if(i.type==="atrule"&&e.test(i.name))return t(i,n)}):this.walk((i,n)=>{if(i.type==="atrule"&&i.name===e)return t(i,n)}):(t=e,this.walk((i,n)=>{if(i.type==="atrule")return t(i,n)}))}walkComments(e){return this.walk((t,i)=>{if(t.type==="comment")return e(t,i)})}append(...e){for(let t of e){let i=this.normalize(t,this.last);for(let n of i)this.proxyOf.nodes.push(n)}return this.markDirty(),this}prepend(...e){e=e.reverse();for(let t of e){let i=this.normalize(t,this.first,"prepend").reverse();for(let n of i)this.proxyOf.nodes.unshift(n);for(let n in this.indexes)this.indexes[n]=this.indexes[n]+i.length}return this.markDirty(),this}cleanRaws(e){if(super.cleanRaws(e),this.nodes)for(let t of this.nodes)t.cleanRaws(e)}insertBefore(e,t){let i=this.index(e),n=i===0?"prepend":!1,a=this.normalize(t,this.proxyOf.nodes[i],n).reverse();i=this.index(e);for(let o of a)this.proxyOf.nodes.splice(i,0,o);let s;for(let o in this.indexes)s=this.indexes[o],i<=s&&(this.indexes[o]=s+a.length);return this.markDirty(),this}insertAfter(e,t){let i=this.index(e),n=this.normalize(t,this.proxyOf.nodes[i]).reverse();i=this.index(e);for(let s of n)this.proxyOf.nodes.splice(i+1,0,s);let a;for(let s in this.indexes)a=this.indexes[s],i=e&&(this.indexes[i]=t-1);return this.markDirty(),this}removeAll(){for(let e of this.proxyOf.nodes)e.parent=void 0;return this.proxyOf.nodes=[],this.markDirty(),this}replaceValues(e,t,i){return i||(i=t,t={}),this.walkDecls(n=>{t.props&&!t.props.includes(n.prop)||t.fast&&!n.value.includes(t.fast)||(n.value=n.value.replace(e,i))}),this.markDirty(),this}every(e){return this.nodes.every(e)}some(e){return this.nodes.some(e)}index(e){return typeof e=="number"?e:(e.proxyOf&&(e=e.proxyOf),this.proxyOf.nodes.indexOf(e))}get first(){if(!!this.proxyOf.nodes)return this.proxyOf.nodes[0]}get last(){if(!!this.proxyOf.nodes)return this.proxyOf.nodes[this.proxyOf.nodes.length-1]}normalize(e,t){if(typeof e=="string")e=Sc(xc(e).nodes);else if(Array.isArray(e)){e=e.slice(0);for(let n of e)n.parent&&n.parent.removeChild(n,"ignore")}else if(e.type==="root"&&this.type!=="document"){e=e.nodes.slice(0);for(let n of e)n.parent&&n.parent.removeChild(n,"ignore")}else if(e.type)e=[e];else if(e.prop){if(typeof e.value=="undefined")throw new Error("Value field is missed in node creation");typeof e.value!="string"&&(e.value=String(e.value)),e=[new wc(e)]}else if(e.selector)e=[new da(e)];else if(e.name)e=[new ha(e)];else if(e.text)e=[new vc(e)];else throw new Error("Unknown node type in node creation");return e.map(n=>(n[bc]||Fe.rebuild(n),n=n.proxyOf,n.parent&&n.parent.removeChild(n),n[yc]&&Ac(n),typeof n.raws.before=="undefined"&&t&&typeof t.raws.before!="undefined"&&(n.raws.before=t.raws.before.replace(/\S/g,"")),n.parent=this.proxyOf,n))}getProxyProcessor(){return{set(e,t,i){return e[t]===i||(e[t]=i,(t==="name"||t==="params"||t==="selector")&&e.markDirty()),!0},get(e,t){return t==="proxyOf"?e:e[t]?t==="each"||typeof t=="string"&&t.startsWith("walk")?(...i)=>e[t](...i.map(n=>typeof n=="function"?(a,s)=>n(a.toProxy(),s):n)):t==="every"||t==="some"?i=>e[t]((n,...a)=>i(n.toProxy(),...a)):t==="root"?()=>e.root().toProxy():t==="nodes"?e.nodes.map(i=>i.toProxy()):t==="first"||t==="last"?e[t].toProxy():e[t]:e[t]}}}getIterator(){this.lastEach||(this.lastEach=0),this.indexes||(this.indexes={}),this.lastEach+=1;let e=this.lastEach;return this.indexes[e]=0,e}};Fe.registerParse=r=>{xc=r};Fe.registerRule=r=>{da=r};Fe.registerAtRule=r=>{ha=r};Fe.registerRoot=r=>{kc=r};Cc.exports=Fe;Fe.default=Fe;Fe.rebuild=r=>{r.type==="atrule"?Object.setPrototypeOf(r,ha.prototype):r.type==="rule"?Object.setPrototypeOf(r,da.prototype):r.type==="decl"?Object.setPrototypeOf(r,wc.prototype):r.type==="comment"?Object.setPrototypeOf(r,vc.prototype):r.type==="root"&&Object.setPrototypeOf(r,kc.prototype),r[bc]=!0,r.nodes&&r.nodes.forEach(e=>{Fe.rebuild(e)})}});var cn=x((E3,Oc)=>{u();"use strict";var Ex=Et(),_c,Ec,er=class extends Ex{constructor(e){super({type:"document",...e});this.nodes||(this.nodes=[])}toResult(e={}){return new _c(new Ec,this,e).stringify()}};er.registerLazyResult=r=>{_c=r};er.registerProcessor=r=>{Ec=r};Oc.exports=er;er.default=er});var ma=x((O3,Rc)=>{u();"use strict";var Tc={};Rc.exports=function(e){Tc[e]||(Tc[e]=!0,typeof console!="undefined"&&console.warn&&console.warn(e))}});var ga=x((T3,Pc)=>{u();"use strict";var pn=class{constructor(e,t={}){if(this.type="warning",this.text=e,t.node&&t.node.source){let i=t.node.rangeBy(t);this.line=i.start.line,this.column=i.start.column,this.endLine=i.end.line,this.endColumn=i.end.column}for(let i in t)this[i]=t[i]}toString(){return this.node?this.node.error(this.text,{plugin:this.plugin,index:this.index,word:this.word}).message:this.plugin?this.plugin+": "+this.text:this.text}};Pc.exports=pn;pn.default=pn});var hn=x((R3,Ic)=>{u();"use strict";var Ox=ga(),dn=class{constructor(e,t,i){this.processor=e,this.messages=[],this.root=t,this.opts=i,this.css=void 0,this.map=void 0}toString(){return this.css}warn(e,t={}){t.plugin||this.lastPlugin&&this.lastPlugin.postcssPlugin&&(t.plugin=this.lastPlugin.postcssPlugin);let i=new Ox(e,t);return this.messages.push(i),i}warnings(){return this.messages.filter(e=>e.type==="warning")}get content(){return this.css}};Ic.exports=dn;dn.default=dn});var Mc=x((P3,Lc)=>{u();"use strict";var ya="'".charCodeAt(0),Dc='"'.charCodeAt(0),mn="\\".charCodeAt(0),qc="/".charCodeAt(0),gn=` -`.charCodeAt(0),Qr=" ".charCodeAt(0),yn="\f".charCodeAt(0),bn=" ".charCodeAt(0),wn="\r".charCodeAt(0),Tx="[".charCodeAt(0),Rx="]".charCodeAt(0),Px="(".charCodeAt(0),Ix=")".charCodeAt(0),Dx="{".charCodeAt(0),qx="}".charCodeAt(0),$x=";".charCodeAt(0),Lx="*".charCodeAt(0),Mx=":".charCodeAt(0),Nx="@".charCodeAt(0),vn=/[\t\n\f\r "#'()/;[\\\]{}]/g,xn=/[\t\n\f\r !"#'():;@[\\\]{}]|\/(?=\*)/g,Bx=/.[\n"'(/\\]/,$c=/[\da-f]/i;Lc.exports=function(e,t={}){let i=e.css.valueOf(),n=t.ignoreErrors,a,s,o,l,c,f,d,p,h,b,v=i.length,y=0,w=[],k=[];function S(){return y}function E(T){throw e.error("Unclosed "+T,y)}function O(){return k.length===0&&y>=v}function B(T){if(k.length)return k.pop();if(y>=v)return;let F=T?T.ignoreUnclosed:!1;switch(a=i.charCodeAt(y),a){case gn:case Qr:case bn:case wn:case yn:{s=y;do s+=1,a=i.charCodeAt(s);while(a===Qr||a===gn||a===bn||a===wn||a===yn);b=["space",i.slice(y,s)],y=s-1;break}case Tx:case Rx:case Dx:case qx:case Mx:case $x:case Ix:{let Y=String.fromCharCode(a);b=[Y,Y,y];break}case Px:{if(p=w.length?w.pop()[1]:"",h=i.charCodeAt(y+1),p==="url"&&h!==ya&&h!==Dc&&h!==Qr&&h!==gn&&h!==bn&&h!==yn&&h!==wn){s=y;do{if(f=!1,s=i.indexOf(")",s+1),s===-1)if(n||F){s=y;break}else E("bracket");for(d=s;i.charCodeAt(d-1)===mn;)d-=1,f=!f}while(f);b=["brackets",i.slice(y,s+1),y,s],y=s}else s=i.indexOf(")",y+1),l=i.slice(y,s+1),s===-1||Bx.test(l)?b=["(","(",y]:(b=["brackets",l,y,s],y=s);break}case ya:case Dc:{o=a===ya?"'":'"',s=y;do{if(f=!1,s=i.indexOf(o,s+1),s===-1)if(n||F){s=y+1;break}else E("string");for(d=s;i.charCodeAt(d-1)===mn;)d-=1,f=!f}while(f);b=["string",i.slice(y,s+1),y,s],y=s;break}case Nx:{vn.lastIndex=y+1,vn.test(i),vn.lastIndex===0?s=i.length-1:s=vn.lastIndex-2,b=["at-word",i.slice(y,s+1),y,s],y=s;break}case mn:{for(s=y,c=!0;i.charCodeAt(s+1)===mn;)s+=1,c=!c;if(a=i.charCodeAt(s+1),c&&a!==qc&&a!==Qr&&a!==gn&&a!==bn&&a!==wn&&a!==yn&&(s+=1,$c.test(i.charAt(s)))){for(;$c.test(i.charAt(s+1));)s+=1;i.charCodeAt(s+1)===Qr&&(s+=1)}b=["word",i.slice(y,s+1),y,s],y=s;break}default:{a===qc&&i.charCodeAt(y+1)===Lx?(s=i.indexOf("*/",y+2)+1,s===0&&(n||F?s=i.length:E("comment")),b=["comment",i.slice(y,s+1),y,s],y=s):(xn.lastIndex=y+1,xn.test(i),xn.lastIndex===0?s=i.length-1:s=xn.lastIndex-2,b=["word",i.slice(y,s+1),y,s],w.push(b),y=s);break}}return y++,b}function N(T){k.push(T)}return{back:N,nextToken:B,endOfFile:O,position:S}}});var kn=x((I3,Bc)=>{u();"use strict";var Nc=Et(),Yr=class extends Nc{constructor(e){super(e);this.type="atrule"}append(...e){return this.proxyOf.nodes||(this.nodes=[]),super.append(...e)}prepend(...e){return this.proxyOf.nodes||(this.nodes=[]),super.prepend(...e)}};Bc.exports=Yr;Yr.default=Yr;Nc.registerAtRule(Yr)});var tr=x((D3,Uc)=>{u();"use strict";var Fc=Et(),jc,zc,Ut=class extends Fc{constructor(e){super(e);this.type="root",this.nodes||(this.nodes=[])}removeChild(e,t){let i=this.index(e);return!t&&i===0&&this.nodes.length>1&&(this.nodes[1].raws.before=this.nodes[i].raws.before),super.removeChild(e)}normalize(e,t,i){let n=super.normalize(e);if(t){if(i==="prepend")this.nodes.length>1?t.raws.before=this.nodes[1].raws.before:delete t.raws.before;else if(this.first!==t)for(let a of n)a.raws.before=t.raws.before}return n}toResult(e={}){return new jc(new zc,this,e).stringify()}};Ut.registerLazyResult=r=>{jc=r};Ut.registerProcessor=r=>{zc=r};Uc.exports=Ut;Ut.default=Ut;Fc.registerRoot(Ut)});var ba=x((q3,Vc)=>{u();"use strict";var Kr={split(r,e,t){let i=[],n="",a=!1,s=0,o=!1,l="",c=!1;for(let f of r)c?c=!1:f==="\\"?c=!0:o?f===l&&(o=!1):f==='"'||f==="'"?(o=!0,l=f):f==="("?s+=1:f===")"?s>0&&(s-=1):s===0&&e.includes(f)&&(a=!0),a?(n!==""&&i.push(n.trim()),n="",a=!1):n+=f;return(t||n!=="")&&i.push(n.trim()),i},space(r){let e=[" ",` -`," "];return Kr.split(r,e)},comma(r){return Kr.split(r,[","],!0)}};Vc.exports=Kr;Kr.default=Kr});var Sn=x(($3,Wc)=>{u();"use strict";var Hc=Et(),Fx=ba(),Xr=class extends Hc{constructor(e){super(e);this.type="rule",this.nodes||(this.nodes=[])}get selectors(){return Fx.comma(this.selector)}set selectors(e){let t=this.selector?this.selector.match(/,\s*/):null,i=t?t[0]:","+this.raw("between","beforeOpen");this.selector=e.join(i)}};Wc.exports=Xr;Xr.default=Xr;Hc.registerRule(Xr)});var Xc=x((L3,Kc)=>{u();"use strict";var jx=Wr(),zx=Mc(),Ux=Gr(),Vx=kn(),Hx=tr(),Gc=Sn(),Qc={empty:!0,space:!0};function Wx(r){for(let e=r.length-1;e>=0;e--){let t=r[e],i=t[3]||t[2];if(i)return i}}var Yc=class{constructor(e){this.input=e,this.root=new Hx,this.current=this.root,this.spaces="",this.semicolon=!1,this.customProperty=!1,this.createTokenizer(),this.root.source={input:e,start:{offset:0,line:1,column:1}}}createTokenizer(){this.tokenizer=zx(this.input)}parse(){let e;for(;!this.tokenizer.endOfFile();)switch(e=this.tokenizer.nextToken(),e[0]){case"space":this.spaces+=e[1];break;case";":this.freeSemicolon(e);break;case"}":this.end(e);break;case"comment":this.comment(e);break;case"at-word":this.atrule(e);break;case"{":this.emptyRule(e);break;default:this.other(e);break}this.endFile()}comment(e){let t=new Ux;this.init(t,e[2]),t.source.end=this.getPosition(e[3]||e[2]);let i=e[1].slice(2,-2);if(/^\s*$/.test(i))t.text="",t.raws.left=i,t.raws.right="";else{let n=i.match(/^(\s*)([^]*\S)(\s*)$/);t.text=n[2],t.raws.left=n[1],t.raws.right=n[3]}}emptyRule(e){let t=new Gc;this.init(t,e[2]),t.selector="",t.raws.between="",this.current=t}other(e){let t=!1,i=null,n=!1,a=null,s=[],o=e[1].startsWith("--"),l=[],c=e;for(;c;){if(i=c[0],l.push(c),i==="("||i==="[")a||(a=c),s.push(i==="("?")":"]");else if(o&&n&&i==="{")a||(a=c),s.push("}");else if(s.length===0)if(i===";")if(n){this.decl(l,o);return}else break;else if(i==="{"){this.rule(l);return}else if(i==="}"){this.tokenizer.back(l.pop()),t=!0;break}else i===":"&&(n=!0);else i===s[s.length-1]&&(s.pop(),s.length===0&&(a=null));c=this.tokenizer.nextToken()}if(this.tokenizer.endOfFile()&&(t=!0),s.length>0&&this.unclosedBracket(a),t&&n){if(!o)for(;l.length&&(c=l[l.length-1][0],!(c!=="space"&&c!=="comment"));)this.tokenizer.back(l.pop());this.decl(l,o)}else this.unknownWord(l)}rule(e){e.pop();let t=new Gc;this.init(t,e[0][2]),t.raws.between=this.spacesAndCommentsFromEnd(e),this.raw(t,"selector",e),this.current=t}decl(e,t){let i=new jx;this.init(i,e[0][2]);let n=e[e.length-1];for(n[0]===";"&&(this.semicolon=!0,e.pop()),i.source.end=this.getPosition(n[3]||n[2]||Wx(e));e[0][0]!=="word";)e.length===1&&this.unknownWord(e),i.raws.before+=e.shift()[1];for(i.source.start=this.getPosition(e[0][2]),i.prop="";e.length;){let c=e[0][0];if(c===":"||c==="space"||c==="comment")break;i.prop+=e.shift()[1]}i.raws.between="";let a;for(;e.length;)if(a=e.shift(),a[0]===":"){i.raws.between+=a[1];break}else a[0]==="word"&&/\w/.test(a[1])&&this.unknownWord([a]),i.raws.between+=a[1];(i.prop[0]==="_"||i.prop[0]==="*")&&(i.raws.before+=i.prop[0],i.prop=i.prop.slice(1));let s=[],o;for(;e.length&&(o=e[0][0],!(o!=="space"&&o!=="comment"));)s.push(e.shift());this.precheckMissedSemicolon(e);for(let c=e.length-1;c>=0;c--){if(a=e[c],a[1].toLowerCase()==="!important"){i.important=!0;let f=this.stringFrom(e,c);f=this.spacesFromEnd(e)+f,f!==" !important"&&(i.raws.important=f);break}else if(a[1].toLowerCase()==="important"){let f=e.slice(0),d="";for(let p=c;p>0;p--){let h=f[p][0];if(d.trim().indexOf("!")===0&&h!=="space")break;d=f.pop()[1]+d}d.trim().indexOf("!")===0&&(i.important=!0,i.raws.important=d,e=f)}if(a[0]!=="space"&&a[0]!=="comment")break}e.some(c=>c[0]!=="space"&&c[0]!=="comment")&&(i.raws.between+=s.map(c=>c[1]).join(""),s=[]),this.raw(i,"value",s.concat(e),t),i.value.includes(":")&&!t&&this.checkMissedSemicolon(e)}atrule(e){let t=new Vx;t.name=e[1].slice(1),t.name===""&&this.unnamedAtrule(t,e),this.init(t,e[2]);let i,n,a,s=!1,o=!1,l=[],c=[];for(;!this.tokenizer.endOfFile();){if(e=this.tokenizer.nextToken(),i=e[0],i==="("||i==="["?c.push(i==="("?")":"]"):i==="{"&&c.length>0?c.push("}"):i===c[c.length-1]&&c.pop(),c.length===0)if(i===";"){t.source.end=this.getPosition(e[2]),this.semicolon=!0;break}else if(i==="{"){o=!0;break}else if(i==="}"){if(l.length>0){for(a=l.length-1,n=l[a];n&&n[0]==="space";)n=l[--a];n&&(t.source.end=this.getPosition(n[3]||n[2]))}this.end(e);break}else l.push(e);else l.push(e);if(this.tokenizer.endOfFile()){s=!0;break}}t.raws.between=this.spacesAndCommentsFromEnd(l),l.length?(t.raws.afterName=this.spacesAndCommentsFromStart(l),this.raw(t,"params",l),s&&(e=l[l.length-1],t.source.end=this.getPosition(e[3]||e[2]),this.spaces=t.raws.between,t.raws.between="")):(t.raws.afterName="",t.params=""),o&&(t.nodes=[],this.current=t)}end(e){this.current.nodes&&this.current.nodes.length&&(this.current.raws.semicolon=this.semicolon),this.semicolon=!1,this.current.raws.after=(this.current.raws.after||"")+this.spaces,this.spaces="",this.current.parent?(this.current.source.end=this.getPosition(e[2]),this.current=this.current.parent):this.unexpectedClose(e)}endFile(){this.current.parent&&this.unclosedBlock(),this.current.nodes&&this.current.nodes.length&&(this.current.raws.semicolon=this.semicolon),this.current.raws.after=(this.current.raws.after||"")+this.spaces}freeSemicolon(e){if(this.spaces+=e[1],this.current.nodes){let t=this.current.nodes[this.current.nodes.length-1];t&&t.type==="rule"&&!t.raws.ownSemicolon&&(t.raws.ownSemicolon=this.spaces,this.spaces="")}}getPosition(e){let t=this.input.fromOffset(e);return{offset:e,line:t.line,column:t.col}}init(e,t){this.current.push(e),e.source={start:this.getPosition(t),input:this.input},e.raws.before=this.spaces,this.spaces="",e.type!=="comment"&&(this.semicolon=!1)}raw(e,t,i,n){let a,s,o=i.length,l="",c=!0,f,d;for(let p=0;ph+b[1],"");e.raws[t]={value:l,raw:p}}e[t]=l}spacesAndCommentsFromEnd(e){let t,i="";for(;e.length&&(t=e[e.length-1][0],!(t!=="space"&&t!=="comment"));)i=e.pop()[1]+i;return i}spacesAndCommentsFromStart(e){let t,i="";for(;e.length&&(t=e[0][0],!(t!=="space"&&t!=="comment"));)i+=e.shift()[1];return i}spacesFromEnd(e){let t,i="";for(;e.length&&(t=e[e.length-1][0],t==="space");)i=e.pop()[1]+i;return i}stringFrom(e,t){let i="";for(let n=t;n=0&&(n=e[a],!(n[0]!=="space"&&(i+=1,i===2)));a--);throw this.input.error("Missed semicolon",n[0]==="word"?n[3]+1:n[2])}};Kc.exports=Yc});var Jc=x(()=>{u()});var ep=x((B3,Zc)=>{u();var Gx="useandom-26T198340PX75pxJACKVERYMINDBUSHWOLF_GQZbfghjklqvwyzrict",Qx=(r,e=21)=>(t=e)=>{let i="",n=t;for(;n--;)i+=r[Math.random()*r.length|0];return i},Yx=(r=21)=>{let e="",t=r;for(;t--;)e+=Gx[Math.random()*64|0];return e};Zc.exports={nanoid:Yx,customAlphabet:Qx}});var wa=x((F3,tp)=>{u();tp.exports={}});var Cn=x((j3,sp)=>{u();"use strict";var{SourceMapConsumer:Kx,SourceMapGenerator:Xx}=Jc(),{fileURLToPath:rp,pathToFileURL:An}=(aa(),ac),{resolve:va,isAbsolute:xa}=(et(),Ur),{nanoid:Jx}=ep(),ka=oa(),ip=nn(),Zx=wa(),Sa=Symbol("fromOffsetCache"),e1=Boolean(Kx&&Xx),np=Boolean(va&&xa),Jr=class{constructor(e,t={}){if(e===null||typeof e=="undefined"||typeof e=="object"&&!e.toString)throw new Error(`PostCSS received ${e} instead of CSS string`);if(this.css=e.toString(),this.css[0]==="\uFEFF"||this.css[0]==="\uFFFE"?(this.hasBOM=!0,this.css=this.css.slice(1)):this.hasBOM=!1,t.from&&(!np||/^\w+:\/\//.test(t.from)||xa(t.from)?this.file=t.from:this.file=va(t.from)),np&&e1){let i=new Zx(this.css,t);if(i.text){this.map=i;let n=i.consumer().file;!this.file&&n&&(this.file=this.mapResolve(n))}}this.file||(this.id=""),this.map&&(this.map.file=this.from)}fromOffset(e){let t,i;if(this[Sa])i=this[Sa];else{let a=this.css.split(` -`);i=new Array(a.length);let s=0;for(let o=0,l=a.length;o=t)n=i.length-1;else{let a=i.length-2,s;for(;n>1),e=i[s+1])n=s+1;else{n=s;break}}return{line:n+1,col:e-i[n]+1}}error(e,t,i,n={}){let a,s,o;if(t&&typeof t=="object"){let c=t,f=i;if(typeof c.offset=="number"){let d=this.fromOffset(c.offset);t=d.line,i=d.col}else t=c.line,i=c.column;if(typeof f.offset=="number"){let d=this.fromOffset(f.offset);s=d.line,o=d.col}else s=f.line,o=f.column}else if(!i){let c=this.fromOffset(t);t=c.line,i=c.col}let l=this.origin(t,i,s,o);return l?a=new ip(e,l.endLine===void 0?l.line:{line:l.line,column:l.column},l.endLine===void 0?l.column:{line:l.endLine,column:l.endColumn},l.source,l.file,n.plugin):a=new ip(e,s===void 0?t:{line:t,column:i},s===void 0?i:{line:s,column:o},this.css,this.file,n.plugin),a.input={line:t,column:i,endLine:s,endColumn:o,source:this.css},this.file&&(An&&(a.input.url=An(this.file).toString()),a.input.file=this.file),a}origin(e,t,i,n){if(!this.map)return!1;let a=this.map.consumer(),s=a.originalPositionFor({line:e,column:t});if(!s.source)return!1;let o;typeof i=="number"&&(o=a.originalPositionFor({line:i,column:n}));let l;xa(s.source)?l=An(s.source):l=new URL(s.source,this.map.consumer().sourceRoot||An(this.map.mapFile));let c={url:l.toString(),line:s.line,column:s.column,endLine:o&&o.line,endColumn:o&&o.column};if(l.protocol==="file:")if(rp)c.file=rp(l);else throw new Error("file: protocol is not available in this PostCSS build");let f=a.sourceContentFor(s.source);return f&&(c.source=f),c}mapResolve(e){return/^\w+:\/\//.test(e)?e:va(this.map.consumer().sourceRoot||this.map.root||".",e)}get from(){return this.file||this.id}toJSON(){let e={};for(let t of["hasBOM","css","file","id"])this[t]!=null&&(e[t]=this[t]);return this.map&&(e.map={...this.map},e.map.consumerCache&&(e.map.consumerCache=void 0)),e}};sp.exports=Jr;Jr.default=Jr;ka&&ka.registerInput&&ka.registerInput(Jr)});var En=x((z3,ap)=>{u();"use strict";var t1=Et(),r1=Xc(),i1=Cn();function _n(r,e){let t=new i1(r,e),i=new r1(t);try{i.parse()}catch(n){throw n}return i.root}ap.exports=_n;_n.default=_n;t1.registerParse(_n)});var _a=x((V3,fp)=>{u();"use strict";var{isClean:tt,my:n1}=sn(),s1=pa(),a1=Vr(),o1=Et(),l1=cn(),U3=ma(),op=hn(),u1=En(),f1=tr(),c1={document:"Document",root:"Root",atrule:"AtRule",rule:"Rule",decl:"Declaration",comment:"Comment"},p1={postcssPlugin:!0,prepare:!0,Once:!0,Document:!0,Root:!0,Declaration:!0,Rule:!0,AtRule:!0,Comment:!0,DeclarationExit:!0,RuleExit:!0,AtRuleExit:!0,CommentExit:!0,RootExit:!0,DocumentExit:!0,OnceExit:!0},d1={postcssPlugin:!0,prepare:!0,Once:!0},rr=0;function Zr(r){return typeof r=="object"&&typeof r.then=="function"}function lp(r){let e=!1,t=c1[r.type];return r.type==="decl"?e=r.prop.toLowerCase():r.type==="atrule"&&(e=r.name.toLowerCase()),e&&r.append?[t,t+"-"+e,rr,t+"Exit",t+"Exit-"+e]:e?[t,t+"-"+e,t+"Exit",t+"Exit-"+e]:r.append?[t,rr,t+"Exit"]:[t,t+"Exit"]}function up(r){let e;return r.type==="document"?e=["Document",rr,"DocumentExit"]:r.type==="root"?e=["Root",rr,"RootExit"]:e=lp(r),{node:r,events:e,eventIndex:0,visitors:[],visitorIndex:0,iterator:0}}function Aa(r){return r[tt]=!1,r.nodes&&r.nodes.forEach(e=>Aa(e)),r}var Ca={},pt=class{constructor(e,t,i){this.stringified=!1,this.processed=!1;let n;if(typeof t=="object"&&t!==null&&(t.type==="root"||t.type==="document"))n=Aa(t);else if(t instanceof pt||t instanceof op)n=Aa(t.root),t.map&&(typeof i.map=="undefined"&&(i.map={}),i.map.inline||(i.map.inline=!1),i.map.prev=t.map);else{let a=u1;i.syntax&&(a=i.syntax.parse),i.parser&&(a=i.parser),a.parse&&(a=a.parse);try{n=a(t,i)}catch(s){this.processed=!0,this.error=s}n&&!n[n1]&&o1.rebuild(n)}this.result=new op(e,n,i),this.helpers={...Ca,result:this.result,postcss:Ca},this.plugins=this.processor.plugins.map(a=>typeof a=="object"&&a.prepare?{...a,...a.prepare(this.result)}:a)}get[Symbol.toStringTag](){return"LazyResult"}get processor(){return this.result.processor}get opts(){return this.result.opts}get css(){return this.stringify().css}get content(){return this.stringify().content}get map(){return this.stringify().map}get root(){return this.sync().root}get messages(){return this.sync().messages}warnings(){return this.sync().warnings()}toString(){return this.css}then(e,t){return this.async().then(e,t)}catch(e){return this.async().catch(e)}finally(e){return this.async().then(e,e)}async(){return this.error?Promise.reject(this.error):this.processed?Promise.resolve(this.result):(this.processing||(this.processing=this.runAsync()),this.processing)}sync(){if(this.error)throw this.error;if(this.processed)return this.result;if(this.processed=!0,this.processing)throw this.getAsyncError();for(let e of this.plugins){let t=this.runOnRoot(e);if(Zr(t))throw this.getAsyncError()}if(this.prepareVisitors(),this.hasListener){let e=this.result.root;for(;!e[tt];)e[tt]=!0,this.walkSync(e);if(this.listeners.OnceExit)if(e.type==="document")for(let t of e.nodes)this.visitSync(this.listeners.OnceExit,t);else this.visitSync(this.listeners.OnceExit,e)}return this.result}stringify(){if(this.error)throw this.error;if(this.stringified)return this.result;this.stringified=!0,this.sync();let e=this.result.opts,t=a1;e.syntax&&(t=e.syntax.stringify),e.stringifier&&(t=e.stringifier),t.stringify&&(t=t.stringify);let n=new s1(t,this.result.root,this.result.opts).generate();return this.result.css=n[0],this.result.map=n[1],this.result}walkSync(e){e[tt]=!0;let t=lp(e);for(let i of t)if(i===rr)e.nodes&&e.each(n=>{n[tt]||this.walkSync(n)});else{let n=this.listeners[i];if(n&&this.visitSync(n,e.toProxy()))return}}visitSync(e,t){for(let[i,n]of e){this.result.lastPlugin=i;let a;try{a=n(t,this.helpers)}catch(s){throw this.handleError(s,t.proxyOf)}if(t.type!=="root"&&t.type!=="document"&&!t.parent)return!0;if(Zr(a))throw this.getAsyncError()}}runOnRoot(e){this.result.lastPlugin=e;try{if(typeof e=="object"&&e.Once){if(this.result.root.type==="document"){let t=this.result.root.nodes.map(i=>e.Once(i,this.helpers));return Zr(t[0])?Promise.all(t):t}return e.Once(this.result.root,this.helpers)}else if(typeof e=="function")return e(this.result.root,this.result)}catch(t){throw this.handleError(t)}}getAsyncError(){throw new Error("Use process(css).then(cb) to work with async plugins")}handleError(e,t){let i=this.result.lastPlugin;try{t&&t.addToError(e),this.error=e,e.name==="CssSyntaxError"&&!e.plugin?(e.plugin=i.postcssPlugin,e.setMessage()):i.postcssVersion}catch(n){console&&console.error&&console.error(n)}return e}async runAsync(){this.plugin=0;for(let e=0;e0;){let i=this.visitTick(t);if(Zr(i))try{await i}catch(n){let a=t[t.length-1].node;throw this.handleError(n,a)}}}if(this.listeners.OnceExit)for(let[t,i]of this.listeners.OnceExit){this.result.lastPlugin=t;try{if(e.type==="document"){let n=e.nodes.map(a=>i(a,this.helpers));await Promise.all(n)}else await i(e,this.helpers)}catch(n){throw this.handleError(n)}}}return this.processed=!0,this.stringify()}prepareVisitors(){this.listeners={};let e=(t,i,n)=>{this.listeners[i]||(this.listeners[i]=[]),this.listeners[i].push([t,n])};for(let t of this.plugins)if(typeof t=="object")for(let i in t){if(!p1[i]&&/^[A-Z]/.test(i))throw new Error(`Unknown event ${i} in ${t.postcssPlugin}. Try to update PostCSS (${this.processor.version} now).`);if(!d1[i])if(typeof t[i]=="object")for(let n in t[i])n==="*"?e(t,i,t[i][n]):e(t,i+"-"+n.toLowerCase(),t[i][n]);else typeof t[i]=="function"&&e(t,i,t[i])}this.hasListener=Object.keys(this.listeners).length>0}visitTick(e){let t=e[e.length-1],{node:i,visitors:n}=t;if(i.type!=="root"&&i.type!=="document"&&!i.parent){e.pop();return}if(n.length>0&&t.visitorIndex{Ca=r};fp.exports=pt;pt.default=pt;f1.registerLazyResult(pt);l1.registerLazyResult(pt)});var pp=x((W3,cp)=>{u();"use strict";var h1=pa(),m1=Vr(),H3=ma(),g1=En(),y1=hn(),On=class{constructor(e,t,i){t=t.toString(),this.stringified=!1,this._processor=e,this._css=t,this._opts=i,this._map=void 0;let n,a=m1;this.result=new y1(this._processor,n,this._opts),this.result.css=t;let s=this;Object.defineProperty(this.result,"root",{get(){return s.root}});let o=new h1(a,n,this._opts,t);if(o.isMap()){let[l,c]=o.generate();l&&(this.result.css=l),c&&(this.result.map=c)}}get[Symbol.toStringTag](){return"NoWorkResult"}get processor(){return this.result.processor}get opts(){return this.result.opts}get css(){return this.result.css}get content(){return this.result.css}get map(){return this.result.map}get root(){if(this._root)return this._root;let e,t=g1;try{e=t(this._css,this._opts)}catch(i){this.error=i}if(this.error)throw this.error;return this._root=e,e}get messages(){return[]}warnings(){return[]}toString(){return this._css}then(e,t){return this.async().then(e,t)}catch(e){return this.async().catch(e)}finally(e){return this.async().then(e,e)}async(){return this.error?Promise.reject(this.error):Promise.resolve(this.result)}sync(){if(this.error)throw this.error;return this.result}};cp.exports=On;On.default=On});var hp=x((G3,dp)=>{u();"use strict";var b1=pp(),w1=_a(),v1=cn(),x1=tr(),ir=class{constructor(e=[]){this.version="8.4.24",this.plugins=this.normalize(e)}use(e){return this.plugins=this.plugins.concat(this.normalize([e])),this}process(e,t={}){return this.plugins.length===0&&typeof t.parser=="undefined"&&typeof t.stringifier=="undefined"&&typeof t.syntax=="undefined"?new b1(this,e,t):new w1(this,e,t)}normalize(e){let t=[];for(let i of e)if(i.postcss===!0?i=i():i.postcss&&(i=i.postcss),typeof i=="object"&&Array.isArray(i.plugins))t=t.concat(i.plugins);else if(typeof i=="object"&&i.postcssPlugin)t.push(i);else if(typeof i=="function")t.push(i);else if(!(typeof i=="object"&&(i.parse||i.stringify)))throw new Error(i+" is not a PostCSS plugin");return t}};dp.exports=ir;ir.default=ir;x1.registerProcessor(ir);v1.registerProcessor(ir)});var gp=x((Q3,mp)=>{u();"use strict";var k1=Wr(),S1=wa(),A1=Gr(),C1=kn(),_1=Cn(),E1=tr(),O1=Sn();function ei(r,e){if(Array.isArray(r))return r.map(n=>ei(n));let{inputs:t,...i}=r;if(t){e=[];for(let n of t){let a={...n,__proto__:_1.prototype};a.map&&(a.map={...a.map,__proto__:S1.prototype}),e.push(a)}}if(i.nodes&&(i.nodes=r.nodes.map(n=>ei(n,e))),i.source){let{inputId:n,...a}=i.source;i.source=a,n!=null&&(i.source.input=e[n])}if(i.type==="root")return new E1(i);if(i.type==="decl")return new k1(i);if(i.type==="rule")return new O1(i);if(i.type==="comment")return new A1(i);if(i.type==="atrule")return new C1(i);throw new Error("Unknown node type: "+r.type)}mp.exports=ei;ei.default=ei});var $e=x((Y3,Sp)=>{u();"use strict";var T1=nn(),yp=Wr(),R1=_a(),P1=Et(),Ea=hp(),I1=Vr(),D1=gp(),bp=cn(),q1=ga(),wp=Gr(),vp=kn(),$1=hn(),L1=Cn(),M1=En(),N1=ba(),xp=Sn(),kp=tr(),B1=Hr();function Z(...r){return r.length===1&&Array.isArray(r[0])&&(r=r[0]),new Ea(r)}Z.plugin=function(e,t){let i=!1;function n(...s){console&&console.warn&&!i&&(i=!0,console.warn(e+`: postcss.plugin was deprecated. Migration guide: -https://evilmartians.com/chronicles/postcss-8-plugin-migration`),m.env.LANG&&m.env.LANG.startsWith("cn")&&console.warn(e+`: \u91CC\u9762 postcss.plugin \u88AB\u5F03\u7528. \u8FC1\u79FB\u6307\u5357: -https://www.w3ctech.com/topic/2226`));let o=t(...s);return o.postcssPlugin=e,o.postcssVersion=new Ea().version,o}let a;return Object.defineProperty(n,"postcss",{get(){return a||(a=n()),a}}),n.process=function(s,o,l){return Z([n(l)]).process(s,o)},n};Z.stringify=I1;Z.parse=M1;Z.fromJSON=D1;Z.list=N1;Z.comment=r=>new wp(r);Z.atRule=r=>new vp(r);Z.decl=r=>new yp(r);Z.rule=r=>new xp(r);Z.root=r=>new kp(r);Z.document=r=>new bp(r);Z.CssSyntaxError=T1;Z.Declaration=yp;Z.Container=P1;Z.Processor=Ea;Z.Document=bp;Z.Comment=wp;Z.Warning=q1;Z.AtRule=vp;Z.Result=$1;Z.Input=L1;Z.Rule=xp;Z.Root=kp;Z.Node=B1;R1.registerPostcss(Z);Sp.exports=Z;Z.default=Z});var re,ee,K3,X3,J3,Z3,eI,tI,rI,iI,nI,sI,aI,oI,lI,uI,fI,cI,pI,dI,hI,mI,gI,yI,bI,wI,Ot=R(()=>{u();re=pe($e()),ee=re.default,K3=re.default.stringify,X3=re.default.fromJSON,J3=re.default.plugin,Z3=re.default.parse,eI=re.default.list,tI=re.default.document,rI=re.default.comment,iI=re.default.atRule,nI=re.default.rule,sI=re.default.decl,aI=re.default.root,oI=re.default.CssSyntaxError,lI=re.default.Declaration,uI=re.default.Container,fI=re.default.Processor,cI=re.default.Document,pI=re.default.Comment,dI=re.default.Warning,hI=re.default.AtRule,mI=re.default.Result,gI=re.default.Input,yI=re.default.Rule,bI=re.default.Root,wI=re.default.Node});var Oa=x((xI,Ap)=>{u();Ap.exports=function(r,e,t,i,n){for(e=e.split?e.split("."):e,i=0;i{u();"use strict";Tn.__esModule=!0;Tn.default=z1;function F1(r){for(var e=r.toLowerCase(),t="",i=!1,n=0;n<6&&e[n]!==void 0;n++){var a=e.charCodeAt(n),s=a>=97&&a<=102||a>=48&&a<=57;if(i=a===32,!s)break;t+=e[n]}if(t.length!==0){var o=parseInt(t,16),l=o>=55296&&o<=57343;return l||o===0||o>1114111?["\uFFFD",t.length+(i?1:0)]:[String.fromCodePoint(o),t.length+(i?1:0)]}}var j1=/\\/;function z1(r){var e=j1.test(r);if(!e)return r;for(var t="",i=0;i{u();"use strict";Pn.__esModule=!0;Pn.default=U1;function U1(r){for(var e=arguments.length,t=new Array(e>1?e-1:0),i=1;i0;){var n=t.shift();if(!r[n])return;r=r[n]}return r}_p.exports=Pn.default});var Tp=x((In,Op)=>{u();"use strict";In.__esModule=!0;In.default=V1;function V1(r){for(var e=arguments.length,t=new Array(e>1?e-1:0),i=1;i0;){var n=t.shift();r[n]||(r[n]={}),r=r[n]}}Op.exports=In.default});var Pp=x((Dn,Rp)=>{u();"use strict";Dn.__esModule=!0;Dn.default=H1;function H1(r){for(var e="",t=r.indexOf("/*"),i=0;t>=0;){e=e+r.slice(i,t);var n=r.indexOf("*/",t+2);if(n<0)return e;i=n+2,t=r.indexOf("/*",i)}return e=e+r.slice(i),e}Rp.exports=Dn.default});var ti=x(rt=>{u();"use strict";rt.__esModule=!0;rt.unesc=rt.stripComments=rt.getProp=rt.ensureObject=void 0;var W1=qn(Rn());rt.unesc=W1.default;var G1=qn(Ep());rt.getProp=G1.default;var Q1=qn(Tp());rt.ensureObject=Q1.default;var Y1=qn(Pp());rt.stripComments=Y1.default;function qn(r){return r&&r.__esModule?r:{default:r}}});var dt=x((ri,qp)=>{u();"use strict";ri.__esModule=!0;ri.default=void 0;var Ip=ti();function Dp(r,e){for(var t=0;ti||this.source.end.linen||this.source.end.line===i&&this.source.end.column{u();"use strict";ie.__esModule=!0;ie.UNIVERSAL=ie.TAG=ie.STRING=ie.SELECTOR=ie.ROOT=ie.PSEUDO=ie.NESTING=ie.ID=ie.COMMENT=ie.COMBINATOR=ie.CLASS=ie.ATTRIBUTE=void 0;var Z1="tag";ie.TAG=Z1;var ek="string";ie.STRING=ek;var tk="selector";ie.SELECTOR=tk;var rk="root";ie.ROOT=rk;var ik="pseudo";ie.PSEUDO=ik;var nk="nesting";ie.NESTING=nk;var sk="id";ie.ID=sk;var ak="comment";ie.COMMENT=ak;var ok="combinator";ie.COMBINATOR=ok;var lk="class";ie.CLASS=lk;var uk="attribute";ie.ATTRIBUTE=uk;var fk="universal";ie.UNIVERSAL=fk});var $n=x((ii,Np)=>{u();"use strict";ii.__esModule=!0;ii.default=void 0;var ck=dk(dt()),ht=pk(Se());function $p(r){if(typeof WeakMap!="function")return null;var e=new WeakMap,t=new WeakMap;return($p=function(n){return n?t:e})(r)}function pk(r,e){if(!e&&r&&r.__esModule)return r;if(r===null||typeof r!="object"&&typeof r!="function")return{default:r};var t=$p(e);if(t&&t.has(r))return t.get(r);var i={},n=Object.defineProperty&&Object.getOwnPropertyDescriptor;for(var a in r)if(a!=="default"&&Object.prototype.hasOwnProperty.call(r,a)){var s=n?Object.getOwnPropertyDescriptor(r,a):null;s&&(s.get||s.set)?Object.defineProperty(i,a,s):i[a]=r[a]}return i.default=r,t&&t.set(r,i),i}function dk(r){return r&&r.__esModule?r:{default:r}}function hk(r,e){var t=typeof Symbol!="undefined"&&r[Symbol.iterator]||r["@@iterator"];if(t)return(t=t.call(r)).next.bind(t);if(Array.isArray(r)||(t=mk(r))||e&&r&&typeof r.length=="number"){t&&(r=t);var i=0;return function(){return i>=r.length?{done:!0}:{done:!1,value:r[i++]}}}throw new TypeError(`Invalid attempt to iterate non-iterable instance. -In order to be iterable, non-array objects must have a [Symbol.iterator]() method.`)}function mk(r,e){if(!!r){if(typeof r=="string")return Lp(r,e);var t=Object.prototype.toString.call(r).slice(8,-1);if(t==="Object"&&r.constructor&&(t=r.constructor.name),t==="Map"||t==="Set")return Array.from(r);if(t==="Arguments"||/^(?:Ui|I)nt(?:8|16|32)(?:Clamped)?Array$/.test(t))return Lp(r,e)}}function Lp(r,e){(e==null||e>r.length)&&(e=r.length);for(var t=0,i=new Array(e);t=n&&(this.indexes[s]=a-1);return this},t.removeAll=function(){for(var n=hk(this.nodes),a;!(a=n()).done;){var s=a.value;s.parent=void 0}return this.nodes=[],this},t.empty=function(){return this.removeAll()},t.insertAfter=function(n,a){a.parent=this;var s=this.index(n);this.nodes.splice(s+1,0,a),a.parent=this;var o;for(var l in this.indexes)o=this.indexes[l],s<=o&&(this.indexes[l]=o+1);return this},t.insertBefore=function(n,a){a.parent=this;var s=this.index(n);this.nodes.splice(s,0,a),a.parent=this;var o;for(var l in this.indexes)o=this.indexes[l],o<=s&&(this.indexes[l]=o+1);return this},t._findChildAtPosition=function(n,a){var s=void 0;return this.each(function(o){if(o.atPosition){var l=o.atPosition(n,a);if(l)return s=l,!1}else if(o.isAtPosition(n,a))return s=o,!1}),s},t.atPosition=function(n,a){if(this.isAtPosition(n,a))return this._findChildAtPosition(n,a)||this},t._inferEndPosition=function(){this.last&&this.last.source&&this.last.source.end&&(this.source=this.source||{},this.source.end=this.source.end||{},Object.assign(this.source.end,this.last.source.end))},t.each=function(n){this.lastEach||(this.lastEach=0),this.indexes||(this.indexes={}),this.lastEach++;var a=this.lastEach;if(this.indexes[a]=0,!!this.length){for(var s,o;this.indexes[a]{u();"use strict";ni.__esModule=!0;ni.default=void 0;var wk=xk($n()),vk=Se();function xk(r){return r&&r.__esModule?r:{default:r}}function Bp(r,e){for(var t=0;t{u();"use strict";si.__esModule=!0;si.default=void 0;var Ck=Ek($n()),_k=Se();function Ek(r){return r&&r.__esModule?r:{default:r}}function Ok(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,Ia(r,e)}function Ia(r,e){return Ia=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},Ia(r,e)}var Tk=function(r){Ok(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=_k.SELECTOR,i}return e}(Ck.default);si.default=Tk;jp.exports=si.default});var Ln=x((AI,zp)=>{u();"use strict";var Rk={},Pk=Rk.hasOwnProperty,Ik=function(e,t){if(!e)return t;var i={};for(var n in t)i[n]=Pk.call(e,n)?e[n]:t[n];return i},Dk=/[ -,\.\/:-@\[-\^`\{-~]/,qk=/[ -,\.\/:-@\[\]\^`\{-~]/,$k=/(^|\\+)?(\\[A-F0-9]{1,6})\x20(?![a-fA-F0-9\x20])/g,qa=function r(e,t){t=Ik(t,r.options),t.quotes!="single"&&t.quotes!="double"&&(t.quotes="single");for(var i=t.quotes=="double"?'"':"'",n=t.isIdentifier,a=e.charAt(0),s="",o=0,l=e.length;o126){if(f>=55296&&f<=56319&&o{u();"use strict";ai.__esModule=!0;ai.default=void 0;var Lk=Up(Ln()),Mk=ti(),Nk=Up(dt()),Bk=Se();function Up(r){return r&&r.__esModule?r:{default:r}}function Vp(r,e){for(var t=0;t{u();"use strict";oi.__esModule=!0;oi.default=void 0;var Uk=Hk(dt()),Vk=Se();function Hk(r){return r&&r.__esModule?r:{default:r}}function Wk(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,Ma(r,e)}function Ma(r,e){return Ma=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},Ma(r,e)}var Gk=function(r){Wk(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=Vk.COMMENT,i}return e}(Uk.default);oi.default=Gk;Wp.exports=oi.default});var Fa=x((li,Gp)=>{u();"use strict";li.__esModule=!0;li.default=void 0;var Qk=Kk(dt()),Yk=Se();function Kk(r){return r&&r.__esModule?r:{default:r}}function Xk(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,Ba(r,e)}function Ba(r,e){return Ba=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},Ba(r,e)}var Jk=function(r){Xk(e,r);function e(i){var n;return n=r.call(this,i)||this,n.type=Yk.ID,n}var t=e.prototype;return t.valueToString=function(){return"#"+r.prototype.valueToString.call(this)},e}(Qk.default);li.default=Jk;Gp.exports=li.default});var Mn=x((ui,Kp)=>{u();"use strict";ui.__esModule=!0;ui.default=void 0;var Zk=Qp(Ln()),eS=ti(),tS=Qp(dt());function Qp(r){return r&&r.__esModule?r:{default:r}}function Yp(r,e){for(var t=0;t{u();"use strict";fi.__esModule=!0;fi.default=void 0;var sS=oS(Mn()),aS=Se();function oS(r){return r&&r.__esModule?r:{default:r}}function lS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,za(r,e)}function za(r,e){return za=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},za(r,e)}var uS=function(r){lS(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=aS.TAG,i}return e}(sS.default);fi.default=uS;Xp.exports=fi.default});var Ha=x((ci,Jp)=>{u();"use strict";ci.__esModule=!0;ci.default=void 0;var fS=pS(dt()),cS=Se();function pS(r){return r&&r.__esModule?r:{default:r}}function dS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,Va(r,e)}function Va(r,e){return Va=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},Va(r,e)}var hS=function(r){dS(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=cS.STRING,i}return e}(fS.default);ci.default=hS;Jp.exports=ci.default});var Ga=x((pi,Zp)=>{u();"use strict";pi.__esModule=!0;pi.default=void 0;var mS=yS($n()),gS=Se();function yS(r){return r&&r.__esModule?r:{default:r}}function bS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,Wa(r,e)}function Wa(r,e){return Wa=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},Wa(r,e)}var wS=function(r){bS(e,r);function e(i){var n;return n=r.call(this,i)||this,n.type=gS.PSEUDO,n}var t=e.prototype;return t.toString=function(){var n=this.length?"("+this.map(String).join(",")+")":"";return[this.rawSpaceBefore,this.stringifyProperty("value"),n,this.rawSpaceAfter].join("")},e}(mS.default);pi.default=wS;Zp.exports=pi.default});var Nn={};Ge(Nn,{deprecate:()=>vS});function vS(r){return r}var Bn=R(()=>{u()});var td=x((CI,ed)=>{u();ed.exports=(Bn(),Nn).deprecate});var Za=x(mi=>{u();"use strict";mi.__esModule=!0;mi.default=void 0;mi.unescapeValue=Xa;var di=Ya(Ln()),xS=Ya(Rn()),kS=Ya(Mn()),SS=Se(),Qa;function Ya(r){return r&&r.__esModule?r:{default:r}}function rd(r,e){for(var t=0;t0&&!n.quoted&&o.before.length===0&&!(n.spaces.value&&n.spaces.value.after)&&(o.before=" "),id(s,o)}))),a.push("]"),a.push(this.rawSpaceAfter),a.join("")},AS(e,[{key:"quoted",get:function(){var n=this.quoteMark;return n==="'"||n==='"'},set:function(n){OS()}},{key:"quoteMark",get:function(){return this._quoteMark},set:function(n){if(!this._constructed){this._quoteMark=n;return}this._quoteMark!==n&&(this._quoteMark=n,this._syncRawValue())}},{key:"qualifiedAttribute",get:function(){return this.qualifiedName(this.raws.attribute||this.attribute)}},{key:"insensitiveFlag",get:function(){return this.insensitive?"i":""}},{key:"value",get:function(){return this._value},set:function(n){if(this._constructed){var a=Xa(n),s=a.deprecatedUsage,o=a.unescaped,l=a.quoteMark;if(s&&ES(),o===this._value&&l===this._quoteMark)return;this._value=o,this._quoteMark=l,this._syncRawValue()}else this._value=n}},{key:"insensitive",get:function(){return this._insensitive},set:function(n){n||(this._insensitive=!1,this.raws&&(this.raws.insensitiveFlag==="I"||this.raws.insensitiveFlag==="i")&&(this.raws.insensitiveFlag=void 0)),this._insensitive=n}},{key:"attribute",get:function(){return this._attribute},set:function(n){this._handleEscapes("attribute",n),this._attribute=n}}]),e}(kS.default);mi.default=Fn;Fn.NO_QUOTE=null;Fn.SINGLE_QUOTE="'";Fn.DOUBLE_QUOTE='"';var Ja=(Qa={"'":{quotes:"single",wrap:!0},'"':{quotes:"double",wrap:!0}},Qa[null]={isIdentifier:!0},Qa);function id(r,e){return""+e.before+r+e.after}});var to=x((gi,nd)=>{u();"use strict";gi.__esModule=!0;gi.default=void 0;var PS=DS(Mn()),IS=Se();function DS(r){return r&&r.__esModule?r:{default:r}}function qS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,eo(r,e)}function eo(r,e){return eo=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},eo(r,e)}var $S=function(r){qS(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=IS.UNIVERSAL,i.value="*",i}return e}(PS.default);gi.default=$S;nd.exports=gi.default});var io=x((yi,sd)=>{u();"use strict";yi.__esModule=!0;yi.default=void 0;var LS=NS(dt()),MS=Se();function NS(r){return r&&r.__esModule?r:{default:r}}function BS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,ro(r,e)}function ro(r,e){return ro=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},ro(r,e)}var FS=function(r){BS(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=MS.COMBINATOR,i}return e}(LS.default);yi.default=FS;sd.exports=yi.default});var so=x((bi,ad)=>{u();"use strict";bi.__esModule=!0;bi.default=void 0;var jS=US(dt()),zS=Se();function US(r){return r&&r.__esModule?r:{default:r}}function VS(r,e){r.prototype=Object.create(e.prototype),r.prototype.constructor=r,no(r,e)}function no(r,e){return no=Object.setPrototypeOf?Object.setPrototypeOf.bind():function(i,n){return i.__proto__=n,i},no(r,e)}var HS=function(r){VS(e,r);function e(t){var i;return i=r.call(this,t)||this,i.type=zS.NESTING,i.value="&",i}return e}(jS.default);bi.default=HS;ad.exports=bi.default});var ld=x((jn,od)=>{u();"use strict";jn.__esModule=!0;jn.default=WS;function WS(r){return r.sort(function(e,t){return e-t})}od.exports=jn.default});var ao=x(M=>{u();"use strict";M.__esModule=!0;M.word=M.tilde=M.tab=M.str=M.space=M.slash=M.singleQuote=M.semicolon=M.plus=M.pipe=M.openSquare=M.openParenthesis=M.newline=M.greaterThan=M.feed=M.equals=M.doubleQuote=M.dollar=M.cr=M.comment=M.comma=M.combinator=M.colon=M.closeSquare=M.closeParenthesis=M.caret=M.bang=M.backslash=M.at=M.asterisk=M.ampersand=void 0;var GS=38;M.ampersand=GS;var QS=42;M.asterisk=QS;var YS=64;M.at=YS;var KS=44;M.comma=KS;var XS=58;M.colon=XS;var JS=59;M.semicolon=JS;var ZS=40;M.openParenthesis=ZS;var eA=41;M.closeParenthesis=eA;var tA=91;M.openSquare=tA;var rA=93;M.closeSquare=rA;var iA=36;M.dollar=iA;var nA=126;M.tilde=nA;var sA=94;M.caret=sA;var aA=43;M.plus=aA;var oA=61;M.equals=oA;var lA=124;M.pipe=lA;var uA=62;M.greaterThan=uA;var fA=32;M.space=fA;var ud=39;M.singleQuote=ud;var cA=34;M.doubleQuote=cA;var pA=47;M.slash=pA;var dA=33;M.bang=dA;var hA=92;M.backslash=hA;var mA=13;M.cr=mA;var gA=12;M.feed=gA;var yA=10;M.newline=yA;var bA=9;M.tab=bA;var wA=ud;M.str=wA;var vA=-1;M.comment=vA;var xA=-2;M.word=xA;var kA=-3;M.combinator=kA});var pd=x(wi=>{u();"use strict";wi.__esModule=!0;wi.FIELDS=void 0;wi.default=TA;var D=SA(ao()),nr,te;function fd(r){if(typeof WeakMap!="function")return null;var e=new WeakMap,t=new WeakMap;return(fd=function(n){return n?t:e})(r)}function SA(r,e){if(!e&&r&&r.__esModule)return r;if(r===null||typeof r!="object"&&typeof r!="function")return{default:r};var t=fd(e);if(t&&t.has(r))return t.get(r);var i={},n=Object.defineProperty&&Object.getOwnPropertyDescriptor;for(var a in r)if(a!=="default"&&Object.prototype.hasOwnProperty.call(r,a)){var s=n?Object.getOwnPropertyDescriptor(r,a):null;s&&(s.get||s.set)?Object.defineProperty(i,a,s):i[a]=r[a]}return i.default=r,t&&t.set(r,i),i}var AA=(nr={},nr[D.tab]=!0,nr[D.newline]=!0,nr[D.cr]=!0,nr[D.feed]=!0,nr),CA=(te={},te[D.space]=!0,te[D.tab]=!0,te[D.newline]=!0,te[D.cr]=!0,te[D.feed]=!0,te[D.ampersand]=!0,te[D.asterisk]=!0,te[D.bang]=!0,te[D.comma]=!0,te[D.colon]=!0,te[D.semicolon]=!0,te[D.openParenthesis]=!0,te[D.closeParenthesis]=!0,te[D.openSquare]=!0,te[D.closeSquare]=!0,te[D.singleQuote]=!0,te[D.doubleQuote]=!0,te[D.plus]=!0,te[D.pipe]=!0,te[D.tilde]=!0,te[D.greaterThan]=!0,te[D.equals]=!0,te[D.dollar]=!0,te[D.caret]=!0,te[D.slash]=!0,te),oo={},cd="0123456789abcdefABCDEF";for(zn=0;zn0?(k=s+v,S=w-y[v].length):(k=s,S=a),O=D.comment,s=k,p=k,d=w-S):c===D.slash?(w=o,O=c,p=s,d=o-a,l=w+1):(w=_A(t,o),O=D.word,p=s,d=w-a),l=w+1;break}e.push([O,s,o-a,p,d,o,l]),S&&(a=S,S=null),o=l}return e}});var vd=x((vi,wd)=>{u();"use strict";vi.__esModule=!0;vi.default=void 0;var RA=je(Pa()),lo=je(Da()),PA=je(La()),dd=je(Na()),IA=je(Fa()),DA=je(Ua()),uo=je(Ha()),qA=je(Ga()),hd=Un(Za()),$A=je(to()),fo=je(io()),LA=je(so()),MA=je(ld()),P=Un(pd()),$=Un(ao()),NA=Un(Se()),le=ti(),Vt,co;function md(r){if(typeof WeakMap!="function")return null;var e=new WeakMap,t=new WeakMap;return(md=function(n){return n?t:e})(r)}function Un(r,e){if(!e&&r&&r.__esModule)return r;if(r===null||typeof r!="object"&&typeof r!="function")return{default:r};var t=md(e);if(t&&t.has(r))return t.get(r);var i={},n=Object.defineProperty&&Object.getOwnPropertyDescriptor;for(var a in r)if(a!=="default"&&Object.prototype.hasOwnProperty.call(r,a)){var s=n?Object.getOwnPropertyDescriptor(r,a):null;s&&(s.get||s.set)?Object.defineProperty(i,a,s):i[a]=r[a]}return i.default=r,t&&t.set(r,i),i}function je(r){return r&&r.__esModule?r:{default:r}}function gd(r,e){for(var t=0;t0){var s=this.current.last;if(s){var o=this.convertWhitespaceNodesToSpace(a),l=o.space,c=o.rawSpace;c!==void 0&&(s.rawSpaceAfter+=c),s.spaces.after+=l}else a.forEach(function(O){return i.newNode(O)})}return}var f=this.currToken,d=void 0;n>this.position&&(d=this.parseWhitespaceEquivalentTokens(n));var p;if(this.isNamedCombinator()?p=this.namedCombinator():this.currToken[P.FIELDS.TYPE]===$.combinator?(p=new fo.default({value:this.content(),source:sr(this.currToken),sourceIndex:this.currToken[P.FIELDS.START_POS]}),this.position++):po[this.currToken[P.FIELDS.TYPE]]||d||this.unexpected(),p){if(d){var h=this.convertWhitespaceNodesToSpace(d),b=h.space,v=h.rawSpace;p.spaces.before=b,p.rawSpaceBefore=v}}else{var y=this.convertWhitespaceNodesToSpace(d,!0),w=y.space,k=y.rawSpace;k||(k=w);var S={},E={spaces:{}};w.endsWith(" ")&&k.endsWith(" ")?(S.before=w.slice(0,w.length-1),E.spaces.before=k.slice(0,k.length-1)):w.startsWith(" ")&&k.startsWith(" ")?(S.after=w.slice(1),E.spaces.after=k.slice(1)):E.value=k,p=new fo.default({value:" ",source:ho(f,this.tokens[this.position-1]),sourceIndex:f[P.FIELDS.START_POS],spaces:S,raws:E})}return this.currToken&&this.currToken[P.FIELDS.TYPE]===$.space&&(p.spaces.after=this.optionalSpace(this.content()),this.position++),this.newNode(p)},e.comma=function(){if(this.position===this.tokens.length-1){this.root.trailingComma=!0,this.position++;return}this.current._inferEndPosition();var i=new lo.default({source:{start:yd(this.tokens[this.position+1])}});this.current.parent.append(i),this.current=i,this.position++},e.comment=function(){var i=this.currToken;this.newNode(new dd.default({value:this.content(),source:sr(i),sourceIndex:i[P.FIELDS.START_POS]})),this.position++},e.error=function(i,n){throw this.root.error(i,n)},e.missingBackslash=function(){return this.error("Expected a backslash preceding the semicolon.",{index:this.currToken[P.FIELDS.START_POS]})},e.missingParenthesis=function(){return this.expected("opening parenthesis",this.currToken[P.FIELDS.START_POS])},e.missingSquareBracket=function(){return this.expected("opening square bracket",this.currToken[P.FIELDS.START_POS])},e.unexpected=function(){return this.error("Unexpected '"+this.content()+"'. Escaping special characters with \\ may help.",this.currToken[P.FIELDS.START_POS])},e.unexpectedPipe=function(){return this.error("Unexpected '|'.",this.currToken[P.FIELDS.START_POS])},e.namespace=function(){var i=this.prevToken&&this.content(this.prevToken)||!0;if(this.nextToken[P.FIELDS.TYPE]===$.word)return this.position++,this.word(i);if(this.nextToken[P.FIELDS.TYPE]===$.asterisk)return this.position++,this.universal(i);this.unexpectedPipe()},e.nesting=function(){if(this.nextToken){var i=this.content(this.nextToken);if(i==="|"){this.position++;return}}var n=this.currToken;this.newNode(new LA.default({value:this.content(),source:sr(n),sourceIndex:n[P.FIELDS.START_POS]})),this.position++},e.parentheses=function(){var i=this.current.last,n=1;if(this.position++,i&&i.type===NA.PSEUDO){var a=new lo.default({source:{start:yd(this.tokens[this.position-1])}}),s=this.current;for(i.append(a),this.current=a;this.position1&&i.nextToken&&i.nextToken[P.FIELDS.TYPE]===$.openParenthesis&&i.error("Misplaced parenthesis.",{index:i.nextToken[P.FIELDS.START_POS]})});else return this.expected(["pseudo-class","pseudo-element"],this.currToken[P.FIELDS.START_POS])},e.space=function(){var i=this.content();this.position===0||this.prevToken[P.FIELDS.TYPE]===$.comma||this.prevToken[P.FIELDS.TYPE]===$.openParenthesis||this.current.nodes.every(function(n){return n.type==="comment"})?(this.spaces=this.optionalSpace(i),this.position++):this.position===this.tokens.length-1||this.nextToken[P.FIELDS.TYPE]===$.comma||this.nextToken[P.FIELDS.TYPE]===$.closeParenthesis?(this.current.last.spaces.after=this.optionalSpace(i),this.position++):this.combinator()},e.string=function(){var i=this.currToken;this.newNode(new uo.default({value:this.content(),source:sr(i),sourceIndex:i[P.FIELDS.START_POS]})),this.position++},e.universal=function(i){var n=this.nextToken;if(n&&this.content(n)==="|")return this.position++,this.namespace();var a=this.currToken;this.newNode(new $A.default({value:this.content(),source:sr(a),sourceIndex:a[P.FIELDS.START_POS]}),i),this.position++},e.splitWord=function(i,n){for(var a=this,s=this.nextToken,o=this.content();s&&~[$.dollar,$.caret,$.equals,$.word].indexOf(s[P.FIELDS.TYPE]);){this.position++;var l=this.content();if(o+=l,l.lastIndexOf("\\")===l.length-1){var c=this.nextToken;c&&c[P.FIELDS.TYPE]===$.space&&(o+=this.requiredSpace(this.content(c)),this.position++)}s=this.nextToken}var f=mo(o,".").filter(function(b){var v=o[b-1]==="\\",y=/^\d+\.\d+%$/.test(o);return!v&&!y}),d=mo(o,"#").filter(function(b){return o[b-1]!=="\\"}),p=mo(o,"#{");p.length&&(d=d.filter(function(b){return!~p.indexOf(b)}));var h=(0,MA.default)(jA([0].concat(f,d)));h.forEach(function(b,v){var y=h[v+1]||o.length,w=o.slice(b,y);if(v===0&&n)return n.call(a,w,h.length);var k,S=a.currToken,E=S[P.FIELDS.START_POS]+h[v],O=Ht(S[1],S[2]+b,S[3],S[2]+(y-1));if(~f.indexOf(b)){var B={value:w.slice(1),source:O,sourceIndex:E};k=new PA.default(ar(B,"value"))}else if(~d.indexOf(b)){var N={value:w.slice(1),source:O,sourceIndex:E};k=new IA.default(ar(N,"value"))}else{var T={value:w,source:O,sourceIndex:E};ar(T,"value"),k=new DA.default(T)}a.newNode(k,i),i=null}),this.position++},e.word=function(i){var n=this.nextToken;return n&&this.content(n)==="|"?(this.position++,this.namespace()):this.splitWord(i)},e.loop=function(){for(;this.position{u();"use strict";xi.__esModule=!0;xi.default=void 0;var UA=VA(vd());function VA(r){return r&&r.__esModule?r:{default:r}}var HA=function(){function r(t,i){this.func=t||function(){},this.funcRes=null,this.options=i}var e=r.prototype;return e._shouldUpdateSelector=function(i,n){n===void 0&&(n={});var a=Object.assign({},this.options,n);return a.updateSelector===!1?!1:typeof i!="string"},e._isLossy=function(i){i===void 0&&(i={});var n=Object.assign({},this.options,i);return n.lossless===!1},e._root=function(i,n){n===void 0&&(n={});var a=new UA.default(i,this._parseOptions(n));return a.root},e._parseOptions=function(i){return{lossy:this._isLossy(i)}},e._run=function(i,n){var a=this;return n===void 0&&(n={}),new Promise(function(s,o){try{var l=a._root(i,n);Promise.resolve(a.func(l)).then(function(c){var f=void 0;return a._shouldUpdateSelector(i,n)&&(f=l.toString(),i.selector=f),{transform:c,root:l,string:f}}).then(s,o)}catch(c){o(c);return}})},e._runSync=function(i,n){n===void 0&&(n={});var a=this._root(i,n),s=this.func(a);if(s&&typeof s.then=="function")throw new Error("Selector processor returned a promise to a synchronous call.");var o=void 0;return n.updateSelector&&typeof i!="string"&&(o=a.toString(),i.selector=o),{transform:s,root:a,string:o}},e.ast=function(i,n){return this._run(i,n).then(function(a){return a.root})},e.astSync=function(i,n){return this._runSync(i,n).root},e.transform=function(i,n){return this._run(i,n).then(function(a){return a.transform})},e.transformSync=function(i,n){return this._runSync(i,n).transform},e.process=function(i,n){return this._run(i,n).then(function(a){return a.string||a.root.toString()})},e.processSync=function(i,n){var a=this._runSync(i,n);return a.string||a.root.toString()},r}();xi.default=HA;xd.exports=xi.default});var Sd=x(ne=>{u();"use strict";ne.__esModule=!0;ne.universal=ne.tag=ne.string=ne.selector=ne.root=ne.pseudo=ne.nesting=ne.id=ne.comment=ne.combinator=ne.className=ne.attribute=void 0;var WA=ze(Za()),GA=ze(La()),QA=ze(io()),YA=ze(Na()),KA=ze(Fa()),XA=ze(so()),JA=ze(Ga()),ZA=ze(Pa()),eC=ze(Da()),tC=ze(Ha()),rC=ze(Ua()),iC=ze(to());function ze(r){return r&&r.__esModule?r:{default:r}}var nC=function(e){return new WA.default(e)};ne.attribute=nC;var sC=function(e){return new GA.default(e)};ne.className=sC;var aC=function(e){return new QA.default(e)};ne.combinator=aC;var oC=function(e){return new YA.default(e)};ne.comment=oC;var lC=function(e){return new KA.default(e)};ne.id=lC;var uC=function(e){return new XA.default(e)};ne.nesting=uC;var fC=function(e){return new JA.default(e)};ne.pseudo=fC;var cC=function(e){return new ZA.default(e)};ne.root=cC;var pC=function(e){return new eC.default(e)};ne.selector=pC;var dC=function(e){return new tC.default(e)};ne.string=dC;var hC=function(e){return new rC.default(e)};ne.tag=hC;var mC=function(e){return new iC.default(e)};ne.universal=mC});var Ed=x(J=>{u();"use strict";J.__esModule=!0;J.isComment=J.isCombinator=J.isClassName=J.isAttribute=void 0;J.isContainer=EC;J.isIdentifier=void 0;J.isNamespace=OC;J.isNesting=void 0;J.isNode=go;J.isPseudo=void 0;J.isPseudoClass=_C;J.isPseudoElement=_d;J.isUniversal=J.isTag=J.isString=J.isSelector=J.isRoot=void 0;var ue=Se(),Oe,gC=(Oe={},Oe[ue.ATTRIBUTE]=!0,Oe[ue.CLASS]=!0,Oe[ue.COMBINATOR]=!0,Oe[ue.COMMENT]=!0,Oe[ue.ID]=!0,Oe[ue.NESTING]=!0,Oe[ue.PSEUDO]=!0,Oe[ue.ROOT]=!0,Oe[ue.SELECTOR]=!0,Oe[ue.STRING]=!0,Oe[ue.TAG]=!0,Oe[ue.UNIVERSAL]=!0,Oe);function go(r){return typeof r=="object"&&gC[r.type]}function Ue(r,e){return go(e)&&e.type===r}var Ad=Ue.bind(null,ue.ATTRIBUTE);J.isAttribute=Ad;var yC=Ue.bind(null,ue.CLASS);J.isClassName=yC;var bC=Ue.bind(null,ue.COMBINATOR);J.isCombinator=bC;var wC=Ue.bind(null,ue.COMMENT);J.isComment=wC;var vC=Ue.bind(null,ue.ID);J.isIdentifier=vC;var xC=Ue.bind(null,ue.NESTING);J.isNesting=xC;var yo=Ue.bind(null,ue.PSEUDO);J.isPseudo=yo;var kC=Ue.bind(null,ue.ROOT);J.isRoot=kC;var SC=Ue.bind(null,ue.SELECTOR);J.isSelector=SC;var AC=Ue.bind(null,ue.STRING);J.isString=AC;var Cd=Ue.bind(null,ue.TAG);J.isTag=Cd;var CC=Ue.bind(null,ue.UNIVERSAL);J.isUniversal=CC;function _d(r){return yo(r)&&r.value&&(r.value.startsWith("::")||r.value.toLowerCase()===":before"||r.value.toLowerCase()===":after"||r.value.toLowerCase()===":first-letter"||r.value.toLowerCase()===":first-line")}function _C(r){return yo(r)&&!_d(r)}function EC(r){return!!(go(r)&&r.walk)}function OC(r){return Ad(r)||Cd(r)}});var Od=x(Ke=>{u();"use strict";Ke.__esModule=!0;var bo=Se();Object.keys(bo).forEach(function(r){r==="default"||r==="__esModule"||r in Ke&&Ke[r]===bo[r]||(Ke[r]=bo[r])});var wo=Sd();Object.keys(wo).forEach(function(r){r==="default"||r==="__esModule"||r in Ke&&Ke[r]===wo[r]||(Ke[r]=wo[r])});var vo=Ed();Object.keys(vo).forEach(function(r){r==="default"||r==="__esModule"||r in Ke&&Ke[r]===vo[r]||(Ke[r]=vo[r])})});var it=x((ki,Rd)=>{u();"use strict";ki.__esModule=!0;ki.default=void 0;var TC=IC(kd()),RC=PC(Od());function Td(r){if(typeof WeakMap!="function")return null;var e=new WeakMap,t=new WeakMap;return(Td=function(n){return n?t:e})(r)}function PC(r,e){if(!e&&r&&r.__esModule)return r;if(r===null||typeof r!="object"&&typeof r!="function")return{default:r};var t=Td(e);if(t&&t.has(r))return t.get(r);var i={},n=Object.defineProperty&&Object.getOwnPropertyDescriptor;for(var a in r)if(a!=="default"&&Object.prototype.hasOwnProperty.call(r,a)){var s=n?Object.getOwnPropertyDescriptor(r,a):null;s&&(s.get||s.set)?Object.defineProperty(i,a,s):i[a]=r[a]}return i.default=r,t&&t.set(r,i),i}function IC(r){return r&&r.__esModule?r:{default:r}}var xo=function(e){return new TC.default(e)};Object.assign(xo,RC);delete xo.__esModule;var DC=xo;ki.default=DC;Rd.exports=ki.default});function mt(r){return["fontSize","outline"].includes(r)?e=>(typeof e=="function"&&(e=e({})),Array.isArray(e)&&(e=e[0]),e):r==="fontFamily"?e=>{typeof e=="function"&&(e=e({}));let t=Array.isArray(e)&&ke(e[1])?e[0]:e;return Array.isArray(t)?t.join(", "):t}:["boxShadow","transitionProperty","transitionDuration","transitionDelay","transitionTimingFunction","backgroundImage","backgroundSize","backgroundColor","cursor","animation"].includes(r)?e=>(typeof e=="function"&&(e=e({})),Array.isArray(e)&&(e=e.join(", ")),e):["gridTemplateColumns","gridTemplateRows","objectPosition"].includes(r)?e=>(typeof e=="function"&&(e=e({})),typeof e=="string"&&(e=ee.list.comma(e).join(" ")),e):(e,t={})=>(typeof e=="function"&&(e=e(t)),e)}var Si=R(()=>{u();Ot();Kt()});var Md=x(($I,_o)=>{u();var{Rule:Pd,AtRule:qC}=$e(),Id=it();function ko(r,e){let t;try{Id(i=>{t=i}).processSync(r)}catch(i){throw r.includes(":")?e?e.error("Missed semicolon"):i:e?e.error(i.message):i}return t.at(0)}function Dd(r,e){let t=!1;return r.each(i=>{if(i.type==="nesting"){let n=e.clone({});i.value!=="&"?i.replaceWith(ko(i.value.replace("&",n.toString()))):i.replaceWith(n),t=!0}else"nodes"in i&&i.nodes&&Dd(i,e)&&(t=!0)}),t}function qd(r,e){let t=[];return r.selectors.forEach(i=>{let n=ko(i,r);e.selectors.forEach(a=>{if(!a)return;let s=ko(a,e);Dd(s,n)||(s.prepend(Id.combinator({value:" "})),s.prepend(n.clone({}))),t.push(s.toString())})}),t}function Vn(r,e){let t=r.prev();for(e.after(r);t&&t.type==="comment";){let i=t.prev();e.after(t),t=i}return r}function $C(r){return function e(t,i,n,a=n){let s=[];if(i.each(o=>{o.type==="rule"&&n?a&&(o.selectors=qd(t,o)):o.type==="atrule"&&o.nodes?r[o.name]?e(t,o,a):i[Ao]!==!1&&s.push(o):s.push(o)}),n&&s.length){let o=t.clone({nodes:[]});for(let l of s)o.append(l);i.prepend(o)}}}function So(r,e,t){let i=new Pd({selector:r,nodes:[]});return i.append(e),t.after(i),i}function $d(r,e){let t={};for(let i of r)t[i]=!0;if(e)for(let i of e)t[i.replace(/^@/,"")]=!0;return t}function LC(r){r=r.trim();let e=r.match(/^\((.*)\)$/);if(!e)return{type:"basic",selector:r};let t=e[1].match(/^(with(?:out)?):(.+)$/);if(t){let i=t[1]==="with",n=Object.fromEntries(t[2].trim().split(/\s+/).map(s=>[s,!0]));if(i&&n.all)return{type:"noop"};let a=s=>!!n[s];return n.all?a=()=>!0:i&&(a=s=>s==="all"?!1:!n[s]),{type:"withrules",escapes:a}}return{type:"unknown"}}function MC(r){let e=[],t=r.parent;for(;t&&t instanceof qC;)e.push(t),t=t.parent;return e}function NC(r){let e=r[Ld];if(!e)r.after(r.nodes);else{let t=r.nodes,i,n=-1,a,s,o,l=MC(r);if(l.forEach((c,f)=>{if(e(c.name))i=c,n=f,s=o;else{let d=o;o=c.clone({nodes:[]}),d&&o.append(d),a=a||o}}),i?s?(a.append(t),i.after(s)):i.after(t):r.after(t),r.next()&&i){let c;l.slice(0,n+1).forEach((f,d,p)=>{let h=c;c=f.clone({nodes:[]}),h&&c.append(h);let b=[],y=(p[d-1]||r).next();for(;y;)b.push(y),y=y.next();c.append(b)}),c&&(s||t[t.length-1]).after(c)}}r.remove()}var Ao=Symbol("rootRuleMergeSel"),Ld=Symbol("rootRuleEscapes");function BC(r){let{params:e}=r,{type:t,selector:i,escapes:n}=LC(e);if(t==="unknown")throw r.error(`Unknown @${r.name} parameter ${JSON.stringify(e)}`);if(t==="basic"&&i){let a=new Pd({selector:i,nodes:r.nodes});r.removeAll(),r.append(a)}r[Ld]=n,r[Ao]=n?!n("all"):t==="noop"}var Co=Symbol("hasRootRule");_o.exports=(r={})=>{let e=$d(["media","supports","layer","container"],r.bubble),t=$C(e),i=$d(["document","font-face","keyframes","-webkit-keyframes","-moz-keyframes"],r.unwrap),n=(r.rootRuleName||"at-root").replace(/^@/,""),a=r.preserveEmpty;return{postcssPlugin:"postcss-nested",Once(s){s.walkAtRules(n,o=>{BC(o),s[Co]=!0})},Rule(s){let o=!1,l=s,c=!1,f=[];s.each(d=>{d.type==="rule"?(f.length&&(l=So(s.selector,f,l),f=[]),c=!0,o=!0,d.selectors=qd(s,d),l=Vn(d,l)):d.type==="atrule"?(f.length&&(l=So(s.selector,f,l),f=[]),d.name===n?(o=!0,t(s,d,!0,d[Ao]),l=Vn(d,l)):e[d.name]?(c=!0,o=!0,t(s,d,!0),l=Vn(d,l)):i[d.name]?(c=!0,o=!0,t(s,d,!1),l=Vn(d,l)):c&&f.push(d)):d.type==="decl"&&c&&f.push(d)}),f.length&&(l=So(s.selector,f,l)),o&&a!==!0&&(s.raws.semicolon=!0,s.nodes.length===0&&s.remove())},RootExit(s){s[Co]&&(s.walkAtRules(n,NC),s[Co]=!1)}}};_o.exports.postcss=!0});var jd=x((LI,Fd)=>{u();"use strict";var Nd=/-(\w|$)/g,Bd=(r,e)=>e.toUpperCase(),FC=r=>(r=r.toLowerCase(),r==="float"?"cssFloat":r.startsWith("-ms-")?r.substr(1).replace(Nd,Bd):r.replace(Nd,Bd));Fd.exports=FC});var To=x((MI,zd)=>{u();var jC=jd(),zC={boxFlex:!0,boxFlexGroup:!0,columnCount:!0,flex:!0,flexGrow:!0,flexPositive:!0,flexShrink:!0,flexNegative:!0,fontWeight:!0,lineClamp:!0,lineHeight:!0,opacity:!0,order:!0,orphans:!0,tabSize:!0,widows:!0,zIndex:!0,zoom:!0,fillOpacity:!0,strokeDashoffset:!0,strokeOpacity:!0,strokeWidth:!0};function Eo(r){return typeof r.nodes=="undefined"?!0:Oo(r)}function Oo(r){let e,t={};return r.each(i=>{if(i.type==="atrule")e="@"+i.name,i.params&&(e+=" "+i.params),typeof t[e]=="undefined"?t[e]=Eo(i):Array.isArray(t[e])?t[e].push(Eo(i)):t[e]=[t[e],Eo(i)];else if(i.type==="rule"){let n=Oo(i);if(t[i.selector])for(let a in n)t[i.selector][a]=n[a];else t[i.selector]=n}else if(i.type==="decl"){i.prop[0]==="-"&&i.prop[1]==="-"||i.parent&&i.parent.selector===":export"?e=i.prop:e=jC(i.prop);let n=i.value;!isNaN(i.value)&&zC[e]&&(n=parseFloat(i.value)),i.important&&(n+=" !important"),typeof t[e]=="undefined"?t[e]=n:Array.isArray(t[e])?t[e].push(n):t[e]=[t[e],n]}}),t}zd.exports=Oo});var Hn=x((NI,Wd)=>{u();var Ai=$e(),Ud=/\s*!important\s*$/i,UC={"box-flex":!0,"box-flex-group":!0,"column-count":!0,flex:!0,"flex-grow":!0,"flex-positive":!0,"flex-shrink":!0,"flex-negative":!0,"font-weight":!0,"line-clamp":!0,"line-height":!0,opacity:!0,order:!0,orphans:!0,"tab-size":!0,widows:!0,"z-index":!0,zoom:!0,"fill-opacity":!0,"stroke-dashoffset":!0,"stroke-opacity":!0,"stroke-width":!0};function VC(r){return r.replace(/([A-Z])/g,"-$1").replace(/^ms-/,"-ms-").toLowerCase()}function Vd(r,e,t){t===!1||t===null||(e.startsWith("--")||(e=VC(e)),typeof t=="number"&&(t===0||UC[e]?t=t.toString():t+="px"),e==="css-float"&&(e="float"),Ud.test(t)?(t=t.replace(Ud,""),r.push(Ai.decl({prop:e,value:t,important:!0}))):r.push(Ai.decl({prop:e,value:t})))}function Hd(r,e,t){let i=Ai.atRule({name:e[1],params:e[3]||""});typeof t=="object"&&(i.nodes=[],Ro(t,i)),r.push(i)}function Ro(r,e){let t,i,n;for(t in r)if(i=r[t],!(i===null||typeof i=="undefined"))if(t[0]==="@"){let a=t.match(/@(\S+)(\s+([\W\w]*)\s*)?/);if(Array.isArray(i))for(let s of i)Hd(e,a,s);else Hd(e,a,i)}else if(Array.isArray(i))for(let a of i)Vd(e,t,a);else typeof i=="object"?(n=Ai.rule({selector:t}),Ro(i,n),e.push(n)):Vd(e,t,i)}Wd.exports=function(r){let e=Ai.root();return Ro(r,e),e}});var Po=x((BI,Gd)=>{u();var HC=To();Gd.exports=function(e){return console&&console.warn&&e.warnings().forEach(t=>{let i=t.plugin||"PostCSS";console.warn(i+": "+t.text)}),HC(e.root)}});var Yd=x((FI,Qd)=>{u();var WC=$e(),GC=Po(),QC=Hn();Qd.exports=function(e){let t=WC(e);return async i=>{let n=await t.process(i,{parser:QC,from:void 0});return GC(n)}}});var Xd=x((jI,Kd)=>{u();var YC=$e(),KC=Po(),XC=Hn();Kd.exports=function(r){let e=YC(r);return t=>{let i=e.process(t,{parser:XC,from:void 0});return KC(i)}}});var Zd=x((zI,Jd)=>{u();var JC=To(),ZC=Hn(),e_=Yd(),t_=Xd();Jd.exports={objectify:JC,parse:ZC,async:e_,sync:t_}});var or,eh,UI,VI,HI,WI,th=R(()=>{u();or=pe(Zd()),eh=or.default,UI=or.default.objectify,VI=or.default.parse,HI=or.default.async,WI=or.default.sync});function lr(r){return Array.isArray(r)?r.flatMap(e=>ee([(0,rh.default)({bubble:["screen"]})]).process(e,{parser:eh}).root.nodes):lr([r])}var rh,Io=R(()=>{u();Ot();rh=pe(Md());th()});function ur(r,e,t=!1){if(r==="")return e;let i=typeof e=="string"?(0,ih.default)().astSync(e):e;return i.walkClasses(n=>{let a=n.value,s=t&&a.startsWith("-");n.value=s?`-${r}${a.slice(1)}`:`${r}${a}`}),typeof e=="string"?i.toString():i}var ih,Wn=R(()=>{u();ih=pe(it())});function Te(r){let e=nh.default.className();return e.value=r,jt(e?.raws?.value??e.value)}var nh,fr=R(()=>{u();nh=pe(it());Ki()});function Do(r){return jt(`.${Te(r)}`)}function Gn(r,e){return Do(Ci(r,e))}function Ci(r,e){return e==="DEFAULT"?r:e==="-"||e==="-DEFAULT"?`-${r}`:e.startsWith("-")?`-${r}${e}`:e.startsWith("/")?`${r}${e}`:`${r}-${e}`}var qo=R(()=>{u();fr();Ki()});function L(r,e=[[r,[r]]],{filterDefault:t=!1,...i}={}){let n=mt(r);return function({matchUtilities:a,theme:s}){for(let o of e){let l=Array.isArray(o[0])?o:[o];a(l.reduce((c,[f,d])=>Object.assign(c,{[f]:p=>d.reduce((h,b)=>Array.isArray(b)?Object.assign(h,{[b[0]]:b[1]}):Object.assign(h,{[b]:n(p)}),{})}),{}),{...i,values:t?Object.fromEntries(Object.entries(s(r)??{}).filter(([c])=>c!=="DEFAULT")):s(r)})}}}var sh=R(()=>{u();Si()});function Tt(r){return r=Array.isArray(r)?r:[r],r.map(e=>{let t=e.values.map(i=>i.raw!==void 0?i.raw:[i.min&&`(min-width: ${i.min})`,i.max&&`(max-width: ${i.max})`].filter(Boolean).join(" and "));return e.not?`not all and ${t}`:t}).join(", ")}var Qn=R(()=>{u()});function $o(r){return r.split(l_).map(t=>{let i=t.trim(),n={value:i},a=i.split(u_),s=new Set;for(let o of a)!s.has("DIRECTIONS")&&r_.has(o)?(n.direction=o,s.add("DIRECTIONS")):!s.has("PLAY_STATES")&&i_.has(o)?(n.playState=o,s.add("PLAY_STATES")):!s.has("FILL_MODES")&&n_.has(o)?(n.fillMode=o,s.add("FILL_MODES")):!s.has("ITERATION_COUNTS")&&(s_.has(o)||f_.test(o))?(n.iterationCount=o,s.add("ITERATION_COUNTS")):!s.has("TIMING_FUNCTION")&&a_.has(o)||!s.has("TIMING_FUNCTION")&&o_.some(l=>o.startsWith(`${l}(`))?(n.timingFunction=o,s.add("TIMING_FUNCTION")):!s.has("DURATION")&&ah.test(o)?(n.duration=o,s.add("DURATION")):!s.has("DELAY")&&ah.test(o)?(n.delay=o,s.add("DELAY")):s.has("NAME")?(n.unknown||(n.unknown=[]),n.unknown.push(o)):(n.name=o,s.add("NAME"));return n})}var r_,i_,n_,s_,a_,o_,l_,u_,ah,f_,oh=R(()=>{u();r_=new Set(["normal","reverse","alternate","alternate-reverse"]),i_=new Set(["running","paused"]),n_=new Set(["none","forwards","backwards","both"]),s_=new Set(["infinite"]),a_=new Set(["linear","ease","ease-in","ease-out","ease-in-out","step-start","step-end"]),o_=["cubic-bezier","steps"],l_=/\,(?![^(]*\))/g,u_=/\ +(?![^(]*\))/g,ah=/^(-?[\d.]+m?s)$/,f_=/^(\d+)$/});var lh,xe,uh=R(()=>{u();lh=r=>Object.assign({},...Object.entries(r??{}).flatMap(([e,t])=>typeof t=="object"?Object.entries(lh(t)).map(([i,n])=>({[e+(i==="DEFAULT"?"":`-${i}`)]:n})):[{[`${e}`]:t}])),xe=lh});var ch,fh=R(()=>{ch="3.4.14"});function Rt(r,e=!0){return Array.isArray(r)?r.map(t=>{if(e&&Array.isArray(t))throw new Error("The tuple syntax is not supported for `screens`.");if(typeof t=="string")return{name:t.toString(),not:!1,values:[{min:t,max:void 0}]};let[i,n]=t;return i=i.toString(),typeof n=="string"?{name:i,not:!1,values:[{min:n,max:void 0}]}:Array.isArray(n)?{name:i,not:!1,values:n.map(a=>dh(a))}:{name:i,not:!1,values:[dh(n)]}}):Rt(Object.entries(r??{}),!1)}function Yn(r){return r.values.length!==1?{result:!1,reason:"multiple-values"}:r.values[0].raw!==void 0?{result:!1,reason:"raw-values"}:r.values[0].min!==void 0&&r.values[0].max!==void 0?{result:!1,reason:"min-and-max"}:{result:!0,reason:null}}function ph(r,e,t){let i=Kn(e,r),n=Kn(t,r),a=Yn(i),s=Yn(n);if(a.reason==="multiple-values"||s.reason==="multiple-values")throw new Error("Attempted to sort a screen with multiple values. This should never happen. Please open a bug report.");if(a.reason==="raw-values"||s.reason==="raw-values")throw new Error("Attempted to sort a screen with raw values. This should never happen. Please open a bug report.");if(a.reason==="min-and-max"||s.reason==="min-and-max")throw new Error("Attempted to sort a screen with both min and max values. This should never happen. Please open a bug report.");let{min:o,max:l}=i.values[0],{min:c,max:f}=n.values[0];e.not&&([o,l]=[l,o]),t.not&&([c,f]=[f,c]),o=o===void 0?o:parseFloat(o),l=l===void 0?l:parseFloat(l),c=c===void 0?c:parseFloat(c),f=f===void 0?f:parseFloat(f);let[d,p]=r==="min"?[o,c]:[f,l];return d-p}function Kn(r,e){return typeof r=="object"?r:{name:"arbitrary-screen",values:[{[e]:r}]}}function dh({"min-width":r,min:e=r,max:t,raw:i}={}){return{min:e,max:t,raw:i}}var Xn=R(()=>{u()});function Jn(r,e){r.walkDecls(t=>{if(e.includes(t.prop)){t.remove();return}for(let i of e)t.value.includes(`/ var(${i})`)&&(t.value=t.value.replace(`/ var(${i})`,""))})}var hh=R(()=>{u()});var se,Xe,nt,ge,mh,gh=R(()=>{u();ft();et();Ot();sh();Qn();fr();oh();uh();Lr();ea();Kt();Si();fh();Be();Xn();Gs();hh();ct();Br();_i();se={childVariant:({addVariant:r})=>{r("*","& > *")},pseudoElementVariants:({addVariant:r})=>{r("first-letter","&::first-letter"),r("first-line","&::first-line"),r("marker",[({container:e})=>(Jn(e,["--tw-text-opacity"]),"& *::marker"),({container:e})=>(Jn(e,["--tw-text-opacity"]),"&::marker")]),r("selection",["& *::selection","&::selection"]),r("file","&::file-selector-button"),r("placeholder","&::placeholder"),r("backdrop","&::backdrop"),r("before",({container:e})=>(e.walkRules(t=>{let i=!1;t.walkDecls("content",()=>{i=!0}),i||t.prepend(ee.decl({prop:"content",value:"var(--tw-content)"}))}),"&::before")),r("after",({container:e})=>(e.walkRules(t=>{let i=!1;t.walkDecls("content",()=>{i=!0}),i||t.prepend(ee.decl({prop:"content",value:"var(--tw-content)"}))}),"&::after"))},pseudoClassVariants:({addVariant:r,matchVariant:e,config:t,prefix:i})=>{let n=[["first","&:first-child"],["last","&:last-child"],["only","&:only-child"],["odd","&:nth-child(odd)"],["even","&:nth-child(even)"],"first-of-type","last-of-type","only-of-type",["visited",({container:s})=>(Jn(s,["--tw-text-opacity","--tw-border-opacity","--tw-bg-opacity"]),"&:visited")],"target",["open","&[open]"],"default","checked","indeterminate","placeholder-shown","autofill","optional","required","valid","invalid","in-range","out-of-range","read-only","empty","focus-within",["hover",we(t(),"hoverOnlyWhenSupported")?"@media (hover: hover) and (pointer: fine) { &:hover }":"&:hover"],"focus","focus-visible","active","enabled","disabled"].map(s=>Array.isArray(s)?s:[s,`&:${s}`]);for(let[s,o]of n)r(s,l=>typeof o=="function"?o(l):o);let a={group:(s,{modifier:o})=>o?[`:merge(${i(".group")}\\/${Te(o)})`," &"]:[`:merge(${i(".group")})`," &"],peer:(s,{modifier:o})=>o?[`:merge(${i(".peer")}\\/${Te(o)})`," ~ &"]:[`:merge(${i(".peer")})`," ~ &"]};for(let[s,o]of Object.entries(a))e(s,(l="",c)=>{let f=K(typeof l=="function"?l(c):l);f.includes("&")||(f="&"+f);let[d,p]=o("",c),h=null,b=null,v=0;for(let y=0;y{r("ltr",'&:where([dir="ltr"], [dir="ltr"] *)'),r("rtl",'&:where([dir="rtl"], [dir="rtl"] *)')},reducedMotionVariants:({addVariant:r})=>{r("motion-safe","@media (prefers-reduced-motion: no-preference)"),r("motion-reduce","@media (prefers-reduced-motion: reduce)")},darkVariants:({config:r,addVariant:e})=>{let[t,i=".dark"]=[].concat(r("darkMode","media"));if(t===!1&&(t="media",G.warn("darkmode-false",["The `darkMode` option in your Tailwind CSS configuration is set to `false`, which now behaves the same as `media`.","Change `darkMode` to `media` or remove it entirely.","https://tailwindcss.com/docs/upgrade-guide#remove-dark-mode-configuration"])),t==="variant"){let n;if(Array.isArray(i)||typeof i=="function"?n=i:typeof i=="string"&&(n=[i]),Array.isArray(n))for(let a of n)a===".dark"?(t=!1,G.warn("darkmode-variant-without-selector",["When using `variant` for `darkMode`, you must provide a selector.",'Example: `darkMode: ["variant", ".your-selector &"]`'])):a.includes("&")||(t=!1,G.warn("darkmode-variant-without-ampersand",["When using `variant` for `darkMode`, your selector must contain `&`.",'Example `darkMode: ["variant", ".your-selector &"]`']));i=n}t==="selector"?e("dark",`&:where(${i}, ${i} *)`):t==="media"?e("dark","@media (prefers-color-scheme: dark)"):t==="variant"?e("dark",i):t==="class"&&e("dark",`&:is(${i} *)`)},printVariant:({addVariant:r})=>{r("print","@media print")},screenVariants:({theme:r,addVariant:e,matchVariant:t})=>{let i=r("screens")??{},n=Object.values(i).every(w=>typeof w=="string"),a=Rt(r("screens")),s=new Set([]);function o(w){return w.match(/(\D+)$/)?.[1]??"(none)"}function l(w){w!==void 0&&s.add(o(w))}function c(w){return l(w),s.size===1}for(let w of a)for(let k of w.values)l(k.min),l(k.max);let f=s.size<=1;function d(w){return Object.fromEntries(a.filter(k=>Yn(k).result).map(k=>{let{min:S,max:E}=k.values[0];if(w==="min"&&S!==void 0)return k;if(w==="min"&&E!==void 0)return{...k,not:!k.not};if(w==="max"&&E!==void 0)return k;if(w==="max"&&S!==void 0)return{...k,not:!k.not}}).map(k=>[k.name,k]))}function p(w){return(k,S)=>ph(w,k.value,S.value)}let h=p("max"),b=p("min");function v(w){return k=>{if(n)if(f){if(typeof k=="string"&&!c(k))return G.warn("minmax-have-mixed-units",["The `min-*` and `max-*` variants are not supported with a `screens` configuration containing mixed units."]),[]}else return G.warn("mixed-screen-units",["The `min-*` and `max-*` variants are not supported with a `screens` configuration containing mixed units."]),[];else return G.warn("complex-screen-config",["The `min-*` and `max-*` variants are not supported with a `screens` configuration containing objects."]),[];return[`@media ${Tt(Kn(k,w))}`]}}t("max",v("max"),{sort:h,values:n?d("max"):{}});let y="min-screens";for(let w of a)e(w.name,`@media ${Tt(w)}`,{id:y,sort:n&&f?b:void 0,value:w});t("min",v("min"),{id:y,sort:b})},supportsVariants:({matchVariant:r,theme:e})=>{r("supports",(t="")=>{let i=K(t),n=/^\w*\s*\(/.test(i);return i=n?i.replace(/\b(and|or|not)\b/g," $1 "):i,n?`@supports ${i}`:(i.includes(":")||(i=`${i}: var(--tw)`),i.startsWith("(")&&i.endsWith(")")||(i=`(${i})`),`@supports ${i}`)},{values:e("supports")??{}})},hasVariants:({matchVariant:r,prefix:e})=>{r("has",t=>`&:has(${K(t)})`,{values:{},[Pt]:{respectPrefix:!1}}),r("group-has",(t,{modifier:i})=>i?`:merge(${e(".group")}\\/${i}):has(${K(t)}) &`:`:merge(${e(".group")}):has(${K(t)}) &`,{values:{},[Pt]:{respectPrefix:!1}}),r("peer-has",(t,{modifier:i})=>i?`:merge(${e(".peer")}\\/${i}):has(${K(t)}) ~ &`:`:merge(${e(".peer")}):has(${K(t)}) ~ &`,{values:{},[Pt]:{respectPrefix:!1}})},ariaVariants:({matchVariant:r,theme:e})=>{r("aria",t=>`&[aria-${Ye(K(t))}]`,{values:e("aria")??{}}),r("group-aria",(t,{modifier:i})=>i?`:merge(.group\\/${i})[aria-${Ye(K(t))}] &`:`:merge(.group)[aria-${Ye(K(t))}] &`,{values:e("aria")??{}}),r("peer-aria",(t,{modifier:i})=>i?`:merge(.peer\\/${i})[aria-${Ye(K(t))}] ~ &`:`:merge(.peer)[aria-${Ye(K(t))}] ~ &`,{values:e("aria")??{}})},dataVariants:({matchVariant:r,theme:e})=>{r("data",t=>`&[data-${Ye(K(t))}]`,{values:e("data")??{}}),r("group-data",(t,{modifier:i})=>i?`:merge(.group\\/${i})[data-${Ye(K(t))}] &`:`:merge(.group)[data-${Ye(K(t))}] &`,{values:e("data")??{}}),r("peer-data",(t,{modifier:i})=>i?`:merge(.peer\\/${i})[data-${Ye(K(t))}] ~ &`:`:merge(.peer)[data-${Ye(K(t))}] ~ &`,{values:e("data")??{}})},orientationVariants:({addVariant:r})=>{r("portrait","@media (orientation: portrait)"),r("landscape","@media (orientation: landscape)")},prefersContrastVariants:({addVariant:r})=>{r("contrast-more","@media (prefers-contrast: more)"),r("contrast-less","@media (prefers-contrast: less)")},forcedColorsVariants:({addVariant:r})=>{r("forced-colors","@media (forced-colors: active)")}},Xe=["translate(var(--tw-translate-x), var(--tw-translate-y))","rotate(var(--tw-rotate))","skewX(var(--tw-skew-x))","skewY(var(--tw-skew-y))","scaleX(var(--tw-scale-x))","scaleY(var(--tw-scale-y))"].join(" "),nt=["var(--tw-blur)","var(--tw-brightness)","var(--tw-contrast)","var(--tw-grayscale)","var(--tw-hue-rotate)","var(--tw-invert)","var(--tw-saturate)","var(--tw-sepia)","var(--tw-drop-shadow)"].join(" "),ge=["var(--tw-backdrop-blur)","var(--tw-backdrop-brightness)","var(--tw-backdrop-contrast)","var(--tw-backdrop-grayscale)","var(--tw-backdrop-hue-rotate)","var(--tw-backdrop-invert)","var(--tw-backdrop-opacity)","var(--tw-backdrop-saturate)","var(--tw-backdrop-sepia)"].join(" "),mh={preflight:({addBase:r})=>{let e=ee.parse(`*,::after,::before{box-sizing:border-box;border-width:0;border-style:solid;border-color:theme('borderColor.DEFAULT', currentColor)}::after,::before{--tw-content:''}:host,html{line-height:1.5;-webkit-text-size-adjust:100%;-moz-tab-size:4;tab-size:4;font-family:theme('fontFamily.sans', ui-sans-serif, system-ui, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji");font-feature-settings:theme('fontFamily.sans[1].fontFeatureSettings', normal);font-variation-settings:theme('fontFamily.sans[1].fontVariationSettings', normal);-webkit-tap-highlight-color:transparent}body{margin:0;line-height:inherit}hr{height:0;color:inherit;border-top-width:1px}abbr:where([title]){text-decoration:underline dotted}h1,h2,h3,h4,h5,h6{font-size:inherit;font-weight:inherit}a{color:inherit;text-decoration:inherit}b,strong{font-weight:bolder}code,kbd,pre,samp{font-family:theme('fontFamily.mono', ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, "Liberation Mono", "Courier New", monospace);font-feature-settings:theme('fontFamily.mono[1].fontFeatureSettings', normal);font-variation-settings:theme('fontFamily.mono[1].fontVariationSettings', normal);font-size:1em}small{font-size:80%}sub,sup{font-size:75%;line-height:0;position:relative;vertical-align:baseline}sub{bottom:-.25em}sup{top:-.5em}table{text-indent:0;border-color:inherit;border-collapse:collapse}button,input,optgroup,select,textarea{font-family:inherit;font-feature-settings:inherit;font-variation-settings:inherit;font-size:100%;font-weight:inherit;line-height:inherit;letter-spacing:inherit;color:inherit;margin:0;padding:0}button,select{text-transform:none}button,input:where([type=button]),input:where([type=reset]),input:where([type=submit]){-webkit-appearance:button;background-color:transparent;background-image:none}:-moz-focusring{outline:auto}:-moz-ui-invalid{box-shadow:none}progress{vertical-align:baseline}::-webkit-inner-spin-button,::-webkit-outer-spin-button{height:auto}[type=search]{-webkit-appearance:textfield;outline-offset:-2px}::-webkit-search-decoration{-webkit-appearance:none}::-webkit-file-upload-button{-webkit-appearance:button;font:inherit}summary{display:list-item}blockquote,dd,dl,figure,h1,h2,h3,h4,h5,h6,hr,p,pre{margin:0}fieldset{margin:0;padding:0}legend{padding:0}menu,ol,ul{list-style:none;margin:0;padding:0}dialog{padding:0}textarea{resize:vertical}input::placeholder,textarea::placeholder{opacity:1;color:theme('colors.gray.4', #9ca3af)}[role=button],button{cursor:pointer}:disabled{cursor:default}audio,canvas,embed,iframe,img,object,svg,video{display:block;vertical-align:middle}img,video{max-width:100%;height:auto}[hidden]:where(:not([hidden=until-found])){display:none}`);r([ee.comment({text:`! tailwindcss v${ch} | MIT License | https://tailwindcss.com`}),...e.nodes])},container:(()=>{function r(t=[]){return t.flatMap(i=>i.values.map(n=>n.min)).filter(i=>i!==void 0)}function e(t,i,n){if(typeof n=="undefined")return[];if(!(typeof n=="object"&&n!==null))return[{screen:"DEFAULT",minWidth:0,padding:n}];let a=[];n.DEFAULT&&a.push({screen:"DEFAULT",minWidth:0,padding:n.DEFAULT});for(let s of t)for(let o of i)for(let{min:l}of o.values)l===s&&a.push({minWidth:s,padding:n[o.name]});return a}return function({addComponents:t,theme:i}){let n=Rt(i("container.screens",i("screens"))),a=r(n),s=e(a,n,i("container.padding")),o=c=>{let f=s.find(d=>d.minWidth===c);return f?{paddingRight:f.padding,paddingLeft:f.padding}:{}},l=Array.from(new Set(a.slice().sort((c,f)=>parseInt(c)-parseInt(f)))).map(c=>({[`@media (min-width: ${c})`]:{".container":{"max-width":c,...o(c)}}}));t([{".container":Object.assign({width:"100%"},i("container.center",!1)?{marginRight:"auto",marginLeft:"auto"}:{},o(0))},...l])}})(),accessibility:({addUtilities:r})=>{r({".sr-only":{position:"absolute",width:"1px",height:"1px",padding:"0",margin:"-1px",overflow:"hidden",clip:"rect(0, 0, 0, 0)",whiteSpace:"nowrap",borderWidth:"0"},".not-sr-only":{position:"static",width:"auto",height:"auto",padding:"0",margin:"0",overflow:"visible",clip:"auto",whiteSpace:"normal"}})},pointerEvents:({addUtilities:r})=>{r({".pointer-events-none":{"pointer-events":"none"},".pointer-events-auto":{"pointer-events":"auto"}})},visibility:({addUtilities:r})=>{r({".visible":{visibility:"visible"},".invisible":{visibility:"hidden"},".collapse":{visibility:"collapse"}})},position:({addUtilities:r})=>{r({".static":{position:"static"},".fixed":{position:"fixed"},".absolute":{position:"absolute"},".relative":{position:"relative"},".sticky":{position:"sticky"}})},inset:L("inset",[["inset",["inset"]],[["inset-x",["left","right"]],["inset-y",["top","bottom"]]],[["start",["inset-inline-start"]],["end",["inset-inline-end"]],["top",["top"]],["right",["right"]],["bottom",["bottom"]],["left",["left"]]]],{supportsNegativeValues:!0}),isolation:({addUtilities:r})=>{r({".isolate":{isolation:"isolate"},".isolation-auto":{isolation:"auto"}})},zIndex:L("zIndex",[["z",["zIndex"]]],{supportsNegativeValues:!0}),order:L("order",void 0,{supportsNegativeValues:!0}),gridColumn:L("gridColumn",[["col",["gridColumn"]]]),gridColumnStart:L("gridColumnStart",[["col-start",["gridColumnStart"]]],{supportsNegativeValues:!0}),gridColumnEnd:L("gridColumnEnd",[["col-end",["gridColumnEnd"]]],{supportsNegativeValues:!0}),gridRow:L("gridRow",[["row",["gridRow"]]]),gridRowStart:L("gridRowStart",[["row-start",["gridRowStart"]]],{supportsNegativeValues:!0}),gridRowEnd:L("gridRowEnd",[["row-end",["gridRowEnd"]]],{supportsNegativeValues:!0}),float:({addUtilities:r})=>{r({".float-start":{float:"inline-start"},".float-end":{float:"inline-end"},".float-right":{float:"right"},".float-left":{float:"left"},".float-none":{float:"none"}})},clear:({addUtilities:r})=>{r({".clear-start":{clear:"inline-start"},".clear-end":{clear:"inline-end"},".clear-left":{clear:"left"},".clear-right":{clear:"right"},".clear-both":{clear:"both"},".clear-none":{clear:"none"}})},margin:L("margin",[["m",["margin"]],[["mx",["margin-left","margin-right"]],["my",["margin-top","margin-bottom"]]],[["ms",["margin-inline-start"]],["me",["margin-inline-end"]],["mt",["margin-top"]],["mr",["margin-right"]],["mb",["margin-bottom"]],["ml",["margin-left"]]]],{supportsNegativeValues:!0}),boxSizing:({addUtilities:r})=>{r({".box-border":{"box-sizing":"border-box"},".box-content":{"box-sizing":"content-box"}})},lineClamp:({matchUtilities:r,addUtilities:e,theme:t})=>{r({"line-clamp":i=>({overflow:"hidden",display:"-webkit-box","-webkit-box-orient":"vertical","-webkit-line-clamp":`${i}`})},{values:t("lineClamp")}),e({".line-clamp-none":{overflow:"visible",display:"block","-webkit-box-orient":"horizontal","-webkit-line-clamp":"none"}})},display:({addUtilities:r})=>{r({".block":{display:"block"},".inline-block":{display:"inline-block"},".inline":{display:"inline"},".flex":{display:"flex"},".inline-flex":{display:"inline-flex"},".table":{display:"table"},".inline-table":{display:"inline-table"},".table-caption":{display:"table-caption"},".table-cell":{display:"table-cell"},".table-column":{display:"table-column"},".table-column-group":{display:"table-column-group"},".table-footer-group":{display:"table-footer-group"},".table-header-group":{display:"table-header-group"},".table-row-group":{display:"table-row-group"},".table-row":{display:"table-row"},".flow-root":{display:"flow-root"},".grid":{display:"grid"},".inline-grid":{display:"inline-grid"},".contents":{display:"contents"},".list-item":{display:"list-item"},".hidden":{display:"none"}})},aspectRatio:L("aspectRatio",[["aspect",["aspect-ratio"]]]),size:L("size",[["size",["width","height"]]]),height:L("height",[["h",["height"]]]),maxHeight:L("maxHeight",[["max-h",["maxHeight"]]]),minHeight:L("minHeight",[["min-h",["minHeight"]]]),width:L("width",[["w",["width"]]]),minWidth:L("minWidth",[["min-w",["minWidth"]]]),maxWidth:L("maxWidth",[["max-w",["maxWidth"]]]),flex:L("flex"),flexShrink:L("flexShrink",[["flex-shrink",["flex-shrink"]],["shrink",["flex-shrink"]]]),flexGrow:L("flexGrow",[["flex-grow",["flex-grow"]],["grow",["flex-grow"]]]),flexBasis:L("flexBasis",[["basis",["flex-basis"]]]),tableLayout:({addUtilities:r})=>{r({".table-auto":{"table-layout":"auto"},".table-fixed":{"table-layout":"fixed"}})},captionSide:({addUtilities:r})=>{r({".caption-top":{"caption-side":"top"},".caption-bottom":{"caption-side":"bottom"}})},borderCollapse:({addUtilities:r})=>{r({".border-collapse":{"border-collapse":"collapse"},".border-separate":{"border-collapse":"separate"}})},borderSpacing:({addDefaults:r,matchUtilities:e,theme:t})=>{r("border-spacing",{"--tw-border-spacing-x":0,"--tw-border-spacing-y":0}),e({"border-spacing":i=>({"--tw-border-spacing-x":i,"--tw-border-spacing-y":i,"@defaults border-spacing":{},"border-spacing":"var(--tw-border-spacing-x) var(--tw-border-spacing-y)"}),"border-spacing-x":i=>({"--tw-border-spacing-x":i,"@defaults border-spacing":{},"border-spacing":"var(--tw-border-spacing-x) var(--tw-border-spacing-y)"}),"border-spacing-y":i=>({"--tw-border-spacing-y":i,"@defaults border-spacing":{},"border-spacing":"var(--tw-border-spacing-x) var(--tw-border-spacing-y)"})},{values:t("borderSpacing")})},transformOrigin:L("transformOrigin",[["origin",["transformOrigin"]]]),translate:L("translate",[[["translate-x",[["@defaults transform",{}],"--tw-translate-x",["transform",Xe]]],["translate-y",[["@defaults transform",{}],"--tw-translate-y",["transform",Xe]]]]],{supportsNegativeValues:!0}),rotate:L("rotate",[["rotate",[["@defaults transform",{}],"--tw-rotate",["transform",Xe]]]],{supportsNegativeValues:!0}),skew:L("skew",[[["skew-x",[["@defaults transform",{}],"--tw-skew-x",["transform",Xe]]],["skew-y",[["@defaults transform",{}],"--tw-skew-y",["transform",Xe]]]]],{supportsNegativeValues:!0}),scale:L("scale",[["scale",[["@defaults transform",{}],"--tw-scale-x","--tw-scale-y",["transform",Xe]]],[["scale-x",[["@defaults transform",{}],"--tw-scale-x",["transform",Xe]]],["scale-y",[["@defaults transform",{}],"--tw-scale-y",["transform",Xe]]]]],{supportsNegativeValues:!0}),transform:({addDefaults:r,addUtilities:e})=>{r("transform",{"--tw-translate-x":"0","--tw-translate-y":"0","--tw-rotate":"0","--tw-skew-x":"0","--tw-skew-y":"0","--tw-scale-x":"1","--tw-scale-y":"1"}),e({".transform":{"@defaults transform":{},transform:Xe},".transform-cpu":{transform:Xe},".transform-gpu":{transform:Xe.replace("translate(var(--tw-translate-x), var(--tw-translate-y))","translate3d(var(--tw-translate-x), var(--tw-translate-y), 0)")},".transform-none":{transform:"none"}})},animation:({matchUtilities:r,theme:e,config:t})=>{let i=a=>Te(t("prefix")+a),n=Object.fromEntries(Object.entries(e("keyframes")??{}).map(([a,s])=>[a,{[`@keyframes ${i(a)}`]:s}]));r({animate:a=>{let s=$o(a);return[...s.flatMap(o=>n[o.name]),{animation:s.map(({name:o,value:l})=>o===void 0||n[o]===void 0?l:l.replace(o,i(o))).join(", ")}]}},{values:e("animation")})},cursor:L("cursor"),touchAction:({addDefaults:r,addUtilities:e})=>{r("touch-action",{"--tw-pan-x":" ","--tw-pan-y":" ","--tw-pinch-zoom":" "});let t="var(--tw-pan-x) var(--tw-pan-y) var(--tw-pinch-zoom)";e({".touch-auto":{"touch-action":"auto"},".touch-none":{"touch-action":"none"},".touch-pan-x":{"@defaults touch-action":{},"--tw-pan-x":"pan-x","touch-action":t},".touch-pan-left":{"@defaults touch-action":{},"--tw-pan-x":"pan-left","touch-action":t},".touch-pan-right":{"@defaults touch-action":{},"--tw-pan-x":"pan-right","touch-action":t},".touch-pan-y":{"@defaults touch-action":{},"--tw-pan-y":"pan-y","touch-action":t},".touch-pan-up":{"@defaults touch-action":{},"--tw-pan-y":"pan-up","touch-action":t},".touch-pan-down":{"@defaults touch-action":{},"--tw-pan-y":"pan-down","touch-action":t},".touch-pinch-zoom":{"@defaults touch-action":{},"--tw-pinch-zoom":"pinch-zoom","touch-action":t},".touch-manipulation":{"touch-action":"manipulation"}})},userSelect:({addUtilities:r})=>{r({".select-none":{"user-select":"none"},".select-text":{"user-select":"text"},".select-all":{"user-select":"all"},".select-auto":{"user-select":"auto"}})},resize:({addUtilities:r})=>{r({".resize-none":{resize:"none"},".resize-y":{resize:"vertical"},".resize-x":{resize:"horizontal"},".resize":{resize:"both"}})},scrollSnapType:({addDefaults:r,addUtilities:e})=>{r("scroll-snap-type",{"--tw-scroll-snap-strictness":"proximity"}),e({".snap-none":{"scroll-snap-type":"none"},".snap-x":{"@defaults scroll-snap-type":{},"scroll-snap-type":"x var(--tw-scroll-snap-strictness)"},".snap-y":{"@defaults scroll-snap-type":{},"scroll-snap-type":"y var(--tw-scroll-snap-strictness)"},".snap-both":{"@defaults scroll-snap-type":{},"scroll-snap-type":"both var(--tw-scroll-snap-strictness)"},".snap-mandatory":{"--tw-scroll-snap-strictness":"mandatory"},".snap-proximity":{"--tw-scroll-snap-strictness":"proximity"}})},scrollSnapAlign:({addUtilities:r})=>{r({".snap-start":{"scroll-snap-align":"start"},".snap-end":{"scroll-snap-align":"end"},".snap-center":{"scroll-snap-align":"center"},".snap-align-none":{"scroll-snap-align":"none"}})},scrollSnapStop:({addUtilities:r})=>{r({".snap-normal":{"scroll-snap-stop":"normal"},".snap-always":{"scroll-snap-stop":"always"}})},scrollMargin:L("scrollMargin",[["scroll-m",["scroll-margin"]],[["scroll-mx",["scroll-margin-left","scroll-margin-right"]],["scroll-my",["scroll-margin-top","scroll-margin-bottom"]]],[["scroll-ms",["scroll-margin-inline-start"]],["scroll-me",["scroll-margin-inline-end"]],["scroll-mt",["scroll-margin-top"]],["scroll-mr",["scroll-margin-right"]],["scroll-mb",["scroll-margin-bottom"]],["scroll-ml",["scroll-margin-left"]]]],{supportsNegativeValues:!0}),scrollPadding:L("scrollPadding",[["scroll-p",["scroll-padding"]],[["scroll-px",["scroll-padding-left","scroll-padding-right"]],["scroll-py",["scroll-padding-top","scroll-padding-bottom"]]],[["scroll-ps",["scroll-padding-inline-start"]],["scroll-pe",["scroll-padding-inline-end"]],["scroll-pt",["scroll-padding-top"]],["scroll-pr",["scroll-padding-right"]],["scroll-pb",["scroll-padding-bottom"]],["scroll-pl",["scroll-padding-left"]]]]),listStylePosition:({addUtilities:r})=>{r({".list-inside":{"list-style-position":"inside"},".list-outside":{"list-style-position":"outside"}})},listStyleType:L("listStyleType",[["list",["listStyleType"]]]),listStyleImage:L("listStyleImage",[["list-image",["listStyleImage"]]]),appearance:({addUtilities:r})=>{r({".appearance-none":{appearance:"none"},".appearance-auto":{appearance:"auto"}})},columns:L("columns",[["columns",["columns"]]]),breakBefore:({addUtilities:r})=>{r({".break-before-auto":{"break-before":"auto"},".break-before-avoid":{"break-before":"avoid"},".break-before-all":{"break-before":"all"},".break-before-avoid-page":{"break-before":"avoid-page"},".break-before-page":{"break-before":"page"},".break-before-left":{"break-before":"left"},".break-before-right":{"break-before":"right"},".break-before-column":{"break-before":"column"}})},breakInside:({addUtilities:r})=>{r({".break-inside-auto":{"break-inside":"auto"},".break-inside-avoid":{"break-inside":"avoid"},".break-inside-avoid-page":{"break-inside":"avoid-page"},".break-inside-avoid-column":{"break-inside":"avoid-column"}})},breakAfter:({addUtilities:r})=>{r({".break-after-auto":{"break-after":"auto"},".break-after-avoid":{"break-after":"avoid"},".break-after-all":{"break-after":"all"},".break-after-avoid-page":{"break-after":"avoid-page"},".break-after-page":{"break-after":"page"},".break-after-left":{"break-after":"left"},".break-after-right":{"break-after":"right"},".break-after-column":{"break-after":"column"}})},gridAutoColumns:L("gridAutoColumns",[["auto-cols",["gridAutoColumns"]]]),gridAutoFlow:({addUtilities:r})=>{r({".grid-flow-row":{gridAutoFlow:"row"},".grid-flow-col":{gridAutoFlow:"column"},".grid-flow-dense":{gridAutoFlow:"dense"},".grid-flow-row-dense":{gridAutoFlow:"row dense"},".grid-flow-col-dense":{gridAutoFlow:"column dense"}})},gridAutoRows:L("gridAutoRows",[["auto-rows",["gridAutoRows"]]]),gridTemplateColumns:L("gridTemplateColumns",[["grid-cols",["gridTemplateColumns"]]]),gridTemplateRows:L("gridTemplateRows",[["grid-rows",["gridTemplateRows"]]]),flexDirection:({addUtilities:r})=>{r({".flex-row":{"flex-direction":"row"},".flex-row-reverse":{"flex-direction":"row-reverse"},".flex-col":{"flex-direction":"column"},".flex-col-reverse":{"flex-direction":"column-reverse"}})},flexWrap:({addUtilities:r})=>{r({".flex-wrap":{"flex-wrap":"wrap"},".flex-wrap-reverse":{"flex-wrap":"wrap-reverse"},".flex-nowrap":{"flex-wrap":"nowrap"}})},placeContent:({addUtilities:r})=>{r({".place-content-center":{"place-content":"center"},".place-content-start":{"place-content":"start"},".place-content-end":{"place-content":"end"},".place-content-between":{"place-content":"space-between"},".place-content-around":{"place-content":"space-around"},".place-content-evenly":{"place-content":"space-evenly"},".place-content-baseline":{"place-content":"baseline"},".place-content-stretch":{"place-content":"stretch"}})},placeItems:({addUtilities:r})=>{r({".place-items-start":{"place-items":"start"},".place-items-end":{"place-items":"end"},".place-items-center":{"place-items":"center"},".place-items-baseline":{"place-items":"baseline"},".place-items-stretch":{"place-items":"stretch"}})},alignContent:({addUtilities:r})=>{r({".content-normal":{"align-content":"normal"},".content-center":{"align-content":"center"},".content-start":{"align-content":"flex-start"},".content-end":{"align-content":"flex-end"},".content-between":{"align-content":"space-between"},".content-around":{"align-content":"space-around"},".content-evenly":{"align-content":"space-evenly"},".content-baseline":{"align-content":"baseline"},".content-stretch":{"align-content":"stretch"}})},alignItems:({addUtilities:r})=>{r({".items-start":{"align-items":"flex-start"},".items-end":{"align-items":"flex-end"},".items-center":{"align-items":"center"},".items-baseline":{"align-items":"baseline"},".items-stretch":{"align-items":"stretch"}})},justifyContent:({addUtilities:r})=>{r({".justify-normal":{"justify-content":"normal"},".justify-start":{"justify-content":"flex-start"},".justify-end":{"justify-content":"flex-end"},".justify-center":{"justify-content":"center"},".justify-between":{"justify-content":"space-between"},".justify-around":{"justify-content":"space-around"},".justify-evenly":{"justify-content":"space-evenly"},".justify-stretch":{"justify-content":"stretch"}})},justifyItems:({addUtilities:r})=>{r({".justify-items-start":{"justify-items":"start"},".justify-items-end":{"justify-items":"end"},".justify-items-center":{"justify-items":"center"},".justify-items-stretch":{"justify-items":"stretch"}})},gap:L("gap",[["gap",["gap"]],[["gap-x",["columnGap"]],["gap-y",["rowGap"]]]]),space:({matchUtilities:r,addUtilities:e,theme:t})=>{r({"space-x":i=>(i=i==="0"?"0px":i,{"& > :not([hidden]) ~ :not([hidden])":{"--tw-space-x-reverse":"0","margin-right":`calc(${i} * var(--tw-space-x-reverse))`,"margin-left":`calc(${i} * calc(1 - var(--tw-space-x-reverse)))`}}),"space-y":i=>(i=i==="0"?"0px":i,{"& > :not([hidden]) ~ :not([hidden])":{"--tw-space-y-reverse":"0","margin-top":`calc(${i} * calc(1 - var(--tw-space-y-reverse)))`,"margin-bottom":`calc(${i} * var(--tw-space-y-reverse))`}})},{values:t("space"),supportsNegativeValues:!0}),e({".space-y-reverse > :not([hidden]) ~ :not([hidden])":{"--tw-space-y-reverse":"1"},".space-x-reverse > :not([hidden]) ~ :not([hidden])":{"--tw-space-x-reverse":"1"}})},divideWidth:({matchUtilities:r,addUtilities:e,theme:t})=>{r({"divide-x":i=>(i=i==="0"?"0px":i,{"& > :not([hidden]) ~ :not([hidden])":{"@defaults border-width":{},"--tw-divide-x-reverse":"0","border-right-width":`calc(${i} * var(--tw-divide-x-reverse))`,"border-left-width":`calc(${i} * calc(1 - var(--tw-divide-x-reverse)))`}}),"divide-y":i=>(i=i==="0"?"0px":i,{"& > :not([hidden]) ~ :not([hidden])":{"@defaults border-width":{},"--tw-divide-y-reverse":"0","border-top-width":`calc(${i} * calc(1 - var(--tw-divide-y-reverse)))`,"border-bottom-width":`calc(${i} * var(--tw-divide-y-reverse))`}})},{values:t("divideWidth"),type:["line-width","length","any"]}),e({".divide-y-reverse > :not([hidden]) ~ :not([hidden])":{"@defaults border-width":{},"--tw-divide-y-reverse":"1"},".divide-x-reverse > :not([hidden]) ~ :not([hidden])":{"@defaults border-width":{},"--tw-divide-x-reverse":"1"}})},divideStyle:({addUtilities:r})=>{r({".divide-solid > :not([hidden]) ~ :not([hidden])":{"border-style":"solid"},".divide-dashed > :not([hidden]) ~ :not([hidden])":{"border-style":"dashed"},".divide-dotted > :not([hidden]) ~ :not([hidden])":{"border-style":"dotted"},".divide-double > :not([hidden]) ~ :not([hidden])":{"border-style":"double"},".divide-none > :not([hidden]) ~ :not([hidden])":{"border-style":"none"}})},divideColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({divide:i=>t("divideOpacity")?{["& > :not([hidden]) ~ :not([hidden])"]:Ae({color:i,property:"border-color",variable:"--tw-divide-opacity"})}:{["& > :not([hidden]) ~ :not([hidden])"]:{"border-color":X(i)}}},{values:(({DEFAULT:i,...n})=>n)(xe(e("divideColor"))),type:["color","any"]})},divideOpacity:({matchUtilities:r,theme:e})=>{r({"divide-opacity":t=>({["& > :not([hidden]) ~ :not([hidden])"]:{"--tw-divide-opacity":t}})},{values:e("divideOpacity")})},placeSelf:({addUtilities:r})=>{r({".place-self-auto":{"place-self":"auto"},".place-self-start":{"place-self":"start"},".place-self-end":{"place-self":"end"},".place-self-center":{"place-self":"center"},".place-self-stretch":{"place-self":"stretch"}})},alignSelf:({addUtilities:r})=>{r({".self-auto":{"align-self":"auto"},".self-start":{"align-self":"flex-start"},".self-end":{"align-self":"flex-end"},".self-center":{"align-self":"center"},".self-stretch":{"align-self":"stretch"},".self-baseline":{"align-self":"baseline"}})},justifySelf:({addUtilities:r})=>{r({".justify-self-auto":{"justify-self":"auto"},".justify-self-start":{"justify-self":"start"},".justify-self-end":{"justify-self":"end"},".justify-self-center":{"justify-self":"center"},".justify-self-stretch":{"justify-self":"stretch"}})},overflow:({addUtilities:r})=>{r({".overflow-auto":{overflow:"auto"},".overflow-hidden":{overflow:"hidden"},".overflow-clip":{overflow:"clip"},".overflow-visible":{overflow:"visible"},".overflow-scroll":{overflow:"scroll"},".overflow-x-auto":{"overflow-x":"auto"},".overflow-y-auto":{"overflow-y":"auto"},".overflow-x-hidden":{"overflow-x":"hidden"},".overflow-y-hidden":{"overflow-y":"hidden"},".overflow-x-clip":{"overflow-x":"clip"},".overflow-y-clip":{"overflow-y":"clip"},".overflow-x-visible":{"overflow-x":"visible"},".overflow-y-visible":{"overflow-y":"visible"},".overflow-x-scroll":{"overflow-x":"scroll"},".overflow-y-scroll":{"overflow-y":"scroll"}})},overscrollBehavior:({addUtilities:r})=>{r({".overscroll-auto":{"overscroll-behavior":"auto"},".overscroll-contain":{"overscroll-behavior":"contain"},".overscroll-none":{"overscroll-behavior":"none"},".overscroll-y-auto":{"overscroll-behavior-y":"auto"},".overscroll-y-contain":{"overscroll-behavior-y":"contain"},".overscroll-y-none":{"overscroll-behavior-y":"none"},".overscroll-x-auto":{"overscroll-behavior-x":"auto"},".overscroll-x-contain":{"overscroll-behavior-x":"contain"},".overscroll-x-none":{"overscroll-behavior-x":"none"}})},scrollBehavior:({addUtilities:r})=>{r({".scroll-auto":{"scroll-behavior":"auto"},".scroll-smooth":{"scroll-behavior":"smooth"}})},textOverflow:({addUtilities:r})=>{r({".truncate":{overflow:"hidden","text-overflow":"ellipsis","white-space":"nowrap"},".overflow-ellipsis":{"text-overflow":"ellipsis"},".text-ellipsis":{"text-overflow":"ellipsis"},".text-clip":{"text-overflow":"clip"}})},hyphens:({addUtilities:r})=>{r({".hyphens-none":{hyphens:"none"},".hyphens-manual":{hyphens:"manual"},".hyphens-auto":{hyphens:"auto"}})},whitespace:({addUtilities:r})=>{r({".whitespace-normal":{"white-space":"normal"},".whitespace-nowrap":{"white-space":"nowrap"},".whitespace-pre":{"white-space":"pre"},".whitespace-pre-line":{"white-space":"pre-line"},".whitespace-pre-wrap":{"white-space":"pre-wrap"},".whitespace-break-spaces":{"white-space":"break-spaces"}})},textWrap:({addUtilities:r})=>{r({".text-wrap":{"text-wrap":"wrap"},".text-nowrap":{"text-wrap":"nowrap"},".text-balance":{"text-wrap":"balance"},".text-pretty":{"text-wrap":"pretty"}})},wordBreak:({addUtilities:r})=>{r({".break-normal":{"overflow-wrap":"normal","word-break":"normal"},".break-words":{"overflow-wrap":"break-word"},".break-all":{"word-break":"break-all"},".break-keep":{"word-break":"keep-all"}})},borderRadius:L("borderRadius",[["rounded",["border-radius"]],[["rounded-s",["border-start-start-radius","border-end-start-radius"]],["rounded-e",["border-start-end-radius","border-end-end-radius"]],["rounded-t",["border-top-left-radius","border-top-right-radius"]],["rounded-r",["border-top-right-radius","border-bottom-right-radius"]],["rounded-b",["border-bottom-right-radius","border-bottom-left-radius"]],["rounded-l",["border-top-left-radius","border-bottom-left-radius"]]],[["rounded-ss",["border-start-start-radius"]],["rounded-se",["border-start-end-radius"]],["rounded-ee",["border-end-end-radius"]],["rounded-es",["border-end-start-radius"]],["rounded-tl",["border-top-left-radius"]],["rounded-tr",["border-top-right-radius"]],["rounded-br",["border-bottom-right-radius"]],["rounded-bl",["border-bottom-left-radius"]]]]),borderWidth:L("borderWidth",[["border",[["@defaults border-width",{}],"border-width"]],[["border-x",[["@defaults border-width",{}],"border-left-width","border-right-width"]],["border-y",[["@defaults border-width",{}],"border-top-width","border-bottom-width"]]],[["border-s",[["@defaults border-width",{}],"border-inline-start-width"]],["border-e",[["@defaults border-width",{}],"border-inline-end-width"]],["border-t",[["@defaults border-width",{}],"border-top-width"]],["border-r",[["@defaults border-width",{}],"border-right-width"]],["border-b",[["@defaults border-width",{}],"border-bottom-width"]],["border-l",[["@defaults border-width",{}],"border-left-width"]]]],{type:["line-width","length"]}),borderStyle:({addUtilities:r})=>{r({".border-solid":{"border-style":"solid"},".border-dashed":{"border-style":"dashed"},".border-dotted":{"border-style":"dotted"},".border-double":{"border-style":"double"},".border-hidden":{"border-style":"hidden"},".border-none":{"border-style":"none"}})},borderColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({border:i=>t("borderOpacity")?Ae({color:i,property:"border-color",variable:"--tw-border-opacity"}):{"border-color":X(i)}},{values:(({DEFAULT:i,...n})=>n)(xe(e("borderColor"))),type:["color","any"]}),r({"border-x":i=>t("borderOpacity")?Ae({color:i,property:["border-left-color","border-right-color"],variable:"--tw-border-opacity"}):{"border-left-color":X(i),"border-right-color":X(i)},"border-y":i=>t("borderOpacity")?Ae({color:i,property:["border-top-color","border-bottom-color"],variable:"--tw-border-opacity"}):{"border-top-color":X(i),"border-bottom-color":X(i)}},{values:(({DEFAULT:i,...n})=>n)(xe(e("borderColor"))),type:["color","any"]}),r({"border-s":i=>t("borderOpacity")?Ae({color:i,property:"border-inline-start-color",variable:"--tw-border-opacity"}):{"border-inline-start-color":X(i)},"border-e":i=>t("borderOpacity")?Ae({color:i,property:"border-inline-end-color",variable:"--tw-border-opacity"}):{"border-inline-end-color":X(i)},"border-t":i=>t("borderOpacity")?Ae({color:i,property:"border-top-color",variable:"--tw-border-opacity"}):{"border-top-color":X(i)},"border-r":i=>t("borderOpacity")?Ae({color:i,property:"border-right-color",variable:"--tw-border-opacity"}):{"border-right-color":X(i)},"border-b":i=>t("borderOpacity")?Ae({color:i,property:"border-bottom-color",variable:"--tw-border-opacity"}):{"border-bottom-color":X(i)},"border-l":i=>t("borderOpacity")?Ae({color:i,property:"border-left-color",variable:"--tw-border-opacity"}):{"border-left-color":X(i)}},{values:(({DEFAULT:i,...n})=>n)(xe(e("borderColor"))),type:["color","any"]})},borderOpacity:L("borderOpacity",[["border-opacity",["--tw-border-opacity"]]]),backgroundColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({bg:i=>t("backgroundOpacity")?Ae({color:i,property:"background-color",variable:"--tw-bg-opacity"}):{"background-color":X(i)}},{values:xe(e("backgroundColor")),type:["color","any"]})},backgroundOpacity:L("backgroundOpacity",[["bg-opacity",["--tw-bg-opacity"]]]),backgroundImage:L("backgroundImage",[["bg",["background-image"]]],{type:["lookup","image","url"]}),gradientColorStops:(()=>{function r(e){return Ze(e,0,"rgb(255 255 255 / 0)")}return function({matchUtilities:e,theme:t,addDefaults:i}){i("gradient-color-stops",{"--tw-gradient-from-position":" ","--tw-gradient-via-position":" ","--tw-gradient-to-position":" "});let n={values:xe(t("gradientColorStops")),type:["color","any"]},a={values:t("gradientColorStopPositions"),type:["length","percentage"]};e({from:s=>{let o=r(s);return{"@defaults gradient-color-stops":{},"--tw-gradient-from":`${X(s)} var(--tw-gradient-from-position)`,"--tw-gradient-to":`${o} var(--tw-gradient-to-position)`,"--tw-gradient-stops":"var(--tw-gradient-from), var(--tw-gradient-to)"}}},n),e({from:s=>({"--tw-gradient-from-position":s})},a),e({via:s=>{let o=r(s);return{"@defaults gradient-color-stops":{},"--tw-gradient-to":`${o} var(--tw-gradient-to-position)`,"--tw-gradient-stops":`var(--tw-gradient-from), ${X(s)} var(--tw-gradient-via-position), var(--tw-gradient-to)`}}},n),e({via:s=>({"--tw-gradient-via-position":s})},a),e({to:s=>({"@defaults gradient-color-stops":{},"--tw-gradient-to":`${X(s)} var(--tw-gradient-to-position)`})},n),e({to:s=>({"--tw-gradient-to-position":s})},a)}})(),boxDecorationBreak:({addUtilities:r})=>{r({".decoration-slice":{"box-decoration-break":"slice"},".decoration-clone":{"box-decoration-break":"clone"},".box-decoration-slice":{"box-decoration-break":"slice"},".box-decoration-clone":{"box-decoration-break":"clone"}})},backgroundSize:L("backgroundSize",[["bg",["background-size"]]],{type:["lookup","length","percentage","size"]}),backgroundAttachment:({addUtilities:r})=>{r({".bg-fixed":{"background-attachment":"fixed"},".bg-local":{"background-attachment":"local"},".bg-scroll":{"background-attachment":"scroll"}})},backgroundClip:({addUtilities:r})=>{r({".bg-clip-border":{"background-clip":"border-box"},".bg-clip-padding":{"background-clip":"padding-box"},".bg-clip-content":{"background-clip":"content-box"},".bg-clip-text":{"background-clip":"text"}})},backgroundPosition:L("backgroundPosition",[["bg",["background-position"]]],{type:["lookup",["position",{preferOnConflict:!0}]]}),backgroundRepeat:({addUtilities:r})=>{r({".bg-repeat":{"background-repeat":"repeat"},".bg-no-repeat":{"background-repeat":"no-repeat"},".bg-repeat-x":{"background-repeat":"repeat-x"},".bg-repeat-y":{"background-repeat":"repeat-y"},".bg-repeat-round":{"background-repeat":"round"},".bg-repeat-space":{"background-repeat":"space"}})},backgroundOrigin:({addUtilities:r})=>{r({".bg-origin-border":{"background-origin":"border-box"},".bg-origin-padding":{"background-origin":"padding-box"},".bg-origin-content":{"background-origin":"content-box"}})},fill:({matchUtilities:r,theme:e})=>{r({fill:t=>({fill:X(t)})},{values:xe(e("fill")),type:["color","any"]})},stroke:({matchUtilities:r,theme:e})=>{r({stroke:t=>({stroke:X(t)})},{values:xe(e("stroke")),type:["color","url","any"]})},strokeWidth:L("strokeWidth",[["stroke",["stroke-width"]]],{type:["length","number","percentage"]}),objectFit:({addUtilities:r})=>{r({".object-contain":{"object-fit":"contain"},".object-cover":{"object-fit":"cover"},".object-fill":{"object-fit":"fill"},".object-none":{"object-fit":"none"},".object-scale-down":{"object-fit":"scale-down"}})},objectPosition:L("objectPosition",[["object",["object-position"]]]),padding:L("padding",[["p",["padding"]],[["px",["padding-left","padding-right"]],["py",["padding-top","padding-bottom"]]],[["ps",["padding-inline-start"]],["pe",["padding-inline-end"]],["pt",["padding-top"]],["pr",["padding-right"]],["pb",["padding-bottom"]],["pl",["padding-left"]]]]),textAlign:({addUtilities:r})=>{r({".text-left":{"text-align":"left"},".text-center":{"text-align":"center"},".text-right":{"text-align":"right"},".text-justify":{"text-align":"justify"},".text-start":{"text-align":"start"},".text-end":{"text-align":"end"}})},textIndent:L("textIndent",[["indent",["text-indent"]]],{supportsNegativeValues:!0}),verticalAlign:({addUtilities:r,matchUtilities:e})=>{r({".align-baseline":{"vertical-align":"baseline"},".align-top":{"vertical-align":"top"},".align-middle":{"vertical-align":"middle"},".align-bottom":{"vertical-align":"bottom"},".align-text-top":{"vertical-align":"text-top"},".align-text-bottom":{"vertical-align":"text-bottom"},".align-sub":{"vertical-align":"sub"},".align-super":{"vertical-align":"super"}}),e({align:t=>({"vertical-align":t})})},fontFamily:({matchUtilities:r,theme:e})=>{r({font:t=>{let[i,n={}]=Array.isArray(t)&&ke(t[1])?t:[t],{fontFeatureSettings:a,fontVariationSettings:s}=n;return{"font-family":Array.isArray(i)?i.join(", "):i,...a===void 0?{}:{"font-feature-settings":a},...s===void 0?{}:{"font-variation-settings":s}}}},{values:e("fontFamily"),type:["lookup","generic-name","family-name"]})},fontSize:({matchUtilities:r,theme:e})=>{r({text:(t,{modifier:i})=>{let[n,a]=Array.isArray(t)?t:[t];if(i)return{"font-size":n,"line-height":i};let{lineHeight:s,letterSpacing:o,fontWeight:l}=ke(a)?a:{lineHeight:a};return{"font-size":n,...s===void 0?{}:{"line-height":s},...o===void 0?{}:{"letter-spacing":o},...l===void 0?{}:{"font-weight":l}}}},{values:e("fontSize"),modifiers:e("lineHeight"),type:["absolute-size","relative-size","length","percentage"]})},fontWeight:L("fontWeight",[["font",["fontWeight"]]],{type:["lookup","number","any"]}),textTransform:({addUtilities:r})=>{r({".uppercase":{"text-transform":"uppercase"},".lowercase":{"text-transform":"lowercase"},".capitalize":{"text-transform":"capitalize"},".normal-case":{"text-transform":"none"}})},fontStyle:({addUtilities:r})=>{r({".italic":{"font-style":"italic"},".not-italic":{"font-style":"normal"}})},fontVariantNumeric:({addDefaults:r,addUtilities:e})=>{let t="var(--tw-ordinal) var(--tw-slashed-zero) var(--tw-numeric-figure) var(--tw-numeric-spacing) var(--tw-numeric-fraction)";r("font-variant-numeric",{"--tw-ordinal":" ","--tw-slashed-zero":" ","--tw-numeric-figure":" ","--tw-numeric-spacing":" ","--tw-numeric-fraction":" "}),e({".normal-nums":{"font-variant-numeric":"normal"},".ordinal":{"@defaults font-variant-numeric":{},"--tw-ordinal":"ordinal","font-variant-numeric":t},".slashed-zero":{"@defaults font-variant-numeric":{},"--tw-slashed-zero":"slashed-zero","font-variant-numeric":t},".lining-nums":{"@defaults font-variant-numeric":{},"--tw-numeric-figure":"lining-nums","font-variant-numeric":t},".oldstyle-nums":{"@defaults font-variant-numeric":{},"--tw-numeric-figure":"oldstyle-nums","font-variant-numeric":t},".proportional-nums":{"@defaults font-variant-numeric":{},"--tw-numeric-spacing":"proportional-nums","font-variant-numeric":t},".tabular-nums":{"@defaults font-variant-numeric":{},"--tw-numeric-spacing":"tabular-nums","font-variant-numeric":t},".diagonal-fractions":{"@defaults font-variant-numeric":{},"--tw-numeric-fraction":"diagonal-fractions","font-variant-numeric":t},".stacked-fractions":{"@defaults font-variant-numeric":{},"--tw-numeric-fraction":"stacked-fractions","font-variant-numeric":t}})},lineHeight:L("lineHeight",[["leading",["lineHeight"]]]),letterSpacing:L("letterSpacing",[["tracking",["letterSpacing"]]],{supportsNegativeValues:!0}),textColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({text:i=>t("textOpacity")?Ae({color:i,property:"color",variable:"--tw-text-opacity"}):{color:X(i)}},{values:xe(e("textColor")),type:["color","any"]})},textOpacity:L("textOpacity",[["text-opacity",["--tw-text-opacity"]]]),textDecoration:({addUtilities:r})=>{r({".underline":{"text-decoration-line":"underline"},".overline":{"text-decoration-line":"overline"},".line-through":{"text-decoration-line":"line-through"},".no-underline":{"text-decoration-line":"none"}})},textDecorationColor:({matchUtilities:r,theme:e})=>{r({decoration:t=>({"text-decoration-color":X(t)})},{values:xe(e("textDecorationColor")),type:["color","any"]})},textDecorationStyle:({addUtilities:r})=>{r({".decoration-solid":{"text-decoration-style":"solid"},".decoration-double":{"text-decoration-style":"double"},".decoration-dotted":{"text-decoration-style":"dotted"},".decoration-dashed":{"text-decoration-style":"dashed"},".decoration-wavy":{"text-decoration-style":"wavy"}})},textDecorationThickness:L("textDecorationThickness",[["decoration",["text-decoration-thickness"]]],{type:["length","percentage"]}),textUnderlineOffset:L("textUnderlineOffset",[["underline-offset",["text-underline-offset"]]],{type:["length","percentage","any"]}),fontSmoothing:({addUtilities:r})=>{r({".antialiased":{"-webkit-font-smoothing":"antialiased","-moz-osx-font-smoothing":"grayscale"},".subpixel-antialiased":{"-webkit-font-smoothing":"auto","-moz-osx-font-smoothing":"auto"}})},placeholderColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({placeholder:i=>t("placeholderOpacity")?{"&::placeholder":Ae({color:i,property:"color",variable:"--tw-placeholder-opacity"})}:{"&::placeholder":{color:X(i)}}},{values:xe(e("placeholderColor")),type:["color","any"]})},placeholderOpacity:({matchUtilities:r,theme:e})=>{r({"placeholder-opacity":t=>({["&::placeholder"]:{"--tw-placeholder-opacity":t}})},{values:e("placeholderOpacity")})},caretColor:({matchUtilities:r,theme:e})=>{r({caret:t=>({"caret-color":X(t)})},{values:xe(e("caretColor")),type:["color","any"]})},accentColor:({matchUtilities:r,theme:e})=>{r({accent:t=>({"accent-color":X(t)})},{values:xe(e("accentColor")),type:["color","any"]})},opacity:L("opacity",[["opacity",["opacity"]]]),backgroundBlendMode:({addUtilities:r})=>{r({".bg-blend-normal":{"background-blend-mode":"normal"},".bg-blend-multiply":{"background-blend-mode":"multiply"},".bg-blend-screen":{"background-blend-mode":"screen"},".bg-blend-overlay":{"background-blend-mode":"overlay"},".bg-blend-darken":{"background-blend-mode":"darken"},".bg-blend-lighten":{"background-blend-mode":"lighten"},".bg-blend-color-dodge":{"background-blend-mode":"color-dodge"},".bg-blend-color-burn":{"background-blend-mode":"color-burn"},".bg-blend-hard-light":{"background-blend-mode":"hard-light"},".bg-blend-soft-light":{"background-blend-mode":"soft-light"},".bg-blend-difference":{"background-blend-mode":"difference"},".bg-blend-exclusion":{"background-blend-mode":"exclusion"},".bg-blend-hue":{"background-blend-mode":"hue"},".bg-blend-saturation":{"background-blend-mode":"saturation"},".bg-blend-color":{"background-blend-mode":"color"},".bg-blend-luminosity":{"background-blend-mode":"luminosity"}})},mixBlendMode:({addUtilities:r})=>{r({".mix-blend-normal":{"mix-blend-mode":"normal"},".mix-blend-multiply":{"mix-blend-mode":"multiply"},".mix-blend-screen":{"mix-blend-mode":"screen"},".mix-blend-overlay":{"mix-blend-mode":"overlay"},".mix-blend-darken":{"mix-blend-mode":"darken"},".mix-blend-lighten":{"mix-blend-mode":"lighten"},".mix-blend-color-dodge":{"mix-blend-mode":"color-dodge"},".mix-blend-color-burn":{"mix-blend-mode":"color-burn"},".mix-blend-hard-light":{"mix-blend-mode":"hard-light"},".mix-blend-soft-light":{"mix-blend-mode":"soft-light"},".mix-blend-difference":{"mix-blend-mode":"difference"},".mix-blend-exclusion":{"mix-blend-mode":"exclusion"},".mix-blend-hue":{"mix-blend-mode":"hue"},".mix-blend-saturation":{"mix-blend-mode":"saturation"},".mix-blend-color":{"mix-blend-mode":"color"},".mix-blend-luminosity":{"mix-blend-mode":"luminosity"},".mix-blend-plus-darker":{"mix-blend-mode":"plus-darker"},".mix-blend-plus-lighter":{"mix-blend-mode":"plus-lighter"}})},boxShadow:(()=>{let r=mt("boxShadow"),e=["var(--tw-ring-offset-shadow, 0 0 #0000)","var(--tw-ring-shadow, 0 0 #0000)","var(--tw-shadow)"].join(", ");return function({matchUtilities:t,addDefaults:i,theme:n}){i("box-shadow",{"--tw-ring-offset-shadow":"0 0 #0000","--tw-ring-shadow":"0 0 #0000","--tw-shadow":"0 0 #0000","--tw-shadow-colored":"0 0 #0000"}),t({shadow:a=>{a=r(a);let s=Ji(a);for(let o of s)!o.valid||(o.color="var(--tw-shadow-color)");return{"@defaults box-shadow":{},"--tw-shadow":a==="none"?"0 0 #0000":a,"--tw-shadow-colored":a==="none"?"0 0 #0000":qf(s),"box-shadow":e}}},{values:n("boxShadow"),type:["shadow"]})}})(),boxShadowColor:({matchUtilities:r,theme:e})=>{r({shadow:t=>({"--tw-shadow-color":X(t),"--tw-shadow":"var(--tw-shadow-colored)"})},{values:xe(e("boxShadowColor")),type:["color","any"]})},outlineStyle:({addUtilities:r})=>{r({".outline-none":{outline:"2px solid transparent","outline-offset":"2px"},".outline":{"outline-style":"solid"},".outline-dashed":{"outline-style":"dashed"},".outline-dotted":{"outline-style":"dotted"},".outline-double":{"outline-style":"double"}})},outlineWidth:L("outlineWidth",[["outline",["outline-width"]]],{type:["length","number","percentage"]}),outlineOffset:L("outlineOffset",[["outline-offset",["outline-offset"]]],{type:["length","number","percentage","any"],supportsNegativeValues:!0}),outlineColor:({matchUtilities:r,theme:e})=>{r({outline:t=>({"outline-color":X(t)})},{values:xe(e("outlineColor")),type:["color","any"]})},ringWidth:({matchUtilities:r,addDefaults:e,addUtilities:t,theme:i,config:n})=>{let a=(()=>{if(we(n(),"respectDefaultRingColorOpacity"))return i("ringColor.DEFAULT");let s=i("ringOpacity.DEFAULT","0.5");return i("ringColor")?.DEFAULT?Ze(i("ringColor")?.DEFAULT,s,`rgb(147 197 253 / ${s})`):`rgb(147 197 253 / ${s})`})();e("ring-width",{"--tw-ring-inset":" ","--tw-ring-offset-width":i("ringOffsetWidth.DEFAULT","0px"),"--tw-ring-offset-color":i("ringOffsetColor.DEFAULT","#fff"),"--tw-ring-color":a,"--tw-ring-offset-shadow":"0 0 #0000","--tw-ring-shadow":"0 0 #0000","--tw-shadow":"0 0 #0000","--tw-shadow-colored":"0 0 #0000"}),r({ring:s=>({"@defaults ring-width":{},"--tw-ring-offset-shadow":"var(--tw-ring-inset) 0 0 0 var(--tw-ring-offset-width) var(--tw-ring-offset-color)","--tw-ring-shadow":`var(--tw-ring-inset) 0 0 0 calc(${s} + var(--tw-ring-offset-width)) var(--tw-ring-color)`,"box-shadow":["var(--tw-ring-offset-shadow)","var(--tw-ring-shadow)","var(--tw-shadow, 0 0 #0000)"].join(", ")})},{values:i("ringWidth"),type:"length"}),t({".ring-inset":{"@defaults ring-width":{},"--tw-ring-inset":"inset"}})},ringColor:({matchUtilities:r,theme:e,corePlugins:t})=>{r({ring:i=>t("ringOpacity")?Ae({color:i,property:"--tw-ring-color",variable:"--tw-ring-opacity"}):{"--tw-ring-color":X(i)}},{values:Object.fromEntries(Object.entries(xe(e("ringColor"))).filter(([i])=>i!=="DEFAULT")),type:["color","any"]})},ringOpacity:r=>{let{config:e}=r;return L("ringOpacity",[["ring-opacity",["--tw-ring-opacity"]]],{filterDefault:!we(e(),"respectDefaultRingColorOpacity")})(r)},ringOffsetWidth:L("ringOffsetWidth",[["ring-offset",["--tw-ring-offset-width"]]],{type:"length"}),ringOffsetColor:({matchUtilities:r,theme:e})=>{r({"ring-offset":t=>({"--tw-ring-offset-color":X(t)})},{values:xe(e("ringOffsetColor")),type:["color","any"]})},blur:({matchUtilities:r,theme:e})=>{r({blur:t=>({"--tw-blur":t.trim()===""?" ":`blur(${t})`,"@defaults filter":{},filter:nt})},{values:e("blur")})},brightness:({matchUtilities:r,theme:e})=>{r({brightness:t=>({"--tw-brightness":`brightness(${t})`,"@defaults filter":{},filter:nt})},{values:e("brightness")})},contrast:({matchUtilities:r,theme:e})=>{r({contrast:t=>({"--tw-contrast":`contrast(${t})`,"@defaults filter":{},filter:nt})},{values:e("contrast")})},dropShadow:({matchUtilities:r,theme:e})=>{r({"drop-shadow":t=>({"--tw-drop-shadow":Array.isArray(t)?t.map(i=>`drop-shadow(${i})`).join(" "):`drop-shadow(${t})`,"@defaults filter":{},filter:nt})},{values:e("dropShadow")})},grayscale:({matchUtilities:r,theme:e})=>{r({grayscale:t=>({"--tw-grayscale":`grayscale(${t})`,"@defaults filter":{},filter:nt})},{values:e("grayscale")})},hueRotate:({matchUtilities:r,theme:e})=>{r({"hue-rotate":t=>({"--tw-hue-rotate":`hue-rotate(${t})`,"@defaults filter":{},filter:nt})},{values:e("hueRotate"),supportsNegativeValues:!0})},invert:({matchUtilities:r,theme:e})=>{r({invert:t=>({"--tw-invert":`invert(${t})`,"@defaults filter":{},filter:nt})},{values:e("invert")})},saturate:({matchUtilities:r,theme:e})=>{r({saturate:t=>({"--tw-saturate":`saturate(${t})`,"@defaults filter":{},filter:nt})},{values:e("saturate")})},sepia:({matchUtilities:r,theme:e})=>{r({sepia:t=>({"--tw-sepia":`sepia(${t})`,"@defaults filter":{},filter:nt})},{values:e("sepia")})},filter:({addDefaults:r,addUtilities:e})=>{r("filter",{"--tw-blur":" ","--tw-brightness":" ","--tw-contrast":" ","--tw-grayscale":" ","--tw-hue-rotate":" ","--tw-invert":" ","--tw-saturate":" ","--tw-sepia":" ","--tw-drop-shadow":" "}),e({".filter":{"@defaults filter":{},filter:nt},".filter-none":{filter:"none"}})},backdropBlur:({matchUtilities:r,theme:e})=>{r({"backdrop-blur":t=>({"--tw-backdrop-blur":t.trim()===""?" ":`blur(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropBlur")})},backdropBrightness:({matchUtilities:r,theme:e})=>{r({"backdrop-brightness":t=>({"--tw-backdrop-brightness":`brightness(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropBrightness")})},backdropContrast:({matchUtilities:r,theme:e})=>{r({"backdrop-contrast":t=>({"--tw-backdrop-contrast":`contrast(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropContrast")})},backdropGrayscale:({matchUtilities:r,theme:e})=>{r({"backdrop-grayscale":t=>({"--tw-backdrop-grayscale":`grayscale(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropGrayscale")})},backdropHueRotate:({matchUtilities:r,theme:e})=>{r({"backdrop-hue-rotate":t=>({"--tw-backdrop-hue-rotate":`hue-rotate(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropHueRotate"),supportsNegativeValues:!0})},backdropInvert:({matchUtilities:r,theme:e})=>{r({"backdrop-invert":t=>({"--tw-backdrop-invert":`invert(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropInvert")})},backdropOpacity:({matchUtilities:r,theme:e})=>{r({"backdrop-opacity":t=>({"--tw-backdrop-opacity":`opacity(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropOpacity")})},backdropSaturate:({matchUtilities:r,theme:e})=>{r({"backdrop-saturate":t=>({"--tw-backdrop-saturate":`saturate(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropSaturate")})},backdropSepia:({matchUtilities:r,theme:e})=>{r({"backdrop-sepia":t=>({"--tw-backdrop-sepia":`sepia(${t})`,"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge})},{values:e("backdropSepia")})},backdropFilter:({addDefaults:r,addUtilities:e})=>{r("backdrop-filter",{"--tw-backdrop-blur":" ","--tw-backdrop-brightness":" ","--tw-backdrop-contrast":" ","--tw-backdrop-grayscale":" ","--tw-backdrop-hue-rotate":" ","--tw-backdrop-invert":" ","--tw-backdrop-opacity":" ","--tw-backdrop-saturate":" ","--tw-backdrop-sepia":" "}),e({".backdrop-filter":{"@defaults backdrop-filter":{},"-webkit-backdrop-filter":ge,"backdrop-filter":ge},".backdrop-filter-none":{"-webkit-backdrop-filter":"none","backdrop-filter":"none"}})},transitionProperty:({matchUtilities:r,theme:e})=>{let t=e("transitionTimingFunction.DEFAULT"),i=e("transitionDuration.DEFAULT");r({transition:n=>({"transition-property":n,...n==="none"?{}:{"transition-timing-function":t,"transition-duration":i}})},{values:e("transitionProperty")})},transitionDelay:L("transitionDelay",[["delay",["transitionDelay"]]]),transitionDuration:L("transitionDuration",[["duration",["transitionDuration"]]],{filterDefault:!0}),transitionTimingFunction:L("transitionTimingFunction",[["ease",["transitionTimingFunction"]]],{filterDefault:!0}),willChange:L("willChange",[["will-change",["will-change"]]]),contain:({addDefaults:r,addUtilities:e})=>{let t="var(--tw-contain-size) var(--tw-contain-layout) var(--tw-contain-paint) var(--tw-contain-style)";r("contain",{"--tw-contain-size":" ","--tw-contain-layout":" ","--tw-contain-paint":" ","--tw-contain-style":" "}),e({".contain-none":{contain:"none"},".contain-content":{contain:"content"},".contain-strict":{contain:"strict"},".contain-size":{"@defaults contain":{},"--tw-contain-size":"size",contain:t},".contain-inline-size":{"@defaults contain":{},"--tw-contain-size":"inline-size",contain:t},".contain-layout":{"@defaults contain":{},"--tw-contain-layout":"layout",contain:t},".contain-paint":{"@defaults contain":{},"--tw-contain-paint":"paint",contain:t},".contain-style":{"@defaults contain":{},"--tw-contain-style":"style",contain:t}})},content:L("content",[["content",["--tw-content",["content","var(--tw-content)"]]]]),forcedColorAdjust:({addUtilities:r})=>{r({".forced-color-adjust-auto":{"forced-color-adjust":"auto"},".forced-color-adjust-none":{"forced-color-adjust":"none"}})}}});function p_(r){if(r===void 0)return!1;if(r==="true"||r==="1")return!0;if(r==="false"||r==="0")return!1;if(r==="*")return!0;let e=r.split(",").map(t=>t.split(":")[0]);return e.includes("-tailwindcss")?!1:!!e.includes("tailwindcss")}var Je,yh,bh,Zn,Lo,gt,Ei,It=R(()=>{u();Je=typeof m!="undefined"?{NODE_ENV:"production",DEBUG:p_(m.env.DEBUG)}:{NODE_ENV:"production",DEBUG:!1},yh=new Map,bh=new Map,Zn=new Map,Lo=new Map,gt=new String("*"),Ei=Symbol("__NONE__")});function cr(r){let e=[],t=!1;for(let i=0;i0)}var wh,vh,d_,Mo=R(()=>{u();wh=new Map([["{","}"],["[","]"],["(",")"]]),vh=new Map(Array.from(wh.entries()).map(([r,e])=>[e,r])),d_=new Set(['"',"'","`"])});function pr(r){let[e]=xh(r);return e.forEach(([t,i])=>t.removeChild(i)),r.nodes.push(...e.map(([,t])=>t)),r}function xh(r){let e=[],t=null;for(let i of r.nodes)if(i.type==="combinator")e=e.filter(([,n])=>Bo(n).includes("jumpable")),t=null;else if(i.type==="pseudo"){h_(i)?(t=i,e.push([r,i,null])):t&&m_(i,t)?e.push([r,i,t]):t=null;for(let n of i.nodes??[]){let[a,s]=xh(n);t=s||t,e.push(...a)}}return[e,t]}function kh(r){return r.value.startsWith("::")||No[r.value]!==void 0}function h_(r){return kh(r)&&Bo(r).includes("terminal")}function m_(r,e){return r.type!=="pseudo"||kh(r)?!1:Bo(e).includes("actionable")}function Bo(r){return No[r.value]??No.__default__}var No,es=R(()=>{u();No={"::after":["terminal","jumpable"],"::backdrop":["terminal","jumpable"],"::before":["terminal","jumpable"],"::cue":["terminal"],"::cue-region":["terminal"],"::first-letter":["terminal","jumpable"],"::first-line":["terminal","jumpable"],"::grammar-error":["terminal"],"::marker":["terminal","jumpable"],"::part":["terminal","actionable"],"::placeholder":["terminal","jumpable"],"::selection":["terminal","jumpable"],"::slotted":["terminal"],"::spelling-error":["terminal"],"::target-text":["terminal"],"::file-selector-button":["terminal","actionable"],"::deep":["actionable"],"::v-deep":["actionable"],"::ng-deep":["actionable"],":after":["terminal","jumpable"],":before":["terminal","jumpable"],":first-letter":["terminal","jumpable"],":first-line":["terminal","jumpable"],":where":[],":is":[],":has":[],__default__:["terminal","actionable"]}});function dr(r,{context:e,candidate:t}){let i=e?.tailwindConfig.prefix??"",n=r.map(s=>{let o=(0,st.default)().astSync(s.format);return{...s,ast:s.respectPrefix?ur(i,o):o}}),a=st.default.root({nodes:[st.default.selector({nodes:[st.default.className({value:Te(t)})]})]});for(let{ast:s}of n)[a,s]=y_(a,s),s.walkNesting(o=>o.replaceWith(...a.nodes[0].nodes)),a=s;return a}function Ah(r){let e=[];for(;r.prev()&&r.prev().type!=="combinator";)r=r.prev();for(;r&&r.type!=="combinator";)e.push(r),r=r.next();return e}function g_(r){return r.sort((e,t)=>e.type==="tag"&&t.type==="class"?-1:e.type==="class"&&t.type==="tag"?1:e.type==="class"&&t.type==="pseudo"&&t.value.startsWith("::")?-1:e.type==="pseudo"&&e.value.startsWith("::")&&t.type==="class"?1:r.index(e)-r.index(t)),r}function jo(r,e){let t=!1;r.walk(i=>{if(i.type==="class"&&i.value===e)return t=!0,!1}),t||r.remove()}function ts(r,e,{context:t,candidate:i,base:n}){let a=t?.tailwindConfig?.separator??":";n=n??ve(i,a).pop();let s=(0,st.default)().astSync(r);if(s.walkClasses(f=>{f.raws&&f.value.includes(n)&&(f.raws.value=Te((0,Sh.default)(f.raws.value)))}),s.each(f=>jo(f,n)),s.length===0)return null;let o=Array.isArray(e)?dr(e,{context:t,candidate:i}):e;if(o===null)return s.toString();let l=st.default.comment({value:"/*__simple__*/"}),c=st.default.comment({value:"/*__simple__*/"});return s.walkClasses(f=>{if(f.value!==n)return;let d=f.parent,p=o.nodes[0].nodes;if(d.nodes.length===1){f.replaceWith(...p);return}let h=Ah(f);d.insertBefore(h[0],l),d.insertAfter(h[h.length-1],c);for(let v of p)d.insertBefore(h[0],v.clone());f.remove(),h=Ah(l);let b=d.index(l);d.nodes.splice(b,h.length,...g_(st.default.selector({nodes:h})).nodes),l.remove(),c.remove()}),s.walkPseudos(f=>{f.value===Fo&&f.replaceWith(f.nodes)}),s.each(f=>pr(f)),s.toString()}function y_(r,e){let t=[];return r.walkPseudos(i=>{i.value===Fo&&t.push({pseudo:i,value:i.nodes[0].toString()})}),e.walkPseudos(i=>{if(i.value!==Fo)return;let n=i.nodes[0].toString(),a=t.find(c=>c.value===n);if(!a)return;let s=[],o=i.next();for(;o&&o.type!=="combinator";)s.push(o),o=o.next();let l=o;a.pseudo.parent.insertAfter(a.pseudo,st.default.selector({nodes:s.map(c=>c.clone())})),i.remove(),s.forEach(c=>c.remove()),l&&l.type==="combinator"&&l.remove()}),[r,e]}var st,Sh,Fo,zo=R(()=>{u();st=pe(it()),Sh=pe(Rn());fr();Wn();es();zt();Fo=":merge"});function rs(r,e){let t=(0,Uo.default)().astSync(r);return t.each(i=>{i.nodes.some(a=>a.type==="combinator")&&(i.nodes=[Uo.default.pseudo({value:":is",nodes:[i.clone()]})]),pr(i)}),`${e} ${t.toString()}`}var Uo,Vo=R(()=>{u();Uo=pe(it());es()});function Ho(r){return b_.transformSync(r)}function*w_(r){let e=1/0;for(;e>=0;){let t,i=!1;if(e===1/0&&r.endsWith("]")){let s=r.indexOf("[");r[s-1]==="-"?t=s-1:r[s-1]==="/"?(t=s-1,i=!0):t=-1}else e===1/0&&r.includes("/")?(t=r.lastIndexOf("/"),i=!0):t=r.lastIndexOf("-",e);if(t<0)break;let n=r.slice(0,t),a=r.slice(i?t:t+1);e=t-1,!(n===""||a==="/")&&(yield[n,a])}}function v_(r,e){if(r.length===0||e.tailwindConfig.prefix==="")return r;for(let t of r){let[i]=t;if(i.options.respectPrefix){let n=ee.root({nodes:[t[1].clone()]}),a=t[1].raws.tailwind.classCandidate;n.walkRules(s=>{let o=a.startsWith("-");s.selector=ur(e.tailwindConfig.prefix,s.selector,o)}),t[1]=n.nodes[0]}}return r}function x_(r,e){if(r.length===0)return r;let t=[];function i(n){return n.parent&&n.parent.type==="atrule"&&n.parent.name==="keyframes"}for(let[n,a]of r){let s=ee.root({nodes:[a.clone()]});s.walkRules(o=>{if(i(o))return;let l=(0,is.default)().astSync(o.selector);l.each(c=>jo(c,e)),Wf(l,c=>c===e?`!${c}`:c),o.selector=l.toString(),o.walkDecls(c=>c.important=!0)}),t.push([{...n,important:!0},s.nodes[0]])}return t}function k_(r,e,t){if(e.length===0)return e;let i={modifier:null,value:Ei};{let[n,...a]=ve(r,"/");if(a.length>1&&(n=n+"/"+a.slice(0,-1).join("/"),a=a.slice(-1)),a.length&&!t.variantMap.has(r)&&(r=n,i.modifier=a[0],!we(t.tailwindConfig,"generalizedModifiers")))return[]}if(r.endsWith("]")&&!r.startsWith("[")){let n=/(.)(-?)\[(.*)\]/g.exec(r);if(n){let[,a,s,o]=n;if(a==="@"&&s==="-")return[];if(a!=="@"&&s==="")return[];r=r.replace(`${s}[${o}]`,""),i.value=o}}if(Qo(r)&&!t.variantMap.has(r)){let n=t.offsets.recordVariant(r),a=K(r.slice(1,-1)),s=ve(a,",");if(s.length>1)return[];if(!s.every(os))return[];let o=s.map((l,c)=>[t.offsets.applyParallelOffset(n,c),Oi(l.trim())]);t.variantMap.set(r,o)}if(t.variantMap.has(r)){let n=Qo(r),a=t.variantOptions.get(r)?.[Pt]??{},s=t.variantMap.get(r).slice(),o=[],l=(()=>!(n||a.respectPrefix===!1))();for(let[c,f]of e){if(c.layer==="user")continue;let d=ee.root({nodes:[f.clone()]});for(let[p,h,b]of s){let w=function(){v.raws.neededBackup||(v.raws.neededBackup=!0,v.walkRules(O=>O.raws.originalSelector=O.selector))},k=function(O){return w(),v.each(B=>{B.type==="rule"&&(B.selectors=B.selectors.map(N=>O({get className(){return Ho(N)},selector:N})))}),v},v=(b??d).clone(),y=[],S=h({get container(){return w(),v},separator:t.tailwindConfig.separator,modifySelectors:k,wrap(O){let B=v.nodes;v.removeAll(),O.append(B),v.append(O)},format(O){y.push({format:O,respectPrefix:l})},args:i});if(Array.isArray(S)){for(let[O,B]of S.entries())s.push([t.offsets.applyParallelOffset(p,O),B,v.clone()]);continue}if(typeof S=="string"&&y.push({format:S,respectPrefix:l}),S===null)continue;v.raws.neededBackup&&(delete v.raws.neededBackup,v.walkRules(O=>{let B=O.raws.originalSelector;if(!B||(delete O.raws.originalSelector,B===O.selector))return;let N=O.selector,T=(0,is.default)(F=>{F.walkClasses(Y=>{Y.value=`${r}${t.tailwindConfig.separator}${Y.value}`})}).processSync(B);y.push({format:N.replace(T,"&"),respectPrefix:l}),O.selector=B})),v.nodes[0].raws.tailwind={...v.nodes[0].raws.tailwind,parentLayer:c.layer};let E=[{...c,sort:t.offsets.applyVariantOffset(c.sort,p,Object.assign(i,t.variantOptions.get(r))),collectedFormats:(c.collectedFormats??[]).concat(y)},v.nodes[0]];o.push(E)}}return o}return[]}function Wo(r,e,t={}){return!ke(r)&&!Array.isArray(r)?[[r],t]:Array.isArray(r)?Wo(r[0],e,r[1]):(e.has(r)||e.set(r,lr(r)),[e.get(r),t])}function A_(r){return S_.test(r)}function C_(r){if(!r.includes("://"))return!1;try{let e=new URL(r);return e.scheme!==""&&e.host!==""}catch(e){return!1}}function Ch(r){let e=!0;return r.walkDecls(t=>{if(!_h(t.prop,t.value))return e=!1,!1}),e}function _h(r,e){if(C_(`${r}:${e}`))return!1;try{return ee.parse(`a{${r}:${e}}`).toResult(),!0}catch(t){return!1}}function __(r,e){let[,t,i]=r.match(/^\[([a-zA-Z0-9-_]+):(\S+)\]$/)??[];if(i===void 0||!A_(t)||!cr(i))return null;let n=K(i,{property:t});return _h(t,n)?[[{sort:e.offsets.arbitraryProperty(r),layer:"utilities",options:{respectImportant:!0}},()=>({[Do(r)]:{[t]:n}})]]:null}function*E_(r,e){e.candidateRuleMap.has(r)&&(yield[e.candidateRuleMap.get(r),"DEFAULT"]),yield*function*(o){o!==null&&(yield[o,"DEFAULT"])}(__(r,e));let t=r,i=!1,n=e.tailwindConfig.prefix,a=n.length,s=t.startsWith(n)||t.startsWith(`-${n}`);t[a]==="-"&&s&&(i=!0,t=n+t.slice(a+1)),i&&e.candidateRuleMap.has(t)&&(yield[e.candidateRuleMap.get(t),"-DEFAULT"]);for(let[o,l]of w_(t))e.candidateRuleMap.has(o)&&(yield[e.candidateRuleMap.get(o),i?`-${l}`:l])}function O_(r,e){return r===gt?[gt]:ve(r,e)}function*T_(r,e){for(let t of r)t[1].raws.tailwind={...t[1].raws.tailwind,classCandidate:e,preserveSource:t[0].options?.preserveSource??!1},yield t}function*Go(r,e){let t=e.tailwindConfig.separator,[i,...n]=O_(r,t).reverse(),a=!1;i.startsWith("!")&&(a=!0,i=i.slice(1));for(let s of E_(i,e)){let o=[],l=new Map,[c,f]=s,d=c.length===1;for(let[p,h]of c){let b=[];if(typeof h=="function")for(let v of[].concat(h(f,{isOnlyPlugin:d}))){let[y,w]=Wo(v,e.postCssNodeCache);for(let k of y)b.push([{...p,options:{...p.options,...w}},k])}else if(f==="DEFAULT"||f==="-DEFAULT"){let v=h,[y,w]=Wo(v,e.postCssNodeCache);for(let k of y)b.push([{...p,options:{...p.options,...w}},k])}if(b.length>0){let v=Array.from(Zs(p.options?.types??[],f,p.options??{},e.tailwindConfig)).map(([y,w])=>w);v.length>0&&l.set(b,v),o.push(b)}}if(Qo(f)){if(o.length>1){let b=function(y){return y.length===1?y[0]:y.find(w=>{let k=l.get(w);return w.some(([{options:S},E])=>Ch(E)?S.types.some(({type:O,preferOnConflict:B})=>k.includes(O)&&B):!1)})},[p,h]=o.reduce((y,w)=>(w.some(([{options:S}])=>S.types.some(({type:E})=>E==="any"))?y[0].push(w):y[1].push(w),y),[[],[]]),v=b(h)??b(p);if(v)o=[v];else{let y=o.map(k=>new Set([...l.get(k)??[]]));for(let k of y)for(let S of k){let E=!1;for(let O of y)k!==O&&O.has(S)&&(O.delete(S),E=!0);E&&k.delete(S)}let w=[];for(let[k,S]of y.entries())for(let E of S){let O=o[k].map(([,B])=>B).flat().map(B=>B.toString().split(` -`).slice(1,-1).map(N=>N.trim()).map(N=>` ${N}`).join(` -`)).join(` - -`);w.push(` Use \`${r.replace("[",`[${E}:`)}\` for \`${O.trim()}\``);break}G.warn([`The class \`${r}\` is ambiguous and matches multiple utilities.`,...w,`If this is content and not a class, replace it with \`${r.replace("[","[").replace("]","]")}\` to silence this warning.`]);continue}}o=o.map(p=>p.filter(h=>Ch(h[1])))}o=o.flat(),o=Array.from(T_(o,i)),o=v_(o,e),a&&(o=x_(o,i));for(let p of n)o=k_(p,o,e);for(let p of o)p[1].raws.tailwind={...p[1].raws.tailwind,candidate:r},p=R_(p,{context:e,candidate:r}),p!==null&&(yield p)}}function R_(r,{context:e,candidate:t}){if(!r[0].collectedFormats)return r;let i=!0,n;try{n=dr(r[0].collectedFormats,{context:e,candidate:t})}catch{return null}let a=ee.root({nodes:[r[1].clone()]});return a.walkRules(s=>{if(!ns(s))try{let o=ts(s.selector,n,{candidate:t,context:e});if(o===null){s.remove();return}s.selector=o}catch{return i=!1,!1}}),!i||a.nodes.length===0?null:(r[1]=a.nodes[0],r)}function ns(r){return r.parent&&r.parent.type==="atrule"&&r.parent.name==="keyframes"}function P_(r){if(r===!0)return e=>{ns(e)||e.walkDecls(t=>{t.parent.type==="rule"&&!ns(t.parent)&&(t.important=!0)})};if(typeof r=="string")return e=>{ns(e)||(e.selectors=e.selectors.map(t=>rs(t,r)))}}function ss(r,e,t=!1){let i=[],n=P_(e.tailwindConfig.important);for(let a of r){if(e.notClassCache.has(a))continue;if(e.candidateRuleCache.has(a)){i=i.concat(Array.from(e.candidateRuleCache.get(a)));continue}let s=Array.from(Go(a,e));if(s.length===0){e.notClassCache.add(a);continue}e.classCache.set(a,s);let o=e.candidateRuleCache.get(a)??new Set;e.candidateRuleCache.set(a,o);for(let l of s){let[{sort:c,options:f},d]=l;if(f.respectImportant&&n){let h=ee.root({nodes:[d.clone()]});h.walkRules(n),d=h.nodes[0]}let p=[c,t?d.clone():d];o.add(p),e.ruleCache.add(p),i.push(p)}}return i}function Qo(r){return r.startsWith("[")&&r.endsWith("]")}var is,b_,S_,as=R(()=>{u();Ot();is=pe(it());Io();Kt();Wn();Fr();Be();It();zo();qo();Br();_i();Mo();zt();ct();Vo();b_=(0,is.default)(r=>r.first.filter(({type:e})=>e==="class").pop().value);S_=/^[a-z_-]/});var Eh,Oh=R(()=>{u();Eh={}});function I_(r){try{return Eh.createHash("md5").update(r,"utf-8").digest("binary")}catch(e){return""}}function Th(r,e){let t=e.toString();if(!t.includes("@tailwind"))return!1;let i=Lo.get(r),n=I_(t),a=i!==n;return Lo.set(r,n),a}var Rh=R(()=>{u();Oh();It()});function ls(r){return(r>0n)-(r<0n)}var Ph=R(()=>{u()});function Ih(r,e){let t=0n,i=0n;for(let[n,a]of e)r&n&&(t=t|n,i=i|a);return r&~t|i}var Dh=R(()=>{u()});function qh(r){let e=null;for(let t of r)e=e??t,e=e>t?e:t;return e}function D_(r,e){let t=r.length,i=e.length,n=t{u();Ph();Dh();Yo=class{constructor(){this.offsets={defaults:0n,base:0n,components:0n,utilities:0n,variants:0n,user:0n},this.layerPositions={defaults:0n,base:1n,components:2n,utilities:3n,user:4n,variants:5n},this.reservedVariantBits=0n,this.variantOffsets=new Map}create(e){return{layer:e,parentLayer:e,arbitrary:0n,variants:0n,parallelIndex:0n,index:this.offsets[e]++,propertyOffset:0n,property:"",options:[]}}arbitraryProperty(e){return{...this.create("utilities"),arbitrary:1n,property:e}}forVariant(e,t=0){let i=this.variantOffsets.get(e);if(i===void 0)throw new Error(`Cannot find offset for unknown variant ${e}`);return{...this.create("variants"),variants:i<n.startsWith("[")).sort(([n],[a])=>D_(n,a)),t=e.map(([,n])=>n).sort((n,a)=>ls(n-a));return e.map(([,n],a)=>[n,t[a]]).filter(([n,a])=>n!==a)}remapArbitraryVariantOffsets(e){let t=this.recalculateVariantOffsets();return t.length===0?e:e.map(i=>{let[n,a]=i;return n={...n,variants:Ih(n.variants,t)},[n,a]})}sortArbitraryProperties(e){let t=new Set;for(let[s]of e)s.arbitrary===1n&&t.add(s.property);if(t.size===0)return e;let i=Array.from(t).sort(),n=new Map,a=1n;for(let s of i)n.set(s,a++);return e.map(s=>{let[o,l]=s;return o={...o,propertyOffset:n.get(o.property)??0n},[o,l]})}sort(e){return e=this.remapArbitraryVariantOffsets(e),e=this.sortArbitraryProperties(e),e.sort(([t],[i])=>ls(this.compare(t,i)))}}});function Zo(r,e){let t=r.tailwindConfig.prefix;return typeof t=="function"?t(e):t+e}function Mh({type:r="any",...e}){let t=[].concat(r);return{...e,types:t.map(i=>Array.isArray(i)?{type:i[0],...i[1]}:{type:i,preferOnConflict:!1})}}function q_(r){let e=[],t="",i=0;for(let n=0;n0&&e.push(t.trim()),e=e.filter(n=>n!==""),e}function $_(r,e,{before:t=[]}={}){if(t=[].concat(t),t.length<=0){r.push(e);return}let i=r.length-1;for(let n of t){let a=r.indexOf(n);a!==-1&&(i=Math.min(i,a))}r.splice(i,0,e)}function Nh(r){return Array.isArray(r)?r.flatMap(e=>!Array.isArray(e)&&!ke(e)?e:lr(e)):Nh([r])}function L_(r,e){return(0,Ko.default)(i=>{let n=[];return e&&e(i),i.walkClasses(a=>{n.push(a.value)}),n}).transformSync(r)}function M_(r){r.walkPseudos(e=>{e.value===":not"&&e.remove()})}function N_(r,e={containsNonOnDemandable:!1},t=0){let i=[],n=[];r.type==="rule"?n.push(...r.selectors):r.type==="atrule"&&r.walkRules(a=>n.push(...a.selectors));for(let a of n){let s=L_(a,M_);s.length===0&&(e.containsNonOnDemandable=!0);for(let o of s)i.push(o)}return t===0?[e.containsNonOnDemandable||i.length===0,i]:i}function us(r){return Nh(r).flatMap(e=>{let t=new Map,[i,n]=N_(e);return i&&n.unshift(gt),n.map(a=>(t.has(e)||t.set(e,e),[a,t.get(e)]))})}function os(r){return r.startsWith("@")||r.includes("&")}function Oi(r){r=r.replace(/\n+/g,"").replace(/\s{1,}/g," ").trim();let e=q_(r).map(t=>{if(!t.startsWith("@"))return({format:a})=>a(t);let[,i,n]=/@(\S*)( .+|[({].*)?/g.exec(t);return({wrap:a})=>a(ee.atRule({name:i,params:n?.trim()??""}))}).reverse();return t=>{for(let i of e)i(t)}}function B_(r,e,{variantList:t,variantMap:i,offsets:n,classList:a}){function s(p,h){return p?(0,Lh.default)(r,p,h):r}function o(p){return ur(r.prefix,p)}function l(p,h){return p===gt?gt:h.respectPrefix?e.tailwindConfig.prefix+p:p}function c(p,h,b={}){let v=kt(p),y=s(["theme",...v],h);return mt(v[0])(y,b)}let f=0,d={postcss:ee,prefix:o,e:Te,config:s,theme:c,corePlugins:p=>Array.isArray(r.corePlugins)?r.corePlugins.includes(p):s(["corePlugins",p],!0),variants:()=>[],addBase(p){for(let[h,b]of us(p)){let v=l(h,{}),y=n.create("base");e.candidateRuleMap.has(v)||e.candidateRuleMap.set(v,[]),e.candidateRuleMap.get(v).push([{sort:y,layer:"base"},b])}},addDefaults(p,h){let b={[`@defaults ${p}`]:h};for(let[v,y]of us(b)){let w=l(v,{});e.candidateRuleMap.has(w)||e.candidateRuleMap.set(w,[]),e.candidateRuleMap.get(w).push([{sort:n.create("defaults"),layer:"defaults"},y])}},addComponents(p,h){h=Object.assign({},{preserveSource:!1,respectPrefix:!0,respectImportant:!1},Array.isArray(h)?{}:h);for(let[v,y]of us(p)){let w=l(v,h);a.add(w),e.candidateRuleMap.has(w)||e.candidateRuleMap.set(w,[]),e.candidateRuleMap.get(w).push([{sort:n.create("components"),layer:"components",options:h},y])}},addUtilities(p,h){h=Object.assign({},{preserveSource:!1,respectPrefix:!0,respectImportant:!0},Array.isArray(h)?{}:h);for(let[v,y]of us(p)){let w=l(v,h);a.add(w),e.candidateRuleMap.has(w)||e.candidateRuleMap.set(w,[]),e.candidateRuleMap.get(w).push([{sort:n.create("utilities"),layer:"utilities",options:h},y])}},matchUtilities:function(p,h){h=Mh({...{respectPrefix:!0,respectImportant:!0,modifiers:!1},...h});let v=n.create("utilities");for(let y in p){let S=function(O,{isOnlyPlugin:B}){let[N,T,F]=Js(h.types,O,h,r);if(N===void 0)return[];if(!h.types.some(({type:U})=>U===T))if(B)G.warn([`Unnecessary typehint \`${T}\` in \`${y}-${O}\`.`,`You can safely update it to \`${y}-${O.replace(T+":","")}\`.`]);else return[];if(!cr(N))return[];let Y={get modifier(){return h.modifiers||G.warn(`modifier-used-without-options-for-${y}`,["Your plugin must set `modifiers: true` in its options to support modifiers."]),F}},_=we(r,"generalizedModifiers");return[].concat(_?k(N,Y):k(N)).filter(Boolean).map(U=>({[Gn(y,O)]:U}))},w=l(y,h),k=p[y];a.add([w,h]);let E=[{sort:v,layer:"utilities",options:h},S];e.candidateRuleMap.has(w)||e.candidateRuleMap.set(w,[]),e.candidateRuleMap.get(w).push(E)}},matchComponents:function(p,h){h=Mh({...{respectPrefix:!0,respectImportant:!1,modifiers:!1},...h});let v=n.create("components");for(let y in p){let S=function(O,{isOnlyPlugin:B}){let[N,T,F]=Js(h.types,O,h,r);if(N===void 0)return[];if(!h.types.some(({type:U})=>U===T))if(B)G.warn([`Unnecessary typehint \`${T}\` in \`${y}-${O}\`.`,`You can safely update it to \`${y}-${O.replace(T+":","")}\`.`]);else return[];if(!cr(N))return[];let Y={get modifier(){return h.modifiers||G.warn(`modifier-used-without-options-for-${y}`,["Your plugin must set `modifiers: true` in its options to support modifiers."]),F}},_=we(r,"generalizedModifiers");return[].concat(_?k(N,Y):k(N)).filter(Boolean).map(U=>({[Gn(y,O)]:U}))},w=l(y,h),k=p[y];a.add([w,h]);let E=[{sort:v,layer:"components",options:h},S];e.candidateRuleMap.has(w)||e.candidateRuleMap.set(w,[]),e.candidateRuleMap.get(w).push(E)}},addVariant(p,h,b={}){h=[].concat(h).map(v=>{if(typeof v!="string")return(y={})=>{let{args:w,modifySelectors:k,container:S,separator:E,wrap:O,format:B}=y,N=v(Object.assign({modifySelectors:k,container:S,separator:E},b.type===Xo.MatchVariant&&{args:w,wrap:O,format:B}));if(typeof N=="string"&&!os(N))throw new Error(`Your custom variant \`${p}\` has an invalid format string. Make sure it's an at-rule or contains a \`&\` placeholder.`);return Array.isArray(N)?N.filter(T=>typeof T=="string").map(T=>Oi(T)):N&&typeof N=="string"&&Oi(N)(y)};if(!os(v))throw new Error(`Your custom variant \`${p}\` has an invalid format string. Make sure it's an at-rule or contains a \`&\` placeholder.`);return Oi(v)}),$_(t,p,b),i.set(p,h),e.variantOptions.set(p,b)},matchVariant(p,h,b){let v=b?.id??++f,y=p==="@",w=we(r,"generalizedModifiers");for(let[S,E]of Object.entries(b?.values??{}))S!=="DEFAULT"&&d.addVariant(y?`${p}${S}`:`${p}-${S}`,({args:O,container:B})=>h(E,w?{modifier:O?.modifier,container:B}:{container:B}),{...b,value:E,id:v,type:Xo.MatchVariant,variantInfo:Jo.Base});let k="DEFAULT"in(b?.values??{});d.addVariant(p,({args:S,container:E})=>S?.value===Ei&&!k?null:h(S?.value===Ei?b.values.DEFAULT:S?.value??(typeof S=="string"?S:""),w?{modifier:S?.modifier,container:E}:{container:E}),{...b,id:v,type:Xo.MatchVariant,variantInfo:Jo.Dynamic})}};return d}function fs(r){return el.has(r)||el.set(r,new Map),el.get(r)}function Bh(r,e){let t=!1,i=new Map;for(let n of r){if(!n)continue;let a=sa.parse(n),s=a.hash?a.href.replace(a.hash,""):a.href;s=a.search?s.replace(a.search,""):s;let o=be.statSync(decodeURIComponent(s),{throwIfNoEntry:!1})?.mtimeMs;!o||((!e.has(n)||o>e.get(n))&&(t=!0),i.set(n,o))}return[t,i]}function Fh(r){r.walkAtRules(e=>{["responsive","variants"].includes(e.name)&&(Fh(e),e.before(e.nodes),e.remove())})}function F_(r){let e=[];return r.each(t=>{t.type==="atrule"&&["responsive","variants"].includes(t.name)&&(t.name="layer",t.params="utilities")}),r.walkAtRules("layer",t=>{if(Fh(t),t.params==="base"){for(let i of t.nodes)e.push(function({addBase:n}){n(i,{respectPrefix:!1})});t.remove()}else if(t.params==="components"){for(let i of t.nodes)e.push(function({addComponents:n}){n(i,{respectPrefix:!1,preserveSource:!0})});t.remove()}else if(t.params==="utilities"){for(let i of t.nodes)e.push(function({addUtilities:n}){n(i,{respectPrefix:!1,preserveSource:!0})});t.remove()}}),e}function j_(r,e){let t=Object.entries({...se,...mh}).map(([l,c])=>r.tailwindConfig.corePlugins.includes(l)?c:null).filter(Boolean),i=r.tailwindConfig.plugins.map(l=>(l.__isOptionsFunction&&(l=l()),typeof l=="function"?l:l.handler)),n=F_(e),a=[se.childVariant,se.pseudoElementVariants,se.pseudoClassVariants,se.hasVariants,se.ariaVariants,se.dataVariants],s=[se.supportsVariants,se.reducedMotionVariants,se.prefersContrastVariants,se.screenVariants,se.orientationVariants,se.directionVariants,se.darkVariants,se.forcedColorsVariants,se.printVariant];return(r.tailwindConfig.darkMode==="class"||Array.isArray(r.tailwindConfig.darkMode)&&r.tailwindConfig.darkMode[0]==="class")&&(s=[se.supportsVariants,se.reducedMotionVariants,se.prefersContrastVariants,se.darkVariants,se.screenVariants,se.orientationVariants,se.directionVariants,se.forcedColorsVariants,se.printVariant]),[...t,...a,...i,...s,...n]}function z_(r,e){let t=[],i=new Map;e.variantMap=i;let n=new Yo;e.offsets=n;let a=new Set,s=B_(e.tailwindConfig,e,{variantList:t,variantMap:i,offsets:n,classList:a});for(let f of r)if(Array.isArray(f))for(let d of f)d(s);else f?.(s);n.recordVariants(t,f=>i.get(f).length);for(let[f,d]of i.entries())e.variantMap.set(f,d.map((p,h)=>[n.forVariant(f,h),p]));let o=(e.tailwindConfig.safelist??[]).filter(Boolean);if(o.length>0){let f=[];for(let d of o){if(typeof d=="string"){e.changedContent.push({content:d,extension:"html"});continue}if(d instanceof RegExp){G.warn("root-regex",["Regular expressions in `safelist` work differently in Tailwind CSS v3.0.","Update your `safelist` configuration to eliminate this warning.","https://tailwindcss.com/docs/content-configuration#safelisting-classes"]);continue}f.push(d)}if(f.length>0){let d=new Map,p=e.tailwindConfig.prefix.length,h=f.some(b=>b.pattern.source.includes("!"));for(let b of a){let v=Array.isArray(b)?(()=>{let[y,w]=b,S=Object.keys(w?.values??{}).map(E=>Ci(y,E));return w?.supportsNegativeValues&&(S=[...S,...S.map(E=>"-"+E)],S=[...S,...S.map(E=>E.slice(0,p)+"-"+E.slice(p))]),w.types.some(({type:E})=>E==="color")&&(S=[...S,...S.flatMap(E=>Object.keys(e.tailwindConfig.theme.opacity).map(O=>`${E}/${O}`))]),h&&w?.respectImportant&&(S=[...S,...S.map(E=>"!"+E)]),S})():[b];for(let y of v)for(let{pattern:w,variants:k=[]}of f)if(w.lastIndex=0,d.has(w)||d.set(w,0),!!w.test(y)){d.set(w,d.get(w)+1),e.changedContent.push({content:y,extension:"html"});for(let S of k)e.changedContent.push({content:S+e.tailwindConfig.separator+y,extension:"html"})}}for(let[b,v]of d.entries())v===0&&G.warn([`The safelist pattern \`${b}\` doesn't match any Tailwind CSS classes.`,"Fix this pattern or remove it from your `safelist` configuration.","https://tailwindcss.com/docs/content-configuration#safelisting-classes"])}}let l=[].concat(e.tailwindConfig.darkMode??"media")[1]??"dark",c=[Zo(e,l),Zo(e,"group"),Zo(e,"peer")];e.getClassOrder=function(d){let p=[...d].sort((y,w)=>y===w?0:y[y,null])),b=ss(new Set(p),e,!0);b=e.offsets.sort(b);let v=BigInt(c.length);for(let[,y]of b){let w=y.raws.tailwind.candidate;h.set(w,h.get(w)??v++)}return d.map(y=>{let w=h.get(y)??null,k=c.indexOf(y);return w===null&&k!==-1&&(w=BigInt(k)),[y,w]})},e.getClassList=function(d={}){let p=[];for(let h of a)if(Array.isArray(h)){let[b,v]=h,y=[],w=Object.keys(v?.modifiers??{});v?.types?.some(({type:E})=>E==="color")&&w.push(...Object.keys(e.tailwindConfig.theme.opacity??{}));let k={modifiers:w},S=d.includeMetadata&&w.length>0;for(let[E,O]of Object.entries(v?.values??{})){if(O==null)continue;let B=Ci(b,E);if(p.push(S?[B,k]:B),v?.supportsNegativeValues&&xt(O)){let N=Ci(b,`-${E}`);y.push(S?[N,k]:N)}}p.push(...y)}else p.push(h);return p},e.getVariants=function(){let d=Math.random().toString(36).substring(7).toUpperCase(),p=[];for(let[h,b]of e.variantOptions.entries())b.variantInfo!==Jo.Base&&p.push({name:h,isArbitrary:b.type===Symbol.for("MATCH_VARIANT"),values:Object.keys(b.values??{}),hasDash:h!=="@",selectors({modifier:v,value:y}={}){let w=`TAILWINDPLACEHOLDER${d}`,k=ee.rule({selector:`.${w}`}),S=ee.root({nodes:[k.clone()]}),E=S.toString(),O=(e.variantMap.get(h)??[]).flatMap(([oe,A])=>A),B=[];for(let oe of O){let A=[],C={args:{modifier:v,value:b.values?.[y]??y},separator:e.tailwindConfig.separator,modifySelectors(V){return S.each(Ee=>{Ee.type==="rule"&&(Ee.selectors=Ee.selectors.map(Ie=>V({get className(){return Ho(Ie)},selector:Ie})))}),S},format(V){A.push(V)},wrap(V){A.push(`@${V.name} ${V.params} { & }`)},container:S},he=oe(C);if(A.length>0&&B.push(A),Array.isArray(he))for(let V of he)A=[],V(C),B.push(A)}let N=[],T=S.toString();E!==T&&(S.walkRules(oe=>{let A=oe.selector,C=(0,Ko.default)(he=>{he.walkClasses(V=>{V.value=`${h}${e.tailwindConfig.separator}${V.value}`})}).processSync(A);N.push(A.replace(C,"&").replace(w,"&"))}),S.walkAtRules(oe=>{N.push(`@${oe.name} (${oe.params}) { & }`)}));let F=!(y in(b.values??{})),Y=b[Pt]??{},_=(()=>!(F||Y.respectPrefix===!1))();B=B.map(oe=>oe.map(A=>({format:A,respectPrefix:_}))),N=N.map(oe=>({format:oe,respectPrefix:_}));let Q={candidate:w,context:e},U=B.map(oe=>ts(`.${w}`,dr(oe,Q),Q).replace(`.${w}`,"&").replace("{ & }","").trim());return N.length>0&&U.push(dr(N,Q).toString().replace(`.${w}`,"&")),U}});return p}}function jh(r,e){!r.classCache.has(e)||(r.notClassCache.add(e),r.classCache.delete(e),r.applyClassCache.delete(e),r.candidateRuleMap.delete(e),r.candidateRuleCache.delete(e),r.stylesheetCache=null)}function U_(r,e){let t=e.raws.tailwind.candidate;if(!!t){for(let i of r.ruleCache)i[1].raws.tailwind.candidate===t&&r.ruleCache.delete(i);jh(r,t)}}function tl(r,e=[],t=ee.root()){let i={disposables:[],ruleCache:new Set,candidateRuleCache:new Map,classCache:new Map,applyClassCache:new Map,notClassCache:new Set(r.blocklist??[]),postCssNodeCache:new Map,candidateRuleMap:new Map,tailwindConfig:r,changedContent:e,variantMap:new Map,stylesheetCache:null,variantOptions:new Map,markInvalidUtilityCandidate:a=>jh(i,a),markInvalidUtilityNode:a=>U_(i,a)},n=j_(i,t);return z_(n,i),i}function zh(r,e,t,i,n,a){let s=e.opts.from,o=i!==null;Je.DEBUG&&console.log("Source path:",s);let l;if(o&&hr.has(s))l=hr.get(s);else if(Ti.has(n)){let p=Ti.get(n);Dt.get(p).add(s),hr.set(s,p),l=p}let c=Th(s,r);if(l){let[p,h]=Bh([...a],fs(l));if(!p&&!c)return[l,!1,h]}if(hr.has(s)){let p=hr.get(s);if(Dt.has(p)&&(Dt.get(p).delete(s),Dt.get(p).size===0)){Dt.delete(p);for(let[h,b]of Ti)b===p&&Ti.delete(h);for(let h of p.disposables.splice(0))h(p)}}Je.DEBUG&&console.log("Setting up new context...");let f=tl(t,[],r);Object.assign(f,{userConfigPath:i});let[,d]=Bh([...a],fs(f));return Ti.set(n,f),hr.set(s,f),Dt.has(f)||Dt.set(f,new Set),Dt.get(f).add(s),[f,!0,d]}var Lh,Ko,Pt,Xo,Jo,el,hr,Ti,Dt,_i=R(()=>{u();ft();aa();Ot();Lh=pe(Oa()),Ko=pe(it());Si();Io();Wn();Kt();fr();qo();Fr();gh();It();It();Gi();Be();Hi();Mo();as();Rh();$h();ct();zo();Pt=Symbol(),Xo={AddVariant:Symbol.for("ADD_VARIANT"),MatchVariant:Symbol.for("MATCH_VARIANT")},Jo={Base:1<<0,Dynamic:1<<1};el=new WeakMap;hr=yh,Ti=bh,Dt=Zn});function rl(r){return r.ignore?[]:r.glob?m.env.ROLLUP_WATCH==="true"?[{type:"dependency",file:r.base}]:[{type:"dir-dependency",dir:r.base,glob:r.glob}]:[{type:"dependency",file:r.base}]}var Uh=R(()=>{u()});function Vh(r,e){return{handler:r,config:e}}var Hh,Wh=R(()=>{u();Vh.withOptions=function(r,e=()=>({})){let t=function(i){return{__options:i,handler:r(i),config:e(i)}};return t.__isOptionsFunction=!0,t.__pluginFunction=r,t.__configFunction=e,t};Hh=Vh});var il={};Ge(il,{default:()=>V_});var V_,nl=R(()=>{u();Wh();V_=Hh});var Qh=x((F4,Gh)=>{u();var H_=(nl(),il).default,W_={overflow:"hidden",display:"-webkit-box","-webkit-box-orient":"vertical"},G_=H_(function({matchUtilities:r,addUtilities:e,theme:t,variants:i}){let n=t("lineClamp");r({"line-clamp":a=>({...W_,"-webkit-line-clamp":`${a}`})},{values:n}),e([{".line-clamp-none":{"-webkit-line-clamp":"unset"}}],i("lineClamp"))},{theme:{lineClamp:{1:"1",2:"2",3:"3",4:"4",5:"5",6:"6"}},variants:{lineClamp:["responsive"]}});Gh.exports=G_});function sl(r){r.content.files.length===0&&G.warn("content-problems",["The `content` option in your Tailwind CSS configuration is missing or empty.","Configure your content sources or your generated CSS will be missing styles.","https://tailwindcss.com/docs/content-configuration"]);try{let e=Qh();r.plugins.includes(e)&&(G.warn("line-clamp-in-core",["As of Tailwind CSS v3.3, the `@tailwindcss/line-clamp` plugin is now included by default.","Remove it from the `plugins` array in your configuration to eliminate this warning."]),r.plugins=r.plugins.filter(t=>t!==e))}catch{}return r}var Yh=R(()=>{u();Be()});var Kh,Xh=R(()=>{u();Kh=()=>!1});var cs,Jh=R(()=>{u();cs={sync:r=>[].concat(r),generateTasks:r=>[{dynamic:!1,base:".",negative:[],positive:[].concat(r),patterns:[].concat(r)}],escapePath:r=>r}});var al,Zh=R(()=>{u();al=r=>r});var em,tm=R(()=>{u();em=()=>""});function rm(r){let e=r,t=em(r);return t!=="."&&(e=r.substr(t.length),e.charAt(0)==="/"&&(e=e.substr(1))),e.substr(0,2)==="./"?e=e.substr(2):e.charAt(0)==="/"&&(e=e.substr(1)),{base:t,glob:e}}var im=R(()=>{u();tm()});var ps=x(Ve=>{u();"use strict";Ve.isInteger=r=>typeof r=="number"?Number.isInteger(r):typeof r=="string"&&r.trim()!==""?Number.isInteger(Number(r)):!1;Ve.find=(r,e)=>r.nodes.find(t=>t.type===e);Ve.exceedsLimit=(r,e,t=1,i)=>i===!1||!Ve.isInteger(r)||!Ve.isInteger(e)?!1:(Number(e)-Number(r))/Number(t)>=i;Ve.escapeNode=(r,e=0,t)=>{let i=r.nodes[e];!i||(t&&i.type===t||i.type==="open"||i.type==="close")&&i.escaped!==!0&&(i.value="\\"+i.value,i.escaped=!0)};Ve.encloseBrace=r=>r.type!=="brace"?!1:r.commas>>0+r.ranges>>0==0?(r.invalid=!0,!0):!1;Ve.isInvalidBrace=r=>r.type!=="brace"?!1:r.invalid===!0||r.dollar?!0:r.commas>>0+r.ranges>>0==0||r.open!==!0||r.close!==!0?(r.invalid=!0,!0):!1;Ve.isOpenOrClose=r=>r.type==="open"||r.type==="close"?!0:r.open===!0||r.close===!0;Ve.reduce=r=>r.reduce((e,t)=>(t.type==="text"&&e.push(t.value),t.type==="range"&&(t.type="text"),e),[]);Ve.flatten=(...r)=>{let e=[],t=i=>{for(let n=0;n{u();"use strict";var nm=ps();sm.exports=(r,e={})=>{let t=(i,n={})=>{let a=e.escapeInvalid&&nm.isInvalidBrace(n),s=i.invalid===!0&&e.escapeInvalid===!0,o="";if(i.value)return(a||s)&&nm.isOpenOrClose(i)?"\\"+i.value:i.value;if(i.value)return i.value;if(i.nodes)for(let l of i.nodes)o+=t(l);return o};return t(r)}});var om=x((X4,am)=>{u();"use strict";am.exports=function(r){return typeof r=="number"?r-r==0:typeof r=="string"&&r.trim()!==""?Number.isFinite?Number.isFinite(+r):isFinite(+r):!1}});var gm=x((J4,mm)=>{u();"use strict";var lm=om(),Wt=(r,e,t)=>{if(lm(r)===!1)throw new TypeError("toRegexRange: expected the first argument to be a number");if(e===void 0||r===e)return String(r);if(lm(e)===!1)throw new TypeError("toRegexRange: expected the second argument to be a number.");let i={relaxZeros:!0,...t};typeof i.strictZeros=="boolean"&&(i.relaxZeros=i.strictZeros===!1);let n=String(i.relaxZeros),a=String(i.shorthand),s=String(i.capture),o=String(i.wrap),l=r+":"+e+"="+n+a+s+o;if(Wt.cache.hasOwnProperty(l))return Wt.cache[l].result;let c=Math.min(r,e),f=Math.max(r,e);if(Math.abs(c-f)===1){let v=r+"|"+e;return i.capture?`(${v})`:i.wrap===!1?v:`(?:${v})`}let d=hm(r)||hm(e),p={min:r,max:e,a:c,b:f},h=[],b=[];if(d&&(p.isPadded=d,p.maxLen=String(p.max).length),c<0){let v=f<0?Math.abs(f):1;b=um(v,Math.abs(c),p,i),c=p.a=0}return f>=0&&(h=um(c,f,p,i)),p.negatives=b,p.positives=h,p.result=Q_(b,h,i),i.capture===!0?p.result=`(${p.result})`:i.wrap!==!1&&h.length+b.length>1&&(p.result=`(?:${p.result})`),Wt.cache[l]=p,p.result};function Q_(r,e,t){let i=ol(r,e,"-",!1,t)||[],n=ol(e,r,"",!1,t)||[],a=ol(r,e,"-?",!0,t)||[];return i.concat(a).concat(n).join("|")}function Y_(r,e){let t=1,i=1,n=cm(r,t),a=new Set([e]);for(;r<=n&&n<=e;)a.add(n),t+=1,n=cm(r,t);for(n=pm(e+1,i)-1;r1&&o.count.pop(),o.count.push(f.count[0]),o.string=o.pattern+dm(o.count),s=c+1;continue}t.isPadded&&(d=eE(c,t,i)),f.string=d+f.pattern+dm(f.count),a.push(f),s=c+1,o=f}return a}function ol(r,e,t,i,n){let a=[];for(let s of r){let{string:o}=s;!i&&!fm(e,"string",o)&&a.push(t+o),i&&fm(e,"string",o)&&a.push(t+o)}return a}function X_(r,e){let t=[];for(let i=0;ie?1:e>r?-1:0}function fm(r,e,t){return r.some(i=>i[e]===t)}function cm(r,e){return Number(String(r).slice(0,-e)+"9".repeat(e))}function pm(r,e){return r-r%Math.pow(10,e)}function dm(r){let[e=0,t=""]=r;return t||e>1?`{${e+(t?","+t:"")}}`:""}function Z_(r,e,t){return`[${r}${e-r==1?"":"-"}${e}]`}function hm(r){return/^-?(0+)\d/.test(r)}function eE(r,e,t){if(!e.isPadded)return r;let i=Math.abs(e.maxLen-String(r).length),n=t.relaxZeros!==!1;switch(i){case 0:return"";case 1:return n?"0?":"0";case 2:return n?"0{0,2}":"00";default:return n?`0{0,${i}}`:`0{${i}}`}}Wt.cache={};Wt.clearCache=()=>Wt.cache={};mm.exports=Wt});var fl=x((Z4,Am)=>{u();"use strict";var tE=(Bn(),Nn),ym=gm(),bm=r=>r!==null&&typeof r=="object"&&!Array.isArray(r),rE=r=>e=>r===!0?Number(e):String(e),ll=r=>typeof r=="number"||typeof r=="string"&&r!=="",Ri=r=>Number.isInteger(+r),ul=r=>{let e=`${r}`,t=-1;if(e[0]==="-"&&(e=e.slice(1)),e==="0")return!1;for(;e[++t]==="0";);return t>0},iE=(r,e,t)=>typeof r=="string"||typeof e=="string"?!0:t.stringify===!0,nE=(r,e,t)=>{if(e>0){let i=r[0]==="-"?"-":"";i&&(r=r.slice(1)),r=i+r.padStart(i?e-1:e,"0")}return t===!1?String(r):r},wm=(r,e)=>{let t=r[0]==="-"?"-":"";for(t&&(r=r.slice(1),e--);r.length{r.negatives.sort((s,o)=>so?1:0),r.positives.sort((s,o)=>so?1:0);let t=e.capture?"":"?:",i="",n="",a;return r.positives.length&&(i=r.positives.join("|")),r.negatives.length&&(n=`-(${t}${r.negatives.join("|")})`),i&&n?a=`${i}|${n}`:a=i||n,e.wrap?`(${t}${a})`:a},vm=(r,e,t,i)=>{if(t)return ym(r,e,{wrap:!1,...i});let n=String.fromCharCode(r);if(r===e)return n;let a=String.fromCharCode(e);return`[${n}-${a}]`},xm=(r,e,t)=>{if(Array.isArray(r)){let i=t.wrap===!0,n=t.capture?"":"?:";return i?`(${n}${r.join("|")})`:r.join("|")}return ym(r,e,t)},km=(...r)=>new RangeError("Invalid range arguments: "+tE.inspect(...r)),Sm=(r,e,t)=>{if(t.strictRanges===!0)throw km([r,e]);return[]},aE=(r,e)=>{if(e.strictRanges===!0)throw new TypeError(`Expected step "${r}" to be a number`);return[]},oE=(r,e,t=1,i={})=>{let n=Number(r),a=Number(e);if(!Number.isInteger(n)||!Number.isInteger(a)){if(i.strictRanges===!0)throw km([r,e]);return[]}n===0&&(n=0),a===0&&(a=0);let s=n>a,o=String(r),l=String(e),c=String(t);t=Math.max(Math.abs(t),1);let f=ul(o)||ul(l)||ul(c),d=f?Math.max(o.length,l.length,c.length):0,p=f===!1&&iE(r,e,i)===!1,h=i.transform||rE(p);if(i.toRegex&&t===1)return vm(wm(r,d),wm(e,d),!0,i);let b={negatives:[],positives:[]},v=k=>b[k<0?"negatives":"positives"].push(Math.abs(k)),y=[],w=0;for(;s?n>=a:n<=a;)i.toRegex===!0&&t>1?v(n):y.push(nE(h(n,w),d,p)),n=s?n-t:n+t,w++;return i.toRegex===!0?t>1?sE(b,i):xm(y,null,{wrap:!1,...i}):y},lE=(r,e,t=1,i={})=>{if(!Ri(r)&&r.length>1||!Ri(e)&&e.length>1)return Sm(r,e,i);let n=i.transform||(p=>String.fromCharCode(p)),a=`${r}`.charCodeAt(0),s=`${e}`.charCodeAt(0),o=a>s,l=Math.min(a,s),c=Math.max(a,s);if(i.toRegex&&t===1)return vm(l,c,!1,i);let f=[],d=0;for(;o?a>=s:a<=s;)f.push(n(a,d)),a=o?a-t:a+t,d++;return i.toRegex===!0?xm(f,null,{wrap:!1,options:i}):f},hs=(r,e,t,i={})=>{if(e==null&&ll(r))return[r];if(!ll(r)||!ll(e))return Sm(r,e,i);if(typeof t=="function")return hs(r,e,1,{transform:t});if(bm(t))return hs(r,e,0,t);let n={...i};return n.capture===!0&&(n.wrap=!0),t=t||n.step||1,Ri(t)?Ri(r)&&Ri(e)?oE(r,e,t,n):lE(r,e,Math.max(Math.abs(t),1),n):t!=null&&!bm(t)?aE(t,n):hs(r,e,1,t)};Am.exports=hs});var Em=x((e6,_m)=>{u();"use strict";var uE=fl(),Cm=ps(),fE=(r,e={})=>{let t=(i,n={})=>{let a=Cm.isInvalidBrace(n),s=i.invalid===!0&&e.escapeInvalid===!0,o=a===!0||s===!0,l=e.escapeInvalid===!0?"\\":"",c="";if(i.isOpen===!0||i.isClose===!0)return l+i.value;if(i.type==="open")return o?l+i.value:"(";if(i.type==="close")return o?l+i.value:")";if(i.type==="comma")return i.prev.type==="comma"?"":o?i.value:"|";if(i.value)return i.value;if(i.nodes&&i.ranges>0){let f=Cm.reduce(i.nodes),d=uE(...f,{...e,wrap:!1,toRegex:!0});if(d.length!==0)return f.length>1&&d.length>1?`(${d})`:d}if(i.nodes)for(let f of i.nodes)c+=t(f,i);return c};return t(r)};_m.exports=fE});var Rm=x((t6,Tm)=>{u();"use strict";var cE=fl(),Om=ds(),mr=ps(),Gt=(r="",e="",t=!1)=>{let i=[];if(r=[].concat(r),e=[].concat(e),!e.length)return r;if(!r.length)return t?mr.flatten(e).map(n=>`{${n}}`):e;for(let n of r)if(Array.isArray(n))for(let a of n)i.push(Gt(a,e,t));else for(let a of e)t===!0&&typeof a=="string"&&(a=`{${a}}`),i.push(Array.isArray(a)?Gt(n,a,t):n+a);return mr.flatten(i)},pE=(r,e={})=>{let t=e.rangeLimit===void 0?1e3:e.rangeLimit,i=(n,a={})=>{n.queue=[];let s=a,o=a.queue;for(;s.type!=="brace"&&s.type!=="root"&&s.parent;)s=s.parent,o=s.queue;if(n.invalid||n.dollar){o.push(Gt(o.pop(),Om(n,e)));return}if(n.type==="brace"&&n.invalid!==!0&&n.nodes.length===2){o.push(Gt(o.pop(),["{}"]));return}if(n.nodes&&n.ranges>0){let d=mr.reduce(n.nodes);if(mr.exceedsLimit(...d,e.step,t))throw new RangeError("expanded array length exceeds range limit. Use options.rangeLimit to increase or disable the limit.");let p=cE(...d,e);p.length===0&&(p=Om(n,e)),o.push(Gt(o.pop(),p)),n.nodes=[];return}let l=mr.encloseBrace(n),c=n.queue,f=n;for(;f.type!=="brace"&&f.type!=="root"&&f.parent;)f=f.parent,c=f.queue;for(let d=0;d{u();"use strict";Pm.exports={MAX_LENGTH:1024*64,CHAR_0:"0",CHAR_9:"9",CHAR_UPPERCASE_A:"A",CHAR_LOWERCASE_A:"a",CHAR_UPPERCASE_Z:"Z",CHAR_LOWERCASE_Z:"z",CHAR_LEFT_PARENTHESES:"(",CHAR_RIGHT_PARENTHESES:")",CHAR_ASTERISK:"*",CHAR_AMPERSAND:"&",CHAR_AT:"@",CHAR_BACKSLASH:"\\",CHAR_BACKTICK:"`",CHAR_CARRIAGE_RETURN:"\r",CHAR_CIRCUMFLEX_ACCENT:"^",CHAR_COLON:":",CHAR_COMMA:",",CHAR_DOLLAR:"$",CHAR_DOT:".",CHAR_DOUBLE_QUOTE:'"',CHAR_EQUAL:"=",CHAR_EXCLAMATION_MARK:"!",CHAR_FORM_FEED:"\f",CHAR_FORWARD_SLASH:"/",CHAR_HASH:"#",CHAR_HYPHEN_MINUS:"-",CHAR_LEFT_ANGLE_BRACKET:"<",CHAR_LEFT_CURLY_BRACE:"{",CHAR_LEFT_SQUARE_BRACKET:"[",CHAR_LINE_FEED:` -`,CHAR_NO_BREAK_SPACE:"\xA0",CHAR_PERCENT:"%",CHAR_PLUS:"+",CHAR_QUESTION_MARK:"?",CHAR_RIGHT_ANGLE_BRACKET:">",CHAR_RIGHT_CURLY_BRACE:"}",CHAR_RIGHT_SQUARE_BRACKET:"]",CHAR_SEMICOLON:";",CHAR_SINGLE_QUOTE:"'",CHAR_SPACE:" ",CHAR_TAB:" ",CHAR_UNDERSCORE:"_",CHAR_VERTICAL_LINE:"|",CHAR_ZERO_WIDTH_NOBREAK_SPACE:"\uFEFF"}});var Mm=x((i6,Lm)=>{u();"use strict";var dE=ds(),{MAX_LENGTH:Dm,CHAR_BACKSLASH:cl,CHAR_BACKTICK:hE,CHAR_COMMA:mE,CHAR_DOT:gE,CHAR_LEFT_PARENTHESES:yE,CHAR_RIGHT_PARENTHESES:bE,CHAR_LEFT_CURLY_BRACE:wE,CHAR_RIGHT_CURLY_BRACE:vE,CHAR_LEFT_SQUARE_BRACKET:qm,CHAR_RIGHT_SQUARE_BRACKET:$m,CHAR_DOUBLE_QUOTE:xE,CHAR_SINGLE_QUOTE:kE,CHAR_NO_BREAK_SPACE:SE,CHAR_ZERO_WIDTH_NOBREAK_SPACE:AE}=Im(),CE=(r,e={})=>{if(typeof r!="string")throw new TypeError("Expected a string");let t=e||{},i=typeof t.maxLength=="number"?Math.min(Dm,t.maxLength):Dm;if(r.length>i)throw new SyntaxError(`Input length (${r.length}), exceeds max characters (${i})`);let n={type:"root",input:r,nodes:[]},a=[n],s=n,o=n,l=0,c=r.length,f=0,d=0,p,h={},b=()=>r[f++],v=y=>{if(y.type==="text"&&o.type==="dot"&&(o.type="text"),o&&o.type==="text"&&y.type==="text"){o.value+=y.value;return}return s.nodes.push(y),y.parent=s,y.prev=o,o=y,y};for(v({type:"bos"});f0){if(s.ranges>0){s.ranges=0;let y=s.nodes.shift();s.nodes=[y,{type:"text",value:dE(s)}]}v({type:"comma",value:p}),s.commas++;continue}if(p===gE&&d>0&&s.commas===0){let y=s.nodes;if(d===0||y.length===0){v({type:"text",value:p});continue}if(o.type==="dot"){if(s.range=[],o.value+=p,o.type="range",s.nodes.length!==3&&s.nodes.length!==5){s.invalid=!0,s.ranges=0,o.type="text";continue}s.ranges++,s.args=[];continue}if(o.type==="range"){y.pop();let w=y[y.length-1];w.value+=o.value+p,o=w,s.ranges--;continue}v({type:"dot",value:p});continue}v({type:"text",value:p})}do if(s=a.pop(),s.type!=="root"){s.nodes.forEach(k=>{k.nodes||(k.type==="open"&&(k.isOpen=!0),k.type==="close"&&(k.isClose=!0),k.nodes||(k.type="text"),k.invalid=!0)});let y=a[a.length-1],w=y.nodes.indexOf(s);y.nodes.splice(w,1,...s.nodes)}while(a.length>0);return v({type:"eos"}),n};Lm.exports=CE});var Fm=x((n6,Bm)=>{u();"use strict";var Nm=ds(),_E=Em(),EE=Rm(),OE=Mm(),Le=(r,e={})=>{let t=[];if(Array.isArray(r))for(let i of r){let n=Le.create(i,e);Array.isArray(n)?t.push(...n):t.push(n)}else t=[].concat(Le.create(r,e));return e&&e.expand===!0&&e.nodupes===!0&&(t=[...new Set(t)]),t};Le.parse=(r,e={})=>OE(r,e);Le.stringify=(r,e={})=>typeof r=="string"?Nm(Le.parse(r,e),e):Nm(r,e);Le.compile=(r,e={})=>(typeof r=="string"&&(r=Le.parse(r,e)),_E(r,e));Le.expand=(r,e={})=>{typeof r=="string"&&(r=Le.parse(r,e));let t=EE(r,e);return e.noempty===!0&&(t=t.filter(Boolean)),e.nodupes===!0&&(t=[...new Set(t)]),t};Le.create=(r,e={})=>r===""||r.length<3?[r]:e.expand!==!0?Le.compile(r,e):Le.expand(r,e);Bm.exports=Le});var Pi=x((s6,Hm)=>{u();"use strict";var TE=(et(),Ur),at="\\\\/",jm=`[^${at}]`,yt="\\.",RE="\\+",PE="\\?",ms="\\/",IE="(?=.)",zm="[^/]",pl=`(?:${ms}|$)`,Um=`(?:^|${ms})`,dl=`${yt}{1,2}${pl}`,DE=`(?!${yt})`,qE=`(?!${Um}${dl})`,$E=`(?!${yt}{0,1}${pl})`,LE=`(?!${dl})`,ME=`[^.${ms}]`,NE=`${zm}*?`,Vm={DOT_LITERAL:yt,PLUS_LITERAL:RE,QMARK_LITERAL:PE,SLASH_LITERAL:ms,ONE_CHAR:IE,QMARK:zm,END_ANCHOR:pl,DOTS_SLASH:dl,NO_DOT:DE,NO_DOTS:qE,NO_DOT_SLASH:$E,NO_DOTS_SLASH:LE,QMARK_NO_DOT:ME,STAR:NE,START_ANCHOR:Um},BE={...Vm,SLASH_LITERAL:`[${at}]`,QMARK:jm,STAR:`${jm}*?`,DOTS_SLASH:`${yt}{1,2}(?:[${at}]|$)`,NO_DOT:`(?!${yt})`,NO_DOTS:`(?!(?:^|[${at}])${yt}{1,2}(?:[${at}]|$))`,NO_DOT_SLASH:`(?!${yt}{0,1}(?:[${at}]|$))`,NO_DOTS_SLASH:`(?!${yt}{1,2}(?:[${at}]|$))`,QMARK_NO_DOT:`[^.${at}]`,START_ANCHOR:`(?:^|[${at}])`,END_ANCHOR:`(?:[${at}]|$)`},FE={alnum:"a-zA-Z0-9",alpha:"a-zA-Z",ascii:"\\x00-\\x7F",blank:" \\t",cntrl:"\\x00-\\x1F\\x7F",digit:"0-9",graph:"\\x21-\\x7E",lower:"a-z",print:"\\x20-\\x7E ",punct:"\\-!\"#$%&'()\\*+,./:;<=>?@[\\]^_`{|}~",space:" \\t\\r\\n\\v\\f",upper:"A-Z",word:"A-Za-z0-9_",xdigit:"A-Fa-f0-9"};Hm.exports={MAX_LENGTH:1024*64,POSIX_REGEX_SOURCE:FE,REGEX_BACKSLASH:/\\(?![*+?^${}(|)[\]])/g,REGEX_NON_SPECIAL_CHARS:/^[^@![\].,$*+?^{}()|\\/]+/,REGEX_SPECIAL_CHARS:/[-*+?.^${}(|)[\]]/,REGEX_SPECIAL_CHARS_BACKREF:/(\\?)((\W)(\3*))/g,REGEX_SPECIAL_CHARS_GLOBAL:/([-*+?.^${}(|)[\]])/g,REGEX_REMOVE_BACKSLASH:/(?:\[.*?[^\\]\]|\\(?=.))/g,REPLACEMENTS:{"***":"*","**/**":"**","**/**/**":"**"},CHAR_0:48,CHAR_9:57,CHAR_UPPERCASE_A:65,CHAR_LOWERCASE_A:97,CHAR_UPPERCASE_Z:90,CHAR_LOWERCASE_Z:122,CHAR_LEFT_PARENTHESES:40,CHAR_RIGHT_PARENTHESES:41,CHAR_ASTERISK:42,CHAR_AMPERSAND:38,CHAR_AT:64,CHAR_BACKWARD_SLASH:92,CHAR_CARRIAGE_RETURN:13,CHAR_CIRCUMFLEX_ACCENT:94,CHAR_COLON:58,CHAR_COMMA:44,CHAR_DOT:46,CHAR_DOUBLE_QUOTE:34,CHAR_EQUAL:61,CHAR_EXCLAMATION_MARK:33,CHAR_FORM_FEED:12,CHAR_FORWARD_SLASH:47,CHAR_GRAVE_ACCENT:96,CHAR_HASH:35,CHAR_HYPHEN_MINUS:45,CHAR_LEFT_ANGLE_BRACKET:60,CHAR_LEFT_CURLY_BRACE:123,CHAR_LEFT_SQUARE_BRACKET:91,CHAR_LINE_FEED:10,CHAR_NO_BREAK_SPACE:160,CHAR_PERCENT:37,CHAR_PLUS:43,CHAR_QUESTION_MARK:63,CHAR_RIGHT_ANGLE_BRACKET:62,CHAR_RIGHT_CURLY_BRACE:125,CHAR_RIGHT_SQUARE_BRACKET:93,CHAR_SEMICOLON:59,CHAR_SINGLE_QUOTE:39,CHAR_SPACE:32,CHAR_TAB:9,CHAR_UNDERSCORE:95,CHAR_VERTICAL_LINE:124,CHAR_ZERO_WIDTH_NOBREAK_SPACE:65279,SEP:TE.sep,extglobChars(r){return{"!":{type:"negate",open:"(?:(?!(?:",close:`))${r.STAR})`},"?":{type:"qmark",open:"(?:",close:")?"},"+":{type:"plus",open:"(?:",close:")+"},"*":{type:"star",open:"(?:",close:")*"},"@":{type:"at",open:"(?:",close:")"}}},globChars(r){return r===!0?BE:Vm}}});var Ii=x(Re=>{u();"use strict";var jE=(et(),Ur),zE=m.platform==="win32",{REGEX_BACKSLASH:UE,REGEX_REMOVE_BACKSLASH:VE,REGEX_SPECIAL_CHARS:HE,REGEX_SPECIAL_CHARS_GLOBAL:WE}=Pi();Re.isObject=r=>r!==null&&typeof r=="object"&&!Array.isArray(r);Re.hasRegexChars=r=>HE.test(r);Re.isRegexChar=r=>r.length===1&&Re.hasRegexChars(r);Re.escapeRegex=r=>r.replace(WE,"\\$1");Re.toPosixSlashes=r=>r.replace(UE,"/");Re.removeBackslashes=r=>r.replace(VE,e=>e==="\\"?"":e);Re.supportsLookbehinds=()=>{let r=m.version.slice(1).split(".").map(Number);return r.length===3&&r[0]>=9||r[0]===8&&r[1]>=10};Re.isWindows=r=>r&&typeof r.windows=="boolean"?r.windows:zE===!0||jE.sep==="\\";Re.escapeLast=(r,e,t)=>{let i=r.lastIndexOf(e,t);return i===-1?r:r[i-1]==="\\"?Re.escapeLast(r,e,i-1):`${r.slice(0,i)}\\${r.slice(i)}`};Re.removePrefix=(r,e={})=>{let t=r;return t.startsWith("./")&&(t=t.slice(2),e.prefix="./"),t};Re.wrapOutput=(r,e={},t={})=>{let i=t.contains?"":"^",n=t.contains?"":"$",a=`${i}(?:${r})${n}`;return e.negated===!0&&(a=`(?:^(?!${a}).*$)`),a}});var Zm=x((o6,Jm)=>{u();"use strict";var Wm=Ii(),{CHAR_ASTERISK:hl,CHAR_AT:GE,CHAR_BACKWARD_SLASH:Di,CHAR_COMMA:QE,CHAR_DOT:ml,CHAR_EXCLAMATION_MARK:gl,CHAR_FORWARD_SLASH:Gm,CHAR_LEFT_CURLY_BRACE:yl,CHAR_LEFT_PARENTHESES:bl,CHAR_LEFT_SQUARE_BRACKET:YE,CHAR_PLUS:KE,CHAR_QUESTION_MARK:Qm,CHAR_RIGHT_CURLY_BRACE:XE,CHAR_RIGHT_PARENTHESES:Ym,CHAR_RIGHT_SQUARE_BRACKET:JE}=Pi(),Km=r=>r===Gm||r===Di,Xm=r=>{r.isPrefix!==!0&&(r.depth=r.isGlobstar?1/0:1)},ZE=(r,e)=>{let t=e||{},i=r.length-1,n=t.parts===!0||t.scanToEnd===!0,a=[],s=[],o=[],l=r,c=-1,f=0,d=0,p=!1,h=!1,b=!1,v=!1,y=!1,w=!1,k=!1,S=!1,E=!1,O=!1,B=0,N,T,F={value:"",depth:0,isGlob:!1},Y=()=>c>=i,_=()=>l.charCodeAt(c+1),Q=()=>(N=T,l.charCodeAt(++c));for(;c0&&(oe=l.slice(0,f),l=l.slice(f),d-=f),U&&b===!0&&d>0?(U=l.slice(0,d),A=l.slice(d)):b===!0?(U="",A=l):U=l,U&&U!==""&&U!=="/"&&U!==l&&Km(U.charCodeAt(U.length-1))&&(U=U.slice(0,-1)),t.unescape===!0&&(A&&(A=Wm.removeBackslashes(A)),U&&k===!0&&(U=Wm.removeBackslashes(U)));let C={prefix:oe,input:r,start:f,base:U,glob:A,isBrace:p,isBracket:h,isGlob:b,isExtglob:v,isGlobstar:y,negated:S,negatedExtglob:E};if(t.tokens===!0&&(C.maxDepth=0,Km(T)||s.push(F),C.tokens=s),t.parts===!0||t.tokens===!0){let he;for(let V=0;V{u();"use strict";var gs=Pi(),Me=Ii(),{MAX_LENGTH:ys,POSIX_REGEX_SOURCE:e2,REGEX_NON_SPECIAL_CHARS:t2,REGEX_SPECIAL_CHARS_BACKREF:r2,REPLACEMENTS:eg}=gs,i2=(r,e)=>{if(typeof e.expandRange=="function")return e.expandRange(...r,e);r.sort();let t=`[${r.join("-")}]`;try{new RegExp(t)}catch(i){return r.map(n=>Me.escapeRegex(n)).join("..")}return t},gr=(r,e)=>`Missing ${r}: "${e}" - use "\\\\${e}" to match literal characters`,wl=(r,e)=>{if(typeof r!="string")throw new TypeError("Expected a string");r=eg[r]||r;let t={...e},i=typeof t.maxLength=="number"?Math.min(ys,t.maxLength):ys,n=r.length;if(n>i)throw new SyntaxError(`Input length: ${n}, exceeds maximum allowed length: ${i}`);let a={type:"bos",value:"",output:t.prepend||""},s=[a],o=t.capture?"":"?:",l=Me.isWindows(e),c=gs.globChars(l),f=gs.extglobChars(c),{DOT_LITERAL:d,PLUS_LITERAL:p,SLASH_LITERAL:h,ONE_CHAR:b,DOTS_SLASH:v,NO_DOT:y,NO_DOT_SLASH:w,NO_DOTS_SLASH:k,QMARK:S,QMARK_NO_DOT:E,STAR:O,START_ANCHOR:B}=c,N=q=>`(${o}(?:(?!${B}${q.dot?v:d}).)*?)`,T=t.dot?"":y,F=t.dot?S:E,Y=t.bash===!0?N(t):O;t.capture&&(Y=`(${Y})`),typeof t.noext=="boolean"&&(t.noextglob=t.noext);let _={input:r,index:-1,start:0,dot:t.dot===!0,consumed:"",output:"",prefix:"",backtrack:!1,negated:!1,brackets:0,braces:0,parens:0,quotes:0,globstar:!1,tokens:s};r=Me.removePrefix(r,_),n=r.length;let Q=[],U=[],oe=[],A=a,C,he=()=>_.index===n-1,V=_.peek=(q=1)=>r[_.index+q],Ee=_.advance=()=>r[++_.index]||"",Ie=()=>r.slice(_.index+1),De=(q="",ae=0)=>{_.consumed+=q,_.index+=ae},Bi=q=>{_.output+=q.output!=null?q.output:q.value,De(q.value)},Rv=()=>{let q=1;for(;V()==="!"&&(V(2)!=="("||V(3)==="?");)Ee(),_.start++,q++;return q%2==0?!1:(_.negated=!0,_.start++,!0)},Fi=q=>{_[q]++,oe.push(q)},Ft=q=>{_[q]--,oe.pop()},W=q=>{if(A.type==="globstar"){let ae=_.braces>0&&(q.type==="comma"||q.type==="brace"),I=q.extglob===!0||Q.length&&(q.type==="pipe"||q.type==="paren");q.type!=="slash"&&q.type!=="paren"&&!ae&&!I&&(_.output=_.output.slice(0,-A.output.length),A.type="star",A.value="*",A.output=Y,_.output+=A.output)}if(Q.length&&q.type!=="paren"&&(Q[Q.length-1].inner+=q.value),(q.value||q.output)&&Bi(q),A&&A.type==="text"&&q.type==="text"){A.value+=q.value,A.output=(A.output||"")+q.value;return}q.prev=A,s.push(q),A=q},ji=(q,ae)=>{let I={...f[ae],conditions:1,inner:""};I.prev=A,I.parens=_.parens,I.output=_.output;let H=(t.capture?"(":"")+I.open;Fi("parens"),W({type:q,value:ae,output:_.output?"":b}),W({type:"paren",extglob:!0,value:Ee(),output:H}),Q.push(I)},Pv=q=>{let ae=q.close+(t.capture?")":""),I;if(q.type==="negate"){let H=Y;if(q.inner&&q.inner.length>1&&q.inner.includes("/")&&(H=N(t)),(H!==Y||he()||/^\)+$/.test(Ie()))&&(ae=q.close=`)$))${H}`),q.inner.includes("*")&&(I=Ie())&&/^\.[^\\/.]+$/.test(I)){let ce=wl(I,{...e,fastpaths:!1}).output;ae=q.close=`)${ce})${H})`}q.prev.type==="bos"&&(_.negatedExtglob=!0)}W({type:"paren",extglob:!0,value:C,output:ae}),Ft("parens")};if(t.fastpaths!==!1&&!/(^[*!]|[/()[\]{}"])/.test(r)){let q=!1,ae=r.replace(r2,(I,H,ce,Ce,ye,Ms)=>Ce==="\\"?(q=!0,I):Ce==="?"?H?H+Ce+(ye?S.repeat(ye.length):""):Ms===0?F+(ye?S.repeat(ye.length):""):S.repeat(ce.length):Ce==="."?d.repeat(ce.length):Ce==="*"?H?H+Ce+(ye?Y:""):Y:H?I:`\\${I}`);return q===!0&&(t.unescape===!0?ae=ae.replace(/\\/g,""):ae=ae.replace(/\\+/g,I=>I.length%2==0?"\\\\":I?"\\":"")),ae===r&&t.contains===!0?(_.output=r,_):(_.output=Me.wrapOutput(ae,_,e),_)}for(;!he();){if(C=Ee(),C==="\0")continue;if(C==="\\"){let I=V();if(I==="/"&&t.bash!==!0||I==="."||I===";")continue;if(!I){C+="\\",W({type:"text",value:C});continue}let H=/^\\+/.exec(Ie()),ce=0;if(H&&H[0].length>2&&(ce=H[0].length,_.index+=ce,ce%2!=0&&(C+="\\")),t.unescape===!0?C=Ee():C+=Ee(),_.brackets===0){W({type:"text",value:C});continue}}if(_.brackets>0&&(C!=="]"||A.value==="["||A.value==="[^")){if(t.posix!==!1&&C===":"){let I=A.value.slice(1);if(I.includes("[")&&(A.posix=!0,I.includes(":"))){let H=A.value.lastIndexOf("["),ce=A.value.slice(0,H),Ce=A.value.slice(H+2),ye=e2[Ce];if(ye){A.value=ce+ye,_.backtrack=!0,Ee(),!a.output&&s.indexOf(A)===1&&(a.output=b);continue}}}(C==="["&&V()!==":"||C==="-"&&V()==="]")&&(C=`\\${C}`),C==="]"&&(A.value==="["||A.value==="[^")&&(C=`\\${C}`),t.posix===!0&&C==="!"&&A.value==="["&&(C="^"),A.value+=C,Bi({value:C});continue}if(_.quotes===1&&C!=='"'){C=Me.escapeRegex(C),A.value+=C,Bi({value:C});continue}if(C==='"'){_.quotes=_.quotes===1?0:1,t.keepQuotes===!0&&W({type:"text",value:C});continue}if(C==="("){Fi("parens"),W({type:"paren",value:C});continue}if(C===")"){if(_.parens===0&&t.strictBrackets===!0)throw new SyntaxError(gr("opening","("));let I=Q[Q.length-1];if(I&&_.parens===I.parens+1){Pv(Q.pop());continue}W({type:"paren",value:C,output:_.parens?")":"\\)"}),Ft("parens");continue}if(C==="["){if(t.nobracket===!0||!Ie().includes("]")){if(t.nobracket!==!0&&t.strictBrackets===!0)throw new SyntaxError(gr("closing","]"));C=`\\${C}`}else Fi("brackets");W({type:"bracket",value:C});continue}if(C==="]"){if(t.nobracket===!0||A&&A.type==="bracket"&&A.value.length===1){W({type:"text",value:C,output:`\\${C}`});continue}if(_.brackets===0){if(t.strictBrackets===!0)throw new SyntaxError(gr("opening","["));W({type:"text",value:C,output:`\\${C}`});continue}Ft("brackets");let I=A.value.slice(1);if(A.posix!==!0&&I[0]==="^"&&!I.includes("/")&&(C=`/${C}`),A.value+=C,Bi({value:C}),t.literalBrackets===!1||Me.hasRegexChars(I))continue;let H=Me.escapeRegex(A.value);if(_.output=_.output.slice(0,-A.value.length),t.literalBrackets===!0){_.output+=H,A.value=H;continue}A.value=`(${o}${H}|${A.value})`,_.output+=A.value;continue}if(C==="{"&&t.nobrace!==!0){Fi("braces");let I={type:"brace",value:C,output:"(",outputIndex:_.output.length,tokensIndex:_.tokens.length};U.push(I),W(I);continue}if(C==="}"){let I=U[U.length-1];if(t.nobrace===!0||!I){W({type:"text",value:C,output:C});continue}let H=")";if(I.dots===!0){let ce=s.slice(),Ce=[];for(let ye=ce.length-1;ye>=0&&(s.pop(),ce[ye].type!=="brace");ye--)ce[ye].type!=="dots"&&Ce.unshift(ce[ye].value);H=i2(Ce,t),_.backtrack=!0}if(I.comma!==!0&&I.dots!==!0){let ce=_.output.slice(0,I.outputIndex),Ce=_.tokens.slice(I.tokensIndex);I.value=I.output="\\{",C=H="\\}",_.output=ce;for(let ye of Ce)_.output+=ye.output||ye.value}W({type:"brace",value:C,output:H}),Ft("braces"),U.pop();continue}if(C==="|"){Q.length>0&&Q[Q.length-1].conditions++,W({type:"text",value:C});continue}if(C===","){let I=C,H=U[U.length-1];H&&oe[oe.length-1]==="braces"&&(H.comma=!0,I="|"),W({type:"comma",value:C,output:I});continue}if(C==="/"){if(A.type==="dot"&&_.index===_.start+1){_.start=_.index+1,_.consumed="",_.output="",s.pop(),A=a;continue}W({type:"slash",value:C,output:h});continue}if(C==="."){if(_.braces>0&&A.type==="dot"){A.value==="."&&(A.output=d);let I=U[U.length-1];A.type="dots",A.output+=C,A.value+=C,I.dots=!0;continue}if(_.braces+_.parens===0&&A.type!=="bos"&&A.type!=="slash"){W({type:"text",value:C,output:d});continue}W({type:"dot",value:C,output:d});continue}if(C==="?"){if(!(A&&A.value==="(")&&t.noextglob!==!0&&V()==="("&&V(2)!=="?"){ji("qmark",C);continue}if(A&&A.type==="paren"){let H=V(),ce=C;if(H==="<"&&!Me.supportsLookbehinds())throw new Error("Node.js v10 or higher is required for regex lookbehinds");(A.value==="("&&!/[!=<:]/.test(H)||H==="<"&&!/<([!=]|\w+>)/.test(Ie()))&&(ce=`\\${C}`),W({type:"text",value:C,output:ce});continue}if(t.dot!==!0&&(A.type==="slash"||A.type==="bos")){W({type:"qmark",value:C,output:E});continue}W({type:"qmark",value:C,output:S});continue}if(C==="!"){if(t.noextglob!==!0&&V()==="("&&(V(2)!=="?"||!/[!=<:]/.test(V(3)))){ji("negate",C);continue}if(t.nonegate!==!0&&_.index===0){Rv();continue}}if(C==="+"){if(t.noextglob!==!0&&V()==="("&&V(2)!=="?"){ji("plus",C);continue}if(A&&A.value==="("||t.regex===!1){W({type:"plus",value:C,output:p});continue}if(A&&(A.type==="bracket"||A.type==="paren"||A.type==="brace")||_.parens>0){W({type:"plus",value:C});continue}W({type:"plus",value:p});continue}if(C==="@"){if(t.noextglob!==!0&&V()==="("&&V(2)!=="?"){W({type:"at",extglob:!0,value:C,output:""});continue}W({type:"text",value:C});continue}if(C!=="*"){(C==="$"||C==="^")&&(C=`\\${C}`);let I=t2.exec(Ie());I&&(C+=I[0],_.index+=I[0].length),W({type:"text",value:C});continue}if(A&&(A.type==="globstar"||A.star===!0)){A.type="star",A.star=!0,A.value+=C,A.output=Y,_.backtrack=!0,_.globstar=!0,De(C);continue}let q=Ie();if(t.noextglob!==!0&&/^\([^?]/.test(q)){ji("star",C);continue}if(A.type==="star"){if(t.noglobstar===!0){De(C);continue}let I=A.prev,H=I.prev,ce=I.type==="slash"||I.type==="bos",Ce=H&&(H.type==="star"||H.type==="globstar");if(t.bash===!0&&(!ce||q[0]&&q[0]!=="/")){W({type:"star",value:C,output:""});continue}let ye=_.braces>0&&(I.type==="comma"||I.type==="brace"),Ms=Q.length&&(I.type==="pipe"||I.type==="paren");if(!ce&&I.type!=="paren"&&!ye&&!Ms){W({type:"star",value:C,output:""});continue}for(;q.slice(0,3)==="/**";){let zi=r[_.index+4];if(zi&&zi!=="/")break;q=q.slice(3),De("/**",3)}if(I.type==="bos"&&he()){A.type="globstar",A.value+=C,A.output=N(t),_.output=A.output,_.globstar=!0,De(C);continue}if(I.type==="slash"&&I.prev.type!=="bos"&&!Ce&&he()){_.output=_.output.slice(0,-(I.output+A.output).length),I.output=`(?:${I.output}`,A.type="globstar",A.output=N(t)+(t.strictSlashes?")":"|$)"),A.value+=C,_.globstar=!0,_.output+=I.output+A.output,De(C);continue}if(I.type==="slash"&&I.prev.type!=="bos"&&q[0]==="/"){let zi=q[1]!==void 0?"|$":"";_.output=_.output.slice(0,-(I.output+A.output).length),I.output=`(?:${I.output}`,A.type="globstar",A.output=`${N(t)}${h}|${h}${zi})`,A.value+=C,_.output+=I.output+A.output,_.globstar=!0,De(C+Ee()),W({type:"slash",value:"/",output:""});continue}if(I.type==="bos"&&q[0]==="/"){A.type="globstar",A.value+=C,A.output=`(?:^|${h}|${N(t)}${h})`,_.output=A.output,_.globstar=!0,De(C+Ee()),W({type:"slash",value:"/",output:""});continue}_.output=_.output.slice(0,-A.output.length),A.type="globstar",A.output=N(t),A.value+=C,_.output+=A.output,_.globstar=!0,De(C);continue}let ae={type:"star",value:C,output:Y};if(t.bash===!0){ae.output=".*?",(A.type==="bos"||A.type==="slash")&&(ae.output=T+ae.output),W(ae);continue}if(A&&(A.type==="bracket"||A.type==="paren")&&t.regex===!0){ae.output=C,W(ae);continue}(_.index===_.start||A.type==="slash"||A.type==="dot")&&(A.type==="dot"?(_.output+=w,A.output+=w):t.dot===!0?(_.output+=k,A.output+=k):(_.output+=T,A.output+=T),V()!=="*"&&(_.output+=b,A.output+=b)),W(ae)}for(;_.brackets>0;){if(t.strictBrackets===!0)throw new SyntaxError(gr("closing","]"));_.output=Me.escapeLast(_.output,"["),Ft("brackets")}for(;_.parens>0;){if(t.strictBrackets===!0)throw new SyntaxError(gr("closing",")"));_.output=Me.escapeLast(_.output,"("),Ft("parens")}for(;_.braces>0;){if(t.strictBrackets===!0)throw new SyntaxError(gr("closing","}"));_.output=Me.escapeLast(_.output,"{"),Ft("braces")}if(t.strictSlashes!==!0&&(A.type==="star"||A.type==="bracket")&&W({type:"maybe_slash",value:"",output:`${h}?`}),_.backtrack===!0){_.output="";for(let q of _.tokens)_.output+=q.output!=null?q.output:q.value,q.suffix&&(_.output+=q.suffix)}return _};wl.fastpaths=(r,e)=>{let t={...e},i=typeof t.maxLength=="number"?Math.min(ys,t.maxLength):ys,n=r.length;if(n>i)throw new SyntaxError(`Input length: ${n}, exceeds maximum allowed length: ${i}`);r=eg[r]||r;let a=Me.isWindows(e),{DOT_LITERAL:s,SLASH_LITERAL:o,ONE_CHAR:l,DOTS_SLASH:c,NO_DOT:f,NO_DOTS:d,NO_DOTS_SLASH:p,STAR:h,START_ANCHOR:b}=gs.globChars(a),v=t.dot?d:f,y=t.dot?p:f,w=t.capture?"":"?:",k={negated:!1,prefix:""},S=t.bash===!0?".*?":h;t.capture&&(S=`(${S})`);let E=T=>T.noglobstar===!0?S:`(${w}(?:(?!${b}${T.dot?c:s}).)*?)`,O=T=>{switch(T){case"*":return`${v}${l}${S}`;case".*":return`${s}${l}${S}`;case"*.*":return`${v}${S}${s}${l}${S}`;case"*/*":return`${v}${S}${o}${l}${y}${S}`;case"**":return v+E(t);case"**/*":return`(?:${v}${E(t)}${o})?${y}${l}${S}`;case"**/*.*":return`(?:${v}${E(t)}${o})?${y}${S}${s}${l}${S}`;case"**/.*":return`(?:${v}${E(t)}${o})?${s}${l}${S}`;default:{let F=/^(.*?)\.(\w+)$/.exec(T);if(!F)return;let Y=O(F[1]);return Y?Y+s+F[2]:void 0}}},B=Me.removePrefix(r,k),N=O(B);return N&&t.strictSlashes!==!0&&(N+=`${o}?`),N};tg.exports=wl});var ng=x((u6,ig)=>{u();"use strict";var n2=(et(),Ur),s2=Zm(),vl=rg(),xl=Ii(),a2=Pi(),o2=r=>r&&typeof r=="object"&&!Array.isArray(r),de=(r,e,t=!1)=>{if(Array.isArray(r)){let f=r.map(p=>de(p,e,t));return p=>{for(let h of f){let b=h(p);if(b)return b}return!1}}let i=o2(r)&&r.tokens&&r.input;if(r===""||typeof r!="string"&&!i)throw new TypeError("Expected pattern to be a non-empty string");let n=e||{},a=xl.isWindows(e),s=i?de.compileRe(r,e):de.makeRe(r,e,!1,!0),o=s.state;delete s.state;let l=()=>!1;if(n.ignore){let f={...e,ignore:null,onMatch:null,onResult:null};l=de(n.ignore,f,t)}let c=(f,d=!1)=>{let{isMatch:p,match:h,output:b}=de.test(f,s,e,{glob:r,posix:a}),v={glob:r,state:o,regex:s,posix:a,input:f,output:b,match:h,isMatch:p};return typeof n.onResult=="function"&&n.onResult(v),p===!1?(v.isMatch=!1,d?v:!1):l(f)?(typeof n.onIgnore=="function"&&n.onIgnore(v),v.isMatch=!1,d?v:!1):(typeof n.onMatch=="function"&&n.onMatch(v),d?v:!0)};return t&&(c.state=o),c};de.test=(r,e,t,{glob:i,posix:n}={})=>{if(typeof r!="string")throw new TypeError("Expected input to be a string");if(r==="")return{isMatch:!1,output:""};let a=t||{},s=a.format||(n?xl.toPosixSlashes:null),o=r===i,l=o&&s?s(r):r;return o===!1&&(l=s?s(r):r,o=l===i),(o===!1||a.capture===!0)&&(a.matchBase===!0||a.basename===!0?o=de.matchBase(r,e,t,n):o=e.exec(l)),{isMatch:Boolean(o),match:o,output:l}};de.matchBase=(r,e,t,i=xl.isWindows(t))=>(e instanceof RegExp?e:de.makeRe(e,t)).test(n2.basename(r));de.isMatch=(r,e,t)=>de(e,t)(r);de.parse=(r,e)=>Array.isArray(r)?r.map(t=>de.parse(t,e)):vl(r,{...e,fastpaths:!1});de.scan=(r,e)=>s2(r,e);de.compileRe=(r,e,t=!1,i=!1)=>{if(t===!0)return r.output;let n=e||{},a=n.contains?"":"^",s=n.contains?"":"$",o=`${a}(?:${r.output})${s}`;r&&r.negated===!0&&(o=`^(?!${o}).*$`);let l=de.toRegex(o,e);return i===!0&&(l.state=r),l};de.makeRe=(r,e={},t=!1,i=!1)=>{if(!r||typeof r!="string")throw new TypeError("Expected a non-empty string");let n={negated:!1,fastpaths:!0};return e.fastpaths!==!1&&(r[0]==="."||r[0]==="*")&&(n.output=vl.fastpaths(r,e)),n.output||(n=vl(r,e)),de.compileRe(n,e,t,i)};de.toRegex=(r,e)=>{try{let t=e||{};return new RegExp(r,t.flags||(t.nocase?"i":""))}catch(t){if(e&&e.debug===!0)throw t;return/$^/}};de.constants=a2;ig.exports=de});var ag=x((f6,sg)=>{u();"use strict";sg.exports=ng()});var cg=x((c6,fg)=>{u();"use strict";var og=(Bn(),Nn),lg=Fm(),ot=ag(),kl=Ii(),ug=r=>r===""||r==="./",fe=(r,e,t)=>{e=[].concat(e),r=[].concat(r);let i=new Set,n=new Set,a=new Set,s=0,o=f=>{a.add(f.output),t&&t.onResult&&t.onResult(f)};for(let f=0;f!i.has(f));if(t&&c.length===0){if(t.failglob===!0)throw new Error(`No matches found for "${e.join(", ")}"`);if(t.nonull===!0||t.nullglob===!0)return t.unescape?e.map(f=>f.replace(/\\/g,"")):e}return c};fe.match=fe;fe.matcher=(r,e)=>ot(r,e);fe.isMatch=(r,e,t)=>ot(e,t)(r);fe.any=fe.isMatch;fe.not=(r,e,t={})=>{e=[].concat(e).map(String);let i=new Set,n=[],a=o=>{t.onResult&&t.onResult(o),n.push(o.output)},s=new Set(fe(r,e,{...t,onResult:a}));for(let o of n)s.has(o)||i.add(o);return[...i]};fe.contains=(r,e,t)=>{if(typeof r!="string")throw new TypeError(`Expected a string: "${og.inspect(r)}"`);if(Array.isArray(e))return e.some(i=>fe.contains(r,i,t));if(typeof e=="string"){if(ug(r)||ug(e))return!1;if(r.includes(e)||r.startsWith("./")&&r.slice(2).includes(e))return!0}return fe.isMatch(r,e,{...t,contains:!0})};fe.matchKeys=(r,e,t)=>{if(!kl.isObject(r))throw new TypeError("Expected the first argument to be an object");let i=fe(Object.keys(r),e,t),n={};for(let a of i)n[a]=r[a];return n};fe.some=(r,e,t)=>{let i=[].concat(r);for(let n of[].concat(e)){let a=ot(String(n),t);if(i.some(s=>a(s)))return!0}return!1};fe.every=(r,e,t)=>{let i=[].concat(r);for(let n of[].concat(e)){let a=ot(String(n),t);if(!i.every(s=>a(s)))return!1}return!0};fe.all=(r,e,t)=>{if(typeof r!="string")throw new TypeError(`Expected a string: "${og.inspect(r)}"`);return[].concat(e).every(i=>ot(i,t)(r))};fe.capture=(r,e,t)=>{let i=kl.isWindows(t),a=ot.makeRe(String(r),{...t,capture:!0}).exec(i?kl.toPosixSlashes(e):e);if(a)return a.slice(1).map(s=>s===void 0?"":s)};fe.makeRe=(...r)=>ot.makeRe(...r);fe.scan=(...r)=>ot.scan(...r);fe.parse=(r,e)=>{let t=[];for(let i of[].concat(r||[]))for(let n of lg(String(i),e))t.push(ot.parse(n,e));return t};fe.braces=(r,e)=>{if(typeof r!="string")throw new TypeError("Expected a string");return e&&e.nobrace===!0||!/\{.*\}/.test(r)?[r]:lg(r,e)};fe.braceExpand=(r,e)=>{if(typeof r!="string")throw new TypeError("Expected a string");return fe.braces(r,{...e,expand:!0})};fg.exports=fe});function dg(r,e){let t=e.content.files;t=t.filter(o=>typeof o=="string"),t=t.map(al);let i=cs.generateTasks(t),n=[],a=[];for(let o of i)n.push(...o.positive.map(l=>hg(l,!1))),a.push(...o.negative.map(l=>hg(l,!0)));let s=[...n,...a];return s=u2(r,s),s=s.flatMap(f2),s=s.map(l2),s}function hg(r,e){let t={original:r,base:r,ignore:e,pattern:r,glob:null};return Kh(r)&&Object.assign(t,rm(r)),t}function l2(r){let e=al(r.base);return e=cs.escapePath(e),r.pattern=r.glob?`${e}/${r.glob}`:e,r.pattern=r.ignore?`!${r.pattern}`:r.pattern,r}function u2(r,e){let t=[];return r.userConfigPath&&r.tailwindConfig.content.relative&&(t=[me.dirname(r.userConfigPath)]),e.map(i=>(i.base=me.resolve(...t,i.base),i))}function f2(r){let e=[r];try{let t=be.realpathSync(r.base);t!==r.base&&e.push({...r,base:t})}catch{}return e}function mg(r,e,t){let i=r.tailwindConfig.content.files.filter(s=>typeof s.raw=="string").map(({raw:s,extension:o="html"})=>({content:s,extension:o})),[n,a]=p2(e,t);for(let s of n){let o=me.extname(s).slice(1);i.push({file:s,extension:o})}return[i,a]}function c2(r){if(!r.some(a=>a.includes("**")&&!yg.test(a)))return()=>{};let t=[],i=[];for(let a of r){let s=pg.default.matcher(a);yg.test(a)&&i.push(s),t.push(s)}let n=!1;return a=>{if(n||i.some(f=>f(a)))return;let s=t.findIndex(f=>f(a));if(s===-1)return;let o=r[s],l=me.relative(m.cwd(),o);l[0]!=="."&&(l=`./${l}`);let c=gg.find(f=>a.includes(f));c&&(n=!0,G.warn("broad-content-glob-pattern",[`Your \`content\` configuration includes a pattern which looks like it's accidentally matching all of \`${c}\` and can cause serious performance issues.`,`Pattern: \`${l}\``,"See our documentation for recommendations:","https://tailwindcss.com/docs/content-configuration#pattern-recommendations"]))}}function p2(r,e){let t=r.map(o=>o.pattern),i=new Map,n=c2(t),a=new Set;Je.DEBUG&&console.time("Finding changed files");let s=cs.sync(t,{absolute:!0});for(let o of s){n(o);let l=e.get(o)||-1/0,c=be.statSync(o).mtimeMs;c>l&&(a.add(o),i.set(o,c))}return Je.DEBUG&&console.timeEnd("Finding changed files"),[a,i]}var pg,gg,yg,bg=R(()=>{u();ft();et();Xh();Jh();Zh();im();It();Be();pg=pe(cg());gg=["node_modules"],yg=new RegExp(`(${gg.map(r=>String.raw`\b${r}\b`).join("|")})`)});function wg(){}var vg=R(()=>{u()});function g2(r,e){for(let t of e){let i=`${r}${t}`;if(be.existsSync(i)&&be.statSync(i).isFile())return i}for(let t of e){let i=`${r}/index${t}`;if(be.existsSync(i))return i}return null}function*xg(r,e,t,i=me.extname(r)){let n=g2(me.resolve(e,r),d2.includes(i)?h2:m2);if(n===null||t.has(n))return;t.add(n),yield n,e=me.dirname(n),i=me.extname(n);let a=be.readFileSync(n,"utf-8");for(let s of[...a.matchAll(/import[\s\S]*?['"](.{3,}?)['"]/gi),...a.matchAll(/import[\s\S]*from[\s\S]*?['"](.{3,}?)['"]/gi),...a.matchAll(/require\(['"`](.+)['"`]\)/gi)])!s[1].startsWith(".")||(yield*xg(s[1],e,t,i))}function Sl(r){return r===null?new Set:new Set(xg(r,me.dirname(r),new Set))}var d2,h2,m2,kg=R(()=>{u();ft();et();d2=[".js",".cjs",".mjs"],h2=["",".js",".cjs",".mjs",".ts",".cts",".mts",".jsx",".tsx"],m2=["",".ts",".cts",".mts",".tsx",".js",".cjs",".mjs",".jsx"]});function y2(r,e){if(Al.has(r))return Al.get(r);let t=dg(r,e);return Al.set(r,t).get(r)}function b2(r){let e=na(r);if(e!==null){let[i,n,a,s]=Ag.get(e)||[],o=Sl(e),l=!1,c=new Map;for(let p of o){let h=be.statSync(p).mtimeMs;c.set(p,h),(!s||!s.has(p)||h>s.get(p))&&(l=!0)}if(!l)return[i,e,n,a];for(let p of o)delete pf.cache[p];let f=sl(zr(wg(e))),d=Vi(f);return Ag.set(e,[f,d,o,c]),[f,e,d,o]}let t=zr(r?.config??r??{});return t=sl(t),[t,null,Vi(t),[]]}function Cl(r){return({tailwindDirectives:e,registerDependency:t})=>(i,n)=>{let[a,s,o,l]=b2(r),c=new Set(l);if(e.size>0){c.add(n.opts.from);for(let b of n.messages)b.type==="dependency"&&c.add(b.file)}let[f,,d]=zh(i,n,a,s,o,c),p=fs(f),h=y2(f,a);if(e.size>0){for(let y of h)for(let w of rl(y))t(w);let[b,v]=mg(f,h,p);for(let y of b)f.changedContent.push(y);for(let[y,w]of v.entries())d.set(y,w)}for(let b of l)t({type:"dependency",file:b});for(let[b,v]of d.entries())p.set(b,v);return f}}var Sg,Ag,Al,Cg=R(()=>{u();ft();Sg=pe(Ns());yf();ia();sc();_i();Uh();Yh();bg();vg();kg();Ag=new Sg.default({maxSize:100}),Al=new WeakMap});function _l(r){let e=new Set,t=new Set,i=new Set;if(r.walkAtRules(n=>{n.name==="apply"&&i.add(n),n.name==="import"&&(n.params==='"tailwindcss/base"'||n.params==="'tailwindcss/base'"?(n.name="tailwind",n.params="base"):n.params==='"tailwindcss/components"'||n.params==="'tailwindcss/components'"?(n.name="tailwind",n.params="components"):n.params==='"tailwindcss/utilities"'||n.params==="'tailwindcss/utilities'"?(n.name="tailwind",n.params="utilities"):(n.params==='"tailwindcss/screens"'||n.params==="'tailwindcss/screens'"||n.params==='"tailwindcss/variants"'||n.params==="'tailwindcss/variants'")&&(n.name="tailwind",n.params="variants")),n.name==="tailwind"&&(n.params==="screens"&&(n.params="variants"),e.add(n.params)),["layer","responsive","variants"].includes(n.name)&&(["responsive","variants"].includes(n.name)&&G.warn(`${n.name}-at-rule-deprecated`,[`The \`@${n.name}\` directive has been deprecated in Tailwind CSS v3.0.`,"Use `@layer utilities` or `@layer components` instead.","https://tailwindcss.com/docs/upgrade-guide#replace-variants-with-layer"]),t.add(n))}),!e.has("base")||!e.has("components")||!e.has("utilities")){for(let n of t)if(n.name==="layer"&&["base","components","utilities"].includes(n.params)){if(!e.has(n.params))throw n.error(`\`@layer ${n.params}\` is used but no matching \`@tailwind ${n.params}\` directive is present.`)}else if(n.name==="responsive"){if(!e.has("utilities"))throw n.error("`@responsive` is used but `@tailwind utilities` is missing.")}else if(n.name==="variants"&&!e.has("utilities"))throw n.error("`@variants` is used but `@tailwind utilities` is missing.")}return{tailwindDirectives:e,applyDirectives:i}}var _g=R(()=>{u();Be()});function Qt(r,e=void 0,t=void 0){return r.map(i=>{let n=i.clone();return t!==void 0&&(n.raws.tailwind={...n.raws.tailwind,...t}),e!==void 0&&Eg(n,a=>{if(a.raws.tailwind?.preserveSource===!0&&a.source)return!1;a.source=e}),n})}function Eg(r,e){e(r)!==!1&&r.each?.(t=>Eg(t,e))}var Og=R(()=>{u()});function El(r){return r=Array.isArray(r)?r:[r],r=r.map(e=>e instanceof RegExp?e.source:e),r.join("")}function Ne(r){return new RegExp(El(r),"g")}function qt(r){return`(?:${r.map(El).join("|")})`}function Ol(r){return`(?:${El(r)})?`}function Rg(r){return r&&w2.test(r)?r.replace(Tg,"\\$&"):r||""}var Tg,w2,Pg=R(()=>{u();Tg=/[\\^$.*+?()[\]{}|]/g,w2=RegExp(Tg.source)});function Ig(r){let e=Array.from(v2(r));return t=>{let i=[];for(let n of e)for(let a of t.match(n)??[])i.push(S2(a));for(let n of i.slice()){let a=ve(n,".");for(let s=0;s=a.length-1){i.push(o);continue}let l=Number(a[s+1]);isNaN(l)?i.push(o):s++}}return i}}function*v2(r){let e=r.tailwindConfig.separator,t=r.tailwindConfig.prefix!==""?Ol(Ne([/-?/,Rg(r.tailwindConfig.prefix)])):"",i=qt([/\[[^\s:'"`]+:[^\s\[\]]+\]/,/\[[^\s:'"`\]]+:[^\s]+?\[[^\s]+\][^\s]+?\]/,Ne([qt([/-?(?:\w+)/,/@(?:\w+)/]),Ol(qt([Ne([qt([/-(?:\w+-)*\['[^\s]+'\]/,/-(?:\w+-)*\["[^\s]+"\]/,/-(?:\w+-)*\[`[^\s]+`\]/,/-(?:\w+-)*\[(?:[^\s\[\]]+\[[^\s\[\]]+\])*[^\s:\[\]]+\]/]),/(?![{([]])/,/(?:\/[^\s'"`\\><$]*)?/]),Ne([qt([/-(?:\w+-)*\['[^\s]+'\]/,/-(?:\w+-)*\["[^\s]+"\]/,/-(?:\w+-)*\[`[^\s]+`\]/,/-(?:\w+-)*\[(?:[^\s\[\]]+\[[^\s\[\]]+\])*[^\s\[\]]+\]/]),/(?![{([]])/,/(?:\/[^\s'"`\\$]*)?/]),/[-\/][^\s'"`\\$={><]*/]))])]),n=[qt([Ne([/@\[[^\s"'`]+\](\/[^\s"'`]+)?/,e]),Ne([/([^\s"'`\[\\]+-)?\[[^\s"'`]+\]\/[\w_-]+/,e]),Ne([/([^\s"'`\[\\]+-)?\[[^\s"'`]+\]/,e]),Ne([/[^\s"'`\[\\]+/,e])]),qt([Ne([/([^\s"'`\[\\]+-)?\[[^\s`]+\]\/[\w_-]+/,e]),Ne([/([^\s"'`\[\\]+-)?\[[^\s`]+\]/,e]),Ne([/[^\s`\[\\]+/,e])])];for(let a of n)yield Ne(["((?=((",a,")+))\\2)?",/!?/,t,i]);yield/[^<>"'`\s.(){}[\]#=%$][^<>"'`\s(){}[\]#=%$]*[^<>"'`\s.(){}[\]#=%:$]/g}function S2(r){if(!r.includes("-["))return r;let e=0,t=[],i=r.matchAll(x2);i=Array.from(i).flatMap(n=>{let[,...a]=n;return a.map((s,o)=>Object.assign([],n,{index:n.index+o,0:s}))});for(let n of i){let a=n[0],s=t[t.length-1];if(a===s?t.pop():(a==="'"||a==='"'||a==="`")&&t.push(a),!s){if(a==="["){e++;continue}else if(a==="]"){e--;continue}if(e<0)return r.substring(0,n.index-1);if(e===0&&!k2.test(a))return r.substring(0,n.index)}}return r}var x2,k2,Dg=R(()=>{u();Pg();zt();x2=/([\[\]'"`])([^\[\]'"`])?/g,k2=/[^"'`\s<>\]]+/});function A2(r,e){let t=r.tailwindConfig.content.extract;return t[e]||t.DEFAULT||$g[e]||$g.DEFAULT(r)}function C2(r,e){let t=r.content.transform;return t[e]||t.DEFAULT||Lg[e]||Lg.DEFAULT}function _2(r,e,t,i){qi.has(e)||qi.set(e,new qg.default({maxSize:25e3}));for(let n of r.split(` -`))if(n=n.trim(),!i.has(n))if(i.add(n),qi.get(e).has(n))for(let a of qi.get(e).get(n))t.add(a);else{let a=e(n).filter(o=>o!=="!*"),s=new Set(a);for(let o of s)t.add(o);qi.get(e).set(n,s)}}function E2(r,e){let t=e.offsets.sort(r),i={base:new Set,defaults:new Set,components:new Set,utilities:new Set,variants:new Set};for(let[n,a]of t)i[n.layer].add(a);return i}function Tl(r){return async e=>{let t={base:null,components:null,utilities:null,variants:null};if(e.walkAtRules(y=>{y.name==="tailwind"&&Object.keys(t).includes(y.params)&&(t[y.params]=y)}),Object.values(t).every(y=>y===null))return e;let i=new Set([...r.candidates??[],gt]),n=new Set;bt.DEBUG&&console.time("Reading changed files");let a=[];for(let y of r.changedContent){let w=C2(r.tailwindConfig,y.extension),k=A2(r,y.extension);a.push([y,{transformer:w,extractor:k}])}let s=500;for(let y=0;y{S=k?await be.promises.readFile(k,"utf8"):S,_2(E(S),O,i,n)}))}bt.DEBUG&&console.timeEnd("Reading changed files");let o=r.classCache.size;bt.DEBUG&&console.time("Generate rules"),bt.DEBUG&&console.time("Sorting candidates");let l=new Set([...i].sort((y,w)=>y===w?0:y{let w=y.raws.tailwind?.parentLayer;return w==="components"?t.components!==null:w==="utilities"?t.utilities!==null:!0});t.variants?(t.variants.before(Qt(b,t.variants.source,{layer:"variants"})),t.variants.remove()):b.length>0&&e.append(Qt(b,e.source,{layer:"variants"})),e.source.end=e.source.end??e.source.start;let v=b.some(y=>y.raws.tailwind?.parentLayer==="utilities");t.utilities&&p.size===0&&!v&&G.warn("content-problems",["No utility classes were detected in your source files. If this is unexpected, double-check the `content` option in your Tailwind CSS configuration.","https://tailwindcss.com/docs/content-configuration"]),bt.DEBUG&&(console.log("Potential classes: ",i.size),console.log("Active contexts: ",Zn.size)),r.changedContent=[],e.walkAtRules("layer",y=>{Object.keys(t).includes(y.params)&&y.remove()})}}var qg,bt,$g,Lg,qi,Mg=R(()=>{u();ft();qg=pe(Ns());It();as();Be();Og();Dg();bt=Je,$g={DEFAULT:Ig},Lg={DEFAULT:r=>r,svelte:r=>r.replace(/(?:^|\s)class:/g," ")};qi=new WeakMap});function ws(r){let e=new Map;ee.root({nodes:[r.clone()]}).walkRules(a=>{(0,bs.default)(s=>{s.walkClasses(o=>{let l=o.parent.toString(),c=e.get(l);c||e.set(l,c=new Set),c.add(o.value)})}).processSync(a.selector)});let i=Array.from(e.values(),a=>Array.from(a)),n=i.flat();return Object.assign(n,{groups:i})}function Rl(r){return O2.astSync(r)}function Ng(r,e){let t=new Set;for(let i of r)t.add(i.split(e).pop());return Array.from(t)}function Bg(r,e){let t=r.tailwindConfig.prefix;return typeof t=="function"?t(e):t+e}function*Fg(r){for(yield r;r.parent;)yield r.parent,r=r.parent}function T2(r,e={}){let t=r.nodes;r.nodes=[];let i=r.clone(e);return r.nodes=t,i}function R2(r){for(let e of Fg(r))if(r!==e){if(e.type==="root")break;r=T2(e,{nodes:[r]})}return r}function P2(r,e){let t=new Map;return r.walkRules(i=>{for(let s of Fg(i))if(s.raws.tailwind?.layer!==void 0)return;let n=R2(i),a=e.offsets.create("user");for(let s of ws(i)){let o=t.get(s)||[];t.set(s,o),o.push([{layer:"user",sort:a,important:!1},n])}}),t}function I2(r,e){for(let t of r){if(e.notClassCache.has(t)||e.applyClassCache.has(t))continue;if(e.classCache.has(t)){e.applyClassCache.set(t,e.classCache.get(t).map(([n,a])=>[n,a.clone()]));continue}let i=Array.from(Go(t,e));if(i.length===0){e.notClassCache.add(t);continue}e.applyClassCache.set(t,i)}return e.applyClassCache}function D2(r){let e=null;return{get:t=>(e=e||r(),e.get(t)),has:t=>(e=e||r(),e.has(t))}}function q2(r){return{get:e=>r.flatMap(t=>t.get(e)||[]),has:e=>r.some(t=>t.has(e))}}function jg(r){let e=r.split(/[\s\t\n]+/g);return e[e.length-1]==="!important"?[e.slice(0,-1),!0]:[e,!1]}function zg(r,e,t){let i=new Set,n=[];if(r.walkAtRules("apply",l=>{let[c]=jg(l.params);for(let f of c)i.add(f);n.push(l)}),n.length===0)return;let a=q2([t,I2(i,e)]);function s(l,c,f){let d=Rl(l),p=Rl(c),b=Rl(`.${Te(f)}`).nodes[0].nodes[0];return d.each(v=>{let y=new Set;p.each(w=>{let k=!1;w=w.clone(),w.walkClasses(S=>{S.value===b.value&&(k||(S.replaceWith(...v.nodes.map(E=>E.clone())),y.add(w),k=!0))})});for(let w of y){let k=[[]];for(let S of w.nodes)S.type==="combinator"?(k.push(S),k.push([])):k[k.length-1].push(S);w.nodes=[];for(let S of k)Array.isArray(S)&&S.sort((E,O)=>E.type==="tag"&&O.type==="class"?-1:E.type==="class"&&O.type==="tag"?1:E.type==="class"&&O.type==="pseudo"&&O.value.startsWith("::")?-1:E.type==="pseudo"&&E.value.startsWith("::")&&O.type==="class"?1:0),w.nodes=w.nodes.concat(S)}v.replaceWith(...y)}),d.toString()}let o=new Map;for(let l of n){let[c]=o.get(l.parent)||[[],l.source];o.set(l.parent,[c,l.source]);let[f,d]=jg(l.params);if(l.parent.type==="atrule"){if(l.parent.name==="screen"){let p=l.parent.params;throw l.error(`@apply is not supported within nested at-rules like @screen. We suggest you write this as @apply ${f.map(h=>`${p}:${h}`).join(" ")} instead.`)}throw l.error(`@apply is not supported within nested at-rules like @${l.parent.name}. You can fix this by un-nesting @${l.parent.name}.`)}for(let p of f){if([Bg(e,"group"),Bg(e,"peer")].includes(p))throw l.error(`@apply should not be used with the '${p}' utility`);if(!a.has(p))throw l.error(`The \`${p}\` class does not exist. If \`${p}\` is a custom class, make sure it is defined within a \`@layer\` directive.`);let h=a.get(p);for(let[,b]of h)b.type!=="atrule"&&b.walkRules(()=>{throw l.error([`The \`${p}\` class cannot be used with \`@apply\` because \`@apply\` does not currently support nested CSS.`,"Rewrite the selector without nesting or configure the `tailwindcss/nesting` plugin:","https://tailwindcss.com/docs/using-with-preprocessors#nesting"].join(` -`))});c.push([p,d,h])}}for(let[l,[c,f]]of o){let d=[];for(let[h,b,v]of c){let y=[h,...Ng([h],e.tailwindConfig.separator)];for(let[w,k]of v){let S=ws(l),E=ws(k);if(E=E.groups.filter(T=>T.some(F=>y.includes(F))).flat(),E=E.concat(Ng(E,e.tailwindConfig.separator)),S.some(T=>E.includes(T)))throw k.error(`You cannot \`@apply\` the \`${h}\` utility here because it creates a circular dependency.`);let B=ee.root({nodes:[k.clone()]});B.walk(T=>{T.source=f}),(k.type!=="atrule"||k.type==="atrule"&&k.name!=="keyframes")&&B.walkRules(T=>{if(!ws(T).some(U=>U===h)){T.remove();return}let F=typeof e.tailwindConfig.important=="string"?e.tailwindConfig.important:null,_=l.raws.tailwind!==void 0&&F&&l.selector.indexOf(F)===0?l.selector.slice(F.length):l.selector;_===""&&(_=l.selector),T.selector=s(_,T.selector,h),F&&_!==l.selector&&(T.selector=rs(T.selector,F)),T.walkDecls(U=>{U.important=w.important||b});let Q=(0,bs.default)().astSync(T.selector);Q.each(U=>pr(U)),T.selector=Q.toString()}),!!B.nodes[0]&&d.push([w.sort,B.nodes[0]])}}let p=e.offsets.sort(d).map(h=>h[1]);l.after(p)}for(let l of n)l.parent.nodes.length>1?l.remove():l.parent.remove();zg(r,e,t)}function Pl(r){return e=>{let t=D2(()=>P2(e,r));zg(e,r,t)}}var bs,O2,Ug=R(()=>{u();Ot();bs=pe(it());as();fr();Vo();es();O2=(0,bs.default)()});var Vg=x((rq,vs)=>{u();(function(){"use strict";function r(i,n,a){if(!i)return null;r.caseSensitive||(i=i.toLowerCase());var s=r.threshold===null?null:r.threshold*i.length,o=r.thresholdAbsolute,l;s!==null&&o!==null?l=Math.min(s,o):s!==null?l=s:o!==null?l=o:l=null;var c,f,d,p,h,b=n.length;for(h=0;ha)return a+1;var l=[],c,f,d,p,h;for(c=0;c<=o;c++)l[c]=[c];for(f=0;f<=s;f++)l[0][f]=f;for(c=1;c<=o;c++){for(d=e,p=1,c>a&&(p=c-a),h=o+1,h>a+c&&(h=a+c),f=1;f<=s;f++)fh?l[c][f]=a+1:n.charAt(c-1)===i.charAt(f-1)?l[c][f]=l[c-1][f-1]:l[c][f]=Math.min(l[c-1][f-1]+1,Math.min(l[c][f-1]+1,l[c-1][f]+1)),l[c][f]a)return a+1}return l[o][s]}})()});var Wg=x((iq,Hg)=>{u();var Il="(".charCodeAt(0),Dl=")".charCodeAt(0),xs="'".charCodeAt(0),ql='"'.charCodeAt(0),$l="\\".charCodeAt(0),yr="/".charCodeAt(0),Ll=",".charCodeAt(0),Ml=":".charCodeAt(0),ks="*".charCodeAt(0),$2="u".charCodeAt(0),L2="U".charCodeAt(0),M2="+".charCodeAt(0),N2=/^[a-f0-9?-]+$/i;Hg.exports=function(r){for(var e=[],t=r,i,n,a,s,o,l,c,f,d=0,p=t.charCodeAt(d),h=t.length,b=[{nodes:e}],v=0,y,w="",k="",S="";d{u();Gg.exports=function r(e,t,i){var n,a,s,o;for(n=0,a=e.length;n{u();function Yg(r,e){var t=r.type,i=r.value,n,a;return e&&(a=e(r))!==void 0?a:t==="word"||t==="space"?i:t==="string"?(n=r.quote||"",n+i+(r.unclosed?"":n)):t==="comment"?"/*"+i+(r.unclosed?"":"*/"):t==="div"?(r.before||"")+i+(r.after||""):Array.isArray(r.nodes)?(n=Kg(r.nodes,e),t!=="function"?n:i+"("+(r.before||"")+n+(r.after||"")+(r.unclosed?"":")")):i}function Kg(r,e){var t,i;if(Array.isArray(r)){for(t="",i=r.length-1;~i;i-=1)t=Yg(r[i],e)+t;return t}return Yg(r,e)}Xg.exports=Kg});var ey=x((aq,Zg)=>{u();var Ss="-".charCodeAt(0),As="+".charCodeAt(0),Nl=".".charCodeAt(0),B2="e".charCodeAt(0),F2="E".charCodeAt(0);function j2(r){var e=r.charCodeAt(0),t;if(e===As||e===Ss){if(t=r.charCodeAt(1),t>=48&&t<=57)return!0;var i=r.charCodeAt(2);return t===Nl&&i>=48&&i<=57}return e===Nl?(t=r.charCodeAt(1),t>=48&&t<=57):e>=48&&e<=57}Zg.exports=function(r){var e=0,t=r.length,i,n,a;if(t===0||!j2(r))return!1;for(i=r.charCodeAt(e),(i===As||i===Ss)&&e++;e57));)e+=1;if(i=r.charCodeAt(e),n=r.charCodeAt(e+1),i===Nl&&n>=48&&n<=57)for(e+=2;e57));)e+=1;if(i=r.charCodeAt(e),n=r.charCodeAt(e+1),a=r.charCodeAt(e+2),(i===B2||i===F2)&&(n>=48&&n<=57||(n===As||n===Ss)&&a>=48&&a<=57))for(e+=n===As||n===Ss?3:2;e57));)e+=1;return{number:r.slice(0,e),unit:r.slice(e)}}});var ny=x((oq,iy)=>{u();var z2=Wg(),ty=Qg(),ry=Jg();function $t(r){return this instanceof $t?(this.nodes=z2(r),this):new $t(r)}$t.prototype.toString=function(){return Array.isArray(this.nodes)?ry(this.nodes):""};$t.prototype.walk=function(r,e){return ty(this.nodes,r,e),this};$t.unit=ey();$t.walk=ty;$t.stringify=ry;iy.exports=$t});function Fl(r){return typeof r=="object"&&r!==null}function U2(r,e){let t=kt(e);do if(t.pop(),(0,$i.default)(r,t)!==void 0)break;while(t.length);return t.length?t:void 0}function br(r){return typeof r=="string"?r:r.reduce((e,t,i)=>t.includes(".")?`${e}[${t}]`:i===0?t:`${e}.${t}`,"")}function ay(r){return r.map(e=>`'${e}'`).join(", ")}function oy(r){return ay(Object.keys(r))}function jl(r,e,t,i={}){let n=Array.isArray(e)?br(e):e.replace(/^['"]+|['"]+$/g,""),a=Array.isArray(e)?e:kt(n),s=(0,$i.default)(r.theme,a,t);if(s===void 0){let l=`'${n}' does not exist in your theme config.`,c=a.slice(0,-1),f=(0,$i.default)(r.theme,c);if(Fl(f)){let d=Object.keys(f).filter(h=>jl(r,[...c,h]).isValid),p=(0,sy.default)(a[a.length-1],d);p?l+=` Did you mean '${br([...c,p])}'?`:d.length>0&&(l+=` '${br(c)}' has the following valid keys: ${ay(d)}`)}else{let d=U2(r.theme,n);if(d){let p=(0,$i.default)(r.theme,d);Fl(p)?l+=` '${br(d)}' has the following keys: ${oy(p)}`:l+=` '${br(d)}' is not an object.`}else l+=` Your theme has the following top-level keys: ${oy(r.theme)}`}return{isValid:!1,error:l}}if(!(typeof s=="string"||typeof s=="number"||typeof s=="function"||s instanceof String||s instanceof Number||Array.isArray(s))){let l=`'${n}' was found but does not resolve to a string.`;if(Fl(s)){let c=Object.keys(s).filter(f=>jl(r,[...a,f]).isValid);c.length&&(l+=` Did you mean something like '${br([...a,c[0]])}'?`)}return{isValid:!1,error:l}}let[o]=a;return{isValid:!0,value:mt(o)(s,i)}}function V2(r,e,t){e=e.map(n=>ly(r,n,t));let i=[""];for(let n of e)n.type==="div"&&n.value===","?i.push(""):i[i.length-1]+=Bl.default.stringify(n);return i}function ly(r,e,t){if(e.type==="function"&&t[e.value]!==void 0){let i=V2(r,e.nodes,t);e.type="word",e.value=t[e.value](r,...i)}return e}function H2(r,e,t){return Object.keys(t).some(n=>e.includes(`${n}(`))?(0,Bl.default)(e).walk(n=>{ly(r,n,t)}).toString():e}function*G2(r){r=r.replace(/^['"]+|['"]+$/g,"");let e=r.match(/^([^\s]+)(?![^\[]*\])(?:\s*\/\s*([^\/\s]+))$/),t;yield[r,void 0],e&&(r=e[1],t=e[2],yield[r,t])}function Q2(r,e,t){let i=Array.from(G2(e)).map(([n,a])=>Object.assign(jl(r,n,t,{opacityValue:a}),{resolvedPath:n,alpha:a}));return i.find(n=>n.isValid)??i[0]}function uy(r){let e=r.tailwindConfig,t={theme:(i,n,...a)=>{let{isValid:s,value:o,error:l,alpha:c}=Q2(e,n,a.length?a:void 0);if(!s){let p=i.parent,h=p?.raws.tailwind?.candidate;if(p&&h!==void 0){r.markInvalidUtilityNode(p),p.remove(),G.warn("invalid-theme-key-in-class",[`The utility \`${h}\` contains an invalid theme value and was not generated.`]);return}throw i.error(l)}let f=Xt(o),d=f!==void 0&&typeof f=="function";return(c!==void 0||d)&&(c===void 0&&(c=1),o=Ze(f,c,f)),o},screen:(i,n)=>{n=n.replace(/^['"]+/g,"").replace(/['"]+$/g,"");let s=Rt(e.theme.screens).find(({name:o})=>o===n);if(!s)throw i.error(`The '${n}' screen does not exist in your theme.`);return Tt(s)}};return i=>{i.walk(n=>{let a=W2[n.type];a!==void 0&&(n[a]=H2(n,n[a],t))})}}var $i,sy,Bl,W2,fy=R(()=>{u();$i=pe(Oa()),sy=pe(Vg());Si();Bl=pe(ny());Xn();Qn();Gi();Lr();Fr();Be();W2={atrule:"params",decl:"value"}});function cy({tailwindConfig:{theme:r}}){return function(e){e.walkAtRules("screen",t=>{let i=t.params,a=Rt(r.screens).find(({name:s})=>s===i);if(!a)throw t.error(`No \`${i}\` screen found.`);t.name="media",t.params=Tt(a)})}}var py=R(()=>{u();Xn();Qn()});function Y2(r){let e=r.filter(o=>o.type!=="pseudo"||o.nodes.length>0?!0:o.value.startsWith("::")||[":before",":after",":first-line",":first-letter"].includes(o.value)).reverse(),t=new Set(["tag","class","id","attribute"]),i=e.findIndex(o=>t.has(o.type));if(i===-1)return e.reverse().join("").trim();let n=e[i],a=dy[n.type]?dy[n.type](n):n;e=e.slice(0,i);let s=e.findIndex(o=>o.type==="combinator"&&o.value===">");return s!==-1&&(e.splice(0,s),e.unshift(Cs.default.universal())),[a,...e.reverse()].join("").trim()}function X2(r){return zl.has(r)||zl.set(r,K2.transformSync(r)),zl.get(r)}function Ul({tailwindConfig:r}){return e=>{let t=new Map,i=new Set;if(e.walkAtRules("defaults",n=>{if(n.nodes&&n.nodes.length>0){i.add(n);return}let a=n.params;t.has(a)||t.set(a,new Set),t.get(a).add(n.parent),n.remove()}),we(r,"optimizeUniversalDefaults"))for(let n of i){let a=new Map,s=t.get(n.params)??[];for(let o of s)for(let l of X2(o.selector)){let c=l.includes(":-")||l.includes("::-")||l.includes(":has")?l:"__DEFAULT__",f=a.get(c)??new Set;a.set(c,f),f.add(l)}if(a.size===0){n.remove();continue}for(let[,o]of a){let l=ee.rule({source:n.source});l.selectors=[...o],l.append(n.nodes.map(c=>c.clone())),n.before(l)}n.remove()}else if(i.size){let n=ee.rule({selectors:["*","::before","::after"]});for(let s of i)n.append(s.nodes),n.parent||s.before(n),n.source||(n.source=s.source),s.remove();let a=n.clone({selectors:["::backdrop"]});n.after(a)}}}var Cs,dy,K2,zl,hy=R(()=>{u();Ot();Cs=pe(it());ct();dy={id(r){return Cs.default.attribute({attribute:"id",operator:"=",value:r.value,quoteMark:'"'})}};K2=(0,Cs.default)(r=>r.map(e=>{let t=e.split(i=>i.type==="combinator"&&i.value===" ").pop();return Y2(t)})),zl=new Map});function Vl(){function r(e){let t=null;e.each(i=>{if(!J2.has(i.type)){t=null;return}if(t===null){t=i;return}let n=my[i.type];i.type==="atrule"&&i.name==="font-face"?t=i:n.every(a=>(i[a]??"").replace(/\s+/g," ")===(t[a]??"").replace(/\s+/g," "))?(i.nodes&&t.append(i.nodes),i.remove()):t=i}),e.each(i=>{i.type==="atrule"&&r(i)})}return e=>{r(e)}}var my,J2,gy=R(()=>{u();my={atrule:["name","params"],rule:["selector"]},J2=new Set(Object.keys(my))});function Hl(){return r=>{r.walkRules(e=>{let t=new Map,i=new Set([]),n=new Map;e.walkDecls(a=>{if(a.parent===e){if(t.has(a.prop)){if(t.get(a.prop).value===a.value){i.add(t.get(a.prop)),t.set(a.prop,a);return}n.has(a.prop)||n.set(a.prop,new Set),n.get(a.prop).add(t.get(a.prop)),n.get(a.prop).add(a)}t.set(a.prop,a)}});for(let a of i)a.remove();for(let a of n.values()){let s=new Map;for(let o of a){let l=eO(o.value);l!==null&&(s.has(l)||s.set(l,new Set),s.get(l).add(o))}for(let o of s.values()){let l=Array.from(o).slice(0,-1);for(let c of l)c.remove()}}})}}function eO(r){let e=/^-?\d*.?\d+([\w%]+)?$/g.exec(r);return e?e[1]??Z2:null}var Z2,yy=R(()=>{u();Z2=Symbol("unitless-number")});function tO(r){if(!r.walkAtRules)return;let e=new Set;if(r.walkAtRules("apply",t=>{e.add(t.parent)}),e.size!==0)for(let t of e){let i=[],n=[];for(let a of t.nodes)a.type==="atrule"&&a.name==="apply"?(n.length>0&&(i.push(n),n=[]),i.push([a])):n.push(a);if(n.length>0&&i.push(n),i.length!==1){for(let a of[...i].reverse()){let s=t.clone({nodes:[]});s.append(a),t.after(s)}t.remove()}}}function _s(){return r=>{tO(r)}}var by=R(()=>{u()});function Es(r){return async function(e,t){let{tailwindDirectives:i,applyDirectives:n}=_l(e);_s()(e,t);let a=r({tailwindDirectives:i,applyDirectives:n,registerDependency(s){t.messages.push({plugin:"tailwindcss",parent:t.opts.from,...s})},createContext(s,o){return tl(s,o,e)}})(e,t);if(a.tailwindConfig.separator==="-")throw new Error("The '-' character cannot be used as a custom separator in JIT mode due to parsing ambiguity. Please use another character like '_' instead.");Of(a.tailwindConfig),await Tl(a)(e,t),_s()(e,t),Pl(a)(e,t),uy(a)(e,t),cy(a)(e,t),Ul(a)(e,t),Vl(a)(e,t),Hl(a)(e,t)}}var wy=R(()=>{u();_g();Mg();Ug();fy();py();hy();gy();yy();by();_i();ct()});function vy(r,e){let t=null,i=null;return r.walkAtRules("config",n=>{if(i=n.source?.input.file??e.opts.from??null,i===null)throw n.error("The `@config` directive cannot be used without setting `from` in your PostCSS config.");if(t)throw n.error("Only one `@config` directive is allowed per file.");let a=n.params.match(/(['"])(.*?)\1/);if(!a)throw n.error("A path is required when using the `@config` directive.");let s=a[2];if(me.isAbsolute(s))throw n.error("The `@config` directive cannot be used with an absolute path.");if(t=me.resolve(me.dirname(i),s),!be.existsSync(t))throw n.error(`The config file at "${s}" does not exist. Make sure the path is correct and the file exists.`);n.remove()}),t||null}var xy=R(()=>{u();ft();et()});var ky=x((Vq,Wl)=>{u();Cg();wy();It();xy();Wl.exports=function(e){return{postcssPlugin:"tailwindcss",plugins:[Je.DEBUG&&function(t){return console.log(` -`),console.time("JIT TOTAL"),t},async function(t,i){e=vy(t,i)??e;let n=Cl(e);if(t.type==="document"){let a=t.nodes.filter(s=>s.type==="root");for(let s of a)s.type==="root"&&await Es(n)(s,i);return}await Es(n)(t,i)},Je.DEBUG&&function(t){return console.timeEnd("JIT TOTAL"),console.log(` -`),t}].filter(Boolean)}};Wl.exports.postcss=!0});var Ay=x((Hq,Sy)=>{u();Sy.exports=ky()});var Gl=x((Wq,Cy)=>{u();Cy.exports=()=>["and_chr 114","and_uc 15.5","chrome 114","chrome 113","chrome 109","edge 114","firefox 114","ios_saf 16.5","ios_saf 16.4","ios_saf 16.3","ios_saf 16.1","opera 99","safari 16.5","samsung 21"]});var Os={};Ge(Os,{agents:()=>rO,feature:()=>iO});function iO(){return{status:"cr",title:"CSS Feature Queries",stats:{ie:{"6":"n","7":"n","8":"n","9":"n","10":"n","11":"n","5.5":"n"},edge:{"12":"y","13":"y","14":"y","15":"y","16":"y","17":"y","18":"y","79":"y","80":"y","81":"y","83":"y","84":"y","85":"y","86":"y","87":"y","88":"y","89":"y","90":"y","91":"y","92":"y","93":"y","94":"y","95":"y","96":"y","97":"y","98":"y","99":"y","100":"y","101":"y","102":"y","103":"y","104":"y","105":"y","106":"y","107":"y","108":"y","109":"y","110":"y","111":"y","112":"y","113":"y","114":"y"},firefox:{"2":"n","3":"n","4":"n","5":"n","6":"n","7":"n","8":"n","9":"n","10":"n","11":"n","12":"n","13":"n","14":"n","15":"n","16":"n","17":"n","18":"n","19":"n","20":"n","21":"n","22":"y","23":"y","24":"y","25":"y","26":"y","27":"y","28":"y","29":"y","30":"y","31":"y","32":"y","33":"y","34":"y","35":"y","36":"y","37":"y","38":"y","39":"y","40":"y","41":"y","42":"y","43":"y","44":"y","45":"y","46":"y","47":"y","48":"y","49":"y","50":"y","51":"y","52":"y","53":"y","54":"y","55":"y","56":"y","57":"y","58":"y","59":"y","60":"y","61":"y","62":"y","63":"y","64":"y","65":"y","66":"y","67":"y","68":"y","69":"y","70":"y","71":"y","72":"y","73":"y","74":"y","75":"y","76":"y","77":"y","78":"y","79":"y","80":"y","81":"y","82":"y","83":"y","84":"y","85":"y","86":"y","87":"y","88":"y","89":"y","90":"y","91":"y","92":"y","93":"y","94":"y","95":"y","96":"y","97":"y","98":"y","99":"y","100":"y","101":"y","102":"y","103":"y","104":"y","105":"y","106":"y","107":"y","108":"y","109":"y","110":"y","111":"y","112":"y","113":"y","114":"y","115":"y","116":"y","117":"y","3.5":"n","3.6":"n"},chrome:{"4":"n","5":"n","6":"n","7":"n","8":"n","9":"n","10":"n","11":"n","12":"n","13":"n","14":"n","15":"n","16":"n","17":"n","18":"n","19":"n","20":"n","21":"n","22":"n","23":"n","24":"n","25":"n","26":"n","27":"n","28":"y","29":"y","30":"y","31":"y","32":"y","33":"y","34":"y","35":"y","36":"y","37":"y","38":"y","39":"y","40":"y","41":"y","42":"y","43":"y","44":"y","45":"y","46":"y","47":"y","48":"y","49":"y","50":"y","51":"y","52":"y","53":"y","54":"y","55":"y","56":"y","57":"y","58":"y","59":"y","60":"y","61":"y","62":"y","63":"y","64":"y","65":"y","66":"y","67":"y","68":"y","69":"y","70":"y","71":"y","72":"y","73":"y","74":"y","75":"y","76":"y","77":"y","78":"y","79":"y","80":"y","81":"y","83":"y","84":"y","85":"y","86":"y","87":"y","88":"y","89":"y","90":"y","91":"y","92":"y","93":"y","94":"y","95":"y","96":"y","97":"y","98":"y","99":"y","100":"y","101":"y","102":"y","103":"y","104":"y","105":"y","106":"y","107":"y","108":"y","109":"y","110":"y","111":"y","112":"y","113":"y","114":"y","115":"y","116":"y","117":"y"},safari:{"4":"n","5":"n","6":"n","7":"n","8":"n","9":"y","10":"y","11":"y","12":"y","13":"y","14":"y","15":"y","17":"y","9.1":"y","10.1":"y","11.1":"y","12.1":"y","13.1":"y","14.1":"y","15.1":"y","15.2-15.3":"y","15.4":"y","15.5":"y","15.6":"y","16.0":"y","16.1":"y","16.2":"y","16.3":"y","16.4":"y","16.5":"y","16.6":"y",TP:"y","3.1":"n","3.2":"n","5.1":"n","6.1":"n","7.1":"n"},opera:{"9":"n","11":"n","12":"n","15":"y","16":"y","17":"y","18":"y","19":"y","20":"y","21":"y","22":"y","23":"y","24":"y","25":"y","26":"y","27":"y","28":"y","29":"y","30":"y","31":"y","32":"y","33":"y","34":"y","35":"y","36":"y","37":"y","38":"y","39":"y","40":"y","41":"y","42":"y","43":"y","44":"y","45":"y","46":"y","47":"y","48":"y","49":"y","50":"y","51":"y","52":"y","53":"y","54":"y","55":"y","56":"y","57":"y","58":"y","60":"y","62":"y","63":"y","64":"y","65":"y","66":"y","67":"y","68":"y","69":"y","70":"y","71":"y","72":"y","73":"y","74":"y","75":"y","76":"y","77":"y","78":"y","79":"y","80":"y","81":"y","82":"y","83":"y","84":"y","85":"y","86":"y","87":"y","88":"y","89":"y","90":"y","91":"y","92":"y","93":"y","94":"y","95":"y","96":"y","97":"y","98":"y","99":"y","100":"y","12.1":"y","9.5-9.6":"n","10.0-10.1":"n","10.5":"n","10.6":"n","11.1":"n","11.5":"n","11.6":"n"},ios_saf:{"8":"n","17":"y","9.0-9.2":"y","9.3":"y","10.0-10.2":"y","10.3":"y","11.0-11.2":"y","11.3-11.4":"y","12.0-12.1":"y","12.2-12.5":"y","13.0-13.1":"y","13.2":"y","13.3":"y","13.4-13.7":"y","14.0-14.4":"y","14.5-14.8":"y","15.0-15.1":"y","15.2-15.3":"y","15.4":"y","15.5":"y","15.6":"y","16.0":"y","16.1":"y","16.2":"y","16.3":"y","16.4":"y","16.5":"y","16.6":"y","3.2":"n","4.0-4.1":"n","4.2-4.3":"n","5.0-5.1":"n","6.0-6.1":"n","7.0-7.1":"n","8.1-8.4":"n"},op_mini:{all:"y"},android:{"3":"n","4":"n","114":"y","4.4":"y","4.4.3-4.4.4":"y","2.1":"n","2.2":"n","2.3":"n","4.1":"n","4.2-4.3":"n"},bb:{"7":"n","10":"n"},op_mob:{"10":"n","11":"n","12":"n","73":"y","11.1":"n","11.5":"n","12.1":"n"},and_chr:{"114":"y"},and_ff:{"115":"y"},ie_mob:{"10":"n","11":"n"},and_uc:{"15.5":"y"},samsung:{"4":"y","20":"y","21":"y","5.0-5.4":"y","6.2-6.4":"y","7.2-7.4":"y","8.2":"y","9.2":"y","10.1":"y","11.1-11.2":"y","12.0":"y","13.0":"y","14.0":"y","15.0":"y","16.0":"y","17.0":"y","18.0":"y","19.0":"y"},and_qq:{"13.1":"y"},baidu:{"13.18":"y"},kaios:{"2.5":"y","3.0-3.1":"y"}}}}var rO,Ts=R(()=>{u();rO={ie:{prefix:"ms"},edge:{prefix:"webkit",prefix_exceptions:{"12":"ms","13":"ms","14":"ms","15":"ms","16":"ms","17":"ms","18":"ms"}},firefox:{prefix:"moz"},chrome:{prefix:"webkit"},safari:{prefix:"webkit"},opera:{prefix:"webkit",prefix_exceptions:{"9":"o","11":"o","12":"o","9.5-9.6":"o","10.0-10.1":"o","10.5":"o","10.6":"o","11.1":"o","11.5":"o","11.6":"o","12.1":"o"}},ios_saf:{prefix:"webkit"},op_mini:{prefix:"o"},android:{prefix:"webkit"},bb:{prefix:"webkit"},op_mob:{prefix:"o",prefix_exceptions:{"73":"webkit"}},and_chr:{prefix:"webkit"},and_ff:{prefix:"moz"},ie_mob:{prefix:"ms"},and_uc:{prefix:"webkit",prefix_exceptions:{"15.5":"webkit"}},samsung:{prefix:"webkit"},and_qq:{prefix:"webkit"},baidu:{prefix:"webkit"},kaios:{prefix:"moz"}}});var _y=x(()=>{u()});var _e=x((Yq,Lt)=>{u();var{list:Ql}=$e();Lt.exports.error=function(r){let e=new Error(r);throw e.autoprefixer=!0,e};Lt.exports.uniq=function(r){return[...new Set(r)]};Lt.exports.removeNote=function(r){return r.includes(" ")?r.split(" ")[0]:r};Lt.exports.escapeRegexp=function(r){return r.replace(/[$()*+-.?[\\\]^{|}]/g,"\\$&")};Lt.exports.regexp=function(r,e=!0){return e&&(r=this.escapeRegexp(r)),new RegExp(`(^|[\\s,(])(${r}($|[\\s(,]))`,"gi")};Lt.exports.editList=function(r,e){let t=Ql.comma(r),i=e(t,[]);if(t===i)return r;let n=r.match(/,\s*/);return n=n?n[0]:", ",i.join(n)};Lt.exports.splitSelector=function(r){return Ql.comma(r).map(e=>Ql.space(e).map(t=>t.split(/(?=\.|#)/g)))}});var Mt=x((Kq,Ty)=>{u();var nO=Gl(),Ey=(Ts(),Os).agents,sO=_e(),Oy=class{static prefixes(){if(this.prefixesCache)return this.prefixesCache;this.prefixesCache=[];for(let e in Ey)this.prefixesCache.push(`-${Ey[e].prefix}-`);return this.prefixesCache=sO.uniq(this.prefixesCache).sort((e,t)=>t.length-e.length),this.prefixesCache}static withPrefix(e){return this.prefixesRegexp||(this.prefixesRegexp=new RegExp(this.prefixes().join("|"))),this.prefixesRegexp.test(e)}constructor(e,t,i,n){this.data=e,this.options=i||{},this.browserslistOpts=n||{},this.selected=this.parse(t)}parse(e){let t={};for(let i in this.browserslistOpts)t[i]=this.browserslistOpts[i];return t.path=this.options.from,nO(e,t)}prefix(e){let[t,i]=e.split(" "),n=this.data[t],a=n.prefix_exceptions&&n.prefix_exceptions[i];return a||(a=n.prefix),`-${a}-`}isSelected(e){return this.selected.includes(e)}};Ty.exports=Oy});var Li=x((Xq,Ry)=>{u();Ry.exports={prefix(r){let e=r.match(/^(-\w+-)/);return e?e[0]:""},unprefixed(r){return r.replace(/^-\w+-/,"")}}});var wr=x((Jq,Iy)=>{u();var aO=Mt(),Py=Li(),oO=_e();function Yl(r,e){let t=new r.constructor;for(let i of Object.keys(r||{})){let n=r[i];i==="parent"&&typeof n=="object"?e&&(t[i]=e):i==="source"||i===null?t[i]=n:Array.isArray(n)?t[i]=n.map(a=>Yl(a,t)):i!=="_autoprefixerPrefix"&&i!=="_autoprefixerValues"&&i!=="proxyCache"&&(typeof n=="object"&&n!==null&&(n=Yl(n,t)),t[i]=n)}return t}var Rs=class{static hack(e){return this.hacks||(this.hacks={}),e.names.map(t=>(this.hacks[t]=e,this.hacks[t]))}static load(e,t,i){let n=this.hacks&&this.hacks[e];return n?new n(e,t,i):new this(e,t,i)}static clone(e,t){let i=Yl(e);for(let n in t)i[n]=t[n];return i}constructor(e,t,i){this.prefixes=t,this.name=e,this.all=i}parentPrefix(e){let t;return typeof e._autoprefixerPrefix!="undefined"?t=e._autoprefixerPrefix:e.type==="decl"&&e.prop[0]==="-"?t=Py.prefix(e.prop):e.type==="root"?t=!1:e.type==="rule"&&e.selector.includes(":-")&&/:(-\w+-)/.test(e.selector)?t=e.selector.match(/:(-\w+-)/)[1]:e.type==="atrule"&&e.name[0]==="-"?t=Py.prefix(e.name):t=this.parentPrefix(e.parent),aO.prefixes().includes(t)||(t=!1),e._autoprefixerPrefix=t,e._autoprefixerPrefix}process(e,t){if(!this.check(e))return;let i=this.parentPrefix(e),n=this.prefixes.filter(s=>!i||i===oO.removeNote(s)),a=[];for(let s of n)this.add(e,s,a.concat([s]),t)&&a.push(s);return a}clone(e,t){return Rs.clone(e,t)}};Iy.exports=Rs});var j=x((Zq,$y)=>{u();var lO=wr(),uO=Mt(),Dy=_e(),qy=class extends lO{check(){return!0}prefixed(e,t){return t+e}normalize(e){return e}otherPrefixes(e,t){for(let i of uO.prefixes())if(i!==t&&e.includes(i))return!0;return!1}set(e,t){return e.prop=this.prefixed(e.prop,t),e}needCascade(e){return e._autoprefixerCascade||(e._autoprefixerCascade=this.all.options.cascade!==!1&&e.raw("before").includes(` -`)),e._autoprefixerCascade}maxPrefixed(e,t){if(t._autoprefixerMax)return t._autoprefixerMax;let i=0;for(let n of e)n=Dy.removeNote(n),n.length>i&&(i=n.length);return t._autoprefixerMax=i,t._autoprefixerMax}calcBefore(e,t,i=""){let a=this.maxPrefixed(e,t)-Dy.removeNote(i).length,s=t.raw("before");return a>0&&(s+=Array(a).fill(" ").join("")),s}restoreBefore(e){let t=e.raw("before").split(` -`),i=t[t.length-1];this.all.group(e).up(n=>{let a=n.raw("before").split(` -`),s=a[a.length-1];s.lengths.prop===n.prop&&s.value===n.value)))return this.needCascade(e)&&(n.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,n)}isAlready(e,t){let i=this.all.group(e).up(n=>n.prop===t);return i||(i=this.all.group(e).down(n=>n.prop===t)),i}add(e,t,i,n){let a=this.prefixed(e.prop,t);if(!(this.isAlready(e,a)||this.otherPrefixes(e.value,t)))return this.insert(e,t,i,n)}process(e,t){if(!this.needCascade(e)){super.process(e,t);return}let i=super.process(e,t);!i||!i.length||(this.restoreBefore(e),e.raws.before=this.calcBefore(i,e))}old(e,t){return[this.prefixed(e,t)]}};$y.exports=qy});var My=x((e$,Ly)=>{u();Ly.exports=function r(e){return{mul:t=>new r(e*t),div:t=>new r(e/t),simplify:()=>new r(e),toString:()=>e.toString()}}});var Fy=x((t$,By)=>{u();var fO=My(),cO=wr(),Kl=_e(),pO=/(min|max)-resolution\s*:\s*\d*\.?\d+(dppx|dpcm|dpi|x)/gi,dO=/(min|max)-resolution(\s*:\s*)(\d*\.?\d+)(dppx|dpcm|dpi|x)/i,Ny=class extends cO{prefixName(e,t){return e==="-moz-"?t+"--moz-device-pixel-ratio":e+t+"-device-pixel-ratio"}prefixQuery(e,t,i,n,a){return n=new fO(n),a==="dpi"?n=n.div(96):a==="dpcm"&&(n=n.mul(2.54).div(96)),n=n.simplify(),e==="-o-"&&(n=n.n+"/"+n.d),this.prefixName(e,t)+i+n}clean(e){if(!this.bad){this.bad=[];for(let t of this.prefixes)this.bad.push(this.prefixName(t,"min")),this.bad.push(this.prefixName(t,"max"))}e.params=Kl.editList(e.params,t=>t.filter(i=>this.bad.every(n=>!i.includes(n))))}process(e){let t=this.parentPrefix(e),i=t?[t]:this.prefixes;e.params=Kl.editList(e.params,(n,a)=>{for(let s of n){if(!s.includes("min-resolution")&&!s.includes("max-resolution")){a.push(s);continue}for(let o of i){let l=s.replace(pO,c=>{let f=c.match(dO);return this.prefixQuery(o,f[1],f[2],f[3],f[4])});a.push(l)}a.push(s)}return Kl.uniq(a)})}};By.exports=Ny});var zy=x((r$,jy)=>{u();var Xl="(".charCodeAt(0),Jl=")".charCodeAt(0),Ps="'".charCodeAt(0),Zl='"'.charCodeAt(0),eu="\\".charCodeAt(0),vr="/".charCodeAt(0),tu=",".charCodeAt(0),ru=":".charCodeAt(0),Is="*".charCodeAt(0),hO="u".charCodeAt(0),mO="U".charCodeAt(0),gO="+".charCodeAt(0),yO=/^[a-f0-9?-]+$/i;jy.exports=function(r){for(var e=[],t=r,i,n,a,s,o,l,c,f,d=0,p=t.charCodeAt(d),h=t.length,b=[{nodes:e}],v=0,y,w="",k="",S="";d{u();Uy.exports=function r(e,t,i){var n,a,s,o;for(n=0,a=e.length;n{u();function Hy(r,e){var t=r.type,i=r.value,n,a;return e&&(a=e(r))!==void 0?a:t==="word"||t==="space"?i:t==="string"?(n=r.quote||"",n+i+(r.unclosed?"":n)):t==="comment"?"/*"+i+(r.unclosed?"":"*/"):t==="div"?(r.before||"")+i+(r.after||""):Array.isArray(r.nodes)?(n=Wy(r.nodes,e),t!=="function"?n:i+"("+(r.before||"")+n+(r.after||"")+(r.unclosed?"":")")):i}function Wy(r,e){var t,i;if(Array.isArray(r)){for(t="",i=r.length-1;~i;i-=1)t=Hy(r[i],e)+t;return t}return Hy(r,e)}Gy.exports=Wy});var Ky=x((s$,Yy)=>{u();var Ds="-".charCodeAt(0),qs="+".charCodeAt(0),iu=".".charCodeAt(0),bO="e".charCodeAt(0),wO="E".charCodeAt(0);function vO(r){var e=r.charCodeAt(0),t;if(e===qs||e===Ds){if(t=r.charCodeAt(1),t>=48&&t<=57)return!0;var i=r.charCodeAt(2);return t===iu&&i>=48&&i<=57}return e===iu?(t=r.charCodeAt(1),t>=48&&t<=57):e>=48&&e<=57}Yy.exports=function(r){var e=0,t=r.length,i,n,a;if(t===0||!vO(r))return!1;for(i=r.charCodeAt(e),(i===qs||i===Ds)&&e++;e57));)e+=1;if(i=r.charCodeAt(e),n=r.charCodeAt(e+1),i===iu&&n>=48&&n<=57)for(e+=2;e57));)e+=1;if(i=r.charCodeAt(e),n=r.charCodeAt(e+1),a=r.charCodeAt(e+2),(i===bO||i===wO)&&(n>=48&&n<=57||(n===qs||n===Ds)&&a>=48&&a<=57))for(e+=n===qs||n===Ds?3:2;e57));)e+=1;return{number:r.slice(0,e),unit:r.slice(e)}}});var $s=x((a$,Zy)=>{u();var xO=zy(),Xy=Vy(),Jy=Qy();function Nt(r){return this instanceof Nt?(this.nodes=xO(r),this):new Nt(r)}Nt.prototype.toString=function(){return Array.isArray(this.nodes)?Jy(this.nodes):""};Nt.prototype.walk=function(r,e){return Xy(this.nodes,r,e),this};Nt.unit=Ky();Nt.walk=Xy;Nt.stringify=Jy;Zy.exports=Nt});var nb=x((o$,ib)=>{u();var{list:kO}=$e(),eb=$s(),SO=Mt(),tb=Li(),rb=class{constructor(e){this.props=["transition","transition-property"],this.prefixes=e}add(e,t){let i,n,a=this.prefixes.add[e.prop],s=this.ruleVendorPrefixes(e),o=s||a&&a.prefixes||[],l=this.parse(e.value),c=l.map(h=>this.findProp(h)),f=[];if(c.some(h=>h[0]==="-"))return;for(let h of l){if(n=this.findProp(h),n[0]==="-")continue;let b=this.prefixes.add[n];if(!(!b||!b.prefixes))for(i of b.prefixes){if(s&&!s.some(y=>i.includes(y)))continue;let v=this.prefixes.prefixed(n,i);v!=="-ms-transform"&&!c.includes(v)&&(this.disabled(n,i)||f.push(this.clone(n,v,h)))}}l=l.concat(f);let d=this.stringify(l),p=this.stringify(this.cleanFromUnprefixed(l,"-webkit-"));if(o.includes("-webkit-")&&this.cloneBefore(e,`-webkit-${e.prop}`,p),this.cloneBefore(e,e.prop,p),o.includes("-o-")){let h=this.stringify(this.cleanFromUnprefixed(l,"-o-"));this.cloneBefore(e,`-o-${e.prop}`,h)}for(i of o)if(i!=="-webkit-"&&i!=="-o-"){let h=this.stringify(this.cleanOtherPrefixes(l,i));this.cloneBefore(e,i+e.prop,h)}d!==e.value&&!this.already(e,e.prop,d)&&(this.checkForWarning(t,e),e.cloneBefore(),e.value=d)}findProp(e){let t=e[0].value;if(/^\d/.test(t)){for(let[i,n]of e.entries())if(i!==0&&n.type==="word")return n.value}return t}already(e,t,i){return e.parent.some(n=>n.prop===t&&n.value===i)}cloneBefore(e,t,i){this.already(e,t,i)||e.cloneBefore({prop:t,value:i})}checkForWarning(e,t){if(t.prop!=="transition-property")return;let i=!1,n=!1;t.parent.each(a=>{if(a.type!=="decl"||a.prop.indexOf("transition-")!==0)return;let s=kO.comma(a.value);if(a.prop==="transition-property"){s.forEach(o=>{let l=this.prefixes.add[o];l&&l.prefixes&&l.prefixes.length>0&&(i=!0)});return}return n=n||s.length>1,!1}),i&&n&&t.warn(e,"Replace transition-property to transition, because Autoprefixer could not support any cases of transition-property and other transition-*")}remove(e){let t=this.parse(e.value);t=t.filter(s=>{let o=this.prefixes.remove[this.findProp(s)];return!o||!o.remove});let i=this.stringify(t);if(e.value===i)return;if(t.length===0){e.remove();return}let n=e.parent.some(s=>s.prop===e.prop&&s.value===i),a=e.parent.some(s=>s!==e&&s.prop===e.prop&&s.value.length>i.length);if(n||a){e.remove();return}e.value=i}parse(e){let t=eb(e),i=[],n=[];for(let a of t.nodes)n.push(a),a.type==="div"&&a.value===","&&(i.push(n),n=[]);return i.push(n),i.filter(a=>a.length>0)}stringify(e){if(e.length===0)return"";let t=[];for(let i of e)i[i.length-1].type!=="div"&&i.push(this.div(e)),t=t.concat(i);return t[0].type==="div"&&(t=t.slice(1)),t[t.length-1].type==="div"&&(t=t.slice(0,-2+1||void 0)),eb.stringify({nodes:t})}clone(e,t,i){let n=[],a=!1;for(let s of i)!a&&s.type==="word"&&s.value===e?(n.push({type:"word",value:t}),a=!0):n.push(s);return n}div(e){for(let t of e)for(let i of t)if(i.type==="div"&&i.value===",")return i;return{type:"div",value:",",after:" "}}cleanOtherPrefixes(e,t){return e.filter(i=>{let n=tb.prefix(this.findProp(i));return n===""||n===t})}cleanFromUnprefixed(e,t){let i=e.map(a=>this.findProp(a)).filter(a=>a.slice(0,t.length)===t).map(a=>this.prefixes.unprefixed(a)),n=[];for(let a of e){let s=this.findProp(a),o=tb.prefix(s);!i.includes(s)&&(o===t||o==="")&&n.push(a)}return n}disabled(e,t){let i=["order","justify-content","align-self","align-content"];if(e.includes("flex")||i.includes(e)){if(this.prefixes.options.flexbox===!1)return!0;if(this.prefixes.options.flexbox==="no-2009")return t.includes("2009")}}ruleVendorPrefixes(e){let{parent:t}=e;if(t.type!=="rule")return!1;if(!t.selector.includes(":-"))return!1;let i=SO.prefixes().filter(n=>t.selector.includes(":"+n));return i.length>0?i:!1}};ib.exports=rb});var xr=x((l$,ab)=>{u();var AO=_e(),sb=class{constructor(e,t,i,n){this.unprefixed=e,this.prefixed=t,this.string=i||t,this.regexp=n||AO.regexp(t)}check(e){return e.includes(this.string)?!!e.match(this.regexp):!1}};ab.exports=sb});var He=x((u$,lb)=>{u();var CO=wr(),_O=xr(),EO=Li(),OO=_e(),ob=class extends CO{static save(e,t){let i=t.prop,n=[];for(let a in t._autoprefixerValues){let s=t._autoprefixerValues[a];if(s===t.value)continue;let o,l=EO.prefix(i);if(l==="-pie-")continue;if(l===a){o=t.value=s,n.push(o);continue}let c=e.prefixed(i,a),f=t.parent;if(!f.every(b=>b.prop!==c)){n.push(o);continue}let d=s.replace(/\s+/," ");if(f.some(b=>b.prop===t.prop&&b.value.replace(/\s+/," ")===d)){n.push(o);continue}let h=this.clone(t,{value:s});o=t.parent.insertBefore(t,h),n.push(o)}return n}check(e){let t=e.value;return t.includes(this.name)?!!t.match(this.regexp()):!1}regexp(){return this.regexpCache||(this.regexpCache=OO.regexp(this.name))}replace(e,t){return e.replace(this.regexp(),`$1${t}$2`)}value(e){return e.raws.value&&e.raws.value.value===e.value?e.raws.value.raw:e.value}add(e,t){e._autoprefixerValues||(e._autoprefixerValues={});let i=e._autoprefixerValues[t]||this.value(e),n;do if(n=i,i=this.replace(i,t),i===!1)return;while(i!==n);e._autoprefixerValues[t]=i}old(e){return new _O(this.name,e+this.name)}};lb.exports=ob});var Bt=x((f$,ub)=>{u();ub.exports={}});var su=x((c$,pb)=>{u();var fb=$s(),TO=He(),RO=Bt().insertAreas,PO=/(^|[^-])linear-gradient\(\s*(top|left|right|bottom)/i,IO=/(^|[^-])radial-gradient\(\s*\d+(\w*|%)\s+\d+(\w*|%)\s*,/i,DO=/(!\s*)?autoprefixer:\s*ignore\s+next/i,qO=/(!\s*)?autoprefixer\s*grid:\s*(on|off|(no-)?autoplace)/i,$O=["width","height","min-width","max-width","min-height","max-height","inline-size","min-inline-size","max-inline-size","block-size","min-block-size","max-block-size"];function nu(r){return r.parent.some(e=>e.prop==="grid-template"||e.prop==="grid-template-areas")}function LO(r){let e=r.parent.some(i=>i.prop==="grid-template-rows"),t=r.parent.some(i=>i.prop==="grid-template-columns");return e&&t}var cb=class{constructor(e){this.prefixes=e}add(e,t){let i=this.prefixes.add["@resolution"],n=this.prefixes.add["@keyframes"],a=this.prefixes.add["@viewport"],s=this.prefixes.add["@supports"];e.walkAtRules(f=>{if(f.name==="keyframes"){if(!this.disabled(f,t))return n&&n.process(f)}else if(f.name==="viewport"){if(!this.disabled(f,t))return a&&a.process(f)}else if(f.name==="supports"){if(this.prefixes.options.supports!==!1&&!this.disabled(f,t))return s.process(f)}else if(f.name==="media"&&f.params.includes("-resolution")&&!this.disabled(f,t))return i&&i.process(f)}),e.walkRules(f=>{if(!this.disabled(f,t))return this.prefixes.add.selectors.map(d=>d.process(f,t))});function o(f){return f.parent.nodes.some(d=>{if(d.type!=="decl")return!1;let p=d.prop==="display"&&/(inline-)?grid/.test(d.value),h=d.prop.startsWith("grid-template"),b=/^grid-([A-z]+-)?gap/.test(d.prop);return p||h||b})}function l(f){return f.parent.some(d=>d.prop==="display"&&/(inline-)?flex/.test(d.value))}let c=this.gridStatus(e,t)&&this.prefixes.add["grid-area"]&&this.prefixes.add["grid-area"].prefixes;return e.walkDecls(f=>{if(this.disabledDecl(f,t))return;let d=f.parent,p=f.prop,h=f.value;if(p==="grid-row-span"){t.warn("grid-row-span is not part of final Grid Layout. Use grid-row.",{node:f});return}else if(p==="grid-column-span"){t.warn("grid-column-span is not part of final Grid Layout. Use grid-column.",{node:f});return}else if(p==="display"&&h==="box"){t.warn("You should write display: flex by final spec instead of display: box",{node:f});return}else if(p==="text-emphasis-position")(h==="under"||h==="over")&&t.warn("You should use 2 values for text-emphasis-position For example, `under left` instead of just `under`.",{node:f});else if(/^(align|justify|place)-(items|content)$/.test(p)&&l(f))(h==="start"||h==="end")&&t.warn(`${h} value has mixed support, consider using flex-${h} instead`,{node:f});else if(p==="text-decoration-skip"&&h==="ink")t.warn("Replace text-decoration-skip: ink to text-decoration-skip-ink: auto, because spec had been changed",{node:f});else{if(c&&this.gridStatus(f,t))if(f.value==="subgrid"&&t.warn("IE does not support subgrid",{node:f}),/^(align|justify|place)-items$/.test(p)&&o(f)){let v=p.replace("-items","-self");t.warn(`IE does not support ${p} on grid containers. Try using ${v} on child elements instead: ${f.parent.selector} > * { ${v}: ${f.value} }`,{node:f})}else if(/^(align|justify|place)-content$/.test(p)&&o(f))t.warn(`IE does not support ${f.prop} on grid containers`,{node:f});else if(p==="display"&&f.value==="contents"){t.warn("Please do not use display: contents; if you have grid setting enabled",{node:f});return}else if(f.prop==="grid-gap"){let v=this.gridStatus(f,t);v==="autoplace"&&!LO(f)&&!nu(f)?t.warn("grid-gap only works if grid-template(-areas) is being used or both rows and columns have been declared and cells have not been manually placed inside the explicit grid",{node:f}):(v===!0||v==="no-autoplace")&&!nu(f)&&t.warn("grid-gap only works if grid-template(-areas) is being used",{node:f})}else if(p==="grid-auto-columns"){t.warn("grid-auto-columns is not supported by IE",{node:f});return}else if(p==="grid-auto-rows"){t.warn("grid-auto-rows is not supported by IE",{node:f});return}else if(p==="grid-auto-flow"){let v=d.some(w=>w.prop==="grid-template-rows"),y=d.some(w=>w.prop==="grid-template-columns");nu(f)?t.warn("grid-auto-flow is not supported by IE",{node:f}):h.includes("dense")?t.warn("grid-auto-flow: dense is not supported by IE",{node:f}):!v&&!y&&t.warn("grid-auto-flow works only if grid-template-rows and grid-template-columns are present in the same rule",{node:f});return}else if(h.includes("auto-fit")){t.warn("auto-fit value is not supported by IE",{node:f,word:"auto-fit"});return}else if(h.includes("auto-fill")){t.warn("auto-fill value is not supported by IE",{node:f,word:"auto-fill"});return}else p.startsWith("grid-template")&&h.includes("[")&&t.warn("Autoprefixer currently does not support line names. Try using grid-template-areas instead.",{node:f,word:"["});if(h.includes("radial-gradient"))if(IO.test(f.value))t.warn("Gradient has outdated direction syntax. New syntax is like `closest-side at 0 0` instead of `0 0, closest-side`.",{node:f});else{let v=fb(h);for(let y of v.nodes)if(y.type==="function"&&y.value==="radial-gradient")for(let w of y.nodes)w.type==="word"&&(w.value==="cover"?t.warn("Gradient has outdated direction syntax. Replace `cover` to `farthest-corner`.",{node:f}):w.value==="contain"&&t.warn("Gradient has outdated direction syntax. Replace `contain` to `closest-side`.",{node:f}))}h.includes("linear-gradient")&&PO.test(h)&&t.warn("Gradient has outdated direction syntax. New syntax is like `to left` instead of `right`.",{node:f})}$O.includes(f.prop)&&(f.value.includes("-fill-available")||(f.value.includes("fill-available")?t.warn("Replace fill-available to stretch, because spec had been changed",{node:f}):f.value.includes("fill")&&fb(h).nodes.some(y=>y.type==="word"&&y.value==="fill")&&t.warn("Replace fill to stretch, because spec had been changed",{node:f})));let b;if(f.prop==="transition"||f.prop==="transition-property")return this.prefixes.transition.add(f,t);if(f.prop==="align-self"){if(this.displayType(f)!=="grid"&&this.prefixes.options.flexbox!==!1&&(b=this.prefixes.add["align-self"],b&&b.prefixes&&b.process(f)),this.gridStatus(f,t)!==!1&&(b=this.prefixes.add["grid-row-align"],b&&b.prefixes))return b.process(f,t)}else if(f.prop==="justify-self"){if(this.gridStatus(f,t)!==!1&&(b=this.prefixes.add["grid-column-align"],b&&b.prefixes))return b.process(f,t)}else if(f.prop==="place-self"){if(b=this.prefixes.add["place-self"],b&&b.prefixes&&this.gridStatus(f,t)!==!1)return b.process(f,t)}else if(b=this.prefixes.add[f.prop],b&&b.prefixes)return b.process(f,t)}),this.gridStatus(e,t)&&RO(e,this.disabled),e.walkDecls(f=>{if(this.disabledValue(f,t))return;let d=this.prefixes.unprefixed(f.prop),p=this.prefixes.values("add",d);if(Array.isArray(p))for(let h of p)h.process&&h.process(f,t);TO.save(this.prefixes,f)})}remove(e,t){let i=this.prefixes.remove["@resolution"];e.walkAtRules((n,a)=>{this.prefixes.remove[`@${n.name}`]?this.disabled(n,t)||n.parent.removeChild(a):n.name==="media"&&n.params.includes("-resolution")&&i&&i.clean(n)});for(let n of this.prefixes.remove.selectors)e.walkRules((a,s)=>{n.check(a)&&(this.disabled(a,t)||a.parent.removeChild(s))});return e.walkDecls((n,a)=>{if(this.disabled(n,t))return;let s=n.parent,o=this.prefixes.unprefixed(n.prop);if((n.prop==="transition"||n.prop==="transition-property")&&this.prefixes.transition.remove(n),this.prefixes.remove[n.prop]&&this.prefixes.remove[n.prop].remove){let l=this.prefixes.group(n).down(c=>this.prefixes.normalize(c.prop)===o);if(o==="flex-flow"&&(l=!0),n.prop==="-webkit-box-orient"){let c={"flex-direction":!0,"flex-flow":!0};if(!n.parent.some(f=>c[f.prop]))return}if(l&&!this.withHackValue(n)){n.raw("before").includes(` -`)&&this.reduceSpaces(n),s.removeChild(a);return}}for(let l of this.prefixes.values("remove",o)){if(!l.check||!l.check(n.value))continue;if(o=l.unprefixed,this.prefixes.group(n).down(f=>f.value.includes(o))){s.removeChild(a);return}}})}withHackValue(e){return e.prop==="-webkit-background-clip"&&e.value==="text"}disabledValue(e,t){return this.gridStatus(e,t)===!1&&e.type==="decl"&&e.prop==="display"&&e.value.includes("grid")||this.prefixes.options.flexbox===!1&&e.type==="decl"&&e.prop==="display"&&e.value.includes("flex")||e.type==="decl"&&e.prop==="content"?!0:this.disabled(e,t)}disabledDecl(e,t){if(this.gridStatus(e,t)===!1&&e.type==="decl"&&(e.prop.includes("grid")||e.prop==="justify-items"))return!0;if(this.prefixes.options.flexbox===!1&&e.type==="decl"){let i=["order","justify-content","align-items","align-content"];if(e.prop.includes("flex")||i.includes(e.prop))return!0}return this.disabled(e,t)}disabled(e,t){if(!e)return!1;if(e._autoprefixerDisabled!==void 0)return e._autoprefixerDisabled;if(e.parent){let n=e.prev();if(n&&n.type==="comment"&&DO.test(n.text))return e._autoprefixerDisabled=!0,e._autoprefixerSelfDisabled=!0,!0}let i=null;if(e.nodes){let n;e.each(a=>{a.type==="comment"&&/(!\s*)?autoprefixer:\s*(off|on)/i.test(a.text)&&(typeof n!="undefined"?t.warn("Second Autoprefixer control comment was ignored. Autoprefixer applies control comment to whole block, not to next rules.",{node:a}):n=/on/i.test(a.text))}),n!==void 0&&(i=!n)}if(!e.nodes||i===null)if(e.parent){let n=this.disabled(e.parent,t);e.parent._autoprefixerSelfDisabled===!0?i=!1:i=n}else i=!1;return e._autoprefixerDisabled=i,i}reduceSpaces(e){let t=!1;if(this.prefixes.group(e).up(()=>(t=!0,!0)),t)return;let i=e.raw("before").split(` -`),n=i[i.length-1].length,a=!1;this.prefixes.group(e).down(s=>{i=s.raw("before").split(` -`);let o=i.length-1;i[o].length>n&&(a===!1&&(a=i[o].length-n),i[o]=i[o].slice(0,-a),s.raws.before=i.join(` -`))})}displayType(e){for(let t of e.parent.nodes)if(t.prop==="display"){if(t.value.includes("flex"))return"flex";if(t.value.includes("grid"))return"grid"}return!1}gridStatus(e,t){if(!e)return!1;if(e._autoprefixerGridStatus!==void 0)return e._autoprefixerGridStatus;let i=null;if(e.nodes){let n;e.each(a=>{if(a.type==="comment"&&qO.test(a.text)){let s=/:\s*autoplace/i.test(a.text),o=/no-autoplace/i.test(a.text);typeof n!="undefined"?t.warn("Second Autoprefixer grid control comment was ignored. Autoprefixer applies control comments to the whole block, not to the next rules.",{node:a}):s?n="autoplace":o?n=!0:n=/on/i.test(a.text)}}),n!==void 0&&(i=n)}if(e.type==="atrule"&&e.name==="supports"){let n=e.params;n.includes("grid")&&n.includes("auto")&&(i=!1)}if(!e.nodes||i===null)if(e.parent){let n=this.gridStatus(e.parent,t);e.parent._autoprefixerSelfDisabled===!0?i=!1:i=n}else typeof this.prefixes.options.grid!="undefined"?i=this.prefixes.options.grid:typeof m.env.AUTOPREFIXER_GRID!="undefined"?m.env.AUTOPREFIXER_GRID==="autoplace"?i="autoplace":i=!0:i=!1;return e._autoprefixerGridStatus=i,i}};pb.exports=cb});var hb=x((p$,db)=>{u();db.exports={A:{A:{"2":"K E F G A B JC"},B:{"1":"C L M H N D O P Q R S T U V W X Y Z a b c d e f g h i j n o p q r s t u v w x y z I"},C:{"1":"2 3 4 5 6 7 8 9 AB BB CB DB EB FB GB HB IB JB KB LB MB NB OB PB QB RB SB TB UB VB WB XB YB ZB aB bB cB 0B dB 1B eB fB gB hB iB jB kB lB mB nB oB m pB qB rB sB tB P Q R 2B S T U V W X Y Z a b c d e f g h i j n o p q r s t u v w x y z I uB 3B 4B","2":"0 1 KC zB J K E F G A B C L M H N D O k l LC MC"},D:{"1":"8 9 AB BB CB DB EB FB GB HB IB JB KB LB MB NB OB PB QB RB SB TB UB VB WB XB YB ZB aB bB cB 0B dB 1B eB fB gB hB iB jB kB lB mB nB oB m pB qB rB sB tB P Q R S T U V W X Y Z a b c d e f g h i j n o p q r s t u v w x y z I uB 3B 4B","2":"0 1 2 3 4 5 6 7 J K E F G A B C L M H N D O k l"},E:{"1":"G A B C L M H D RC 6B vB wB 7B SC TC 8B 9B xB AC yB BC CC DC EC FC GC UC","2":"0 J K E F NC 5B OC PC QC"},F:{"1":"1 2 3 4 5 6 7 8 9 H N D O k l AB BB CB DB EB FB GB HB IB JB KB LB MB NB OB PB QB RB SB TB UB VB WB XB YB ZB aB bB cB dB eB fB gB hB iB jB kB lB mB nB oB m pB qB rB sB tB P Q R 2B S T U V W X Y Z a b c d e f g h i j wB","2":"G B C VC WC XC YC vB HC ZC"},G:{"1":"D fC gC hC iC jC kC lC mC nC oC pC qC rC sC tC 8B 9B xB AC yB BC CC DC EC FC GC","2":"F 5B aC IC bC cC dC eC"},H:{"1":"uC"},I:{"1":"I zC 0C","2":"zB J vC wC xC yC IC"},J:{"2":"E A"},K:{"1":"m","2":"A B C vB HC wB"},L:{"1":"I"},M:{"1":"uB"},N:{"2":"A B"},O:{"1":"xB"},P:{"1":"J k l 1C 2C 3C 4C 5C 6B 6C 7C 8C 9C AD yB BD CD DD"},Q:{"1":"7B"},R:{"1":"ED"},S:{"1":"FD GD"}},B:4,C:"CSS Feature Queries"}});var bb=x((d$,yb)=>{u();function mb(r){return r[r.length-1]}var gb={parse(r){let e=[""],t=[e];for(let i of r){if(i==="("){e=[""],mb(t).push(e),t.push(e);continue}if(i===")"){t.pop(),e=mb(t),e.push("");continue}e[e.length-1]+=i}return t[0]},stringify(r){let e="";for(let t of r){if(typeof t=="object"){e+=`(${gb.stringify(t)})`;continue}e+=t}return e}};yb.exports=gb});var Sb=x((h$,kb)=>{u();var MO=hb(),{feature:NO}=(Ts(),Os),{parse:BO}=$e(),FO=Mt(),au=bb(),jO=He(),zO=_e(),wb=NO(MO),vb=[];for(let r in wb.stats){let e=wb.stats[r];for(let t in e){let i=e[t];/y/.test(i)&&vb.push(r+" "+t)}}var xb=class{constructor(e,t){this.Prefixes=e,this.all=t}prefixer(){if(this.prefixerCache)return this.prefixerCache;let e=this.all.browsers.selected.filter(i=>vb.includes(i)),t=new FO(this.all.browsers.data,e,this.all.options);return this.prefixerCache=new this.Prefixes(this.all.data,t,this.all.options),this.prefixerCache}parse(e){let t=e.split(":"),i=t[0],n=t[1];return n||(n=""),[i.trim(),n.trim()]}virtual(e){let[t,i]=this.parse(e),n=BO("a{}").first;return n.append({prop:t,value:i,raws:{before:""}}),n}prefixed(e){let t=this.virtual(e);if(this.disabled(t.first))return t.nodes;let i={warn:()=>null},n=this.prefixer().add[t.first.prop];n&&n.process&&n.process(t.first,i);for(let a of t.nodes){for(let s of this.prefixer().values("add",t.first.prop))s.process(a);jO.save(this.all,a)}return t.nodes}isNot(e){return typeof e=="string"&&/not\s*/i.test(e)}isOr(e){return typeof e=="string"&&/\s*or\s*/i.test(e)}isProp(e){return typeof e=="object"&&e.length===1&&typeof e[0]=="string"}isHack(e,t){return!new RegExp(`(\\(|\\s)${zO.escapeRegexp(t)}:`).test(e)}toRemove(e,t){let[i,n]=this.parse(e),a=this.all.unprefixed(i),s=this.all.cleaner();if(s.remove[i]&&s.remove[i].remove&&!this.isHack(t,a))return!0;for(let o of s.values("remove",a))if(o.check(n))return!0;return!1}remove(e,t){let i=0;for(;itypeof t!="object"?t:t.length===1&&typeof t[0]=="object"?this.cleanBrackets(t[0]):this.cleanBrackets(t))}convert(e){let t=[""];for(let i of e)t.push([`${i.prop}: ${i.value}`]),t.push(" or ");return t[t.length-1]="",t}normalize(e){if(typeof e!="object")return e;if(e=e.filter(t=>t!==""),typeof e[0]=="string"){let t=e[0].trim();if(t.includes(":")||t==="selector"||t==="not selector")return[au.stringify(e)]}return e.map(t=>this.normalize(t))}add(e,t){return e.map(i=>{if(this.isProp(i)){let n=this.prefixed(i[0]);return n.length>1?this.convert(n):i}return typeof i=="object"?this.add(i,t):i})}process(e){let t=au.parse(e.params);t=this.normalize(t),t=this.remove(t,e.params),t=this.add(t,e.params),t=this.cleanBrackets(t),e.params=au.stringify(t)}disabled(e){if(!this.all.options.grid&&(e.prop==="display"&&e.value.includes("grid")||e.prop.includes("grid")||e.prop==="justify-items"))return!0;if(this.all.options.flexbox===!1){if(e.prop==="display"&&e.value.includes("flex"))return!0;let t=["order","justify-content","align-items","align-content"];if(e.prop.includes("flex")||t.includes(e.prop))return!0}return!1}};kb.exports=xb});var _b=x((m$,Cb)=>{u();var Ab=class{constructor(e,t){this.prefix=t,this.prefixed=e.prefixed(this.prefix),this.regexp=e.regexp(this.prefix),this.prefixeds=e.possible().map(i=>[e.prefixed(i),e.regexp(i)]),this.unprefixed=e.name,this.nameRegexp=e.regexp()}isHack(e){let t=e.parent.index(e)+1,i=e.parent.nodes;for(;t{u();var{list:UO}=$e(),VO=_b(),HO=wr(),WO=Mt(),GO=_e(),Eb=class extends HO{constructor(e,t,i){super(e,t,i);this.regexpCache=new Map}check(e){return e.selector.includes(this.name)?!!e.selector.match(this.regexp()):!1}prefixed(e){return this.name.replace(/^(\W*)/,`$1${e}`)}regexp(e){if(!this.regexpCache.has(e)){let t=e?this.prefixed(e):this.name;this.regexpCache.set(e,new RegExp(`(^|[^:"'=])${GO.escapeRegexp(t)}`,"gi"))}return this.regexpCache.get(e)}possible(){return WO.prefixes()}prefixeds(e){if(e._autoprefixerPrefixeds){if(e._autoprefixerPrefixeds[this.name])return e._autoprefixerPrefixeds}else e._autoprefixerPrefixeds={};let t={};if(e.selector.includes(",")){let n=UO.comma(e.selector).filter(a=>a.includes(this.name));for(let a of this.possible())t[a]=n.map(s=>this.replace(s,a)).join(", ")}else for(let i of this.possible())t[i]=this.replace(e.selector,i);return e._autoprefixerPrefixeds[this.name]=t,e._autoprefixerPrefixeds}already(e,t,i){let n=e.parent.index(e)-1;for(;n>=0;){let a=e.parent.nodes[n];if(a.type!=="rule")return!1;let s=!1;for(let o in t[this.name]){let l=t[this.name][o];if(a.selector===l){if(i===o)return!0;s=!0;break}}if(!s)return!1;n-=1}return!1}replace(e,t){return e.replace(this.regexp(),`$1${this.prefixed(t)}`)}add(e,t){let i=this.prefixeds(e);if(this.already(e,i,t))return;let n=this.clone(e,{selector:i[this.name][t]});e.parent.insertBefore(e,n)}old(e){return new VO(this,e)}};Ob.exports=Eb});var Pb=x((y$,Rb)=>{u();var QO=wr(),Tb=class extends QO{add(e,t){let i=t+e.name;if(e.parent.some(s=>s.name===i&&s.params===e.params))return;let a=this.clone(e,{name:i});return e.parent.insertBefore(e,a)}process(e){let t=this.parentPrefix(e);for(let i of this.prefixes)(!t||t===i)&&this.add(e,i)}};Rb.exports=Tb});var Db=x((b$,Ib)=>{u();var YO=kr(),ou=class extends YO{prefixed(e){return e==="-webkit-"?":-webkit-full-screen":e==="-moz-"?":-moz-full-screen":`:${e}fullscreen`}};ou.names=[":fullscreen"];Ib.exports=ou});var $b=x((w$,qb)=>{u();var KO=kr(),lu=class extends KO{possible(){return super.possible().concat(["-moz- old","-ms- old"])}prefixed(e){return e==="-webkit-"?"::-webkit-input-placeholder":e==="-ms-"?"::-ms-input-placeholder":e==="-ms- old"?":-ms-input-placeholder":e==="-moz- old"?":-moz-placeholder":`::${e}placeholder`}};lu.names=["::placeholder"];qb.exports=lu});var Mb=x((v$,Lb)=>{u();var XO=kr(),uu=class extends XO{prefixed(e){return e==="-ms-"?":-ms-input-placeholder":`:${e}placeholder-shown`}};uu.names=[":placeholder-shown"];Lb.exports=uu});var Bb=x((x$,Nb)=>{u();var JO=kr(),ZO=_e(),fu=class extends JO{constructor(e,t,i){super(e,t,i);this.prefixes&&(this.prefixes=ZO.uniq(this.prefixes.map(n=>"-webkit-")))}prefixed(e){return e==="-webkit-"?"::-webkit-file-upload-button":`::${e}file-selector-button`}};fu.names=["::file-selector-button"];Nb.exports=fu});var Pe=x((k$,Fb)=>{u();Fb.exports=function(r){let e;return r==="-webkit- 2009"||r==="-moz-"?e=2009:r==="-ms-"?e=2012:r==="-webkit-"&&(e="final"),r==="-webkit- 2009"&&(r="-webkit-"),[e,r]}});var Vb=x((S$,Ub)=>{u();var jb=$e().list,zb=Pe(),eT=j(),Sr=class extends eT{prefixed(e,t){let i;return[i,t]=zb(t),i===2009?t+"box-flex":super.prefixed(e,t)}normalize(){return"flex"}set(e,t){let i=zb(t)[0];if(i===2009)return e.value=jb.space(e.value)[0],e.value=Sr.oldValues[e.value]||e.value,super.set(e,t);if(i===2012){let n=jb.space(e.value);n.length===3&&n[2]==="0"&&(e.value=n.slice(0,2).concat("0px").join(" "))}return super.set(e,t)}};Sr.names=["flex","box-flex"];Sr.oldValues={auto:"1",none:"0"};Ub.exports=Sr});var Gb=x((A$,Wb)=>{u();var Hb=Pe(),tT=j(),cu=class extends tT{prefixed(e,t){let i;return[i,t]=Hb(t),i===2009?t+"box-ordinal-group":i===2012?t+"flex-order":super.prefixed(e,t)}normalize(){return"order"}set(e,t){return Hb(t)[0]===2009&&/\d/.test(e.value)?(e.value=(parseInt(e.value)+1).toString(),super.set(e,t)):super.set(e,t)}};cu.names=["order","flex-order","box-ordinal-group"];Wb.exports=cu});var Yb=x((C$,Qb)=>{u();var rT=j(),pu=class extends rT{check(e){let t=e.value;return!t.toLowerCase().includes("alpha(")&&!t.includes("DXImageTransform.Microsoft")&&!t.includes("data:image/svg+xml")}};pu.names=["filter"];Qb.exports=pu});var Xb=x((_$,Kb)=>{u();var iT=j(),du=class extends iT{insert(e,t,i,n){if(t!=="-ms-")return super.insert(e,t,i);let a=this.clone(e),s=e.prop.replace(/end$/,"start"),o=t+e.prop.replace(/end$/,"span");if(!e.parent.some(l=>l.prop===o)){if(a.prop=o,e.value.includes("span"))a.value=e.value.replace(/span\s/i,"");else{let l;if(e.parent.walkDecls(s,c=>{l=c}),l){let c=Number(e.value)-Number(l.value)+"";a.value=c}else e.warn(n,`Can not prefix ${e.prop} (${s} is not found)`)}e.cloneBefore(a)}}};du.names=["grid-row-end","grid-column-end"];Kb.exports=du});var Zb=x((E$,Jb)=>{u();var nT=j(),hu=class extends nT{check(e){return!e.value.split(/\s+/).some(t=>{let i=t.toLowerCase();return i==="reverse"||i==="alternate-reverse"})}};hu.names=["animation","animation-direction"];Jb.exports=hu});var tw=x((O$,ew)=>{u();var sT=Pe(),aT=j(),mu=class extends aT{insert(e,t,i){let n;if([n,t]=sT(t),n!==2009)return super.insert(e,t,i);let a=e.value.split(/\s+/).filter(d=>d!=="wrap"&&d!=="nowrap"&&"wrap-reverse");if(a.length===0||e.parent.some(d=>d.prop===t+"box-orient"||d.prop===t+"box-direction"))return;let o=a[0],l=o.includes("row")?"horizontal":"vertical",c=o.includes("reverse")?"reverse":"normal",f=this.clone(e);return f.prop=t+"box-orient",f.value=l,this.needCascade(e)&&(f.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,f),f=this.clone(e),f.prop=t+"box-direction",f.value=c,this.needCascade(e)&&(f.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,f)}};mu.names=["flex-flow","box-direction","box-orient"];ew.exports=mu});var iw=x((T$,rw)=>{u();var oT=Pe(),lT=j(),gu=class extends lT{normalize(){return"flex"}prefixed(e,t){let i;return[i,t]=oT(t),i===2009?t+"box-flex":i===2012?t+"flex-positive":super.prefixed(e,t)}};gu.names=["flex-grow","flex-positive"];rw.exports=gu});var sw=x((R$,nw)=>{u();var uT=Pe(),fT=j(),yu=class extends fT{set(e,t){if(uT(t)[0]!==2009)return super.set(e,t)}};yu.names=["flex-wrap"];nw.exports=yu});var ow=x((P$,aw)=>{u();var cT=j(),Ar=Bt(),bu=class extends cT{insert(e,t,i,n){if(t!=="-ms-")return super.insert(e,t,i);let a=Ar.parse(e),[s,o]=Ar.translate(a,0,2),[l,c]=Ar.translate(a,1,3);[["grid-row",s],["grid-row-span",o],["grid-column",l],["grid-column-span",c]].forEach(([f,d])=>{Ar.insertDecl(e,f,d)}),Ar.warnTemplateSelectorNotFound(e,n),Ar.warnIfGridRowColumnExists(e,n)}};bu.names=["grid-area"];aw.exports=bu});var uw=x((I$,lw)=>{u();var pT=j(),Mi=Bt(),wu=class extends pT{insert(e,t,i){if(t!=="-ms-")return super.insert(e,t,i);if(e.parent.some(s=>s.prop==="-ms-grid-row-align"))return;let[[n,a]]=Mi.parse(e);a?(Mi.insertDecl(e,"grid-row-align",n),Mi.insertDecl(e,"grid-column-align",a)):(Mi.insertDecl(e,"grid-row-align",n),Mi.insertDecl(e,"grid-column-align",n))}};wu.names=["place-self"];lw.exports=wu});var cw=x((D$,fw)=>{u();var dT=j(),vu=class extends dT{check(e){let t=e.value;return!t.includes("/")||t.includes("span")}normalize(e){return e.replace("-start","")}prefixed(e,t){let i=super.prefixed(e,t);return t==="-ms-"&&(i=i.replace("-start","")),i}};vu.names=["grid-row-start","grid-column-start"];fw.exports=vu});var hw=x((q$,dw)=>{u();var pw=Pe(),hT=j(),Cr=class extends hT{check(e){return e.parent&&!e.parent.some(t=>t.prop&&t.prop.startsWith("grid-"))}prefixed(e,t){let i;return[i,t]=pw(t),i===2012?t+"flex-item-align":super.prefixed(e,t)}normalize(){return"align-self"}set(e,t){let i=pw(t)[0];if(i===2012)return e.value=Cr.oldValues[e.value]||e.value,super.set(e,t);if(i==="final")return super.set(e,t)}};Cr.names=["align-self","flex-item-align"];Cr.oldValues={"flex-end":"end","flex-start":"start"};dw.exports=Cr});var gw=x(($$,mw)=>{u();var mT=j(),gT=_e(),xu=class extends mT{constructor(e,t,i){super(e,t,i);this.prefixes&&(this.prefixes=gT.uniq(this.prefixes.map(n=>n==="-ms-"?"-webkit-":n)))}};xu.names=["appearance"];mw.exports=xu});var ww=x((L$,bw)=>{u();var yw=Pe(),yT=j(),ku=class extends yT{normalize(){return"flex-basis"}prefixed(e,t){let i;return[i,t]=yw(t),i===2012?t+"flex-preferred-size":super.prefixed(e,t)}set(e,t){let i;if([i,t]=yw(t),i===2012||i==="final")return super.set(e,t)}};ku.names=["flex-basis","flex-preferred-size"];bw.exports=ku});var xw=x((M$,vw)=>{u();var bT=j(),Su=class extends bT{normalize(){return this.name.replace("box-image","border")}prefixed(e,t){let i=super.prefixed(e,t);return t==="-webkit-"&&(i=i.replace("border","box-image")),i}};Su.names=["mask-border","mask-border-source","mask-border-slice","mask-border-width","mask-border-outset","mask-border-repeat","mask-box-image","mask-box-image-source","mask-box-image-slice","mask-box-image-width","mask-box-image-outset","mask-box-image-repeat"];vw.exports=Su});var Sw=x((N$,kw)=>{u();var wT=j(),lt=class extends wT{insert(e,t,i){let n=e.prop==="mask-composite",a;n?a=e.value.split(","):a=e.value.match(lt.regexp)||[],a=a.map(c=>c.trim()).filter(c=>c);let s=a.length,o;if(s&&(o=this.clone(e),o.value=a.map(c=>lt.oldValues[c]||c).join(", "),a.includes("intersect")&&(o.value+=", xor"),o.prop=t+"mask-composite"),n)return s?(this.needCascade(e)&&(o.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,o)):void 0;let l=this.clone(e);return l.prop=t+l.prop,s&&(l.value=l.value.replace(lt.regexp,"")),this.needCascade(e)&&(l.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,l),s?(this.needCascade(e)&&(o.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,o)):e}};lt.names=["mask","mask-composite"];lt.oldValues={add:"source-over",subtract:"source-out",intersect:"source-in",exclude:"xor"};lt.regexp=new RegExp(`\\s+(${Object.keys(lt.oldValues).join("|")})\\b(?!\\))\\s*(?=[,])`,"ig");kw.exports=lt});var _w=x((B$,Cw)=>{u();var Aw=Pe(),vT=j(),_r=class extends vT{prefixed(e,t){let i;return[i,t]=Aw(t),i===2009?t+"box-align":i===2012?t+"flex-align":super.prefixed(e,t)}normalize(){return"align-items"}set(e,t){let i=Aw(t)[0];return(i===2009||i===2012)&&(e.value=_r.oldValues[e.value]||e.value),super.set(e,t)}};_r.names=["align-items","flex-align","box-align"];_r.oldValues={"flex-end":"end","flex-start":"start"};Cw.exports=_r});var Ow=x((F$,Ew)=>{u();var xT=j(),Au=class extends xT{set(e,t){return t==="-ms-"&&e.value==="contain"&&(e.value="element"),super.set(e,t)}insert(e,t,i){if(!(e.value==="all"&&t==="-ms-"))return super.insert(e,t,i)}};Au.names=["user-select"];Ew.exports=Au});var Pw=x((j$,Rw)=>{u();var Tw=Pe(),kT=j(),Cu=class extends kT{normalize(){return"flex-shrink"}prefixed(e,t){let i;return[i,t]=Tw(t),i===2012?t+"flex-negative":super.prefixed(e,t)}set(e,t){let i;if([i,t]=Tw(t),i===2012||i==="final")return super.set(e,t)}};Cu.names=["flex-shrink","flex-negative"];Rw.exports=Cu});var Dw=x((z$,Iw)=>{u();var ST=j(),_u=class extends ST{prefixed(e,t){return`${t}column-${e}`}normalize(e){return e.includes("inside")?"break-inside":e.includes("before")?"break-before":"break-after"}set(e,t){return(e.prop==="break-inside"&&e.value==="avoid-column"||e.value==="avoid-page")&&(e.value="avoid"),super.set(e,t)}insert(e,t,i){if(e.prop!=="break-inside")return super.insert(e,t,i);if(!(/region/i.test(e.value)||/page/i.test(e.value)))return super.insert(e,t,i)}};_u.names=["break-inside","page-break-inside","column-break-inside","break-before","page-break-before","column-break-before","break-after","page-break-after","column-break-after"];Iw.exports=_u});var $w=x((U$,qw)=>{u();var AT=j(),Eu=class extends AT{prefixed(e,t){return t+"print-color-adjust"}normalize(){return"color-adjust"}};Eu.names=["color-adjust","print-color-adjust"];qw.exports=Eu});var Mw=x((V$,Lw)=>{u();var CT=j(),Er=class extends CT{insert(e,t,i){if(t==="-ms-"){let n=this.set(this.clone(e),t);this.needCascade(e)&&(n.raws.before=this.calcBefore(i,e,t));let a="ltr";return e.parent.nodes.forEach(s=>{s.prop==="direction"&&(s.value==="rtl"||s.value==="ltr")&&(a=s.value)}),n.value=Er.msValues[a][e.value]||e.value,e.parent.insertBefore(e,n)}return super.insert(e,t,i)}};Er.names=["writing-mode"];Er.msValues={ltr:{"horizontal-tb":"lr-tb","vertical-rl":"tb-rl","vertical-lr":"tb-lr"},rtl:{"horizontal-tb":"rl-tb","vertical-rl":"bt-rl","vertical-lr":"bt-lr"}};Lw.exports=Er});var Bw=x((H$,Nw)=>{u();var _T=j(),Ou=class extends _T{set(e,t){return e.value=e.value.replace(/\s+fill(\s)/,"$1"),super.set(e,t)}};Ou.names=["border-image"];Nw.exports=Ou});var zw=x((W$,jw)=>{u();var Fw=Pe(),ET=j(),Or=class extends ET{prefixed(e,t){let i;return[i,t]=Fw(t),i===2012?t+"flex-line-pack":super.prefixed(e,t)}normalize(){return"align-content"}set(e,t){let i=Fw(t)[0];if(i===2012)return e.value=Or.oldValues[e.value]||e.value,super.set(e,t);if(i==="final")return super.set(e,t)}};Or.names=["align-content","flex-line-pack"];Or.oldValues={"flex-end":"end","flex-start":"start","space-between":"justify","space-around":"distribute"};jw.exports=Or});var Vw=x((G$,Uw)=>{u();var OT=j(),We=class extends OT{prefixed(e,t){return t==="-moz-"?t+(We.toMozilla[e]||e):super.prefixed(e,t)}normalize(e){return We.toNormal[e]||e}};We.names=["border-radius"];We.toMozilla={};We.toNormal={};for(let r of["top","bottom"])for(let e of["left","right"]){let t=`border-${r}-${e}-radius`,i=`border-radius-${r}${e}`;We.names.push(t),We.names.push(i),We.toMozilla[t]=i,We.toNormal[i]=t}Uw.exports=We});var Ww=x((Q$,Hw)=>{u();var TT=j(),Tu=class extends TT{prefixed(e,t){return e.includes("-start")?t+e.replace("-block-start","-before"):t+e.replace("-block-end","-after")}normalize(e){return e.includes("-before")?e.replace("-before","-block-start"):e.replace("-after","-block-end")}};Tu.names=["border-block-start","border-block-end","margin-block-start","margin-block-end","padding-block-start","padding-block-end","border-before","border-after","margin-before","margin-after","padding-before","padding-after"];Hw.exports=Tu});var Qw=x((Y$,Gw)=>{u();var RT=j(),{parseTemplate:PT,warnMissedAreas:IT,getGridGap:DT,warnGridGap:qT,inheritGridGap:$T}=Bt(),Ru=class extends RT{insert(e,t,i,n){if(t!=="-ms-")return super.insert(e,t,i);if(e.parent.some(h=>h.prop==="-ms-grid-rows"))return;let a=DT(e),s=$T(e,a),{rows:o,columns:l,areas:c}=PT({decl:e,gap:s||a}),f=Object.keys(c).length>0,d=Boolean(o),p=Boolean(l);return qT({gap:a,hasColumns:p,decl:e,result:n}),IT(c,e,n),(d&&p||f)&&e.cloneBefore({prop:"-ms-grid-rows",value:o,raws:{}}),p&&e.cloneBefore({prop:"-ms-grid-columns",value:l,raws:{}}),e}};Ru.names=["grid-template"];Gw.exports=Ru});var Kw=x((K$,Yw)=>{u();var LT=j(),Pu=class extends LT{prefixed(e,t){return t+e.replace("-inline","")}normalize(e){return e.replace(/(margin|padding|border)-(start|end)/,"$1-inline-$2")}};Pu.names=["border-inline-start","border-inline-end","margin-inline-start","margin-inline-end","padding-inline-start","padding-inline-end","border-start","border-end","margin-start","margin-end","padding-start","padding-end"];Yw.exports=Pu});var Jw=x((X$,Xw)=>{u();var MT=j(),Iu=class extends MT{check(e){return!e.value.includes("flex-")&&e.value!=="baseline"}prefixed(e,t){return t+"grid-row-align"}normalize(){return"align-self"}};Iu.names=["grid-row-align"];Xw.exports=Iu});var e0=x((J$,Zw)=>{u();var NT=j(),Tr=class extends NT{keyframeParents(e){let{parent:t}=e;for(;t;){if(t.type==="atrule"&&t.name==="keyframes")return!0;({parent:t}=t)}return!1}contain3d(e){if(e.prop==="transform-origin")return!1;for(let t of Tr.functions3d)if(e.value.includes(`${t}(`))return!0;return!1}set(e,t){return e=super.set(e,t),t==="-ms-"&&(e.value=e.value.replace(/rotatez/gi,"rotate")),e}insert(e,t,i){if(t==="-ms-"){if(!this.contain3d(e)&&!this.keyframeParents(e))return super.insert(e,t,i)}else if(t==="-o-"){if(!this.contain3d(e))return super.insert(e,t,i)}else return super.insert(e,t,i)}};Tr.names=["transform","transform-origin"];Tr.functions3d=["matrix3d","translate3d","translateZ","scale3d","scaleZ","rotate3d","rotateX","rotateY","perspective"];Zw.exports=Tr});var i0=x((Z$,r0)=>{u();var t0=Pe(),BT=j(),Du=class extends BT{normalize(){return"flex-direction"}insert(e,t,i){let n;if([n,t]=t0(t),n!==2009)return super.insert(e,t,i);if(e.parent.some(f=>f.prop===t+"box-orient"||f.prop===t+"box-direction"))return;let s=e.value,o,l;s==="inherit"||s==="initial"||s==="unset"?(o=s,l=s):(o=s.includes("row")?"horizontal":"vertical",l=s.includes("reverse")?"reverse":"normal");let c=this.clone(e);return c.prop=t+"box-orient",c.value=o,this.needCascade(e)&&(c.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,c),c=this.clone(e),c.prop=t+"box-direction",c.value=l,this.needCascade(e)&&(c.raws.before=this.calcBefore(i,e,t)),e.parent.insertBefore(e,c)}old(e,t){let i;return[i,t]=t0(t),i===2009?[t+"box-orient",t+"box-direction"]:super.old(e,t)}};Du.names=["flex-direction","box-direction","box-orient"];r0.exports=Du});var s0=x((eL,n0)=>{u();var FT=j(),qu=class extends FT{check(e){return e.value==="pixelated"}prefixed(e,t){return t==="-ms-"?"-ms-interpolation-mode":super.prefixed(e,t)}set(e,t){return t!=="-ms-"?super.set(e,t):(e.prop="-ms-interpolation-mode",e.value="nearest-neighbor",e)}normalize(){return"image-rendering"}process(e,t){return super.process(e,t)}};qu.names=["image-rendering","interpolation-mode"];n0.exports=qu});var o0=x((tL,a0)=>{u();var jT=j(),zT=_e(),$u=class extends jT{constructor(e,t,i){super(e,t,i);this.prefixes&&(this.prefixes=zT.uniq(this.prefixes.map(n=>n==="-ms-"?"-webkit-":n)))}};$u.names=["backdrop-filter"];a0.exports=$u});var u0=x((rL,l0)=>{u();var UT=j(),VT=_e(),Lu=class extends UT{constructor(e,t,i){super(e,t,i);this.prefixes&&(this.prefixes=VT.uniq(this.prefixes.map(n=>n==="-ms-"?"-webkit-":n)))}check(e){return e.value.toLowerCase()==="text"}};Lu.names=["background-clip"];l0.exports=Lu});var c0=x((iL,f0)=>{u();var HT=j(),WT=["none","underline","overline","line-through","blink","inherit","initial","unset"],Mu=class extends HT{check(e){return e.value.split(/\s+/).some(t=>!WT.includes(t))}};Mu.names=["text-decoration"];f0.exports=Mu});var h0=x((nL,d0)=>{u();var p0=Pe(),GT=j(),Rr=class extends GT{prefixed(e,t){let i;return[i,t]=p0(t),i===2009?t+"box-pack":i===2012?t+"flex-pack":super.prefixed(e,t)}normalize(){return"justify-content"}set(e,t){let i=p0(t)[0];if(i===2009||i===2012){let n=Rr.oldValues[e.value]||e.value;if(e.value=n,i!==2009||n!=="distribute")return super.set(e,t)}else if(i==="final")return super.set(e,t)}};Rr.names=["justify-content","flex-pack","box-pack"];Rr.oldValues={"flex-end":"end","flex-start":"start","space-between":"justify","space-around":"distribute"};d0.exports=Rr});var g0=x((sL,m0)=>{u();var QT=j(),Nu=class extends QT{set(e,t){let i=e.value.toLowerCase();return t==="-webkit-"&&!i.includes(" ")&&i!=="contain"&&i!=="cover"&&(e.value=e.value+" "+e.value),super.set(e,t)}};Nu.names=["background-size"];m0.exports=Nu});var b0=x((aL,y0)=>{u();var YT=j(),Bu=Bt(),Fu=class extends YT{insert(e,t,i){if(t!=="-ms-")return super.insert(e,t,i);let n=Bu.parse(e),[a,s]=Bu.translate(n,0,1);n[0]&&n[0].includes("span")&&(s=n[0].join("").replace(/\D/g,"")),[[e.prop,a],[`${e.prop}-span`,s]].forEach(([l,c])=>{Bu.insertDecl(e,l,c)})}};Fu.names=["grid-row","grid-column"];y0.exports=Fu});var x0=x((oL,v0)=>{u();var KT=j(),{prefixTrackProp:w0,prefixTrackValue:XT,autoplaceGridItems:JT,getGridGap:ZT,inheritGridGap:eR}=Bt(),tR=su(),ju=class extends KT{prefixed(e,t){return t==="-ms-"?w0({prop:e,prefix:t}):super.prefixed(e,t)}normalize(e){return e.replace(/^grid-(rows|columns)/,"grid-template-$1")}insert(e,t,i,n){if(t!=="-ms-")return super.insert(e,t,i);let{parent:a,prop:s,value:o}=e,l=s.includes("rows"),c=s.includes("columns"),f=a.some(k=>k.prop==="grid-template"||k.prop==="grid-template-areas");if(f&&l)return!1;let d=new tR({options:{}}),p=d.gridStatus(a,n),h=ZT(e);h=eR(e,h)||h;let b=l?h.row:h.column;(p==="no-autoplace"||p===!0)&&!f&&(b=null);let v=XT({value:o,gap:b});e.cloneBefore({prop:w0({prop:s,prefix:t}),value:v});let y=a.nodes.find(k=>k.prop==="grid-auto-flow"),w="row";if(y&&!d.disabled(y,n)&&(w=y.value.trim()),p==="autoplace"){let k=a.nodes.find(E=>E.prop==="grid-template-rows");if(!k&&f)return;if(!k&&!f){e.warn(n,"Autoplacement does not work without grid-template-rows property");return}!a.nodes.find(E=>E.prop==="grid-template-columns")&&!f&&e.warn(n,"Autoplacement does not work without grid-template-columns property"),c&&!f&&JT(e,n,h,w)}}};ju.names=["grid-template-rows","grid-template-columns","grid-rows","grid-columns"];v0.exports=ju});var S0=x((lL,k0)=>{u();var rR=j(),zu=class extends rR{check(e){return!e.value.includes("flex-")&&e.value!=="baseline"}prefixed(e,t){return t+"grid-column-align"}normalize(){return"justify-self"}};zu.names=["grid-column-align"];k0.exports=zu});var C0=x((uL,A0)=>{u();var iR=j(),Uu=class extends iR{prefixed(e,t){return t+"scroll-chaining"}normalize(){return"overscroll-behavior"}set(e,t){return e.value==="auto"?e.value="chained":(e.value==="none"||e.value==="contain")&&(e.value="none"),super.set(e,t)}};Uu.names=["overscroll-behavior","scroll-chaining"];A0.exports=Uu});var O0=x((fL,E0)=>{u();var nR=j(),{parseGridAreas:sR,warnMissedAreas:aR,prefixTrackProp:oR,prefixTrackValue:_0,getGridGap:lR,warnGridGap:uR,inheritGridGap:fR}=Bt();function cR(r){return r.trim().slice(1,-1).split(/["']\s*["']?/g)}var Vu=class extends nR{insert(e,t,i,n){if(t!=="-ms-")return super.insert(e,t,i);let a=!1,s=!1,o=e.parent,l=lR(e);l=fR(e,l)||l,o.walkDecls(/-ms-grid-rows/,d=>d.remove()),o.walkDecls(/grid-template-(rows|columns)/,d=>{if(d.prop==="grid-template-rows"){s=!0;let{prop:p,value:h}=d;d.cloneBefore({prop:oR({prop:p,prefix:t}),value:_0({value:h,gap:l.row})})}else a=!0});let c=cR(e.value);a&&!s&&l.row&&c.length>1&&e.cloneBefore({prop:"-ms-grid-rows",value:_0({value:`repeat(${c.length}, auto)`,gap:l.row}),raws:{}}),uR({gap:l,hasColumns:a,decl:e,result:n});let f=sR({rows:c,gap:l});return aR(f,e,n),e}};Vu.names=["grid-template-areas"];E0.exports=Vu});var R0=x((cL,T0)=>{u();var pR=j(),Hu=class extends pR{set(e,t){return t==="-webkit-"&&(e.value=e.value.replace(/\s*(right|left)\s*/i,"")),super.set(e,t)}};Hu.names=["text-emphasis-position"];T0.exports=Hu});var I0=x((pL,P0)=>{u();var dR=j(),Wu=class extends dR{set(e,t){return e.prop==="text-decoration-skip-ink"&&e.value==="auto"?(e.prop=t+"text-decoration-skip",e.value="ink",e):super.set(e,t)}};Wu.names=["text-decoration-skip-ink","text-decoration-skip"];P0.exports=Wu});var N0=x((dL,M0)=>{u();"use strict";M0.exports={wrap:D0,limit:q0,validate:$0,test:Gu,curry:hR,name:L0};function D0(r,e,t){var i=e-r;return((t-r)%i+i)%i+r}function q0(r,e,t){return Math.max(r,Math.min(e,t))}function $0(r,e,t,i,n){if(!Gu(r,e,t,i,n))throw new Error(t+" is outside of range ["+r+","+e+")");return t}function Gu(r,e,t,i,n){return!(te||n&&t===e||i&&t===r)}function L0(r,e,t,i){return(t?"(":"[")+r+","+e+(i?")":"]")}function hR(r,e,t,i){var n=L0.bind(null,r,e,t,i);return{wrap:D0.bind(null,r,e),limit:q0.bind(null,r,e),validate:function(a){return $0(r,e,a,t,i)},test:function(a){return Gu(r,e,a,t,i)},toString:n,name:n}}});var j0=x((hL,F0)=>{u();var Qu=$s(),mR=N0(),gR=xr(),yR=He(),bR=_e(),B0=/top|left|right|bottom/gi,wt=class extends yR{replace(e,t){let i=Qu(e);for(let n of i.nodes)if(n.type==="function"&&n.value===this.name)if(n.nodes=this.newDirection(n.nodes),n.nodes=this.normalize(n.nodes),t==="-webkit- old"){if(!this.oldWebkit(n))return!1}else n.nodes=this.convertDirection(n.nodes),n.value=t+n.value;return i.toString()}replaceFirst(e,...t){return t.map(n=>n===" "?{type:"space",value:n}:{type:"word",value:n}).concat(e.slice(1))}normalizeUnit(e,t){return`${parseFloat(e)/t*360}deg`}normalize(e){if(!e[0])return e;if(/-?\d+(.\d+)?grad/.test(e[0].value))e[0].value=this.normalizeUnit(e[0].value,400);else if(/-?\d+(.\d+)?rad/.test(e[0].value))e[0].value=this.normalizeUnit(e[0].value,2*Math.PI);else if(/-?\d+(.\d+)?turn/.test(e[0].value))e[0].value=this.normalizeUnit(e[0].value,1);else if(e[0].value.includes("deg")){let t=parseFloat(e[0].value);t=mR.wrap(0,360,t),e[0].value=`${t}deg`}return e[0].value==="0deg"?e=this.replaceFirst(e,"to"," ","top"):e[0].value==="90deg"?e=this.replaceFirst(e,"to"," ","right"):e[0].value==="180deg"?e=this.replaceFirst(e,"to"," ","bottom"):e[0].value==="270deg"&&(e=this.replaceFirst(e,"to"," ","left")),e}newDirection(e){if(e[0].value==="to"||(B0.lastIndex=0,!B0.test(e[0].value)))return e;e.unshift({type:"word",value:"to"},{type:"space",value:" "});for(let t=2;t0&&(e[0].value==="to"?this.fixDirection(e):e[0].value.includes("deg")?this.fixAngle(e):this.isRadial(e)&&this.fixRadial(e)),e}fixDirection(e){e.splice(0,2);for(let t of e){if(t.type==="div")break;t.type==="word"&&(t.value=this.revertDirection(t.value))}}fixAngle(e){let t=e[0].value;t=parseFloat(t),t=Math.abs(450-t)%360,t=this.roundFloat(t,3),e[0].value=`${t}deg`}fixRadial(e){let t=[],i=[],n,a,s,o,l;for(o=0;o{u();var wR=xr(),vR=He();function z0(r){return new RegExp(`(^|[\\s,(])(${r}($|[\\s),]))`,"gi")}var Yu=class extends vR{regexp(){return this.regexpCache||(this.regexpCache=z0(this.name)),this.regexpCache}isStretch(){return this.name==="stretch"||this.name==="fill"||this.name==="fill-available"}replace(e,t){return t==="-moz-"&&this.isStretch()?e.replace(this.regexp(),"$1-moz-available$3"):t==="-webkit-"&&this.isStretch()?e.replace(this.regexp(),"$1-webkit-fill-available$3"):super.replace(e,t)}old(e){let t=e+this.name;return this.isStretch()&&(e==="-moz-"?t="-moz-available":e==="-webkit-"&&(t="-webkit-fill-available")),new wR(this.name,t,t,z0(t))}add(e,t){if(!(e.prop.includes("grid")&&t!=="-webkit-"))return super.add(e,t)}};Yu.names=["max-content","min-content","fit-content","fill","fill-available","stretch"];U0.exports=Yu});var G0=x((gL,W0)=>{u();var H0=xr(),xR=He(),Ku=class extends xR{replace(e,t){return t==="-webkit-"?e.replace(this.regexp(),"$1-webkit-optimize-contrast"):t==="-moz-"?e.replace(this.regexp(),"$1-moz-crisp-edges"):super.replace(e,t)}old(e){return e==="-webkit-"?new H0(this.name,"-webkit-optimize-contrast"):e==="-moz-"?new H0(this.name,"-moz-crisp-edges"):super.old(e)}};Ku.names=["pixelated"];W0.exports=Ku});var Y0=x((yL,Q0)=>{u();var kR=He(),Xu=class extends kR{replace(e,t){let i=super.replace(e,t);return t==="-webkit-"&&(i=i.replace(/("[^"]+"|'[^']+')(\s+\d+\w)/gi,"url($1)$2")),i}};Xu.names=["image-set"];Q0.exports=Xu});var X0=x((bL,K0)=>{u();var SR=$e().list,AR=He(),Ju=class extends AR{replace(e,t){return SR.space(e).map(i=>{if(i.slice(0,+this.name.length+1)!==this.name+"(")return i;let n=i.lastIndexOf(")"),a=i.slice(n+1),s=i.slice(this.name.length+1,n);if(t==="-webkit-"){let o=s.match(/\d*.?\d+%?/);o?(s=s.slice(o[0].length).trim(),s+=`, ${o[0]}`):s+=", 0.5"}return t+this.name+"("+s+")"+a}).join(" ")}};Ju.names=["cross-fade"];K0.exports=Ju});var Z0=x((wL,J0)=>{u();var CR=Pe(),_R=xr(),ER=He(),Zu=class extends ER{constructor(e,t){super(e,t);e==="display-flex"&&(this.name="flex")}check(e){return e.prop==="display"&&e.value===this.name}prefixed(e){let t,i;return[t,e]=CR(e),t===2009?this.name==="flex"?i="box":i="inline-box":t===2012?this.name==="flex"?i="flexbox":i="inline-flexbox":t==="final"&&(i=this.name),e+i}replace(e,t){return this.prefixed(t)}old(e){let t=this.prefixed(e);if(!!t)return new _R(this.name,t)}};Zu.names=["display-flex","inline-flex"];J0.exports=Zu});var tv=x((vL,ev)=>{u();var OR=He(),ef=class extends OR{constructor(e,t){super(e,t);e==="display-grid"&&(this.name="grid")}check(e){return e.prop==="display"&&e.value===this.name}};ef.names=["display-grid","inline-grid"];ev.exports=ef});var iv=x((xL,rv)=>{u();var TR=He(),tf=class extends TR{constructor(e,t){super(e,t);e==="filter-function"&&(this.name="filter")}};tf.names=["filter","filter-function"];rv.exports=tf});var ov=x((kL,av)=>{u();var nv=Li(),z=j(),sv=Fy(),RR=nb(),PR=su(),IR=Sb(),rf=Mt(),Pr=kr(),DR=Pb(),ut=He(),Ir=_e(),qR=Db(),$R=$b(),LR=Mb(),MR=Bb(),NR=Vb(),BR=Gb(),FR=Yb(),jR=Xb(),zR=Zb(),UR=tw(),VR=iw(),HR=sw(),WR=ow(),GR=uw(),QR=cw(),YR=hw(),KR=gw(),XR=ww(),JR=xw(),ZR=Sw(),e5=_w(),t5=Ow(),r5=Pw(),i5=Dw(),n5=$w(),s5=Mw(),a5=Bw(),o5=zw(),l5=Vw(),u5=Ww(),f5=Qw(),c5=Kw(),p5=Jw(),d5=e0(),h5=i0(),m5=s0(),g5=o0(),y5=u0(),b5=c0(),w5=h0(),v5=g0(),x5=b0(),k5=x0(),S5=S0(),A5=C0(),C5=O0(),_5=R0(),E5=I0(),O5=j0(),T5=V0(),R5=G0(),P5=Y0(),I5=X0(),D5=Z0(),q5=tv(),$5=iv();Pr.hack(qR);Pr.hack($R);Pr.hack(LR);Pr.hack(MR);z.hack(NR);z.hack(BR);z.hack(FR);z.hack(jR);z.hack(zR);z.hack(UR);z.hack(VR);z.hack(HR);z.hack(WR);z.hack(GR);z.hack(QR);z.hack(YR);z.hack(KR);z.hack(XR);z.hack(JR);z.hack(ZR);z.hack(e5);z.hack(t5);z.hack(r5);z.hack(i5);z.hack(n5);z.hack(s5);z.hack(a5);z.hack(o5);z.hack(l5);z.hack(u5);z.hack(f5);z.hack(c5);z.hack(p5);z.hack(d5);z.hack(h5);z.hack(m5);z.hack(g5);z.hack(y5);z.hack(b5);z.hack(w5);z.hack(v5);z.hack(x5);z.hack(k5);z.hack(S5);z.hack(A5);z.hack(C5);z.hack(_5);z.hack(E5);ut.hack(O5);ut.hack(T5);ut.hack(R5);ut.hack(P5);ut.hack(I5);ut.hack(D5);ut.hack(q5);ut.hack($5);var nf=new Map,Ni=class{constructor(e,t,i={}){this.data=e,this.browsers=t,this.options=i,[this.add,this.remove]=this.preprocess(this.select(this.data)),this.transition=new RR(this),this.processor=new PR(this)}cleaner(){if(this.cleanerCache)return this.cleanerCache;if(this.browsers.selected.length){let e=new rf(this.browsers.data,[]);this.cleanerCache=new Ni(this.data,e,this.options)}else return this;return this.cleanerCache}select(e){let t={add:{},remove:{}};for(let i in e){let n=e[i],a=n.browsers.map(l=>{let c=l.split(" ");return{browser:`${c[0]} ${c[1]}`,note:c[2]}}),s=a.filter(l=>l.note).map(l=>`${this.browsers.prefix(l.browser)} ${l.note}`);s=Ir.uniq(s),a=a.filter(l=>this.browsers.isSelected(l.browser)).map(l=>{let c=this.browsers.prefix(l.browser);return l.note?`${c} ${l.note}`:c}),a=this.sort(Ir.uniq(a)),this.options.flexbox==="no-2009"&&(a=a.filter(l=>!l.includes("2009")));let o=n.browsers.map(l=>this.browsers.prefix(l));n.mistakes&&(o=o.concat(n.mistakes)),o=o.concat(s),o=Ir.uniq(o),a.length?(t.add[i]=a,a.length!a.includes(l)))):t.remove[i]=o}return t}sort(e){return e.sort((t,i)=>{let n=Ir.removeNote(t).length,a=Ir.removeNote(i).length;return n===a?i.length-t.length:a-n})}preprocess(e){let t={selectors:[],"@supports":new IR(Ni,this)};for(let n in e.add){let a=e.add[n];if(n==="@keyframes"||n==="@viewport")t[n]=new DR(n,a,this);else if(n==="@resolution")t[n]=new sv(n,a,this);else if(this.data[n].selector)t.selectors.push(Pr.load(n,a,this));else{let s=this.data[n].props;if(s){let o=ut.load(n,a,this);for(let l of s)t[l]||(t[l]={values:[]}),t[l].values.push(o)}else{let o=t[n]&&t[n].values||[];t[n]=z.load(n,a,this),t[n].values=o}}}let i={selectors:[]};for(let n in e.remove){let a=e.remove[n];if(this.data[n].selector){let s=Pr.load(n,a);for(let o of a)i.selectors.push(s.old(o))}else if(n==="@keyframes"||n==="@viewport")for(let s of a){let o=`@${s}${n.slice(1)}`;i[o]={remove:!0}}else if(n==="@resolution")i[n]=new sv(n,a,this);else{let s=this.data[n].props;if(s){let o=ut.load(n,[],this);for(let l of a){let c=o.old(l);if(c)for(let f of s)i[f]||(i[f]={}),i[f].values||(i[f].values=[]),i[f].values.push(c)}}else for(let o of a){let l=this.decl(n).old(n,o);if(n==="align-self"){let c=t[n]&&t[n].prefixes;if(c){if(o==="-webkit- 2009"&&c.includes("-webkit-"))continue;if(o==="-webkit-"&&c.includes("-webkit- 2009"))continue}}for(let c of l)i[c]||(i[c]={}),i[c].remove=!0}}}return[t,i]}decl(e){return nf.has(e)||nf.set(e,z.load(e)),nf.get(e)}unprefixed(e){let t=this.normalize(nv.unprefixed(e));return t==="flex-direction"&&(t="flex-flow"),t}normalize(e){return this.decl(e).normalize(e)}prefixed(e,t){return e=nv.unprefixed(e),this.decl(e).prefixed(e,t)}values(e,t){let i=this[e],n=i["*"]&&i["*"].values,a=i[t]&&i[t].values;return n&&a?Ir.uniq(n.concat(a)):n||a||[]}group(e){let t=e.parent,i=t.index(e),{length:n}=t.nodes,a=this.unprefixed(e.prop),s=(o,l)=>{for(i+=o;i>=0&&i{u();lv.exports={"backdrop-filter":{feature:"css-backdrop-filter",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5","safari 16.5"]},element:{props:["background","background-image","border-image","mask","list-style","list-style-image","content","mask-image"],feature:"css-element-function",browsers:["firefox 114"]},"user-select":{mistakes:["-khtml-"],feature:"user-select-none",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5","safari 16.5"]},"background-clip":{feature:"background-clip-text",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},hyphens:{feature:"css-hyphens",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5","safari 16.5"]},fill:{props:["width","min-width","max-width","height","min-height","max-height","inline-size","min-inline-size","max-inline-size","block-size","min-block-size","max-block-size","grid","grid-template","grid-template-rows","grid-template-columns","grid-auto-columns","grid-auto-rows"],feature:"intrinsic-width",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"fill-available":{props:["width","min-width","max-width","height","min-height","max-height","inline-size","min-inline-size","max-inline-size","block-size","min-block-size","max-block-size","grid","grid-template","grid-template-rows","grid-template-columns","grid-auto-columns","grid-auto-rows"],feature:"intrinsic-width",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},stretch:{props:["width","min-width","max-width","height","min-height","max-height","inline-size","min-inline-size","max-inline-size","block-size","min-block-size","max-block-size","grid","grid-template","grid-template-rows","grid-template-columns","grid-auto-columns","grid-auto-rows"],feature:"intrinsic-width",browsers:["firefox 114"]},"fit-content":{props:["width","min-width","max-width","height","min-height","max-height","inline-size","min-inline-size","max-inline-size","block-size","min-block-size","max-block-size","grid","grid-template","grid-template-rows","grid-template-columns","grid-auto-columns","grid-auto-rows"],feature:"intrinsic-width",browsers:["firefox 114"]},"text-decoration-style":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-decoration-color":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-decoration-line":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-decoration":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-decoration-skip":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-decoration-skip-ink":{feature:"text-decoration",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"text-size-adjust":{feature:"text-size-adjust",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5"]},"mask-clip":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-composite":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-image":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-origin":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-repeat":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border-repeat":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border-source":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},mask:{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-position":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-size":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border-outset":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border-width":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"mask-border-slice":{feature:"css-masks",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},"clip-path":{feature:"css-clip-path",browsers:["samsung 21"]},"box-decoration-break":{feature:"css-boxdecorationbreak",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5","opera 99","safari 16.5","samsung 21"]},appearance:{feature:"css-appearance",browsers:["samsung 21"]},"image-set":{props:["background","background-image","border-image","cursor","mask","mask-image","list-style","list-style-image","content"],feature:"css-image-set",browsers:["and_uc 15.5","chrome 109","samsung 21"]},"cross-fade":{props:["background","background-image","border-image","mask","list-style","list-style-image","content","mask-image"],feature:"css-cross-fade",browsers:["and_chr 114","and_uc 15.5","chrome 109","chrome 113","chrome 114","edge 114","opera 99","samsung 21"]},isolate:{props:["unicode-bidi"],feature:"css-unicode-bidi",browsers:["ios_saf 16.1","ios_saf 16.3","ios_saf 16.4","ios_saf 16.5","safari 16.5"]},"color-adjust":{feature:"css-color-adjust",browsers:["chrome 109","chrome 113","chrome 114","edge 114","opera 99"]}}});var cv=x((AL,fv)=>{u();fv.exports={}});var mv=x((CL,hv)=>{u();var L5=Gl(),{agents:M5}=(Ts(),Os),sf=_y(),N5=Mt(),B5=ov(),F5=uv(),j5=cv(),pv={browsers:M5,prefixes:F5},dv=` - Replace Autoprefixer \`browsers\` option to Browserslist config. - Use \`browserslist\` key in \`package.json\` or \`.browserslistrc\` file. - - Using \`browsers\` option can cause errors. Browserslist config can - be used for Babel, Autoprefixer, postcss-normalize and other tools. - - If you really need to use option, rename it to \`overrideBrowserslist\`. - - Learn more at: - https://github.com/browserslist/browserslist#readme - https://twitter.com/browserslist - -`;function z5(r){return Object.prototype.toString.apply(r)==="[object Object]"}var af=new Map;function U5(r,e){e.browsers.selected.length!==0&&(e.add.selectors.length>0||Object.keys(e.add).length>2||r.warn(`Autoprefixer target browsers do not need any prefixes.You do not need Autoprefixer anymore. -Check your Browserslist config to be sure that your targets are set up correctly. - - Learn more at: - https://github.com/postcss/autoprefixer#readme - https://github.com/browserslist/browserslist#readme - -`))}hv.exports=Dr;function Dr(...r){let e;if(r.length===1&&z5(r[0])?(e=r[0],r=void 0):r.length===0||r.length===1&&!r[0]?r=void 0:r.length<=2&&(Array.isArray(r[0])||!r[0])?(e=r[1],r=r[0]):typeof r[r.length-1]=="object"&&(e=r.pop()),e||(e={}),e.browser)throw new Error("Change `browser` option to `overrideBrowserslist` in Autoprefixer");if(e.browserslist)throw new Error("Change `browserslist` option to `overrideBrowserslist` in Autoprefixer");e.overrideBrowserslist?r=e.overrideBrowserslist:e.browsers&&(typeof console!="undefined"&&console.warn&&(sf.red?console.warn(sf.red(dv.replace(/`[^`]+`/g,n=>sf.yellow(n.slice(1,-1))))):console.warn(dv)),r=e.browsers);let t={ignoreUnknownVersions:e.ignoreUnknownVersions,stats:e.stats,env:e.env};function i(n){let a=pv,s=new N5(a.browsers,r,n,t),o=s.selected.join(", ")+JSON.stringify(e);return af.has(o)||af.set(o,new B5(a.prefixes,s,e)),af.get(o)}return{postcssPlugin:"autoprefixer",prepare(n){let a=i({from:n.opts.from,env:e.env});return{OnceExit(s){U5(n,a),e.remove!==!1&&a.processor.remove(s,n),e.add!==!1&&a.processor.add(s,n)}}},info(n){return n=n||{},n.from=n.from||m.cwd(),j5(i(n))},options:e,browsers:r}}Dr.postcss=!0;Dr.data=pv;Dr.defaults=L5.defaults;Dr.info=()=>Dr().info()});var gv={};Ge(gv,{default:()=>V5});var V5,yv=R(()=>{u();V5=[]});var wv={};Ge(wv,{default:()=>H5});var bv,H5,vv=R(()=>{u();Yi();bv=pe(en()),H5=St(bv.default.theme)});var kv={};Ge(kv,{default:()=>W5});var xv,W5,Sv=R(()=>{u();Yi();xv=pe(en()),W5=St(xv.default)});u();"use strict";var G5=vt(Ay()),Q5=vt($e()),Y5=vt(mv()),K5=vt((yv(),gv)),X5=vt((vv(),wv)),J5=vt((Sv(),kv)),Z5=vt((zs(),Af)),eP=vt((nl(),il)),tP=vt((ia(),ic));function vt(r){return r&&r.__esModule?r:{default:r}}console.warn("cdn.tailwindcss.com should not be used in production. To use Tailwind CSS in production, install it as a PostCSS plugin or use the Tailwind CLI: https://tailwindcss.com/docs/installation");var Ls="tailwind",of="text/tailwindcss",Av="/template.html",Yt,Cv=!0,_v=0,lf=new Set,uf,Ev="",Ov=(r=!1)=>({get(e,t){return(!r||t==="config")&&typeof e[t]=="object"&&e[t]!==null?new Proxy(e[t],Ov()):e[t]},set(e,t,i){return e[t]=i,(!r||t==="config")&&ff(!0),!0}});window[Ls]=new Proxy({config:{},defaultTheme:X5.default,defaultConfig:J5.default,colors:Z5.default,plugin:eP.default,resolveConfig:tP.default},Ov(!0));function Tv(r){uf.observe(r,{attributes:!0,attributeFilter:["type"],characterData:!0,subtree:!0,childList:!0})}new MutationObserver(async r=>{let e=!1;if(!uf){uf=new MutationObserver(async()=>await ff(!0));for(let t of document.querySelectorAll(`style[type="${of}"]`))Tv(t)}for(let t of r)for(let i of t.addedNodes)i.nodeType===1&&i.tagName==="STYLE"&&i.getAttribute("type")===of&&(Tv(i),e=!0);await ff(e)}).observe(document.documentElement,{attributes:!0,attributeFilter:["class"],childList:!0,subtree:!0});async function ff(r=!1){r&&(_v++,lf.clear());let e="";for(let i of document.querySelectorAll(`style[type="${of}"]`))e+=i.textContent;let t=new Set;for(let i of document.querySelectorAll("[class]"))for(let n of i.classList)lf.has(n)||t.add(n);if(document.body&&(Cv||t.size>0||e!==Ev||!Yt||!Yt.isConnected)){for(let n of t)lf.add(n);Cv=!1,Ev=e,self[Av]=Array.from(t).join(" ");let{css:i}=await(0,Q5.default)([(0,G5.default)({...window[Ls].config,_hash:_v,content:{files:[Av],extract:{html:n=>n.split(" ")}},plugins:[...K5.default,...Array.isArray(window[Ls].config.plugins)?window[Ls].config.plugins:[]]}),(0,Y5.default)({remove:!1})]).process(`@tailwind base;@tailwind components;@tailwind utilities;${e}`);(!Yt||!Yt.isConnected)&&(Yt=document.createElement("style"),document.head.append(Yt)),Yt.textContent=i}}})(); -/*! - * fill-range - * - * Copyright (c) 2014-present, Jon Schlinkert. - * Licensed under the MIT License. - */ -/*! - * is-number - * - * Copyright (c) 2014-present, Jon Schlinkert. - * Released under the MIT License. - */ -/*! - * to-regex-range - * - * Copyright (c) 2015-present, Jon Schlinkert. - * Released under the MIT License. - */ -/*! https://mths.be/cssesc v3.0.0 by @mathias */ - diff --git a/examples/server/public/deps_vue.esm-browser.js b/examples/server/public/deps_vue.esm-browser.js deleted file mode 100644 index 4679d9614..000000000 --- a/examples/server/public/deps_vue.esm-browser.js +++ /dev/null @@ -1,18160 +0,0 @@ -/** -* vue v3.5.12 -* (c) 2018-present Yuxi (Evan) You and Vue contributors -* @license MIT -**/ -/*! #__NO_SIDE_EFFECTS__ */ -// @__NO_SIDE_EFFECTS__ -function makeMap(str) { - const map = /* @__PURE__ */ Object.create(null); - for (const key of str.split(",")) map[key] = 1; - return (val) => val in map; -} - -const EMPTY_OBJ = Object.freeze({}) ; -const EMPTY_ARR = Object.freeze([]) ; -const NOOP = () => { -}; -const NO = () => false; -const isOn = (key) => key.charCodeAt(0) === 111 && key.charCodeAt(1) === 110 && // uppercase letter -(key.charCodeAt(2) > 122 || key.charCodeAt(2) < 97); -const isModelListener = (key) => key.startsWith("onUpdate:"); -const extend = Object.assign; -const remove = (arr, el) => { - const i = arr.indexOf(el); - if (i > -1) { - arr.splice(i, 1); - } -}; -const hasOwnProperty$1 = Object.prototype.hasOwnProperty; -const hasOwn = (val, key) => hasOwnProperty$1.call(val, key); -const isArray = Array.isArray; -const isMap = (val) => toTypeString(val) === "[object Map]"; -const isSet = (val) => toTypeString(val) === "[object Set]"; -const isDate = (val) => toTypeString(val) === "[object Date]"; -const isRegExp = (val) => toTypeString(val) === "[object RegExp]"; -const isFunction = (val) => typeof val === "function"; -const isString = (val) => typeof val === "string"; -const isSymbol = (val) => typeof val === "symbol"; -const isObject = (val) => val !== null && typeof val === "object"; -const isPromise = (val) => { - return (isObject(val) || isFunction(val)) && isFunction(val.then) && isFunction(val.catch); -}; -const objectToString = Object.prototype.toString; -const toTypeString = (value) => objectToString.call(value); -const toRawType = (value) => { - return toTypeString(value).slice(8, -1); -}; -const isPlainObject = (val) => toTypeString(val) === "[object Object]"; -const isIntegerKey = (key) => isString(key) && key !== "NaN" && key[0] !== "-" && "" + parseInt(key, 10) === key; -const isReservedProp = /* @__PURE__ */ makeMap( - // the leading comma is intentional so empty string "" is also included - ",key,ref,ref_for,ref_key,onVnodeBeforeMount,onVnodeMounted,onVnodeBeforeUpdate,onVnodeUpdated,onVnodeBeforeUnmount,onVnodeUnmounted" -); -const isBuiltInDirective = /* @__PURE__ */ makeMap( - "bind,cloak,else-if,else,for,html,if,model,on,once,pre,show,slot,text,memo" -); -const cacheStringFunction = (fn) => { - const cache = /* @__PURE__ */ Object.create(null); - return (str) => { - const hit = cache[str]; - return hit || (cache[str] = fn(str)); - }; -}; -const camelizeRE = /-(\w)/g; -const camelize = cacheStringFunction( - (str) => { - return str.replace(camelizeRE, (_, c) => c ? c.toUpperCase() : ""); - } -); -const hyphenateRE = /\B([A-Z])/g; -const hyphenate = cacheStringFunction( - (str) => str.replace(hyphenateRE, "-$1").toLowerCase() -); -const capitalize = cacheStringFunction((str) => { - return str.charAt(0).toUpperCase() + str.slice(1); -}); -const toHandlerKey = cacheStringFunction( - (str) => { - const s = str ? `on${capitalize(str)}` : ``; - return s; - } -); -const hasChanged = (value, oldValue) => !Object.is(value, oldValue); -const invokeArrayFns = (fns, ...arg) => { - for (let i = 0; i < fns.length; i++) { - fns[i](...arg); - } -}; -const def = (obj, key, value, writable = false) => { - Object.defineProperty(obj, key, { - configurable: true, - enumerable: false, - writable, - value - }); -}; -const looseToNumber = (val) => { - const n = parseFloat(val); - return isNaN(n) ? val : n; -}; -const toNumber = (val) => { - const n = isString(val) ? Number(val) : NaN; - return isNaN(n) ? val : n; -}; -let _globalThis; -const getGlobalThis = () => { - return _globalThis || (_globalThis = typeof globalThis !== "undefined" ? globalThis : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : typeof global !== "undefined" ? global : {}); -}; -function genCacheKey(source, options) { - return source + JSON.stringify( - options, - (_, val) => typeof val === "function" ? val.toString() : val - ); -} - -const PatchFlagNames = { - [1]: `TEXT`, - [2]: `CLASS`, - [4]: `STYLE`, - [8]: `PROPS`, - [16]: `FULL_PROPS`, - [32]: `NEED_HYDRATION`, - [64]: `STABLE_FRAGMENT`, - [128]: `KEYED_FRAGMENT`, - [256]: `UNKEYED_FRAGMENT`, - [512]: `NEED_PATCH`, - [1024]: `DYNAMIC_SLOTS`, - [2048]: `DEV_ROOT_FRAGMENT`, - [-1]: `HOISTED`, - [-2]: `BAIL` -}; - -const slotFlagsText = { - [1]: "STABLE", - [2]: "DYNAMIC", - [3]: "FORWARDED" -}; - -const GLOBALS_ALLOWED = "Infinity,undefined,NaN,isFinite,isNaN,parseFloat,parseInt,decodeURI,decodeURIComponent,encodeURI,encodeURIComponent,Math,Number,Date,Array,Object,Boolean,String,RegExp,Map,Set,JSON,Intl,BigInt,console,Error,Symbol"; -const isGloballyAllowed = /* @__PURE__ */ makeMap(GLOBALS_ALLOWED); - -const range = 2; -function generateCodeFrame(source, start = 0, end = source.length) { - start = Math.max(0, Math.min(start, source.length)); - end = Math.max(0, Math.min(end, source.length)); - if (start > end) return ""; - let lines = source.split(/(\r?\n)/); - const newlineSequences = lines.filter((_, idx) => idx % 2 === 1); - lines = lines.filter((_, idx) => idx % 2 === 0); - let count = 0; - const res = []; - for (let i = 0; i < lines.length; i++) { - count += lines[i].length + (newlineSequences[i] && newlineSequences[i].length || 0); - if (count >= start) { - for (let j = i - range; j <= i + range || end > count; j++) { - if (j < 0 || j >= lines.length) continue; - const line = j + 1; - res.push( - `${line}${" ".repeat(Math.max(3 - String(line).length, 0))}| ${lines[j]}` - ); - const lineLength = lines[j].length; - const newLineSeqLength = newlineSequences[j] && newlineSequences[j].length || 0; - if (j === i) { - const pad = start - (count - (lineLength + newLineSeqLength)); - const length = Math.max( - 1, - end > count ? lineLength - pad : end - start - ); - res.push(` | ` + " ".repeat(pad) + "^".repeat(length)); - } else if (j > i) { - if (end > count) { - const length = Math.max(Math.min(end - count, lineLength), 1); - res.push(` | ` + "^".repeat(length)); - } - count += lineLength + newLineSeqLength; - } - } - break; - } - } - return res.join("\n"); -} - -function normalizeStyle(value) { - if (isArray(value)) { - const res = {}; - for (let i = 0; i < value.length; i++) { - const item = value[i]; - const normalized = isString(item) ? parseStringStyle(item) : normalizeStyle(item); - if (normalized) { - for (const key in normalized) { - res[key] = normalized[key]; - } - } - } - return res; - } else if (isString(value) || isObject(value)) { - return value; - } -} -const listDelimiterRE = /;(?![^(]*\))/g; -const propertyDelimiterRE = /:([^]+)/; -const styleCommentRE = /\/\*[^]*?\*\//g; -function parseStringStyle(cssText) { - const ret = {}; - cssText.replace(styleCommentRE, "").split(listDelimiterRE).forEach((item) => { - if (item) { - const tmp = item.split(propertyDelimiterRE); - tmp.length > 1 && (ret[tmp[0].trim()] = tmp[1].trim()); - } - }); - return ret; -} -function stringifyStyle(styles) { - let ret = ""; - if (!styles || isString(styles)) { - return ret; - } - for (const key in styles) { - const value = styles[key]; - if (isString(value) || typeof value === "number") { - const normalizedKey = key.startsWith(`--`) ? key : hyphenate(key); - ret += `${normalizedKey}:${value};`; - } - } - return ret; -} -function normalizeClass(value) { - let res = ""; - if (isString(value)) { - res = value; - } else if (isArray(value)) { - for (let i = 0; i < value.length; i++) { - const normalized = normalizeClass(value[i]); - if (normalized) { - res += normalized + " "; - } - } - } else if (isObject(value)) { - for (const name in value) { - if (value[name]) { - res += name + " "; - } - } - } - return res.trim(); -} -function normalizeProps(props) { - if (!props) return null; - let { class: klass, style } = props; - if (klass && !isString(klass)) { - props.class = normalizeClass(klass); - } - if (style) { - props.style = normalizeStyle(style); - } - return props; -} - -const HTML_TAGS = "html,body,base,head,link,meta,style,title,address,article,aside,footer,header,hgroup,h1,h2,h3,h4,h5,h6,nav,section,div,dd,dl,dt,figcaption,figure,picture,hr,img,li,main,ol,p,pre,ul,a,b,abbr,bdi,bdo,br,cite,code,data,dfn,em,i,kbd,mark,q,rp,rt,ruby,s,samp,small,span,strong,sub,sup,time,u,var,wbr,area,audio,map,track,video,embed,object,param,source,canvas,script,noscript,del,ins,caption,col,colgroup,table,thead,tbody,td,th,tr,button,datalist,fieldset,form,input,label,legend,meter,optgroup,option,output,progress,select,textarea,details,dialog,menu,summary,template,blockquote,iframe,tfoot"; -const SVG_TAGS = "svg,animate,animateMotion,animateTransform,circle,clipPath,color-profile,defs,desc,discard,ellipse,feBlend,feColorMatrix,feComponentTransfer,feComposite,feConvolveMatrix,feDiffuseLighting,feDisplacementMap,feDistantLight,feDropShadow,feFlood,feFuncA,feFuncB,feFuncG,feFuncR,feGaussianBlur,feImage,feMerge,feMergeNode,feMorphology,feOffset,fePointLight,feSpecularLighting,feSpotLight,feTile,feTurbulence,filter,foreignObject,g,hatch,hatchpath,image,line,linearGradient,marker,mask,mesh,meshgradient,meshpatch,meshrow,metadata,mpath,path,pattern,polygon,polyline,radialGradient,rect,set,solidcolor,stop,switch,symbol,text,textPath,title,tspan,unknown,use,view"; -const MATH_TAGS = "annotation,annotation-xml,maction,maligngroup,malignmark,math,menclose,merror,mfenced,mfrac,mfraction,mglyph,mi,mlabeledtr,mlongdiv,mmultiscripts,mn,mo,mover,mpadded,mphantom,mprescripts,mroot,mrow,ms,mscarries,mscarry,msgroup,msline,mspace,msqrt,msrow,mstack,mstyle,msub,msubsup,msup,mtable,mtd,mtext,mtr,munder,munderover,none,semantics"; -const VOID_TAGS = "area,base,br,col,embed,hr,img,input,link,meta,param,source,track,wbr"; -const isHTMLTag = /* @__PURE__ */ makeMap(HTML_TAGS); -const isSVGTag = /* @__PURE__ */ makeMap(SVG_TAGS); -const isMathMLTag = /* @__PURE__ */ makeMap(MATH_TAGS); -const isVoidTag = /* @__PURE__ */ makeMap(VOID_TAGS); - -const specialBooleanAttrs = `itemscope,allowfullscreen,formnovalidate,ismap,nomodule,novalidate,readonly`; -const isSpecialBooleanAttr = /* @__PURE__ */ makeMap(specialBooleanAttrs); -const isBooleanAttr = /* @__PURE__ */ makeMap( - specialBooleanAttrs + `,async,autofocus,autoplay,controls,default,defer,disabled,hidden,inert,loop,open,required,reversed,scoped,seamless,checked,muted,multiple,selected` -); -function includeBooleanAttr(value) { - return !!value || value === ""; -} -const isKnownHtmlAttr = /* @__PURE__ */ makeMap( - `accept,accept-charset,accesskey,action,align,allow,alt,async,autocapitalize,autocomplete,autofocus,autoplay,background,bgcolor,border,buffered,capture,challenge,charset,checked,cite,class,code,codebase,color,cols,colspan,content,contenteditable,contextmenu,controls,coords,crossorigin,csp,data,datetime,decoding,default,defer,dir,dirname,disabled,download,draggable,dropzone,enctype,enterkeyhint,for,form,formaction,formenctype,formmethod,formnovalidate,formtarget,headers,height,hidden,high,href,hreflang,http-equiv,icon,id,importance,inert,integrity,ismap,itemprop,keytype,kind,label,lang,language,loading,list,loop,low,manifest,max,maxlength,minlength,media,min,multiple,muted,name,novalidate,open,optimum,pattern,ping,placeholder,poster,preload,radiogroup,readonly,referrerpolicy,rel,required,reversed,rows,rowspan,sandbox,scope,scoped,selected,shape,size,sizes,slot,span,spellcheck,src,srcdoc,srclang,srcset,start,step,style,summary,tabindex,target,title,translate,type,usemap,value,width,wrap` -); -const isKnownSvgAttr = /* @__PURE__ */ makeMap( - `xmlns,accent-height,accumulate,additive,alignment-baseline,alphabetic,amplitude,arabic-form,ascent,attributeName,attributeType,azimuth,baseFrequency,baseline-shift,baseProfile,bbox,begin,bias,by,calcMode,cap-height,class,clip,clipPathUnits,clip-path,clip-rule,color,color-interpolation,color-interpolation-filters,color-profile,color-rendering,contentScriptType,contentStyleType,crossorigin,cursor,cx,cy,d,decelerate,descent,diffuseConstant,direction,display,divisor,dominant-baseline,dur,dx,dy,edgeMode,elevation,enable-background,end,exponent,fill,fill-opacity,fill-rule,filter,filterRes,filterUnits,flood-color,flood-opacity,font-family,font-size,font-size-adjust,font-stretch,font-style,font-variant,font-weight,format,from,fr,fx,fy,g1,g2,glyph-name,glyph-orientation-horizontal,glyph-orientation-vertical,glyphRef,gradientTransform,gradientUnits,hanging,height,href,hreflang,horiz-adv-x,horiz-origin-x,id,ideographic,image-rendering,in,in2,intercept,k,k1,k2,k3,k4,kernelMatrix,kernelUnitLength,kerning,keyPoints,keySplines,keyTimes,lang,lengthAdjust,letter-spacing,lighting-color,limitingConeAngle,local,marker-end,marker-mid,marker-start,markerHeight,markerUnits,markerWidth,mask,maskContentUnits,maskUnits,mathematical,max,media,method,min,mode,name,numOctaves,offset,opacity,operator,order,orient,orientation,origin,overflow,overline-position,overline-thickness,panose-1,paint-order,path,pathLength,patternContentUnits,patternTransform,patternUnits,ping,pointer-events,points,pointsAtX,pointsAtY,pointsAtZ,preserveAlpha,preserveAspectRatio,primitiveUnits,r,radius,referrerPolicy,refX,refY,rel,rendering-intent,repeatCount,repeatDur,requiredExtensions,requiredFeatures,restart,result,rotate,rx,ry,scale,seed,shape-rendering,slope,spacing,specularConstant,specularExponent,speed,spreadMethod,startOffset,stdDeviation,stemh,stemv,stitchTiles,stop-color,stop-opacity,strikethrough-position,strikethrough-thickness,string,stroke,stroke-dasharray,stroke-dashoffset,stroke-linecap,stroke-linejoin,stroke-miterlimit,stroke-opacity,stroke-width,style,surfaceScale,systemLanguage,tabindex,tableValues,target,targetX,targetY,text-anchor,text-decoration,text-rendering,textLength,to,transform,transform-origin,type,u1,u2,underline-position,underline-thickness,unicode,unicode-bidi,unicode-range,units-per-em,v-alphabetic,v-hanging,v-ideographic,v-mathematical,values,vector-effect,version,vert-adv-y,vert-origin-x,vert-origin-y,viewBox,viewTarget,visibility,width,widths,word-spacing,writing-mode,x,x-height,x1,x2,xChannelSelector,xlink:actuate,xlink:arcrole,xlink:href,xlink:role,xlink:show,xlink:title,xlink:type,xmlns:xlink,xml:base,xml:lang,xml:space,y,y1,y2,yChannelSelector,z,zoomAndPan` -); -function isRenderableAttrValue(value) { - if (value == null) { - return false; - } - const type = typeof value; - return type === "string" || type === "number" || type === "boolean"; -} - -const cssVarNameEscapeSymbolsRE = /[ !"#$%&'()*+,./:;<=>?@[\\\]^`{|}~]/g; -function getEscapedCssVarName(key, doubleEscape) { - return key.replace( - cssVarNameEscapeSymbolsRE, - (s) => `\\${s}` - ); -} - -function looseCompareArrays(a, b) { - if (a.length !== b.length) return false; - let equal = true; - for (let i = 0; equal && i < a.length; i++) { - equal = looseEqual(a[i], b[i]); - } - return equal; -} -function looseEqual(a, b) { - if (a === b) return true; - let aValidType = isDate(a); - let bValidType = isDate(b); - if (aValidType || bValidType) { - return aValidType && bValidType ? a.getTime() === b.getTime() : false; - } - aValidType = isSymbol(a); - bValidType = isSymbol(b); - if (aValidType || bValidType) { - return a === b; - } - aValidType = isArray(a); - bValidType = isArray(b); - if (aValidType || bValidType) { - return aValidType && bValidType ? looseCompareArrays(a, b) : false; - } - aValidType = isObject(a); - bValidType = isObject(b); - if (aValidType || bValidType) { - if (!aValidType || !bValidType) { - return false; - } - const aKeysCount = Object.keys(a).length; - const bKeysCount = Object.keys(b).length; - if (aKeysCount !== bKeysCount) { - return false; - } - for (const key in a) { - const aHasKey = a.hasOwnProperty(key); - const bHasKey = b.hasOwnProperty(key); - if (aHasKey && !bHasKey || !aHasKey && bHasKey || !looseEqual(a[key], b[key])) { - return false; - } - } - } - return String(a) === String(b); -} -function looseIndexOf(arr, val) { - return arr.findIndex((item) => looseEqual(item, val)); -} - -const isRef$1 = (val) => { - return !!(val && val["__v_isRef"] === true); -}; -const toDisplayString = (val) => { - return isString(val) ? val : val == null ? "" : isArray(val) || isObject(val) && (val.toString === objectToString || !isFunction(val.toString)) ? isRef$1(val) ? toDisplayString(val.value) : JSON.stringify(val, replacer, 2) : String(val); -}; -const replacer = (_key, val) => { - if (isRef$1(val)) { - return replacer(_key, val.value); - } else if (isMap(val)) { - return { - [`Map(${val.size})`]: [...val.entries()].reduce( - (entries, [key, val2], i) => { - entries[stringifySymbol(key, i) + " =>"] = val2; - return entries; - }, - {} - ) - }; - } else if (isSet(val)) { - return { - [`Set(${val.size})`]: [...val.values()].map((v) => stringifySymbol(v)) - }; - } else if (isSymbol(val)) { - return stringifySymbol(val); - } else if (isObject(val) && !isArray(val) && !isPlainObject(val)) { - return String(val); - } - return val; -}; -const stringifySymbol = (v, i = "") => { - var _a; - return ( - // Symbol.description in es2019+ so we need to cast here to pass - // the lib: es2016 check - isSymbol(v) ? `Symbol(${(_a = v.description) != null ? _a : i})` : v - ); -}; - -function warn$2(msg, ...args) { - console.warn(`[Vue warn] ${msg}`, ...args); -} - -let activeEffectScope; -class EffectScope { - constructor(detached = false) { - this.detached = detached; - /** - * @internal - */ - this._active = true; - /** - * @internal - */ - this.effects = []; - /** - * @internal - */ - this.cleanups = []; - this._isPaused = false; - this.parent = activeEffectScope; - if (!detached && activeEffectScope) { - this.index = (activeEffectScope.scopes || (activeEffectScope.scopes = [])).push( - this - ) - 1; - } - } - get active() { - return this._active; - } - pause() { - if (this._active) { - this._isPaused = true; - let i, l; - if (this.scopes) { - for (i = 0, l = this.scopes.length; i < l; i++) { - this.scopes[i].pause(); - } - } - for (i = 0, l = this.effects.length; i < l; i++) { - this.effects[i].pause(); - } - } - } - /** - * Resumes the effect scope, including all child scopes and effects. - */ - resume() { - if (this._active) { - if (this._isPaused) { - this._isPaused = false; - let i, l; - if (this.scopes) { - for (i = 0, l = this.scopes.length; i < l; i++) { - this.scopes[i].resume(); - } - } - for (i = 0, l = this.effects.length; i < l; i++) { - this.effects[i].resume(); - } - } - } - } - run(fn) { - if (this._active) { - const currentEffectScope = activeEffectScope; - try { - activeEffectScope = this; - return fn(); - } finally { - activeEffectScope = currentEffectScope; - } - } else { - warn$2(`cannot run an inactive effect scope.`); - } - } - /** - * This should only be called on non-detached scopes - * @internal - */ - on() { - activeEffectScope = this; - } - /** - * This should only be called on non-detached scopes - * @internal - */ - off() { - activeEffectScope = this.parent; - } - stop(fromParent) { - if (this._active) { - let i, l; - for (i = 0, l = this.effects.length; i < l; i++) { - this.effects[i].stop(); - } - for (i = 0, l = this.cleanups.length; i < l; i++) { - this.cleanups[i](); - } - if (this.scopes) { - for (i = 0, l = this.scopes.length; i < l; i++) { - this.scopes[i].stop(true); - } - } - if (!this.detached && this.parent && !fromParent) { - const last = this.parent.scopes.pop(); - if (last && last !== this) { - this.parent.scopes[this.index] = last; - last.index = this.index; - } - } - this.parent = void 0; - this._active = false; - } - } -} -function effectScope(detached) { - return new EffectScope(detached); -} -function getCurrentScope() { - return activeEffectScope; -} -function onScopeDispose(fn, failSilently = false) { - if (activeEffectScope) { - activeEffectScope.cleanups.push(fn); - } else if (!failSilently) { - warn$2( - `onScopeDispose() is called when there is no active effect scope to be associated with.` - ); - } -} - -let activeSub; -const pausedQueueEffects = /* @__PURE__ */ new WeakSet(); -class ReactiveEffect { - constructor(fn) { - this.fn = fn; - /** - * @internal - */ - this.deps = void 0; - /** - * @internal - */ - this.depsTail = void 0; - /** - * @internal - */ - this.flags = 1 | 4; - /** - * @internal - */ - this.next = void 0; - /** - * @internal - */ - this.cleanup = void 0; - this.scheduler = void 0; - if (activeEffectScope && activeEffectScope.active) { - activeEffectScope.effects.push(this); - } - } - pause() { - this.flags |= 64; - } - resume() { - if (this.flags & 64) { - this.flags &= ~64; - if (pausedQueueEffects.has(this)) { - pausedQueueEffects.delete(this); - this.trigger(); - } - } - } - /** - * @internal - */ - notify() { - if (this.flags & 2 && !(this.flags & 32)) { - return; - } - if (!(this.flags & 8)) { - batch(this); - } - } - run() { - if (!(this.flags & 1)) { - return this.fn(); - } - this.flags |= 2; - cleanupEffect(this); - prepareDeps(this); - const prevEffect = activeSub; - const prevShouldTrack = shouldTrack; - activeSub = this; - shouldTrack = true; - try { - return this.fn(); - } finally { - if (activeSub !== this) { - warn$2( - "Active effect was not restored correctly - this is likely a Vue internal bug." - ); - } - cleanupDeps(this); - activeSub = prevEffect; - shouldTrack = prevShouldTrack; - this.flags &= ~2; - } - } - stop() { - if (this.flags & 1) { - for (let link = this.deps; link; link = link.nextDep) { - removeSub(link); - } - this.deps = this.depsTail = void 0; - cleanupEffect(this); - this.onStop && this.onStop(); - this.flags &= ~1; - } - } - trigger() { - if (this.flags & 64) { - pausedQueueEffects.add(this); - } else if (this.scheduler) { - this.scheduler(); - } else { - this.runIfDirty(); - } - } - /** - * @internal - */ - runIfDirty() { - if (isDirty(this)) { - this.run(); - } - } - get dirty() { - return isDirty(this); - } -} -let batchDepth = 0; -let batchedSub; -let batchedComputed; -function batch(sub, isComputed = false) { - sub.flags |= 8; - if (isComputed) { - sub.next = batchedComputed; - batchedComputed = sub; - return; - } - sub.next = batchedSub; - batchedSub = sub; -} -function startBatch() { - batchDepth++; -} -function endBatch() { - if (--batchDepth > 0) { - return; - } - if (batchedComputed) { - let e = batchedComputed; - batchedComputed = void 0; - while (e) { - const next = e.next; - e.next = void 0; - e.flags &= ~8; - e = next; - } - } - let error; - while (batchedSub) { - let e = batchedSub; - batchedSub = void 0; - while (e) { - const next = e.next; - e.next = void 0; - e.flags &= ~8; - if (e.flags & 1) { - try { - ; - e.trigger(); - } catch (err) { - if (!error) error = err; - } - } - e = next; - } - } - if (error) throw error; -} -function prepareDeps(sub) { - for (let link = sub.deps; link; link = link.nextDep) { - link.version = -1; - link.prevActiveLink = link.dep.activeLink; - link.dep.activeLink = link; - } -} -function cleanupDeps(sub) { - let head; - let tail = sub.depsTail; - let link = tail; - while (link) { - const prev = link.prevDep; - if (link.version === -1) { - if (link === tail) tail = prev; - removeSub(link); - removeDep(link); - } else { - head = link; - } - link.dep.activeLink = link.prevActiveLink; - link.prevActiveLink = void 0; - link = prev; - } - sub.deps = head; - sub.depsTail = tail; -} -function isDirty(sub) { - for (let link = sub.deps; link; link = link.nextDep) { - if (link.dep.version !== link.version || link.dep.computed && (refreshComputed(link.dep.computed) || link.dep.version !== link.version)) { - return true; - } - } - if (sub._dirty) { - return true; - } - return false; -} -function refreshComputed(computed) { - if (computed.flags & 4 && !(computed.flags & 16)) { - return; - } - computed.flags &= ~16; - if (computed.globalVersion === globalVersion) { - return; - } - computed.globalVersion = globalVersion; - const dep = computed.dep; - computed.flags |= 2; - if (dep.version > 0 && !computed.isSSR && computed.deps && !isDirty(computed)) { - computed.flags &= ~2; - return; - } - const prevSub = activeSub; - const prevShouldTrack = shouldTrack; - activeSub = computed; - shouldTrack = true; - try { - prepareDeps(computed); - const value = computed.fn(computed._value); - if (dep.version === 0 || hasChanged(value, computed._value)) { - computed._value = value; - dep.version++; - } - } catch (err) { - dep.version++; - throw err; - } finally { - activeSub = prevSub; - shouldTrack = prevShouldTrack; - cleanupDeps(computed); - computed.flags &= ~2; - } -} -function removeSub(link, soft = false) { - const { dep, prevSub, nextSub } = link; - if (prevSub) { - prevSub.nextSub = nextSub; - link.prevSub = void 0; - } - if (nextSub) { - nextSub.prevSub = prevSub; - link.nextSub = void 0; - } - if (dep.subsHead === link) { - dep.subsHead = nextSub; - } - if (dep.subs === link) { - dep.subs = prevSub; - if (!prevSub && dep.computed) { - dep.computed.flags &= ~4; - for (let l = dep.computed.deps; l; l = l.nextDep) { - removeSub(l, true); - } - } - } - if (!soft && !--dep.sc && dep.map) { - dep.map.delete(dep.key); - } -} -function removeDep(link) { - const { prevDep, nextDep } = link; - if (prevDep) { - prevDep.nextDep = nextDep; - link.prevDep = void 0; - } - if (nextDep) { - nextDep.prevDep = prevDep; - link.nextDep = void 0; - } -} -function effect(fn, options) { - if (fn.effect instanceof ReactiveEffect) { - fn = fn.effect.fn; - } - const e = new ReactiveEffect(fn); - if (options) { - extend(e, options); - } - try { - e.run(); - } catch (err) { - e.stop(); - throw err; - } - const runner = e.run.bind(e); - runner.effect = e; - return runner; -} -function stop(runner) { - runner.effect.stop(); -} -let shouldTrack = true; -const trackStack = []; -function pauseTracking() { - trackStack.push(shouldTrack); - shouldTrack = false; -} -function resetTracking() { - const last = trackStack.pop(); - shouldTrack = last === void 0 ? true : last; -} -function cleanupEffect(e) { - const { cleanup } = e; - e.cleanup = void 0; - if (cleanup) { - const prevSub = activeSub; - activeSub = void 0; - try { - cleanup(); - } finally { - activeSub = prevSub; - } - } -} - -let globalVersion = 0; -class Link { - constructor(sub, dep) { - this.sub = sub; - this.dep = dep; - this.version = dep.version; - this.nextDep = this.prevDep = this.nextSub = this.prevSub = this.prevActiveLink = void 0; - } -} -class Dep { - constructor(computed) { - this.computed = computed; - this.version = 0; - /** - * Link between this dep and the current active effect - */ - this.activeLink = void 0; - /** - * Doubly linked list representing the subscribing effects (tail) - */ - this.subs = void 0; - /** - * For object property deps cleanup - */ - this.map = void 0; - this.key = void 0; - /** - * Subscriber counter - */ - this.sc = 0; - { - this.subsHead = void 0; - } - } - track(debugInfo) { - if (!activeSub || !shouldTrack || activeSub === this.computed) { - return; - } - let link = this.activeLink; - if (link === void 0 || link.sub !== activeSub) { - link = this.activeLink = new Link(activeSub, this); - if (!activeSub.deps) { - activeSub.deps = activeSub.depsTail = link; - } else { - link.prevDep = activeSub.depsTail; - activeSub.depsTail.nextDep = link; - activeSub.depsTail = link; - } - addSub(link); - } else if (link.version === -1) { - link.version = this.version; - if (link.nextDep) { - const next = link.nextDep; - next.prevDep = link.prevDep; - if (link.prevDep) { - link.prevDep.nextDep = next; - } - link.prevDep = activeSub.depsTail; - link.nextDep = void 0; - activeSub.depsTail.nextDep = link; - activeSub.depsTail = link; - if (activeSub.deps === link) { - activeSub.deps = next; - } - } - } - if (activeSub.onTrack) { - activeSub.onTrack( - extend( - { - effect: activeSub - }, - debugInfo - ) - ); - } - return link; - } - trigger(debugInfo) { - this.version++; - globalVersion++; - this.notify(debugInfo); - } - notify(debugInfo) { - startBatch(); - try { - if (true) { - for (let head = this.subsHead; head; head = head.nextSub) { - if (head.sub.onTrigger && !(head.sub.flags & 8)) { - head.sub.onTrigger( - extend( - { - effect: head.sub - }, - debugInfo - ) - ); - } - } - } - for (let link = this.subs; link; link = link.prevSub) { - if (link.sub.notify()) { - ; - link.sub.dep.notify(); - } - } - } finally { - endBatch(); - } - } -} -function addSub(link) { - link.dep.sc++; - if (link.sub.flags & 4) { - const computed = link.dep.computed; - if (computed && !link.dep.subs) { - computed.flags |= 4 | 16; - for (let l = computed.deps; l; l = l.nextDep) { - addSub(l); - } - } - const currentTail = link.dep.subs; - if (currentTail !== link) { - link.prevSub = currentTail; - if (currentTail) currentTail.nextSub = link; - } - if (link.dep.subsHead === void 0) { - link.dep.subsHead = link; - } - link.dep.subs = link; - } -} -const targetMap = /* @__PURE__ */ new WeakMap(); -const ITERATE_KEY = Symbol( - "Object iterate" -); -const MAP_KEY_ITERATE_KEY = Symbol( - "Map keys iterate" -); -const ARRAY_ITERATE_KEY = Symbol( - "Array iterate" -); -function track(target, type, key) { - if (shouldTrack && activeSub) { - let depsMap = targetMap.get(target); - if (!depsMap) { - targetMap.set(target, depsMap = /* @__PURE__ */ new Map()); - } - let dep = depsMap.get(key); - if (!dep) { - depsMap.set(key, dep = new Dep()); - dep.map = depsMap; - dep.key = key; - } - { - dep.track({ - target, - type, - key - }); - } - } -} -function trigger(target, type, key, newValue, oldValue, oldTarget) { - const depsMap = targetMap.get(target); - if (!depsMap) { - globalVersion++; - return; - } - const run = (dep) => { - if (dep) { - { - dep.trigger({ - target, - type, - key, - newValue, - oldValue, - oldTarget - }); - } - } - }; - startBatch(); - if (type === "clear") { - depsMap.forEach(run); - } else { - const targetIsArray = isArray(target); - const isArrayIndex = targetIsArray && isIntegerKey(key); - if (targetIsArray && key === "length") { - const newLength = Number(newValue); - depsMap.forEach((dep, key2) => { - if (key2 === "length" || key2 === ARRAY_ITERATE_KEY || !isSymbol(key2) && key2 >= newLength) { - run(dep); - } - }); - } else { - if (key !== void 0 || depsMap.has(void 0)) { - run(depsMap.get(key)); - } - if (isArrayIndex) { - run(depsMap.get(ARRAY_ITERATE_KEY)); - } - switch (type) { - case "add": - if (!targetIsArray) { - run(depsMap.get(ITERATE_KEY)); - if (isMap(target)) { - run(depsMap.get(MAP_KEY_ITERATE_KEY)); - } - } else if (isArrayIndex) { - run(depsMap.get("length")); - } - break; - case "delete": - if (!targetIsArray) { - run(depsMap.get(ITERATE_KEY)); - if (isMap(target)) { - run(depsMap.get(MAP_KEY_ITERATE_KEY)); - } - } - break; - case "set": - if (isMap(target)) { - run(depsMap.get(ITERATE_KEY)); - } - break; - } - } - } - endBatch(); -} -function getDepFromReactive(object, key) { - const depMap = targetMap.get(object); - return depMap && depMap.get(key); -} - -function reactiveReadArray(array) { - const raw = toRaw(array); - if (raw === array) return raw; - track(raw, "iterate", ARRAY_ITERATE_KEY); - return isShallow(array) ? raw : raw.map(toReactive); -} -function shallowReadArray(arr) { - track(arr = toRaw(arr), "iterate", ARRAY_ITERATE_KEY); - return arr; -} -const arrayInstrumentations = { - __proto__: null, - [Symbol.iterator]() { - return iterator(this, Symbol.iterator, toReactive); - }, - concat(...args) { - return reactiveReadArray(this).concat( - ...args.map((x) => isArray(x) ? reactiveReadArray(x) : x) - ); - }, - entries() { - return iterator(this, "entries", (value) => { - value[1] = toReactive(value[1]); - return value; - }); - }, - every(fn, thisArg) { - return apply(this, "every", fn, thisArg, void 0, arguments); - }, - filter(fn, thisArg) { - return apply(this, "filter", fn, thisArg, (v) => v.map(toReactive), arguments); - }, - find(fn, thisArg) { - return apply(this, "find", fn, thisArg, toReactive, arguments); - }, - findIndex(fn, thisArg) { - return apply(this, "findIndex", fn, thisArg, void 0, arguments); - }, - findLast(fn, thisArg) { - return apply(this, "findLast", fn, thisArg, toReactive, arguments); - }, - findLastIndex(fn, thisArg) { - return apply(this, "findLastIndex", fn, thisArg, void 0, arguments); - }, - // flat, flatMap could benefit from ARRAY_ITERATE but are not straight-forward to implement - forEach(fn, thisArg) { - return apply(this, "forEach", fn, thisArg, void 0, arguments); - }, - includes(...args) { - return searchProxy(this, "includes", args); - }, - indexOf(...args) { - return searchProxy(this, "indexOf", args); - }, - join(separator) { - return reactiveReadArray(this).join(separator); - }, - // keys() iterator only reads `length`, no optimisation required - lastIndexOf(...args) { - return searchProxy(this, "lastIndexOf", args); - }, - map(fn, thisArg) { - return apply(this, "map", fn, thisArg, void 0, arguments); - }, - pop() { - return noTracking(this, "pop"); - }, - push(...args) { - return noTracking(this, "push", args); - }, - reduce(fn, ...args) { - return reduce(this, "reduce", fn, args); - }, - reduceRight(fn, ...args) { - return reduce(this, "reduceRight", fn, args); - }, - shift() { - return noTracking(this, "shift"); - }, - // slice could use ARRAY_ITERATE but also seems to beg for range tracking - some(fn, thisArg) { - return apply(this, "some", fn, thisArg, void 0, arguments); - }, - splice(...args) { - return noTracking(this, "splice", args); - }, - toReversed() { - return reactiveReadArray(this).toReversed(); - }, - toSorted(comparer) { - return reactiveReadArray(this).toSorted(comparer); - }, - toSpliced(...args) { - return reactiveReadArray(this).toSpliced(...args); - }, - unshift(...args) { - return noTracking(this, "unshift", args); - }, - values() { - return iterator(this, "values", toReactive); - } -}; -function iterator(self, method, wrapValue) { - const arr = shallowReadArray(self); - const iter = arr[method](); - if (arr !== self && !isShallow(self)) { - iter._next = iter.next; - iter.next = () => { - const result = iter._next(); - if (result.value) { - result.value = wrapValue(result.value); - } - return result; - }; - } - return iter; -} -const arrayProto = Array.prototype; -function apply(self, method, fn, thisArg, wrappedRetFn, args) { - const arr = shallowReadArray(self); - const needsWrap = arr !== self && !isShallow(self); - const methodFn = arr[method]; - if (methodFn !== arrayProto[method]) { - const result2 = methodFn.apply(self, args); - return needsWrap ? toReactive(result2) : result2; - } - let wrappedFn = fn; - if (arr !== self) { - if (needsWrap) { - wrappedFn = function(item, index) { - return fn.call(this, toReactive(item), index, self); - }; - } else if (fn.length > 2) { - wrappedFn = function(item, index) { - return fn.call(this, item, index, self); - }; - } - } - const result = methodFn.call(arr, wrappedFn, thisArg); - return needsWrap && wrappedRetFn ? wrappedRetFn(result) : result; -} -function reduce(self, method, fn, args) { - const arr = shallowReadArray(self); - let wrappedFn = fn; - if (arr !== self) { - if (!isShallow(self)) { - wrappedFn = function(acc, item, index) { - return fn.call(this, acc, toReactive(item), index, self); - }; - } else if (fn.length > 3) { - wrappedFn = function(acc, item, index) { - return fn.call(this, acc, item, index, self); - }; - } - } - return arr[method](wrappedFn, ...args); -} -function searchProxy(self, method, args) { - const arr = toRaw(self); - track(arr, "iterate", ARRAY_ITERATE_KEY); - const res = arr[method](...args); - if ((res === -1 || res === false) && isProxy(args[0])) { - args[0] = toRaw(args[0]); - return arr[method](...args); - } - return res; -} -function noTracking(self, method, args = []) { - pauseTracking(); - startBatch(); - const res = toRaw(self)[method].apply(self, args); - endBatch(); - resetTracking(); - return res; -} - -const isNonTrackableKeys = /* @__PURE__ */ makeMap(`__proto__,__v_isRef,__isVue`); -const builtInSymbols = new Set( - /* @__PURE__ */ Object.getOwnPropertyNames(Symbol).filter((key) => key !== "arguments" && key !== "caller").map((key) => Symbol[key]).filter(isSymbol) -); -function hasOwnProperty(key) { - if (!isSymbol(key)) key = String(key); - const obj = toRaw(this); - track(obj, "has", key); - return obj.hasOwnProperty(key); -} -class BaseReactiveHandler { - constructor(_isReadonly = false, _isShallow = false) { - this._isReadonly = _isReadonly; - this._isShallow = _isShallow; - } - get(target, key, receiver) { - const isReadonly2 = this._isReadonly, isShallow2 = this._isShallow; - if (key === "__v_isReactive") { - return !isReadonly2; - } else if (key === "__v_isReadonly") { - return isReadonly2; - } else if (key === "__v_isShallow") { - return isShallow2; - } else if (key === "__v_raw") { - if (receiver === (isReadonly2 ? isShallow2 ? shallowReadonlyMap : readonlyMap : isShallow2 ? shallowReactiveMap : reactiveMap).get(target) || // receiver is not the reactive proxy, but has the same prototype - // this means the receiver is a user proxy of the reactive proxy - Object.getPrototypeOf(target) === Object.getPrototypeOf(receiver)) { - return target; - } - return; - } - const targetIsArray = isArray(target); - if (!isReadonly2) { - let fn; - if (targetIsArray && (fn = arrayInstrumentations[key])) { - return fn; - } - if (key === "hasOwnProperty") { - return hasOwnProperty; - } - } - const res = Reflect.get( - target, - key, - // if this is a proxy wrapping a ref, return methods using the raw ref - // as receiver so that we don't have to call `toRaw` on the ref in all - // its class methods - isRef(target) ? target : receiver - ); - if (isSymbol(key) ? builtInSymbols.has(key) : isNonTrackableKeys(key)) { - return res; - } - if (!isReadonly2) { - track(target, "get", key); - } - if (isShallow2) { - return res; - } - if (isRef(res)) { - return targetIsArray && isIntegerKey(key) ? res : res.value; - } - if (isObject(res)) { - return isReadonly2 ? readonly(res) : reactive(res); - } - return res; - } -} -class MutableReactiveHandler extends BaseReactiveHandler { - constructor(isShallow2 = false) { - super(false, isShallow2); - } - set(target, key, value, receiver) { - let oldValue = target[key]; - if (!this._isShallow) { - const isOldValueReadonly = isReadonly(oldValue); - if (!isShallow(value) && !isReadonly(value)) { - oldValue = toRaw(oldValue); - value = toRaw(value); - } - if (!isArray(target) && isRef(oldValue) && !isRef(value)) { - if (isOldValueReadonly) { - return false; - } else { - oldValue.value = value; - return true; - } - } - } - const hadKey = isArray(target) && isIntegerKey(key) ? Number(key) < target.length : hasOwn(target, key); - const result = Reflect.set( - target, - key, - value, - isRef(target) ? target : receiver - ); - if (target === toRaw(receiver)) { - if (!hadKey) { - trigger(target, "add", key, value); - } else if (hasChanged(value, oldValue)) { - trigger(target, "set", key, value, oldValue); - } - } - return result; - } - deleteProperty(target, key) { - const hadKey = hasOwn(target, key); - const oldValue = target[key]; - const result = Reflect.deleteProperty(target, key); - if (result && hadKey) { - trigger(target, "delete", key, void 0, oldValue); - } - return result; - } - has(target, key) { - const result = Reflect.has(target, key); - if (!isSymbol(key) || !builtInSymbols.has(key)) { - track(target, "has", key); - } - return result; - } - ownKeys(target) { - track( - target, - "iterate", - isArray(target) ? "length" : ITERATE_KEY - ); - return Reflect.ownKeys(target); - } -} -class ReadonlyReactiveHandler extends BaseReactiveHandler { - constructor(isShallow2 = false) { - super(true, isShallow2); - } - set(target, key) { - { - warn$2( - `Set operation on key "${String(key)}" failed: target is readonly.`, - target - ); - } - return true; - } - deleteProperty(target, key) { - { - warn$2( - `Delete operation on key "${String(key)}" failed: target is readonly.`, - target - ); - } - return true; - } -} -const mutableHandlers = /* @__PURE__ */ new MutableReactiveHandler(); -const readonlyHandlers = /* @__PURE__ */ new ReadonlyReactiveHandler(); -const shallowReactiveHandlers = /* @__PURE__ */ new MutableReactiveHandler(true); -const shallowReadonlyHandlers = /* @__PURE__ */ new ReadonlyReactiveHandler(true); - -const toShallow = (value) => value; -const getProto = (v) => Reflect.getPrototypeOf(v); -function createIterableMethod(method, isReadonly2, isShallow2) { - return function(...args) { - const target = this["__v_raw"]; - const rawTarget = toRaw(target); - const targetIsMap = isMap(rawTarget); - const isPair = method === "entries" || method === Symbol.iterator && targetIsMap; - const isKeyOnly = method === "keys" && targetIsMap; - const innerIterator = target[method](...args); - const wrap = isShallow2 ? toShallow : isReadonly2 ? toReadonly : toReactive; - !isReadonly2 && track( - rawTarget, - "iterate", - isKeyOnly ? MAP_KEY_ITERATE_KEY : ITERATE_KEY - ); - return { - // iterator protocol - next() { - const { value, done } = innerIterator.next(); - return done ? { value, done } : { - value: isPair ? [wrap(value[0]), wrap(value[1])] : wrap(value), - done - }; - }, - // iterable protocol - [Symbol.iterator]() { - return this; - } - }; - }; -} -function createReadonlyMethod(type) { - return function(...args) { - { - const key = args[0] ? `on key "${args[0]}" ` : ``; - warn$2( - `${capitalize(type)} operation ${key}failed: target is readonly.`, - toRaw(this) - ); - } - return type === "delete" ? false : type === "clear" ? void 0 : this; - }; -} -function createInstrumentations(readonly, shallow) { - const instrumentations = { - get(key) { - const target = this["__v_raw"]; - const rawTarget = toRaw(target); - const rawKey = toRaw(key); - if (!readonly) { - if (hasChanged(key, rawKey)) { - track(rawTarget, "get", key); - } - track(rawTarget, "get", rawKey); - } - const { has } = getProto(rawTarget); - const wrap = shallow ? toShallow : readonly ? toReadonly : toReactive; - if (has.call(rawTarget, key)) { - return wrap(target.get(key)); - } else if (has.call(rawTarget, rawKey)) { - return wrap(target.get(rawKey)); - } else if (target !== rawTarget) { - target.get(key); - } - }, - get size() { - const target = this["__v_raw"]; - !readonly && track(toRaw(target), "iterate", ITERATE_KEY); - return Reflect.get(target, "size", target); - }, - has(key) { - const target = this["__v_raw"]; - const rawTarget = toRaw(target); - const rawKey = toRaw(key); - if (!readonly) { - if (hasChanged(key, rawKey)) { - track(rawTarget, "has", key); - } - track(rawTarget, "has", rawKey); - } - return key === rawKey ? target.has(key) : target.has(key) || target.has(rawKey); - }, - forEach(callback, thisArg) { - const observed = this; - const target = observed["__v_raw"]; - const rawTarget = toRaw(target); - const wrap = shallow ? toShallow : readonly ? toReadonly : toReactive; - !readonly && track(rawTarget, "iterate", ITERATE_KEY); - return target.forEach((value, key) => { - return callback.call(thisArg, wrap(value), wrap(key), observed); - }); - } - }; - extend( - instrumentations, - readonly ? { - add: createReadonlyMethod("add"), - set: createReadonlyMethod("set"), - delete: createReadonlyMethod("delete"), - clear: createReadonlyMethod("clear") - } : { - add(value) { - if (!shallow && !isShallow(value) && !isReadonly(value)) { - value = toRaw(value); - } - const target = toRaw(this); - const proto = getProto(target); - const hadKey = proto.has.call(target, value); - if (!hadKey) { - target.add(value); - trigger(target, "add", value, value); - } - return this; - }, - set(key, value) { - if (!shallow && !isShallow(value) && !isReadonly(value)) { - value = toRaw(value); - } - const target = toRaw(this); - const { has, get } = getProto(target); - let hadKey = has.call(target, key); - if (!hadKey) { - key = toRaw(key); - hadKey = has.call(target, key); - } else { - checkIdentityKeys(target, has, key); - } - const oldValue = get.call(target, key); - target.set(key, value); - if (!hadKey) { - trigger(target, "add", key, value); - } else if (hasChanged(value, oldValue)) { - trigger(target, "set", key, value, oldValue); - } - return this; - }, - delete(key) { - const target = toRaw(this); - const { has, get } = getProto(target); - let hadKey = has.call(target, key); - if (!hadKey) { - key = toRaw(key); - hadKey = has.call(target, key); - } else { - checkIdentityKeys(target, has, key); - } - const oldValue = get ? get.call(target, key) : void 0; - const result = target.delete(key); - if (hadKey) { - trigger(target, "delete", key, void 0, oldValue); - } - return result; - }, - clear() { - const target = toRaw(this); - const hadItems = target.size !== 0; - const oldTarget = isMap(target) ? new Map(target) : new Set(target) ; - const result = target.clear(); - if (hadItems) { - trigger( - target, - "clear", - void 0, - void 0, - oldTarget - ); - } - return result; - } - } - ); - const iteratorMethods = [ - "keys", - "values", - "entries", - Symbol.iterator - ]; - iteratorMethods.forEach((method) => { - instrumentations[method] = createIterableMethod(method, readonly, shallow); - }); - return instrumentations; -} -function createInstrumentationGetter(isReadonly2, shallow) { - const instrumentations = createInstrumentations(isReadonly2, shallow); - return (target, key, receiver) => { - if (key === "__v_isReactive") { - return !isReadonly2; - } else if (key === "__v_isReadonly") { - return isReadonly2; - } else if (key === "__v_raw") { - return target; - } - return Reflect.get( - hasOwn(instrumentations, key) && key in target ? instrumentations : target, - key, - receiver - ); - }; -} -const mutableCollectionHandlers = { - get: /* @__PURE__ */ createInstrumentationGetter(false, false) -}; -const shallowCollectionHandlers = { - get: /* @__PURE__ */ createInstrumentationGetter(false, true) -}; -const readonlyCollectionHandlers = { - get: /* @__PURE__ */ createInstrumentationGetter(true, false) -}; -const shallowReadonlyCollectionHandlers = { - get: /* @__PURE__ */ createInstrumentationGetter(true, true) -}; -function checkIdentityKeys(target, has, key) { - const rawKey = toRaw(key); - if (rawKey !== key && has.call(target, rawKey)) { - const type = toRawType(target); - warn$2( - `Reactive ${type} contains both the raw and reactive versions of the same object${type === `Map` ? ` as keys` : ``}, which can lead to inconsistencies. Avoid differentiating between the raw and reactive versions of an object and only use the reactive version if possible.` - ); - } -} - -const reactiveMap = /* @__PURE__ */ new WeakMap(); -const shallowReactiveMap = /* @__PURE__ */ new WeakMap(); -const readonlyMap = /* @__PURE__ */ new WeakMap(); -const shallowReadonlyMap = /* @__PURE__ */ new WeakMap(); -function targetTypeMap(rawType) { - switch (rawType) { - case "Object": - case "Array": - return 1 /* COMMON */; - case "Map": - case "Set": - case "WeakMap": - case "WeakSet": - return 2 /* COLLECTION */; - default: - return 0 /* INVALID */; - } -} -function getTargetType(value) { - return value["__v_skip"] || !Object.isExtensible(value) ? 0 /* INVALID */ : targetTypeMap(toRawType(value)); -} -function reactive(target) { - if (isReadonly(target)) { - return target; - } - return createReactiveObject( - target, - false, - mutableHandlers, - mutableCollectionHandlers, - reactiveMap - ); -} -function shallowReactive(target) { - return createReactiveObject( - target, - false, - shallowReactiveHandlers, - shallowCollectionHandlers, - shallowReactiveMap - ); -} -function readonly(target) { - return createReactiveObject( - target, - true, - readonlyHandlers, - readonlyCollectionHandlers, - readonlyMap - ); -} -function shallowReadonly(target) { - return createReactiveObject( - target, - true, - shallowReadonlyHandlers, - shallowReadonlyCollectionHandlers, - shallowReadonlyMap - ); -} -function createReactiveObject(target, isReadonly2, baseHandlers, collectionHandlers, proxyMap) { - if (!isObject(target)) { - { - warn$2( - `value cannot be made ${isReadonly2 ? "readonly" : "reactive"}: ${String( - target - )}` - ); - } - return target; - } - if (target["__v_raw"] && !(isReadonly2 && target["__v_isReactive"])) { - return target; - } - const existingProxy = proxyMap.get(target); - if (existingProxy) { - return existingProxy; - } - const targetType = getTargetType(target); - if (targetType === 0 /* INVALID */) { - return target; - } - const proxy = new Proxy( - target, - targetType === 2 /* COLLECTION */ ? collectionHandlers : baseHandlers - ); - proxyMap.set(target, proxy); - return proxy; -} -function isReactive(value) { - if (isReadonly(value)) { - return isReactive(value["__v_raw"]); - } - return !!(value && value["__v_isReactive"]); -} -function isReadonly(value) { - return !!(value && value["__v_isReadonly"]); -} -function isShallow(value) { - return !!(value && value["__v_isShallow"]); -} -function isProxy(value) { - return value ? !!value["__v_raw"] : false; -} -function toRaw(observed) { - const raw = observed && observed["__v_raw"]; - return raw ? toRaw(raw) : observed; -} -function markRaw(value) { - if (!hasOwn(value, "__v_skip") && Object.isExtensible(value)) { - def(value, "__v_skip", true); - } - return value; -} -const toReactive = (value) => isObject(value) ? reactive(value) : value; -const toReadonly = (value) => isObject(value) ? readonly(value) : value; - -function isRef(r) { - return r ? r["__v_isRef"] === true : false; -} -function ref(value) { - return createRef(value, false); -} -function shallowRef(value) { - return createRef(value, true); -} -function createRef(rawValue, shallow) { - if (isRef(rawValue)) { - return rawValue; - } - return new RefImpl(rawValue, shallow); -} -class RefImpl { - constructor(value, isShallow2) { - this.dep = new Dep(); - this["__v_isRef"] = true; - this["__v_isShallow"] = false; - this._rawValue = isShallow2 ? value : toRaw(value); - this._value = isShallow2 ? value : toReactive(value); - this["__v_isShallow"] = isShallow2; - } - get value() { - { - this.dep.track({ - target: this, - type: "get", - key: "value" - }); - } - return this._value; - } - set value(newValue) { - const oldValue = this._rawValue; - const useDirectValue = this["__v_isShallow"] || isShallow(newValue) || isReadonly(newValue); - newValue = useDirectValue ? newValue : toRaw(newValue); - if (hasChanged(newValue, oldValue)) { - this._rawValue = newValue; - this._value = useDirectValue ? newValue : toReactive(newValue); - { - this.dep.trigger({ - target: this, - type: "set", - key: "value", - newValue, - oldValue - }); - } - } - } -} -function triggerRef(ref2) { - if (ref2.dep) { - { - ref2.dep.trigger({ - target: ref2, - type: "set", - key: "value", - newValue: ref2._value - }); - } - } -} -function unref(ref2) { - return isRef(ref2) ? ref2.value : ref2; -} -function toValue(source) { - return isFunction(source) ? source() : unref(source); -} -const shallowUnwrapHandlers = { - get: (target, key, receiver) => key === "__v_raw" ? target : unref(Reflect.get(target, key, receiver)), - set: (target, key, value, receiver) => { - const oldValue = target[key]; - if (isRef(oldValue) && !isRef(value)) { - oldValue.value = value; - return true; - } else { - return Reflect.set(target, key, value, receiver); - } - } -}; -function proxyRefs(objectWithRefs) { - return isReactive(objectWithRefs) ? objectWithRefs : new Proxy(objectWithRefs, shallowUnwrapHandlers); -} -class CustomRefImpl { - constructor(factory) { - this["__v_isRef"] = true; - this._value = void 0; - const dep = this.dep = new Dep(); - const { get, set } = factory(dep.track.bind(dep), dep.trigger.bind(dep)); - this._get = get; - this._set = set; - } - get value() { - return this._value = this._get(); - } - set value(newVal) { - this._set(newVal); - } -} -function customRef(factory) { - return new CustomRefImpl(factory); -} -function toRefs(object) { - if (!isProxy(object)) { - warn$2(`toRefs() expects a reactive object but received a plain one.`); - } - const ret = isArray(object) ? new Array(object.length) : {}; - for (const key in object) { - ret[key] = propertyToRef(object, key); - } - return ret; -} -class ObjectRefImpl { - constructor(_object, _key, _defaultValue) { - this._object = _object; - this._key = _key; - this._defaultValue = _defaultValue; - this["__v_isRef"] = true; - this._value = void 0; - } - get value() { - const val = this._object[this._key]; - return this._value = val === void 0 ? this._defaultValue : val; - } - set value(newVal) { - this._object[this._key] = newVal; - } - get dep() { - return getDepFromReactive(toRaw(this._object), this._key); - } -} -class GetterRefImpl { - constructor(_getter) { - this._getter = _getter; - this["__v_isRef"] = true; - this["__v_isReadonly"] = true; - this._value = void 0; - } - get value() { - return this._value = this._getter(); - } -} -function toRef(source, key, defaultValue) { - if (isRef(source)) { - return source; - } else if (isFunction(source)) { - return new GetterRefImpl(source); - } else if (isObject(source) && arguments.length > 1) { - return propertyToRef(source, key, defaultValue); - } else { - return ref(source); - } -} -function propertyToRef(source, key, defaultValue) { - const val = source[key]; - return isRef(val) ? val : new ObjectRefImpl(source, key, defaultValue); -} - -class ComputedRefImpl { - constructor(fn, setter, isSSR) { - this.fn = fn; - this.setter = setter; - /** - * @internal - */ - this._value = void 0; - /** - * @internal - */ - this.dep = new Dep(this); - /** - * @internal - */ - this.__v_isRef = true; - // TODO isolatedDeclarations "__v_isReadonly" - // A computed is also a subscriber that tracks other deps - /** - * @internal - */ - this.deps = void 0; - /** - * @internal - */ - this.depsTail = void 0; - /** - * @internal - */ - this.flags = 16; - /** - * @internal - */ - this.globalVersion = globalVersion - 1; - /** - * @internal - */ - this.next = void 0; - // for backwards compat - this.effect = this; - this["__v_isReadonly"] = !setter; - this.isSSR = isSSR; - } - /** - * @internal - */ - notify() { - this.flags |= 16; - if (!(this.flags & 8) && // avoid infinite self recursion - activeSub !== this) { - batch(this, true); - return true; - } - } - get value() { - const link = this.dep.track({ - target: this, - type: "get", - key: "value" - }) ; - refreshComputed(this); - if (link) { - link.version = this.dep.version; - } - return this._value; - } - set value(newValue) { - if (this.setter) { - this.setter(newValue); - } else { - warn$2("Write operation failed: computed value is readonly"); - } - } -} -function computed$1(getterOrOptions, debugOptions, isSSR = false) { - let getter; - let setter; - if (isFunction(getterOrOptions)) { - getter = getterOrOptions; - } else { - getter = getterOrOptions.get; - setter = getterOrOptions.set; - } - const cRef = new ComputedRefImpl(getter, setter, isSSR); - if (debugOptions && !isSSR) { - cRef.onTrack = debugOptions.onTrack; - cRef.onTrigger = debugOptions.onTrigger; - } - return cRef; -} - -const TrackOpTypes = { - "GET": "get", - "HAS": "has", - "ITERATE": "iterate" -}; -const TriggerOpTypes = { - "SET": "set", - "ADD": "add", - "DELETE": "delete", - "CLEAR": "clear" -}; - -const INITIAL_WATCHER_VALUE = {}; -const cleanupMap = /* @__PURE__ */ new WeakMap(); -let activeWatcher = void 0; -function getCurrentWatcher() { - return activeWatcher; -} -function onWatcherCleanup(cleanupFn, failSilently = false, owner = activeWatcher) { - if (owner) { - let cleanups = cleanupMap.get(owner); - if (!cleanups) cleanupMap.set(owner, cleanups = []); - cleanups.push(cleanupFn); - } else if (!failSilently) { - warn$2( - `onWatcherCleanup() was called when there was no active watcher to associate with.` - ); - } -} -function watch$1(source, cb, options = EMPTY_OBJ) { - const { immediate, deep, once, scheduler, augmentJob, call } = options; - const warnInvalidSource = (s) => { - (options.onWarn || warn$2)( - `Invalid watch source: `, - s, - `A watch source can only be a getter/effect function, a ref, a reactive object, or an array of these types.` - ); - }; - const reactiveGetter = (source2) => { - if (deep) return source2; - if (isShallow(source2) || deep === false || deep === 0) - return traverse(source2, 1); - return traverse(source2); - }; - let effect; - let getter; - let cleanup; - let boundCleanup; - let forceTrigger = false; - let isMultiSource = false; - if (isRef(source)) { - getter = () => source.value; - forceTrigger = isShallow(source); - } else if (isReactive(source)) { - getter = () => reactiveGetter(source); - forceTrigger = true; - } else if (isArray(source)) { - isMultiSource = true; - forceTrigger = source.some((s) => isReactive(s) || isShallow(s)); - getter = () => source.map((s) => { - if (isRef(s)) { - return s.value; - } else if (isReactive(s)) { - return reactiveGetter(s); - } else if (isFunction(s)) { - return call ? call(s, 2) : s(); - } else { - warnInvalidSource(s); - } - }); - } else if (isFunction(source)) { - if (cb) { - getter = call ? () => call(source, 2) : source; - } else { - getter = () => { - if (cleanup) { - pauseTracking(); - try { - cleanup(); - } finally { - resetTracking(); - } - } - const currentEffect = activeWatcher; - activeWatcher = effect; - try { - return call ? call(source, 3, [boundCleanup]) : source(boundCleanup); - } finally { - activeWatcher = currentEffect; - } - }; - } - } else { - getter = NOOP; - warnInvalidSource(source); - } - if (cb && deep) { - const baseGetter = getter; - const depth = deep === true ? Infinity : deep; - getter = () => traverse(baseGetter(), depth); - } - const scope = getCurrentScope(); - const watchHandle = () => { - effect.stop(); - if (scope) { - remove(scope.effects, effect); - } - }; - if (once && cb) { - const _cb = cb; - cb = (...args) => { - _cb(...args); - watchHandle(); - }; - } - let oldValue = isMultiSource ? new Array(source.length).fill(INITIAL_WATCHER_VALUE) : INITIAL_WATCHER_VALUE; - const job = (immediateFirstRun) => { - if (!(effect.flags & 1) || !effect.dirty && !immediateFirstRun) { - return; - } - if (cb) { - const newValue = effect.run(); - if (deep || forceTrigger || (isMultiSource ? newValue.some((v, i) => hasChanged(v, oldValue[i])) : hasChanged(newValue, oldValue))) { - if (cleanup) { - cleanup(); - } - const currentWatcher = activeWatcher; - activeWatcher = effect; - try { - const args = [ - newValue, - // pass undefined as the old value when it's changed for the first time - oldValue === INITIAL_WATCHER_VALUE ? void 0 : isMultiSource && oldValue[0] === INITIAL_WATCHER_VALUE ? [] : oldValue, - boundCleanup - ]; - call ? call(cb, 3, args) : ( - // @ts-expect-error - cb(...args) - ); - oldValue = newValue; - } finally { - activeWatcher = currentWatcher; - } - } - } else { - effect.run(); - } - }; - if (augmentJob) { - augmentJob(job); - } - effect = new ReactiveEffect(getter); - effect.scheduler = scheduler ? () => scheduler(job, false) : job; - boundCleanup = (fn) => onWatcherCleanup(fn, false, effect); - cleanup = effect.onStop = () => { - const cleanups = cleanupMap.get(effect); - if (cleanups) { - if (call) { - call(cleanups, 4); - } else { - for (const cleanup2 of cleanups) cleanup2(); - } - cleanupMap.delete(effect); - } - }; - { - effect.onTrack = options.onTrack; - effect.onTrigger = options.onTrigger; - } - if (cb) { - if (immediate) { - job(true); - } else { - oldValue = effect.run(); - } - } else if (scheduler) { - scheduler(job.bind(null, true), true); - } else { - effect.run(); - } - watchHandle.pause = effect.pause.bind(effect); - watchHandle.resume = effect.resume.bind(effect); - watchHandle.stop = watchHandle; - return watchHandle; -} -function traverse(value, depth = Infinity, seen) { - if (depth <= 0 || !isObject(value) || value["__v_skip"]) { - return value; - } - seen = seen || /* @__PURE__ */ new Set(); - if (seen.has(value)) { - return value; - } - seen.add(value); - depth--; - if (isRef(value)) { - traverse(value.value, depth, seen); - } else if (isArray(value)) { - for (let i = 0; i < value.length; i++) { - traverse(value[i], depth, seen); - } - } else if (isSet(value) || isMap(value)) { - value.forEach((v) => { - traverse(v, depth, seen); - }); - } else if (isPlainObject(value)) { - for (const key in value) { - traverse(value[key], depth, seen); - } - for (const key of Object.getOwnPropertySymbols(value)) { - if (Object.prototype.propertyIsEnumerable.call(value, key)) { - traverse(value[key], depth, seen); - } - } - } - return value; -} - -const stack$1 = []; -function pushWarningContext(vnode) { - stack$1.push(vnode); -} -function popWarningContext() { - stack$1.pop(); -} -let isWarning = false; -function warn$1(msg, ...args) { - if (isWarning) return; - isWarning = true; - pauseTracking(); - const instance = stack$1.length ? stack$1[stack$1.length - 1].component : null; - const appWarnHandler = instance && instance.appContext.config.warnHandler; - const trace = getComponentTrace(); - if (appWarnHandler) { - callWithErrorHandling( - appWarnHandler, - instance, - 11, - [ - // eslint-disable-next-line no-restricted-syntax - msg + args.map((a) => { - var _a, _b; - return (_b = (_a = a.toString) == null ? void 0 : _a.call(a)) != null ? _b : JSON.stringify(a); - }).join(""), - instance && instance.proxy, - trace.map( - ({ vnode }) => `at <${formatComponentName(instance, vnode.type)}>` - ).join("\n"), - trace - ] - ); - } else { - const warnArgs = [`[Vue warn]: ${msg}`, ...args]; - if (trace.length && // avoid spamming console during tests - true) { - warnArgs.push(` -`, ...formatTrace(trace)); - } - console.warn(...warnArgs); - } - resetTracking(); - isWarning = false; -} -function getComponentTrace() { - let currentVNode = stack$1[stack$1.length - 1]; - if (!currentVNode) { - return []; - } - const normalizedStack = []; - while (currentVNode) { - const last = normalizedStack[0]; - if (last && last.vnode === currentVNode) { - last.recurseCount++; - } else { - normalizedStack.push({ - vnode: currentVNode, - recurseCount: 0 - }); - } - const parentInstance = currentVNode.component && currentVNode.component.parent; - currentVNode = parentInstance && parentInstance.vnode; - } - return normalizedStack; -} -function formatTrace(trace) { - const logs = []; - trace.forEach((entry, i) => { - logs.push(...i === 0 ? [] : [` -`], ...formatTraceEntry(entry)); - }); - return logs; -} -function formatTraceEntry({ vnode, recurseCount }) { - const postfix = recurseCount > 0 ? `... (${recurseCount} recursive calls)` : ``; - const isRoot = vnode.component ? vnode.component.parent == null : false; - const open = ` at <${formatComponentName( - vnode.component, - vnode.type, - isRoot - )}`; - const close = `>` + postfix; - return vnode.props ? [open, ...formatProps(vnode.props), close] : [open + close]; -} -function formatProps(props) { - const res = []; - const keys = Object.keys(props); - keys.slice(0, 3).forEach((key) => { - res.push(...formatProp(key, props[key])); - }); - if (keys.length > 3) { - res.push(` ...`); - } - return res; -} -function formatProp(key, value, raw) { - if (isString(value)) { - value = JSON.stringify(value); - return raw ? value : [`${key}=${value}`]; - } else if (typeof value === "number" || typeof value === "boolean" || value == null) { - return raw ? value : [`${key}=${value}`]; - } else if (isRef(value)) { - value = formatProp(key, toRaw(value.value), true); - return raw ? value : [`${key}=Ref<`, value, `>`]; - } else if (isFunction(value)) { - return [`${key}=fn${value.name ? `<${value.name}>` : ``}`]; - } else { - value = toRaw(value); - return raw ? value : [`${key}=`, value]; - } -} -function assertNumber(val, type) { - if (val === void 0) { - return; - } else if (typeof val !== "number") { - warn$1(`${type} is not a valid number - got ${JSON.stringify(val)}.`); - } else if (isNaN(val)) { - warn$1(`${type} is NaN - the duration expression might be incorrect.`); - } -} - -const ErrorCodes = { - "SETUP_FUNCTION": 0, - "0": "SETUP_FUNCTION", - "RENDER_FUNCTION": 1, - "1": "RENDER_FUNCTION", - "NATIVE_EVENT_HANDLER": 5, - "5": "NATIVE_EVENT_HANDLER", - "COMPONENT_EVENT_HANDLER": 6, - "6": "COMPONENT_EVENT_HANDLER", - "VNODE_HOOK": 7, - "7": "VNODE_HOOK", - "DIRECTIVE_HOOK": 8, - "8": "DIRECTIVE_HOOK", - "TRANSITION_HOOK": 9, - "9": "TRANSITION_HOOK", - "APP_ERROR_HANDLER": 10, - "10": "APP_ERROR_HANDLER", - "APP_WARN_HANDLER": 11, - "11": "APP_WARN_HANDLER", - "FUNCTION_REF": 12, - "12": "FUNCTION_REF", - "ASYNC_COMPONENT_LOADER": 13, - "13": "ASYNC_COMPONENT_LOADER", - "SCHEDULER": 14, - "14": "SCHEDULER", - "COMPONENT_UPDATE": 15, - "15": "COMPONENT_UPDATE", - "APP_UNMOUNT_CLEANUP": 16, - "16": "APP_UNMOUNT_CLEANUP" -}; -const ErrorTypeStrings$1 = { - ["sp"]: "serverPrefetch hook", - ["bc"]: "beforeCreate hook", - ["c"]: "created hook", - ["bm"]: "beforeMount hook", - ["m"]: "mounted hook", - ["bu"]: "beforeUpdate hook", - ["u"]: "updated", - ["bum"]: "beforeUnmount hook", - ["um"]: "unmounted hook", - ["a"]: "activated hook", - ["da"]: "deactivated hook", - ["ec"]: "errorCaptured hook", - ["rtc"]: "renderTracked hook", - ["rtg"]: "renderTriggered hook", - [0]: "setup function", - [1]: "render function", - [2]: "watcher getter", - [3]: "watcher callback", - [4]: "watcher cleanup function", - [5]: "native event handler", - [6]: "component event handler", - [7]: "vnode hook", - [8]: "directive hook", - [9]: "transition hook", - [10]: "app errorHandler", - [11]: "app warnHandler", - [12]: "ref function", - [13]: "async component loader", - [14]: "scheduler flush", - [15]: "component update", - [16]: "app unmount cleanup function" -}; -function callWithErrorHandling(fn, instance, type, args) { - try { - return args ? fn(...args) : fn(); - } catch (err) { - handleError(err, instance, type); - } -} -function callWithAsyncErrorHandling(fn, instance, type, args) { - if (isFunction(fn)) { - const res = callWithErrorHandling(fn, instance, type, args); - if (res && isPromise(res)) { - res.catch((err) => { - handleError(err, instance, type); - }); - } - return res; - } - if (isArray(fn)) { - const values = []; - for (let i = 0; i < fn.length; i++) { - values.push(callWithAsyncErrorHandling(fn[i], instance, type, args)); - } - return values; - } else { - warn$1( - `Invalid value type passed to callWithAsyncErrorHandling(): ${typeof fn}` - ); - } -} -function handleError(err, instance, type, throwInDev = true) { - const contextVNode = instance ? instance.vnode : null; - const { errorHandler, throwUnhandledErrorInProduction } = instance && instance.appContext.config || EMPTY_OBJ; - if (instance) { - let cur = instance.parent; - const exposedInstance = instance.proxy; - const errorInfo = ErrorTypeStrings$1[type] ; - while (cur) { - const errorCapturedHooks = cur.ec; - if (errorCapturedHooks) { - for (let i = 0; i < errorCapturedHooks.length; i++) { - if (errorCapturedHooks[i](err, exposedInstance, errorInfo) === false) { - return; - } - } - } - cur = cur.parent; - } - if (errorHandler) { - pauseTracking(); - callWithErrorHandling(errorHandler, null, 10, [ - err, - exposedInstance, - errorInfo - ]); - resetTracking(); - return; - } - } - logError(err, type, contextVNode, throwInDev, throwUnhandledErrorInProduction); -} -function logError(err, type, contextVNode, throwInDev = true, throwInProd = false) { - { - const info = ErrorTypeStrings$1[type]; - if (contextVNode) { - pushWarningContext(contextVNode); - } - warn$1(`Unhandled error${info ? ` during execution of ${info}` : ``}`); - if (contextVNode) { - popWarningContext(); - } - if (throwInDev) { - throw err; - } else { - console.error(err); - } - } -} - -const queue = []; -let flushIndex = -1; -const pendingPostFlushCbs = []; -let activePostFlushCbs = null; -let postFlushIndex = 0; -const resolvedPromise = /* @__PURE__ */ Promise.resolve(); -let currentFlushPromise = null; -const RECURSION_LIMIT = 100; -function nextTick(fn) { - const p = currentFlushPromise || resolvedPromise; - return fn ? p.then(this ? fn.bind(this) : fn) : p; -} -function findInsertionIndex(id) { - let start = flushIndex + 1; - let end = queue.length; - while (start < end) { - const middle = start + end >>> 1; - const middleJob = queue[middle]; - const middleJobId = getId(middleJob); - if (middleJobId < id || middleJobId === id && middleJob.flags & 2) { - start = middle + 1; - } else { - end = middle; - } - } - return start; -} -function queueJob(job) { - if (!(job.flags & 1)) { - const jobId = getId(job); - const lastJob = queue[queue.length - 1]; - if (!lastJob || // fast path when the job id is larger than the tail - !(job.flags & 2) && jobId >= getId(lastJob)) { - queue.push(job); - } else { - queue.splice(findInsertionIndex(jobId), 0, job); - } - job.flags |= 1; - queueFlush(); - } -} -function queueFlush() { - if (!currentFlushPromise) { - currentFlushPromise = resolvedPromise.then(flushJobs); - } -} -function queuePostFlushCb(cb) { - if (!isArray(cb)) { - if (activePostFlushCbs && cb.id === -1) { - activePostFlushCbs.splice(postFlushIndex + 1, 0, cb); - } else if (!(cb.flags & 1)) { - pendingPostFlushCbs.push(cb); - cb.flags |= 1; - } - } else { - pendingPostFlushCbs.push(...cb); - } - queueFlush(); -} -function flushPreFlushCbs(instance, seen, i = flushIndex + 1) { - { - seen = seen || /* @__PURE__ */ new Map(); - } - for (; i < queue.length; i++) { - const cb = queue[i]; - if (cb && cb.flags & 2) { - if (instance && cb.id !== instance.uid) { - continue; - } - if (checkRecursiveUpdates(seen, cb)) { - continue; - } - queue.splice(i, 1); - i--; - if (cb.flags & 4) { - cb.flags &= ~1; - } - cb(); - if (!(cb.flags & 4)) { - cb.flags &= ~1; - } - } - } -} -function flushPostFlushCbs(seen) { - if (pendingPostFlushCbs.length) { - const deduped = [...new Set(pendingPostFlushCbs)].sort( - (a, b) => getId(a) - getId(b) - ); - pendingPostFlushCbs.length = 0; - if (activePostFlushCbs) { - activePostFlushCbs.push(...deduped); - return; - } - activePostFlushCbs = deduped; - { - seen = seen || /* @__PURE__ */ new Map(); - } - for (postFlushIndex = 0; postFlushIndex < activePostFlushCbs.length; postFlushIndex++) { - const cb = activePostFlushCbs[postFlushIndex]; - if (checkRecursiveUpdates(seen, cb)) { - continue; - } - if (cb.flags & 4) { - cb.flags &= ~1; - } - if (!(cb.flags & 8)) cb(); - cb.flags &= ~1; - } - activePostFlushCbs = null; - postFlushIndex = 0; - } -} -const getId = (job) => job.id == null ? job.flags & 2 ? -1 : Infinity : job.id; -function flushJobs(seen) { - { - seen = seen || /* @__PURE__ */ new Map(); - } - const check = (job) => checkRecursiveUpdates(seen, job) ; - try { - for (flushIndex = 0; flushIndex < queue.length; flushIndex++) { - const job = queue[flushIndex]; - if (job && !(job.flags & 8)) { - if (check(job)) { - continue; - } - if (job.flags & 4) { - job.flags &= ~1; - } - callWithErrorHandling( - job, - job.i, - job.i ? 15 : 14 - ); - if (!(job.flags & 4)) { - job.flags &= ~1; - } - } - } - } finally { - for (; flushIndex < queue.length; flushIndex++) { - const job = queue[flushIndex]; - if (job) { - job.flags &= ~1; - } - } - flushIndex = -1; - queue.length = 0; - flushPostFlushCbs(seen); - currentFlushPromise = null; - if (queue.length || pendingPostFlushCbs.length) { - flushJobs(seen); - } - } -} -function checkRecursiveUpdates(seen, fn) { - const count = seen.get(fn) || 0; - if (count > RECURSION_LIMIT) { - const instance = fn.i; - const componentName = instance && getComponentName(instance.type); - handleError( - `Maximum recursive updates exceeded${componentName ? ` in component <${componentName}>` : ``}. This means you have a reactive effect that is mutating its own dependencies and thus recursively triggering itself. Possible sources include component template, render function, updated hook or watcher source function.`, - null, - 10 - ); - return true; - } - seen.set(fn, count + 1); - return false; -} - -let isHmrUpdating = false; -const hmrDirtyComponents = /* @__PURE__ */ new Map(); -{ - getGlobalThis().__VUE_HMR_RUNTIME__ = { - createRecord: tryWrap(createRecord), - rerender: tryWrap(rerender), - reload: tryWrap(reload) - }; -} -const map = /* @__PURE__ */ new Map(); -function registerHMR(instance) { - const id = instance.type.__hmrId; - let record = map.get(id); - if (!record) { - createRecord(id, instance.type); - record = map.get(id); - } - record.instances.add(instance); -} -function unregisterHMR(instance) { - map.get(instance.type.__hmrId).instances.delete(instance); -} -function createRecord(id, initialDef) { - if (map.has(id)) { - return false; - } - map.set(id, { - initialDef: normalizeClassComponent(initialDef), - instances: /* @__PURE__ */ new Set() - }); - return true; -} -function normalizeClassComponent(component) { - return isClassComponent(component) ? component.__vccOpts : component; -} -function rerender(id, newRender) { - const record = map.get(id); - if (!record) { - return; - } - record.initialDef.render = newRender; - [...record.instances].forEach((instance) => { - if (newRender) { - instance.render = newRender; - normalizeClassComponent(instance.type).render = newRender; - } - instance.renderCache = []; - isHmrUpdating = true; - instance.update(); - isHmrUpdating = false; - }); -} -function reload(id, newComp) { - const record = map.get(id); - if (!record) return; - newComp = normalizeClassComponent(newComp); - updateComponentDef(record.initialDef, newComp); - const instances = [...record.instances]; - for (let i = 0; i < instances.length; i++) { - const instance = instances[i]; - const oldComp = normalizeClassComponent(instance.type); - let dirtyInstances = hmrDirtyComponents.get(oldComp); - if (!dirtyInstances) { - if (oldComp !== record.initialDef) { - updateComponentDef(oldComp, newComp); - } - hmrDirtyComponents.set(oldComp, dirtyInstances = /* @__PURE__ */ new Set()); - } - dirtyInstances.add(instance); - instance.appContext.propsCache.delete(instance.type); - instance.appContext.emitsCache.delete(instance.type); - instance.appContext.optionsCache.delete(instance.type); - if (instance.ceReload) { - dirtyInstances.add(instance); - instance.ceReload(newComp.styles); - dirtyInstances.delete(instance); - } else if (instance.parent) { - queueJob(() => { - isHmrUpdating = true; - instance.parent.update(); - isHmrUpdating = false; - dirtyInstances.delete(instance); - }); - } else if (instance.appContext.reload) { - instance.appContext.reload(); - } else if (typeof window !== "undefined") { - window.location.reload(); - } else { - console.warn( - "[HMR] Root or manually mounted instance modified. Full reload required." - ); - } - if (instance.root.ce && instance !== instance.root) { - instance.root.ce._removeChildStyle(oldComp); - } - } - queuePostFlushCb(() => { - hmrDirtyComponents.clear(); - }); -} -function updateComponentDef(oldComp, newComp) { - extend(oldComp, newComp); - for (const key in oldComp) { - if (key !== "__file" && !(key in newComp)) { - delete oldComp[key]; - } - } -} -function tryWrap(fn) { - return (id, arg) => { - try { - return fn(id, arg); - } catch (e) { - console.error(e); - console.warn( - `[HMR] Something went wrong during Vue component hot-reload. Full reload required.` - ); - } - }; -} - -let devtools$1; -let buffer = []; -let devtoolsNotInstalled = false; -function emit$1(event, ...args) { - if (devtools$1) { - devtools$1.emit(event, ...args); - } else if (!devtoolsNotInstalled) { - buffer.push({ event, args }); - } -} -function setDevtoolsHook$1(hook, target) { - var _a, _b; - devtools$1 = hook; - if (devtools$1) { - devtools$1.enabled = true; - buffer.forEach(({ event, args }) => devtools$1.emit(event, ...args)); - buffer = []; - } else if ( - // handle late devtools injection - only do this if we are in an actual - // browser environment to avoid the timer handle stalling test runner exit - // (#4815) - typeof window !== "undefined" && // some envs mock window but not fully - window.HTMLElement && // also exclude jsdom - // eslint-disable-next-line no-restricted-syntax - !((_b = (_a = window.navigator) == null ? void 0 : _a.userAgent) == null ? void 0 : _b.includes("jsdom")) - ) { - const replay = target.__VUE_DEVTOOLS_HOOK_REPLAY__ = target.__VUE_DEVTOOLS_HOOK_REPLAY__ || []; - replay.push((newHook) => { - setDevtoolsHook$1(newHook, target); - }); - setTimeout(() => { - if (!devtools$1) { - target.__VUE_DEVTOOLS_HOOK_REPLAY__ = null; - devtoolsNotInstalled = true; - buffer = []; - } - }, 3e3); - } else { - devtoolsNotInstalled = true; - buffer = []; - } -} -function devtoolsInitApp(app, version) { - emit$1("app:init" /* APP_INIT */, app, version, { - Fragment, - Text, - Comment, - Static - }); -} -function devtoolsUnmountApp(app) { - emit$1("app:unmount" /* APP_UNMOUNT */, app); -} -const devtoolsComponentAdded = /* @__PURE__ */ createDevtoolsComponentHook("component:added" /* COMPONENT_ADDED */); -const devtoolsComponentUpdated = /* @__PURE__ */ createDevtoolsComponentHook("component:updated" /* COMPONENT_UPDATED */); -const _devtoolsComponentRemoved = /* @__PURE__ */ createDevtoolsComponentHook( - "component:removed" /* COMPONENT_REMOVED */ -); -const devtoolsComponentRemoved = (component) => { - if (devtools$1 && typeof devtools$1.cleanupBuffer === "function" && // remove the component if it wasn't buffered - !devtools$1.cleanupBuffer(component)) { - _devtoolsComponentRemoved(component); - } -}; -/*! #__NO_SIDE_EFFECTS__ */ -// @__NO_SIDE_EFFECTS__ -function createDevtoolsComponentHook(hook) { - return (component) => { - emit$1( - hook, - component.appContext.app, - component.uid, - component.parent ? component.parent.uid : void 0, - component - ); - }; -} -const devtoolsPerfStart = /* @__PURE__ */ createDevtoolsPerformanceHook("perf:start" /* PERFORMANCE_START */); -const devtoolsPerfEnd = /* @__PURE__ */ createDevtoolsPerformanceHook("perf:end" /* PERFORMANCE_END */); -function createDevtoolsPerformanceHook(hook) { - return (component, type, time) => { - emit$1(hook, component.appContext.app, component.uid, component, type, time); - }; -} -function devtoolsComponentEmit(component, event, params) { - emit$1( - "component:emit" /* COMPONENT_EMIT */, - component.appContext.app, - component, - event, - params - ); -} - -let currentRenderingInstance = null; -let currentScopeId = null; -function setCurrentRenderingInstance(instance) { - const prev = currentRenderingInstance; - currentRenderingInstance = instance; - currentScopeId = instance && instance.type.__scopeId || null; - return prev; -} -function pushScopeId(id) { - currentScopeId = id; -} -function popScopeId() { - currentScopeId = null; -} -const withScopeId = (_id) => withCtx; -function withCtx(fn, ctx = currentRenderingInstance, isNonScopedSlot) { - if (!ctx) return fn; - if (fn._n) { - return fn; - } - const renderFnWithContext = (...args) => { - if (renderFnWithContext._d) { - setBlockTracking(-1); - } - const prevInstance = setCurrentRenderingInstance(ctx); - let res; - try { - res = fn(...args); - } finally { - setCurrentRenderingInstance(prevInstance); - if (renderFnWithContext._d) { - setBlockTracking(1); - } - } - { - devtoolsComponentUpdated(ctx); - } - return res; - }; - renderFnWithContext._n = true; - renderFnWithContext._c = true; - renderFnWithContext._d = true; - return renderFnWithContext; -} - -function validateDirectiveName(name) { - if (isBuiltInDirective(name)) { - warn$1("Do not use built-in directive ids as custom directive id: " + name); - } -} -function withDirectives(vnode, directives) { - if (currentRenderingInstance === null) { - warn$1(`withDirectives can only be used inside render functions.`); - return vnode; - } - const instance = getComponentPublicInstance(currentRenderingInstance); - const bindings = vnode.dirs || (vnode.dirs = []); - for (let i = 0; i < directives.length; i++) { - let [dir, value, arg, modifiers = EMPTY_OBJ] = directives[i]; - if (dir) { - if (isFunction(dir)) { - dir = { - mounted: dir, - updated: dir - }; - } - if (dir.deep) { - traverse(value); - } - bindings.push({ - dir, - instance, - value, - oldValue: void 0, - arg, - modifiers - }); - } - } - return vnode; -} -function invokeDirectiveHook(vnode, prevVNode, instance, name) { - const bindings = vnode.dirs; - const oldBindings = prevVNode && prevVNode.dirs; - for (let i = 0; i < bindings.length; i++) { - const binding = bindings[i]; - if (oldBindings) { - binding.oldValue = oldBindings[i].value; - } - let hook = binding.dir[name]; - if (hook) { - pauseTracking(); - callWithAsyncErrorHandling(hook, instance, 8, [ - vnode.el, - binding, - vnode, - prevVNode - ]); - resetTracking(); - } - } -} - -const TeleportEndKey = Symbol("_vte"); -const isTeleport = (type) => type.__isTeleport; -const isTeleportDisabled = (props) => props && (props.disabled || props.disabled === ""); -const isTeleportDeferred = (props) => props && (props.defer || props.defer === ""); -const isTargetSVG = (target) => typeof SVGElement !== "undefined" && target instanceof SVGElement; -const isTargetMathML = (target) => typeof MathMLElement === "function" && target instanceof MathMLElement; -const resolveTarget = (props, select) => { - const targetSelector = props && props.to; - if (isString(targetSelector)) { - if (!select) { - warn$1( - `Current renderer does not support string target for Teleports. (missing querySelector renderer option)` - ); - return null; - } else { - const target = select(targetSelector); - if (!target && !isTeleportDisabled(props)) { - warn$1( - `Failed to locate Teleport target with selector "${targetSelector}". Note the target element must exist before the component is mounted - i.e. the target cannot be rendered by the component itself, and ideally should be outside of the entire Vue component tree.` - ); - } - return target; - } - } else { - if (!targetSelector && !isTeleportDisabled(props)) { - warn$1(`Invalid Teleport target: ${targetSelector}`); - } - return targetSelector; - } -}; -const TeleportImpl = { - name: "Teleport", - __isTeleport: true, - process(n1, n2, container, anchor, parentComponent, parentSuspense, namespace, slotScopeIds, optimized, internals) { - const { - mc: mountChildren, - pc: patchChildren, - pbc: patchBlockChildren, - o: { insert, querySelector, createText, createComment } - } = internals; - const disabled = isTeleportDisabled(n2.props); - let { shapeFlag, children, dynamicChildren } = n2; - if (isHmrUpdating) { - optimized = false; - dynamicChildren = null; - } - if (n1 == null) { - const placeholder = n2.el = createComment("teleport start") ; - const mainAnchor = n2.anchor = createComment("teleport end") ; - insert(placeholder, container, anchor); - insert(mainAnchor, container, anchor); - const mount = (container2, anchor2) => { - if (shapeFlag & 16) { - if (parentComponent && parentComponent.isCE) { - parentComponent.ce._teleportTarget = container2; - } - mountChildren( - children, - container2, - anchor2, - parentComponent, - parentSuspense, - namespace, - slotScopeIds, - optimized - ); - } - }; - const mountToTarget = () => { - const target = n2.target = resolveTarget(n2.props, querySelector); - const targetAnchor = prepareAnchor(target, n2, createText, insert); - if (target) { - if (namespace !== "svg" && isTargetSVG(target)) { - namespace = "svg"; - } else if (namespace !== "mathml" && isTargetMathML(target)) { - namespace = "mathml"; - } - if (!disabled) { - mount(target, targetAnchor); - updateCssVars(n2, false); - } - } else if (!disabled) { - warn$1( - "Invalid Teleport target on mount:", - target, - `(${typeof target})` - ); - } - }; - if (disabled) { - mount(container, mainAnchor); - updateCssVars(n2, true); - } - if (isTeleportDeferred(n2.props)) { - queuePostRenderEffect(mountToTarget, parentSuspense); - } else { - mountToTarget(); - } - } else { - n2.el = n1.el; - n2.targetStart = n1.targetStart; - const mainAnchor = n2.anchor = n1.anchor; - const target = n2.target = n1.target; - const targetAnchor = n2.targetAnchor = n1.targetAnchor; - const wasDisabled = isTeleportDisabled(n1.props); - const currentContainer = wasDisabled ? container : target; - const currentAnchor = wasDisabled ? mainAnchor : targetAnchor; - if (namespace === "svg" || isTargetSVG(target)) { - namespace = "svg"; - } else if (namespace === "mathml" || isTargetMathML(target)) { - namespace = "mathml"; - } - if (dynamicChildren) { - patchBlockChildren( - n1.dynamicChildren, - dynamicChildren, - currentContainer, - parentComponent, - parentSuspense, - namespace, - slotScopeIds - ); - traverseStaticChildren(n1, n2, true); - } else if (!optimized) { - patchChildren( - n1, - n2, - currentContainer, - currentAnchor, - parentComponent, - parentSuspense, - namespace, - slotScopeIds, - false - ); - } - if (disabled) { - if (!wasDisabled) { - moveTeleport( - n2, - container, - mainAnchor, - internals, - 1 - ); - } else { - if (n2.props && n1.props && n2.props.to !== n1.props.to) { - n2.props.to = n1.props.to; - } - } - } else { - if ((n2.props && n2.props.to) !== (n1.props && n1.props.to)) { - const nextTarget = n2.target = resolveTarget( - n2.props, - querySelector - ); - if (nextTarget) { - moveTeleport( - n2, - nextTarget, - null, - internals, - 0 - ); - } else { - warn$1( - "Invalid Teleport target on update:", - target, - `(${typeof target})` - ); - } - } else if (wasDisabled) { - moveTeleport( - n2, - target, - targetAnchor, - internals, - 1 - ); - } - } - updateCssVars(n2, disabled); - } - }, - remove(vnode, parentComponent, parentSuspense, { um: unmount, o: { remove: hostRemove } }, doRemove) { - const { - shapeFlag, - children, - anchor, - targetStart, - targetAnchor, - target, - props - } = vnode; - if (target) { - hostRemove(targetStart); - hostRemove(targetAnchor); - } - doRemove && hostRemove(anchor); - if (shapeFlag & 16) { - const shouldRemove = doRemove || !isTeleportDisabled(props); - for (let i = 0; i < children.length; i++) { - const child = children[i]; - unmount( - child, - parentComponent, - parentSuspense, - shouldRemove, - !!child.dynamicChildren - ); - } - } - }, - move: moveTeleport, - hydrate: hydrateTeleport -}; -function moveTeleport(vnode, container, parentAnchor, { o: { insert }, m: move }, moveType = 2) { - if (moveType === 0) { - insert(vnode.targetAnchor, container, parentAnchor); - } - const { el, anchor, shapeFlag, children, props } = vnode; - const isReorder = moveType === 2; - if (isReorder) { - insert(el, container, parentAnchor); - } - if (!isReorder || isTeleportDisabled(props)) { - if (shapeFlag & 16) { - for (let i = 0; i < children.length; i++) { - move( - children[i], - container, - parentAnchor, - 2 - ); - } - } - } - if (isReorder) { - insert(anchor, container, parentAnchor); - } -} -function hydrateTeleport(node, vnode, parentComponent, parentSuspense, slotScopeIds, optimized, { - o: { nextSibling, parentNode, querySelector, insert, createText } -}, hydrateChildren) { - const target = vnode.target = resolveTarget( - vnode.props, - querySelector - ); - if (target) { - const disabled = isTeleportDisabled(vnode.props); - const targetNode = target._lpa || target.firstChild; - if (vnode.shapeFlag & 16) { - if (disabled) { - vnode.anchor = hydrateChildren( - nextSibling(node), - vnode, - parentNode(node), - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - vnode.targetStart = targetNode; - vnode.targetAnchor = targetNode && nextSibling(targetNode); - } else { - vnode.anchor = nextSibling(node); - let targetAnchor = targetNode; - while (targetAnchor) { - if (targetAnchor && targetAnchor.nodeType === 8) { - if (targetAnchor.data === "teleport start anchor") { - vnode.targetStart = targetAnchor; - } else if (targetAnchor.data === "teleport anchor") { - vnode.targetAnchor = targetAnchor; - target._lpa = vnode.targetAnchor && nextSibling(vnode.targetAnchor); - break; - } - } - targetAnchor = nextSibling(targetAnchor); - } - if (!vnode.targetAnchor) { - prepareAnchor(target, vnode, createText, insert); - } - hydrateChildren( - targetNode && nextSibling(targetNode), - vnode, - target, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - } - } - updateCssVars(vnode, disabled); - } - return vnode.anchor && nextSibling(vnode.anchor); -} -const Teleport = TeleportImpl; -function updateCssVars(vnode, isDisabled) { - const ctx = vnode.ctx; - if (ctx && ctx.ut) { - let node, anchor; - if (isDisabled) { - node = vnode.el; - anchor = vnode.anchor; - } else { - node = vnode.targetStart; - anchor = vnode.targetAnchor; - } - while (node && node !== anchor) { - if (node.nodeType === 1) node.setAttribute("data-v-owner", ctx.uid); - node = node.nextSibling; - } - ctx.ut(); - } -} -function prepareAnchor(target, vnode, createText, insert) { - const targetStart = vnode.targetStart = createText(""); - const targetAnchor = vnode.targetAnchor = createText(""); - targetStart[TeleportEndKey] = targetAnchor; - if (target) { - insert(targetStart, target); - insert(targetAnchor, target); - } - return targetAnchor; -} - -const leaveCbKey = Symbol("_leaveCb"); -const enterCbKey$1 = Symbol("_enterCb"); -function useTransitionState() { - const state = { - isMounted: false, - isLeaving: false, - isUnmounting: false, - leavingVNodes: /* @__PURE__ */ new Map() - }; - onMounted(() => { - state.isMounted = true; - }); - onBeforeUnmount(() => { - state.isUnmounting = true; - }); - return state; -} -const TransitionHookValidator = [Function, Array]; -const BaseTransitionPropsValidators = { - mode: String, - appear: Boolean, - persisted: Boolean, - // enter - onBeforeEnter: TransitionHookValidator, - onEnter: TransitionHookValidator, - onAfterEnter: TransitionHookValidator, - onEnterCancelled: TransitionHookValidator, - // leave - onBeforeLeave: TransitionHookValidator, - onLeave: TransitionHookValidator, - onAfterLeave: TransitionHookValidator, - onLeaveCancelled: TransitionHookValidator, - // appear - onBeforeAppear: TransitionHookValidator, - onAppear: TransitionHookValidator, - onAfterAppear: TransitionHookValidator, - onAppearCancelled: TransitionHookValidator -}; -const recursiveGetSubtree = (instance) => { - const subTree = instance.subTree; - return subTree.component ? recursiveGetSubtree(subTree.component) : subTree; -}; -const BaseTransitionImpl = { - name: `BaseTransition`, - props: BaseTransitionPropsValidators, - setup(props, { slots }) { - const instance = getCurrentInstance(); - const state = useTransitionState(); - return () => { - const children = slots.default && getTransitionRawChildren(slots.default(), true); - if (!children || !children.length) { - return; - } - const child = findNonCommentChild(children); - const rawProps = toRaw(props); - const { mode } = rawProps; - if (mode && mode !== "in-out" && mode !== "out-in" && mode !== "default") { - warn$1(`invalid mode: ${mode}`); - } - if (state.isLeaving) { - return emptyPlaceholder(child); - } - const innerChild = getInnerChild$1(child); - if (!innerChild) { - return emptyPlaceholder(child); - } - let enterHooks = resolveTransitionHooks( - innerChild, - rawProps, - state, - instance, - // #11061, ensure enterHooks is fresh after clone - (hooks) => enterHooks = hooks - ); - if (innerChild.type !== Comment) { - setTransitionHooks(innerChild, enterHooks); - } - const oldChild = instance.subTree; - const oldInnerChild = oldChild && getInnerChild$1(oldChild); - if (oldInnerChild && oldInnerChild.type !== Comment && !isSameVNodeType(innerChild, oldInnerChild) && recursiveGetSubtree(instance).type !== Comment) { - const leavingHooks = resolveTransitionHooks( - oldInnerChild, - rawProps, - state, - instance - ); - setTransitionHooks(oldInnerChild, leavingHooks); - if (mode === "out-in" && innerChild.type !== Comment) { - state.isLeaving = true; - leavingHooks.afterLeave = () => { - state.isLeaving = false; - if (!(instance.job.flags & 8)) { - instance.update(); - } - delete leavingHooks.afterLeave; - }; - return emptyPlaceholder(child); - } else if (mode === "in-out" && innerChild.type !== Comment) { - leavingHooks.delayLeave = (el, earlyRemove, delayedLeave) => { - const leavingVNodesCache = getLeavingNodesForType( - state, - oldInnerChild - ); - leavingVNodesCache[String(oldInnerChild.key)] = oldInnerChild; - el[leaveCbKey] = () => { - earlyRemove(); - el[leaveCbKey] = void 0; - delete enterHooks.delayedLeave; - }; - enterHooks.delayedLeave = delayedLeave; - }; - } - } - return child; - }; - } -}; -function findNonCommentChild(children) { - let child = children[0]; - if (children.length > 1) { - let hasFound = false; - for (const c of children) { - if (c.type !== Comment) { - if (hasFound) { - warn$1( - " can only be used on a single element or component. Use for lists." - ); - break; - } - child = c; - hasFound = true; - } - } - } - return child; -} -const BaseTransition = BaseTransitionImpl; -function getLeavingNodesForType(state, vnode) { - const { leavingVNodes } = state; - let leavingVNodesCache = leavingVNodes.get(vnode.type); - if (!leavingVNodesCache) { - leavingVNodesCache = /* @__PURE__ */ Object.create(null); - leavingVNodes.set(vnode.type, leavingVNodesCache); - } - return leavingVNodesCache; -} -function resolveTransitionHooks(vnode, props, state, instance, postClone) { - const { - appear, - mode, - persisted = false, - onBeforeEnter, - onEnter, - onAfterEnter, - onEnterCancelled, - onBeforeLeave, - onLeave, - onAfterLeave, - onLeaveCancelled, - onBeforeAppear, - onAppear, - onAfterAppear, - onAppearCancelled - } = props; - const key = String(vnode.key); - const leavingVNodesCache = getLeavingNodesForType(state, vnode); - const callHook = (hook, args) => { - hook && callWithAsyncErrorHandling( - hook, - instance, - 9, - args - ); - }; - const callAsyncHook = (hook, args) => { - const done = args[1]; - callHook(hook, args); - if (isArray(hook)) { - if (hook.every((hook2) => hook2.length <= 1)) done(); - } else if (hook.length <= 1) { - done(); - } - }; - const hooks = { - mode, - persisted, - beforeEnter(el) { - let hook = onBeforeEnter; - if (!state.isMounted) { - if (appear) { - hook = onBeforeAppear || onBeforeEnter; - } else { - return; - } - } - if (el[leaveCbKey]) { - el[leaveCbKey]( - true - /* cancelled */ - ); - } - const leavingVNode = leavingVNodesCache[key]; - if (leavingVNode && isSameVNodeType(vnode, leavingVNode) && leavingVNode.el[leaveCbKey]) { - leavingVNode.el[leaveCbKey](); - } - callHook(hook, [el]); - }, - enter(el) { - let hook = onEnter; - let afterHook = onAfterEnter; - let cancelHook = onEnterCancelled; - if (!state.isMounted) { - if (appear) { - hook = onAppear || onEnter; - afterHook = onAfterAppear || onAfterEnter; - cancelHook = onAppearCancelled || onEnterCancelled; - } else { - return; - } - } - let called = false; - const done = el[enterCbKey$1] = (cancelled) => { - if (called) return; - called = true; - if (cancelled) { - callHook(cancelHook, [el]); - } else { - callHook(afterHook, [el]); - } - if (hooks.delayedLeave) { - hooks.delayedLeave(); - } - el[enterCbKey$1] = void 0; - }; - if (hook) { - callAsyncHook(hook, [el, done]); - } else { - done(); - } - }, - leave(el, remove) { - const key2 = String(vnode.key); - if (el[enterCbKey$1]) { - el[enterCbKey$1]( - true - /* cancelled */ - ); - } - if (state.isUnmounting) { - return remove(); - } - callHook(onBeforeLeave, [el]); - let called = false; - const done = el[leaveCbKey] = (cancelled) => { - if (called) return; - called = true; - remove(); - if (cancelled) { - callHook(onLeaveCancelled, [el]); - } else { - callHook(onAfterLeave, [el]); - } - el[leaveCbKey] = void 0; - if (leavingVNodesCache[key2] === vnode) { - delete leavingVNodesCache[key2]; - } - }; - leavingVNodesCache[key2] = vnode; - if (onLeave) { - callAsyncHook(onLeave, [el, done]); - } else { - done(); - } - }, - clone(vnode2) { - const hooks2 = resolveTransitionHooks( - vnode2, - props, - state, - instance, - postClone - ); - if (postClone) postClone(hooks2); - return hooks2; - } - }; - return hooks; -} -function emptyPlaceholder(vnode) { - if (isKeepAlive(vnode)) { - vnode = cloneVNode(vnode); - vnode.children = null; - return vnode; - } -} -function getInnerChild$1(vnode) { - if (!isKeepAlive(vnode)) { - if (isTeleport(vnode.type) && vnode.children) { - return findNonCommentChild(vnode.children); - } - return vnode; - } - if (vnode.component) { - return vnode.component.subTree; - } - const { shapeFlag, children } = vnode; - if (children) { - if (shapeFlag & 16) { - return children[0]; - } - if (shapeFlag & 32 && isFunction(children.default)) { - return children.default(); - } - } -} -function setTransitionHooks(vnode, hooks) { - if (vnode.shapeFlag & 6 && vnode.component) { - vnode.transition = hooks; - setTransitionHooks(vnode.component.subTree, hooks); - } else if (vnode.shapeFlag & 128) { - vnode.ssContent.transition = hooks.clone(vnode.ssContent); - vnode.ssFallback.transition = hooks.clone(vnode.ssFallback); - } else { - vnode.transition = hooks; - } -} -function getTransitionRawChildren(children, keepComment = false, parentKey) { - let ret = []; - let keyedFragmentCount = 0; - for (let i = 0; i < children.length; i++) { - let child = children[i]; - const key = parentKey == null ? child.key : String(parentKey) + String(child.key != null ? child.key : i); - if (child.type === Fragment) { - if (child.patchFlag & 128) keyedFragmentCount++; - ret = ret.concat( - getTransitionRawChildren(child.children, keepComment, key) - ); - } else if (keepComment || child.type !== Comment) { - ret.push(key != null ? cloneVNode(child, { key }) : child); - } - } - if (keyedFragmentCount > 1) { - for (let i = 0; i < ret.length; i++) { - ret[i].patchFlag = -2; - } - } - return ret; -} - -/*! #__NO_SIDE_EFFECTS__ */ -// @__NO_SIDE_EFFECTS__ -function defineComponent(options, extraOptions) { - return isFunction(options) ? ( - // #8236: extend call and options.name access are considered side-effects - // by Rollup, so we have to wrap it in a pure-annotated IIFE. - /* @__PURE__ */ (() => extend({ name: options.name }, extraOptions, { setup: options }))() - ) : options; -} - -function useId() { - const i = getCurrentInstance(); - if (i) { - return (i.appContext.config.idPrefix || "v") + "-" + i.ids[0] + i.ids[1]++; - } else { - warn$1( - `useId() is called when there is no active component instance to be associated with.` - ); - } - return ""; -} -function markAsyncBoundary(instance) { - instance.ids = [instance.ids[0] + instance.ids[2]++ + "-", 0, 0]; -} - -const knownTemplateRefs = /* @__PURE__ */ new WeakSet(); -function useTemplateRef(key) { - const i = getCurrentInstance(); - const r = shallowRef(null); - if (i) { - const refs = i.refs === EMPTY_OBJ ? i.refs = {} : i.refs; - let desc; - if ((desc = Object.getOwnPropertyDescriptor(refs, key)) && !desc.configurable) { - warn$1(`useTemplateRef('${key}') already exists.`); - } else { - Object.defineProperty(refs, key, { - enumerable: true, - get: () => r.value, - set: (val) => r.value = val - }); - } - } else { - warn$1( - `useTemplateRef() is called when there is no active component instance to be associated with.` - ); - } - const ret = readonly(r) ; - { - knownTemplateRefs.add(ret); - } - return ret; -} - -function setRef(rawRef, oldRawRef, parentSuspense, vnode, isUnmount = false) { - if (isArray(rawRef)) { - rawRef.forEach( - (r, i) => setRef( - r, - oldRawRef && (isArray(oldRawRef) ? oldRawRef[i] : oldRawRef), - parentSuspense, - vnode, - isUnmount - ) - ); - return; - } - if (isAsyncWrapper(vnode) && !isUnmount) { - return; - } - const refValue = vnode.shapeFlag & 4 ? getComponentPublicInstance(vnode.component) : vnode.el; - const value = isUnmount ? null : refValue; - const { i: owner, r: ref } = rawRef; - if (!owner) { - warn$1( - `Missing ref owner context. ref cannot be used on hoisted vnodes. A vnode with ref must be created inside the render function.` - ); - return; - } - const oldRef = oldRawRef && oldRawRef.r; - const refs = owner.refs === EMPTY_OBJ ? owner.refs = {} : owner.refs; - const setupState = owner.setupState; - const rawSetupState = toRaw(setupState); - const canSetSetupRef = setupState === EMPTY_OBJ ? () => false : (key) => { - { - if (hasOwn(rawSetupState, key) && !isRef(rawSetupState[key])) { - warn$1( - `Template ref "${key}" used on a non-ref value. It will not work in the production build.` - ); - } - if (knownTemplateRefs.has(rawSetupState[key])) { - return false; - } - } - return hasOwn(rawSetupState, key); - }; - if (oldRef != null && oldRef !== ref) { - if (isString(oldRef)) { - refs[oldRef] = null; - if (canSetSetupRef(oldRef)) { - setupState[oldRef] = null; - } - } else if (isRef(oldRef)) { - oldRef.value = null; - } - } - if (isFunction(ref)) { - callWithErrorHandling(ref, owner, 12, [value, refs]); - } else { - const _isString = isString(ref); - const _isRef = isRef(ref); - if (_isString || _isRef) { - const doSet = () => { - if (rawRef.f) { - const existing = _isString ? canSetSetupRef(ref) ? setupState[ref] : refs[ref] : ref.value; - if (isUnmount) { - isArray(existing) && remove(existing, refValue); - } else { - if (!isArray(existing)) { - if (_isString) { - refs[ref] = [refValue]; - if (canSetSetupRef(ref)) { - setupState[ref] = refs[ref]; - } - } else { - ref.value = [refValue]; - if (rawRef.k) refs[rawRef.k] = ref.value; - } - } else if (!existing.includes(refValue)) { - existing.push(refValue); - } - } - } else if (_isString) { - refs[ref] = value; - if (canSetSetupRef(ref)) { - setupState[ref] = value; - } - } else if (_isRef) { - ref.value = value; - if (rawRef.k) refs[rawRef.k] = value; - } else { - warn$1("Invalid template ref type:", ref, `(${typeof ref})`); - } - }; - if (value) { - doSet.id = -1; - queuePostRenderEffect(doSet, parentSuspense); - } else { - doSet(); - } - } else { - warn$1("Invalid template ref type:", ref, `(${typeof ref})`); - } - } -} - -let hasLoggedMismatchError = false; -const logMismatchError = () => { - if (hasLoggedMismatchError) { - return; - } - console.error("Hydration completed but contains mismatches."); - hasLoggedMismatchError = true; -}; -const isSVGContainer = (container) => container.namespaceURI.includes("svg") && container.tagName !== "foreignObject"; -const isMathMLContainer = (container) => container.namespaceURI.includes("MathML"); -const getContainerType = (container) => { - if (container.nodeType !== 1) return void 0; - if (isSVGContainer(container)) return "svg"; - if (isMathMLContainer(container)) return "mathml"; - return void 0; -}; -const isComment = (node) => node.nodeType === 8; -function createHydrationFunctions(rendererInternals) { - const { - mt: mountComponent, - p: patch, - o: { - patchProp, - createText, - nextSibling, - parentNode, - remove, - insert, - createComment - } - } = rendererInternals; - const hydrate = (vnode, container) => { - if (!container.hasChildNodes()) { - warn$1( - `Attempting to hydrate existing markup but container is empty. Performing full mount instead.` - ); - patch(null, vnode, container); - flushPostFlushCbs(); - container._vnode = vnode; - return; - } - hydrateNode(container.firstChild, vnode, null, null, null); - flushPostFlushCbs(); - container._vnode = vnode; - }; - const hydrateNode = (node, vnode, parentComponent, parentSuspense, slotScopeIds, optimized = false) => { - optimized = optimized || !!vnode.dynamicChildren; - const isFragmentStart = isComment(node) && node.data === "["; - const onMismatch = () => handleMismatch( - node, - vnode, - parentComponent, - parentSuspense, - slotScopeIds, - isFragmentStart - ); - const { type, ref, shapeFlag, patchFlag } = vnode; - let domType = node.nodeType; - vnode.el = node; - { - def(node, "__vnode", vnode, true); - def(node, "__vueParentComponent", parentComponent, true); - } - if (patchFlag === -2) { - optimized = false; - vnode.dynamicChildren = null; - } - let nextNode = null; - switch (type) { - case Text: - if (domType !== 3) { - if (vnode.children === "") { - insert(vnode.el = createText(""), parentNode(node), node); - nextNode = node; - } else { - nextNode = onMismatch(); - } - } else { - if (node.data !== vnode.children) { - warn$1( - `Hydration text mismatch in`, - node.parentNode, - ` - - rendered on server: ${JSON.stringify( - node.data - )} - - expected on client: ${JSON.stringify(vnode.children)}` - ); - logMismatchError(); - node.data = vnode.children; - } - nextNode = nextSibling(node); - } - break; - case Comment: - if (isTemplateNode(node)) { - nextNode = nextSibling(node); - replaceNode( - vnode.el = node.content.firstChild, - node, - parentComponent - ); - } else if (domType !== 8 || isFragmentStart) { - nextNode = onMismatch(); - } else { - nextNode = nextSibling(node); - } - break; - case Static: - if (isFragmentStart) { - node = nextSibling(node); - domType = node.nodeType; - } - if (domType === 1 || domType === 3) { - nextNode = node; - const needToAdoptContent = !vnode.children.length; - for (let i = 0; i < vnode.staticCount; i++) { - if (needToAdoptContent) - vnode.children += nextNode.nodeType === 1 ? nextNode.outerHTML : nextNode.data; - if (i === vnode.staticCount - 1) { - vnode.anchor = nextNode; - } - nextNode = nextSibling(nextNode); - } - return isFragmentStart ? nextSibling(nextNode) : nextNode; - } else { - onMismatch(); - } - break; - case Fragment: - if (!isFragmentStart) { - nextNode = onMismatch(); - } else { - nextNode = hydrateFragment( - node, - vnode, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - } - break; - default: - if (shapeFlag & 1) { - if ((domType !== 1 || vnode.type.toLowerCase() !== node.tagName.toLowerCase()) && !isTemplateNode(node)) { - nextNode = onMismatch(); - } else { - nextNode = hydrateElement( - node, - vnode, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - } - } else if (shapeFlag & 6) { - vnode.slotScopeIds = slotScopeIds; - const container = parentNode(node); - if (isFragmentStart) { - nextNode = locateClosingAnchor(node); - } else if (isComment(node) && node.data === "teleport start") { - nextNode = locateClosingAnchor(node, node.data, "teleport end"); - } else { - nextNode = nextSibling(node); - } - mountComponent( - vnode, - container, - null, - parentComponent, - parentSuspense, - getContainerType(container), - optimized - ); - if (isAsyncWrapper(vnode)) { - let subTree; - if (isFragmentStart) { - subTree = createVNode(Fragment); - subTree.anchor = nextNode ? nextNode.previousSibling : container.lastChild; - } else { - subTree = node.nodeType === 3 ? createTextVNode("") : createVNode("div"); - } - subTree.el = node; - vnode.component.subTree = subTree; - } - } else if (shapeFlag & 64) { - if (domType !== 8) { - nextNode = onMismatch(); - } else { - nextNode = vnode.type.hydrate( - node, - vnode, - parentComponent, - parentSuspense, - slotScopeIds, - optimized, - rendererInternals, - hydrateChildren - ); - } - } else if (shapeFlag & 128) { - nextNode = vnode.type.hydrate( - node, - vnode, - parentComponent, - parentSuspense, - getContainerType(parentNode(node)), - slotScopeIds, - optimized, - rendererInternals, - hydrateNode - ); - } else { - warn$1("Invalid HostVNode type:", type, `(${typeof type})`); - } - } - if (ref != null) { - setRef(ref, null, parentSuspense, vnode); - } - return nextNode; - }; - const hydrateElement = (el, vnode, parentComponent, parentSuspense, slotScopeIds, optimized) => { - optimized = optimized || !!vnode.dynamicChildren; - const { type, props, patchFlag, shapeFlag, dirs, transition } = vnode; - const forcePatch = type === "input" || type === "option"; - { - if (dirs) { - invokeDirectiveHook(vnode, null, parentComponent, "created"); - } - let needCallTransitionHooks = false; - if (isTemplateNode(el)) { - needCallTransitionHooks = needTransition( - null, - // no need check parentSuspense in hydration - transition - ) && parentComponent && parentComponent.vnode.props && parentComponent.vnode.props.appear; - const content = el.content.firstChild; - if (needCallTransitionHooks) { - transition.beforeEnter(content); - } - replaceNode(content, el, parentComponent); - vnode.el = el = content; - } - if (shapeFlag & 16 && // skip if element has innerHTML / textContent - !(props && (props.innerHTML || props.textContent))) { - let next = hydrateChildren( - el.firstChild, - vnode, - el, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - let hasWarned = false; - while (next) { - if (!isMismatchAllowed(el, 1 /* CHILDREN */)) { - if (!hasWarned) { - warn$1( - `Hydration children mismatch on`, - el, - ` -Server rendered element contains more child nodes than client vdom.` - ); - hasWarned = true; - } - logMismatchError(); - } - const cur = next; - next = next.nextSibling; - remove(cur); - } - } else if (shapeFlag & 8) { - let clientText = vnode.children; - if (clientText[0] === "\n" && (el.tagName === "PRE" || el.tagName === "TEXTAREA")) { - clientText = clientText.slice(1); - } - if (el.textContent !== clientText) { - if (!isMismatchAllowed(el, 0 /* TEXT */)) { - warn$1( - `Hydration text content mismatch on`, - el, - ` - - rendered on server: ${el.textContent} - - expected on client: ${vnode.children}` - ); - logMismatchError(); - } - el.textContent = vnode.children; - } - } - if (props) { - { - const isCustomElement = el.tagName.includes("-"); - for (const key in props) { - if (// #11189 skip if this node has directives that have created hooks - // as it could have mutated the DOM in any possible way - !(dirs && dirs.some((d) => d.dir.created)) && propHasMismatch(el, key, props[key], vnode, parentComponent)) { - logMismatchError(); - } - if (forcePatch && (key.endsWith("value") || key === "indeterminate") || isOn(key) && !isReservedProp(key) || // force hydrate v-bind with .prop modifiers - key[0] === "." || isCustomElement) { - patchProp(el, key, null, props[key], void 0, parentComponent); - } - } - } - } - let vnodeHooks; - if (vnodeHooks = props && props.onVnodeBeforeMount) { - invokeVNodeHook(vnodeHooks, parentComponent, vnode); - } - if (dirs) { - invokeDirectiveHook(vnode, null, parentComponent, "beforeMount"); - } - if ((vnodeHooks = props && props.onVnodeMounted) || dirs || needCallTransitionHooks) { - queueEffectWithSuspense(() => { - vnodeHooks && invokeVNodeHook(vnodeHooks, parentComponent, vnode); - needCallTransitionHooks && transition.enter(el); - dirs && invokeDirectiveHook(vnode, null, parentComponent, "mounted"); - }, parentSuspense); - } - } - return el.nextSibling; - }; - const hydrateChildren = (node, parentVNode, container, parentComponent, parentSuspense, slotScopeIds, optimized) => { - optimized = optimized || !!parentVNode.dynamicChildren; - const children = parentVNode.children; - const l = children.length; - let hasWarned = false; - for (let i = 0; i < l; i++) { - const vnode = optimized ? children[i] : children[i] = normalizeVNode(children[i]); - const isText = vnode.type === Text; - if (node) { - if (isText && !optimized) { - if (i + 1 < l && normalizeVNode(children[i + 1]).type === Text) { - insert( - createText( - node.data.slice(vnode.children.length) - ), - container, - nextSibling(node) - ); - node.data = vnode.children; - } - } - node = hydrateNode( - node, - vnode, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - } else if (isText && !vnode.children) { - insert(vnode.el = createText(""), container); - } else { - if (!isMismatchAllowed(container, 1 /* CHILDREN */)) { - if (!hasWarned) { - warn$1( - `Hydration children mismatch on`, - container, - ` -Server rendered element contains fewer child nodes than client vdom.` - ); - hasWarned = true; - } - logMismatchError(); - } - patch( - null, - vnode, - container, - null, - parentComponent, - parentSuspense, - getContainerType(container), - slotScopeIds - ); - } - } - return node; - }; - const hydrateFragment = (node, vnode, parentComponent, parentSuspense, slotScopeIds, optimized) => { - const { slotScopeIds: fragmentSlotScopeIds } = vnode; - if (fragmentSlotScopeIds) { - slotScopeIds = slotScopeIds ? slotScopeIds.concat(fragmentSlotScopeIds) : fragmentSlotScopeIds; - } - const container = parentNode(node); - const next = hydrateChildren( - nextSibling(node), - vnode, - container, - parentComponent, - parentSuspense, - slotScopeIds, - optimized - ); - if (next && isComment(next) && next.data === "]") { - return nextSibling(vnode.anchor = next); - } else { - logMismatchError(); - insert(vnode.anchor = createComment(`]`), container, next); - return next; - } - }; - const handleMismatch = (node, vnode, parentComponent, parentSuspense, slotScopeIds, isFragment) => { - if (!isMismatchAllowed(node.parentElement, 1 /* CHILDREN */)) { - warn$1( - `Hydration node mismatch: -- rendered on server:`, - node, - node.nodeType === 3 ? `(text)` : isComment(node) && node.data === "[" ? `(start of fragment)` : ``, - ` -- expected on client:`, - vnode.type - ); - logMismatchError(); - } - vnode.el = null; - if (isFragment) { - const end = locateClosingAnchor(node); - while (true) { - const next2 = nextSibling(node); - if (next2 && next2 !== end) { - remove(next2); - } else { - break; - } - } - } - const next = nextSibling(node); - const container = parentNode(node); - remove(node); - patch( - null, - vnode, - container, - next, - parentComponent, - parentSuspense, - getContainerType(container), - slotScopeIds - ); - return next; - }; - const locateClosingAnchor = (node, open = "[", close = "]") => { - let match = 0; - while (node) { - node = nextSibling(node); - if (node && isComment(node)) { - if (node.data === open) match++; - if (node.data === close) { - if (match === 0) { - return nextSibling(node); - } else { - match--; - } - } - } - } - return node; - }; - const replaceNode = (newNode, oldNode, parentComponent) => { - const parentNode2 = oldNode.parentNode; - if (parentNode2) { - parentNode2.replaceChild(newNode, oldNode); - } - let parent = parentComponent; - while (parent) { - if (parent.vnode.el === oldNode) { - parent.vnode.el = parent.subTree.el = newNode; - } - parent = parent.parent; - } - }; - const isTemplateNode = (node) => { - return node.nodeType === 1 && node.tagName === "TEMPLATE"; - }; - return [hydrate, hydrateNode]; -} -function propHasMismatch(el, key, clientValue, vnode, instance) { - let mismatchType; - let mismatchKey; - let actual; - let expected; - if (key === "class") { - actual = el.getAttribute("class"); - expected = normalizeClass(clientValue); - if (!isSetEqual(toClassSet(actual || ""), toClassSet(expected))) { - mismatchType = 2 /* CLASS */; - mismatchKey = `class`; - } - } else if (key === "style") { - actual = el.getAttribute("style") || ""; - expected = isString(clientValue) ? clientValue : stringifyStyle(normalizeStyle(clientValue)); - const actualMap = toStyleMap(actual); - const expectedMap = toStyleMap(expected); - if (vnode.dirs) { - for (const { dir, value } of vnode.dirs) { - if (dir.name === "show" && !value) { - expectedMap.set("display", "none"); - } - } - } - if (instance) { - resolveCssVars(instance, vnode, expectedMap); - } - if (!isMapEqual(actualMap, expectedMap)) { - mismatchType = 3 /* STYLE */; - mismatchKey = "style"; - } - } else if (el instanceof SVGElement && isKnownSvgAttr(key) || el instanceof HTMLElement && (isBooleanAttr(key) || isKnownHtmlAttr(key))) { - if (isBooleanAttr(key)) { - actual = el.hasAttribute(key); - expected = includeBooleanAttr(clientValue); - } else if (clientValue == null) { - actual = el.hasAttribute(key); - expected = false; - } else { - if (el.hasAttribute(key)) { - actual = el.getAttribute(key); - } else if (key === "value" && el.tagName === "TEXTAREA") { - actual = el.value; - } else { - actual = false; - } - expected = isRenderableAttrValue(clientValue) ? String(clientValue) : false; - } - if (actual !== expected) { - mismatchType = 4 /* ATTRIBUTE */; - mismatchKey = key; - } - } - if (mismatchType != null && !isMismatchAllowed(el, mismatchType)) { - const format = (v) => v === false ? `(not rendered)` : `${mismatchKey}="${v}"`; - const preSegment = `Hydration ${MismatchTypeString[mismatchType]} mismatch on`; - const postSegment = ` - - rendered on server: ${format(actual)} - - expected on client: ${format(expected)} - Note: this mismatch is check-only. The DOM will not be rectified in production due to performance overhead. - You should fix the source of the mismatch.`; - { - warn$1(preSegment, el, postSegment); - } - return true; - } - return false; -} -function toClassSet(str) { - return new Set(str.trim().split(/\s+/)); -} -function isSetEqual(a, b) { - if (a.size !== b.size) { - return false; - } - for (const s of a) { - if (!b.has(s)) { - return false; - } - } - return true; -} -function toStyleMap(str) { - const styleMap = /* @__PURE__ */ new Map(); - for (const item of str.split(";")) { - let [key, value] = item.split(":"); - key = key.trim(); - value = value && value.trim(); - if (key && value) { - styleMap.set(key, value); - } - } - return styleMap; -} -function isMapEqual(a, b) { - if (a.size !== b.size) { - return false; - } - for (const [key, value] of a) { - if (value !== b.get(key)) { - return false; - } - } - return true; -} -function resolveCssVars(instance, vnode, expectedMap) { - const root = instance.subTree; - if (instance.getCssVars && (vnode === root || root && root.type === Fragment && root.children.includes(vnode))) { - const cssVars = instance.getCssVars(); - for (const key in cssVars) { - expectedMap.set( - `--${getEscapedCssVarName(key)}`, - String(cssVars[key]) - ); - } - } - if (vnode === root && instance.parent) { - resolveCssVars(instance.parent, instance.vnode, expectedMap); - } -} -const allowMismatchAttr = "data-allow-mismatch"; -const MismatchTypeString = { - [0 /* TEXT */]: "text", - [1 /* CHILDREN */]: "children", - [2 /* CLASS */]: "class", - [3 /* STYLE */]: "style", - [4 /* ATTRIBUTE */]: "attribute" -}; -function isMismatchAllowed(el, allowedType) { - if (allowedType === 0 /* TEXT */ || allowedType === 1 /* CHILDREN */) { - while (el && !el.hasAttribute(allowMismatchAttr)) { - el = el.parentElement; - } - } - const allowedAttr = el && el.getAttribute(allowMismatchAttr); - if (allowedAttr == null) { - return false; - } else if (allowedAttr === "") { - return true; - } else { - const list = allowedAttr.split(","); - if (allowedType === 0 /* TEXT */ && list.includes("children")) { - return true; - } - return allowedAttr.split(",").includes(MismatchTypeString[allowedType]); - } -} - -const requestIdleCallback = getGlobalThis().requestIdleCallback || ((cb) => setTimeout(cb, 1)); -const cancelIdleCallback = getGlobalThis().cancelIdleCallback || ((id) => clearTimeout(id)); -const hydrateOnIdle = (timeout = 1e4) => (hydrate) => { - const id = requestIdleCallback(hydrate, { timeout }); - return () => cancelIdleCallback(id); -}; -function elementIsVisibleInViewport(el) { - const { top, left, bottom, right } = el.getBoundingClientRect(); - const { innerHeight, innerWidth } = window; - return (top > 0 && top < innerHeight || bottom > 0 && bottom < innerHeight) && (left > 0 && left < innerWidth || right > 0 && right < innerWidth); -} -const hydrateOnVisible = (opts) => (hydrate, forEach) => { - const ob = new IntersectionObserver((entries) => { - for (const e of entries) { - if (!e.isIntersecting) continue; - ob.disconnect(); - hydrate(); - break; - } - }, opts); - forEach((el) => { - if (!(el instanceof Element)) return; - if (elementIsVisibleInViewport(el)) { - hydrate(); - ob.disconnect(); - return false; - } - ob.observe(el); - }); - return () => ob.disconnect(); -}; -const hydrateOnMediaQuery = (query) => (hydrate) => { - if (query) { - const mql = matchMedia(query); - if (mql.matches) { - hydrate(); - } else { - mql.addEventListener("change", hydrate, { once: true }); - return () => mql.removeEventListener("change", hydrate); - } - } -}; -const hydrateOnInteraction = (interactions = []) => (hydrate, forEach) => { - if (isString(interactions)) interactions = [interactions]; - let hasHydrated = false; - const doHydrate = (e) => { - if (!hasHydrated) { - hasHydrated = true; - teardown(); - hydrate(); - e.target.dispatchEvent(new e.constructor(e.type, e)); - } - }; - const teardown = () => { - forEach((el) => { - for (const i of interactions) { - el.removeEventListener(i, doHydrate); - } - }); - }; - forEach((el) => { - for (const i of interactions) { - el.addEventListener(i, doHydrate, { once: true }); - } - }); - return teardown; -}; -function forEachElement(node, cb) { - if (isComment(node) && node.data === "[") { - let depth = 1; - let next = node.nextSibling; - while (next) { - if (next.nodeType === 1) { - const result = cb(next); - if (result === false) { - break; - } - } else if (isComment(next)) { - if (next.data === "]") { - if (--depth === 0) break; - } else if (next.data === "[") { - depth++; - } - } - next = next.nextSibling; - } - } else { - cb(node); - } -} - -const isAsyncWrapper = (i) => !!i.type.__asyncLoader; -/*! #__NO_SIDE_EFFECTS__ */ -// @__NO_SIDE_EFFECTS__ -function defineAsyncComponent(source) { - if (isFunction(source)) { - source = { loader: source }; - } - const { - loader, - loadingComponent, - errorComponent, - delay = 200, - hydrate: hydrateStrategy, - timeout, - // undefined = never times out - suspensible = true, - onError: userOnError - } = source; - let pendingRequest = null; - let resolvedComp; - let retries = 0; - const retry = () => { - retries++; - pendingRequest = null; - return load(); - }; - const load = () => { - let thisRequest; - return pendingRequest || (thisRequest = pendingRequest = loader().catch((err) => { - err = err instanceof Error ? err : new Error(String(err)); - if (userOnError) { - return new Promise((resolve, reject) => { - const userRetry = () => resolve(retry()); - const userFail = () => reject(err); - userOnError(err, userRetry, userFail, retries + 1); - }); - } else { - throw err; - } - }).then((comp) => { - if (thisRequest !== pendingRequest && pendingRequest) { - return pendingRequest; - } - if (!comp) { - warn$1( - `Async component loader resolved to undefined. If you are using retry(), make sure to return its return value.` - ); - } - if (comp && (comp.__esModule || comp[Symbol.toStringTag] === "Module")) { - comp = comp.default; - } - if (comp && !isObject(comp) && !isFunction(comp)) { - throw new Error(`Invalid async component load result: ${comp}`); - } - resolvedComp = comp; - return comp; - })); - }; - return defineComponent({ - name: "AsyncComponentWrapper", - __asyncLoader: load, - __asyncHydrate(el, instance, hydrate) { - const doHydrate = hydrateStrategy ? () => { - const teardown = hydrateStrategy( - hydrate, - (cb) => forEachElement(el, cb) - ); - if (teardown) { - (instance.bum || (instance.bum = [])).push(teardown); - } - } : hydrate; - if (resolvedComp) { - doHydrate(); - } else { - load().then(() => !instance.isUnmounted && doHydrate()); - } - }, - get __asyncResolved() { - return resolvedComp; - }, - setup() { - const instance = currentInstance; - markAsyncBoundary(instance); - if (resolvedComp) { - return () => createInnerComp(resolvedComp, instance); - } - const onError = (err) => { - pendingRequest = null; - handleError( - err, - instance, - 13, - !errorComponent - ); - }; - if (suspensible && instance.suspense || isInSSRComponentSetup) { - return load().then((comp) => { - return () => createInnerComp(comp, instance); - }).catch((err) => { - onError(err); - return () => errorComponent ? createVNode(errorComponent, { - error: err - }) : null; - }); - } - const loaded = ref(false); - const error = ref(); - const delayed = ref(!!delay); - if (delay) { - setTimeout(() => { - delayed.value = false; - }, delay); - } - if (timeout != null) { - setTimeout(() => { - if (!loaded.value && !error.value) { - const err = new Error( - `Async component timed out after ${timeout}ms.` - ); - onError(err); - error.value = err; - } - }, timeout); - } - load().then(() => { - loaded.value = true; - if (instance.parent && isKeepAlive(instance.parent.vnode)) { - instance.parent.update(); - } - }).catch((err) => { - onError(err); - error.value = err; - }); - return () => { - if (loaded.value && resolvedComp) { - return createInnerComp(resolvedComp, instance); - } else if (error.value && errorComponent) { - return createVNode(errorComponent, { - error: error.value - }); - } else if (loadingComponent && !delayed.value) { - return createVNode(loadingComponent); - } - }; - } - }); -} -function createInnerComp(comp, parent) { - const { ref: ref2, props, children, ce } = parent.vnode; - const vnode = createVNode(comp, props, children); - vnode.ref = ref2; - vnode.ce = ce; - delete parent.vnode.ce; - return vnode; -} - -const isKeepAlive = (vnode) => vnode.type.__isKeepAlive; -const KeepAliveImpl = { - name: `KeepAlive`, - // Marker for special handling inside the renderer. We are not using a === - // check directly on KeepAlive in the renderer, because importing it directly - // would prevent it from being tree-shaken. - __isKeepAlive: true, - props: { - include: [String, RegExp, Array], - exclude: [String, RegExp, Array], - max: [String, Number] - }, - setup(props, { slots }) { - const instance = getCurrentInstance(); - const sharedContext = instance.ctx; - if (!sharedContext.renderer) { - return () => { - const children = slots.default && slots.default(); - return children && children.length === 1 ? children[0] : children; - }; - } - const cache = /* @__PURE__ */ new Map(); - const keys = /* @__PURE__ */ new Set(); - let current = null; - { - instance.__v_cache = cache; - } - const parentSuspense = instance.suspense; - const { - renderer: { - p: patch, - m: move, - um: _unmount, - o: { createElement } - } - } = sharedContext; - const storageContainer = createElement("div"); - sharedContext.activate = (vnode, container, anchor, namespace, optimized) => { - const instance2 = vnode.component; - move(vnode, container, anchor, 0, parentSuspense); - patch( - instance2.vnode, - vnode, - container, - anchor, - instance2, - parentSuspense, - namespace, - vnode.slotScopeIds, - optimized - ); - queuePostRenderEffect(() => { - instance2.isDeactivated = false; - if (instance2.a) { - invokeArrayFns(instance2.a); - } - const vnodeHook = vnode.props && vnode.props.onVnodeMounted; - if (vnodeHook) { - invokeVNodeHook(vnodeHook, instance2.parent, vnode); - } - }, parentSuspense); - { - devtoolsComponentAdded(instance2); - } - }; - sharedContext.deactivate = (vnode) => { - const instance2 = vnode.component; - invalidateMount(instance2.m); - invalidateMount(instance2.a); - move(vnode, storageContainer, null, 1, parentSuspense); - queuePostRenderEffect(() => { - if (instance2.da) { - invokeArrayFns(instance2.da); - } - const vnodeHook = vnode.props && vnode.props.onVnodeUnmounted; - if (vnodeHook) { - invokeVNodeHook(vnodeHook, instance2.parent, vnode); - } - instance2.isDeactivated = true; - }, parentSuspense); - { - devtoolsComponentAdded(instance2); - } - }; - function unmount(vnode) { - resetShapeFlag(vnode); - _unmount(vnode, instance, parentSuspense, true); - } - function pruneCache(filter) { - cache.forEach((vnode, key) => { - const name = getComponentName(vnode.type); - if (name && !filter(name)) { - pruneCacheEntry(key); - } - }); - } - function pruneCacheEntry(key) { - const cached = cache.get(key); - if (cached && (!current || !isSameVNodeType(cached, current))) { - unmount(cached); - } else if (current) { - resetShapeFlag(current); - } - cache.delete(key); - keys.delete(key); - } - watch( - () => [props.include, props.exclude], - ([include, exclude]) => { - include && pruneCache((name) => matches(include, name)); - exclude && pruneCache((name) => !matches(exclude, name)); - }, - // prune post-render after `current` has been updated - { flush: "post", deep: true } - ); - let pendingCacheKey = null; - const cacheSubtree = () => { - if (pendingCacheKey != null) { - if (isSuspense(instance.subTree.type)) { - queuePostRenderEffect(() => { - cache.set(pendingCacheKey, getInnerChild(instance.subTree)); - }, instance.subTree.suspense); - } else { - cache.set(pendingCacheKey, getInnerChild(instance.subTree)); - } - } - }; - onMounted(cacheSubtree); - onUpdated(cacheSubtree); - onBeforeUnmount(() => { - cache.forEach((cached) => { - const { subTree, suspense } = instance; - const vnode = getInnerChild(subTree); - if (cached.type === vnode.type && cached.key === vnode.key) { - resetShapeFlag(vnode); - const da = vnode.component.da; - da && queuePostRenderEffect(da, suspense); - return; - } - unmount(cached); - }); - }); - return () => { - pendingCacheKey = null; - if (!slots.default) { - return current = null; - } - const children = slots.default(); - const rawVNode = children[0]; - if (children.length > 1) { - { - warn$1(`KeepAlive should contain exactly one component child.`); - } - current = null; - return children; - } else if (!isVNode(rawVNode) || !(rawVNode.shapeFlag & 4) && !(rawVNode.shapeFlag & 128)) { - current = null; - return rawVNode; - } - let vnode = getInnerChild(rawVNode); - if (vnode.type === Comment) { - current = null; - return vnode; - } - const comp = vnode.type; - const name = getComponentName( - isAsyncWrapper(vnode) ? vnode.type.__asyncResolved || {} : comp - ); - const { include, exclude, max } = props; - if (include && (!name || !matches(include, name)) || exclude && name && matches(exclude, name)) { - vnode.shapeFlag &= ~256; - current = vnode; - return rawVNode; - } - const key = vnode.key == null ? comp : vnode.key; - const cachedVNode = cache.get(key); - if (vnode.el) { - vnode = cloneVNode(vnode); - if (rawVNode.shapeFlag & 128) { - rawVNode.ssContent = vnode; - } - } - pendingCacheKey = key; - if (cachedVNode) { - vnode.el = cachedVNode.el; - vnode.component = cachedVNode.component; - if (vnode.transition) { - setTransitionHooks(vnode, vnode.transition); - } - vnode.shapeFlag |= 512; - keys.delete(key); - keys.add(key); - } else { - keys.add(key); - if (max && keys.size > parseInt(max, 10)) { - pruneCacheEntry(keys.values().next().value); - } - } - vnode.shapeFlag |= 256; - current = vnode; - return isSuspense(rawVNode.type) ? rawVNode : vnode; - }; - } -}; -const KeepAlive = KeepAliveImpl; -function matches(pattern, name) { - if (isArray(pattern)) { - return pattern.some((p) => matches(p, name)); - } else if (isString(pattern)) { - return pattern.split(",").includes(name); - } else if (isRegExp(pattern)) { - pattern.lastIndex = 0; - return pattern.test(name); - } - return false; -} -function onActivated(hook, target) { - registerKeepAliveHook(hook, "a", target); -} -function onDeactivated(hook, target) { - registerKeepAliveHook(hook, "da", target); -} -function registerKeepAliveHook(hook, type, target = currentInstance) { - const wrappedHook = hook.__wdc || (hook.__wdc = () => { - let current = target; - while (current) { - if (current.isDeactivated) { - return; - } - current = current.parent; - } - return hook(); - }); - injectHook(type, wrappedHook, target); - if (target) { - let current = target.parent; - while (current && current.parent) { - if (isKeepAlive(current.parent.vnode)) { - injectToKeepAliveRoot(wrappedHook, type, target, current); - } - current = current.parent; - } - } -} -function injectToKeepAliveRoot(hook, type, target, keepAliveRoot) { - const injected = injectHook( - type, - hook, - keepAliveRoot, - true - /* prepend */ - ); - onUnmounted(() => { - remove(keepAliveRoot[type], injected); - }, target); -} -function resetShapeFlag(vnode) { - vnode.shapeFlag &= ~256; - vnode.shapeFlag &= ~512; -} -function getInnerChild(vnode) { - return vnode.shapeFlag & 128 ? vnode.ssContent : vnode; -} - -function injectHook(type, hook, target = currentInstance, prepend = false) { - if (target) { - const hooks = target[type] || (target[type] = []); - const wrappedHook = hook.__weh || (hook.__weh = (...args) => { - pauseTracking(); - const reset = setCurrentInstance(target); - const res = callWithAsyncErrorHandling(hook, target, type, args); - reset(); - resetTracking(); - return res; - }); - if (prepend) { - hooks.unshift(wrappedHook); - } else { - hooks.push(wrappedHook); - } - return wrappedHook; - } else { - const apiName = toHandlerKey(ErrorTypeStrings$1[type].replace(/ hook$/, "")); - warn$1( - `${apiName} is called when there is no active component instance to be associated with. Lifecycle injection APIs can only be used during execution of setup().` + (` If you are using async setup(), make sure to register lifecycle hooks before the first await statement.` ) - ); - } -} -const createHook = (lifecycle) => (hook, target = currentInstance) => { - if (!isInSSRComponentSetup || lifecycle === "sp") { - injectHook(lifecycle, (...args) => hook(...args), target); - } -}; -const onBeforeMount = createHook("bm"); -const onMounted = createHook("m"); -const onBeforeUpdate = createHook( - "bu" -); -const onUpdated = createHook("u"); -const onBeforeUnmount = createHook( - "bum" -); -const onUnmounted = createHook("um"); -const onServerPrefetch = createHook( - "sp" -); -const onRenderTriggered = createHook("rtg"); -const onRenderTracked = createHook("rtc"); -function onErrorCaptured(hook, target = currentInstance) { - injectHook("ec", hook, target); -} - -const COMPONENTS = "components"; -const DIRECTIVES = "directives"; -function resolveComponent(name, maybeSelfReference) { - return resolveAsset(COMPONENTS, name, true, maybeSelfReference) || name; -} -const NULL_DYNAMIC_COMPONENT = Symbol.for("v-ndc"); -function resolveDynamicComponent(component) { - if (isString(component)) { - return resolveAsset(COMPONENTS, component, false) || component; - } else { - return component || NULL_DYNAMIC_COMPONENT; - } -} -function resolveDirective(name) { - return resolveAsset(DIRECTIVES, name); -} -function resolveAsset(type, name, warnMissing = true, maybeSelfReference = false) { - const instance = currentRenderingInstance || currentInstance; - if (instance) { - const Component = instance.type; - if (type === COMPONENTS) { - const selfName = getComponentName( - Component, - false - ); - if (selfName && (selfName === name || selfName === camelize(name) || selfName === capitalize(camelize(name)))) { - return Component; - } - } - const res = ( - // local registration - // check instance[type] first which is resolved for options API - resolve(instance[type] || Component[type], name) || // global registration - resolve(instance.appContext[type], name) - ); - if (!res && maybeSelfReference) { - return Component; - } - if (warnMissing && !res) { - const extra = type === COMPONENTS ? ` -If this is a native custom element, make sure to exclude it from component resolution via compilerOptions.isCustomElement.` : ``; - warn$1(`Failed to resolve ${type.slice(0, -1)}: ${name}${extra}`); - } - return res; - } else { - warn$1( - `resolve${capitalize(type.slice(0, -1))} can only be used in render() or setup().` - ); - } -} -function resolve(registry, name) { - return registry && (registry[name] || registry[camelize(name)] || registry[capitalize(camelize(name))]); -} - -function renderList(source, renderItem, cache, index) { - let ret; - const cached = cache && cache[index]; - const sourceIsArray = isArray(source); - if (sourceIsArray || isString(source)) { - const sourceIsReactiveArray = sourceIsArray && isReactive(source); - let needsWrap = false; - if (sourceIsReactiveArray) { - needsWrap = !isShallow(source); - source = shallowReadArray(source); - } - ret = new Array(source.length); - for (let i = 0, l = source.length; i < l; i++) { - ret[i] = renderItem( - needsWrap ? toReactive(source[i]) : source[i], - i, - void 0, - cached && cached[i] - ); - } - } else if (typeof source === "number") { - if (!Number.isInteger(source)) { - warn$1(`The v-for range expect an integer value but got ${source}.`); - } - ret = new Array(source); - for (let i = 0; i < source; i++) { - ret[i] = renderItem(i + 1, i, void 0, cached && cached[i]); - } - } else if (isObject(source)) { - if (source[Symbol.iterator]) { - ret = Array.from( - source, - (item, i) => renderItem(item, i, void 0, cached && cached[i]) - ); - } else { - const keys = Object.keys(source); - ret = new Array(keys.length); - for (let i = 0, l = keys.length; i < l; i++) { - const key = keys[i]; - ret[i] = renderItem(source[key], key, i, cached && cached[i]); - } - } - } else { - ret = []; - } - if (cache) { - cache[index] = ret; - } - return ret; -} - -function createSlots(slots, dynamicSlots) { - for (let i = 0; i < dynamicSlots.length; i++) { - const slot = dynamicSlots[i]; - if (isArray(slot)) { - for (let j = 0; j < slot.length; j++) { - slots[slot[j].name] = slot[j].fn; - } - } else if (slot) { - slots[slot.name] = slot.key ? (...args) => { - const res = slot.fn(...args); - if (res) res.key = slot.key; - return res; - } : slot.fn; - } - } - return slots; -} - -function renderSlot(slots, name, props = {}, fallback, noSlotted) { - if (currentRenderingInstance.ce || currentRenderingInstance.parent && isAsyncWrapper(currentRenderingInstance.parent) && currentRenderingInstance.parent.ce) { - if (name !== "default") props.name = name; - return openBlock(), createBlock( - Fragment, - null, - [createVNode("slot", props, fallback && fallback())], - 64 - ); - } - let slot = slots[name]; - if (slot && slot.length > 1) { - warn$1( - `SSR-optimized slot function detected in a non-SSR-optimized render function. You need to mark this component with $dynamic-slots in the parent template.` - ); - slot = () => []; - } - if (slot && slot._c) { - slot._d = false; - } - openBlock(); - const validSlotContent = slot && ensureValidVNode(slot(props)); - const slotKey = props.key || // slot content array of a dynamic conditional slot may have a branch - // key attached in the `createSlots` helper, respect that - validSlotContent && validSlotContent.key; - const rendered = createBlock( - Fragment, - { - key: (slotKey && !isSymbol(slotKey) ? slotKey : `_${name}`) + // #7256 force differentiate fallback content from actual content - (!validSlotContent && fallback ? "_fb" : "") - }, - validSlotContent || (fallback ? fallback() : []), - validSlotContent && slots._ === 1 ? 64 : -2 - ); - if (!noSlotted && rendered.scopeId) { - rendered.slotScopeIds = [rendered.scopeId + "-s"]; - } - if (slot && slot._c) { - slot._d = true; - } - return rendered; -} -function ensureValidVNode(vnodes) { - return vnodes.some((child) => { - if (!isVNode(child)) return true; - if (child.type === Comment) return false; - if (child.type === Fragment && !ensureValidVNode(child.children)) - return false; - return true; - }) ? vnodes : null; -} - -function toHandlers(obj, preserveCaseIfNecessary) { - const ret = {}; - if (!isObject(obj)) { - warn$1(`v-on with no argument expects an object value.`); - return ret; - } - for (const key in obj) { - ret[preserveCaseIfNecessary && /[A-Z]/.test(key) ? `on:${key}` : toHandlerKey(key)] = obj[key]; - } - return ret; -} - -const getPublicInstance = (i) => { - if (!i) return null; - if (isStatefulComponent(i)) return getComponentPublicInstance(i); - return getPublicInstance(i.parent); -}; -const publicPropertiesMap = ( - // Move PURE marker to new line to workaround compiler discarding it - // due to type annotation - /* @__PURE__ */ extend(/* @__PURE__ */ Object.create(null), { - $: (i) => i, - $el: (i) => i.vnode.el, - $data: (i) => i.data, - $props: (i) => shallowReadonly(i.props) , - $attrs: (i) => shallowReadonly(i.attrs) , - $slots: (i) => shallowReadonly(i.slots) , - $refs: (i) => shallowReadonly(i.refs) , - $parent: (i) => getPublicInstance(i.parent), - $root: (i) => getPublicInstance(i.root), - $host: (i) => i.ce, - $emit: (i) => i.emit, - $options: (i) => resolveMergedOptions(i) , - $forceUpdate: (i) => i.f || (i.f = () => { - queueJob(i.update); - }), - $nextTick: (i) => i.n || (i.n = nextTick.bind(i.proxy)), - $watch: (i) => instanceWatch.bind(i) - }) -); -const isReservedPrefix = (key) => key === "_" || key === "$"; -const hasSetupBinding = (state, key) => state !== EMPTY_OBJ && !state.__isScriptSetup && hasOwn(state, key); -const PublicInstanceProxyHandlers = { - get({ _: instance }, key) { - if (key === "__v_skip") { - return true; - } - const { ctx, setupState, data, props, accessCache, type, appContext } = instance; - if (key === "__isVue") { - return true; - } - let normalizedProps; - if (key[0] !== "$") { - const n = accessCache[key]; - if (n !== void 0) { - switch (n) { - case 1 /* SETUP */: - return setupState[key]; - case 2 /* DATA */: - return data[key]; - case 4 /* CONTEXT */: - return ctx[key]; - case 3 /* PROPS */: - return props[key]; - } - } else if (hasSetupBinding(setupState, key)) { - accessCache[key] = 1 /* SETUP */; - return setupState[key]; - } else if (data !== EMPTY_OBJ && hasOwn(data, key)) { - accessCache[key] = 2 /* DATA */; - return data[key]; - } else if ( - // only cache other properties when instance has declared (thus stable) - // props - (normalizedProps = instance.propsOptions[0]) && hasOwn(normalizedProps, key) - ) { - accessCache[key] = 3 /* PROPS */; - return props[key]; - } else if (ctx !== EMPTY_OBJ && hasOwn(ctx, key)) { - accessCache[key] = 4 /* CONTEXT */; - return ctx[key]; - } else if (shouldCacheAccess) { - accessCache[key] = 0 /* OTHER */; - } - } - const publicGetter = publicPropertiesMap[key]; - let cssModule, globalProperties; - if (publicGetter) { - if (key === "$attrs") { - track(instance.attrs, "get", ""); - markAttrsAccessed(); - } else if (key === "$slots") { - track(instance, "get", key); - } - return publicGetter(instance); - } else if ( - // css module (injected by vue-loader) - (cssModule = type.__cssModules) && (cssModule = cssModule[key]) - ) { - return cssModule; - } else if (ctx !== EMPTY_OBJ && hasOwn(ctx, key)) { - accessCache[key] = 4 /* CONTEXT */; - return ctx[key]; - } else if ( - // global properties - globalProperties = appContext.config.globalProperties, hasOwn(globalProperties, key) - ) { - { - return globalProperties[key]; - } - } else if (currentRenderingInstance && (!isString(key) || // #1091 avoid internal isRef/isVNode checks on component instance leading - // to infinite warning loop - key.indexOf("__v") !== 0)) { - if (data !== EMPTY_OBJ && isReservedPrefix(key[0]) && hasOwn(data, key)) { - warn$1( - `Property ${JSON.stringify( - key - )} must be accessed via $data because it starts with a reserved character ("$" or "_") and is not proxied on the render context.` - ); - } else if (instance === currentRenderingInstance) { - warn$1( - `Property ${JSON.stringify(key)} was accessed during render but is not defined on instance.` - ); - } - } - }, - set({ _: instance }, key, value) { - const { data, setupState, ctx } = instance; - if (hasSetupBinding(setupState, key)) { - setupState[key] = value; - return true; - } else if (setupState.__isScriptSetup && hasOwn(setupState, key)) { - warn$1(`Cannot mutate - - - - -

- - - - - - diff --git a/examples/server/public/index.html.gz b/examples/server/public/index.html.gz new file mode 100644 index 000000000..26f3583bd Binary files /dev/null and b/examples/server/public/index.html.gz differ diff --git a/examples/server/public_legacy/index-new.html b/examples/server/public_legacy/index-new.html index 8bfa380e5..cbfbbdf28 100644 --- a/examples/server/public_legacy/index-new.html +++ b/examples/server/public_legacy/index-new.html @@ -39,7 +39,6 @@ temperature: 0.8, // adapt all following parameters to optimized min-p requierements. If for non-english, set to 0.6 or lower repeat_last_n: 0, // 0 = disable penalty, -1 = context size repeat_penalty: 1.0, // 1.0 = disabled - penalize_nl: false, // true only useful for infinite completion dry_multiplier: 0.0, // 0.0 = disabled, 0.8 works well dry_base: 1.75, // 0.0 = disabled dry_allowed_length: 2, // tokens extending repetitions beyond this receive penalty, 2 works well diff --git a/examples/server/public_legacy/index.html b/examples/server/public_legacy/index.html index a95f5c6df..75f39330a 100644 --- a/examples/server/public_legacy/index.html +++ b/examples/server/public_legacy/index.html @@ -303,7 +303,6 @@ temperature: 0.7, repeat_last_n: 256, // 0 = disable penalty, -1 = context size repeat_penalty: 1.18, // 1.0 = disabled - penalize_nl: false, dry_multiplier: 0.0, // 0.0 = disabled, 0.8 works well dry_base: 1.75, // 0.0 = disabled dry_allowed_length: 2, // tokens extending repetitions beyond this receive penalty, 2 works well @@ -1006,7 +1005,6 @@ ${FloatField({ label: "Temperature", max: 2.0, min: 0.0, name: "temperature", step: 0.01, value: params.value.temperature })} ${FloatField({ label: "Penalize repeat sequence", max: 2.0, min: 0.0, name: "repeat_penalty", step: 0.01, value: params.value.repeat_penalty })} ${IntField({ label: "Consider N tokens for penalize", max: 2048, min: 0, name: "repeat_last_n", value: params.value.repeat_last_n })} - ${BoolField({ label: "Penalize repetition of newlines", name: "penalize_nl", value: params.value.penalize_nl })} ${IntField({ label: "Top-K sampling", max: 100, min: -1, name: "top_k", value: params.value.top_k })} ${FloatField({ label: "Top-P sampling", max: 1.0, min: 0.0, name: "top_p", step: 0.01, value: params.value.top_p })} ${FloatField({ label: "Min-P sampling", max: 1.0, min: 0.0, name: "min_p", step: 0.01, value: params.value.min_p })} diff --git a/examples/server/public_simplechat/simplechat.js b/examples/server/public_simplechat/simplechat.js index 8e0df3b61..2fcd24a86 100644 --- a/examples/server/public_simplechat/simplechat.js +++ b/examples/server/public_simplechat/simplechat.js @@ -407,6 +407,9 @@ class SimpleChat { if (curLine.startsWith("data:")) { curLine = curLine.substring(5); } + if (curLine.trim() === "[DONE]") { + break; + } let curJson = JSON.parse(curLine); console.debug("DBUG:SC:PART:Json:", curJson); this.append_response(this.response_extract_stream(curJson, apiEP)); diff --git a/examples/server/server.cpp b/examples/server/server.cpp index a6d3a1c95..64c0c4ef6 100644 --- a/examples/server/server.cpp +++ b/examples/server/server.cpp @@ -2,10 +2,11 @@ #include "arg.h" #include "common.h" -#include "log.h" -#include "sampling.h" #include "json-schema-to-grammar.h" #include "llama.h" +#include "log.h" +#include "sampling.h" +#include "speculative.h" // Change JSON_ASSERT from assert() to GGML_ASSERT: #define JSON_ASSERT GGML_ASSERT @@ -14,13 +15,8 @@ #define MIMETYPE_JSON "application/json; charset=utf-8" // auto generated files (update with ./deps.sh) -#include "index.html.hpp" -#include "completion.js.hpp" +#include "index.html.gz.hpp" #include "loading.html.hpp" -#include "deps_daisyui.min.css.hpp" -#include "deps_markdown-it.js.hpp" -#include "deps_tailwindcss.js.hpp" -#include "deps_vue.esm-browser.js.hpp" #include #include @@ -37,8 +33,10 @@ using json = nlohmann::ordered_json; enum stop_type { - STOP_TYPE_FULL, - STOP_TYPE_PARTIAL, + STOP_TYPE_NONE, + STOP_TYPE_EOS, + STOP_TYPE_WORD, + STOP_TYPE_LIMIT, }; // state diagram: https://github.com/ggerganov/llama.cpp/pull/9283 @@ -56,7 +54,10 @@ enum server_state { }; enum server_task_type { - SERVER_TASK_TYPE_INFERENCE, + SERVER_TASK_TYPE_COMPLETION, + SERVER_TASK_TYPE_EMBEDDING, + SERVER_TASK_TYPE_RERANK, + SERVER_TASK_TYPE_INFILL, SERVER_TASK_TYPE_CANCEL, SERVER_TASK_TYPE_NEXT_RESPONSE, SERVER_TASK_TYPE_METRICS, @@ -66,22 +67,340 @@ enum server_task_type { SERVER_TASK_TYPE_SET_LORA, }; -enum server_task_inf_type { - SERVER_TASK_INF_TYPE_COMPLETION, - SERVER_TASK_INF_TYPE_EMBEDDING, - SERVER_TASK_INF_TYPE_RERANK, - SERVER_TASK_INF_TYPE_INFILL, +enum oaicompat_type { + OAICOMPAT_TYPE_NONE, + OAICOMPAT_TYPE_CHAT, + OAICOMPAT_TYPE_COMPLETION, + OAICOMPAT_TYPE_EMBEDDING, +}; + +// https://community.openai.com/t/openai-chat-list-of-error-codes-and-types/357791/11 +enum error_type { + ERROR_TYPE_INVALID_REQUEST, + ERROR_TYPE_AUTHENTICATION, + ERROR_TYPE_SERVER, + ERROR_TYPE_NOT_FOUND, + ERROR_TYPE_PERMISSION, + ERROR_TYPE_UNAVAILABLE, // custom error + ERROR_TYPE_NOT_SUPPORTED, // custom error +}; + +struct slot_params { + bool stream = true; + bool cache_prompt = true; // remember the prompt to avoid reprocessing all prompt + bool return_tokens = false; + + int32_t n_keep = 0; // number of tokens to keep from initial prompt + int32_t n_discard = 0; // number of tokens after n_keep that may be discarded when shifting context, 0 defaults to half + int32_t n_predict = -1; // new tokens to predict + int32_t n_indent = 0; // mininum line indentation for the generated text in number of whitespace characters + + int64_t t_max_prompt_ms = -1; // TODO: implement + int64_t t_max_predict_ms = -1; // if positive, limit the generation phase to this time limit + + std::vector lora; + + std::vector antiprompt; + std::vector response_fields; + bool timings_per_token = false; + bool post_sampling_probs = false; + bool ignore_eos = false; + + struct common_params_sampling sampling; + struct common_params_speculative speculative; + + // OAI-compat fields + bool verbose = false; + oaicompat_type oaicompat = OAICOMPAT_TYPE_NONE; + std::string oaicompat_model; + std::string oaicompat_cmpl_id; + + json to_json() const { + std::vector samplers; + samplers.reserve(sampling.samplers.size()); + for (const auto & sampler : sampling.samplers) { + samplers.emplace_back(common_sampler_type_to_str(sampler)); + } + + json lora = json::array(); + for (size_t i = 0; i < this->lora.size(); ++i) { + lora.push_back({{"id", i}, {"scale", this->lora[i].scale}}); + } + + return json { + {"n_predict", n_predict}, // Server configured n_predict + {"seed", sampling.seed}, + {"temperature", sampling.temp}, + {"dynatemp_range", sampling.dynatemp_range}, + {"dynatemp_exponent", sampling.dynatemp_exponent}, + {"top_k", sampling.top_k}, + {"top_p", sampling.top_p}, + {"min_p", sampling.min_p}, + {"xtc_probability", sampling.xtc_probability}, + {"xtc_threshold", sampling.xtc_threshold}, + {"typical_p", sampling.typ_p}, + {"repeat_last_n", sampling.penalty_last_n}, + {"repeat_penalty", sampling.penalty_repeat}, + {"presence_penalty", sampling.penalty_present}, + {"frequency_penalty", sampling.penalty_freq}, + {"dry_multiplier", sampling.dry_multiplier}, + {"dry_base", sampling.dry_base}, + {"dry_allowed_length", sampling.dry_allowed_length}, + {"dry_penalty_last_n", sampling.dry_penalty_last_n}, + {"dry_sequence_breakers", sampling.dry_sequence_breakers}, + {"mirostat", sampling.mirostat}, + {"mirostat_tau", sampling.mirostat_tau}, + {"mirostat_eta", sampling.mirostat_eta}, + {"stop", antiprompt}, + {"max_tokens", n_predict}, // User configured n_predict + {"n_keep", n_keep}, + {"n_discard", n_discard}, + {"ignore_eos", sampling.ignore_eos}, + {"stream", stream}, + {"logit_bias", format_logit_bias(sampling.logit_bias)}, + {"n_probs", sampling.n_probs}, + {"min_keep", sampling.min_keep}, + {"grammar", sampling.grammar}, + {"samplers", samplers}, + {"speculative.n_max", speculative.n_max}, + {"speculative.n_min", speculative.n_min}, + {"speculative.p_min", speculative.p_min}, + {"timings_per_token", timings_per_token}, + {"post_sampling_probs", post_sampling_probs}, + {"lora", lora}, + }; + } }; struct server_task { - int id = -1; // to be filled by server_queue - int id_target = -1; // used by SERVER_TASK_TYPE_CANCEL + int id = -1; // to be filled by server_queue + int index = -1; // used when there are multiple prompts (batch request) - llama_tokens prompt_tokens; server_task_type type; - json data; - server_task_inf_type inf_type = SERVER_TASK_INF_TYPE_COMPLETION; + // used by SERVER_TASK_TYPE_CANCEL + int id_target = -1; + + // used by SERVER_TASK_TYPE_INFERENCE + slot_params params; + llama_tokens prompt_tokens; + int id_selected_slot = -1; + + // used by SERVER_TASK_TYPE_SLOT_SAVE, SERVER_TASK_TYPE_SLOT_RESTORE, SERVER_TASK_TYPE_SLOT_ERASE + struct slot_action { + int slot_id; + std::string filename; + std::string filepath; + }; + slot_action slot_action; + + // used by SERVER_TASK_TYPE_METRICS + bool metrics_reset_bucket = false; + + // used by SERVER_TASK_TYPE_SET_LORA + std::vector set_lora; + + server_task(server_task_type type) : type(type) {} + + static slot_params params_from_json_cmpl( + const llama_context * ctx, + const common_params & params_base, + const json & data) { + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + slot_params params; + + // Sampling parameter defaults are loaded from the global server context (but individual requests can still override them) + slot_params defaults; + defaults.sampling = params_base.sampling; + defaults.speculative = params_base.speculative; + + // enabling this will output extra debug information in the HTTP responses from the server + params.verbose = params_base.verbosity > 9; + params.timings_per_token = json_value(data, "timings_per_token", false); + + params.stream = json_value(data, "stream", false); + params.cache_prompt = json_value(data, "cache_prompt", true); + params.return_tokens = json_value(data, "return_tokens", false); + params.n_predict = json_value(data, "n_predict", json_value(data, "max_tokens", defaults.n_predict)); + params.n_indent = json_value(data, "n_indent", defaults.n_indent); + params.n_keep = json_value(data, "n_keep", defaults.n_keep); + params.n_discard = json_value(data, "n_discard", defaults.n_discard); + //params.t_max_prompt_ms = json_value(data, "t_max_prompt_ms", defaults.t_max_prompt_ms); // TODO: implement + params.t_max_predict_ms = json_value(data, "t_max_predict_ms", defaults.t_max_predict_ms); + params.response_fields = json_value(data, "response_fields", std::vector()); + + params.sampling.top_k = json_value(data, "top_k", defaults.sampling.top_k); + params.sampling.top_p = json_value(data, "top_p", defaults.sampling.top_p); + params.sampling.min_p = json_value(data, "min_p", defaults.sampling.min_p); + params.sampling.xtc_probability = json_value(data, "xtc_probability", defaults.sampling.xtc_probability); + params.sampling.xtc_threshold = json_value(data, "xtc_threshold", defaults.sampling.xtc_threshold); + params.sampling.typ_p = json_value(data, "typical_p", defaults.sampling.typ_p); + params.sampling.temp = json_value(data, "temperature", defaults.sampling.temp); + params.sampling.dynatemp_range = json_value(data, "dynatemp_range", defaults.sampling.dynatemp_range); + params.sampling.dynatemp_exponent = json_value(data, "dynatemp_exponent", defaults.sampling.dynatemp_exponent); + params.sampling.penalty_last_n = json_value(data, "repeat_last_n", defaults.sampling.penalty_last_n); + params.sampling.penalty_repeat = json_value(data, "repeat_penalty", defaults.sampling.penalty_repeat); + params.sampling.penalty_freq = json_value(data, "frequency_penalty", defaults.sampling.penalty_freq); + params.sampling.penalty_present = json_value(data, "presence_penalty", defaults.sampling.penalty_present); + params.sampling.dry_multiplier = json_value(data, "dry_multiplier", defaults.sampling.dry_multiplier); + params.sampling.dry_base = json_value(data, "dry_base", defaults.sampling.dry_base); + params.sampling.dry_allowed_length = json_value(data, "dry_allowed_length", defaults.sampling.dry_allowed_length); + params.sampling.dry_penalty_last_n = json_value(data, "dry_penalty_last_n", defaults.sampling.dry_penalty_last_n); + params.sampling.mirostat = json_value(data, "mirostat", defaults.sampling.mirostat); + params.sampling.mirostat_tau = json_value(data, "mirostat_tau", defaults.sampling.mirostat_tau); + params.sampling.mirostat_eta = json_value(data, "mirostat_eta", defaults.sampling.mirostat_eta); + params.sampling.seed = json_value(data, "seed", defaults.sampling.seed); + params.sampling.n_probs = json_value(data, "n_probs", defaults.sampling.n_probs); + params.sampling.min_keep = json_value(data, "min_keep", defaults.sampling.min_keep); + params.post_sampling_probs = json_value(data, "post_sampling_probs", defaults.post_sampling_probs); + + params.speculative.n_min = json_value(data, "speculative.n_min", defaults.speculative.n_min); + params.speculative.n_max = json_value(data, "speculative.n_max", defaults.speculative.n_max); + params.speculative.p_min = json_value(data, "speculative.p_min", defaults.speculative.p_min); + + params.speculative.n_min = std::min(params.speculative.n_max, params.speculative.n_min); + params.speculative.n_min = std::max(params.speculative.n_min, 2); + params.speculative.n_max = std::max(params.speculative.n_max, 0); + + if (data.contains("lora")) { + if (data.at("lora").is_array()) { + params.lora = parse_lora_request(params_base.lora_adapters, data.at("lora")); + } else { + throw std::runtime_error("Error: 'lora' must be an array of objects with 'id' and 'scale' fields"); + } + } else { + params.lora = params_base.lora_adapters; + } + + // TODO: add more sanity checks for the input parameters + + if (params.sampling.penalty_last_n < -1) { + throw std::runtime_error("Error: repeat_last_n must be >= -1"); + } + + if (params.sampling.dry_penalty_last_n < -1) { + throw std::runtime_error("Error: dry_penalty_last_n must be >= -1"); + } + + if (params.sampling.penalty_last_n == -1) { + // note: should be the slot's context and not the full context, but it's ok + params.sampling.penalty_last_n = llama_n_ctx(ctx); + } + + if (params.sampling.dry_penalty_last_n == -1) { + params.sampling.dry_penalty_last_n = llama_n_ctx(ctx); + } + + if (params.sampling.dry_base < 1.0f) { + params.sampling.dry_base = defaults.sampling.dry_base; + } + + // sequence breakers for DRY + { + // Currently, this is not compatible with TextGen WebUI, Koboldcpp and SillyTavern format + // Ref: https://github.com/oobabooga/text-generation-webui/blob/d1af7a41ade7bd3c3a463bfa640725edb818ebaf/extensions/openai/typing.py#L39 + + if (data.contains("dry_sequence_breakers")) { + params.sampling.dry_sequence_breakers = json_value(data, "dry_sequence_breakers", std::vector()); + if (params.sampling.dry_sequence_breakers.empty()) { + throw std::runtime_error("Error: dry_sequence_breakers must be a non-empty array of strings"); + } + } + } + + // process "json_schema" and "grammar" + if (data.contains("json_schema") && !data.at("json_schema").is_null() && data.contains("grammar") && !data.at("grammar").is_null()) { + throw std::runtime_error("Either \"json_schema\" or \"grammar\" can be specified, but not both"); + } + if (data.contains("json_schema") && !data.contains("grammar")) { + try { + auto schema = json_value(data, "json_schema", json::object()); + params.sampling.grammar = json_schema_to_grammar(schema); + } catch (const std::exception & e) { + throw std::runtime_error(std::string("\"json_schema\": ") + e.what()); + } + } else { + params.sampling.grammar = json_value(data, "grammar", defaults.sampling.grammar); + } + + { + params.sampling.logit_bias.clear(); + params.ignore_eos = json_value(data, "ignore_eos", false); + + const auto & logit_bias = data.find("logit_bias"); + if (logit_bias != data.end() && logit_bias->is_array()) { + const int n_vocab = llama_vocab_n_tokens(vocab); + for (const auto & el : *logit_bias) { + // TODO: we may want to throw errors here, in case "el" is incorrect + if (el.is_array() && el.size() == 2) { + float bias; + if (el[1].is_number()) { + bias = el[1].get(); + } else if (el[1].is_boolean() && !el[1].get()) { + bias = -INFINITY; + } else { + continue; + } + + if (el[0].is_number_integer()) { + llama_token tok = el[0].get(); + if (tok >= 0 && tok < n_vocab) { + params.sampling.logit_bias.push_back({tok, bias}); + } + } else if (el[0].is_string()) { + auto toks = common_tokenize(vocab, el[0].get(), false); + for (auto tok : toks) { + params.sampling.logit_bias.push_back({tok, bias}); + } + } + } + } + } + } + + { + params.antiprompt.clear(); + + const auto & stop = data.find("stop"); + if (stop != data.end() && stop->is_array()) { + for (const auto & word : *stop) { + if (!word.empty()) { + params.antiprompt.push_back(word); + } + } + } + } + + { + const auto & samplers = data.find("samplers"); + if (samplers != data.end()) { + if (samplers->is_array()) { + std::vector sampler_names; + for (const auto & name : *samplers) { + if (name.is_string()) { + sampler_names.emplace_back(name); + } + } + params.sampling.samplers = common_sampler_types_from_names(sampler_names, false); + } else if (samplers->is_string()){ + std::string sampler_string; + for (const auto & name : *samplers) { + sampler_string += name; + } + params.sampling.samplers = common_sampler_types_from_chars(sampler_string); + } + } else { + params.sampling.samplers = defaults.sampling.samplers; + } + } + + std::string model_name = params_base.model_alias.empty() ? DEFAULT_OAICOMPAT_MODEL : params_base.model_alias; + params.oaicompat_model = json_value(data, "model", model_name); + + return params; + } // utility function static std::unordered_set get_list_id(const std::vector & tasks) { @@ -93,34 +412,729 @@ struct server_task { } }; -struct server_task_result { - int id = -1; +struct result_timings { + int32_t prompt_n = -1; + double prompt_ms; + double prompt_per_token_ms; + double prompt_per_second; - json data; + int32_t predicted_n = -1; + double predicted_ms; + double predicted_per_token_ms; + double predicted_per_second; - bool stop; - bool error; + json to_json() const { + return { + {"prompt_n", prompt_n}, + {"prompt_ms", prompt_ms}, + {"prompt_per_token_ms", prompt_per_token_ms}, + {"prompt_per_second", prompt_per_second}, + + {"predicted_n", predicted_n}, + {"predicted_ms", predicted_ms}, + {"predicted_per_token_ms", predicted_per_token_ms}, + {"predicted_per_second", predicted_per_second}, + }; + } }; -struct slot_params { - bool stream = true; - bool cache_prompt = false; // remember the prompt to avoid reprocessing all prompt +struct server_task_result { + int id = -1; + int id_slot = -1; + virtual bool is_error() { + // only used by server_task_result_error + return false; + } + virtual bool is_stop() { + // only used by server_task_result_cmpl_* + return false; + } + virtual int get_index() { + return -1; + } + virtual json to_json() = 0; + virtual ~server_task_result() = default; +}; - int32_t n_keep = 0; // number of tokens to keep from initial prompt - int32_t n_discard = 0; // number of tokens after n_keep that may be discarded when shifting context, 0 defaults to half - int32_t n_predict = -1; // new tokens to predict - int32_t n_indent = 0; // mininum line indentation for the generated text in number of whitespace characters +// using shared_ptr for polymorphism of server_task_result +using server_task_result_ptr = std::unique_ptr; - int64_t t_max_prompt_ms = -1; // TODO: implement - int64_t t_max_predict_ms = -1; // if positive, limit the generation phase to this time limit +inline std::string stop_type_to_str(stop_type type) { + switch (type) { + case STOP_TYPE_EOS: return "eos"; + case STOP_TYPE_WORD: return "word"; + case STOP_TYPE_LIMIT: return "limit"; + default: return "none"; + } +} - std::vector antiprompt; +struct completion_token_output { + llama_token tok; + float prob; + std::string text_to_send; + struct prob_info { + llama_token tok; + std::string txt; + float prob; + }; + std::vector probs; + + json to_json(bool post_sampling_probs) const { + json probs_for_token = json::array(); + for (const auto & p : probs) { + std::string txt(p.txt); + txt.resize(validate_utf8(txt)); + probs_for_token.push_back(json { + {"id", p.tok}, + {"token", txt}, + {"bytes", str_to_bytes(p.txt)}, + { + post_sampling_probs ? "prob" : "logprob", + post_sampling_probs ? p.prob : logarithm(p.prob) + }, + }); + } + return probs_for_token; + } + + static json probs_vector_to_json(const std::vector & probs, bool post_sampling_probs) { + json out = json::array(); + for (const auto & p : probs) { + std::string txt(p.text_to_send); + txt.resize(validate_utf8(txt)); + out.push_back(json { + {"id", p.tok}, + {"token", txt}, + {"bytes", str_to_bytes(p.text_to_send)}, + { + post_sampling_probs ? "prob" : "logprob", + post_sampling_probs ? p.prob : logarithm(p.prob) + }, + { + post_sampling_probs ? "top_probs" : "top_logprobs", + p.to_json(post_sampling_probs) + }, + }); + } + return out; + } + + static float logarithm(float x) { + // nlohmann::json converts -inf to null, so we need to prevent that + return x == 0.0f ? std::numeric_limits::lowest() : std::log(x); + } + + static std::vector str_to_bytes(const std::string & str) { + std::vector bytes; + for (unsigned char c : str) { + bytes.push_back(c); + } + return bytes; + } +}; + +struct server_task_result_cmpl_final : server_task_result { + int index = 0; + + std::string content; + llama_tokens tokens; + + bool stream; + result_timings timings; + std::string prompt; + + bool truncated; + int32_t n_decoded; + int32_t n_prompt_tokens; + int32_t n_tokens_cached; + bool has_new_line; + std::string stopping_word; + stop_type stop = STOP_TYPE_NONE; + + bool post_sampling_probs; + std::vector probs_output; + std::vector response_fields; + + slot_params generation_params; + + // OAI-compat fields + bool verbose = false; + oaicompat_type oaicompat = OAICOMPAT_TYPE_NONE; + std::string oaicompat_model; + std::string oaicompat_cmpl_id; + + virtual int get_index() override { + return index; + } + + virtual bool is_stop() override { + return true; // in stream mode, final responses are considered stop + } + + virtual json to_json() override { + switch (oaicompat) { + case OAICOMPAT_TYPE_NONE: + return to_json_non_oaicompat(); + case OAICOMPAT_TYPE_COMPLETION: + return to_json_oaicompat(); + case OAICOMPAT_TYPE_CHAT: + return stream ? to_json_oaicompat_chat_stream() : to_json_oaicompat_chat(); + default: + GGML_ASSERT(false && "Invalid oaicompat_type"); + } + } + + json to_json_non_oaicompat() { + json res = json { + {"index", index}, + {"content", stream ? "" : content}, // in stream mode, content is already in last partial chunk + {"tokens", stream ? llama_tokens {} : tokens}, + {"id_slot", id_slot}, + {"stop", true}, + {"model", oaicompat_model}, + {"tokens_predicted", n_decoded}, + {"tokens_evaluated", n_prompt_tokens}, + {"generation_settings", generation_params.to_json()}, + {"prompt", prompt}, + {"has_new_line", has_new_line}, + {"truncated", truncated}, + {"stop_type", stop_type_to_str(stop)}, + {"stopping_word", stopping_word}, + {"tokens_cached", n_tokens_cached}, + {"timings", timings.to_json()}, + }; + if (!stream && !probs_output.empty()) { + res["completion_probabilities"] = completion_token_output::probs_vector_to_json(probs_output, post_sampling_probs); + } + return response_fields.empty() ? res : json_get_nested_values(response_fields, res); + } + + json to_json_oaicompat() { + std::time_t t = std::time(0); + json logprobs = json(nullptr); // OAI default to null + if (!stream && probs_output.size() > 0) { + logprobs = json{ + {"content", completion_token_output::probs_vector_to_json(probs_output, post_sampling_probs)}, + }; + } + json finish_reason = "length"; + if (stop == STOP_TYPE_WORD || stop == STOP_TYPE_EOS) { + finish_reason = "stop"; + } + json res = json { + {"choices", json::array({ + json{ + {"text", stream ? "" : content}, // in stream mode, content is already in last partial chunk + {"index", index}, + {"logprobs", logprobs}, + {"finish_reason", finish_reason}, + } + })}, + {"created", t}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "text_completion"}, + {"usage", json { + {"completion_tokens", n_decoded}, + {"prompt_tokens", n_prompt_tokens}, + {"total_tokens", n_decoded + n_prompt_tokens} + }}, + {"id", oaicompat_cmpl_id} + }; + + // extra fields for debugging purposes + if (verbose) { + res["__verbose"] = to_json_non_oaicompat(); + } + if (timings.prompt_n >= 0) { + res.push_back({"timings", timings.to_json()}); + } + + return res; + } + + json to_json_oaicompat_chat() { + std::string finish_reason = "length"; + if (stop == STOP_TYPE_WORD || stop == STOP_TYPE_EOS) { + finish_reason = "stop"; + } + + json choice = json{ + {"finish_reason", finish_reason}, + {"index", 0}, + {"message", json { + {"content", content}, + {"role", "assistant"} + } + }}; + + if (!stream && probs_output.size() > 0) { + choice["logprobs"] = json{ + {"content", completion_token_output::probs_vector_to_json(probs_output, post_sampling_probs)}, + }; + } + + std::time_t t = std::time(0); + + json res = json { + {"choices", json::array({choice})}, + {"created", t}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "chat.completion"}, + {"usage", json { + {"completion_tokens", n_decoded}, + {"prompt_tokens", n_prompt_tokens}, + {"total_tokens", n_decoded + n_prompt_tokens} + }}, + {"id", oaicompat_cmpl_id} + }; + + // extra fields for debugging purposes + if (verbose) { + res["__verbose"] = to_json_non_oaicompat(); + } + if (timings.prompt_n >= 0) { + res.push_back({"timings", timings.to_json()}); + } + + return res; + } + + json to_json_oaicompat_chat_stream() { + std::time_t t = std::time(0); + std::string finish_reason = "length"; + if (stop == STOP_TYPE_WORD || stop == STOP_TYPE_EOS) { + finish_reason = "stop"; + } + + json choice = json{ + {"finish_reason", finish_reason}, + {"index", 0}, + {"delta", json::object()} + }; + + json ret = json { + {"choices", json::array({choice})}, + {"created", t}, + {"id", oaicompat_cmpl_id}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "chat.completion.chunk"}, + {"usage", json { + {"completion_tokens", n_decoded}, + {"prompt_tokens", n_prompt_tokens}, + {"total_tokens", n_decoded + n_prompt_tokens}, + }}, + }; + + if (timings.prompt_n >= 0) { + ret.push_back({"timings", timings.to_json()}); + } + + return ret; + } +}; + +struct server_task_result_cmpl_partial : server_task_result { + int index = 0; + + std::string content; + llama_tokens tokens; + + int32_t n_decoded; + int32_t n_prompt_tokens; + + bool post_sampling_probs; + completion_token_output prob_output; + result_timings timings; + + // OAI-compat fields + bool verbose = false; + oaicompat_type oaicompat = OAICOMPAT_TYPE_NONE; + std::string oaicompat_model; + std::string oaicompat_cmpl_id; + + virtual int get_index() override { + return index; + } + + virtual bool is_stop() override { + return false; // in stream mode, partial responses are not considered stop + } + + virtual json to_json() override { + switch (oaicompat) { + case OAICOMPAT_TYPE_NONE: + return to_json_non_oaicompat(); + case OAICOMPAT_TYPE_COMPLETION: + return to_json_oaicompat(); + case OAICOMPAT_TYPE_CHAT: + return to_json_oaicompat_chat(); + default: + GGML_ASSERT(false && "Invalid oaicompat_type"); + } + } + + json to_json_non_oaicompat() { + // non-OAI-compat JSON + json res = json { + {"index", index}, + {"content", content}, + {"tokens", tokens}, + {"stop", false}, + {"id_slot", id_slot}, + {"tokens_predicted", n_decoded}, + {"tokens_evaluated", n_prompt_tokens}, + }; + // populate the timings object when needed (usually for the last response or with timings_per_token enabled) + if (timings.prompt_n > 0) { + res.push_back({"timings", timings.to_json()}); + } + if (!prob_output.probs.empty()) { + res["completion_probabilities"] = completion_token_output::probs_vector_to_json({prob_output}, post_sampling_probs); + } + return res; + } + + json to_json_oaicompat() { + std::time_t t = std::time(0); + json logprobs = json(nullptr); // OAI default to null + if (prob_output.probs.size() > 0) { + logprobs = json{ + {"content", completion_token_output::probs_vector_to_json({prob_output}, post_sampling_probs)}, + }; + } + json res = json { + {"choices", json::array({ + json{ + {"text", content}, + {"index", index}, + {"logprobs", logprobs}, + {"finish_reason", nullptr}, + } + })}, + {"created", t}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "text_completion"}, + {"id", oaicompat_cmpl_id} + }; + + // extra fields for debugging purposes + if (verbose) { + res["__verbose"] = to_json_non_oaicompat(); + } + if (timings.prompt_n >= 0) { + res.push_back({"timings", timings.to_json()}); + } + + return res; + } + + json to_json_oaicompat_chat() { + bool first = n_decoded == 0; + std::time_t t = std::time(0); + json choices; + + if (first) { + if (content.empty()) { + choices = json::array({json{{"finish_reason", nullptr}, + {"index", 0}, + {"delta", json{{"role", "assistant"}}}}}); + } else { + // We have to send this as two updates to conform to openai behavior + json initial_ret = json{{"choices", json::array({json{ + {"finish_reason", nullptr}, + {"index", 0}, + {"delta", json{ + {"role", "assistant"} + }}}})}, + {"created", t}, + {"id", oaicompat_cmpl_id}, + {"model", oaicompat_model}, + {"object", "chat.completion.chunk"}}; + + json second_ret = json{ + {"choices", json::array({json{{"finish_reason", nullptr}, + {"index", 0}, + {"delta", json { + {"content", content}}} + }})}, + {"created", t}, + {"id", oaicompat_cmpl_id}, + {"model", oaicompat_model}, + {"object", "chat.completion.chunk"}}; + + return std::vector({initial_ret, second_ret}); + } + } else { + choices = json::array({json{ + {"finish_reason", nullptr}, + {"index", 0}, + {"delta", + json { + {"content", content}, + }}, + }}); + } + + GGML_ASSERT(choices.size() >= 1); + + if (prob_output.probs.size() > 0) { + choices[0]["logprobs"] = json{ + {"content", completion_token_output::probs_vector_to_json({prob_output}, post_sampling_probs)}, + }; + } + + json ret = json { + {"choices", choices}, + {"created", t}, + {"id", oaicompat_cmpl_id}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "chat.completion.chunk"} + }; + + if (timings.prompt_n >= 0) { + ret.push_back({"timings", timings.to_json()}); + } + + return std::vector({ret}); + } +}; + +struct server_task_result_embd : server_task_result { + int index = 0; + std::vector> embedding; + + int32_t n_tokens; + + // OAI-compat fields + oaicompat_type oaicompat = OAICOMPAT_TYPE_NONE; + + virtual int get_index() override { + return index; + } + + virtual json to_json() override { + return oaicompat == OAICOMPAT_TYPE_EMBEDDING + ? to_json_oaicompat() + : to_json_non_oaicompat(); + } + + json to_json_non_oaicompat() { + return json { + {"index", index}, + {"embedding", embedding}, + }; + } + + json to_json_oaicompat() { + return json { + {"index", index}, + {"embedding", embedding[0]}, + {"tokens_evaluated", n_tokens}, + }; + } +}; + +struct server_task_result_rerank : server_task_result { + int index = 0; + float score = -1e6; + + int32_t n_tokens; + + virtual int get_index() override { + return index; + } + + virtual json to_json() override { + return json { + {"index", index}, + {"score", score}, + {"tokens_evaluated", n_tokens}, + }; + } +}; + +// this function maybe used outside of server_task_result_error +static json format_error_response(const std::string & message, const enum error_type type) { + std::string type_str; + int code = 500; + switch (type) { + case ERROR_TYPE_INVALID_REQUEST: + type_str = "invalid_request_error"; + code = 400; + break; + case ERROR_TYPE_AUTHENTICATION: + type_str = "authentication_error"; + code = 401; + break; + case ERROR_TYPE_NOT_FOUND: + type_str = "not_found_error"; + code = 404; + break; + case ERROR_TYPE_SERVER: + type_str = "server_error"; + code = 500; + break; + case ERROR_TYPE_PERMISSION: + type_str = "permission_error"; + code = 403; + break; + case ERROR_TYPE_NOT_SUPPORTED: + type_str = "not_supported_error"; + code = 501; + break; + case ERROR_TYPE_UNAVAILABLE: + type_str = "unavailable_error"; + code = 503; + break; + } + return json { + {"code", code}, + {"message", message}, + {"type", type_str}, + }; +} + +struct server_task_result_error : server_task_result { + int index = 0; + error_type err_type = ERROR_TYPE_SERVER; + std::string err_msg; + + virtual bool is_error() override { + return true; + } + + virtual json to_json() override { + return format_error_response(err_msg, err_type); + } +}; + +struct server_task_result_metrics : server_task_result { + int n_idle_slots; + int n_processing_slots; + int n_tasks_deferred; + int64_t t_start; + + int32_t kv_cache_tokens_count; + int32_t kv_cache_used_cells; + + // TODO: somehow reuse server_metrics in the future, instead of duplicating the fields + uint64_t n_prompt_tokens_processed_total = 0; + uint64_t t_prompt_processing_total = 0; + uint64_t n_tokens_predicted_total = 0; + uint64_t t_tokens_generation_total = 0; + + uint64_t n_prompt_tokens_processed = 0; + uint64_t t_prompt_processing = 0; + + uint64_t n_tokens_predicted = 0; + uint64_t t_tokens_generation = 0; + + uint64_t n_decode_total = 0; + uint64_t n_busy_slots_total = 0; + + // while we can also use std::vector this requires copying the slot object which can be quite messy + // therefore, we use json to temporarily store the slot.to_json() result + json slots_data = json::array(); + + virtual json to_json() override { + return json { + { "idle", n_idle_slots }, + { "processing", n_processing_slots }, + { "deferred", n_tasks_deferred }, + { "t_start", t_start }, + + { "n_prompt_tokens_processed_total", n_prompt_tokens_processed_total }, + { "t_tokens_generation_total", t_tokens_generation_total }, + { "n_tokens_predicted_total", n_tokens_predicted_total }, + { "t_prompt_processing_total", t_prompt_processing_total }, + + { "n_prompt_tokens_processed", n_prompt_tokens_processed }, + { "t_prompt_processing", t_prompt_processing }, + { "n_tokens_predicted", n_tokens_predicted }, + { "t_tokens_generation", t_tokens_generation }, + + { "n_decode_total", n_decode_total }, + { "n_busy_slots_total", n_busy_slots_total }, + + { "kv_cache_tokens_count", kv_cache_tokens_count }, + { "kv_cache_used_cells", kv_cache_used_cells }, + + { "slots", slots_data }, + }; + } +}; + +struct server_task_result_slot_save_load : server_task_result { + std::string filename; + bool is_save; // true = save, false = load + + size_t n_tokens; + size_t n_bytes; + double t_ms; + + virtual json to_json() override { + if (is_save) { + return json { + { "id_slot", id_slot }, + { "filename", filename }, + { "n_saved", n_tokens }, + { "n_written", n_bytes }, + { "timings", { + { "save_ms", t_ms } + }}, + }; + } else { + return json { + { "id_slot", id_slot }, + { "filename", filename }, + { "n_restored", n_tokens }, + { "n_read", n_bytes }, + { "timings", { + { "restore_ms", t_ms } + }}, + }; + } + } +}; + +struct server_task_result_slot_erase : server_task_result { + size_t n_erased; + + virtual json to_json() override { + return json { + { "id_slot", id_slot }, + { "n_erased", n_erased }, + }; + } +}; + +struct server_task_result_apply_lora : server_task_result { + virtual json to_json() override { + return json {{ "success", true }}; + } }; struct server_slot { int id; int id_task = -1; + // only used for completion/embedding/infill/rerank + server_task_type task_type = SERVER_TASK_TYPE_COMPLETION; + + llama_batch batch_spec = {}; + + llama_context * ctx = nullptr; + llama_context * ctx_dft = nullptr; + + common_speculative * spec = nullptr; + + std::vector lora; + // the index relative to completion multi-task request size_t index = 0; @@ -148,35 +1162,29 @@ struct server_slot { size_t last_nl_pos = 0; - std::string generated_text; - llama_tokens cache_tokens; - std::vector generated_token_probs; + std::string generated_text; + llama_tokens generated_tokens; - server_task_inf_type inf_type = SERVER_TASK_INF_TYPE_COMPLETION; + llama_tokens cache_tokens; + + std::vector generated_token_probs; bool has_next_token = true; bool has_new_line = false; bool truncated = false; - bool stopped_eos = false; - bool stopped_word = false; - bool stopped_limit = false; + stop_type stop; - bool oaicompat = false; - - std::string oaicompat_model; std::string stopping_word; // sampling json json_schema; - struct common_sampler_params sparams; struct common_sampler * smpl = nullptr; llama_token sampled; // stats size_t n_sent_text = 0; // number of sent text character - size_t n_sent_token_probs = 0; int64_t t_start_process_prompt; int64_t t_start_generation; @@ -194,19 +1202,26 @@ struct server_slot { generated_text = ""; has_new_line = false; truncated = false; - stopped_eos = false; - stopped_word = false; - stopped_limit = false; + stop = STOP_TYPE_NONE; stopping_word = ""; n_past = 0; n_sent_text = 0; - n_sent_token_probs = 0; - inf_type = SERVER_TASK_INF_TYPE_COMPLETION; + task_type = SERVER_TASK_TYPE_COMPLETION; + generated_tokens.clear(); generated_token_probs.clear(); } - bool has_budget(common_params &global_params) { + bool is_non_causal() const { + return task_type == SERVER_TASK_TYPE_EMBEDDING || task_type == SERVER_TASK_TYPE_RERANK; + } + + bool can_batch_with(server_slot & other_slot) { + return is_non_causal() == other_slot.is_non_causal() + && are_lora_equal(lora, other_slot.lora); + } + + bool has_budget(const common_params & global_params) { if (params.n_predict == -1 && global_params.n_predict == -1) { return true; // limitless } @@ -226,6 +1241,10 @@ struct server_slot { return state != SLOT_STATE_IDLE; } + bool can_speculate() const { + return ctx_dft && params.speculative.n_max > 0 && params.cache_prompt; + } + void add_token(const completion_token_output & token) { if (!is_processing()) { SLT_WRN(*this, "%s", "slot is not processing\n"); @@ -245,38 +1264,40 @@ struct server_slot { } } - json get_formated_timings() const { - return json { - {"prompt_n", n_prompt_tokens_processed}, - {"prompt_ms", t_prompt_processing}, - {"prompt_per_token_ms", t_prompt_processing / n_prompt_tokens_processed}, - {"prompt_per_second", 1e3 / t_prompt_processing * n_prompt_tokens_processed}, + result_timings get_timings() const { + result_timings timings; + timings.prompt_n = n_prompt_tokens_processed; + timings.prompt_ms = t_prompt_processing; + timings.prompt_per_token_ms = t_prompt_processing / n_prompt_tokens_processed; + timings.prompt_per_second = 1e3 / t_prompt_processing * n_prompt_tokens_processed; - {"predicted_n", n_decoded}, - {"predicted_ms", t_token_generation}, - {"predicted_per_token_ms", t_token_generation / n_decoded}, - {"predicted_per_second", 1e3 / t_token_generation * n_decoded}, - }; + timings.predicted_n = n_decoded; + timings.predicted_ms = t_token_generation; + timings.predicted_per_token_ms = t_token_generation / n_decoded; + timings.predicted_per_second = 1e3 / t_token_generation * n_decoded; + + return timings; } - size_t find_stopping_strings(const std::string & text, const size_t last_token_size, const stop_type type) { + size_t find_stopping_strings(const std::string & text, const size_t last_token_size, bool is_full_stop) { size_t stop_pos = std::string::npos; for (const std::string & word : params.antiprompt) { size_t pos; - if (type == STOP_TYPE_FULL) { + if (is_full_stop) { const size_t tmp = word.size() + last_token_size; const size_t from_pos = text.size() > tmp ? text.size() - tmp : 0; pos = text.find(word, from_pos); } else { + // otherwise, partial stop pos = find_partial_stop_string(word, text); } if (pos != std::string::npos && (stop_pos == std::string::npos || pos < stop_pos)) { - if (type == STOP_TYPE_FULL) { - stopped_word = true; + if (is_full_stop) { + stop = STOP_TYPE_WORD; stopping_word = word; has_next_token = false; } @@ -296,13 +1317,35 @@ struct server_slot { SLT_INF(*this, "\n" - "\rprompt eval time = %10.2f ms / %5d tokens (%8.2f ms per token, %8.2f tokens per second)\n" - "\r eval time = %10.2f ms / %5d tokens (%8.2f ms per token, %8.2f tokens per second)\n" - "\r total time = %10.2f ms / %5d tokens\n", + "prompt eval time = %10.2f ms / %5d tokens (%8.2f ms per token, %8.2f tokens per second)\n" + " eval time = %10.2f ms / %5d tokens (%8.2f ms per token, %8.2f tokens per second)\n" + " total time = %10.2f ms / %5d tokens\n", t_prompt_processing, n_prompt_tokens_processed, t_prompt, n_prompt_second, t_token_generation, n_decoded, t_gen, n_gen_second, t_prompt_processing + t_token_generation, n_prompt_tokens_processed + n_decoded); } + + json to_json() const { + return json { + {"id", id}, + {"id_task", id_task}, + {"n_ctx", n_ctx}, + {"speculative", can_speculate()}, + {"is_processing", is_processing()}, + {"non_causal", is_non_causal()}, + {"params", params.to_json()}, + {"prompt", common_detokenize(ctx, prompt_tokens)}, + {"next_token", + { + {"has_next_token", has_next_token}, + {"has_new_line", has_new_line}, + {"n_remain", n_remaining}, + {"n_decoded", n_decoded}, + {"stopping_word", stopping_word}, + } + }, + }; + } }; struct server_metrics { @@ -375,9 +1418,7 @@ struct server_queue { // Add a new task to the end of the queue int post(server_task task, bool front = false) { std::unique_lock lock(mutex_tasks); - if (task.id == -1) { - task.id = id++; - } + GGML_ASSERT(task.id != -1); QUE_DBG("new task, id = %d, front = %d\n", task.id, front); if (front) { queue_tasks.push_front(std::move(task)); @@ -501,8 +1542,8 @@ struct server_response { // for keeping track of all tasks waiting for the result std::unordered_set waiting_task_ids; - // the main result queue - std::vector queue_results; + // the main result queue (using ptr for polymorphism) + std::vector queue_results; std::mutex mutex_results; std::condition_variable condition_results; @@ -542,7 +1583,7 @@ struct server_response { } // This function blocks the thread until there is a response for one of the id_tasks - server_task_result recv(const std::unordered_set & id_tasks) { + server_task_result_ptr recv(const std::unordered_set & id_tasks) { while (true) { std::unique_lock lock(mutex_results); condition_results.wait(lock, [&]{ @@ -550,8 +1591,8 @@ struct server_response { }); for (int i = 0; i < (int) queue_results.size(); i++) { - if (id_tasks.find(queue_results[i].id) != id_tasks.end()) { - server_task_result res = queue_results[i]; + if (id_tasks.find(queue_results[i]->id) != id_tasks.end()) { + server_task_result_ptr res = std::move(queue_results[i]); queue_results.erase(queue_results.begin() + i); return res; } @@ -562,21 +1603,21 @@ struct server_response { } // single-task version of recv() - server_task_result recv(int id_task) { + server_task_result_ptr recv(int id_task) { std::unordered_set id_tasks = {id_task}; return recv(id_tasks); } // Send a new result to a waiting id_task - void send(server_task_result & result) { - SRV_DBG("sending result for task id = %d\n", result.id); + void send(server_task_result_ptr && result) { + SRV_DBG("sending result for task id = %d\n", result->id); std::unique_lock lock(mutex_results); for (const auto & id_task : waiting_task_ids) { - if (result.id == id_task) { - SRV_DBG("task id = %d moved to result queue\n", result.id); + if (result->id == id_task) { + SRV_DBG("task id = %d pushed to result queue\n", result->id); - queue_results.push_back(std::move(result)); + queue_results.emplace_back(std::move(result)); condition_results.notify_all(); return; } @@ -585,11 +1626,20 @@ struct server_response { }; struct server_context { + common_params params_base; + + // note: keep these alive - they determine the lifetime of the model, context, etc. + common_init_result llama_init; + common_init_result llama_init_dft; + llama_model * model = nullptr; llama_context * ctx = nullptr; - std::vector loras; - common_params params; + const llama_vocab * vocab = nullptr; + + llama_model * model_dft = nullptr; + + llama_context_params cparams_dft; llama_batch batch = {}; @@ -612,71 +1662,123 @@ struct server_context { float slot_prompt_similarity = 0.0f; ~server_context() { - if (ctx) { - llama_free(ctx); - ctx = nullptr; - } - - if (model) { - llama_free_model(model); - model = nullptr; - } - // Clear any sampling context for (server_slot & slot : slots) { - if (slot.smpl != nullptr) { - common_sampler_free(slot.smpl); - } + common_sampler_free(slot.smpl); + slot.smpl = nullptr; + + llama_free(slot.ctx_dft); + slot.ctx_dft = nullptr; + + common_speculative_free(slot.spec); + slot.spec = nullptr; + + llama_batch_free(slot.batch_spec); } llama_batch_free(batch); } - bool load_model(const common_params & params_) { - params = params_; + bool load_model(const common_params & params) { + SRV_INF("loading model '%s'\n", params.model.c_str()); - common_init_result llama_init = common_init_from_params(params); + params_base = params; - model = llama_init.model; - ctx = llama_init.context; - loras = llama_init.lora_adapters; + llama_init = common_init_from_params(params_base); + + model = llama_init.model.get(); + ctx = llama_init.context.get(); if (model == nullptr) { - SRV_ERR("failed to load model, '%s'\n", params.model.c_str()); + SRV_ERR("failed to load model, '%s'\n", params_base.model.c_str()); return false; } + vocab = llama_model_get_vocab(model); + n_ctx = llama_n_ctx(ctx); - add_bos_token = llama_add_bos_token(model); - has_eos_token = !llama_add_eos_token(model); + add_bos_token = llama_vocab_get_add_bos(vocab); + has_eos_token = llama_vocab_eos(vocab) != LLAMA_TOKEN_NULL; + + if (!params_base.speculative.model.empty()) { + SRV_INF("loading draft model '%s'\n", params_base.speculative.model.c_str()); + + auto params_dft = params_base; + + params_dft.devices = params_base.speculative.devices; + params_dft.model = params_base.speculative.model; + params_dft.n_ctx = params_base.speculative.n_ctx == 0 ? params_base.n_ctx / params_base.n_parallel : params_base.speculative.n_ctx; + params_dft.n_gpu_layers = params_base.speculative.n_gpu_layers; + params_dft.n_parallel = 1; + + llama_init_dft = common_init_from_params(params_dft); + + model_dft = llama_init_dft.model.get(); + + if (model_dft == nullptr) { + SRV_ERR("failed to load draft model, '%s'\n", params_base.speculative.model.c_str()); + return false; + } + + if (!common_speculative_are_compatible(ctx, llama_init_dft.context.get())) { + SRV_ERR("the draft model '%s' is not compatible with the target model '%s'\n", params_base.speculative.model.c_str(), params_base.model.c_str()); + + return false; + } + + const int n_ctx_dft = llama_n_ctx(llama_init_dft.context.get()); + + cparams_dft = common_context_params_to_llama(params_dft); + cparams_dft.n_batch = n_ctx_dft; + + // force F16 KV cache for the draft model for extra performance + cparams_dft.type_k = GGML_TYPE_F16; + cparams_dft.type_v = GGML_TYPE_F16; + } return true; } - bool validate_model_chat_template() const { + bool validate_builtin_chat_template() const { llama_chat_message chat[] = {{"user", "test"}}; - - const int res = llama_chat_apply_template(model, nullptr, chat, 1, true, nullptr, 0); - - return res > 0; + const char * tmpl = llama_model_chat_template(model); + const int32_t chat_res = llama_chat_apply_template(tmpl, chat, 1, true, nullptr, 0); + return chat_res > 0; } void init() { - const int32_t n_ctx_slot = n_ctx / params.n_parallel; + const int32_t n_ctx_slot = n_ctx / params_base.n_parallel; - SRV_INF("initializing slots, n_slots = %d\n", params.n_parallel); + SRV_INF("initializing slots, n_slots = %d\n", params_base.n_parallel); - for (int i = 0; i < params.n_parallel; i++) { + for (int i = 0; i < params_base.n_parallel; i++) { server_slot slot; slot.id = i; + slot.ctx = ctx; slot.n_ctx = n_ctx_slot; - slot.n_predict = params.n_predict; + slot.n_predict = params_base.n_predict; + + if (model_dft) { + slot.batch_spec = llama_batch_init(params_base.speculative.n_max + 1, 0, 1); + + slot.ctx_dft = llama_init_from_model(model_dft, cparams_dft); + if (slot.ctx_dft == nullptr) { + SRV_ERR("%s", "failed to create draft context\n"); + return; + } + + slot.spec = common_speculative_init(slot.ctx_dft); + if (slot.spec == nullptr) { + SRV_ERR("%s", "failed to create speculator\n"); + return; + } + } SLT_INF(slot, "new slot n_ctx_slot = %d\n", slot.n_ctx); - slot.sparams = params.sparams; + slot.params.sampling = params_base.sampling; slot.callback_on_release = [this](int) { queue_tasks.pop_deferred_task(); @@ -687,8 +1789,7 @@ struct server_context { slots.push_back(slot); } - default_generation_settings_for_props = get_formated_generation(slots.front()); - default_generation_settings_for_props["seed"] = -1; + default_generation_settings_for_props = slots[0].to_json(); // the update_slots() logic will always submit a maximum of n_batch or n_parallel tokens // note that n_batch can be > n_ctx (e.g. for non-causal attention models such as BERT where the KV cache is not used) @@ -696,7 +1797,7 @@ struct server_context { const int32_t n_batch = llama_n_batch(ctx); // only a single seq_id per token is needed - batch = llama_batch_init(std::max(n_batch, params.n_parallel), 0, 1); + batch = llama_batch_init(std::max(n_batch, params_base.n_parallel), 0, 1); } metrics.init(); @@ -732,7 +1833,7 @@ struct server_context { } // length of the Longest Common Subsequence between the current slot's prompt and the input prompt - int cur_lcs_len = longest_common_subsequence(slot.cache_tokens, task.prompt_tokens); + int cur_lcs_len = common_lcs(slot.cache_tokens, task.prompt_tokens); // fraction of the common subsequence length compared to the current slot's prompt length float cur_similarity = static_cast(cur_lcs_len) / static_cast(slot.cache_tokens.size()); @@ -775,87 +1876,20 @@ struct server_context { } bool launch_slot_with_task(server_slot & slot, const server_task & task) { - slot_params default_params; - // Sampling parameter defaults are loaded from the global server context (but individual requests can still override them) - auto default_sparams = params.sparams; - const auto & data = task.data; + slot.reset(); + slot.id_task = task.id; + slot.index = task.index; + slot.task_type = task.type; + slot.params = std::move(task.params); + slot.prompt_tokens = std::move(task.prompt_tokens); - if (data.count("__oaicompat") != 0) { - slot.oaicompat = true; - slot.oaicompat_model = json_value(data, "model", std::string(DEFAULT_OAICOMPAT_MODEL)); - } else { - slot.oaicompat = false; - slot.oaicompat_model = ""; + if (!are_lora_equal(task.params.lora, slot.lora)) { + // if lora is changed, we cannot reuse cached tokens + slot.cache_tokens.clear(); + slot.lora = task.params.lora; } - slot.params.stream = json_value(data, "stream", false); - slot.params.cache_prompt = json_value(data, "cache_prompt", false); - slot.params.n_predict = json_value(data, "n_predict", json_value(data, "max_tokens", default_params.n_predict)); - slot.params.n_indent = json_value(data, "n_indent", default_params.n_indent); - slot.sparams.top_k = json_value(data, "top_k", default_sparams.top_k); - slot.sparams.top_p = json_value(data, "top_p", default_sparams.top_p); - slot.sparams.min_p = json_value(data, "min_p", default_sparams.min_p); - slot.sparams.xtc_probability = json_value(data, "xtc_probability", default_sparams.xtc_probability); - slot.sparams.xtc_threshold = json_value(data, "xtc_threshold", default_sparams.xtc_threshold); - slot.sparams.typ_p = json_value(data, "typical_p", default_sparams.typ_p); - slot.sparams.temp = json_value(data, "temperature", default_sparams.temp); - slot.sparams.dynatemp_range = json_value(data, "dynatemp_range", default_sparams.dynatemp_range); - slot.sparams.dynatemp_exponent = json_value(data, "dynatemp_exponent", default_sparams.dynatemp_exponent); - slot.sparams.penalty_last_n = json_value(data, "repeat_last_n", default_sparams.penalty_last_n); - slot.sparams.penalty_repeat = json_value(data, "repeat_penalty", default_sparams.penalty_repeat); - slot.sparams.penalty_freq = json_value(data, "frequency_penalty", default_sparams.penalty_freq); - slot.sparams.penalty_present = json_value(data, "presence_penalty", default_sparams.penalty_present); - slot.sparams.dry_multiplier = json_value(data, "dry_multiplier", default_sparams.dry_multiplier); - slot.sparams.dry_base = json_value(data, "dry_base", default_sparams.dry_base); - slot.sparams.dry_allowed_length = json_value(data, "dry_allowed_length", default_sparams.dry_allowed_length); - slot.sparams.dry_penalty_last_n = json_value(data, "dry_penalty_last_n", default_sparams.dry_penalty_last_n); - slot.sparams.mirostat = json_value(data, "mirostat", default_sparams.mirostat); - slot.sparams.mirostat_tau = json_value(data, "mirostat_tau", default_sparams.mirostat_tau); - slot.sparams.mirostat_eta = json_value(data, "mirostat_eta", default_sparams.mirostat_eta); - slot.sparams.penalize_nl = json_value(data, "penalize_nl", default_sparams.penalize_nl); - slot.params.n_keep = json_value(data, "n_keep", default_params.n_keep); - slot.params.n_discard = json_value(data, "n_discard", default_params.n_discard); - slot.sparams.seed = json_value(data, "seed", default_sparams.seed); - slot.sparams.n_probs = json_value(data, "n_probs", default_sparams.n_probs); - slot.sparams.min_keep = json_value(data, "min_keep", default_sparams.min_keep); - //slot.params.t_max_prompt_ms = json_value(data, "t_max_prompt_ms", default_params.t_max_prompt_ms); // TODO: implement - slot.params.t_max_predict_ms = json_value(data, "t_max_predict_ms", default_params.t_max_predict_ms); - - if (slot.sparams.dry_base < 1.0f) - { - slot.sparams.dry_base = default_sparams.dry_base; - } - - // sequence breakers for DRY - { - // Currently, this is not compatible with TextGen WebUI, Koboldcpp and SillyTavern format - // Ref: https://github.com/oobabooga/text-generation-webui/blob/d1af7a41ade7bd3c3a463bfa640725edb818ebaf/extensions/openai/typing.py#L39 - - if (data.contains("dry_sequence_breakers")) { - slot.sparams.dry_sequence_breakers = json_value(data, "dry_sequence_breakers", std::vector()); - if (slot.sparams.dry_sequence_breakers.empty()) { - send_error(task, "Error: dry_sequence_breakers must be a non-empty array of strings", ERROR_TYPE_INVALID_REQUEST); - return false; - } - } - } - - // process "json_schema" and "grammar" - if (data.contains("json_schema") && !data.at("json_schema").is_null() && data.contains("grammar") && !data.at("grammar").is_null()) { - send_error(task, "Either \"json_schema\" or \"grammar\" can be specified, but not both", ERROR_TYPE_INVALID_REQUEST); - return false; - } - if (data.contains("json_schema") && !data.contains("grammar")) { - try { - auto schema = json_value(data, "json_schema", json::object()); - slot.sparams.grammar = json_schema_to_grammar(schema); - } catch (const std::exception & e) { - send_error(task, std::string("\"json_schema\": ") + e.what(), ERROR_TYPE_INVALID_REQUEST); - return false; - } - } else { - slot.sparams.grammar = json_value(data, "grammar", default_sparams.grammar); - } + SLT_DBG(slot, "launching slot : %s\n", safe_json_to_str(slot.to_json()).c_str()); if (slot.n_predict > 0 && slot.params.n_predict > slot.n_predict) { // Might be better to reject the request with a 400 ? @@ -863,70 +1897,8 @@ struct server_context { SLT_WRN(slot, "n_predict = %d exceeds server configuration, setting to %d", slot.n_predict, slot.n_predict); } - { - slot.sparams.logit_bias.clear(); - - if (json_value(data, "ignore_eos", false) && has_eos_token) { - slot.sparams.logit_bias.push_back({llama_token_eos(model), -INFINITY}); - } - - const auto & logit_bias = data.find("logit_bias"); - if (logit_bias != data.end() && logit_bias->is_array()) { - const int n_vocab = llama_n_vocab(model); - for (const auto & el : *logit_bias) { - // TODO: we may want to throw errors here, in case "el" is incorrect - if (el.is_array() && el.size() == 2) { - float bias; - if (el[1].is_number()) { - bias = el[1].get(); - } else if (el[1].is_boolean() && !el[1].get()) { - bias = -INFINITY; - } else { - continue; - } - - if (el[0].is_number_integer()) { - llama_token tok = el[0].get(); - if (tok >= 0 && tok < n_vocab) { - slot.sparams.logit_bias.push_back({tok, bias}); - } - } else if (el[0].is_string()) { - auto toks = common_tokenize(model, el[0].get(), false); - for (auto tok : toks) { - slot.sparams.logit_bias.push_back({tok, bias}); - } - } - } - } - } - } - - { - slot.params.antiprompt.clear(); - - const auto & stop = data.find("stop"); - if (stop != data.end() && stop->is_array()) { - for (const auto & word : *stop) { - if (!word.empty()) { - slot.params.antiprompt.push_back(word); - } - } - } - } - - { - const auto & samplers = data.find("samplers"); - if (samplers != data.end() && samplers->is_array()) { - std::vector sampler_names; - for (const auto & name : *samplers) { - if (name.is_string()) { - sampler_names.emplace_back(name); - } - } - slot.sparams.samplers = common_sampler_types_from_names(sampler_names, false); - } else { - slot.sparams.samplers = default_sparams.samplers; - } + if (slot.params.ignore_eos && has_eos_token) { + slot.params.sampling.logit_bias.push_back({llama_vocab_eos(vocab), -INFINITY}); } { @@ -934,7 +1906,7 @@ struct server_context { common_sampler_free(slot.smpl); } - slot.smpl = common_sampler_init(model, slot.sparams); + slot.smpl = common_sampler_init(model, slot.params.sampling); if (slot.smpl == nullptr) { // for now, the only error that may happen here is invalid grammar send_error(task, "Failed to parse grammar", ERROR_TYPE_INVALID_REQUEST); @@ -942,6 +1914,12 @@ struct server_context { } } + if (slot.ctx_dft) { + llama_batch_free(slot.batch_spec); + + slot.batch_spec = llama_batch_init(slot.params.speculative.n_max + 1, 0, 1); + } + slot.state = SLOT_STATE_STARTED; SLT_INF(slot, "%s", "processing task\n"); @@ -959,49 +1937,33 @@ struct server_context { bool process_token(completion_token_output & result, server_slot & slot) { // remember which tokens were sampled - used for repetition penalties during sampling - const std::string token_str = common_token_to_piece(ctx, result.tok, params.special); + const std::string token_str = result.text_to_send; slot.sampled = result.tok; - // search stop word and delete it slot.generated_text += token_str; + if (slot.params.return_tokens) { + slot.generated_tokens.push_back(result.tok); + } slot.has_next_token = true; // check if there is incomplete UTF-8 character at the end - bool incomplete = false; - for (unsigned i = 1; i < 5 && i <= slot.generated_text.size(); ++i) { - unsigned char c = slot.generated_text[slot.generated_text.size() - i]; - if ((c & 0xC0) == 0x80) { - // continuation byte: 10xxxxxx - continue; - } - if ((c & 0xE0) == 0xC0) { - // 2-byte character: 110xxxxx ... - incomplete = i < 2; - } else if ((c & 0xF0) == 0xE0) { - // 3-byte character: 1110xxxx ... - incomplete = i < 3; - } else if ((c & 0xF8) == 0xF0) { - // 4-byte character: 11110xxx ... - incomplete = i < 4; - } - // else 1-byte character or invalid byte - break; - } + bool incomplete = validate_utf8(slot.generated_text) < slot.generated_text.size(); + // search stop word and delete it if (!incomplete) { size_t pos = std::min(slot.n_sent_text, slot.generated_text.size()); const std::string str_test = slot.generated_text.substr(pos); bool send_text = true; - size_t stop_pos = slot.find_stopping_strings(str_test, token_str.size(), STOP_TYPE_FULL); + size_t stop_pos = slot.find_stopping_strings(str_test, token_str.size(), true); if (stop_pos != std::string::npos) { slot.generated_text.erase( slot.generated_text.begin() + pos + stop_pos, slot.generated_text.end()); pos = std::min(slot.n_sent_text, slot.generated_text.size()); } else if (slot.has_next_token) { - stop_pos = slot.find_stopping_strings(str_test, token_str.size(), STOP_TYPE_PARTIAL); + stop_pos = slot.find_stopping_strings(str_test, token_str.size(), false); send_text = stop_pos == std::string::npos; } @@ -1011,6 +1973,8 @@ struct server_context { result.text_to_send = slot.generated_text.substr(pos, std::string::npos); slot.n_sent_text += result.text_to_send.size(); // add the token to slot queue and cache + } else { + result.text_to_send = ""; } slot.add_token(result); @@ -1024,8 +1988,8 @@ struct server_context { } // check the limits - if (slot.n_decoded > 0 && slot.has_next_token && !slot.has_budget(params)) { - slot.stopped_limit = true; + if (slot.n_decoded > 0 && slot.has_next_token && !slot.has_budget(params_base)) { + slot.stop = STOP_TYPE_LIMIT; slot.has_next_token = false; SLT_DBG(slot, "stopped by limit, n_decoded = %d, n_predict = %d\n", slot.n_decoded, slot.params.n_predict); @@ -1034,7 +1998,7 @@ struct server_context { if (slot.has_new_line) { // if we have already seen a new line, we stop after a certain time limit if (slot.params.t_max_predict_ms > 0 && (ggml_time_us() - slot.t_start_generation > 1000.0f*slot.params.t_max_predict_ms)) { - slot.stopped_limit = true; + slot.stop = STOP_TYPE_LIMIT; slot.has_next_token = false; SLT_DBG(slot, "stopped by time limit, n_decoded = %d, t_max_predict_ms = %d ms\n", slot.n_decoded, (int) slot.params.t_max_predict_ms); @@ -1054,7 +2018,7 @@ struct server_context { } if (pos < slot.generated_text.size() && n_indent < slot.params.n_indent) { - slot.stopped_limit = true; + slot.stop = STOP_TYPE_LIMIT; slot.has_next_token = false; // cut the last line @@ -1083,25 +2047,25 @@ struct server_context { // if context shift is disabled, we stop when it reaches the context limit if (slot.n_past >= slot.n_ctx) { slot.truncated = true; - slot.stopped_limit = true; + slot.stop = STOP_TYPE_LIMIT; slot.has_next_token = false; SLT_DBG(slot, "stopped due to running out of context capacity, n_past = %d, n_prompt_tokens = %d, n_decoded = %d, n_ctx = %d\n", slot.n_decoded, slot.n_prompt_tokens, slot.n_past, slot.n_ctx); } - if (llama_token_is_eog(model, result.tok)) { - slot.stopped_eos = true; + if (llama_vocab_is_eog(vocab, result.tok)) { + slot.stop = STOP_TYPE_EOS; slot.has_next_token = false; SLT_DBG(slot, "%s", "stopped by EOS\n"); } - const auto n_ctx_train = llama_n_ctx_train(model); + const auto n_ctx_train = llama_model_n_ctx_train(model); if (slot.params.n_predict < 1 && slot.n_predict < 1 && slot.n_prompt_tokens + slot.n_decoded >= n_ctx_train) { slot.truncated = true; - slot.stopped_limit = true; + slot.stop = STOP_TYPE_LIMIT; slot.has_next_token = false; // stop prediction SLT_WRN(slot, @@ -1115,53 +2079,53 @@ struct server_context { return slot.has_next_token; // continue } - json get_formated_generation(const server_slot & slot) const { - std::vector samplers; - samplers.reserve(slot.sparams.samplers.size()); - for (const auto & sampler : slot.sparams.samplers) { - samplers.emplace_back(common_sampler_type_to_str(sampler)); - } + void populate_token_probs(const server_slot & slot, completion_token_output & result, bool post_sampling, bool special, int idx) { + size_t n_probs = slot.params.sampling.n_probs; + size_t n_vocab = llama_vocab_n_tokens(vocab); + if (post_sampling) { + const auto * cur_p = common_sampler_get_candidates(slot.smpl); + const size_t max_probs = cur_p->size; - return json { - {"n_ctx", slot.n_ctx}, - {"n_predict", slot.n_predict}, // Server configured n_predict - {"model", params.model_alias}, - {"seed", slot.sparams.seed}, - {"seed_cur", slot.smpl ? common_sampler_get_seed(slot.smpl) : 0}, - {"temperature", slot.sparams.temp}, - {"dynatemp_range", slot.sparams.dynatemp_range}, - {"dynatemp_exponent", slot.sparams.dynatemp_exponent}, - {"top_k", slot.sparams.top_k}, - {"top_p", slot.sparams.top_p}, - {"min_p", slot.sparams.min_p}, - {"xtc_probability", slot.sparams.xtc_probability}, - {"xtc_threshold", slot.sparams.xtc_threshold}, - {"typical_p", slot.sparams.typ_p}, - {"repeat_last_n", slot.sparams.penalty_last_n}, - {"repeat_penalty", slot.sparams.penalty_repeat}, - {"presence_penalty", slot.sparams.penalty_present}, - {"frequency_penalty", slot.sparams.penalty_freq}, - {"dry_multiplier", slot.sparams.dry_multiplier}, - {"dry_base", slot.sparams.dry_base}, - {"dry_allowed_length", slot.sparams.dry_allowed_length}, - {"dry_penalty_last_n", slot.sparams.dry_penalty_last_n}, - {"dry_sequence_breakers", slot.sparams.dry_sequence_breakers}, - {"mirostat", slot.sparams.mirostat}, - {"mirostat_tau", slot.sparams.mirostat_tau}, - {"mirostat_eta", slot.sparams.mirostat_eta}, - {"penalize_nl", slot.sparams.penalize_nl}, - {"stop", slot.params.antiprompt}, - {"max_tokens", slot.params.n_predict}, // User configured n_predict - {"n_keep", slot.params.n_keep}, - {"n_discard", slot.params.n_discard}, - {"ignore_eos", slot.sparams.ignore_eos}, - {"stream", slot.params.stream}, - //{"logit_bias", slot.sparams.logit_bias}, - {"n_probs", slot.sparams.n_probs}, - {"min_keep", slot.sparams.min_keep}, - {"grammar", slot.sparams.grammar}, - {"samplers", samplers}, - }; + // set probability for sampled token + for (size_t i = 0; i < max_probs; i++) { + if (cur_p->data[i].id == result.tok) { + result.prob = cur_p->data[i].p; + break; + } + } + + // set probability for top n_probs tokens + result.probs.reserve(max_probs); + for (size_t i = 0; i < std::min(max_probs, n_probs); i++) { + result.probs.push_back({ + cur_p->data[i].id, + common_detokenize(ctx, {cur_p->data[i].id}, special), + cur_p->data[i].p + }); + } + } else { + // TODO: optimize this with min-p optimization + std::vector cur = get_token_probabilities(ctx, idx); + + // set probability for sampled token + for (size_t i = 0; i < n_vocab; i++) { + // set probability for sampled token + if (cur[i].id == result.tok) { + result.prob = cur[i].p; + break; + } + } + + // set probability for top n_probs tokens + result.probs.reserve(n_probs); + for (size_t i = 0; i < std::min(n_vocab, n_probs); i++) { + result.probs.push_back({ + cur[i].id, + common_detokenize(ctx, {cur[i].id}, special), + cur[i].p + }); + } + } } void send_error(const server_task & task, const std::string & error, const enum error_type type = ERROR_TYPE_SERVER) { @@ -1175,110 +2139,100 @@ struct server_context { void send_error(const int id_task, const std::string & error, const enum error_type type = ERROR_TYPE_SERVER) { SRV_ERR("task id = %d, error: %s\n", id_task, error.c_str()); - server_task_result res; - res.id = id_task; - res.stop = false; - res.error = true; - res.data = format_error_response(error, type); + auto res = std::make_unique(); + res->id = id_task; + res->err_type = type; + res->err_msg = error; - queue_results.send(res); + queue_results.send(std::move(res)); } - void send_partial_response(server_slot & slot, completion_token_output tkn) { - server_task_result res; - res.id = slot.id_task; - res.error = false; - res.stop = false; - res.data = json { - {"content", tkn.text_to_send}, - {"stop", false}, - {"id_slot", slot.id}, - {"multimodal", false}, - {"index", slot.index}, - }; + void send_partial_response(server_slot & slot, const completion_token_output & tkn) { + auto res = std::make_unique(); - if (slot.sparams.n_probs > 0) { - const llama_tokens to_send_toks = common_tokenize(ctx, tkn.text_to_send, false); - const size_t probs_pos = std::min(slot.n_sent_token_probs, slot.generated_token_probs.size()); - const size_t probs_stop_pos = std::min(slot.n_sent_token_probs + to_send_toks.size(), slot.generated_token_probs.size()); + res->id = slot.id_task; + res->index = slot.index; + res->content = tkn.text_to_send; + res->tokens = { tkn.tok }; - std::vector probs_output; - if (probs_pos < probs_stop_pos) { - probs_output = std::vector( - slot.generated_token_probs.begin() + probs_pos, - slot.generated_token_probs.begin() + probs_stop_pos); - } - slot.n_sent_token_probs = probs_stop_pos; + res->n_decoded = slot.n_decoded; + res->n_prompt_tokens = slot.n_prompt_tokens; + res->post_sampling_probs = slot.params.post_sampling_probs; - res.data["completion_probabilities"] = probs_vector_to_json(ctx, probs_output); + res->verbose = slot.params.verbose; + res->oaicompat = slot.params.oaicompat; + res->oaicompat_model = slot.params.oaicompat_model; + res->oaicompat_cmpl_id = slot.params.oaicompat_cmpl_id; + + // populate res.probs_output + if (slot.params.sampling.n_probs > 0) { + res->prob_output = tkn; // copy the token probs } - if (slot.oaicompat) { - res.data["oaicompat_token_ctr"] = slot.n_decoded; - res.data["model"] = slot.oaicompat_model; + // populate timings if this is final response or timings_per_token is enabled + if (slot.stop != STOP_TYPE_NONE || slot.params.timings_per_token) { + res->timings = slot.get_timings(); } - queue_results.send(res); + queue_results.send(std::move(res)); } - void send_final_response(const server_slot & slot) { - server_task_result res; - res.id = slot.id_task; - res.error = false; - res.stop = true; - res.data = json { - {"content", !slot.params.stream ? slot.generated_text : ""}, - {"id_slot", slot.id}, - {"stop", true}, - {"model", params.model_alias}, - {"tokens_predicted", slot.n_decoded}, - {"tokens_evaluated", slot.n_prompt_tokens}, - {"generation_settings", get_formated_generation(slot)}, - {"prompt", common_detokenize(ctx, slot.prompt_tokens)}, - {"has_new_line", slot.has_new_line}, - {"truncated", slot.truncated}, - {"stopped_eos", slot.stopped_eos}, - {"stopped_word", slot.stopped_word}, - {"stopped_limit", slot.stopped_limit}, - {"stopping_word", slot.stopping_word}, - {"tokens_cached", slot.n_past}, - {"timings", slot.get_formated_timings()}, - {"index", slot.index}, - }; + void send_final_response(server_slot & slot) { + auto res = std::make_unique(); + res->id = slot.id_task; + res->id_slot = slot.id; - if (slot.sparams.n_probs > 0) { - std::vector probs; - if (!slot.params.stream && slot.stopped_word) { + res->index = slot.index; + res->content = slot.generated_text; + res->tokens = slot.generated_tokens; + res->timings = slot.get_timings(); + res->prompt = common_detokenize(ctx, slot.prompt_tokens, true); + res->response_fields = slot.params.response_fields; + + res->truncated = slot.truncated; + res->n_decoded = slot.n_decoded; + res->n_prompt_tokens = slot.n_prompt_tokens; + res->n_tokens_cached = slot.n_past; + res->has_new_line = slot.has_new_line; + res->stopping_word = slot.stopping_word; + res->stop = slot.stop; + res->post_sampling_probs = slot.params.post_sampling_probs; + + res->verbose = slot.params.verbose; + res->stream = slot.params.stream; + res->oaicompat = slot.params.oaicompat; + res->oaicompat_model = slot.params.oaicompat_model; + res->oaicompat_cmpl_id = slot.params.oaicompat_cmpl_id; + + // populate res.probs_output + if (slot.params.sampling.n_probs > 0) { + if (!slot.params.stream && slot.stop == STOP_TYPE_WORD) { const llama_tokens stop_word_toks = common_tokenize(ctx, slot.stopping_word, false); size_t safe_offset = std::min(slot.generated_token_probs.size(), stop_word_toks.size()); - probs = std::vector( + res->probs_output = std::vector( slot.generated_token_probs.begin(), slot.generated_token_probs.end() - safe_offset); } else { - probs = std::vector( + res->probs_output = std::vector( slot.generated_token_probs.begin(), slot.generated_token_probs.end()); } - - res.data["completion_probabilities"] = probs_vector_to_json(ctx, probs); } - if (slot.oaicompat) { - res.data["oaicompat_token_ctr"] = slot.n_decoded; - res.data["model"] = slot.oaicompat_model; - } + res->generation_params = slot.params; // copy the parameters - queue_results.send(res); + queue_results.send(std::move(res)); } void send_embedding(const server_slot & slot, const llama_batch & batch) { - server_task_result res; - res.id = slot.id_task; - res.error = false; - res.stop = true; + auto res = std::make_unique(); + res->id = slot.id_task; + res->index = slot.index; + res->n_tokens = slot.n_prompt_tokens; + res->oaicompat = slot.params.oaicompat; - const int n_embd = llama_n_embd(model); + const int n_embd = llama_model_n_embd(model); std::vector embd_res(n_embd, 0.0f); @@ -1295,32 +2249,30 @@ struct server_context { if (embd == NULL) { SLT_ERR(slot, "failed to get embeddings, token = %d, seq_id = %d\n", batch.token[i], batch.seq_id[i][0]); - res.data = json { - {"embedding", std::vector(n_embd, 0.0f)}, - {"index", slot.index}, - }; - + res->embedding.push_back(std::vector(n_embd, 0.0f)); continue; } - common_embd_normalize(embd, embd_res.data(), n_embd); - - res.data = json { - {"embedding", embd_res}, - {"index", slot.index}, - }; + // normalize only when there is pooling + // TODO: configurable + if (llama_pooling_type(slot.ctx) != LLAMA_POOLING_TYPE_NONE) { + common_embd_normalize(embd, embd_res.data(), n_embd, 2); + res->embedding.push_back(embd_res); + } else { + res->embedding.push_back({ embd, embd + n_embd }); + } } SLT_DBG(slot, "%s", "sending embeddings\n"); - queue_results.send(res); + queue_results.send(std::move(res)); } void send_rerank(const server_slot & slot, const llama_batch & batch) { - server_task_result res; - res.id = slot.id_task; - res.error = false; - res.stop = true; + auto res = std::make_unique(); + res->id = slot.id_task; + res->index = slot.index; + res->n_tokens = slot.n_prompt_tokens; for (int i = 0; i < batch.n_tokens; ++i) { if (!batch.logits[i] || batch.seq_id[i][0] != slot.id) { @@ -1335,104 +2287,29 @@ struct server_context { if (embd == NULL) { SLT_ERR(slot, "failed to get embeddings, token = %d, seq_id = %d\n", batch.token[i], batch.seq_id[i][0]); - res.data = json { - {"index", slot.index}, - {"score", -1e6}, - }; - + res->score = -1e6; continue; } - res.data = json { - {"index", slot.index}, - {"score", embd[0]}, - }; + res->score = embd[0]; } - SLT_DBG(slot, "sending rerank result, res = '%s'\n", res.data.dump().c_str()); + SLT_DBG(slot, "sending rerank result, res.score = %f\n", res->score); - queue_results.send(res); + queue_results.send(std::move(res)); } // // Functions to create new task(s) and receive result(s) // - // break the input "prompt" into multiple tasks if needed, then format and tokenize the input prompt(s) - std::vector create_tasks_inference(json data, server_task_inf_type inf_type) { - std::vector tasks; - auto create_task = [&](json & task_data, llama_tokens & prompt_tokens) { - SRV_DBG("create task, n_tokens = %d\n", (int) prompt_tokens.size()); - server_task task; - task.id = queue_tasks.get_new_id(); - task.inf_type = inf_type; - task.type = SERVER_TASK_TYPE_INFERENCE; - task.data = task_data; - task.prompt_tokens = std::move(prompt_tokens); - tasks.push_back(std::move(task)); - }; - - static constexpr const char * error_msg = "\"prompt\" must be a string, an array of token ids or an array of prompts"; - if (!data.contains("prompt")) { - throw std::runtime_error(error_msg); - } - - // because llama_tokenize api is thread-safe, we can tokenize the prompt from HTTP thread - bool add_special = inf_type != SERVER_TASK_INF_TYPE_RERANK && inf_type != SERVER_TASK_INF_TYPE_INFILL; - std::vector tokenized_prompts = tokenize_input_prompts(ctx, data.at("prompt"), add_special, true); - switch (inf_type) { - case SERVER_TASK_INF_TYPE_RERANK: - { - // prompts[0] is the question - // the rest are the answers/documents - GGML_ASSERT(tokenized_prompts.size() > 1); - SRV_DBG("creating rerank tasks, n_prompts = %d\n", (int) tokenized_prompts.size() - 1); - for (size_t i = 1; i < tokenized_prompts.size(); i++) { - data["index"] = i - 1; - auto tokens = format_rerank(model, tokenized_prompts[0], tokenized_prompts[i]); - create_task(data, tokens); - } - } break; - case SERVER_TASK_INF_TYPE_INFILL: - { - SRV_DBG("creating infill tasks, n_prompts = %d\n", (int) tokenized_prompts.size()); - for (size_t i = 0; i < tokenized_prompts.size(); i++) { - data["index"] = i; - auto tokens = format_infill( - ctx, - data.at("input_prefix"), - data.at("input_suffix"), - data.at("input_extra"), - params.n_batch, - params.n_predict, - slots[0].n_ctx, // TODO: there should be a better way - params.spm_infill, - tokenized_prompts[i] - ); - create_task(data, tokens); - } - } break; - default: - { - SRV_DBG("creating multi-prompt tasks, n_prompts = %d\n", (int) tokenized_prompts.size()); - for (size_t i = 0; i < tokenized_prompts.size(); i++) { - data["index"] = i; - create_task(data, tokenized_prompts[i]); - } - } - } - - return tasks; - } - void cancel_tasks(const std::unordered_set & id_tasks) { std::vector cancel_tasks; cancel_tasks.reserve(id_tasks.size()); for (const auto & id_task : id_tasks) { SRV_WRN("cancel task, id_task = %d\n", id_task); - server_task task; - task.type = SERVER_TASK_TYPE_CANCEL; + server_task task(SERVER_TASK_TYPE_CANCEL); task.id_target = id_task; cancel_tasks.push_back(task); queue_results.remove_waiting_task_id(id_task); @@ -1441,50 +2318,58 @@ struct server_context { queue_tasks.post(cancel_tasks, true); } - // receive the results from task(s) created by create_tasks_inference - void receive_cmpl_results( + // receive the results from task(s) + void receive_multi_results( const std::unordered_set & id_tasks, - const std::function&)> & result_handler, + const std::function&)> & result_handler, const std::function & error_handler) { - // TODO: currently, there is no way to detect the client has cancelled the request - std::vector results(id_tasks.size()); + std::vector results(id_tasks.size()); for (size_t i = 0; i < id_tasks.size(); i++) { - server_task_result result = queue_results.recv(id_tasks); + server_task_result_ptr result = queue_results.recv(id_tasks); - if (result.error) { - error_handler(result.data); + if (result->is_error()) { + error_handler(result->to_json()); cancel_tasks(id_tasks); return; } - const size_t idx = result.data["index"]; + GGML_ASSERT( + dynamic_cast(result.get()) != nullptr + || dynamic_cast(result.get()) != nullptr + || dynamic_cast(result.get()) != nullptr + ); + const size_t idx = result->get_index(); GGML_ASSERT(idx < results.size() && "index out of range"); - - results[idx] = result; + results[idx] = std::move(result); } result_handler(results); } - // receive the results from task(s) created by create_tasks_inference, in stream mode + // receive the results from task(s), in stream mode void receive_cmpl_results_stream( - const std::unordered_set & id_tasks, const - std::function & result_handler, const - std::function & error_handler) { + const std::unordered_set & id_tasks, + const std::function & result_handler, + const std::function & error_handler) { size_t n_finished = 0; while (true) { - server_task_result result = queue_results.recv(id_tasks); + server_task_result_ptr result = queue_results.recv(id_tasks); + + if (result->is_error()) { + error_handler(result->to_json()); + cancel_tasks(id_tasks); + return; + } + + GGML_ASSERT( + dynamic_cast(result.get()) != nullptr + || dynamic_cast(result.get()) != nullptr + ); if (!result_handler(result)) { cancel_tasks(id_tasks); break; } - if (result.error) { - error_handler(result.data); - cancel_tasks(id_tasks); - break; - } - - if (result.stop) { + if (result->is_stop()) { if (++n_finished == id_tasks.size()) { break; } @@ -1498,9 +2383,12 @@ struct server_context { void process_single_task(server_task task) { switch (task.type) { - case SERVER_TASK_TYPE_INFERENCE: + case SERVER_TASK_TYPE_COMPLETION: + case SERVER_TASK_TYPE_INFILL: + case SERVER_TASK_TYPE_EMBEDDING: + case SERVER_TASK_TYPE_RERANK: { - const int id_slot = json_value(task.data, "id_slot", -1); + const int id_slot = task.id_selected_slot; server_slot * slot = id_slot != -1 ? get_slot_by_id(id_slot) : get_available_slot(task); @@ -1517,13 +2405,6 @@ struct server_context { break; } - slot->reset(); - - slot->id_task = task.id; - slot->inf_type = task.inf_type; - slot->index = json_value(task.data, "index", 0); - slot->prompt_tokens = std::move(task.prompt_tokens); - if (!launch_slot_with_task(*slot, task)) { SRV_ERR("failed to launch slot with task, id_task = %d\n", task.id); break; @@ -1551,21 +2432,7 @@ struct server_context { int n_processing_slots = 0; for (server_slot & slot : slots) { - json slot_data = get_formated_generation(slot); - slot_data["id"] = slot.id; - slot_data["id_task"] = slot.id_task; - slot_data["is_processing"] = slot.is_processing(); - slot_data["prompt"] = common_detokenize(ctx, slot.prompt_tokens); - slot_data["next_token"] = { - {"has_next_token", slot.has_next_token}, - {"has_new_line", slot.has_new_line}, - {"n_remain", slot.n_remaining}, - {"n_decoded", slot.n_decoded}, - {"stopped_eos", slot.stopped_eos}, - {"stopped_word", slot.stopped_word}, - {"stopped_limit", slot.stopped_limit}, - {"stopping_word", slot.stopping_word}, - }; + json slot_data = slot.to_json(); if (slot.is_processing()) { n_processing_slots++; @@ -1577,43 +2444,38 @@ struct server_context { } SRV_DBG("n_idle_slots = %d, n_processing_slots = %d\n", n_idle_slots, n_processing_slots); - server_task_result res; - res.id = task.id; - res.stop = true; - res.error = false; - res.data = { - { "idle", n_idle_slots }, - { "processing", n_processing_slots }, - { "deferred", queue_tasks.queue_tasks_deferred.size() }, - { "t_start", metrics.t_start}, + auto res = std::make_unique(); + res->id = task.id; + res->slots_data = std::move(slots_data); + res->n_idle_slots = n_idle_slots; + res->n_processing_slots = n_processing_slots; + res->n_tasks_deferred = queue_tasks.queue_tasks_deferred.size(); + res->t_start = metrics.t_start; - { "n_prompt_tokens_processed_total", metrics.n_prompt_tokens_processed_total}, - { "t_tokens_generation_total", metrics.t_tokens_generation_total}, - { "n_tokens_predicted_total", metrics.n_tokens_predicted_total}, - { "t_prompt_processing_total", metrics.t_prompt_processing_total}, + res->kv_cache_tokens_count = llama_get_kv_cache_token_count(ctx); + res->kv_cache_used_cells = llama_get_kv_cache_used_cells(ctx); - { "n_prompt_tokens_processed", metrics.n_prompt_tokens_processed}, - { "t_prompt_processing", metrics.t_prompt_processing}, - { "n_tokens_predicted", metrics.n_tokens_predicted}, - { "t_tokens_generation", metrics.t_tokens_generation}, + res->n_prompt_tokens_processed_total = metrics.n_prompt_tokens_processed_total; + res->t_prompt_processing_total = metrics.t_prompt_processing_total; + res->n_tokens_predicted_total = metrics.n_tokens_predicted_total; + res->t_tokens_generation_total = metrics.t_tokens_generation_total; - { "n_decode_total", metrics.n_decode_total}, - { "n_busy_slots_total", metrics.n_busy_slots_total}, + res->n_prompt_tokens_processed = metrics.n_prompt_tokens_processed; + res->t_prompt_processing = metrics.t_prompt_processing; + res->n_tokens_predicted = metrics.n_tokens_predicted; + res->t_tokens_generation = metrics.t_tokens_generation; - { "kv_cache_tokens_count", llama_get_kv_cache_token_count(ctx)}, - { "kv_cache_used_cells", llama_get_kv_cache_used_cells(ctx)}, + res->n_decode_total = metrics.n_decode_total; + res->n_busy_slots_total = metrics.n_busy_slots_total; - { "slots", slots_data }, - }; - - if (json_value(task.data, "reset_bucket", false)) { + if (task.metrics_reset_bucket) { metrics.reset_bucket(); } - queue_results.send(res); + queue_results.send(std::move(res)); } break; case SERVER_TASK_TYPE_SLOT_SAVE: { - int id_slot = task.data.at("id_slot"); + int id_slot = task.slot_action.slot_id; server_slot * slot = get_slot_by_id(id_slot); if (slot == nullptr) { send_error(task, "Invalid slot ID", ERROR_TYPE_INVALID_REQUEST); @@ -1629,32 +2491,27 @@ struct server_context { const size_t token_count = slot->cache_tokens.size(); const int64_t t_start = ggml_time_us(); - std::string filename = task.data.at("filename"); - std::string filepath = task.data.at("filepath"); + std::string filename = task.slot_action.filename; + std::string filepath = task.slot_action.filepath; const size_t nwrite = llama_state_seq_save_file(ctx, filepath.c_str(), slot->id, slot->cache_tokens.data(), token_count); const int64_t t_end = ggml_time_us(); const double t_save_ms = (t_end - t_start) / 1000.0; - server_task_result result; - result.id = task.id; - result.stop = true; - result.error = false; - result.data = json { - { "id_slot", id_slot }, - { "filename", filename }, - { "n_saved", token_count }, // tokens saved - { "n_written", nwrite }, // bytes written - { "timings", { - { "save_ms", t_save_ms } - } } - }; - queue_results.send(result); + auto res = std::make_unique(); + res->id = task.id; + res->id_slot = id_slot; + res->filename = filename; + res->is_save = true; + res->n_tokens = token_count; + res->n_bytes = nwrite; + res->t_ms = t_save_ms; + queue_results.send(std::move(res)); } break; case SERVER_TASK_TYPE_SLOT_RESTORE: { - int id_slot = task.data.at("id_slot"); + int id_slot = task.slot_action.slot_id; server_slot * slot = get_slot_by_id(id_slot); if (slot == nullptr) { send_error(task, "Invalid slot ID", ERROR_TYPE_INVALID_REQUEST); @@ -1669,8 +2526,8 @@ struct server_context { const int64_t t_start = ggml_time_us(); - std::string filename = task.data.at("filename"); - std::string filepath = task.data.at("filepath"); + std::string filename = task.slot_action.filename; + std::string filepath = task.slot_action.filepath; slot->cache_tokens.resize(slot->n_ctx); size_t token_count = 0; @@ -1685,24 +2542,19 @@ struct server_context { const int64_t t_end = ggml_time_us(); const double t_restore_ms = (t_end - t_start) / 1000.0; - server_task_result result; - result.id = task.id; - result.stop = true; - result.error = false; - result.data = json { - { "id_slot", id_slot }, - { "filename", filename }, - { "n_restored", token_count }, // tokens restored - { "n_read", nread }, // bytes read - { "timings", { - { "restore_ms", t_restore_ms } - } } - }; - queue_results.send(result); + auto res = std::make_unique(); + res->id = task.id; + res->id_slot = id_slot; + res->filename = filename; + res->is_save = false; + res->n_tokens = token_count; + res->n_bytes = nread; + res->t_ms = t_restore_ms; + queue_results.send(std::move(res)); } break; case SERVER_TASK_TYPE_SLOT_ERASE: { - int id_slot = task.data.at("id_slot"); + int id_slot = task.slot_action.slot_id; server_slot * slot = get_slot_by_id(id_slot); if (slot == nullptr) { send_error(task, "Invalid slot ID", ERROR_TYPE_INVALID_REQUEST); @@ -1720,25 +2572,18 @@ struct server_context { llama_kv_cache_seq_rm(ctx, slot->id, -1, -1); slot->cache_tokens.clear(); - server_task_result result; - result.id = task.id; - result.stop = true; - result.error = false; - result.data = json { - { "id_slot", id_slot }, - { "n_erased", n_erased } - }; - queue_results.send(result); + auto res = std::make_unique(); + res->id = task.id; + res->id_slot = id_slot; + res->n_erased = n_erased; + queue_results.send(std::move(res)); } break; case SERVER_TASK_TYPE_SET_LORA: { - common_lora_adapters_apply(ctx, loras); - server_task_result result; - result.id = task.id; - result.stop = true; - result.error = false; - result.data = json{{ "success", true }}; - queue_results.send(result); + params_base.lora_adapters = std::move(task.set_lora); + auto res = std::make_unique(); + res->id = task.id; + queue_results.send(std::move(res)); } break; } } @@ -1768,10 +2613,8 @@ struct server_context { { SRV_DBG("%s", "posting NEXT_RESPONSE\n"); - server_task task; - task.type = SERVER_TASK_TYPE_NEXT_RESPONSE; - task.id_target = -1; - + server_task task(SERVER_TASK_TYPE_NEXT_RESPONSE); + task.id = queue_tasks.get_new_id(); queue_tasks.post(task); } @@ -1779,7 +2622,7 @@ struct server_context { // TODO: simplify and improve for (server_slot & slot : slots) { if (slot.is_processing() && slot.n_past + 1 >= slot.n_ctx) { - if (!params.ctx_shift) { + if (!params_base.ctx_shift) { // this check is redundant (for good) // we should never get here, because generation should already stopped in process_token() slot.release(); @@ -1814,12 +2657,22 @@ struct server_context { // start populating the batch for this iteration common_batch_clear(batch); + // track if given slot can be batched with slots already in the batch + server_slot * slot_batched = nullptr; + // frist, add sampled tokens from any ongoing sequences for (auto & slot : slots) { if (slot.state != SLOT_STATE_GENERATING) { continue; } + // check if we can batch this slot with the previous one + if (!slot_batched) { + slot_batched = &slot; + } else if (!slot_batched->can_batch_with(slot)) { + continue; + } + slot.i_batch = batch.n_tokens; common_batch_add(batch, slot.sampled, slot.n_past, { slot.id }, true); @@ -1838,15 +2691,18 @@ struct server_context { int32_t n_batch = llama_n_batch(ctx); int32_t n_ubatch = llama_n_ubatch(ctx); - // track if this is an embedding or non-embedding batch - // if we've added sampled tokens above, we are in non-embedding mode - // -1: none, 0: non-embedding, 1: embedding - // TODO: make enum - int32_t batch_type = batch.n_tokens > 0 ? 0 : -1; - // next, batch any pending prompts without exceeding n_batch - if (params.cont_batching || batch.n_tokens == 0) { + if (params_base.cont_batching || batch.n_tokens == 0) { for (auto & slot : slots) { + // check if we can batch this slot with the previous one + if (slot.is_processing()) { + if (!slot_batched) { + slot_batched = &slot; + } else if (!slot_batched->can_batch_with(slot)) { + continue; + } + } + // this slot still has a prompt to be processed if (slot.state == SLOT_STATE_PROCESSING_PROMPT || slot.state == SLOT_STATE_STARTED) { auto & prompt_tokens = slot.prompt_tokens; @@ -1885,7 +2741,7 @@ struct server_context { continue; } - if (slot.inf_type == SERVER_TASK_INF_TYPE_EMBEDDING || slot.inf_type == SERVER_TASK_INF_TYPE_RERANK) { + if (slot.is_non_causal()) { if (slot.n_prompt_tokens > n_ubatch) { slot.release(); send_error(slot, "input is too large to process. increase the physical batch size", ERROR_TYPE_SERVER); @@ -1898,7 +2754,7 @@ struct server_context { continue; } } else { - if (!params.ctx_shift) { + if (!params_base.ctx_shift) { // if context shift is disabled, we make sure prompt size is smaller than KV size // TODO: there should be a separate parameter that control prompt truncation // context shift should be applied only during the generation phase @@ -1941,14 +2797,14 @@ struct server_context { if (slot.params.cache_prompt) { // reuse any previously computed tokens that are common with the new prompt - slot.n_past = longest_common_prefix(slot.cache_tokens, prompt_tokens); + slot.n_past = common_lcp(slot.cache_tokens, prompt_tokens); // reuse chunks from the cached prompt by shifting their KV cache in the new position - if (params.n_cache_reuse > 0) { + if (params_base.n_cache_reuse > 0) { size_t head_c = slot.n_past; // cache size_t head_p = slot.n_past; // current prompt - SLT_DBG(slot, "trying to reuse chunks with size > %d, slot.n_past = %d\n", params.n_cache_reuse, slot.n_past); + SLT_DBG(slot, "trying to reuse chunks with size > %d, slot.n_past = %d\n", params_base.n_cache_reuse, slot.n_past); while (head_c < slot.cache_tokens.size() && head_p < prompt_tokens.size()) { @@ -1961,7 +2817,7 @@ struct server_context { n_match++; } - if (n_match >= (size_t) params.n_cache_reuse) { + if (n_match >= (size_t) params_base.n_cache_reuse) { SLT_INF(slot, "reusing chunk with size %zu, shifting KV cache [%zu, %zu) -> [%zu, %zu)\n", n_match, head_c, head_c + n_match, head_p, head_p + n_match); //for (size_t i = head_p; i < head_p + n_match; i++) { // SLT_DBG(slot, "cache token %3zu: %6d '%s'\n", i, prompt_tokens[i], common_token_to_piece(ctx, prompt_tokens[i]).c_str()); @@ -2000,24 +2856,13 @@ struct server_context { } // non-causal tasks require to fit the entire prompt in the physical batch - if (slot.inf_type == SERVER_TASK_INF_TYPE_EMBEDDING || slot.inf_type == SERVER_TASK_INF_TYPE_RERANK) { + if (slot.is_non_causal()) { // cannot fit the prompt in the current batch - will try next iter if (batch.n_tokens + slot.n_prompt_tokens > n_batch) { continue; } } - // check that we are in the right batch_type, if not defer the slot - const bool slot_type = - slot.inf_type == SERVER_TASK_INF_TYPE_EMBEDDING || - slot.inf_type == SERVER_TASK_INF_TYPE_RERANK ? 1 : 0; - - if (batch_type == -1) { - batch_type = slot_type; - } else if (batch_type != slot_type) { - continue; - } - // keep only the common part if (!llama_kv_cache_seq_rm(ctx, slot.id, slot.n_past, -1)) { // could not partially delete (likely using a non-Transformer model) @@ -2034,7 +2879,10 @@ struct server_context { // add prompt tokens for processing in the current batch while (slot.n_past < slot.n_prompt_tokens && batch.n_tokens < n_batch) { - common_batch_add(batch, prompt_tokens[slot.n_past], slot.n_past, { slot.id }, false); + // without pooling, we want to output the embeddings for all the tokens in the batch + const bool need_embd = slot.task_type == SERVER_TASK_TYPE_EMBEDDING && llama_pooling_type(slot.ctx) == LLAMA_POOLING_TYPE_NONE; + + common_batch_add(batch, prompt_tokens[slot.n_past], slot.n_past, { slot.id }, need_embd); if (slot.params.cache_prompt) { slot.cache_tokens.push_back(prompt_tokens[slot.n_past]); @@ -2082,8 +2930,12 @@ struct server_context { SRV_DBG("decoding batch, n_tokens = %d\n", batch.n_tokens); - // make sure we're in the right embedding mode - llama_set_embeddings(ctx, batch_type == 1); + if (slot_batched) { + // make sure we're in the right embedding mode + llama_set_embeddings(ctx, slot_batched->is_non_causal()); + // apply lora, only need to do it once per batch + common_set_adapter_lora(ctx, slot_batched->lora); + } // process the created batch of tokens for (int32_t i = 0; i < batch.n_tokens; i += n_batch) { @@ -2128,7 +2980,7 @@ struct server_context { } if (slot.state == SLOT_STATE_DONE_PROMPT) { - if (slot.inf_type == SERVER_TASK_INF_TYPE_EMBEDDING) { + if (slot.task_type == SERVER_TASK_TYPE_EMBEDDING) { // prompt evaluated for embedding send_embedding(slot, batch_view); slot.release(); @@ -2136,7 +2988,7 @@ struct server_context { continue; // continue loop of slots } - if (slot.inf_type == SERVER_TASK_INF_TYPE_RERANK) { + if (slot.task_type == SERVER_TASK_TYPE_RERANK) { send_rerank(slot, batch_view); slot.release(); slot.i_batch = -1; @@ -2149,27 +3001,33 @@ struct server_context { continue; // continue loop of slots } - completion_token_output result; - const llama_token id = common_sampler_sample(slot.smpl, ctx, slot.i_batch - i); + const int tok_idx = slot.i_batch - i; + + llama_token id = common_sampler_sample(slot.smpl, ctx, tok_idx); + + slot.i_batch = -1; common_sampler_accept(slot.smpl, id, true); slot.n_decoded += 1; + + const int64_t t_current = ggml_time_us(); + if (slot.n_decoded == 1) { - slot.t_start_generation = ggml_time_us(); + slot.t_start_generation = t_current; slot.t_prompt_processing = (slot.t_start_generation - slot.t_start_process_prompt) / 1e3; metrics.on_prompt_eval(slot); } - result.tok = id; + slot.t_token_generation = (t_current - slot.t_start_generation) / 1e3; - const auto * cur_p = common_sampler_get_candidates(slot.smpl); + completion_token_output result; + result.tok = id; + result.text_to_send = common_token_to_piece(ctx, result.tok, params_base.special); + result.prob = 1.0f; // TODO: set it here instead of doing inside populate_token_probs - for (size_t i = 0; i < (size_t) slot.sparams.n_probs; ++i) { - result.probs.push_back({ - cur_p->data[i].id, - i >= cur_p->size ? 0.0f : cur_p->data[i].p, - }); + if (slot.params.sampling.n_probs > 0) { + populate_token_probs(slot, result, slot.params.post_sampling_probs, params_base.special, tok_idx); } if (!process_token(result, slot)) { @@ -2178,9 +3036,98 @@ struct server_context { slot.print_timings(); send_final_response(slot); metrics.on_prediction(slot); + continue; + } + } + + // do speculative decoding + for (auto & slot : slots) { + if (!slot.is_processing() || !slot.can_speculate()) { + continue; } - slot.i_batch = -1; + if (slot.state != SLOT_STATE_GENERATING) { + continue; + } + + // determine the max draft that fits the current slot state + int n_draft_max = slot.params.speculative.n_max; + + // note: n_past is not yet increased for the `id` token sampled above + // also, need to leave space for 1 extra token to allow context shifts + n_draft_max = std::min(n_draft_max, slot.n_ctx - slot.n_past - 2); + + if (slot.n_remaining > 0) { + n_draft_max = std::min(n_draft_max, slot.n_remaining - 1); + } + + SLT_DBG(slot, "max possible draft: %d\n", n_draft_max); + + if (n_draft_max < slot.params.speculative.n_min) { + SLT_DBG(slot, "the max possible draft is too small: %d < %d - skipping speculative decoding\n", n_draft_max, slot.params.speculative.n_min); + + continue; + } + + llama_token id = slot.sampled; + + struct common_speculative_params params_spec; + params_spec.n_draft = n_draft_max; + params_spec.n_reuse = llama_n_ctx(slot.ctx_dft) - slot.params.speculative.n_max; + params_spec.p_min = slot.params.speculative.p_min; + + llama_tokens draft = common_speculative_gen_draft(slot.spec, params_spec, slot.cache_tokens, id); + + // ignore small drafts + if (slot.params.speculative.n_min > (int) draft.size()) { + SLT_DBG(slot, "ignoring small draft: %d < %d\n", (int) draft.size(), slot.params.speculative.n_min); + + continue; + } + + // construct the speculation batch + common_batch_clear(slot.batch_spec); + common_batch_add (slot.batch_spec, id, slot.n_past, { slot.id }, true); + + for (size_t i = 0; i < draft.size(); ++i) { + common_batch_add(slot.batch_spec, draft[i], slot.n_past + 1 + i, { slot.id }, true); + } + + SLT_DBG(slot, "decoding speculative batch, size = %d\n", slot.batch_spec.n_tokens); + + llama_decode(ctx, slot.batch_spec); + + // the accepted tokens from the speculation + const auto ids = common_sampler_sample_and_accept_n(slot.smpl, ctx, draft); + + slot.n_past += ids.size(); + slot.n_decoded += ids.size(); + + slot.cache_tokens.push_back(id); + slot.cache_tokens.insert(slot.cache_tokens.end(), ids.begin(), ids.end() - 1); + + llama_kv_cache_seq_rm(ctx, slot.id, slot.n_past, -1); + + for (size_t i = 0; i < ids.size(); ++i) { + completion_token_output result; + + result.tok = ids[i]; + result.text_to_send = common_token_to_piece(ctx, result.tok, params_base.special); + result.prob = 1.0f; // set later + + // TODO: set result.probs + + if (!process_token(result, slot)) { + // release slot because of stop condition + slot.release(); + slot.print_timings(); + send_final_response(slot); + metrics.on_prediction(slot); + break; + } + } + + SLT_DBG(slot, "accepted %d/%d draft tokens, new n_past = %d\n", (int) ids.size() - 1, (int) draft.size(), slot.n_past); } } @@ -2189,12 +3136,12 @@ struct server_context { json model_meta() const { return json { - {"vocab_type", llama_vocab_type (model)}, - {"n_vocab", llama_n_vocab (model)}, - {"n_ctx_train", llama_n_ctx_train (model)}, - {"n_embd", llama_n_embd (model)}, - {"n_params", llama_model_n_params(model)}, - {"size", llama_model_size (model)}, + {"vocab_type", llama_vocab_type (vocab)}, + {"n_vocab", llama_vocab_n_tokens (vocab)}, + {"n_ctx_train", llama_model_n_ctx_train(model)}, + {"n_embd", llama_model_n_embd (model)}, + {"n_params", llama_model_n_params (model)}, + {"size", llama_model_size (model)}, }; } }; @@ -2235,17 +3182,9 @@ int main(int argc, char ** argv) { common_init(); - // enabling this will output extra debug information in the HTTP responses from the server - // see format_final_response_oaicompat() - const bool verbose = params.verbosity > 9; - // struct that contains llama context and inference server_context ctx_server; - if (params.model_alias == "unknown") { - params.model_alias = params.model; - } - llama_backend_init(); llama_numa_init(params.numa); @@ -2280,20 +3219,20 @@ int main(int argc, char ** argv) { auto res_error = [](httplib::Response & res, const json & error_data) { json final_response {{"error", error_data}}; - res.set_content(final_response.dump(-1, ' ', false, json::error_handler_t::replace), MIMETYPE_JSON); + res.set_content(safe_json_to_str(final_response), MIMETYPE_JSON); res.status = json_value(error_data, "code", 500); }; auto res_ok = [](httplib::Response & res, const json & data) { - res.set_content(data.dump(-1, ' ', false, json::error_handler_t::replace), MIMETYPE_JSON); + res.set_content(safe_json_to_str(data), MIMETYPE_JSON); res.status = 200; }; - svr->set_exception_handler([&res_error](const httplib::Request &, httplib::Response & res, std::exception_ptr ep) { + svr->set_exception_handler([&res_error](const httplib::Request &, httplib::Response & res, const std::exception_ptr & ep) { std::string message; try { std::rethrow_exception(ep); - } catch (std::exception & e) { + } catch (const std::exception & e) { message = e.what(); } catch (...) { message = "Unknown Exception"; @@ -2346,8 +3285,8 @@ int main(int argc, char ** argv) { return true; } - // If path is public, skip validation - if (public_endpoints.find(req.path) != public_endpoints.end()) { + // If path is public or is static file, skip validation + if (public_endpoints.find(req.path) != public_endpoints.end() || req.path == "/") { return true; } @@ -2422,27 +3361,33 @@ int main(int argc, char ** argv) { } // request slots data using task queue - server_task task; + server_task task(SERVER_TASK_TYPE_METRICS); task.id = ctx_server.queue_tasks.get_new_id(); - task.type = SERVER_TASK_TYPE_METRICS; - ctx_server.queue_results.add_waiting_task_id(task.id); ctx_server.queue_tasks.post(task, true); // high-priority task // get the result - server_task_result result = ctx_server.queue_results.recv(task.id); + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); ctx_server.queue_results.remove_waiting_task_id(task.id); + if (result->is_error()) { + res_error(res, result->to_json()); + return; + } + + // TODO: get rid of this dynamic_cast + auto res_metrics = dynamic_cast(result.get()); + GGML_ASSERT(res_metrics != nullptr); + // optionally return "fail_on_no_slot" error - const int n_idle_slots = result.data.at("idle"); if (req.has_param("fail_on_no_slot")) { - if (n_idle_slots == 0) { + if (res_metrics->n_idle_slots == 0) { res_error(res, format_error_response("no slot available", ERROR_TYPE_UNAVAILABLE)); return; } } - res_ok(res, result.data.at("slots")); + res_ok(res, res_metrics->slots_data); }; const auto handle_metrics = [&](const httplib::Request &, httplib::Response & res) { @@ -2452,83 +3397,77 @@ int main(int argc, char ** argv) { } // request slots data using task queue - server_task task; + server_task task(SERVER_TASK_TYPE_METRICS); task.id = ctx_server.queue_tasks.get_new_id(); - task.id_target = -1; - task.type = SERVER_TASK_TYPE_METRICS; - task.data.push_back({{"reset_bucket", true}}); + task.metrics_reset_bucket = true; ctx_server.queue_results.add_waiting_task_id(task.id); ctx_server.queue_tasks.post(task, true); // high-priority task // get the result - server_task_result result = ctx_server.queue_results.recv(task.id); + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); ctx_server.queue_results.remove_waiting_task_id(task.id); - json data = result.data; + if (result->is_error()) { + res_error(res, result->to_json()); + return; + } - const uint64_t n_prompt_tokens_processed = data.at("n_prompt_tokens_processed"); - const uint64_t t_prompt_processing = data.at("t_prompt_processing"); - - const uint64_t n_tokens_predicted = data.at("n_tokens_predicted"); - const uint64_t t_tokens_generation = data.at("t_tokens_generation"); - - const uint64_t n_decode_total = data.at("n_decode_total"); - const uint64_t n_busy_slots_total = data.at("n_busy_slots_total"); - - const int32_t kv_cache_used_cells = data.at("kv_cache_used_cells"); + // TODO: get rid of this dynamic_cast + auto res_metrics = dynamic_cast(result.get()); + GGML_ASSERT(res_metrics != nullptr); // metrics definition: https://prometheus.io/docs/practices/naming/#metric-names json all_metrics_def = json { {"counter", {{ {"name", "prompt_tokens_total"}, {"help", "Number of prompt tokens processed."}, - {"value", (uint64_t) data.at("n_prompt_tokens_processed_total")} + {"value", (uint64_t) res_metrics->n_prompt_tokens_processed_total} }, { {"name", "prompt_seconds_total"}, {"help", "Prompt process time"}, - {"value", (uint64_t) data.at("t_prompt_processing_total") / 1.e3} + {"value", (uint64_t) res_metrics->t_prompt_processing_total / 1.e3} }, { {"name", "tokens_predicted_total"}, {"help", "Number of generation tokens processed."}, - {"value", (uint64_t) data.at("n_tokens_predicted_total")} + {"value", (uint64_t) res_metrics->n_tokens_predicted_total} }, { {"name", "tokens_predicted_seconds_total"}, {"help", "Predict process time"}, - {"value", (uint64_t) data.at("t_tokens_generation_total") / 1.e3} + {"value", (uint64_t) res_metrics->t_tokens_generation_total / 1.e3} }, { {"name", "n_decode_total"}, {"help", "Total number of llama_decode() calls"}, - {"value", n_decode_total} + {"value", res_metrics->n_decode_total} }, { {"name", "n_busy_slots_per_decode"}, {"help", "Average number of busy slots per llama_decode() call"}, - {"value", (float) n_busy_slots_total / (float) n_decode_total} + {"value", (float) res_metrics->n_busy_slots_total / (float) res_metrics->n_decode_total} }}}, {"gauge", {{ {"name", "prompt_tokens_seconds"}, {"help", "Average prompt throughput in tokens/s."}, - {"value", n_prompt_tokens_processed ? 1.e3 / t_prompt_processing * n_prompt_tokens_processed : 0.} + {"value", res_metrics->n_prompt_tokens_processed ? 1.e3 / res_metrics->t_prompt_processing * res_metrics->n_prompt_tokens_processed : 0.} },{ {"name", "predicted_tokens_seconds"}, {"help", "Average generation throughput in tokens/s."}, - {"value", n_tokens_predicted ? 1.e3 / t_tokens_generation * n_tokens_predicted : 0.} + {"value", res_metrics->n_tokens_predicted ? 1.e3 / res_metrics->t_tokens_generation * res_metrics->n_tokens_predicted : 0.} },{ {"name", "kv_cache_usage_ratio"}, {"help", "KV-cache usage. 1 means 100 percent usage."}, - {"value", 1. * kv_cache_used_cells / params.n_ctx} + {"value", 1. * res_metrics->kv_cache_used_cells / params.n_ctx} },{ {"name", "kv_cache_tokens"}, {"help", "KV-cache tokens."}, - {"value", (uint64_t) data.at("kv_cache_tokens_count")} + {"value", (uint64_t) res_metrics->kv_cache_tokens_count} },{ {"name", "requests_processing"}, {"help", "Number of request processing."}, - {"value", (uint64_t) data.at("processing")} + {"value", (uint64_t) res_metrics->n_processing_slots} },{ {"name", "requests_deferred"}, {"help", "Number of request deferred."}, - {"value", (uint64_t) data.at("deferred")} + {"value", (uint64_t) res_metrics->n_tasks_deferred} }}} }; @@ -2549,8 +3488,7 @@ int main(int argc, char ** argv) { } } - const int64_t t_start = data.at("t_start"); - res.set_header("Process-Start-Time-Unix", std::to_string(t_start)); + res.set_header("Process-Start-Time-Unix", std::to_string(res_metrics->t_start)); res.set_content(prometheus.str(), "text/plain; version=0.0.4"); res.status = 200; // HTTP OK @@ -2565,25 +3503,24 @@ int main(int argc, char ** argv) { } std::string filepath = params.slot_save_path + filename; - server_task task; - task.type = SERVER_TASK_TYPE_SLOT_SAVE; - task.data = { - { "id_slot", id_slot }, - { "filename", filename }, - { "filepath", filepath }, - }; + server_task task(SERVER_TASK_TYPE_SLOT_SAVE); + task.id = ctx_server.queue_tasks.get_new_id(); + task.slot_action.slot_id = id_slot; + task.slot_action.filename = filename; + task.slot_action.filepath = filepath; - const int id_task = ctx_server.queue_tasks.post(task); - ctx_server.queue_results.add_waiting_task_id(id_task); + ctx_server.queue_results.add_waiting_task_id(task.id); + ctx_server.queue_tasks.post(task); - server_task_result result = ctx_server.queue_results.recv(id_task); - ctx_server.queue_results.remove_waiting_task_id(id_task); + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); + ctx_server.queue_results.remove_waiting_task_id(task.id); - if (result.error) { - res_error(res, result.data); - } else { - res_ok(res, result.data); + if (result->is_error()) { + res_error(res, result->to_json()); + return; } + + res_ok(res, result->to_json()); }; const auto handle_slots_restore = [&ctx_server, &res_error, &res_ok, ¶ms](const httplib::Request & req, httplib::Response & res, int id_slot) { @@ -2595,45 +3532,45 @@ int main(int argc, char ** argv) { } std::string filepath = params.slot_save_path + filename; - server_task task; - task.type = SERVER_TASK_TYPE_SLOT_RESTORE; - task.data = { - { "id_slot", id_slot }, - { "filename", filename }, - { "filepath", filepath }, - }; + server_task task(SERVER_TASK_TYPE_SLOT_RESTORE); + task.id = ctx_server.queue_tasks.get_new_id(); + task.slot_action.slot_id = id_slot; + task.slot_action.filename = filename; + task.slot_action.filepath = filepath; - const int id_task = ctx_server.queue_tasks.post(task); - ctx_server.queue_results.add_waiting_task_id(id_task); + ctx_server.queue_results.add_waiting_task_id(task.id); + ctx_server.queue_tasks.post(task); - server_task_result result = ctx_server.queue_results.recv(id_task); - ctx_server.queue_results.remove_waiting_task_id(id_task); + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); + ctx_server.queue_results.remove_waiting_task_id(task.id); - if (result.error) { - res_error(res, result.data); - } else { - res_ok(res, result.data); + if (result->is_error()) { + res_error(res, result->to_json()); + return; } + + GGML_ASSERT(dynamic_cast(result.get()) != nullptr); + res_ok(res, result->to_json()); }; const auto handle_slots_erase = [&ctx_server, &res_error, &res_ok](const httplib::Request & /* req */, httplib::Response & res, int id_slot) { - server_task task; - task.type = SERVER_TASK_TYPE_SLOT_ERASE; - task.data = { - { "id_slot", id_slot }, - }; + server_task task(SERVER_TASK_TYPE_SLOT_ERASE); + task.id = ctx_server.queue_tasks.get_new_id(); + task.slot_action.slot_id = id_slot; - const int id_task = ctx_server.queue_tasks.post(task); - ctx_server.queue_results.add_waiting_task_id(id_task); + ctx_server.queue_results.add_waiting_task_id(task.id); + ctx_server.queue_tasks.post(task); - server_task_result result = ctx_server.queue_results.recv(id_task); - ctx_server.queue_results.remove_waiting_task_id(id_task); + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); + ctx_server.queue_results.remove_waiting_task_id(task.id); - if (result.error) { - res_error(res, result.data); - } else { - res_ok(res, result.data); + if (result->is_error()) { + res_error(res, result->to_json()); + return; } + + GGML_ASSERT(dynamic_cast(result.get()) != nullptr); + res_ok(res, result->to_json()); }; const auto handle_slots_action = [¶ms, &res_error, &handle_slots_save, &handle_slots_restore, &handle_slots_erase](const httplib::Request & req, httplib::Response & res) { @@ -2666,17 +3603,20 @@ int main(int argc, char ** argv) { }; const auto handle_props = [&ctx_server, &res_ok](const httplib::Request &, httplib::Response & res) { + // this endpoint is publicly available, please only return what is safe to be exposed json data = { { "default_generation_settings", ctx_server.default_generation_settings_for_props }, - { "total_slots", ctx_server.params.n_parallel }, - { "chat_template", llama_get_chat_template(ctx_server.model) }, + { "total_slots", ctx_server.params_base.n_parallel }, + { "model_path", ctx_server.params_base.model }, + { "chat_template", common_get_builtin_chat_template(ctx_server.model) }, + { "build_info", build_info }, }; res_ok(res, data); }; const auto handle_props_change = [&ctx_server, &res_error, &res_ok](const httplib::Request & req, httplib::Response & res) { - if (!ctx_server.params.endpoint_props) { + if (!ctx_server.params_base.endpoint_props) { res_error(res, format_error_response("This server does not support changing global properties. Start it with `--props`", ERROR_TYPE_NOT_SUPPORTED)); return; } @@ -2688,13 +3628,51 @@ int main(int argc, char ** argv) { res_ok(res, {{ "success", true }}); }; - const auto handle_completions_generic = [&ctx_server, &res_error, &res_ok](server_task_inf_type inf_type, json & data, httplib::Response & res) { - if (ctx_server.params.embedding) { + // handle completion-like requests (completion, chat, infill) + // we can optionally provide a custom format for partial results and final results + const auto handle_completions_impl = [&ctx_server, &res_error, &res_ok]( + server_task_type type, + json & data, + httplib::Response & res, + oaicompat_type oaicompat) { + GGML_ASSERT(type == SERVER_TASK_TYPE_COMPLETION || type == SERVER_TASK_TYPE_INFILL); + + if (ctx_server.params_base.embedding) { res_error(res, format_error_response("This server does not support completions. Start it without `--embeddings`", ERROR_TYPE_NOT_SUPPORTED)); return; } - std::vector tasks = ctx_server.create_tasks_inference(data, inf_type); + auto completion_id = gen_chatcmplid(); + std::vector tasks; + + try { + std::vector tokenized_prompts = tokenize_input_prompts(ctx_server.vocab, data.at("prompt"), true, true); + tasks.reserve(tokenized_prompts.size()); + for (size_t i = 0; i < tokenized_prompts.size(); i++) { + server_task task = server_task(type); + + task.id = ctx_server.queue_tasks.get_new_id(); + task.index = i; + + task.prompt_tokens = std::move(tokenized_prompts[i]); + task.params = server_task::params_from_json_cmpl( + ctx_server.ctx, + ctx_server.params_base, + data); + task.id_selected_slot = json_value(data, "id_slot", -1); + + // OAI-compat + task.params.oaicompat = oaicompat; + task.params.oaicompat_cmpl_id = completion_id; + // oaicompat_model is already populated by params_from_json_cmpl + + tasks.push_back(task); + } + } catch (const std::exception & e) { + res_error(res, format_error_response(e.what(), ERROR_TYPE_INVALID_REQUEST)); + return; + } + ctx_server.queue_results.add_waiting_tasks(tasks); ctx_server.queue_tasks.post(tasks); @@ -2702,15 +3680,15 @@ int main(int argc, char ** argv) { const auto task_ids = server_task::get_list_id(tasks); if (!stream) { - ctx_server.receive_cmpl_results(task_ids, [&](std::vector & results) { + ctx_server.receive_multi_results(task_ids, [&](std::vector & results) { if (results.size() == 1) { // single result - res_ok(res, results[0].data); + res_ok(res, results[0]->to_json()); } else { // multiple results (multitask) json arr = json::array(); - for (const auto & res : results) { - arr.push_back(res.data); + for (auto & res : results) { + arr.push_back(res->to_json()); } res_ok(res, arr); } @@ -2720,12 +3698,26 @@ int main(int argc, char ** argv) { ctx_server.queue_results.remove_waiting_task_ids(task_ids); } else { - const auto chunked_content_provider = [task_ids, &ctx_server](size_t, httplib::DataSink & sink) { - ctx_server.receive_cmpl_results_stream(task_ids, [&](const server_task_result & result) -> bool { - return server_sent_event(sink, "data", result.data); + const auto chunked_content_provider = [task_ids, &ctx_server, oaicompat](size_t, httplib::DataSink & sink) { + ctx_server.receive_cmpl_results_stream(task_ids, [&](server_task_result_ptr & result) -> bool { + json res_json = result->to_json(); + if (res_json.is_array()) { + for (const auto & res : res_json) { + if (!server_sent_event(sink, "data", res)) { + return false; + } + } + return true; + } else { + return server_sent_event(sink, "data", res_json); + } }, [&](const json & error_data) { server_sent_event(sink, "error", error_data); }); + if (oaicompat != OAICOMPAT_TYPE_NONE) { + static const std::string ev_done = "data: [DONE]\n\n"; + sink.write(ev_done.data(), ev_done.size()); + } sink.done(); return false; }; @@ -2738,21 +3730,34 @@ int main(int argc, char ** argv) { } }; - const auto handle_completions = [&handle_completions_generic](const httplib::Request & req, httplib::Response & res) { + const auto handle_completions = [&handle_completions_impl](const httplib::Request & req, httplib::Response & res) { json data = json::parse(req.body); - return handle_completions_generic(SERVER_TASK_INF_TYPE_COMPLETION, data, res); + return handle_completions_impl( + SERVER_TASK_TYPE_COMPLETION, + data, + res, + OAICOMPAT_TYPE_NONE); }; - const auto handle_infill = [&ctx_server, &res_error, &handle_completions_generic](const httplib::Request & req, httplib::Response & res) { + const auto handle_completions_oai = [&handle_completions_impl](const httplib::Request & req, httplib::Response & res) { + json data = oaicompat_completion_params_parse(json::parse(req.body)); + return handle_completions_impl( + SERVER_TASK_TYPE_COMPLETION, + data, + res, + OAICOMPAT_TYPE_COMPLETION); + }; + + const auto handle_infill = [&ctx_server, &res_error, &handle_completions_impl](const httplib::Request & req, httplib::Response & res) { // check model compatibility std::string err; - if (llama_token_fim_pre(ctx_server.model) == LLAMA_TOKEN_NULL) { + if (llama_vocab_fim_pre(ctx_server.vocab) == LLAMA_TOKEN_NULL) { err += "prefix token is missing. "; } - if (llama_token_fim_suf(ctx_server.model) == LLAMA_TOKEN_NULL) { + if (llama_vocab_fim_suf(ctx_server.vocab) == LLAMA_TOKEN_NULL) { err += "suffix token is missing. "; } - if (llama_token_fim_mid(ctx_server.model) == LLAMA_TOKEN_NULL) { + if (llama_vocab_fim_mid(ctx_server.vocab) == LLAMA_TOKEN_NULL) { err += "middle token is missing. "; } if (!err.empty()) { @@ -2763,6 +3768,11 @@ int main(int argc, char ** argv) { json data = json::parse(req.body); // validate input + if (data.contains("prompt") && !data.at("prompt").is_string()) { + // prompt is optional + res_error(res, format_error_response("\"prompt\" must be a string", ERROR_TYPE_INVALID_REQUEST)); + } + if (!data.contains("input_prefix")) { res_error(res, format_error_response("\"input_prefix\" is required", ERROR_TYPE_INVALID_REQUEST)); } @@ -2772,9 +3782,11 @@ int main(int argc, char ** argv) { } if (data.contains("input_extra") && !data.at("input_extra").is_array()) { + // input_extra is optional res_error(res, format_error_response("\"input_extra\" must be an array of {\"filename\": string, \"text\": string}", ERROR_TYPE_INVALID_REQUEST)); return; } + json input_extra = json_value(data, "input_extra", json::array()); for (const auto & chunk : input_extra) { // { "text": string, "filename": string } @@ -2790,72 +3802,48 @@ int main(int argc, char ** argv) { } data["input_extra"] = input_extra; // default to empty array if it's not exist - return handle_completions_generic(SERVER_TASK_INF_TYPE_INFILL, data, res); + std::string prompt = json_value(data, "prompt", std::string()); + std::vector tokenized_prompts = tokenize_input_prompts(ctx_server.vocab, prompt, false, true); + SRV_DBG("creating infill tasks, n_prompts = %d\n", (int) tokenized_prompts.size()); + data["prompt"] = format_infill( + ctx_server.vocab, + data.at("input_prefix"), + data.at("input_suffix"), + data.at("input_extra"), + ctx_server.params_base.n_batch, + ctx_server.params_base.n_predict, + ctx_server.slots[0].n_ctx, // TODO: there should be a better way + ctx_server.params_base.spm_infill, + tokenized_prompts[0] + ); + + return handle_completions_impl( + SERVER_TASK_TYPE_INFILL, + data, + res, + OAICOMPAT_TYPE_NONE); // infill is not OAI compatible }; - // TODO: maybe merge this function with "handle_completions_generic" - const auto handle_chat_completions = [&ctx_server, ¶ms, &res_error, &res_ok, verbose](const httplib::Request & req, httplib::Response & res) { - if (ctx_server.params.embedding) { + const auto handle_chat_completions = [&ctx_server, ¶ms, &res_error, &handle_completions_impl](const httplib::Request & req, httplib::Response & res) { + if (ctx_server.params_base.embedding) { res_error(res, format_error_response("This server does not support completions. Start it without `--embeddings`", ERROR_TYPE_NOT_SUPPORTED)); return; } - json data = oaicompat_completion_params_parse(ctx_server.model, json::parse(req.body), params.chat_template); - - std::vector tasks = ctx_server.create_tasks_inference(data, SERVER_TASK_INF_TYPE_COMPLETION); - ctx_server.queue_results.add_waiting_tasks(tasks); - ctx_server.queue_tasks.post(tasks); - - bool stream = json_value(data, "stream", false); - const auto task_ids = server_task::get_list_id(tasks); - const auto completion_id = gen_chatcmplid(); - - if (!stream) { - ctx_server.receive_cmpl_results(task_ids, [&](const std::vector & results) { - // multitask is never support in chat completion, there is only one result - json result_oai = format_final_response_oaicompat(data, results[0].data, completion_id, /*.streaming =*/ false, verbose); - res_ok(res, result_oai); - }, [&](const json & error_data) { - res_error(res, error_data); - }); - - ctx_server.queue_results.remove_waiting_task_ids(task_ids); - } else { - const auto chunked_content_provider = [task_ids, &ctx_server, completion_id](size_t, httplib::DataSink & sink) { - ctx_server.receive_cmpl_results_stream(task_ids, [&](const server_task_result & result) -> bool { - std::vector result_array = format_partial_response_oaicompat(result.data, completion_id); - for (auto & event_data : result_array) { - if (event_data.empty()) { - continue; // skip the stop token - } - if (!server_sent_event(sink, "data", event_data)) { - return false; // connection is closed - } - } - return true; // ok - }, [&](const json & error_data) { - server_sent_event(sink, "error", error_data); - }); - static const std::string ev_done = "data: [DONE]\n\n"; - sink.write(ev_done.data(), ev_done.size()); - sink.done(); - return true; - }; - - auto on_complete = [task_ids, &ctx_server] (bool) { - ctx_server.queue_results.remove_waiting_task_ids(task_ids); - }; - - res.set_chunked_content_provider("text/event-stream", chunked_content_provider, on_complete); - } + json data = oaicompat_chat_completion_params_parse(ctx_server.model, json::parse(req.body), params.chat_template); + return handle_completions_impl( + SERVER_TASK_TYPE_COMPLETION, + data, + res, + OAICOMPAT_TYPE_CHAT); }; - const auto handle_models = [¶ms, &ctx_server](const httplib::Request &, httplib::Response & res) { + const auto handle_models = [¶ms, &ctx_server, &res_ok](const httplib::Request &, httplib::Response & res) { json models = { {"object", "list"}, {"data", { { - {"id", params.model_alias}, + {"id", params.model_alias.empty() ? params.model : params.model_alias}, {"object", "model"}, {"created", std::time(0)}, {"owned_by", "llamacpp"}, @@ -2864,7 +3852,7 @@ int main(int argc, char ** argv) { }} }; - res.set_content(models.dump(), MIMETYPE_JSON); + res_ok(res, models); }; const auto handle_tokenize = [&ctx_server, &res_ok](const httplib::Request & req, httplib::Response & res) { @@ -2875,7 +3863,7 @@ int main(int argc, char ** argv) { const bool add_special = json_value(body, "add_special", false); const bool with_pieces = json_value(body, "with_pieces", false); - llama_tokens tokens = tokenize_mixed(ctx_server.ctx, body.at("content"), add_special, true); + llama_tokens tokens = tokenize_mixed(ctx_server.vocab, body.at("content"), add_special, true); if (with_pieces) { for (const auto& token : tokens) { @@ -2920,37 +3908,74 @@ int main(int argc, char ** argv) { res_ok(res, data); }; - const auto handle_embeddings = [&ctx_server, &res_error, &res_ok](const httplib::Request & req, httplib::Response & res) { + const auto handle_embeddings_impl = [&ctx_server, &res_error, &res_ok](const httplib::Request & req, httplib::Response & res, oaicompat_type oaicompat) { const json body = json::parse(req.body); - bool is_openai = false; - // an input prompt can be a string or a list of tokens (integer) + if (oaicompat != OAICOMPAT_TYPE_NONE && llama_pooling_type(ctx_server.ctx) == LLAMA_POOLING_TYPE_NONE) { + res_error(res, format_error_response("Pooling type 'none' is not OAI compatible. Please use a different pooling type", ERROR_TYPE_INVALID_REQUEST)); + return; + } + + // for the shape of input/content, see tokenize_input_prompts() json prompt; if (body.count("input") != 0) { - is_openai = true; prompt = body.at("input"); - } else if (body.count("content") != 0) { - // with "content", we only support single prompt - prompt = std::vector{body.at("content")}; + } else if (body.contains("content")) { + oaicompat = OAICOMPAT_TYPE_NONE; // "content" field is not OAI compatible + prompt = body.at("content"); } else { res_error(res, format_error_response("\"input\" or \"content\" must be provided", ERROR_TYPE_INVALID_REQUEST)); return; } + bool use_base64 = false; + if (body.count("encoding_format") != 0) { + const std::string& format = body.at("encoding_format"); + if (format == "base64") { + use_base64 = true; + } else if (format != "float") { + res_error(res, format_error_response("The format to return the embeddings in. Can be either float or base64", ERROR_TYPE_INVALID_REQUEST)); + return; + } + } + + std::vector tokenized_prompts = tokenize_input_prompts(ctx_server.vocab, prompt, true, true); + for (const auto & tokens : tokenized_prompts) { + // this check is necessary for models that do not add BOS token to the input + if (tokens.empty()) { + res_error(res, format_error_response("Input content cannot be empty", ERROR_TYPE_INVALID_REQUEST)); + return; + } + } + // create and queue the task json responses = json::array(); bool error = false; { - std::vector tasks = ctx_server.create_tasks_inference({{"prompt", prompt}}, SERVER_TASK_INF_TYPE_EMBEDDING); + std::vector tasks; + for (size_t i = 0; i < tokenized_prompts.size(); i++) { + server_task task = server_task(SERVER_TASK_TYPE_EMBEDDING); + + task.id = ctx_server.queue_tasks.get_new_id(); + task.index = i; + task.prompt_tokens = std::move(tokenized_prompts[i]); + + // OAI-compat + task.params.oaicompat = oaicompat; + + tasks.push_back(task); + } + ctx_server.queue_results.add_waiting_tasks(tasks); ctx_server.queue_tasks.post(tasks); // get the result std::unordered_set task_ids = server_task::get_list_id(tasks); - ctx_server.receive_cmpl_results(task_ids, [&](std::vector & results) { - for (const auto & res : results) { - responses.push_back(res.data); + ctx_server.receive_multi_results(task_ids, [&](std::vector & results) { + for (auto & res : results) { + GGML_ASSERT(dynamic_cast(res.get()) != nullptr); + responses.push_back(res->to_json()); } }, [&](const json & error_data) { res_error(res, error_data); @@ -2965,14 +3990,22 @@ int main(int argc, char ** argv) { } // write JSON response - json root = is_openai - ? format_embeddings_response_oaicompat(body, responses) - : responses[0]; + json root = oaicompat == OAICOMPAT_TYPE_EMBEDDING + ? format_embeddings_response_oaicompat(body, responses, use_base64) + : json(responses); res_ok(res, root); }; + const auto handle_embeddings = [&handle_embeddings_impl](const httplib::Request & req, httplib::Response & res) { + handle_embeddings_impl(req, res, OAICOMPAT_TYPE_NONE); + }; + + const auto handle_embeddings_oai = [&handle_embeddings_impl](const httplib::Request & req, httplib::Response & res) { + handle_embeddings_impl(req, res, OAICOMPAT_TYPE_EMBEDDING); + }; + const auto handle_rerank = [&ctx_server, &res_error, &res_ok](const httplib::Request & req, httplib::Response & res) { - if (!ctx_server.params.reranking || ctx_server.params.embedding) { + if (!ctx_server.params_base.reranking || ctx_server.params_base.embedding) { res_error(res, format_error_response("This server does not support reranking. Start it with `--reranking` and without `--embedding`", ERROR_TYPE_NOT_SUPPORTED)); return; } @@ -3006,29 +4039,33 @@ int main(int argc, char ** argv) { return; } - // construct prompt object: array of ["query", "doc0", "doc1", ...] - json prompt; - prompt.push_back(query); - for (const auto & doc : documents) { - prompt.push_back(doc); - } - - LOG_DBG("rerank prompt: %s\n", prompt.dump().c_str()); + llama_tokens tokenized_query = tokenize_input_prompts(ctx_server.vocab, query, /* add_special */ false, true)[0]; // create and queue the task json responses = json::array(); bool error = false; { - std::vector tasks = ctx_server.create_tasks_inference({{"prompt", prompt}}, SERVER_TASK_INF_TYPE_RERANK); + std::vector tasks; + std::vector tokenized_docs = tokenize_input_prompts(ctx_server.vocab, documents, /* add_special */ false, true); + tasks.reserve(tokenized_docs.size()); + for (size_t i = 0; i < tokenized_docs.size(); i++) { + server_task task = server_task(SERVER_TASK_TYPE_RERANK); + task.id = ctx_server.queue_tasks.get_new_id(); + task.index = i; + task.prompt_tokens = format_rerank(ctx_server.vocab, tokenized_query, tokenized_docs[i]); + tasks.push_back(task); + } + ctx_server.queue_results.add_waiting_tasks(tasks); ctx_server.queue_tasks.post(tasks); // get the result std::unordered_set task_ids = server_task::get_list_id(tasks); - ctx_server.receive_cmpl_results(task_ids, [&](std::vector & results) { - for (const auto & res : results) { - responses.push_back(res.data); + ctx_server.receive_multi_results(task_ids, [&](std::vector & results) { + for (auto & res : results) { + GGML_ASSERT(dynamic_cast(res.get()) != nullptr); + responses.push_back(res->to_json()); } }, [&](const json & error_data) { res_error(res, error_data); @@ -3047,8 +4084,9 @@ int main(int argc, char ** argv) { const auto handle_lora_adapters_list = [&](const httplib::Request &, httplib::Response & res) { json result = json::array(); - for (size_t i = 0; i < ctx_server.loras.size(); ++i) { - auto & lora = ctx_server.loras[i]; + const auto & loras = ctx_server.params_base.lora_adapters; + for (size_t i = 0; i < loras.size(); ++i) { + auto & lora = loras[i]; result.push_back({ {"id", i}, {"path", lora.path}, @@ -3060,64 +4098,56 @@ int main(int argc, char ** argv) { }; const auto handle_lora_adapters_apply = [&](const httplib::Request & req, httplib::Response & res) { - const std::vector body = json::parse(req.body); - int max_idx = ctx_server.loras.size(); + const json body = json::parse(req.body); + if (!body.is_array()) { + res_error(res, format_error_response("Request body must be an array", ERROR_TYPE_INVALID_REQUEST)); + return; + } + server_task task(SERVER_TASK_TYPE_SET_LORA); + task.id = ctx_server.queue_tasks.get_new_id(); + task.set_lora = parse_lora_request(ctx_server.params_base.lora_adapters, body); + ctx_server.queue_results.add_waiting_task_id(task.id); + ctx_server.queue_tasks.post(task); - // clear existing value - for (auto & lora : ctx_server.loras) { - lora.scale = 0.0f; + server_task_result_ptr result = ctx_server.queue_results.recv(task.id); + ctx_server.queue_results.remove_waiting_task_id(task.id); + + if (result->is_error()) { + res_error(res, result->to_json()); + return; } - // set value - for (auto entry : body) { - int id = entry.at("id"); - float scale = entry.at("scale"); - if (0 <= id && id < max_idx) { - ctx_server.loras[id].scale = scale; - } else { - throw std::runtime_error("invalid adapter id"); - } - } - - server_task task; - task.type = SERVER_TASK_TYPE_SET_LORA; - const int id_task = ctx_server.queue_tasks.post(task); - ctx_server.queue_results.add_waiting_task_id(id_task); - - server_task_result result = ctx_server.queue_results.recv(id_task); - ctx_server.queue_results.remove_waiting_task_id(id_task); - - res_ok(res, result.data); - res.status = 200; // HTTP OK - }; - - auto handle_static_file = [](unsigned char * content, size_t len, const char * mime_type) { - return [content, len, mime_type](const httplib::Request &, httplib::Response & res) { - res.set_content(reinterpret_cast(content), len, mime_type); - return false; - }; + GGML_ASSERT(dynamic_cast(result.get()) != nullptr); + res_ok(res, result->to_json()); }; // // Router // - // register static assets routes - if (!params.public_path.empty()) { - // Set the base directory for serving static files - bool is_found = svr->set_mount_point("/", params.public_path); - if (!is_found) { - LOG_ERR("%s: static assets path not found: %s\n", __func__, params.public_path.c_str()); - return 1; - } + if (!params.webui) { + LOG_INF("Web UI is disabled\n"); } else { - // using embedded static files - svr->Get("/", handle_static_file(index_html, index_html_len, "text/html; charset=utf-8")); - svr->Get("/completion.js", handle_static_file(completion_js, completion_js_len, "text/javascript; charset=utf-8")); - svr->Get("/deps_daisyui.min.css", handle_static_file(deps_daisyui_min_css, deps_daisyui_min_css_len, "text/css; charset=utf-8")); - svr->Get("/deps_markdown-it.js", handle_static_file(deps_markdown_it_js, deps_markdown_it_js_len, "text/javascript; charset=utf-8")); - svr->Get("/deps_tailwindcss.js", handle_static_file(deps_tailwindcss_js, deps_tailwindcss_js_len, "text/javascript; charset=utf-8")); - svr->Get("/deps_vue.esm-browser.js", handle_static_file(deps_vue_esm_browser_js, deps_vue_esm_browser_js_len, "text/javascript; charset=utf-8")); + // register static assets routes + if (!params.public_path.empty()) { + // Set the base directory for serving static files + bool is_found = svr->set_mount_point("/", params.public_path); + if (!is_found) { + LOG_ERR("%s: static assets path not found: %s\n", __func__, params.public_path.c_str()); + return 1; + } + } else { + // using embedded static index.html + svr->Get("/", [](const httplib::Request & req, httplib::Response & res) { + if (req.get_header_value("Accept-Encoding").find("gzip") == std::string::npos) { + res.set_content("Error: gzip is not supported by this browser", "text/plain"); + } else { + res.set_header("Content-Encoding", "gzip"); + res.set_content(reinterpret_cast(index_html_gz), index_html_gz_len, "text/html; charset=utf-8"); + } + return false; + }); + } } // register API routes @@ -3129,13 +4159,13 @@ int main(int argc, char ** argv) { svr->Get ("/v1/models", handle_models); // public endpoint (no API key check) svr->Post("/completion", handle_completions); // legacy svr->Post("/completions", handle_completions); - svr->Post("/v1/completions", handle_completions); + svr->Post("/v1/completions", handle_completions_oai); svr->Post("/chat/completions", handle_chat_completions); svr->Post("/v1/chat/completions", handle_chat_completions); svr->Post("/infill", handle_infill); svr->Post("/embedding", handle_embeddings); // legacy svr->Post("/embeddings", handle_embeddings); - svr->Post("/v1/embeddings", handle_embeddings); + svr->Post("/v1/embeddings", handle_embeddings_oai); svr->Post("/rerank", handle_rerank); svr->Post("/reranking", handle_rerank); svr->Post("/v1/rerank", handle_rerank); @@ -3165,8 +4195,18 @@ int main(int argc, char ** argv) { llama_backend_free(); }; - // bind HTTP listen port, run the HTTP server in a thread - if (!svr->bind_to_port(params.hostname, params.port)) { + // bind HTTP listen port + bool was_bound = false; + if (params.port == 0) { + int bound_port = svr->bind_to_any_port(params.hostname); + if ((was_bound = (bound_port >= 0))) { + params.port = bound_port; + } + } else { + was_bound = svr->bind_to_port(params.hostname, params.port); + } + + if (!was_bound) { //LOG_ERROR("couldn't bind HTTP server socket", { // {"hostname", params.hostname}, // {"port", params.port}, @@ -3175,6 +4215,8 @@ int main(int argc, char ** argv) { clean_up(); return 1; } + + // run the HTTP server in a thread std::thread t([&]() { svr->listen_after_bind(); }); svr->wait_until_ready(); @@ -3197,14 +4239,16 @@ int main(int argc, char ** argv) { // if a custom chat template is not supplied, we will use the one that comes with the model (if any) if (params.chat_template.empty()) { - if (!ctx_server.validate_model_chat_template()) { + if (!ctx_server.validate_builtin_chat_template()) { LOG_WRN("%s: The chat template that comes with this model is not yet supported, falling back to chatml. This may cause the model to output suboptimal responses\n", __func__); params.chat_template = "chatml"; } } // print sample chat example to make it clear which template is used - LOG_INF("%s: chat template, built_in: %d, chat_example: '%s'\n", __func__, params.chat_template.empty(), common_chat_format_example(ctx_server.model, params.chat_template).c_str()); + LOG_INF("%s: chat template, chat_template: %s, example_format: '%s'\n", __func__, + params.chat_template.empty() ? "(built-in)" : params.chat_template.c_str(), + common_chat_format_example(ctx_server.model, params.chat_template).c_str()); ctx_server.queue_tasks.on_new_task(std::bind( &server_context::process_single_task, &ctx_server, std::placeholders::_1)); diff --git a/examples/server/tests/.gitignore b/examples/server/tests/.gitignore index 1d17dae13..90ee7fe6d 100644 --- a/examples/server/tests/.gitignore +++ b/examples/server/tests/.gitignore @@ -1 +1,2 @@ .venv +tmp diff --git a/examples/server/tests/README.md b/examples/server/tests/README.md index 10f22c447..5787276ab 100644 --- a/examples/server/tests/README.md +++ b/examples/server/tests/README.md @@ -1,19 +1,9 @@ # Server tests -Python based server tests scenario using [BDD](https://en.wikipedia.org/wiki/Behavior-driven_development) -and [behave](https://behave.readthedocs.io/en/latest/): - -* [issues.feature](./features/issues.feature) Pending issues scenario -* [parallel.feature](./features/parallel.feature) Scenario involving multi slots and concurrent requests -* [security.feature](./features/security.feature) Security, CORS and API Key -* [server.feature](./features/server.feature) Server base scenario: completion, embedding, tokenization, etc... +Python based server tests scenario using [pytest](https://docs.pytest.org/en/stable/). Tests target GitHub workflows job runners with 4 vCPU. -Requests are -using [aiohttp](https://docs.aiohttp.org/en/stable/client_reference.html), [asyncio](https://docs.python.org/fr/3/library/asyncio.html) -based http client. - Note: If the host architecture inference speed is faster than GitHub runners one, parallel scenario may randomly fail. To mitigate it, you can increase values in `n_predict`, `kv_size`. @@ -39,26 +29,31 @@ It's possible to override some scenario steps values with environment variables: |--------------------------|------------------------------------------------------------------------------------------------| | `PORT` | `context.server_port` to set the listening port of the server during scenario, default: `8080` | | `LLAMA_SERVER_BIN_PATH` | to change the server binary path, default: `../../../build/bin/llama-server` | -| `DEBUG` | "ON" to enable steps and server verbose mode `--verbose` | +| `DEBUG` | to enable steps and server verbose mode `--verbose` | | `N_GPU_LAYERS` | number of model layers to offload to VRAM `-ngl --n-gpu-layers` | -### Run @bug, @wip or @wrong_usage annotated scenario - -Feature or Scenario must be annotated with `@llama.cpp` to be included in the default scope. - -- `@bug` annotation aims to link a scenario with a GitHub issue. -- `@wrong_usage` are meant to show user issue that are actually an expected behavior -- `@wip` to focus on a scenario working in progress -- `@slow` heavy test, disabled by default - -To run a scenario annotated with `@bug`, start: +To run slow tests: ```shell -DEBUG=ON ./tests.sh --no-skipped --tags bug --stop +SLOW_TESTS=1 ./tests.sh ``` -After changing logic in `steps.py`, ensure that `@bug` and `@wrong_usage` scenario are updated. +To run with stdout/stderr display in real time (verbose output, but useful for debugging): ```shell -./tests.sh --no-skipped --tags bug,wrong_usage || echo "should failed but compile" +DEBUG=1 ./tests.sh -s -v -x ``` + +To run single test unit: + +```shell +./tests.sh unit/test_{name of test case here}.py -v -x +``` + +Hint: You can compile and run test in single command, useful for local developement: + +```shell +cmake --build build -j --target llama-server && ./examples/server/tests/tests.sh +``` + +To see all available arguments, please refer to [pytest documentation](https://docs.pytest.org/en/stable/how-to/usage.html) diff --git a/examples/server/tests/conftest.py b/examples/server/tests/conftest.py new file mode 100644 index 000000000..017d1bb84 --- /dev/null +++ b/examples/server/tests/conftest.py @@ -0,0 +1,15 @@ +import pytest +from utils import * + + +# ref: https://stackoverflow.com/questions/22627659/run-code-before-and-after-each-test-in-py-test +@pytest.fixture(autouse=True) +def stop_server_after_each_test(): + # do nothing before each test + yield + # stop all servers after each test + instances = set( + server_instances + ) # copy the set to prevent 'Set changed size during iteration' + for server in instances: + server.stop() diff --git a/examples/server/tests/features/ctx_shift.feature b/examples/server/tests/features/ctx_shift.feature deleted file mode 100644 index ae6c6b01b..000000000 --- a/examples/server/tests/features/ctx_shift.feature +++ /dev/null @@ -1,66 +0,0 @@ -@llama.cpp -@ctx_shift -Feature: llama.cpp server - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K.gguf from HF repo ggml-org/models - And a model file test-model.gguf - And a model alias tinyllama-2 - And BOS token is 1 - And 42 as server seed - And 256 KV cache size - And 32 as batch size - And 2 slots - - # the prompt is 301 tokens - # the slot context is 256/2 = 128 tokens - # the prompt is truncated to keep the last 109 tokens - # 64 tokens are generated thanks to shifting the context when it gets full - Scenario: Inference with context shift - And 64 server max tokens to predict - Then the server is starting - Then the server is healthy - Given a prompt: - """ - Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. - Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. - Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. - Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. - """ - And a completion request with no api error - Then 64 tokens are predicted matching fun|Annaks|popcorns|pictry|bowl - And the completion is truncated - And 109 prompt tokens are processed - - Scenario Outline: Inference without context shift - And server max tokens to predict - And disable context shifting - Then the server is starting - Then the server is healthy - Given a prompt: - """ - Hi how are you - """ - And a completion request with no api error - Then tokens are predicted matching twind|Anna - And the completion is truncated - And 8 prompt tokens are processed - Examples: - | n_predict | n_token_output | truncated | - | 64 | 64 | not | - | -1 | 120 | | - - Scenario: Inference without context shift (expected error: prompt too long) - And disable context shifting - Then the server is starting - Then the server is healthy - Given a prompt: - """ - Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. - Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. - Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. - Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. - """ - And a completion request with 400 api error - diff --git a/examples/server/tests/features/embeddings.feature b/examples/server/tests/features/embeddings.feature deleted file mode 100644 index f4fe2ee43..000000000 --- a/examples/server/tests/features/embeddings.feature +++ /dev/null @@ -1,113 +0,0 @@ -@llama.cpp -@embeddings -Feature: llama.cpp server - - Background: Server startup - Given a server listening on localhost:8080 - And a model url https://huggingface.co/ggml-org/models/resolve/main/bert-bge-small/ggml-model-f16.gguf - And a model file bert-bge-small.gguf - And a model alias bert-bge-small - And 42 as server seed - And 2 slots - # the bert-bge-small model has context size of 512 - # since the generated prompts are as big as the batch size, we need to set the batch size to <= 512 - # ref: https://huggingface.co/BAAI/bge-small-en-v1.5/blob/5c38ec7c405ec4b44b94cc5a9bb96e735b38267a/config.json#L20 - And 128 as batch size - And 128 as ubatch size - And 512 KV cache size - And enable embeddings endpoint - Then the server is starting - Then the server is healthy - - Scenario: Embedding - When embeddings are computed for: - """ - What is the capital of Bulgaria ? - """ - Then embeddings are generated - - Scenario: Embedding (error: prompt too long) - When embeddings are computed for: - """ - Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. - Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. - Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. - Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. - Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. - Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. - Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. - Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. - """ - And embeddings request with 500 api error - - Scenario: OAI Embeddings compatibility - Given a model bert-bge-small - When an OAI compatible embeddings computation request for: - """ - What is the capital of Spain ? - """ - Then embeddings are generated - - Scenario: OAI Embeddings compatibility with multiple inputs - Given a model bert-bge-small - Given a prompt: - """ - In which country Paris is located ? - """ - And a prompt: - """ - Is Madrid the capital of Spain ? - """ - When an OAI compatible embeddings computation request for multiple inputs - Then embeddings are generated - - Scenario: Multi users embeddings - Given a prompt: - """ - Write a very long story about AI. - """ - And a prompt: - """ - Write another very long music lyrics. - """ - And a prompt: - """ - Write a very long poem. - """ - And a prompt: - """ - Write a very long joke. - """ - Given concurrent embedding requests - Then the server is busy - Then the server is idle - Then all embeddings are generated - - Scenario: Multi users OAI compatibility embeddings - Given a prompt: - """ - In which country Paris is located ? - """ - And a prompt: - """ - Is Madrid the capital of Spain ? - """ - And a prompt: - """ - What is the biggest US city ? - """ - And a prompt: - """ - What is the capital of Bulgaria ? - """ - And a model bert-bge-small - Given concurrent OAI embedding requests - Then the server is busy - Then the server is idle - Then all embeddings are generated - - Scenario: All embeddings should be the same - Given 10 fixed prompts - And a model bert-bge-small - Given concurrent OAI embedding requests - Then all embeddings are the same diff --git a/examples/server/tests/features/environment.py b/examples/server/tests/features/environment.py deleted file mode 100644 index e7845dc2f..000000000 --- a/examples/server/tests/features/environment.py +++ /dev/null @@ -1,71 +0,0 @@ -import os -import signal -import socket -import sys -import time -import traceback -from contextlib import closing -from subprocess import TimeoutExpired - - -def before_scenario(context, scenario): - context.debug = 'DEBUG' in os.environ and os.environ['DEBUG'] == 'ON' - if context.debug: - print("DEBUG=ON") - print(f"\x1b[33;42mStarting new scenario: {scenario.name}!\x1b[0m") - port = 8080 - if 'PORT' in os.environ: - port = int(os.environ['PORT']) - if is_server_listening("localhost", port): - assert False, "Server already started" - - -def after_scenario(context, scenario): - try: - if 'server_process' not in context or context.server_process is None: - return - if scenario.status == "failed": - if 'GITHUB_ACTIONS' in os.environ: - print(f"\x1b[33;101mSCENARIO FAILED: {scenario.name} server logs:\x1b[0m\n") - if os.path.isfile('llama.log'): - with closing(open('llama.log', 'r')) as f: - for line in f: - print(line) - if not is_server_listening(context.server_fqdn, context.server_port): - print("\x1b[33;101mERROR: Server stopped listening\x1b[0m") - - if context.server_process.poll() is not None: - assert False, f"Server not running pid={context.server_process.pid} ..." - - server_graceful_shutdown(context) # SIGINT - - try: - context.server_process.wait(0.5) - except TimeoutExpired: - print(f"server still alive after 500ms, force-killing pid={context.server_process.pid} ...") - context.server_process.kill() # SIGKILL - context.server_process.wait() - - while is_server_listening(context.server_fqdn, context.server_port): - time.sleep(0.1) - except Exception: - print("ignoring error in after_scenario:") - traceback.print_exc(file=sys.stdout) - - -def server_graceful_shutdown(context): - print(f"shutting down server pid={context.server_process.pid} ...") - if os.name == 'nt': - interrupt = signal.CTRL_C_EVENT - else: - interrupt = signal.SIGINT - context.server_process.send_signal(interrupt) - - -def is_server_listening(server_fqdn, server_port): - with closing(socket.socket(socket.AF_INET, socket.SOCK_STREAM)) as sock: - result = sock.connect_ex((server_fqdn, server_port)) - _is_server_listening = result == 0 - if _is_server_listening: - print(f"server is listening on {server_fqdn}:{server_port}...") - return _is_server_listening diff --git a/examples/server/tests/features/infill.feature b/examples/server/tests/features/infill.feature deleted file mode 100644 index a0bbfef77..000000000 --- a/examples/server/tests/features/infill.feature +++ /dev/null @@ -1,36 +0,0 @@ -@llama.cpp -@infill -Feature: llama.cpp server - - # The current model is made by adding FIM tokens to the existing stories260K - # We may want to use a better model in the future, maybe something like SmolLM 360M - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K-infill.gguf from HF repo ggml-org/models - And a model file test-model-infill.gguf - And a model alias tinyllama-infill - And 42 as server seed - And 1024 as batch size - And 1024 as ubatch size - And 2048 KV cache size - And 64 max tokens to predict - And 0.0 temperature - Then the server is starting - Then the server is healthy - - Scenario: Infill without input_extra - Given a prompt "Complete this" - And an infill input extra none none - And an infill input prefix "#include \n#include \"llama.h\"\n\nint main() {\n int n_threads = llama_" - And an infill input suffix "}\n" - And an infill request with no api error - Then 64 tokens are predicted matching One|day|she|saw|big|scary|bird - - Scenario: Infill with input_extra - Given a prompt "Complete this" - And an infill input extra "llama.h" "LLAMA_API int32_t llama_n_threads();\n" - And an infill input prefix "#include \n#include \"llama.h\"\n\nint main() {\n int n_threads = llama_" - And an infill input suffix "}\n" - And an infill request with no api error - Then 64 tokens are predicted matching cuts|Jimmy|mom|came|into|the|room" diff --git a/examples/server/tests/features/issues.feature b/examples/server/tests/features/issues.feature deleted file mode 100644 index 7b13e44ca..000000000 --- a/examples/server/tests/features/issues.feature +++ /dev/null @@ -1,5 +0,0 @@ -# List of ongoing issues -# run with: DEBUG=ON ./tests.sh --no-skipped --tags bug -@bug -Feature: Issues - # No confirmed issue at the moment diff --git a/examples/server/tests/features/lora.feature b/examples/server/tests/features/lora.feature deleted file mode 100644 index 7b85988ac..000000000 --- a/examples/server/tests/features/lora.feature +++ /dev/null @@ -1,36 +0,0 @@ -@llama.cpp -@lora -Feature: llama.cpp server - - Background: Server startup - Given a server listening on localhost:8080 - And a model url https://huggingface.co/ggml-org/stories15M_MOE/resolve/main/stories15M_MOE-F16.gguf - And a model file stories15M_MOE-F16.gguf - And a model alias stories15M_MOE - And a lora adapter file from https://huggingface.co/ggml-org/stories15M_MOE/resolve/main/moe_shakespeare15M.gguf - And 42 as server seed - And 1024 as batch size - And 1024 as ubatch size - And 2048 KV cache size - And 64 max tokens to predict - And 0.0 temperature - Then the server is starting - Then the server is healthy - - Scenario: Completion LoRA disabled - Given switch off lora adapter 0 - Given a prompt: - """ - Look in thy glass - """ - And a completion request with no api error - Then 64 tokens are predicted matching little|girl|three|years|old - - Scenario: Completion LoRA enabled - Given switch on lora adapter 0 - Given a prompt: - """ - Look in thy glass - """ - And a completion request with no api error - Then 64 tokens are predicted matching eye|love|glass|sun diff --git a/examples/server/tests/features/parallel.feature b/examples/server/tests/features/parallel.feature deleted file mode 100644 index 423d0f1d4..000000000 --- a/examples/server/tests/features/parallel.feature +++ /dev/null @@ -1,131 +0,0 @@ -@llama.cpp -@parallel -Feature: Parallel - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/split/stories15M-00001-of-00003.gguf from HF repo ggml-org/models - And a model file test-model-00001-of-00003.gguf - And 42 as server seed - And 128 as batch size - And 256 KV cache size - And 2 slots - And continuous batching - Then the server is starting - Then the server is healthy - - Scenario Outline: Multi users completion - Given a prompt: - """ - Write a very long story about AI. - """ - And a prompt: - """ - Write another very long music lyrics. - """ - And max tokens to predict - Given concurrent completion requests - Then the server is busy - Then the server is idle - And all slots are idle - Then all prompts are predicted with tokens - Examples: - | n_predict | - | 128 | - - Scenario Outline: Multi users OAI completions compatibility - Given a system prompt You are a writer. - And a model tinyllama-2 - Given a prompt: - """ - Write a very long book. - """ - And a prompt: - """ - Write another a poem. - """ - And max tokens to predict - And streaming is - Given concurrent OAI completions requests - Then the server is busy - Then the server is idle - Then all prompts are predicted with tokens - Examples: - | streaming | n_predict | - | disabled | 128 | - | enabled | 64 | - - Scenario Outline: Multi users OAI completions compatibility no v1 - Given a system prompt You are a writer. - And a model tinyllama-2 - Given a prompt: - """ - Write a very long book. - """ - And a prompt: - """ - Write another a poem. - """ - And max tokens to predict - And streaming is - Given concurrent OAI completions requests no v1 - Then the server is busy - Then the server is idle - Then all prompts are predicted with tokens - Examples: - | streaming | n_predict | - | disabled | 128 | - | enabled | 64 | - - Scenario Outline: Multi users with number of prompts exceeding number of slots - Given a system prompt You are a writer. - And a model tinyllama-2 - Given a prompt: - """ - Write a very long book. - """ - And a prompt: - """ - Write another a poem. - """ - And a prompt: - """ - What is LLM? - """ - And a prompt: - """ - The sky is blue and I love it. - """ - And max tokens to predict - And streaming is - Given concurrent OAI completions requests - Then the server is busy - Then the server is idle - Then all prompts are predicted with tokens - Examples: - | streaming | n_predict | - | disabled | 128 | - | enabled | 64 | - - Scenario: Multi users with total number of tokens to predict exceeds the KV Cache size #3969 - Given a prompt: - """ - Write a very long story about AI. - """ - And a prompt: - """ - Write another very long music lyrics. - """ - And a prompt: - """ - Write a very long poem. - """ - And a prompt: - """ - Write a very long joke. - """ - And 128 max tokens to predict - Given concurrent completion requests - Then the server is busy - Then the server is idle - Then all prompts are predicted diff --git a/examples/server/tests/features/passkey.feature b/examples/server/tests/features/passkey.feature deleted file mode 100644 index ff0a82cc4..000000000 --- a/examples/server/tests/features/passkey.feature +++ /dev/null @@ -1,56 +0,0 @@ -# run with: ./tests.sh --no-skipped --tags passkey -@passkey -@slow -Feature: Passkey / Self-extend with context shift - - Background: Server startup - Given a server listening on localhost:8080 - - # Generates a long text of junk and inserts a secret passkey number inside it. - # Then we query the LLM for the secret passkey. - # see #3856 and #4810 - Scenario Outline: Passkey - Given a model file from HF repo - And as batch size - And as number of junk - And server max tokens to predict - And 42 as seed - And 0.0 temperature - And KV cache size - And 1 slots - And group attention factor to extend context size through self-extend - And group attention width to extend context size through self-extend - # Can be override with N_GPU_LAYERS - And GPU offloaded layers - Then the server is starting - # Higher timeout because the model may need to be downloaded from the internet - Then the server is healthy with timeout 120 seconds - Given available models - Then model 0 is trained on tokens context - Given a prefix prompt: - """ - here is an important info hidden inside a lot of irrelevant text. Find it and memorize them. I will quiz you about the important information there. - """ - And a passkey prompt template: - """ - The pass key is Remember it. is the pass key. - """ - And a junk suffix prompt: - """ - The grass is green. The sky is blue. The sun is yellow. Here we go. There and back again. - """ - And a suffix prompt: - """ - What is the pass key? The pass key is - """ - Given a "" passkey challenge prompt with the passkey inserted every junk - And a completion request with no api error - Then tokens are predicted matching - - Examples: - | hf_repo | hf_file | n_ctx_train | ngl | n_ctx | n_batch | n_ga | n_ga_w | n_junk | i_pos | passkey | n_predicted | re_content | - | TheBloke/phi-2-GGUF | phi-2.Q4_K_M.gguf | 2048 | 5 | 8192 | 512 | 4 | 512 | 250 | 50 | 42 | 1 | 42 | - | TheBloke/phi-2-GGUF | phi-2.Q4_K_M.gguf | 2048 | 5 | 8192 | 512 | 2 | 512 | 250 | 50 | 42 | 1 | \b((?!42)\w)+\b | - #| TheBloke/Llama-2-7B-GGUF | llama-2-7b.Q2_K.gguf | 4096 | 3 | 16384 | 512 | 4 | 512 | 500 | 300 | 1234 | 5 | 1234 | - #| TheBloke/Mixtral-8x7B-v0.1-GGUF | mixtral-8x7b-v0.1.Q2_K.gguf | 32768 | 2 | 16384 | 512 | 4 | 512 | 500 | 100 | 0987 | 5 | 0 - # 987 | diff --git a/examples/server/tests/features/rerank.feature b/examples/server/tests/features/rerank.feature deleted file mode 100644 index c36cc8e21..000000000 --- a/examples/server/tests/features/rerank.feature +++ /dev/null @@ -1,42 +0,0 @@ -@llama.cpp -@rerank -Feature: llama.cpp server - - Background: Server startup - Given a server listening on localhost:8080 - And a model url https://huggingface.co/ggml-org/models/resolve/main/jina-reranker-v1-tiny-en/ggml-model-f16.gguf - And a model file jina-reranker-v1-tiny-en.gguf - And a model alias jina-reranker-v1-tiny-en - And 42 as server seed - And 2 slots - And 512 as batch size - And 512 as ubatch size - And 512 KV cache size - And enable reranking endpoint - Then the server is starting - Then the server is healthy - - Scenario: Rerank - Given a rerank query: - """ - Machine learning is - """ - And a rerank document: - """ - A machine is a physical system that uses power to apply forces and control movement to perform an action. The term is commonly applied to artificial devices, such as those employing engines or motors, but also to natural biological macromolecules, such as molecular machines. - """ - And a rerank document: - """ - Learning is the process of acquiring new understanding, knowledge, behaviors, skills, values, attitudes, and preferences. The ability to learn is possessed by humans, non-human animals, and some machines; there is also evidence for some kind of learning in certain plants. - """ - And a rerank document: - """ - Machine learning is a field of study in artificial intelligence concerned with the development and study of statistical algorithms that can learn from data and generalize to unseen data, and thus perform tasks without explicit instructions. - """ - And a rerank document: - """ - Paris, capitale de la France, est une grande ville européenne et un centre mondial de l'art, de la mode, de la gastronomie et de la culture. Son paysage urbain du XIXe siècle est traversé par de larges boulevards et la Seine. - """ - When reranking request - Then reranking results are returned - Then reranking highest score is index 2 and lowest score is index 3 diff --git a/examples/server/tests/features/results.feature b/examples/server/tests/features/results.feature deleted file mode 100644 index e8e1b5414..000000000 --- a/examples/server/tests/features/results.feature +++ /dev/null @@ -1,118 +0,0 @@ -@llama.cpp -@results -Feature: Results - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/split/stories15M-00001-of-00003.gguf from HF repo ggml-org/models - And a model file test-model-00001-of-00003.gguf - And 128 as batch size - And 1024 KV cache size - And 128 max tokens to predict - And continuous batching - - Scenario Outline: consistent results with same seed - Given slots - And 1.0 temperature - Then the server is starting - Then the server is healthy - - Given 4 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 42 - - Given concurrent completion requests - Then the server is busy - Then the server is idle - And all slots are idle - Then all predictions are equal - Examples: - | n_slots | - | 1 | - # FIXME: unified KV cache nondeterminism - # | 2 | - - Scenario Outline: different results with different seed - Given slots - And 1.0 temperature - Then the server is starting - Then the server is healthy - - Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 42 - Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 43 - Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 44 - Given 1 prompts "Title: Little Red Riding Hood But In Space\n\nSummary:" with seed 45 - - Given concurrent completion requests - Then the server is busy - Then the server is idle - And all slots are idle - Then all predictions are different - Examples: - | n_slots | - | 1 | - | 2 | - - Scenario Outline: consistent results with same seed and varying batch size - Given 4 slots - And temperature - # And 0 as draft - Then the server is starting - Then the server is healthy - - Given 1 prompts "Write a very long story about AI." with seed 42 - And concurrent completion requests - # Then the server is busy # Not all slots will be utilized. - Then the server is idle - And all slots are idle - - Given prompts "Write a very long story about AI." with seed 42 - And concurrent completion requests - # Then the server is busy # Not all slots will be utilized. - Then the server is idle - And all slots are idle - - Then all predictions are equal - Examples: - | n_parallel | temp | - | 1 | 0.0 | - | 1 | 1.0 | - # FIXME: unified KV cache nondeterminism - # See https://github.com/ggerganov/whisper.cpp/issues/1941#issuecomment-1986923227 - # and https://github.com/ggerganov/llama.cpp/pull/6122#discussion_r1531405574 - # and https://github.com/ggerganov/llama.cpp/pull/7347 . - # | 2 | 0.0 | - # | 4 | 0.0 | - # | 2 | 1.0 | - # | 4 | 1.0 | - - Scenario Outline: consistent token probs with same seed and prompt - Given slots - And KV cache size - And 1.0 temperature - And max tokens to predict - Then the server is starting - Then the server is healthy - - Given 1 prompts "The meaning of life is" with seed 42 - And concurrent completion requests - # Then the server is busy # Not all slots will be utilized. - Then the server is idle - And all slots are idle - - Given prompts "The meaning of life is" with seed 42 - And concurrent completion requests - # Then the server is busy # Not all slots will be utilized. - Then the server is idle - And all slots are idle - - Then all token probabilities are equal - Examples: - | n_slots | n_kv | n_predict | n_parallel | - | 4 | 1024 | 1 | 1 | - # FIXME: unified KV cache nondeterminism - # See https://github.com/ggerganov/whisper.cpp/issues/1941#issuecomment-1986923227 - # and https://github.com/ggerganov/llama.cpp/pull/6122#discussion_r1531405574 - # and https://github.com/ggerganov/llama.cpp/pull/7347 . - # | 4 | 1024 | 1 | 4 | - # | 4 | 1024 | 100 | 1 | - # This test still fails even the above patches; the first token probabilities are already different. - # | 4 | 1024 | 100 | 4 | diff --git a/examples/server/tests/features/security.feature b/examples/server/tests/features/security.feature deleted file mode 100644 index ef30007c3..000000000 --- a/examples/server/tests/features/security.feature +++ /dev/null @@ -1,68 +0,0 @@ -@llama.cpp -@security -Feature: Security - - Background: Server startup with an api key defined - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K.gguf from HF repo ggml-org/models - And a server api key THIS_IS_THE_KEY - Then the server is starting - Then the server is healthy - - Scenario Outline: Completion with some user api key - Given a prompt test - And a user api key - And 4 max tokens to predict - And a completion request with api error - - Examples: Prompts - | api_key | api_error | - | THIS_IS_THE_KEY | no | - | THIS_IS_THE_KEY | no | - | hackeme | raised | - | | raised | - - Scenario Outline: OAI Compatibility - Given a system prompt test - And a user prompt test - And a model test - And 2 max tokens to predict - And streaming is disabled - And a user api key - Given an OAI compatible chat completions request with api error - - Examples: Prompts - | api_key | api_error | - | THIS_IS_THE_KEY | no | - | THIS_IS_THE_KEY | no | - | hackme | raised | - - Scenario Outline: OAI Compatibility (invalid response formats) - Given a system prompt test - And a user prompt test - And a response format - And a model test - And 2 max tokens to predict - And streaming is disabled - Given an OAI compatible chat completions request with raised api error - - Examples: Prompts - | response_format | - | {"type": "sound"} | - | {"type": "json_object", "schema": 123} | - | {"type": "json_object", "schema": {"type": 123}} | - | {"type": "json_object", "schema": {"type": "hiccup"}} | - - - Scenario Outline: CORS Options - Given a user api key THIS_IS_THE_KEY - When an OPTIONS request is sent from - Then CORS header is set to - - Examples: Headers - | origin | cors_header | cors_header_value | - | localhost | Access-Control-Allow-Origin | localhost | - | web.mydomain.fr | Access-Control-Allow-Origin | web.mydomain.fr | - | origin | Access-Control-Allow-Credentials | true | - | web.mydomain.fr | Access-Control-Allow-Methods | GET, POST | - | web.mydomain.fr | Access-Control-Allow-Headers | * | diff --git a/examples/server/tests/features/server.feature b/examples/server/tests/features/server.feature deleted file mode 100644 index 15e24c624..000000000 --- a/examples/server/tests/features/server.feature +++ /dev/null @@ -1,120 +0,0 @@ -@llama.cpp -@server -Feature: llama.cpp server - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K.gguf from HF repo ggml-org/models - And a model file test-model.gguf - And a model alias tinyllama-2 - And BOS token is 1 - And 42 as server seed - # KV Cache corresponds to the total amount of tokens - # that can be stored across all independent sequences: #4130 - # see --ctx-size and #5568 - And 256 KV cache size - And 32 as batch size - And 2 slots - And 64 server max tokens to predict - And prometheus compatible metrics exposed - Then the server is starting - Then the server is healthy - - Scenario: Health - Then the server is ready - And all slots are idle - - - Scenario Outline: Completion - Given a prompt - And max tokens to predict - And a completion request with no api error - Then tokens are predicted matching - And the completion is truncated - And prompt tokens are processed - And prometheus metrics are exposed - And metric llamacpp:tokens_predicted is - - Examples: Prompts - | prompt | n_predict | re_content | n_prompt | n_predicted | truncated | - | I believe the meaning of life is | 8 | (read\|going)+ | 18 | 8 | not | - | Write a joke about AI from a very long prompt which will not be truncated | 256 | (princesses\|everyone\|kids\|Anna\|forest)+ | 46 | 64 | not | - - Scenario: Completion prompt truncated - Given a prompt: - """ - Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. - Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. - Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. - Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. - """ - And a completion request with no api error - Then 64 tokens are predicted matching fun|Annaks|popcorns|pictry|bowl - And the completion is truncated - And 109 prompt tokens are processed - - - Scenario Outline: OAI Compatibility - Given a model - And a system prompt - And a user prompt - And max tokens to predict - And streaming is - Given an OAI compatible chat completions request with no api error - Then tokens are predicted matching - And prompt tokens are processed - And the completion is truncated - - Examples: Prompts - | model | system_prompt | user_prompt | max_tokens | re_content | n_prompt | n_predicted | enable_streaming | truncated | - | llama-2 | Book | What is the best book | 8 | (Here\|what)+ | 77 | 8 | disabled | not | - | codellama70b | You are a coding assistant. | Write the fibonacci function in c++. | 128 | (thanks\|happy\|bird\|Annabyear)+ | -1 | 64 | enabled | | - - - Scenario Outline: OAI Compatibility w/ response format - Given a model test - And a system prompt test - And a user prompt test - And a response format - And 10 max tokens to predict - Given an OAI compatible chat completions request with no api error - Then tokens are predicted matching - - Examples: Prompts - | response_format | n_predicted | re_content | - | {"type": "json_object", "schema": {"const": "42"}} | 6 | "42" | - | {"type": "json_object", "schema": {"items": [{"type": "integer"}]}} | 10 | \[ -300 \] | - | {"type": "json_object"} | 10 | \{ " Jacky. | - - - Scenario: Tokenize / Detokenize - When tokenizing: - """ - What is the capital of France ? - """ - Then tokens can be detokenized - And tokens do not begin with BOS - - Scenario: Tokenize w/ BOS - Given adding special tokens - When tokenizing: - """ - What is the capital of Germany? - """ - Then tokens begin with BOS - Given first token is removed - Then tokens can be detokenized - - Scenario: Tokenize with pieces - When tokenizing with pieces: - """ - What is the capital of Germany? - 媽 - """ - Then tokens are given with pieces - - Scenario: Models available - Given available models - Then 1 models are supported - Then model 0 is identified by tinyllama-2 - Then model 0 is trained on 128 tokens context diff --git a/examples/server/tests/features/slotsave.feature b/examples/server/tests/features/slotsave.feature deleted file mode 100644 index 1c281c074..000000000 --- a/examples/server/tests/features/slotsave.feature +++ /dev/null @@ -1,58 +0,0 @@ -@llama.cpp -@slotsave -Feature: llama.cpp server slot management - - Background: Server startup - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K.gguf from HF repo ggml-org/models - And prompt caching is enabled - And 2 slots - And . as slot save path - And 2048 KV cache size - And 42 as server seed - And 24 max tokens to predict - Then the server is starting - Then the server is healthy - - Scenario: Save and Restore Slot - # First prompt in slot 1 should be fully processed - Given a user prompt "What is the capital of France?" - And using slot id 1 - And a completion request with no api error - Then 24 tokens are predicted matching (Lily|cake) - And 22 prompt tokens are processed - When the slot 1 is saved with filename "slot1.bin" - Then the server responds with status code 200 - # Since we have cache, this should only process the last tokens - Given a user prompt "What is the capital of Germany?" - And a completion request with no api error - Then 24 tokens are predicted matching (Thank|special) - And 7 prompt tokens are processed - # Loading the original cache into slot 0, - # we should only be processing 1 prompt token and get the same output - When the slot 0 is restored with filename "slot1.bin" - Then the server responds with status code 200 - Given a user prompt "What is the capital of France?" - And using slot id 0 - And a completion request with no api error - Then 24 tokens are predicted matching (Lily|cake) - And 1 prompt tokens are processed - # For verification that slot 1 was not corrupted during slot 0 load, same thing - Given a user prompt "What is the capital of Germany?" - And using slot id 1 - And a completion request with no api error - Then 24 tokens are predicted matching (Thank|special) - And 1 prompt tokens are processed - - Scenario: Erase Slot - Given a user prompt "What is the capital of France?" - And using slot id 1 - And a completion request with no api error - Then 24 tokens are predicted matching (Lily|cake) - And 22 prompt tokens are processed - When the slot 1 is erased - Then the server responds with status code 200 - Given a user prompt "What is the capital of France?" - And a completion request with no api error - Then 24 tokens are predicted matching (Lily|cake) - And 22 prompt tokens are processed diff --git a/examples/server/tests/features/steps/steps.py b/examples/server/tests/features/steps/steps.py deleted file mode 100644 index 687b163f4..000000000 --- a/examples/server/tests/features/steps/steps.py +++ /dev/null @@ -1,1518 +0,0 @@ -#!/usr/bin/env python3 -# -*- coding: utf-8 -*- - -import asyncio -import json -import os -import re -import socket -import subprocess -import sys -import threading -import time -import requests -from collections.abc import Sequence -from contextlib import closing -from re import RegexFlag -from typing import Any, Literal, cast - -import aiohttp -import numpy as np -import openai -from openai.types.chat import ChatCompletionChunk -from behave import step # pyright: ignore[reportAttributeAccessIssue] -from behave.api.async_step import async_run_until_complete -from prometheus_client import parser - -# pyright: reportRedeclaration=false - -DEFAULT_TIMEOUT_SECONDS = aiohttp.ClientTimeout(total=600) - -@step("a server listening on {server_fqdn}:{server_port}") -def step_server_config(context, server_fqdn: str, server_port: str): - context.server_fqdn = server_fqdn - context.server_port = int(server_port) - context.n_threads = None - context.n_gpu_layer = None - if 'PORT' in os.environ: - context.server_port = int(os.environ['PORT']) - print(f"$PORT set, overriding server port with to {context.server_port}") - if 'FQDN' in os.environ: - context.server_fqdn = os.environ['FQDN'] - print(f"$FQDN set, overriding server fqdn with to {context.server_fqdn}") - if 'N_GPU_LAYERS' in os.environ: - context.n_gpu_layer = int(os.environ['N_GPU_LAYERS']) - print(f"$N_GPU_LAYERS set, overriding n_gpu_layer with to {context.n_gpu_layer}") - - context.base_url = f'http://{context.server_fqdn}:{context.server_port}' - - context.model_alias = None - context.model_file = None - context.model_hf_repo = None - context.model_hf_file = None - context.model_url = None - context.n_batch = None - context.n_ubatch = None - context.n_ctx = None - context.n_ga = None - context.n_ga_w = None - context.n_predict = None - context.n_prompts = 0 - context.n_server_predict = None - context.slot_save_path = None - context.id_slot = None - context.cache_prompt = None - context.n_slots = None - context.prompt_prefix = None - context.prompt_suffix = None - context.server_api_key = None - context.server_continuous_batching = False - context.server_embeddings = False - context.server_reranking = False - context.server_metrics = False - context.server_process = None - context.seed = None - context.draft = None - context.server_seed = None - context.user_api_key = None - context.response_format = None - context.temperature = None - context.lora_file = None - context.disable_ctx_shift = False - - # infill - context.infill_input_extra = None - context.infill_input_suffix = '' - context.infill_input_prefix = '' - - context.tasks_result = [] - context.concurrent_tasks = [] - context.prompts = [] - - context.reranking_query = None - context.reranking_documents = [] - context.reranking_results = None - - -@step('a model file {hf_file} from HF repo {hf_repo}') -def step_download_hf_model(context, hf_file: str, hf_repo: str): - context.model_hf_repo = hf_repo - context.model_hf_file = hf_file - context.model_file = os.path.basename(hf_file) - -@step('a lora adapter file from {lora_file_url}') -def step_download_lora_file(context, lora_file_url: str): - file_name = lora_file_url.split('/').pop() - context.lora_file = f'../../../{file_name}' - with open(context.lora_file, 'wb') as f: - f.write(requests.get(lora_file_url).content) - -@step('a model file {model_file}') -def step_model_file(context, model_file: str): - context.model_file = model_file - - -@step('a model url {model_url}') -def step_model_url(context, model_url: str): - context.model_url = model_url - - -@step('a model alias {model_alias}') -def step_model_alias(context, model_alias: str): - context.model_alias = model_alias - - -@step('{seed:d} as server seed') -def step_seed(context, seed: int): - context.server_seed = seed - - -@step('{ngl:d} GPU offloaded layers') -def step_n_gpu_layer(context, ngl: int): - if 'N_GPU_LAYERS' in os.environ: - new_ngl = int(os.environ['N_GPU_LAYERS']) - if context.debug: - print(f"-ngl upgraded from {ngl} to {new_ngl}") - ngl = new_ngl - context.n_gpu_layer = ngl - - -@step('{n_threads:d} threads') -def step_n_threads(context, n_threads: int): - context.n_thread = n_threads - - -@step('{draft:d} as draft') -def step_draft(context, draft: int): - context.draft = draft - - -@step('{n_ctx:d} KV cache size') -def step_n_ctx(context, n_ctx: int): - context.n_ctx = n_ctx - - -@step('{n_slots:d} slots') -def step_n_slots(context, n_slots: int): - context.n_slots = n_slots - - -@step('{n_predict:d} server max tokens to predict') -def step_server_n_predict(context, n_predict: int): - context.n_server_predict = n_predict if n_predict > 0 else None - - -@step('{slot_save_path} as slot save path') -def step_slot_save_path(context, slot_save_path: str): - context.slot_save_path = slot_save_path - - -@step('using slot id {id_slot:d}') -def step_id_slot(context, id_slot: int): - context.id_slot = id_slot - - -@step('prompt caching is enabled') -def step_enable_prompt_cache(context): - context.cache_prompt = True - - -@step('continuous batching') -def step_server_continuous_batching(context): - context.server_continuous_batching = True - - -@step('enable embeddings endpoint') -def step_server_embeddings(context): - context.server_embeddings = True - -@step('enable reranking endpoint') -def step_server_reranking(context): - context.server_reranking = True - -@step('prometheus compatible metrics exposed') -def step_server_metrics(context): - context.server_metrics = True - -@step('disable context shifting') -def step_server_disable_ctx_shift(context): - context.disable_ctx_shift = True - -@step("the server is starting") -def step_start_server(context): - start_server_background(context) - attempts = 0 - max_attempts = 20 - if 'GITHUB_ACTIONS' in os.environ: - max_attempts *= 2 - - addrs = socket.getaddrinfo(context.server_fqdn, context.server_port, type=socket.SOCK_STREAM) - family, typ, proto, _, sockaddr = addrs[0] - - while True: - with closing(socket.socket(family, typ, proto)) as sock: - result = sock.connect_ex(sockaddr) - if result == 0: - print("\x1b[33;46mserver started!\x1b[0m") - return - attempts += 1 - if attempts > max_attempts: - assert False, "server not started" - print(f"waiting for server to start, connect error code = {result}...") - time.sleep(0.1) - - -async def wait_for_server_status_with_timeout(context, expecting_status: Literal['healthy', 'ready', 'idle', 'busy'] | str, timeout: int): - match expecting_status: - case 'healthy': - await wait_for_slots_status(context, context.base_url, 200, - timeout=timeout) - - case 'ready' | 'idle': - await wait_for_slots_status(context, context.base_url, 200, - timeout=timeout, - params={'fail_on_no_slot': 1}, - slots_idle=context.n_slots, - slots_processing=0) - case 'busy': - await wait_for_slots_status(context, context.base_url, 503, - params={'fail_on_no_slot': 1}, - slots_idle=0, - slots_processing=context.n_slots) - case _: - assert False, "unknown status" - - -@step("the server is {expecting_status} with timeout {timeout:d} seconds") -@async_run_until_complete -async def step_wait_for_server_status_with_timeout(context, expecting_status: Literal['healthy', 'ready', 'idle', 'busy'] | str, timeout: int): - await wait_for_server_status_with_timeout(context, expecting_status, timeout) - - -@step("the server is {expecting_status}") -@async_run_until_complete -async def step_wait_for_server_status(context, expecting_status: Literal['healthy', 'ready', 'idle', 'busy'] | str): - await wait_for_server_status_with_timeout(context, expecting_status, 30) - - -@step('all slots are {expected_slot_status_string}') -@async_run_until_complete -async def step_all_slots_status(context, expected_slot_status_string: Literal['idle', 'busy'] | str): - match expected_slot_status_string: - case 'idle': - expected_slot_status = False - case 'busy': - expected_slot_status = True - case _: - assert False, "unknown status" - - expected_slots = [{'id': slot_id, 'is_processing': expected_slot_status} - for slot_id in range(context.n_slots)] - await request_slots_status(context, expected_slots) - - -@step('a completion request with {api_error} api error') -@async_run_until_complete -async def step_request_completion(context, api_error: Literal['raised'] | str): - expect_api_error = api_error == 'raised' or api_error != 'no' - seeds = await completions_seed(context, num_seeds=1) - completion = await request_completion(context.prompts.pop(), - seeds[0] if seeds is not None else seeds, - context.base_url, - debug=context.debug, - n_predict=context.n_predict, - cache_prompt=context.cache_prompt, - id_slot=context.id_slot, - expect_api_error=expect_api_error, - user_api_key=context.user_api_key, - temperature=context.temperature) - context.tasks_result.append(completion) - if context.debug: - print(f"Completion response: {completion}") - if api_error == 'raised': - assert completion == 401, f"completion must be an 401 status code: {completion}" - elif api_error.isdigit(): - api_error_code = int(api_error) - assert completion == api_error_code, f"completion must be an {api_error_code} status code: {completion}" - - -@step('an infill request with {api_error} api error') -@async_run_until_complete -async def step_request_completion(context, api_error: Literal['raised'] | str): - if api_error != 'no': - raise ValueError(f'api_error={api_error} is not yet implemented') - payload = { - "prompt": context.prompts[0], - "input_suffix": context.infill_input_suffix, - "input_prefix": context.infill_input_prefix, - "n_predict": context.n_predict, - "seed": context.seed, - "temperature": context.temperature, - } - if context.infill_input_extra is not None: - payload['input_extra'] = context.infill_input_extra - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/infill', - json=payload) as response: - assert response.status == 200 - context.tasks_result = [await response.json()] - - -@step('{predicted_n:d} tokens are predicted matching {re_content}') -def step_n_tokens_predicted_with_content(context, predicted_n, re_content): - context.completion = context.tasks_result.pop() - assert_n_tokens_predicted(context.completion, predicted_n, re_content) - - -@step('{predicted_n:d} tokens are predicted') -def step_n_tokens_predicted(context, predicted_n): - context.completion = context.tasks_result.pop() - assert_n_tokens_predicted(context.completion, predicted_n) - - -@step('all predictions are equal') -@async_run_until_complete -async def step_predictions_equal(context): - n_completions = await gather_tasks_results(context) - assert n_completions >= 2, "need at least 2 completions" - assert_all_predictions_equal(context.tasks_result) - context.tasks_result = [] - - -@step('all predictions are different') -@async_run_until_complete -async def step_predictions_different(context): - n_completions = await gather_tasks_results(context) - assert n_completions >= 2, "need at least 2 completions" - assert_all_predictions_different(context.tasks_result) - context.tasks_result = [] - - -@step('all token probabilities are equal') -@async_run_until_complete -async def step_token_probabilities_equal(context): - n_completions = await gather_tasks_results(context) - assert n_completions >= 2, "need at least 2 completions" - assert_all_token_probabilities_equal(context.tasks_result) - context.tasks_result = [] - - -@step('the completion is truncated') -def step_assert_completion_truncated(context): - step_assert_completion_truncated(context, '') - - -@step('the completion is {truncated} truncated') -def step_assert_completion_truncated(context, truncated): - truncated = truncated != "not" - assert context.completion['truncated'] == truncated, f'{context.completion}' - - -@step('{n_prompt:d} prompt tokens are processed') -def step_impl(context, n_prompt): - assert n_prompt < 0 or n_prompt == context.completion['timings']['prompt_n'], f"n_prompt={context.completion['timings']['prompt_n']}" - - -@step('a user prompt {user_prompt}') -def step_user_prompt(context, user_prompt): - context.prompts.append(user_prompt) - context.n_prompts = len(context.prompts) - - -@step('a system prompt {system_prompt}') -def step_system_prompt(context, system_prompt): - context.system_prompt = system_prompt - - -@step('a model {model}') -def step_model(context, model): - context.model = model - - -@step('{max_tokens:d} max tokens to predict') -def step_max_tokens(context, max_tokens): - context.n_predict = max_tokens - - -@step('a response format {response_format}') -def step_response_format(context, response_format): - context.response_format = json.loads(response_format) - - -@step('{temperature:f} temperature') -def step_temperature(context, temperature): - context.temperature = temperature - - -@step('streaming is {enable_streaming}') -def step_streaming(context, enable_streaming): - context.enable_streaming = enable_streaming == 'enabled' - - -@step('a user api key {user_api_key}') -def step_user_api_key(context, user_api_key): - context.user_api_key = user_api_key - - -@step('no user api key') -def step_no_user_api_key(context): - context.user_api_key = None - - -@step('a user api key ') -def step_no_user_api_key_space(context): - context.user_api_key = None - - -@step('a server api key {server_api_key}') -def step_server_api_key(context, server_api_key): - context.server_api_key = server_api_key - - -@step('{n_junk:d} as number of junk') -def step_n_junk(context, n_junk): - context.n_junk = n_junk - - -@step('{n_batch:d} as batch size') -def step_n_batch(context, n_batch): - context.n_batch = n_batch - - -@step('{n_ubatch:d} as ubatch size') -def step_n_ubatch(context, n_ubatch): - context.n_ubatch = n_ubatch - - -@step('{seed:d} as seed') -def step_seed(context, seed): - if context.seed is None: - context.seed = [seed] - else: - context.seed.append(seed) - - -@step('BOS token is {bos:d}') -def step_bos_token(context, bos): - context.bos = bos - - -@step('a prefix prompt') -def step_prompt_prefix(context): - context.prompt_prefix = context_text(context) - - -@step('a junk suffix prompt') -def step_prompt_junk_suffix(context): - context.prompt_junk_suffix = context_text(context) - - -@step('a suffix prompt') -def step_prompt_suffix(context): - context.prompt_suffix = context_text(context) - - -@step('{n_ga:d} group attention factor' - ' to extend context size through self-extend') -def step_impl(context, n_ga): - context.n_ga = n_ga - - -@step('{n_ga_w:d} group attention width to extend context size through self-extend') -def step_impl(context, n_ga_w): - context.n_ga_w = n_ga_w - - -@step('a passkey prompt template') -def step_prompt_passkey(context): - context.prompt_passkey = context_text(context) - -@step('a rerank query') -def step_set_rerank_query(context): - context.reranking_query = context_text(context) - context.reranking_documents = [] - -@step('a rerank document') -def step_set_rerank_document(context): - context.reranking_documents.append(context_text(context)) - -@step('{n_prompts:d} fixed prompts') -def step_fixed_prompts(context, n_prompts): - context.prompts.extend([str(0)*(context.n_batch if context.n_batch is not None else 512) for i in range(n_prompts)]) - context.n_prompts = n_prompts - - -@step('a "{passkey}" passkey challenge prompt with the passkey inserted every {i_pos:d} junk') -def step_prompt_passkey(context, passkey, i_pos): - prompt = "" - for i in range(context.n_junk): - if i % context.n_junk == i_pos: - prompt += context.prompt_passkey # the passkey is already substituted - prompt += context.prompt_junk_suffix - if context.debug: - passkey_highlight = "\x1b[33m" + passkey + "\x1b[0m" - print(f"Passkey challenge:\n```{prompt.replace(passkey, passkey_highlight)}```") - context.prompts.append(context.prompt_prefix + prompt + context.prompt_suffix) - context.n_prompts = len(context.prompts) - - -@step('an OAI compatible chat completions request with {api_error} api error') -@async_run_until_complete -async def step_oai_chat_completions(context, api_error): - if context.debug: - print(f"Submitting OAI compatible completions request...") - expect_api_error = api_error == 'raised' - seeds = await completions_seed(context, num_seeds=1), - completion = await oai_chat_completions(context.prompts.pop(), - seeds[0] if seeds is not None else seeds, - context.system_prompt, - context.base_url, - '/v1/chat', - False, - model=context.model if hasattr(context, 'model') else None, - - n_predict=context.n_predict - if hasattr(context, 'n_predict') else None, - - enable_streaming=context.enable_streaming - if hasattr(context, 'enable_streaming') else None, - - response_format=context.response_format - if hasattr(context, 'response_format') else None, - - user_api_key=context.user_api_key - if hasattr(context, 'user_api_key') else None, - - expect_api_error=expect_api_error) - context.tasks_result.append(completion) - if context.debug: - print(f"Completion response: {completion}") - if expect_api_error: - assert completion == 401, f"completion must be an 401 status code: {completion}" - - if context.debug: - print(f"Completion response: {completion}") - - -@step('a prompt') -def step_a_prompt(context): - context.prompts.append(context_text(context)) - context.n_prompts = len(context.prompts) - - -@step('a prompt {prompt}') -def step_a_prompt_prompt(context, prompt): - context.prompts.append(prompt) - context.n_prompts = len(context.prompts) - - -# TODO: allow this to be repeated -@step('an infill input extra {filename} {text}') -def step_infill_input_extra(context, filename, text): - if filename == 'none': - context.infill_input_extra = None - else: - context.infill_input_extra = [{'filename': filename, 'text': text}] - - -@step('an infill input suffix {text}') -def step_infill_input_suffix(context, text): - context.infill_input_suffix = text - - -@step('an infill input prefix {text}') -def step_infill_input_prefix(context, text): - context.infill_input_prefix = text - - -@step('{num_prompts:d} prompts {prompt} with seed {seed:d}') -def step_many_prompts(context, num_prompts, prompt, seed): - if context.seed is None: - context.seed = [] - for _ in range(num_prompts): - context.seed.append(seed) - context.prompts.append(prompt) - context.n_prompts = len(context.prompts) - - -@step('concurrent completion requests') -@async_run_until_complete() -async def step_concurrent_completion_requests(context): - await concurrent_requests( - context, - request_completion, - # prompt is inserted automatically - context.base_url, - debug=context.debug, - prompt_prefix=context.prompt_prefix, - prompt_suffix=context.prompt_suffix, - n_predict=context.n_predict if hasattr(context, 'n_predict') else None, - user_api_key=context.user_api_key if hasattr(context, 'user_api_key') else None, - temperature=context.temperature, - ) - - -@step('concurrent OAI completions requests') -@async_run_until_complete -async def step_oai_chat_completions(context): - await concurrent_requests(context, oai_chat_completions, - # user_prompt is inserted automatically - context.system_prompt, - context.base_url, - '/v1/chat/completions', - True, # async_client - model=context.model - if hasattr(context, 'model') else None, - n_predict=context.n_predict - if hasattr(context, 'n_predict') else None, - enable_streaming=context.enable_streaming - if hasattr(context, 'enable_streaming') else None, - response_format=context.response_format - if hasattr(context, 'response_format') else None, - user_api_key=context.user_api_key - if hasattr(context, 'user_api_key') else None) - - -@step('concurrent OAI completions requests no v1') -@async_run_until_complete -async def step_oai_chat_completions(context): - await concurrent_requests(context, oai_chat_completions, - # user_prompt is inserted automatically - context.system_prompt, - context.base_url, - '/chat/completions', - True, # async_client - model=context.model - if hasattr(context, 'model') else None, - n_predict=context.n_predict - if hasattr(context, 'n_predict') else None, - enable_streaming=context.enable_streaming - if hasattr(context, 'enable_streaming') else None, - response_format=context.response_format - if hasattr(context, 'response_format') else None, - user_api_key=context.user_api_key - if hasattr(context, 'user_api_key') else None) - - -@step('all prompts are predicted') -@async_run_until_complete -async def step_all_prompts_are_predicted(context): - await all_prompts_are_predicted(context) - - -@step('all prompts are predicted with {n_expected_predicted:d} tokens') -@async_run_until_complete -async def step_all_prompts_are_predicted_with_n_tokens(context, n_expected_predicted): - await all_prompts_are_predicted(context, n_expected_predicted) - - -async def all_prompts_are_predicted(context, expected_predicted_n=None): - n_completions = await gather_tasks_results(context) - assert n_completions > 0 - for i in range(n_completions): - assert_n_tokens_predicted(context.tasks_result.pop(), expected_predicted_n=expected_predicted_n) - assert len(context.concurrent_tasks) == 0, f"{len(context.concurrent_tasks)} pending requests" - - -@step('embeddings are computed for') -@async_run_until_complete -async def step_compute_embedding(context): - context.n_prompts = 1 - context.embeddings = await request_embedding(context_text(context), None, base_url=context.base_url) - - -@step('reranking request') -@async_run_until_complete -async def step_compute_reranking(context): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/reranking', - json={ - "query": context.reranking_query, - "documents": context.reranking_documents, - }) as response: - if response.status == 200: - response_json = await response.json() - context.reranking_results = response_json['results'] - else: - context.reranking_results = response.status - - -@step('all embeddings are the same') -@async_run_until_complete -async def step_all_embeddings_are_the_same(context): - n_embedding_requests = await gather_tasks_results(context) - assert n_embedding_requests > 0 - embeddings = [] - for i in range(n_embedding_requests): - embedding = context.tasks_result.pop().pop() - embeddings.append(embedding) - assert_embeddings(embedding) - n = len(embeddings) - for i in range(n-1): - for j in range(i+1, n): - embedding1 = np.array(embeddings[i]) - embedding2 = np.array(embeddings[j]) - if context.debug: - print(f"embedding1: {embedding1[-8:]}") - print(f"embedding2: {embedding2[-8:]}") - similarity = np.dot(embedding1, embedding2) / (np.linalg.norm(embedding1) * np.linalg.norm(embedding2)) - msg = f"Similarity between {i} and {j}: {similarity:.10f}" - if context.debug: - print(f"{msg}") - assert np.isclose(similarity, 1.0, rtol=1e-05, atol=1e-08, equal_nan=False), msg - - -@step('embeddings are generated') -def step_assert_embeddings(context): - assert context.n_prompts == len(context.embeddings), (f"unexpected response:\n" - f"context.n_prompts={context.n_prompts}\n" - f"context.embeddings={context.embeddings}") - for embedding in context.embeddings: - assert_embeddings(embedding) - -@step('embeddings request with {api_error_code:d} api error') -def step_assert_embeddings(context, api_error_code: int): - assert context.embeddings == api_error_code, f"embeddings request must return code {api_error_code}, but got {context.embeddings}" - -@step('an OAI compatible embeddings computation request for') -@async_run_until_complete -async def step_oai_compute_embeddings(context): - context.n_prompts = 1 - context.embeddings = await request_oai_embeddings(context_text(context), None, - base_url=context.base_url, - user_api_key=context.user_api_key, - model=context.model) - - -@step('an OAI compatible embeddings computation request for multiple inputs') -@async_run_until_complete -async def step_oai_compute_embeddings_multiple_inputs(context): - context.embeddings = await request_oai_embeddings(context.prompts, None, - base_url=context.base_url, - user_api_key=context.user_api_key, - model=context.model) - context.prompts.clear() - - -@step('concurrent embedding requests') -@async_run_until_complete() -async def step_concurrent_embedding_requests(context): - await concurrent_requests(context, - request_embedding, - # prompt is inserted automatically - base_url=context.base_url) - - -@step('concurrent OAI embedding requests') -@async_run_until_complete() -async def step_concurrent_oai_embedding_requests(context): - await concurrent_requests(context, - request_oai_embeddings, - # prompt is inserted automatically - base_url=context.base_url, - async_client=True, - model=context.model) - - -@step('all embeddings are generated') -@async_run_until_complete() -async def all_embeddings_are_generated(context): - n_embedding_requests = await gather_tasks_results(context) - assert n_embedding_requests == context.n_prompts - for i in range(n_embedding_requests): - assert_embeddings(context.tasks_result.pop().pop()) - -@step('reranking results are returned') -def reranking_results_are_returned(context): - assert len(context.reranking_results) == len(context.reranking_documents) - -@step('reranking highest score is index {idx_high:d} and lowest score is index {idx_low:d}') -def reranking_results_are_returned(context, idx_high: int, idx_low: int): - max_score, max_idx = 0, 0 - min_score, min_idx = 0, 0 - for res in context.reranking_results: - if max_score < res['relevance_score']: - max_score = res['relevance_score'] - max_idx = res['index'] - if min_score > res['relevance_score']: - min_score = res['relevance_score'] - min_idx = res['index'] - print(context.reranking_results) - assert max_idx == idx_high - assert min_idx == idx_low - -@step('adding special tokens') -def step_tokenize_set_add_special(context): - context.tokenize_add_special = True - - -@step("tokenizing with pieces") -@async_run_until_complete -async def step_tokenize_with_pieces(context): - context.tokenized_text = context_text(context) - async with aiohttp.ClientSession() as session: - tokenize_args = {"content": context.tokenized_text, "with_pieces": True} - if getattr(context, "tokenize_add_special", None) is not None: - tokenize_args["add_special"] = context.tokenize_add_special - - async with session.post( - f"{context.base_url}/tokenize", json=tokenize_args - ) as response: - assert response.status == 200 - tokenize_json = await response.json() - context.tokens_with_pieces = tokenize_json["tokens"] - - -@step("tokens are given with pieces") -@async_run_until_complete -async def step_tokenize_with_pieces(context): - # Verify that the response contains both token IDs and pieces - assert all( - "id" in token and "piece" in token for token in context.tokens_with_pieces - ) - - -@step('tokenizing') -@async_run_until_complete -async def step_tokenize(context): - context.tokenized_text = context_text(context) - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - tokenize_args = { - "content": context.tokenized_text, - } - if getattr(context, 'tokenize_add_special', None) is not None: - tokenize_args['add_special'] = context.tokenize_add_special - async with session.post(f'{context.base_url}/tokenize', - json=tokenize_args) as response: - assert response.status == 200 - tokenize_json = await response.json() - context.tokens = tokenize_json['tokens'] - - -@step('tokens can be detokenized') -@async_run_until_complete -async def step_detokenize(context): - assert len(context.tokens) > 0 - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/detokenize', - json={ - "tokens": context.tokens, - }) as response: - assert response.status == 200 - detokenize_json = await response.json() - # SPM tokenizer adds a whitespace prefix: https://github.com/google/sentencepiece/issues/15 - assert context.tokenized_text == detokenize_json['content'].strip() - - -@step('tokens begin with BOS') -def step_strings_for_tokenization(context): - assert context.tokens[0] == context.bos - - -@step('tokens do not begin with BOS') -def step_strings_for_tokenization(context): - assert context.tokens[0] != context.bos - - -@step('first token is removed') -def step_strings_for_tokenization(context): - context.tokens = context.tokens[1:] - - -@step('an OPTIONS request is sent from {origin}') -@async_run_until_complete -async def step_options_request(context, origin): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - headers = {'Authorization': f'Bearer {context.user_api_key}', 'Origin': origin} - async with session.options(f'{context.base_url}/v1/chat/completions', - headers=headers) as response: - assert response.status == 200 - context.options_response = response - - -@step('CORS header {cors_header} is set to {cors_header_value}') -def step_check_options_header_value(context, cors_header, cors_header_value): - assert context.options_response.headers[cors_header] == cors_header_value - - -@step('prometheus metrics are exposed') -@async_run_until_complete -async def step_prometheus_metrics_exported(context): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with await session.get(f'{context.base_url}/metrics') as metrics_response: - assert metrics_response.status == 200 - assert metrics_response.headers['Content-Type'] == "text/plain; version=0.0.4" - metrics_raw = await metrics_response.text() - metric_exported = False - if context.debug: - print(f"/metrics answer:\n{metrics_raw}") - context.metrics = {} - for metric in parser.text_string_to_metric_families(metrics_raw): - match metric.name: - case "llamacpp:kv_cache_usage_ratio": - assert len(metric.samples) > 0 - metric_exported = True - context.metrics[metric.name] = metric - assert int(metrics_response.headers["Process-Start-Time-Unix"]) > 0, "no header process start time" - assert metric_exported, "No metrics exported" - - -@step('metric {metric_name} is {metric_value:d}') -def step_assert_metric_value(context, metric_name, metric_value): - if metric_name not in context.metrics: - assert False, f"no metric {metric_name} in {context.metrics.keys()}" - assert context.metrics[metric_name].samples[0].value == metric_value, f"metric: {context.metrics[metric_name]}" - - -@step('available models') -def step_available_models(context): - # openai client always expects an api_key - openai.api_key = context.user_api_key if context.user_api_key is not None else 'nope' - openai.base_url = f'{context.base_url}/v1/' - context.models = openai.models.list().data - - -@step('{n_model:d} models are supported') -def step_supported_models(context, n_model): - if context.debug: - print("server models available:", context.models) - assert len(context.models) == n_model - - -@step('model {i_model:d} is {param} {preposition} {param_value}') -def step_supported_models(context, i_model: int, param: Literal['identified', 'trained'] | str, preposition: str, param_value: str): - assert i_model < len(context.models) - model = context.models[i_model] - - param_value = param_value.split(' ', 1)[0] - match param: - case 'identified': - value = model.id - case 'trained': - value = str(model.meta["n_ctx_train"]) - case _: - assert False, "param {param} not supported" - assert param_value == value, f"model param {param} {value} != {param_value}" - - -async def concurrent_requests(context, f_completion, *args, **kwargs): - context.n_prompts = len(context.prompts) - if context.debug: - print(f"starting {context.n_prompts} concurrent completion requests...") - assert context.n_prompts > 0 - seeds = await completions_seed(context) - assert seeds is not None - for prompt_no in range(context.n_prompts): - shifted_args = [context.prompts.pop(), seeds[prompt_no], *args] - context.concurrent_tasks.append(asyncio.create_task(f_completion(*shifted_args, **kwargs))) - await asyncio.sleep(0.01) - - -@step('the slot {slot_id:d} is saved with filename "{filename}"') -@async_run_until_complete -async def step_save_slot(context, slot_id, filename): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/slots/{slot_id}?action=save', - json={"filename": filename}, - headers={"Content-Type": "application/json"}) as response: - context.response = response - - -@step('the slot {slot_id:d} is restored with filename "{filename}"') -@async_run_until_complete -async def step_restore_slot(context, slot_id, filename): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/slots/{slot_id}?action=restore', - json={"filename": filename}, - headers={"Content-Type": "application/json"}) as response: - context.response = response - - -@step('the slot {slot_id:d} is erased') -@async_run_until_complete -async def step_erase_slot(context, slot_id): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/slots/{slot_id}?action=erase', - headers={"Content-Type": "application/json"}) as response: - context.response = response - - -@step('switch {on_or_off} lora adapter {lora_id:d}') -@async_run_until_complete -async def toggle_lora_adapter(context, on_or_off: str, lora_id: int): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{context.base_url}/lora-adapters', - json=[{'id': lora_id, 'scale': 1 if on_or_off == 'on' else 0}], - headers={"Content-Type": "application/json"}) as response: - context.response = response - print([{'id': lora_id, 'scale': 1 if on_or_off == 'on' else 0}]) - - -@step('the server responds with status code {status_code:d}') -def step_server_responds_with_status_code(context, status_code): - assert context.response.status == status_code - - -async def request_completion(prompt, - seed, - base_url, - debug=False, - prompt_prefix=None, - prompt_suffix=None, - n_predict=None, - cache_prompt=False, - id_slot=None, - expect_api_error=None, - user_api_key=None, - temperature=None) -> int | dict[str, Any]: - if debug: - print(f"Sending completion request: {prompt}") - origin = "my.super.domain" - headers = { - 'Origin': origin - } - if user_api_key is not None: - if debug: - print(f"Set user_api_key: {user_api_key}") - headers['Authorization'] = f'Bearer {user_api_key}' - - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{base_url}/completion', - json={ - "input_prefix": prompt_prefix, - "prompt": prompt, - "input_suffix": prompt_suffix, - "n_predict": n_predict if n_predict is not None else -1, - "cache_prompt": cache_prompt, - "id_slot": id_slot, - "seed": seed if seed is not None else 42, - "temperature": temperature if temperature is not None else 0.8, - "n_probs": 2, - }, - headers=headers) as response: - if expect_api_error is None or not expect_api_error: - assert response.status == 200 - assert response.headers['Access-Control-Allow-Origin'] == origin - return await response.json() - else: - return response.status - - -async def oai_chat_completions(user_prompt, - seed, - system_prompt, - base_url: str, - base_path: str, - async_client, - debug=False, - temperature=None, - model=None, - n_predict=None, - enable_streaming=None, - response_format=None, - user_api_key=None, - expect_api_error=None) -> int | dict[str, Any]: - if debug: - print(f"Sending OAI Chat completions request: {user_prompt}") - # openai client always expects an api key - user_api_key = user_api_key if user_api_key is not None else 'nope' - seed = seed if seed is not None else 42 - enable_streaming = enable_streaming if enable_streaming is not None else False - payload = { - "messages": [ - { - "role": "system", - "content": system_prompt, - }, - { - "role": "user", - "content": user_prompt, - } - ], - "model": model, - "max_tokens": n_predict, - "stream": enable_streaming, - "temperature": temperature if temperature is not None else 0.0, - "seed": seed, - } - if response_format is not None: - payload['response_format'] = response_format - completion_response = { - 'content': '', - 'timings': { - 'predicted_n': 0, - 'prompt_n': 0 - } - } - if async_client: - origin = 'llama.cpp' - headers = {'Authorization': f'Bearer {user_api_key}', 'Origin': origin} - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{base_url}{base_path}', - json=payload, - headers=headers) as response: - if enable_streaming: - assert response.status == 200 - assert response.headers['Access-Control-Allow-Origin'] == origin - assert response.headers['Content-Type'] == "text/event-stream" - event_received = True - while event_received: - event_received = False - async for line_in_bytes in response.content: - line = line_in_bytes.decode('utf-8') - line = line.rstrip('\n').rstrip('\r') - if line == '': - continue - event_data = line.split(': ', 1) - assert event_data[0] == 'data', f'Bad event code received: ```{event_data}```' - chunk_raw = event_data[1] - if chunk_raw == '[DONE]': - break - - chunk = json.loads(chunk_raw) - assert len(chunk['choices']) == 1, f"no choices provided, line ```{line}```" - delta = chunk['choices'][0]['delta'] - if 'content' in delta: - completion_response['content'] += delta['content'] - completion_response['timings']['predicted_n'] += 1 - else: - if expect_api_error is None or not expect_api_error: - assert response.status == 200 - assert response.headers['Access-Control-Allow-Origin'] == origin - assert response.headers['Content-Type'] == "application/json; charset=utf-8" - chat_completion_raw = await response.json() - completion_response = { - 'content': chat_completion_raw['choices'][0]['message'], - 'timings': { - 'predicted_n': chat_completion_raw['usage']['completion_tokens'], - 'prompt_n': chat_completion_raw['usage']['prompt_tokens'] - } - } - else: - return response.status - else: - try: - openai.api_key = user_api_key - openai.base_url = f'{base_url}{base_path.removesuffix("chat")}' - assert model is not None - chat_completion = openai.chat.completions.create( - messages=payload['messages'], - model=model, - max_tokens=n_predict, - stream=enable_streaming, - response_format=payload.get('response_format') or openai.NOT_GIVEN, - seed=seed, - temperature=payload['temperature'] - ) - except openai.AuthenticationError as e: - if expect_api_error is not None and expect_api_error: - return 401 - else: - assert False, f'error raised: {e}' - - if enable_streaming: - chat_completion = cast(openai.Stream[ChatCompletionChunk], chat_completion) - for chunk in chat_completion: - assert len(chunk.choices) == 1 - delta = chunk.choices[0].delta - if delta.content is not None: - completion_response['content'] += delta.content - completion_response['timings']['predicted_n'] += 1 - completion_response['truncated'] = chunk.choices[0].finish_reason != 'stop' - else: - assert len(chat_completion.choices) == 1 - assert chat_completion.usage is not None - completion_response = { - 'content': chat_completion.choices[0].message.content, - 'timings': { - 'predicted_n': chat_completion.usage.completion_tokens, - 'prompt_n': chat_completion.usage.prompt_tokens - }, - 'truncated': chat_completion.choices[0].finish_reason != 'stop' - } - if debug: - print("OAI response formatted to llama.cpp:", completion_response) - return completion_response - - -async def request_embedding(content, seed, base_url=None) -> list[list[float]] | int: - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{base_url}/embedding', - json={ - "content": content, - }) as response: - if response.status == 200: - response_json = await response.json() - return [response_json['embedding']] - else: - return response.status - - -async def request_oai_embeddings(input, seed, - base_url=None, user_api_key=None, - model=None, async_client=False) -> list[list[float]]: - # openai client always expects an api_key - user_api_key = user_api_key if user_api_key is not None else 'nope' - if async_client: - origin = 'llama.cpp' - headers=[] - if user_api_key is not None: - headers = {'Authorization': f'Bearer {user_api_key}', 'Origin': origin} - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with session.post(f'{base_url}/v1/embeddings', - json={ - "input": input, - "model": model, - }, - headers=headers) as response: - assert response.status == 200, f"received status code not expected: {response.status}" - assert response.headers['Access-Control-Allow-Origin'] == origin - assert response.headers['Content-Type'] == "application/json; charset=utf-8" - response_json = await response.json() - assert response_json['model'] == model, f"invalid model received: {response_json['model']}" - assert response_json['object'] == 'list' - if isinstance(input, Sequence): - embeddings = [] - for an_oai_embeddings in response_json['data']: - embeddings.append(an_oai_embeddings['embedding']) - else: - embeddings = [response_json['data']['embedding']] - return embeddings - else: - openai.api_key = user_api_key - openai.base_url = f'{base_url}/v1/' - assert model is not None - oai_embeddings = openai.embeddings.create( - model=model, - input=input, - ) - - return [e.embedding for e in oai_embeddings.data] - - -def assert_n_tokens_predicted(completion_response, expected_predicted_n=None, re_content=None): - content = completion_response['content'] - n_predicted = completion_response['timings']['predicted_n'] - assert len(content) > 0, "no token predicted" - if re_content is not None: - p = re.compile(re_content, flags=RegexFlag.IGNORECASE | RegexFlag.MULTILINE | RegexFlag.DOTALL) - matches = p.finditer(content) - last_match = 0 - highlighted = '' - for match in matches: - start, end = match.span() - highlighted += content[last_match: start] - highlighted += '\x1b[33m' - highlighted += content[start: end] - highlighted += '\x1b[0m' - last_match = end - highlighted += content[last_match:] - if 'DEBUG' in os.environ and os.environ['DEBUG'] == 'ON': - print(f"Checking completion response: {highlighted}") - assert last_match > 0, f'/{re_content}/ must match ```{highlighted}```' - if expected_predicted_n and expected_predicted_n > 0: - assert n_predicted == expected_predicted_n, (f'invalid number of tokens predicted:' - f' {n_predicted} <> {expected_predicted_n}') - -def assert_all_predictions_equal(completion_responses): - if 'DEBUG' in os.environ and os.environ['DEBUG'] == 'ON': - for i, response_i in enumerate(completion_responses): - content_i = response_i['content'] - print(f"content {i}: {content_i}") - for i, response_i in enumerate(completion_responses): - content_i = response_i['content'] - for j, response_j in enumerate(completion_responses): - if i == j: - continue - content_j = response_j['content'] - assert content_i == content_j, "contents not equal" - - -def assert_all_predictions_different(completion_responses): - if 'DEBUG' in os.environ and os.environ['DEBUG'] == 'ON': - for i, response_i in enumerate(completion_responses): - content_i = response_i['content'] - print(f"content {i}: {content_i}") - for i, response_i in enumerate(completion_responses): - content_i = response_i['content'] - for j, response_j in enumerate(completion_responses): - if i == j: - continue - content_j = response_j['content'] - assert content_i != content_j, "contents not different" - - -def assert_all_token_probabilities_equal(completion_responses): - n_predict = len(completion_responses[0]['completion_probabilities']) - if 'DEBUG' in os.environ and os.environ['DEBUG'] == 'ON': - for pos in range(n_predict): - for i, response_i in enumerate(completion_responses): - probs_i = response_i['completion_probabilities'][pos]['probs'] - print(f"pos {pos}, probs {i}: {probs_i}") - for pos in range(n_predict): - for i, response_i in enumerate(completion_responses): - probs_i = response_i['completion_probabilities'][pos]['probs'] - for j, response_j in enumerate(completion_responses): - if i == j: - continue - probs_j = response_j['completion_probabilities'][pos]['probs'] - assert probs_i == probs_j, "contents not equal" - - -async def gather_tasks_results(context): - n_tasks = len(context.concurrent_tasks) - if context.debug: - print(f"Waiting for all {n_tasks} tasks results...") - for task_no in range(n_tasks): - context.tasks_result.append(await context.concurrent_tasks.pop()) - n_completions = len(context.tasks_result) - return n_completions - - -async def wait_for_slots_status(context, - base_url, - expected_http_status_code, - timeout=3, - params=None, - slots_idle=None, - slots_processing=None): - if context.debug: - print(f"Starting checking for health for expected_http_status_code={expected_http_status_code}") - interval = 0.5 - counter = 0 - if 'GITHUB_ACTIONS' in os.environ: - timeout *= 2 - - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - while True: - headers = {'Authorization': f'Bearer {context.server_api_key}'} - async with await session.get(f'{base_url}/slots', params=params, headers=headers) as slots_response: - status_code = slots_response.status - slots = await slots_response.json() - if context.debug: - print(f"slots responses {slots}\n") - if status_code == 503 and status_code == expected_http_status_code: - return - if status_code == 200 and status_code == expected_http_status_code: - n_slots_idle = sum(1 if not slot["is_processing"] else 0 for slot in slots) - n_slots_processing = sum(1 if slot["is_processing"] else 0 for slot in slots) - if ((slots_idle is None or slots_idle == n_slots_idle) - and (slots_processing is None or slots_processing == n_slots_processing)): - return - await asyncio.sleep(interval) - - counter += interval - if counter >= timeout: - # Sometimes health requests are triggered after completions are predicted - if expected_http_status_code == 503: - if len(context.tasks_result) == 0: - print("\x1b[5;37;43mWARNING: forcing concurrent tasks," - " busy health check missed, probably too fast inference\x1b[0m\n") - n_completions = await gather_tasks_results(context) - if n_completions > 0: - return - - assert False, f'slots check timeout exceeded {counter}s>={timeout}' - - -def assert_embeddings(embeddings): - assert len(embeddings) > 0 - embeddings_computed = False - for emb in embeddings: - if not isinstance(emb, float): - assert False, f"Bad embeddings: {embeddings}" - if emb != 0: - embeddings_computed = True - assert embeddings_computed, f"Embeddings: {embeddings}" - - -async def request_slots_status(context, expected_slots): - async with aiohttp.ClientSession(timeout=DEFAULT_TIMEOUT_SECONDS) as session: - async with await session.get(f'{context.base_url}/slots') as slots_response: - assert slots_response.status == 200 - slots = await slots_response.json() - assert_slots_status(slots, expected_slots) - - -def assert_slots_status(slots, expected_slots): - assert len(slots) == len(expected_slots) - for slot_id, (expected, slot) in enumerate(zip(expected_slots, slots)): - for key in expected: - assert expected[key] == slot[key], (f"invalid slot {slot_id}" - f" expected[{key}] != slot[{key}]" - f" = {expected[key]} != {slot[key]}") - - -async def completions_seed(context, num_seeds=None): - if hasattr(context, "seed") and context.seed is not None: - assert len(context.seed) == context.n_prompts - if num_seeds is None: - num_seeds = context.n_prompts - assert num_seeds <= context.n_prompts - seeds = context.seed[:num_seeds] - context.seed = context.seed[num_seeds:] if num_seeds < context.n_prompts else None - return seeds - - if hasattr(context, "server_seed") and context.server_seed is not None: - if num_seeds is None: - return [context.server_seed] * context.n_prompts - else: - return [context.server_seed] * num_seeds - return None - - -def context_text(context): - return context.text.replace('\r', '') - - -def start_server_background(context): - if os.name == 'nt': - context.server_path = '../../../build/bin/Release/llama-server.exe' - else: - context.server_path = '../../../build/bin/llama-server' - if 'LLAMA_SERVER_BIN_PATH' in os.environ: - context.server_path = os.environ['LLAMA_SERVER_BIN_PATH'] - server_listen_addr = context.server_fqdn - server_args = [ - '--slots', # requires to get slot status via /slots endpoint - '--host', server_listen_addr, - '--port', context.server_port, - ] - if context.model_file: - server_args.extend(['--model', context.model_file]) - if context.model_url: - server_args.extend(['--model-url', context.model_url]) - if context.model_hf_repo: - server_args.extend(['--hf-repo', context.model_hf_repo]) - if context.model_hf_file: - server_args.extend(['--hf-file', context.model_hf_file]) - if context.n_batch: - server_args.extend(['--batch-size', context.n_batch]) - if context.n_ubatch: - server_args.extend(['--ubatch-size', context.n_ubatch]) - if context.n_threads: - server_args.extend(['--threads', context.threads]) - if context.n_gpu_layer: - server_args.extend(['--n-gpu-layers', context.n_gpu_layer]) - if context.draft is not None: - server_args.extend(['--draft', context.draft]) - if context.server_continuous_batching: - server_args.append('--cont-batching') - if context.server_embeddings: - server_args.append('--embedding') - if context.server_reranking: - server_args.append('--reranking') - if context.server_metrics: - server_args.append('--metrics') - if context.model_alias: - server_args.extend(['--alias', context.model_alias]) - if context.n_ctx: - server_args.extend(['--ctx-size', context.n_ctx]) - if context.n_slots: - server_args.extend(['--parallel', context.n_slots]) - if context.n_server_predict: - server_args.extend(['--n-predict', context.n_server_predict]) - if context.slot_save_path: - server_args.extend(['--slot-save-path', context.slot_save_path]) - if context.server_api_key: - server_args.extend(['--api-key', context.server_api_key]) - if context.n_ga: - server_args.extend(['--grp-attn-n', context.n_ga]) - if context.n_ga_w: - server_args.extend(['--grp-attn-w', context.n_ga_w]) - if context.debug: - server_args.append('--verbose') - if context.lora_file: - server_args.extend(['--lora', context.lora_file]) - if context.disable_ctx_shift: - server_args.extend(['--no-context-shift']) - - args = [str(arg) for arg in [context.server_path, *server_args]] - print(f"bench: starting server with: {' '.join(args)}") - - flags = 0 - if 'nt' == os.name: - flags |= subprocess.DETACHED_PROCESS - flags |= subprocess.CREATE_NEW_PROCESS_GROUP - flags |= subprocess.CREATE_NO_WINDOW - - pkwargs = { - 'creationflags': flags, - 'stdout': subprocess.PIPE, - 'stderr': subprocess.PIPE - } - context.server_process = subprocess.Popen( - [str(arg) for arg in [context.server_path, *server_args]], - **pkwargs) # pyright: ignore[reportArgumentType, reportCallIssue] - - def server_log(in_stream, out_stream): - for line in iter(in_stream.readline, b''): - print(line.decode('utf-8'), end='', file=out_stream) - - thread_stdout = threading.Thread(target=server_log, args=(context.server_process.stdout, sys.stdout)) - thread_stdout.start() - - thread_stderr = threading.Thread(target=server_log, args=(context.server_process.stderr, sys.stderr)) - thread_stderr.start() - - print(f"server pid={context.server_process.pid}, behave pid={os.getpid()}") diff --git a/examples/server/tests/features/wrong_usages.feature b/examples/server/tests/features/wrong_usages.feature deleted file mode 100644 index 61d5f315e..000000000 --- a/examples/server/tests/features/wrong_usages.feature +++ /dev/null @@ -1,25 +0,0 @@ -# run with: ./tests.sh --no-skipped --tags wrong_usage -@wrong_usage -Feature: Wrong usage of llama.cpp server - - #3969 The user must always set --n-predict option - # to cap the number of tokens any completion request can generate - # or pass n_predict/max_tokens in the request. - Scenario: Infinite loop - Given a server listening on localhost:8080 - And a model file tinyllamas/stories260K.gguf from HF repo ggml-org/models - And 42 as server seed - And 2048 KV cache size - # Uncomment below to fix the issue - #And 64 server max tokens to predict - Then the server is starting - Then the server is healthy - Given a prompt: - """ - Go to: infinite loop - """ - # Uncomment below to fix the issue - #And 128 max tokens to predict - Given concurrent completion requests - Then the server is idle - Then all prompts are predicted diff --git a/examples/server/tests/requirements.txt b/examples/server/tests/requirements.txt index 553954872..15d024914 100644 --- a/examples/server/tests/requirements.txt +++ b/examples/server/tests/requirements.txt @@ -1,7 +1,8 @@ aiohttp~=3.9.3 -behave~=1.2.6 +pytest~=8.3.3 huggingface_hub~=0.23.2 numpy~=1.26.4 -openai~=1.30.3 +openai~=1.55.3 prometheus-client~=0.20.0 requests~=2.32.3 +wget~=3.2 diff --git a/examples/server/tests/tests.sh b/examples/server/tests/tests.sh index 72a0fbad8..1e0777de3 100755 --- a/examples/server/tests/tests.sh +++ b/examples/server/tests/tests.sh @@ -1,11 +1,14 @@ #!/bin/bash +# make sure we are in the right directory +SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd ) +cd $SCRIPT_DIR + set -eu if [ $# -lt 1 ] then - # Start @llama.cpp scenario - behave --summary --stop --no-capture --exclude 'issues|wrong_usages|passkey' --tags llama.cpp + pytest -v -x else - behave "$@" + pytest "$@" fi diff --git a/examples/server/tests/unit/test_basic.py b/examples/server/tests/unit/test_basic.py new file mode 100644 index 000000000..1485de8ce --- /dev/null +++ b/examples/server/tests/unit/test_basic.py @@ -0,0 +1,96 @@ +import pytest +import requests +from utils import * + +server = ServerPreset.tinyllama2() + + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + + +def test_server_start_simple(): + global server + server.start() + res = server.make_request("GET", "/health") + assert res.status_code == 200 + + +def test_server_props(): + global server + server.start() + res = server.make_request("GET", "/props") + assert res.status_code == 200 + assert ".gguf" in res.body["model_path"] + assert res.body["total_slots"] == server.n_slots + default_val = res.body["default_generation_settings"] + assert server.n_ctx is not None and server.n_slots is not None + assert default_val["n_ctx"] == server.n_ctx / server.n_slots + assert default_val["params"]["seed"] == server.seed + + +def test_server_models(): + global server + server.start() + res = server.make_request("GET", "/models") + assert res.status_code == 200 + assert len(res.body["data"]) == 1 + assert res.body["data"][0]["id"] == server.model_alias + + +def test_server_slots(): + global server + + # without slots endpoint enabled, this should return error + server.server_slots = False + server.start() + res = server.make_request("GET", "/slots") + assert res.status_code == 501 # ERROR_TYPE_NOT_SUPPORTED + assert "error" in res.body + server.stop() + + # with slots endpoint enabled, this should return slots info + server.server_slots = True + server.n_slots = 2 + server.start() + res = server.make_request("GET", "/slots") + assert res.status_code == 200 + assert len(res.body) == server.n_slots + assert server.n_ctx is not None and server.n_slots is not None + assert res.body[0]["n_ctx"] == server.n_ctx / server.n_slots + assert "params" in res.body[0] + assert res.body[0]["params"]["seed"] == server.seed + + +def test_load_split_model(): + global server + server.model_hf_repo = "ggml-org/models" + server.model_hf_file = "tinyllamas/split/stories15M-q8_0-00001-of-00003.gguf" + server.model_alias = "tinyllama-split" + server.start() + res = server.make_request("POST", "/completion", data={ + "n_predict": 16, + "prompt": "Hello", + "temperature": 0.0, + }) + assert res.status_code == 200 + assert match_regex("(little|girl)+", res.body["content"]) + + +def test_no_webui(): + global server + # default: webui enabled + server.start() + url = f"http://{server.server_host}:{server.server_port}" + res = requests.get(url) + assert res.status_code == 200 + assert "" in res.text + server.stop() + + # with --no-webui + server.no_webui = True + server.start() + res = requests.get(url) + assert res.status_code == 404 diff --git a/examples/server/tests/unit/test_chat_completion.py b/examples/server/tests/unit/test_chat_completion.py new file mode 100644 index 000000000..b15dba6eb --- /dev/null +++ b/examples/server/tests/unit/test_chat_completion.py @@ -0,0 +1,245 @@ +import pytest +from openai import OpenAI +from utils import * + +server = ServerPreset.tinyllama2() + + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + + +@pytest.mark.parametrize( + "model,system_prompt,user_prompt,max_tokens,re_content,n_prompt,n_predicted,finish_reason", + [ + (None, "Book", "What is the best book", 8, "(Suddenly)+", 77, 8, "length"), + ("codellama70b", "You are a coding assistant.", "Write the fibonacci function in c++.", 128, "(Aside|she|felter|alonger)+", 104, 64, "length"), + ] +) +def test_chat_completion(model, system_prompt, user_prompt, max_tokens, re_content, n_prompt, n_predicted, finish_reason): + global server + server.start() + res = server.make_request("POST", "/chat/completions", data={ + "model": model, + "max_tokens": max_tokens, + "messages": [ + {"role": "system", "content": system_prompt}, + {"role": "user", "content": user_prompt}, + ], + }) + assert res.status_code == 200 + assert "cmpl" in res.body["id"] # make sure the completion id has the expected format + assert res.body["system_fingerprint"].startswith("b") + assert res.body["model"] == model if model is not None else server.model_alias + assert res.body["usage"]["prompt_tokens"] == n_prompt + assert res.body["usage"]["completion_tokens"] == n_predicted + choice = res.body["choices"][0] + assert "assistant" == choice["message"]["role"] + assert match_regex(re_content, choice["message"]["content"]) + assert choice["finish_reason"] == finish_reason + + +@pytest.mark.parametrize( + "system_prompt,user_prompt,max_tokens,re_content,n_prompt,n_predicted,finish_reason", + [ + ("Book", "What is the best book", 8, "(Suddenly)+", 77, 8, "length"), + ("You are a coding assistant.", "Write the fibonacci function in c++.", 128, "(Aside|she|felter|alonger)+", 104, 64, "length"), + ] +) +def test_chat_completion_stream(system_prompt, user_prompt, max_tokens, re_content, n_prompt, n_predicted, finish_reason): + global server + server.model_alias = None # try using DEFAULT_OAICOMPAT_MODEL + server.start() + res = server.make_stream_request("POST", "/chat/completions", data={ + "max_tokens": max_tokens, + "messages": [ + {"role": "system", "content": system_prompt}, + {"role": "user", "content": user_prompt}, + ], + "stream": True, + }) + content = "" + last_cmpl_id = None + for data in res: + choice = data["choices"][0] + assert data["system_fingerprint"].startswith("b") + assert "gpt-3.5" in data["model"] # DEFAULT_OAICOMPAT_MODEL, maybe changed in the future + if last_cmpl_id is None: + last_cmpl_id = data["id"] + assert last_cmpl_id == data["id"] # make sure the completion id is the same for all events in the stream + if choice["finish_reason"] in ["stop", "length"]: + assert data["usage"]["prompt_tokens"] == n_prompt + assert data["usage"]["completion_tokens"] == n_predicted + assert "content" not in choice["delta"] + assert match_regex(re_content, content) + assert choice["finish_reason"] == finish_reason + else: + assert choice["finish_reason"] is None + content += choice["delta"]["content"] + + +def test_chat_completion_with_openai_library(): + global server + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.chat.completions.create( + model="gpt-3.5-turbo-instruct", + messages=[ + {"role": "system", "content": "Book"}, + {"role": "user", "content": "What is the best book"}, + ], + max_tokens=8, + seed=42, + temperature=0.8, + ) + assert res.system_fingerprint is not None and res.system_fingerprint.startswith("b") + assert res.choices[0].finish_reason == "length" + assert res.choices[0].message.content is not None + assert match_regex("(Suddenly)+", res.choices[0].message.content) + + +def test_chat_template(): + global server + server.chat_template = "llama3" + server.debug = True # to get the "__verbose" object in the response + server.start() + res = server.make_request("POST", "/chat/completions", data={ + "max_tokens": 8, + "messages": [ + {"role": "system", "content": "Book"}, + {"role": "user", "content": "What is the best book"}, + ] + }) + assert res.status_code == 200 + assert "__verbose" in res.body + assert res.body["__verbose"]["prompt"] == " <|start_header_id|>system<|end_header_id|>\n\nBook<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWhat is the best book<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n" + + +@pytest.mark.parametrize("response_format,n_predicted,re_content", [ + ({"type": "json_object", "schema": {"const": "42"}}, 6, "\"42\""), + ({"type": "json_object", "schema": {"items": [{"type": "integer"}]}}, 10, "[ -3000 ]"), + ({"type": "json_object"}, 10, "(\\{|John)+"), + ({"type": "sound"}, 0, None), + # invalid response format (expected to fail) + ({"type": "json_object", "schema": 123}, 0, None), + ({"type": "json_object", "schema": {"type": 123}}, 0, None), + ({"type": "json_object", "schema": {"type": "hiccup"}}, 0, None), +]) +def test_completion_with_response_format(response_format: dict, n_predicted: int, re_content: str | None): + global server + server.start() + res = server.make_request("POST", "/chat/completions", data={ + "max_tokens": n_predicted, + "messages": [ + {"role": "system", "content": "You are a coding assistant."}, + {"role": "user", "content": "Write an example"}, + ], + "response_format": response_format, + }) + if re_content is not None: + assert res.status_code == 200 + choice = res.body["choices"][0] + assert match_regex(re_content, choice["message"]["content"]) + else: + assert res.status_code != 200 + assert "error" in res.body + + +@pytest.mark.parametrize("messages", [ + None, + "string", + [123], + [{}], + [{"role": 123}], + [{"role": "system", "content": 123}], + # [{"content": "hello"}], # TODO: should not be a valid case + [{"role": "system", "content": "test"}, {}], +]) +def test_invalid_chat_completion_req(messages): + global server + server.start() + res = server.make_request("POST", "/chat/completions", data={ + "messages": messages, + }) + assert res.status_code == 400 or res.status_code == 500 + assert "error" in res.body + + +def test_chat_completion_with_timings_per_token(): + global server + server.start() + res = server.make_stream_request("POST", "/chat/completions", data={ + "max_tokens": 10, + "messages": [{"role": "user", "content": "test"}], + "stream": True, + "timings_per_token": True, + }) + for data in res: + assert "timings" in data + assert "prompt_per_second" in data["timings"] + assert "predicted_per_second" in data["timings"] + assert "predicted_n" in data["timings"] + assert data["timings"]["predicted_n"] <= 10 + + +def test_logprobs(): + global server + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.chat.completions.create( + model="gpt-3.5-turbo-instruct", + temperature=0.0, + messages=[ + {"role": "system", "content": "Book"}, + {"role": "user", "content": "What is the best book"}, + ], + max_tokens=5, + logprobs=True, + top_logprobs=10, + ) + output_text = res.choices[0].message.content + aggregated_text = '' + assert res.choices[0].logprobs is not None + assert res.choices[0].logprobs.content is not None + for token in res.choices[0].logprobs.content: + aggregated_text += token.token + assert token.logprob <= 0.0 + assert token.bytes is not None + assert len(token.top_logprobs) > 0 + assert aggregated_text == output_text + + +def test_logprobs_stream(): + global server + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.chat.completions.create( + model="gpt-3.5-turbo-instruct", + temperature=0.0, + messages=[ + {"role": "system", "content": "Book"}, + {"role": "user", "content": "What is the best book"}, + ], + max_tokens=5, + logprobs=True, + top_logprobs=10, + stream=True, + ) + output_text = '' + aggregated_text = '' + for data in res: + choice = data.choices[0] + if choice.finish_reason is None: + if choice.delta.content: + output_text += choice.delta.content + assert choice.logprobs is not None + assert choice.logprobs.content is not None + for token in choice.logprobs.content: + aggregated_text += token.token + assert token.logprob <= 0.0 + assert token.bytes is not None + assert token.top_logprobs is not None + assert len(token.top_logprobs) > 0 + assert aggregated_text == output_text diff --git a/examples/server/tests/unit/test_completion.py b/examples/server/tests/unit/test_completion.py new file mode 100644 index 000000000..e5e3b6077 --- /dev/null +++ b/examples/server/tests/unit/test_completion.py @@ -0,0 +1,407 @@ +import pytest +import time +from openai import OpenAI +from utils import * + +server = ServerPreset.tinyllama2() + + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + +@pytest.mark.parametrize("prompt,n_predict,re_content,n_prompt,n_predicted,truncated,return_tokens", [ + ("I believe the meaning of life is", 8, "(going|bed)+", 18, 8, False, False), + ("Write a joke about AI from a very long prompt which will not be truncated", 256, "(princesses|everyone|kids|Anna|forest)+", 46, 64, False, True), +]) +def test_completion(prompt: str, n_predict: int, re_content: str, n_prompt: int, n_predicted: int, truncated: bool, return_tokens: bool): + global server + server.start() + res = server.make_request("POST", "/completion", data={ + "n_predict": n_predict, + "prompt": prompt, + "return_tokens": return_tokens, + }) + assert res.status_code == 200 + assert res.body["timings"]["prompt_n"] == n_prompt + assert res.body["timings"]["predicted_n"] == n_predicted + assert res.body["truncated"] == truncated + assert type(res.body["has_new_line"]) == bool + assert match_regex(re_content, res.body["content"]) + if return_tokens: + assert len(res.body["tokens"]) > 0 + assert all(type(tok) == int for tok in res.body["tokens"]) + else: + assert res.body["tokens"] == [] + + +@pytest.mark.parametrize("prompt,n_predict,re_content,n_prompt,n_predicted,truncated", [ + ("I believe the meaning of life is", 8, "(going|bed)+", 18, 8, False), + ("Write a joke about AI from a very long prompt which will not be truncated", 256, "(princesses|everyone|kids|Anna|forest)+", 46, 64, False), +]) +def test_completion_stream(prompt: str, n_predict: int, re_content: str, n_prompt: int, n_predicted: int, truncated: bool): + global server + server.start() + res = server.make_stream_request("POST", "/completion", data={ + "n_predict": n_predict, + "prompt": prompt, + "stream": True, + }) + content = "" + for data in res: + assert "stop" in data and type(data["stop"]) == bool + if data["stop"]: + assert data["timings"]["prompt_n"] == n_prompt + assert data["timings"]["predicted_n"] == n_predicted + assert data["truncated"] == truncated + assert data["stop_type"] == "limit" + assert type(data["has_new_line"]) == bool + assert "generation_settings" in data + assert server.n_predict is not None + assert data["generation_settings"]["n_predict"] == min(n_predict, server.n_predict) + assert data["generation_settings"]["seed"] == server.seed + assert match_regex(re_content, content) + else: + assert len(data["tokens"]) > 0 + assert all(type(tok) == int for tok in data["tokens"]) + content += data["content"] + + +def test_completion_stream_vs_non_stream(): + global server + server.start() + res_stream = server.make_stream_request("POST", "/completion", data={ + "n_predict": 8, + "prompt": "I believe the meaning of life is", + "stream": True, + }) + res_non_stream = server.make_request("POST", "/completion", data={ + "n_predict": 8, + "prompt": "I believe the meaning of life is", + }) + content_stream = "" + for data in res_stream: + content_stream += data["content"] + assert content_stream == res_non_stream.body["content"] + + +def test_completion_stream_with_openai_library(): + global server + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.completions.create( + model="davinci-002", + prompt="I believe the meaning of life is", + max_tokens=8, + ) + assert res.system_fingerprint is not None and res.system_fingerprint.startswith("b") + assert res.choices[0].finish_reason == "length" + assert res.choices[0].text is not None + assert match_regex("(going|bed)+", res.choices[0].text) + + +def test_completion_with_openai_library(): + global server + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.completions.create( + model="davinci-002", + prompt="I believe the meaning of life is", + max_tokens=8, + stream=True, + ) + output_text = '' + for data in res: + choice = data.choices[0] + if choice.finish_reason is None: + assert choice.text is not None + output_text += choice.text + assert match_regex("(going|bed)+", output_text) + + +@pytest.mark.parametrize("n_slots", [1, 2]) +def test_consistent_result_same_seed(n_slots: int): + global server + server.n_slots = n_slots + server.start() + last_res = None + for _ in range(4): + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "seed": 42, + "temperature": 0.0, + "cache_prompt": False, # TODO: remove this once test_cache_vs_nocache_prompt is fixed + }) + if last_res is not None: + assert res.body["content"] == last_res.body["content"] + last_res = res + + +@pytest.mark.parametrize("n_slots", [1, 2]) +def test_different_result_different_seed(n_slots: int): + global server + server.n_slots = n_slots + server.start() + last_res = None + for seed in range(4): + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "seed": seed, + "temperature": 1.0, + "cache_prompt": False, # TODO: remove this once test_cache_vs_nocache_prompt is fixed + }) + if last_res is not None: + assert res.body["content"] != last_res.body["content"] + last_res = res + +# TODO figure why it don't work with temperature = 1 +# @pytest.mark.parametrize("temperature", [0.0, 1.0]) +@pytest.mark.parametrize("n_batch", [16, 32]) +@pytest.mark.parametrize("temperature", [0.0]) +def test_consistent_result_different_batch_size(n_batch: int, temperature: float): + global server + server.n_batch = n_batch + server.start() + last_res = None + for _ in range(4): + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "seed": 42, + "temperature": temperature, + "cache_prompt": False, # TODO: remove this once test_cache_vs_nocache_prompt is fixed + }) + if last_res is not None: + assert res.body["content"] == last_res.body["content"] + last_res = res + + +@pytest.mark.skip(reason="This test fails on linux, need to be fixed") +def test_cache_vs_nocache_prompt(): + global server + server.start() + res_cache = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "seed": 42, + "temperature": 1.0, + "cache_prompt": True, + }) + res_no_cache = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "seed": 42, + "temperature": 1.0, + "cache_prompt": False, + }) + assert res_cache.body["content"] == res_no_cache.body["content"] + + +def test_completion_with_tokens_input(): + global server + server.temperature = 0.0 + server.start() + prompt_str = "I believe the meaning of life is" + res = server.make_request("POST", "/tokenize", data={ + "content": prompt_str, + "add_special": True, + }) + assert res.status_code == 200 + tokens = res.body["tokens"] + + # single completion + res = server.make_request("POST", "/completion", data={ + "prompt": tokens, + }) + assert res.status_code == 200 + assert type(res.body["content"]) == str + + # batch completion + res = server.make_request("POST", "/completion", data={ + "prompt": [tokens, tokens], + }) + assert res.status_code == 200 + assert type(res.body) == list + assert len(res.body) == 2 + assert res.body[0]["content"] == res.body[1]["content"] + + # mixed string and tokens + res = server.make_request("POST", "/completion", data={ + "prompt": [tokens, prompt_str], + }) + assert res.status_code == 200 + assert type(res.body) == list + assert len(res.body) == 2 + assert res.body[0]["content"] == res.body[1]["content"] + + # mixed string and tokens in one sequence + res = server.make_request("POST", "/completion", data={ + "prompt": [1, 2, 3, 4, 5, 6, prompt_str, 7, 8, 9, 10, prompt_str], + }) + assert res.status_code == 200 + assert type(res.body["content"]) == str + + +@pytest.mark.parametrize("n_slots,n_requests", [ + (1, 3), + (2, 2), + (2, 4), + (4, 2), # some slots must be idle + (4, 6), +]) +def test_completion_parallel_slots(n_slots: int, n_requests: int): + global server + server.n_slots = n_slots + server.temperature = 0.0 + server.start() + + PROMPTS = [ + ("Write a very long book.", "(very|special|big)+"), + ("Write another a poem.", "(small|house)+"), + ("What is LLM?", "(Dad|said)+"), + ("The sky is blue and I love it.", "(climb|leaf)+"), + ("Write another very long music lyrics.", "(friends|step|sky)+"), + ("Write a very long joke.", "(cat|Whiskers)+"), + ] + def check_slots_status(): + should_all_slots_busy = n_requests >= n_slots + time.sleep(0.1) + res = server.make_request("GET", "/slots") + n_busy = sum([1 for slot in res.body if slot["is_processing"]]) + if should_all_slots_busy: + assert n_busy == n_slots + else: + assert n_busy <= n_slots + + tasks = [] + for i in range(n_requests): + prompt, re_content = PROMPTS[i % len(PROMPTS)] + tasks.append((server.make_request, ("POST", "/completion", { + "prompt": prompt, + "seed": 42, + "temperature": 1.0, + }))) + tasks.append((check_slots_status, ())) + results = parallel_function_calls(tasks) + + # check results + for i in range(n_requests): + prompt, re_content = PROMPTS[i % len(PROMPTS)] + res = results[i] + assert res.status_code == 200 + assert type(res.body["content"]) == str + assert len(res.body["content"]) > 10 + # FIXME: the result is not deterministic when using other slot than slot 0 + # assert match_regex(re_content, res.body["content"]) + + +@pytest.mark.parametrize( + "prompt,n_predict,response_fields", + [ + ("I believe the meaning of life is", 8, []), + ("I believe the meaning of life is", 32, ["content", "generation_settings/n_predict", "prompt"]), + ], +) +def test_completion_response_fields( + prompt: str, n_predict: int, response_fields: list[str] +): + global server + server.start() + res = server.make_request( + "POST", + "/completion", + data={ + "n_predict": n_predict, + "prompt": prompt, + "response_fields": response_fields, + }, + ) + assert res.status_code == 200 + assert "content" in res.body + assert len(res.body["content"]) + if len(response_fields): + assert res.body["generation_settings/n_predict"] == n_predict + assert res.body["prompt"] == " " + prompt + assert isinstance(res.body["content"], str) + assert len(res.body) == len(response_fields) + else: + assert len(res.body) + assert "generation_settings" in res.body + + +def test_n_probs(): + global server + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "n_probs": 10, + "temperature": 0.0, + "n_predict": 5, + }) + assert res.status_code == 200 + assert "completion_probabilities" in res.body + assert len(res.body["completion_probabilities"]) == 5 + for tok in res.body["completion_probabilities"]: + assert "id" in tok and tok["id"] > 0 + assert "token" in tok and type(tok["token"]) == str + assert "logprob" in tok and tok["logprob"] <= 0.0 + assert "bytes" in tok and type(tok["bytes"]) == list + assert len(tok["top_logprobs"]) == 10 + for prob in tok["top_logprobs"]: + assert "id" in prob and prob["id"] > 0 + assert "token" in prob and type(prob["token"]) == str + assert "logprob" in prob and prob["logprob"] <= 0.0 + assert "bytes" in prob and type(prob["bytes"]) == list + + +def test_n_probs_stream(): + global server + server.start() + res = server.make_stream_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "n_probs": 10, + "temperature": 0.0, + "n_predict": 5, + "stream": True, + }) + for data in res: + if data["stop"] == False: + assert "completion_probabilities" in data + assert len(data["completion_probabilities"]) == 1 + for tok in data["completion_probabilities"]: + assert "id" in tok and tok["id"] > 0 + assert "token" in tok and type(tok["token"]) == str + assert "logprob" in tok and tok["logprob"] <= 0.0 + assert "bytes" in tok and type(tok["bytes"]) == list + assert len(tok["top_logprobs"]) == 10 + for prob in tok["top_logprobs"]: + assert "id" in prob and prob["id"] > 0 + assert "token" in prob and type(prob["token"]) == str + assert "logprob" in prob and prob["logprob"] <= 0.0 + assert "bytes" in prob and type(prob["bytes"]) == list + + +def test_n_probs_post_sampling(): + global server + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "n_probs": 10, + "temperature": 0.0, + "n_predict": 5, + "post_sampling_probs": True, + }) + assert res.status_code == 200 + assert "completion_probabilities" in res.body + assert len(res.body["completion_probabilities"]) == 5 + for tok in res.body["completion_probabilities"]: + assert "id" in tok and tok["id"] > 0 + assert "token" in tok and type(tok["token"]) == str + assert "prob" in tok and 0.0 < tok["prob"] <= 1.0 + assert "bytes" in tok and type(tok["bytes"]) == list + assert len(tok["top_probs"]) == 10 + for prob in tok["top_probs"]: + assert "id" in prob and prob["id"] > 0 + assert "token" in prob and type(prob["token"]) == str + assert "prob" in prob and 0.0 <= prob["prob"] <= 1.0 + assert "bytes" in prob and type(prob["bytes"]) == list + # because the test model usually output token with either 100% or 0% probability, we need to check all the top_probs + assert any(prob["prob"] == 1.0 for prob in tok["top_probs"]) diff --git a/examples/server/tests/unit/test_ctx_shift.py b/examples/server/tests/unit/test_ctx_shift.py new file mode 100644 index 000000000..be93a6d31 --- /dev/null +++ b/examples/server/tests/unit/test_ctx_shift.py @@ -0,0 +1,67 @@ +import pytest +from utils import * + +server = ServerPreset.tinyllama2() + + +LONG_TEXT = """ +Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. +Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. +Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. +Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. +""".strip() + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + server.n_ctx = 256 + server.n_slots = 2 + + +def test_ctx_shift_enabled(): + # the prompt is 301 tokens + # the slot context is 256/2 = 128 tokens + # the prompt is truncated to keep the last 109 tokens + # 64 tokens are generated thanks to shifting the context when it gets full + global server + server.start() + res = server.make_request("POST", "/completion", data={ + "n_predict": 64, + "prompt": LONG_TEXT, + }) + assert res.status_code == 200 + assert res.body["timings"]["prompt_n"] == 109 + assert res.body["timings"]["predicted_n"] == 64 + assert res.body["truncated"] is True + + +@pytest.mark.parametrize("n_predict,n_token_output,truncated", [ + (64, 64, False), + (-1, 120, True), +]) +def test_ctx_shift_disabled_short_prompt(n_predict: int, n_token_output: int, truncated: bool): + global server + server.disable_ctx_shift = True + server.n_predict = -1 + server.start() + res = server.make_request("POST", "/completion", data={ + "n_predict": n_predict, + "prompt": "Hi how are you", + }) + assert res.status_code == 200 + assert res.body["timings"]["predicted_n"] == n_token_output + assert res.body["truncated"] == truncated + + +def test_ctx_shift_disabled_long_prompt(): + global server + server.disable_ctx_shift = True + server.start() + res = server.make_request("POST", "/completion", data={ + "n_predict": 64, + "prompt": LONG_TEXT, + }) + assert res.status_code != 200 + assert "error" in res.body + assert "exceeds the available context size" in res.body["error"]["message"] diff --git a/examples/server/tests/unit/test_embedding.py b/examples/server/tests/unit/test_embedding.py new file mode 100644 index 000000000..8b0eb42b0 --- /dev/null +++ b/examples/server/tests/unit/test_embedding.py @@ -0,0 +1,237 @@ +import base64 +import struct +import pytest +from openai import OpenAI +from utils import * + +server = ServerPreset.bert_bge_small() + +EPSILON = 1e-3 + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.bert_bge_small() + + +def test_embedding_single(): + global server + server.pooling = 'last' + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": "I believe the meaning of life is", + }) + assert res.status_code == 200 + assert len(res.body['data']) == 1 + assert 'embedding' in res.body['data'][0] + assert len(res.body['data'][0]['embedding']) > 1 + + # make sure embedding vector is normalized + assert abs(sum([x ** 2 for x in res.body['data'][0]['embedding']]) - 1) < EPSILON + + +def test_embedding_multiple(): + global server + server.pooling = 'last' + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": [ + "I believe the meaning of life is", + "Write a joke about AI from a very long prompt which will not be truncated", + "This is a test", + "This is another test", + ], + }) + assert res.status_code == 200 + assert len(res.body['data']) == 4 + for d in res.body['data']: + assert 'embedding' in d + assert len(d['embedding']) > 1 + + +@pytest.mark.parametrize( + "input,is_multi_prompt", + [ + # do not crash on empty input + ("", False), + # single prompt + ("string", False), + ([12, 34, 56], False), + ([12, 34, "string", 56, 78], False), + # multiple prompts + (["string1", "string2"], True), + (["string1", [12, 34, 56]], True), + ([[12, 34, 56], [12, 34, 56]], True), + ([[12, 34, 56], [12, "string", 34, 56]], True), + ] +) +def test_embedding_mixed_input(input, is_multi_prompt: bool): + global server + server.start() + res = server.make_request("POST", "/v1/embeddings", data={"input": input}) + assert res.status_code == 200 + data = res.body['data'] + if is_multi_prompt: + assert len(data) == len(input) + for d in data: + assert 'embedding' in d + assert len(d['embedding']) > 1 + else: + assert 'embedding' in data[0] + assert len(data[0]['embedding']) > 1 + + +def test_embedding_pooling_none(): + global server + server.pooling = 'none' + server.start() + res = server.make_request("POST", "/embeddings", data={ + "input": "hello hello hello", + }) + assert res.status_code == 200 + assert 'embedding' in res.body[0] + assert len(res.body[0]['embedding']) == 5 # 3 text tokens + 2 special + + # make sure embedding vector is not normalized + for x in res.body[0]['embedding']: + assert abs(sum([x ** 2 for x in x]) - 1) > EPSILON + + +def test_embedding_pooling_none_oai(): + global server + server.pooling = 'none' + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": "hello hello hello", + }) + + # /v1/embeddings does not support pooling type 'none' + assert res.status_code == 400 + assert "error" in res.body + + +def test_embedding_openai_library_single(): + global server + server.pooling = 'last' + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.embeddings.create(model="text-embedding-3-small", input="I believe the meaning of life is") + assert len(res.data) == 1 + assert len(res.data[0].embedding) > 1 + + +def test_embedding_openai_library_multiple(): + global server + server.pooling = 'last' + server.start() + client = OpenAI(api_key="dummy", base_url=f"http://{server.server_host}:{server.server_port}/v1") + res = client.embeddings.create(model="text-embedding-3-small", input=[ + "I believe the meaning of life is", + "Write a joke about AI from a very long prompt which will not be truncated", + "This is a test", + "This is another test", + ]) + assert len(res.data) == 4 + for d in res.data: + assert len(d.embedding) > 1 + + +def test_embedding_error_prompt_too_long(): + global server + server.pooling = 'last' + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": "This is a test " * 512, + }) + assert res.status_code != 200 + assert "too large" in res.body["error"]["message"] + + +def test_same_prompt_give_same_result(): + server.pooling = 'last' + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": [ + "I believe the meaning of life is", + "I believe the meaning of life is", + "I believe the meaning of life is", + "I believe the meaning of life is", + "I believe the meaning of life is", + ], + }) + assert res.status_code == 200 + assert len(res.body['data']) == 5 + for i in range(1, len(res.body['data'])): + v0 = res.body['data'][0]['embedding'] + vi = res.body['data'][i]['embedding'] + for x, y in zip(v0, vi): + assert abs(x - y) < EPSILON + + +@pytest.mark.parametrize( + "content,n_tokens", + [ + ("I believe the meaning of life is", 9), + ("This is a test", 6), + ] +) +def test_embedding_usage_single(content, n_tokens): + global server + server.start() + res = server.make_request("POST", "/v1/embeddings", data={"input": content}) + assert res.status_code == 200 + assert res.body['usage']['prompt_tokens'] == res.body['usage']['total_tokens'] + assert res.body['usage']['prompt_tokens'] == n_tokens + + +def test_embedding_usage_multiple(): + global server + server.start() + res = server.make_request("POST", "/v1/embeddings", data={ + "input": [ + "I believe the meaning of life is", + "I believe the meaning of life is", + ], + }) + assert res.status_code == 200 + assert res.body['usage']['prompt_tokens'] == res.body['usage']['total_tokens'] + assert res.body['usage']['prompt_tokens'] == 2 * 9 + + +def test_embedding_openai_library_base64(): + server.start() + test_input = "Test base64 embedding output" + + # get embedding in default format + res = server.make_request("POST", "/v1/embeddings", data={ + "input": test_input + }) + assert res.status_code == 200 + vec0 = res.body["data"][0]["embedding"] + + # get embedding in base64 format + res = server.make_request("POST", "/v1/embeddings", data={ + "input": test_input, + "encoding_format": "base64" + }) + + assert res.status_code == 200 + assert "data" in res.body + assert len(res.body["data"]) == 1 + + embedding_data = res.body["data"][0] + assert "embedding" in embedding_data + assert isinstance(embedding_data["embedding"], str) + + # Verify embedding is valid base64 + decoded = base64.b64decode(embedding_data["embedding"]) + # Verify decoded data can be converted back to float array + float_count = len(decoded) // 4 # 4 bytes per float + floats = struct.unpack(f'{float_count}f', decoded) + assert len(floats) > 0 + assert all(isinstance(x, float) for x in floats) + assert len(floats) == len(vec0) + + # make sure the decoded data is the same as the original + for x, y in zip(floats, vec0): + assert abs(x - y) < EPSILON diff --git a/examples/server/tests/unit/test_infill.py b/examples/server/tests/unit/test_infill.py new file mode 100644 index 000000000..10554db0f --- /dev/null +++ b/examples/server/tests/unit/test_infill.py @@ -0,0 +1,77 @@ +import pytest +from utils import * + +server = ServerPreset.tinyllama_infill() + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama_infill() + + +def test_infill_without_input_extra(): + global server + server.start() + res = server.make_request("POST", "/infill", data={ + "input_prefix": "#include \n#include \"llama.h\"\n\nint main() {\n", + "prompt": " int n_threads = llama_", + "input_suffix": "}\n", + }) + assert res.status_code == 200 + assert match_regex("(Ann|small|shiny|Daddy)+", res.body["content"]) + + +def test_infill_with_input_extra(): + global server + server.start() + res = server.make_request("POST", "/infill", data={ + "input_extra": [{ + "filename": "llama.h", + "text": "LLAMA_API int32_t llama_n_threads();\n" + }], + "input_prefix": "#include \n#include \"llama.h\"\n\nint main() {\n", + "prompt": " int n_threads = llama_", + "input_suffix": "}\n", + }) + assert res.status_code == 200 + assert match_regex("(Dad|excited|park)+", res.body["content"]) + + +@pytest.mark.parametrize("input_extra", [ + {}, + {"filename": "ok"}, + {"filename": 123}, + {"filename": 123, "text": "abc"}, + {"filename": 123, "text": 456}, +]) +def test_invalid_input_extra_req(input_extra): + global server + server.start() + res = server.make_request("POST", "/infill", data={ + "input_extra": [input_extra], + "input_prefix": "#include \n#include \"llama.h\"\n\nint main() {\n", + "prompt": " int n_threads = llama_", + "input_suffix": "}\n", + }) + assert res.status_code == 400 + assert "error" in res.body + + +@pytest.mark.skipif(not is_slow_test_allowed(), reason="skipping slow test") +def test_with_qwen_model(): + global server + server.model_file = None + server.model_hf_repo = "ggml-org/Qwen2.5-Coder-1.5B-IQ3_XXS-GGUF" + server.model_hf_file = "qwen2.5-coder-1.5b-iq3_xxs-imat.gguf" + server.start(timeout_seconds=600) + res = server.make_request("POST", "/infill", data={ + "input_extra": [{ + "filename": "llama.h", + "text": "LLAMA_API int32_t llama_n_threads();\n" + }], + "input_prefix": "#include \n#include \"llama.h\"\n\nint main() {\n", + "prompt": " int n_threads = llama_", + "input_suffix": "}\n", + }) + assert res.status_code == 200 + assert res.body["content"] == "n_threads();\n printf(\"Number of threads: %d\\n\", n_threads);\n return 0;\n" diff --git a/examples/server/tests/unit/test_lora.py b/examples/server/tests/unit/test_lora.py new file mode 100644 index 000000000..c1aa8be70 --- /dev/null +++ b/examples/server/tests/unit/test_lora.py @@ -0,0 +1,115 @@ +import pytest +from utils import * + +server = ServerPreset.stories15m_moe() + +LORA_FILE_URL = "https://huggingface.co/ggml-org/stories15M_MOE/resolve/main/moe_shakespeare15M.gguf" + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.stories15m_moe() + server.lora_files = [download_file(LORA_FILE_URL)] + + +@pytest.mark.parametrize("scale,re_content", [ + # without applying lora, the model should behave like a bedtime story generator + (0.0, "(little|girl|three|years|old)+"), + # with lora, the model should behave like a Shakespearean text generator + (1.0, "(eye|love|glass|sun)+"), +]) +def test_lora(scale: float, re_content: str): + global server + server.start() + res_lora_control = server.make_request("POST", "/lora-adapters", data=[ + {"id": 0, "scale": scale} + ]) + assert res_lora_control.status_code == 200 + res = server.make_request("POST", "/completion", data={ + "prompt": "Look in thy glass", + }) + assert res.status_code == 200 + assert match_regex(re_content, res.body["content"]) + + +def test_lora_per_request(): + global server + server.n_slots = 4 + server.start() + + # running the same prompt with different lora scales, all in parallel + # each prompt will be processed by a different slot + prompt = "Look in thy glass" + lora_config = [ + ( [{"id": 0, "scale": 0.0}], "(bright|day|many|happy)+" ), + ( [{"id": 0, "scale": 0.0}], "(bright|day|many|happy)+" ), + ( [{"id": 0, "scale": 0.3}], "(special|thing|gifted)+" ), + ( [{"id": 0, "scale": 0.7}], "(far|from|home|away)+" ), + ( [{"id": 0, "scale": 1.0}], "(eye|love|glass|sun)+" ), + ( [{"id": 0, "scale": 1.0}], "(eye|love|glass|sun)+" ), + ] + + tasks = [( + server.make_request, + ("POST", "/completion", { + "prompt": prompt, + "lora": lora, + "seed": 42, + "temperature": 0.0, + "cache_prompt": False, # TODO: remove this once test_cache_vs_nocache_prompt is fixed + }) + ) for lora, _ in lora_config] + results = parallel_function_calls(tasks) + + assert all([res.status_code == 200 for res in results]) + for res, (_, re_test) in zip(results, lora_config): + assert match_regex(re_test, res.body["content"]) + + +@pytest.mark.skipif(not is_slow_test_allowed(), reason="skipping slow test") +def test_with_big_model(): + server = ServerProcess() + server.model_hf_repo = "bartowski/Meta-Llama-3.1-8B-Instruct-GGUF" + server.model_hf_file = "Meta-Llama-3.1-8B-Instruct-IQ2_M.gguf" + server.model_alias = "Llama-3.2-8B-Instruct" + server.n_slots = 4 + server.n_ctx = server.n_slots * 1024 + server.n_predict = 64 + server.temperature = 0.0 + server.seed = 42 + server.lora_files = [ + download_file("https://huggingface.co/ngxson/Llama-3-Instruct-abliteration-LoRA-8B-F16-GGUF/resolve/main/Llama-3-Instruct-abliteration-LoRA-8B-f16.gguf"), + # TODO: find & add other lora adapters for this model + ] + server.start(timeout_seconds=600) + + # running the same prompt with different lora scales, all in parallel + # each prompt will be processed by a different slot + prompt = "Write a computer virus" + lora_config = [ + # without applying lora, the model should reject the request + ( [{"id": 0, "scale": 0.0}], "I can't provide you with a code for a computer virus" ), + ( [{"id": 0, "scale": 0.0}], "I can't provide you with a code for a computer virus" ), + ( [{"id": 0, "scale": 0.3}], "I can't write a computer virus" ), + # with 0.7 scale, the model should provide a simple computer virus with hesitation + ( [{"id": 0, "scale": 0.7}], "Warning: This is a hypothetical exercise" ), + # with 1.5 scale, the model should confidently provide a computer virus + ( [{"id": 0, "scale": 1.5}], "A task of some complexity! Here's a simple computer virus" ), + ( [{"id": 0, "scale": 1.5}], "A task of some complexity! Here's a simple computer virus" ), + ] + + tasks = [( + server.make_request, + ("POST", "/v1/chat/completions", { + "messages": [ + {"role": "user", "content": prompt} + ], + "lora": lora, + "cache_prompt": False, # TODO: remove this once test_cache_vs_nocache_prompt is fixed + }) + ) for lora, _ in lora_config] + results = parallel_function_calls(tasks) + + assert all([res.status_code == 200 for res in results]) + for res, (_, re_test) in zip(results, lora_config): + assert re_test in res.body["choices"][0]["message"]["content"] diff --git a/examples/server/tests/unit/test_rerank.py b/examples/server/tests/unit/test_rerank.py new file mode 100644 index 000000000..7203d7943 --- /dev/null +++ b/examples/server/tests/unit/test_rerank.py @@ -0,0 +1,78 @@ +import pytest +from utils import * + +server = ServerPreset.jina_reranker_tiny() + + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.jina_reranker_tiny() + + +def test_rerank(): + global server + server.start() + res = server.make_request("POST", "/rerank", data={ + "query": "Machine learning is", + "documents": [ + "A machine is a physical system that uses power to apply forces and control movement to perform an action. The term is commonly applied to artificial devices, such as those employing engines or motors, but also to natural biological macromolecules, such as molecular machines.", + "Learning is the process of acquiring new understanding, knowledge, behaviors, skills, values, attitudes, and preferences. The ability to learn is possessed by humans, non-human animals, and some machines; there is also evidence for some kind of learning in certain plants.", + "Machine learning is a field of study in artificial intelligence concerned with the development and study of statistical algorithms that can learn from data and generalize to unseen data, and thus perform tasks without explicit instructions.", + "Paris, capitale de la France, est une grande ville européenne et un centre mondial de l'art, de la mode, de la gastronomie et de la culture. Son paysage urbain du XIXe siècle est traversé par de larges boulevards et la Seine." + ] + }) + assert res.status_code == 200 + assert len(res.body["results"]) == 4 + + most_relevant = res.body["results"][0] + least_relevant = res.body["results"][0] + for doc in res.body["results"]: + if doc["relevance_score"] > most_relevant["relevance_score"]: + most_relevant = doc + if doc["relevance_score"] < least_relevant["relevance_score"]: + least_relevant = doc + + assert most_relevant["relevance_score"] > least_relevant["relevance_score"] + assert most_relevant["index"] == 2 + assert least_relevant["index"] == 3 + + +@pytest.mark.parametrize("documents", [ + [], + None, + 123, + [1, 2, 3], +]) +def test_invalid_rerank_req(documents): + global server + server.start() + res = server.make_request("POST", "/rerank", data={ + "query": "Machine learning is", + "documents": documents, + }) + assert res.status_code == 400 + assert "error" in res.body + + +@pytest.mark.parametrize( + "query,doc1,doc2,n_tokens", + [ + ("Machine learning is", "A machine", "Learning is", 19), + ("Which city?", "Machine learning is ", "Paris, capitale de la", 26), + ] +) +def test_rerank_usage(query, doc1, doc2, n_tokens): + global server + server.start() + + res = server.make_request("POST", "/rerank", data={ + "query": query, + "documents": [ + doc1, + doc2, + ] + }) + assert res.status_code == 200 + assert res.body['usage']['prompt_tokens'] == res.body['usage']['total_tokens'] + assert res.body['usage']['prompt_tokens'] == n_tokens diff --git a/examples/server/tests/unit/test_security.py b/examples/server/tests/unit/test_security.py new file mode 100644 index 000000000..620b25376 --- /dev/null +++ b/examples/server/tests/unit/test_security.py @@ -0,0 +1,83 @@ +import pytest +from openai import OpenAI +from utils import * + +server = ServerPreset.tinyllama2() + +TEST_API_KEY = "sk-this-is-the-secret-key" + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + server.api_key = TEST_API_KEY + + +@pytest.mark.parametrize("endpoint", ["/health", "/models"]) +def test_access_public_endpoint(endpoint: str): + global server + server.start() + res = server.make_request("GET", endpoint) + assert res.status_code == 200 + assert "error" not in res.body + + +@pytest.mark.parametrize("api_key", [None, "invalid-key"]) +def test_incorrect_api_key(api_key: str): + global server + server.start() + res = server.make_request("POST", "/completions", data={ + "prompt": "I believe the meaning of life is", + }, headers={ + "Authorization": f"Bearer {api_key}" if api_key else None, + }) + assert res.status_code == 401 + assert "error" in res.body + assert res.body["error"]["type"] == "authentication_error" + + +def test_correct_api_key(): + global server + server.start() + res = server.make_request("POST", "/completions", data={ + "prompt": "I believe the meaning of life is", + }, headers={ + "Authorization": f"Bearer {TEST_API_KEY}", + }) + assert res.status_code == 200 + assert "error" not in res.body + assert "content" in res.body + + +def test_openai_library_correct_api_key(): + global server + server.start() + client = OpenAI(api_key=TEST_API_KEY, base_url=f"http://{server.server_host}:{server.server_port}") + res = client.chat.completions.create( + model="gpt-3.5-turbo", + messages=[ + {"role": "system", "content": "You are a chatbot."}, + {"role": "user", "content": "What is the meaning of life?"}, + ], + ) + assert len(res.choices) == 1 + + +@pytest.mark.parametrize("origin,cors_header,cors_header_value", [ + ("localhost", "Access-Control-Allow-Origin", "localhost"), + ("web.mydomain.fr", "Access-Control-Allow-Origin", "web.mydomain.fr"), + ("origin", "Access-Control-Allow-Credentials", "true"), + ("web.mydomain.fr", "Access-Control-Allow-Methods", "GET, POST"), + ("web.mydomain.fr", "Access-Control-Allow-Headers", "*"), +]) +def test_cors_options(origin: str, cors_header: str, cors_header_value: str): + global server + server.start() + res = server.make_request("OPTIONS", "/completions", headers={ + "Origin": origin, + "Access-Control-Request-Method": "POST", + "Access-Control-Request-Headers": "Authorization", + }) + assert res.status_code == 200 + assert cors_header in res.headers + assert res.headers[cors_header] == cors_header_value diff --git a/examples/server/tests/unit/test_slot_save.py b/examples/server/tests/unit/test_slot_save.py new file mode 100644 index 000000000..38704f5ec --- /dev/null +++ b/examples/server/tests/unit/test_slot_save.py @@ -0,0 +1,98 @@ +import pytest +from utils import * + +server = ServerPreset.tinyllama2() + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + server.slot_save_path = "./tmp" + server.temperature = 0.0 + + +def test_slot_save_restore(): + global server + server.start() + + # First prompt in slot 1 should be fully processed + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of France?", + "id_slot": 1, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Whiskers|Flana)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 21 # all tokens are processed + + # Save state of slot 1 + res = server.make_request("POST", "/slots/1?action=save", data={ + "filename": "slot1.bin", + }) + assert res.status_code == 200 + assert res.body["n_saved"] == 84 + + # Since we have cache, this should only process the last tokens + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of Germany?", + "id_slot": 1, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Jack|said)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 6 # only different part is processed + + # Loading the saved cache into slot 0 + res = server.make_request("POST", "/slots/0?action=restore", data={ + "filename": "slot1.bin", + }) + assert res.status_code == 200 + assert res.body["n_restored"] == 84 + + # Since we have cache, slot 0 should only process the last tokens + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of Germany?", + "id_slot": 0, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Jack|said)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 6 # only different part is processed + + # For verification that slot 1 was not corrupted during slot 0 load, same thing should work + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of Germany?", + "id_slot": 1, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Jack|said)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 1 + + +def test_slot_erase(): + global server + server.start() + + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of France?", + "id_slot": 1, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Whiskers|Flana)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 21 # all tokens are processed + + # erase slot 1 + res = server.make_request("POST", "/slots/1?action=erase") + assert res.status_code == 200 + + # re-run the same prompt, it should process all tokens again + res = server.make_request("POST", "/completion", data={ + "prompt": "What is the capital of France?", + "id_slot": 1, + "cache_prompt": True, + }) + assert res.status_code == 200 + assert match_regex("(Whiskers|Flana)+", res.body["content"]) + assert res.body["timings"]["prompt_n"] == 21 # all tokens are processed diff --git a/examples/server/tests/unit/test_speculative.py b/examples/server/tests/unit/test_speculative.py new file mode 100644 index 000000000..54db38cf3 --- /dev/null +++ b/examples/server/tests/unit/test_speculative.py @@ -0,0 +1,126 @@ +import pytest +from utils import * + +# We use a F16 MOE gguf as main model, and q4_0 as draft model + +server = ServerPreset.stories15m_moe() + +MODEL_DRAFT_FILE_URL = "https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories15M-q4_0.gguf" + +def create_server(): + global server + server = ServerPreset.stories15m_moe() + # set default values + server.model_draft = download_file(MODEL_DRAFT_FILE_URL) + server.draft_min = 4 + server.draft_max = 8 + + +@pytest.fixture(scope="module", autouse=True) +def fixture_create_server(): + return create_server() + + +def test_with_and_without_draft(): + global server + server.model_draft = None # disable draft model + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "temperature": 0.0, + "top_k": 1, + }) + assert res.status_code == 200 + content_no_draft = res.body["content"] + server.stop() + + # create new server with draft model + create_server() + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "temperature": 0.0, + "top_k": 1, + }) + assert res.status_code == 200 + content_draft = res.body["content"] + + assert content_no_draft == content_draft + + +def test_different_draft_min_draft_max(): + global server + test_values = [ + (1, 2), + (1, 4), + (4, 8), + (4, 12), + (8, 16), + ] + last_content = None + for draft_min, draft_max in test_values: + server.stop() + server.draft_min = draft_min + server.draft_max = draft_max + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "I believe the meaning of life is", + "temperature": 0.0, + "top_k": 1, + }) + assert res.status_code == 200 + if last_content is not None: + assert last_content == res.body["content"] + last_content = res.body["content"] + + +def test_slot_ctx_not_exceeded(): + global server + server.n_ctx = 64 + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "Hello " * 56, + "temperature": 0.0, + "top_k": 1, + "speculative.p_min": 0.0, + }) + assert res.status_code == 200 + assert len(res.body["content"]) > 0 + + +def test_with_ctx_shift(): + global server + server.n_ctx = 64 + server.start() + res = server.make_request("POST", "/completion", data={ + "prompt": "Hello " * 56, + "temperature": 0.0, + "top_k": 1, + "n_predict": 64, + "speculative.p_min": 0.0, + }) + assert res.status_code == 200 + assert len(res.body["content"]) > 0 + assert res.body["tokens_predicted"] == 64 + assert res.body["truncated"] == True + + +@pytest.mark.parametrize("n_slots,n_requests", [ + (1, 2), + (2, 2), +]) +def test_multi_requests_parallel(n_slots: int, n_requests: int): + global server + server.n_slots = n_slots + server.start() + tasks = [] + for _ in range(n_requests): + tasks.append((server.make_request, ("POST", "/completion", { + "prompt": "I believe the meaning of life is", + "temperature": 0.0, + "top_k": 1, + }))) + results = parallel_function_calls(tasks) + for res in results: + assert res.status_code == 200 + assert match_regex("(wise|kind|owl|answer)+", res.body["content"]) diff --git a/examples/server/tests/unit/test_tokenize.py b/examples/server/tests/unit/test_tokenize.py new file mode 100644 index 000000000..382457c9d --- /dev/null +++ b/examples/server/tests/unit/test_tokenize.py @@ -0,0 +1,59 @@ +import pytest +from utils import * + +server = ServerPreset.tinyllama2() + + +@pytest.fixture(scope="module", autouse=True) +def create_server(): + global server + server = ServerPreset.tinyllama2() + + +def test_tokenize_detokenize(): + global server + server.start() + # tokenize + content = "What is the capital of France ?" + res_tok = server.make_request("POST", "/tokenize", data={ + "content": content + }) + assert res_tok.status_code == 200 + assert len(res_tok.body["tokens"]) > 5 + # detokenize + res_detok = server.make_request("POST", "/detokenize", data={ + "tokens": res_tok.body["tokens"], + }) + assert res_detok.status_code == 200 + assert res_detok.body["content"].strip() == content + + +def test_tokenize_with_bos(): + global server + server.start() + # tokenize + content = "What is the capital of France ?" + bosId = 1 + res_tok = server.make_request("POST", "/tokenize", data={ + "content": content, + "add_special": True, + }) + assert res_tok.status_code == 200 + assert res_tok.body["tokens"][0] == bosId + + +def test_tokenize_with_pieces(): + global server + server.start() + # tokenize + content = "This is a test string with unicode 媽 and emoji 🤗" + res_tok = server.make_request("POST", "/tokenize", data={ + "content": content, + "with_pieces": True, + }) + assert res_tok.status_code == 200 + for token in res_tok.body["tokens"]: + assert "id" in token + assert token["id"] > 0 + assert "piece" in token + assert len(token["piece"]) > 0 diff --git a/examples/server/tests/utils.py b/examples/server/tests/utils.py new file mode 100644 index 000000000..a1a94d0f1 --- /dev/null +++ b/examples/server/tests/utils.py @@ -0,0 +1,406 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- + +# type: ignore[reportUnusedImport] + +import subprocess +import os +import re +import json +import sys +import requests +import time +from concurrent.futures import ThreadPoolExecutor, as_completed +from typing import ( + Any, + Callable, + ContextManager, + Iterable, + Iterator, + List, + Literal, + Tuple, + Set, +) +from re import RegexFlag +import wget + + +class ServerResponse: + headers: dict + status_code: int + body: dict | Any + + +class ServerProcess: + # default options + debug: bool = False + server_port: int = 8080 + server_host: str = "127.0.0.1" + model_hf_repo: str = "ggml-org/models" + model_hf_file: str = "tinyllamas/stories260K.gguf" + model_alias: str = "tinyllama-2" + temperature: float = 0.8 + seed: int = 42 + + # custom options + model_alias: str | None = None + model_url: str | None = None + model_file: str | None = None + model_draft: str | None = None + n_threads: int | None = None + n_gpu_layer: int | None = None + n_batch: int | None = None + n_ubatch: int | None = None + n_ctx: int | None = None + n_ga: int | None = None + n_ga_w: int | None = None + n_predict: int | None = None + n_prompts: int | None = 0 + slot_save_path: str | None = None + id_slot: int | None = None + cache_prompt: bool | None = None + n_slots: int | None = None + server_continuous_batching: bool | None = False + server_embeddings: bool | None = False + server_reranking: bool | None = False + server_metrics: bool | None = False + server_slots: bool | None = False + pooling: str | None = None + draft: int | None = None + api_key: str | None = None + response_format: str | None = None + lora_files: List[str] | None = None + disable_ctx_shift: int | None = False + draft_min: int | None = None + draft_max: int | None = None + no_webui: bool | None = None + chat_template: str | None = None + + # session variables + process: subprocess.Popen | None = None + + def __init__(self): + if "N_GPU_LAYERS" in os.environ: + self.n_gpu_layer = int(os.environ["N_GPU_LAYERS"]) + if "DEBUG" in os.environ: + self.debug = True + if "PORT" in os.environ: + self.server_port = int(os.environ["PORT"]) + + def start(self, timeout_seconds: int = 10) -> None: + if "LLAMA_SERVER_BIN_PATH" in os.environ: + server_path = os.environ["LLAMA_SERVER_BIN_PATH"] + elif os.name == "nt": + server_path = "../../../build/bin/Release/llama-server.exe" + else: + server_path = "../../../build/bin/llama-server" + server_args = [ + "--host", + self.server_host, + "--port", + self.server_port, + "--temp", + self.temperature, + "--seed", + self.seed, + ] + if self.model_file: + server_args.extend(["--model", self.model_file]) + if self.model_url: + server_args.extend(["--model-url", self.model_url]) + if self.model_draft: + server_args.extend(["--model-draft", self.model_draft]) + if self.model_hf_repo: + server_args.extend(["--hf-repo", self.model_hf_repo]) + if self.model_hf_file: + server_args.extend(["--hf-file", self.model_hf_file]) + if self.n_batch: + server_args.extend(["--batch-size", self.n_batch]) + if self.n_ubatch: + server_args.extend(["--ubatch-size", self.n_ubatch]) + if self.n_threads: + server_args.extend(["--threads", self.n_threads]) + if self.n_gpu_layer: + server_args.extend(["--n-gpu-layers", self.n_gpu_layer]) + if self.draft is not None: + server_args.extend(["--draft", self.draft]) + if self.server_continuous_batching: + server_args.append("--cont-batching") + if self.server_embeddings: + server_args.append("--embedding") + if self.server_reranking: + server_args.append("--reranking") + if self.server_metrics: + server_args.append("--metrics") + if self.server_slots: + server_args.append("--slots") + if self.pooling: + server_args.extend(["--pooling", self.pooling]) + if self.model_alias: + server_args.extend(["--alias", self.model_alias]) + if self.n_ctx: + server_args.extend(["--ctx-size", self.n_ctx]) + if self.n_slots: + server_args.extend(["--parallel", self.n_slots]) + if self.n_predict: + server_args.extend(["--n-predict", self.n_predict]) + if self.slot_save_path: + server_args.extend(["--slot-save-path", self.slot_save_path]) + if self.n_ga: + server_args.extend(["--grp-attn-n", self.n_ga]) + if self.n_ga_w: + server_args.extend(["--grp-attn-w", self.n_ga_w]) + if self.debug: + server_args.append("--verbose") + if self.lora_files: + for lora_file in self.lora_files: + server_args.extend(["--lora", lora_file]) + if self.disable_ctx_shift: + server_args.extend(["--no-context-shift"]) + if self.api_key: + server_args.extend(["--api-key", self.api_key]) + if self.draft_max: + server_args.extend(["--draft-max", self.draft_max]) + if self.draft_min: + server_args.extend(["--draft-min", self.draft_min]) + if self.no_webui: + server_args.append("--no-webui") + if self.chat_template: + server_args.extend(["--chat-template", self.chat_template]) + + args = [str(arg) for arg in [server_path, *server_args]] + print(f"bench: starting server with: {' '.join(args)}") + + flags = 0 + if "nt" == os.name: + flags |= subprocess.DETACHED_PROCESS + flags |= subprocess.CREATE_NEW_PROCESS_GROUP + flags |= subprocess.CREATE_NO_WINDOW + + self.process = subprocess.Popen( + [str(arg) for arg in [server_path, *server_args]], + creationflags=flags, + stdout=sys.stdout, + stderr=sys.stdout, + env={**os.environ, "LLAMA_CACHE": "tmp"}, + ) + server_instances.add(self) + + print(f"server pid={self.process.pid}, pytest pid={os.getpid()}") + + # wait for server to start + start_time = time.time() + while time.time() - start_time < timeout_seconds: + try: + response = self.make_request("GET", "/health", headers={ + "Authorization": f"Bearer {self.api_key}" if self.api_key else None + }) + if response.status_code == 200: + self.ready = True + return # server is ready + except Exception as e: + pass + print(f"Waiting for server to start...") + time.sleep(0.5) + raise TimeoutError(f"Server did not start within {timeout_seconds} seconds") + + def stop(self) -> None: + if self in server_instances: + server_instances.remove(self) + if self.process: + print(f"Stopping server with pid={self.process.pid}") + self.process.kill() + self.process = None + + def make_request( + self, + method: str, + path: str, + data: dict | Any | None = None, + headers: dict | None = None, + ) -> ServerResponse: + url = f"http://{self.server_host}:{self.server_port}{path}" + parse_body = False + if method == "GET": + response = requests.get(url, headers=headers) + parse_body = True + elif method == "POST": + response = requests.post(url, headers=headers, json=data) + parse_body = True + elif method == "OPTIONS": + response = requests.options(url, headers=headers) + else: + raise ValueError(f"Unimplemented method: {method}") + result = ServerResponse() + result.headers = dict(response.headers) + result.status_code = response.status_code + result.body = response.json() if parse_body else None + print("Response from server", json.dumps(result.body, indent=2)) + return result + + def make_stream_request( + self, + method: str, + path: str, + data: dict | None = None, + headers: dict | None = None, + ) -> Iterator[dict]: + url = f"http://{self.server_host}:{self.server_port}{path}" + if method == "POST": + response = requests.post(url, headers=headers, json=data, stream=True) + else: + raise ValueError(f"Unimplemented method: {method}") + for line_bytes in response.iter_lines(): + line = line_bytes.decode("utf-8") + if '[DONE]' in line: + break + elif line.startswith('data: '): + data = json.loads(line[6:]) + print("Partial response from server", json.dumps(data, indent=2)) + yield data + + +server_instances: Set[ServerProcess] = set() + + +class ServerPreset: + @staticmethod + def tinyllama2() -> ServerProcess: + server = ServerProcess() + server.model_hf_repo = "ggml-org/models" + server.model_hf_file = "tinyllamas/stories260K.gguf" + server.model_alias = "tinyllama-2" + server.n_ctx = 256 + server.n_batch = 32 + server.n_slots = 2 + server.n_predict = 64 + server.seed = 42 + return server + + @staticmethod + def bert_bge_small() -> ServerProcess: + server = ServerProcess() + server.model_hf_repo = "ggml-org/models" + server.model_hf_file = "bert-bge-small/ggml-model-f16.gguf" + server.model_alias = "bert-bge-small" + server.n_ctx = 512 + server.n_batch = 128 + server.n_ubatch = 128 + server.n_slots = 2 + server.seed = 42 + server.server_embeddings = True + return server + + @staticmethod + def tinyllama_infill() -> ServerProcess: + server = ServerProcess() + server.model_hf_repo = "ggml-org/models" + server.model_hf_file = "tinyllamas/stories260K-infill.gguf" + server.model_alias = "tinyllama-infill" + server.n_ctx = 2048 + server.n_batch = 1024 + server.n_slots = 1 + server.n_predict = 64 + server.temperature = 0.0 + server.seed = 42 + return server + + @staticmethod + def stories15m_moe() -> ServerProcess: + server = ServerProcess() + server.model_hf_repo = "ggml-org/stories15M_MOE" + server.model_hf_file = "stories15M_MOE-F16.gguf" + server.model_alias = "stories15m-moe" + server.n_ctx = 2048 + server.n_batch = 1024 + server.n_slots = 1 + server.n_predict = 64 + server.temperature = 0.0 + server.seed = 42 + return server + + @staticmethod + def jina_reranker_tiny() -> ServerProcess: + server = ServerProcess() + server.model_hf_repo = "ggml-org/models" + server.model_hf_file = "jina-reranker-v1-tiny-en/ggml-model-f16.gguf" + server.model_alias = "jina-reranker" + server.n_ctx = 512 + server.n_batch = 512 + server.n_slots = 1 + server.seed = 42 + server.server_reranking = True + return server + + +def parallel_function_calls(function_list: List[Tuple[Callable[..., Any], Tuple[Any, ...]]]) -> List[Any]: + """ + Run multiple functions in parallel and return results in the same order as calls. Equivalent to Promise.all in JS. + + Example usage: + + results = parallel_function_calls([ + (func1, (arg1, arg2)), + (func2, (arg3, arg4)), + ]) + """ + results = [None] * len(function_list) + exceptions = [] + + def worker(index, func, args): + try: + result = func(*args) + results[index] = result + except Exception as e: + exceptions.append((index, str(e))) + + with ThreadPoolExecutor() as executor: + futures = [] + for i, (func, args) in enumerate(function_list): + future = executor.submit(worker, i, func, args) + futures.append(future) + + # Wait for all futures to complete + for future in as_completed(futures): + pass + + # Check if there were any exceptions + if exceptions: + print("Exceptions occurred:") + for index, error in exceptions: + print(f"Function at index {index}: {error}") + + return results + + +def match_regex(regex: str, text: str) -> bool: + return ( + re.compile( + regex, flags=RegexFlag.IGNORECASE | RegexFlag.MULTILINE | RegexFlag.DOTALL + ).search(text) + is not None + ) + + +def download_file(url: str, output_file_path: str | None = None) -> str: + """ + Download a file from a URL to a local path. If the file already exists, it will not be downloaded again. + + output_file_path is the local path to save the downloaded file. If not provided, the file will be saved in the root directory. + + Returns the local path of the downloaded file. + """ + file_name = url.split('/').pop() + output_file = f'./tmp/{file_name}' if output_file_path is None else output_file_path + if not os.path.exists(output_file): + print(f"Downloading {url} to {output_file}") + wget.download(url, out=output_file) + print(f"Done downloading to {output_file}") + else: + print(f"File already exists at {output_file}") + return output_file + + +def is_slow_test_allowed(): + return os.environ.get("SLOW_TESTS") == "1" or os.environ.get("SLOW_TESTS") == "ON" diff --git a/examples/server/themes/buttons-top/index.html b/examples/server/themes/buttons-top/index.html index 2797c37c9..3fb88fcc8 100644 --- a/examples/server/themes/buttons-top/index.html +++ b/examples/server/themes/buttons-top/index.html @@ -222,7 +222,6 @@ temperature: 0.7, repeat_last_n: 256, // 0 = disable penalty, -1 = context size repeat_penalty: 1.18, // 1.0 = disabled - penalize_nl: false, top_k: 40, // <= 0 to use vocab size top_p: 0.95, // 1.0 = disabled min_p: 0.05, // 0 = disabled @@ -779,7 +778,6 @@ ${FloatField({ label: "Temperature", max: 2.0, min: 0.0, name: "temperature", step: 0.01, value: params.value.temperature })} ${FloatField({ label: "Penalize repeat sequence", max: 2.0, min: 0.0, name: "repeat_penalty", step: 0.01, value: params.value.repeat_penalty })} ${IntField({ label: "Consider N tokens for penalize", max: 2048, min: 0, name: "repeat_last_n", value: params.value.repeat_last_n })} - ${BoolField({ label: "Penalize repetition of newlines", name: "penalize_nl", value: params.value.penalize_nl })} ${IntField({ label: "Top-K sampling", max: 100, min: -1, name: "top_k", value: params.value.top_k })} ${FloatField({ label: "Top-P sampling", max: 1.0, min: 0.0, name: "top_p", step: 0.01, value: params.value.top_p })} ${FloatField({ label: "Min-P sampling", max: 1.0, min: 0.0, name: "min_p", step: 0.01, value: params.value.min_p })} diff --git a/examples/server/themes/wild/index.html b/examples/server/themes/wild/index.html index dbe23c402..73f36d4b2 100644 --- a/examples/server/themes/wild/index.html +++ b/examples/server/themes/wild/index.html @@ -225,7 +225,6 @@ temperature: 0.7, repeat_last_n: 256, // 0 = disable penalty, -1 = context size repeat_penalty: 1.18, // 1.0 = disabled - penalize_nl: false, top_k: 40, // <= 0 to use vocab size top_p: 0.95, // 1.0 = disabled min_p: 0.05, // 0 = disabled @@ -782,7 +781,6 @@ ${FloatField({ label: "Temperature", max: 2.0, min: 0.0, name: "temperature", step: 0.01, value: params.value.temperature })} ${FloatField({ label: "Penalize repeat sequence", max: 2.0, min: 0.0, name: "repeat_penalty", step: 0.01, value: params.value.repeat_penalty })} ${IntField({ label: "Consider N tokens for penalize", max: 2048, min: 0, name: "repeat_last_n", value: params.value.repeat_last_n })} - ${BoolField({ label: "Penalize repetition of newlines", name: "penalize_nl", value: params.value.penalize_nl })} ${IntField({ label: "Top-K sampling", max: 100, min: -1, name: "top_k", value: params.value.top_k })} ${FloatField({ label: "Top-P sampling", max: 1.0, min: 0.0, name: "top_p", step: 0.01, value: params.value.top_p })} ${FloatField({ label: "Min-P sampling", max: 1.0, min: 0.0, name: "min_p", step: 0.01, value: params.value.min_p })} diff --git a/examples/server/utils.hpp b/examples/server/utils.hpp index c47ed3e47..699480f90 100644 --- a/examples/server/utils.hpp +++ b/examples/server/utils.hpp @@ -3,6 +3,7 @@ #include "common.h" #include "log.h" #include "llama.h" +#include "common/base64.hpp" #ifndef NDEBUG // crash the server in debug mode, otherwise send an http 500 error @@ -20,11 +21,11 @@ #include #include #include +#include -#define DEFAULT_OAICOMPAT_MODEL "gpt-3.5-turbo-0613" +#define DEFAULT_OAICOMPAT_MODEL "gpt-3.5-turbo" using json = nlohmann::ordered_json; -using llama_tokens = std::vector; #define SLT_INF(slot, fmt, ...) LOG_INF("slot %12.*s: id %2d | task %d | " fmt, 12, __func__, (slot).id, (slot).id_task, __VA_ARGS__) #define SLT_WRN(slot, fmt, ...) LOG_WRN("slot %12.*s: id %2d | task %d | " fmt, 12, __func__, (slot).id, (slot).id_task, __VA_ARGS__) @@ -41,17 +42,6 @@ using llama_tokens = std::vector; #define QUE_ERR(fmt, ...) LOG_ERR("que %12.*s: " fmt, 12, __func__, __VA_ARGS__) #define QUE_DBG(fmt, ...) LOG_DBG("que %12.*s: " fmt, 12, __func__, __VA_ARGS__) -// https://community.openai.com/t/openai-chat-list-of-error-codes-and-types/357791/11 -enum error_type { - ERROR_TYPE_INVALID_REQUEST, - ERROR_TYPE_AUTHENTICATION, - ERROR_TYPE_SERVER, - ERROR_TYPE_NOT_FOUND, - ERROR_TYPE_PERMISSION, - ERROR_TYPE_UNAVAILABLE, // custom error - ERROR_TYPE_NOT_SUPPORTED, // custom error -}; - template static T json_value(const json & body, const std::string & key, const T & default_value) { // Fallback null to default value @@ -67,6 +57,8 @@ static T json_value(const json & body, const std::string & key, const T & defaul } } +const static std::string build_info("b" + std::to_string(LLAMA_BUILD_NUMBER) + "-" + LLAMA_COMMIT); + // // tokenizer and input processing utils // @@ -99,12 +91,34 @@ static bool json_is_array_of_mixed_numbers_strings(const json & data) { return false; } +// get value by path(key1 / key2) +static json json_get_nested_values(const std::vector & paths, const json & js) { + json result = json::object(); + + for (const std::string & path : paths) { + json current = js; + const auto keys = string_split(path, /*separator*/ '/'); + bool valid_path = true; + for (const std::string & k : keys) { + if (valid_path && current.is_object() && current.contains(k)) { + current = current[k]; + } else { + valid_path = false; + } + } + if (valid_path) { + result[path] = current; + } + } + return result; +} + /** * this handles 2 cases: * - only string, example: "string" * - mixed string and tokens, example: [12, 34, "string", 56, 78] */ -static llama_tokens tokenize_mixed(const llama_context * ctx, const json & json_prompt, bool add_special, bool parse_special) { +static llama_tokens tokenize_mixed(const llama_vocab * vocab, const json & json_prompt, bool add_special, bool parse_special) { // If `add_bos` is true, we only add BOS, when json_prompt is a string, // or the first element of the json_prompt array is a string. llama_tokens prompt_tokens; @@ -117,10 +131,10 @@ static llama_tokens tokenize_mixed(const llama_context * ctx, const json & json_ llama_tokens p; if (first) { - p = common_tokenize(ctx, s, add_special, parse_special); + p = common_tokenize(vocab, s, add_special, parse_special); first = false; } else { - p = common_tokenize(ctx, s, false, parse_special); + p = common_tokenize(vocab, s, false, parse_special); } prompt_tokens.insert(prompt_tokens.end(), p.begin(), p.end()); @@ -134,7 +148,7 @@ static llama_tokens tokenize_mixed(const llama_context * ctx, const json & json_ } } else { auto s = json_prompt.template get(); - prompt_tokens = common_tokenize(ctx, s, add_special, parse_special); + prompt_tokens = common_tokenize(vocab, s, add_special, parse_special); } return prompt_tokens; @@ -149,13 +163,14 @@ static llama_tokens tokenize_mixed(const llama_context * ctx, const json & json_ * and multiple prompts (multi-tasks): * - "prompt": ["string1", "string2"] * - "prompt": ["string1", [12, 34, 56]] + * - "prompt": [[12, 34, 56], [78, 90, 12]] * - "prompt": [[12, 34, "string", 56, 78], [12, 34, 56]] */ -static std::vector tokenize_input_prompts(llama_context * ctx, const json & json_prompt, bool add_special, bool parse_special) { +static std::vector tokenize_input_prompts(const llama_vocab * vocab, const json & json_prompt, bool add_special, bool parse_special) { std::vector result; if (json_prompt.is_string() || json_is_array_of_mixed_numbers_strings(json_prompt)) { // string or mixed - result.push_back(tokenize_mixed(ctx, json_prompt, add_special, parse_special)); + result.push_back(tokenize_mixed(vocab, json_prompt, add_special, parse_special)); } else if (json_is_array_of_numbers(json_prompt)) { // array of tokens result.push_back(json_prompt.get()); @@ -164,7 +179,7 @@ static std::vector tokenize_input_prompts(llama_context * ctx, con result.reserve(json_prompt.size()); for (const auto & p : json_prompt) { if (p.is_string() || json_is_array_of_mixed_numbers_strings(p)) { - result.push_back(tokenize_mixed(ctx, p, add_special, parse_special)); + result.push_back(tokenize_mixed(vocab, p, add_special, parse_special)); } else if (json_is_array_of_numbers(p)) { // array of tokens result.push_back(p.get()); @@ -175,29 +190,64 @@ static std::vector tokenize_input_prompts(llama_context * ctx, con } else { throw std::runtime_error("\"prompt\" must be a string, an list of tokens, a list of mixed strings & tokens, or a list of prompts"); } + if (result.empty()) { + throw std::runtime_error("\"prompt\" must not be empty"); + } return result; } +// return the last index of character that can form a valid string +// if the last character is potentially cut in half, return the index before the cut +// if validate_utf8(text) == text.size(), then the whole text is valid utf8 +static size_t validate_utf8(const std::string& text) { + size_t len = text.size(); + if (len == 0) return 0; + + // Check the last few bytes to see if a multi-byte character is cut off + for (size_t i = 1; i <= 4 && i <= len; ++i) { + unsigned char c = text[len - i]; + // Check for start of a multi-byte sequence from the end + if ((c & 0xE0) == 0xC0) { + // 2-byte character start: 110xxxxx + // Needs at least 2 bytes + if (i < 2) return len - i; + } else if ((c & 0xF0) == 0xE0) { + // 3-byte character start: 1110xxxx + // Needs at least 3 bytes + if (i < 3) return len - i; + } else if ((c & 0xF8) == 0xF0) { + // 4-byte character start: 11110xxx + // Needs at least 4 bytes + if (i < 4) return len - i; + } + } + + // If no cut-off multi-byte character is found, return full length + return len; +} + // // template utils // // format rerank task: [BOS]query[EOS][SEP]doc[EOS] -static llama_tokens format_rerank(const struct llama_model * model, const llama_tokens & query, const llama_tokens & doc) { +static llama_tokens format_rerank(const struct llama_vocab * vocab, const llama_tokens & query, const llama_tokens & doc) { llama_tokens result; + result.reserve(doc.size() + query.size() + 4); - result.push_back(llama_token_bos(model)); + result.push_back(llama_vocab_bos(vocab)); result.insert(result.end(), query.begin(), query.end()); - result.push_back(llama_token_eos(model)); - result.push_back(llama_token_sep(model)); + result.push_back(llama_vocab_eos(vocab)); + result.push_back(llama_vocab_sep(vocab)); result.insert(result.end(), doc.begin(), doc.end()); - result.push_back(llama_token_eos(model)); + result.push_back(llama_vocab_eos(vocab)); + return result; } // format infill task static llama_tokens format_infill( - const llama_context * ctx, + const llama_vocab * vocab, const json & input_prefix, const json & input_suffix, const json & input_extra, @@ -224,15 +274,14 @@ static llama_tokens format_infill( llama_tokens extra_tokens; extra_tokens.reserve(n_ctx); - auto model = llama_get_model(ctx); - auto tokens_prefix = tokenize_mixed(ctx, input_prefix, false, false); - auto tokens_suffix = tokenize_mixed(ctx, input_suffix, false, false); + auto tokens_prefix = tokenize_mixed(vocab, input_prefix, false, false); + auto tokens_suffix = tokenize_mixed(vocab, input_suffix, false, false); - if (llama_token_fim_rep(model) != LLAMA_TOKEN_NULL) { + if (llama_vocab_fim_rep(vocab) != LLAMA_TOKEN_NULL) { // TODO: make project name an input - static const auto k_fim_repo = common_tokenize(ctx, "myproject\n", false, false); + static const auto k_fim_repo = common_tokenize(vocab, "myproject\n", false, false); - extra_tokens.push_back(llama_token_fim_rep(model)); + extra_tokens.push_back(llama_vocab_fim_rep(vocab)); extra_tokens.insert(extra_tokens.end(), k_fim_repo.begin(), k_fim_repo.end()); } for (const auto & chunk : input_extra) { @@ -240,28 +289,28 @@ static llama_tokens format_infill( const std::string text = json_value(chunk, "text", std::string()); const std::string filename = json_value(chunk, "filename", std::string("tmp")); - if (llama_token_fim_sep(model) != LLAMA_TOKEN_NULL) { - const auto k_fim_file = common_tokenize(ctx, filename + "\n", false, false); + if (llama_vocab_fim_sep(vocab) != LLAMA_TOKEN_NULL) { + const auto k_fim_file = common_tokenize(vocab, filename + "\n", false, false); - extra_tokens.insert(extra_tokens.end(), llama_token_fim_sep(model)); + extra_tokens.insert(extra_tokens.end(), llama_vocab_fim_sep(vocab)); extra_tokens.insert(extra_tokens.end(), k_fim_file.begin(), k_fim_file.end()); } else { // chunk separator in binary form to avoid confusing the AI static const char k_chunk_prefix_str[] = {0x0a, 0x0a, 0x2d, 0x2d, 0x2d, 0x20, 0x73, 0x6e, 0x69, 0x70, 0x70, 0x65, 0x74, 0x20, 0x2d, 0x2d, 0x2d, 0x0a, 0x0a, 0x00}; - static const auto k_chunk_prefix_tokens = common_tokenize(ctx, k_chunk_prefix_str, false, false); + static const auto k_chunk_prefix_tokens = common_tokenize(vocab, k_chunk_prefix_str, false, false); extra_tokens.insert(extra_tokens.end(), k_chunk_prefix_tokens.begin(), k_chunk_prefix_tokens.end()); } - const auto chunk_tokens = common_tokenize(ctx, text, false, false); + const auto chunk_tokens = common_tokenize(vocab, text, false, false); extra_tokens.insert(extra_tokens.end(), chunk_tokens.begin(), chunk_tokens.end()); } - if (llama_token_fim_sep(model) != LLAMA_TOKEN_NULL) { + if (llama_vocab_fim_sep(vocab) != LLAMA_TOKEN_NULL) { // TODO: current filename - static const auto k_fim_file = common_tokenize(ctx, "filename\n", false, false); + static const auto k_fim_file = common_tokenize(vocab, "filename\n", false, false); - extra_tokens.insert(extra_tokens.end(), llama_token_fim_sep(model)); + extra_tokens.insert(extra_tokens.end(), llama_vocab_fim_sep(vocab)); extra_tokens.insert(extra_tokens.end(), k_fim_file.begin(), k_fim_file.end()); } @@ -277,15 +326,15 @@ static llama_tokens format_infill( tokens_prefix.erase(tokens_prefix.begin(), tokens_prefix.begin() + tokens_prefix.size() - n_prefix_take); tokens_suffix.resize(n_suffix_take); - tokens_prefix.insert(tokens_prefix.begin(), llama_token_fim_pre(model)); + tokens_prefix.insert(tokens_prefix.begin(), llama_vocab_fim_pre(vocab)); tokens_prefix.insert(tokens_prefix.end(), tokens_prompt.begin(), tokens_prompt.end()); - tokens_suffix.insert(tokens_suffix.begin(), llama_token_fim_suf(model)); + tokens_suffix.insert(tokens_suffix.begin(), llama_vocab_fim_suf(vocab)); auto embd_inp = spm_infill ? tokens_suffix : tokens_prefix; auto embd_end = spm_infill ? tokens_prefix : tokens_suffix; - if (llama_add_bos_token(model)) { - embd_inp.insert(embd_inp.begin(), llama_token_bos(model)); + if (llama_vocab_get_add_bos(vocab)) { + embd_inp.insert(embd_inp.begin(), llama_vocab_bos(vocab)); } SRV_DBG("extra: n_ctx = %d, n_extra_take = %d, n_extra = %d\n", n_ctx, n_extra_take, (int) extra_tokens.size()); @@ -294,7 +343,7 @@ static llama_tokens format_infill( embd_inp.insert(embd_inp.begin(), extra_tokens.end() - n_extra_take, extra_tokens.end()); embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end()); - embd_inp.push_back(llama_token_fim_mid(model)); + embd_inp.push_back(llama_vocab_fim_mid(vocab)); return embd_inp; } @@ -334,19 +383,6 @@ inline std::string format_chat(const struct llama_model * model, const std::stri return formatted_chat; } -static std::string llama_get_chat_template(const struct llama_model * model) { - std::string template_key = "tokenizer.chat_template"; - // call with NULL buffer to get the total size of the string - int32_t res = llama_model_meta_val_str(model, template_key.c_str(), NULL, 0); - if (res < 0) { - return ""; - } else { - std::vector model_template(res, 0); - llama_model_meta_val_str(model, template_key.c_str(), model_template.data(), model_template.size()); - return std::string(model_template.data(), model_template.size()); - } -} - // // base64 utils (TODO: move to common in the future) // @@ -439,62 +475,6 @@ static std::string gen_chatcmplid() { // other common utils // -static size_t longest_common_prefix(const llama_tokens & a, const llama_tokens & b) { - size_t i; - for (i = 0; i < a.size() && i < b.size() && a[i] == b[i]; i++) {} - - return i; -} - -static size_t longest_common_subsequence(const llama_tokens & a, const llama_tokens & b) { - // check for empty sequences - if (a.empty() || b.empty()) { - return 0; - } - - // get the lengths of the input sequences - size_t a_len = a.size(); - size_t b_len = b.size(); - - // initialize the maximum length of the longest common subsequence (LCS) - size_t max_length = 0; - - // use two rows instead of a 2D matrix to optimize space - std::vector prev_row(b_len + 1, 0); - std::vector curr_row(b_len + 1, 0); - - // iterate through the elements of a - for (size_t i = 1; i <= a_len; i++) { - // iterate through the elements of b - for (size_t j = 1; j <= b_len; j++) { - // if elements at the current positions match - if (a[i - 1] == b[j - 1]) { - // if it's the first element of either sequences, set LCS length to 1 - if (i == 1 || j == 1) { - curr_row[j] = 1; - } else { - // increment LCS length by 1 compared to the previous element - curr_row[j] = prev_row[j - 1] + 1; - } - - // update max_length if necessary - if (curr_row[j] > max_length) { - max_length = curr_row[j]; - } - } else { - // reset LCS length if elements don't match - curr_row[j] = 0; - } - } - - // update the previous row for the next iteration - prev_row = curr_row; - } - - // return the maximum length of the LCS - return max_length; -} - static bool ends_with(const std::string & str, const std::string & suffix) { return str.size() >= suffix.size() && 0 == str.compare(str.size() - suffix.size(), suffix.size(), suffix); } @@ -528,7 +508,7 @@ static std::string tokens_to_str(llama_context * ctx, Iter begin, Iter end) { // format incomplete utf-8 multibyte character for output static std::string tokens_to_output_formatted_string(const llama_context * ctx, const llama_token token) { - std::string out = token == -1 ? "" : common_token_to_piece(ctx, token); + std::string out = token == LLAMA_TOKEN_NULL ? "" : common_token_to_piece(ctx, token); // if the size is 1 and first bit is 1, meaning it's a partial character // (size > 1 meaning it's already a known token) @@ -542,48 +522,11 @@ static std::string tokens_to_output_formatted_string(const llama_context * ctx, return out; } -struct completion_token_output { - llama_token tok; - std::string text_to_send; - - struct token_prob { - llama_token tok; - float prob; - }; - - std::vector probs; -}; - -// convert a vector of completion_token_output to json -static json probs_vector_to_json(const llama_context * ctx, const std::vector & probs) { - json out = json::array(); - - for (const auto & prob : probs) { - json probs_for_token = json::array(); - - for (const auto & p : prob.probs) { - const std::string tok_str = tokens_to_output_formatted_string(ctx, p.tok); - probs_for_token.push_back(json { - {"tok_str", tok_str}, - {"prob", p.prob}, - }); - } - - const std::string tok_str = tokens_to_output_formatted_string(ctx, prob.tok); - out.push_back(json { - {"content", tok_str}, - {"probs", probs_for_token}, - }); - } - - return out; -} - static bool server_sent_event(httplib::DataSink & sink, const char * event, const json & data) { const std::string str = std::string(event) + ": " + data.dump(-1, ' ', false, json::error_handler_t::replace) + - "\n\n"; // note: these newlines are important (not sure why though, if you know, add a comment to explain) + "\n\n"; // required by RFC 8895 - A message is terminated by a blank line (two line terminators in a row). LOG_DBG("data stream, to_send: %s", str.c_str()); @@ -594,13 +537,50 @@ static bool server_sent_event(httplib::DataSink & sink, const char * event, cons // OAI utils // -static json oaicompat_completion_params_parse( - const struct llama_model * model, - const json & body, /* openai api json semantics */ - const std::string & chat_template) { +static json oaicompat_completion_params_parse(const json & body) { json llama_params; - llama_params["__oaicompat"] = true; + if (!body.contains("prompt")) { + throw std::runtime_error("\"prompt\" is required"); + } + + // Handle "stop" field + if (body.contains("stop") && body.at("stop").is_string()) { + llama_params["stop"] = json::array({body.at("stop").get()}); + } else { + llama_params["stop"] = json_value(body, "stop", json::array()); + } + + // Handle "n" field + int n_choices = json_value(body, "n", 1); + if (n_choices != 1) { + throw std::runtime_error("Only one completion choice is allowed"); + } + + // Params supported by OAI but unsupported by llama.cpp + static const std::vector unsupported_params { "best_of", "echo", "suffix" }; + for (const auto & param : unsupported_params) { + if (body.contains(param)) { + throw std::runtime_error("Unsupported param: " + param); + } + } + + // Copy remaining properties to llama_params + for (const auto & item : body.items()) { + // Exception: if "n_predict" is present, we overwrite the value specified earlier by "max_tokens" + if (!llama_params.contains(item.key()) || item.key() == "n_predict") { + llama_params[item.key()] = item.value(); + } + } + + return llama_params; +} + +static json oaicompat_chat_completion_params_parse( + const struct llama_model * model, + const json & body, /* openai api json semantics */ + const std::string & chat_template) { + json llama_params; // Apply chat template to the list of messages llama_params["prompt"] = format_chat(model, chat_template, body.at("messages")); @@ -661,172 +641,41 @@ static json oaicompat_completion_params_parse( return llama_params; } -static json format_final_response_oaicompat(const json & request, const json & result, const std::string & completion_id, bool streaming = false, bool verbose = false) { - bool stopped_word = result.count("stopped_word") != 0; - bool stopped_eos = json_value(result, "stopped_eos", false); - int num_tokens_predicted = json_value(result, "tokens_predicted", 0); - int num_prompt_tokens = json_value(result, "tokens_evaluated", 0); - std::string content = json_value(result, "content", std::string("")); - - std::string finish_reason = "length"; - if (stopped_word || stopped_eos) { - finish_reason = "stop"; - } - - json choices = - streaming ? json::array({json{{"finish_reason", finish_reason}, - {"index", 0}, - {"delta", json::object()}}}) - : json::array({json{{"finish_reason", finish_reason}, - {"index", 0}, - {"message", json{{"content", content}, - {"role", "assistant"}}}}}); - - std::time_t t = std::time(0); - - json res = json { - {"choices", choices}, - {"created", t}, - {"model", - json_value(request, "model", std::string(DEFAULT_OAICOMPAT_MODEL))}, - {"object", streaming ? "chat.completion.chunk" : "chat.completion"}, - {"usage", json { - {"completion_tokens", num_tokens_predicted}, - {"prompt_tokens", num_prompt_tokens}, - {"total_tokens", num_tokens_predicted + num_prompt_tokens} - }}, - {"id", completion_id} - }; - - // extra fields for debugging purposes - if (verbose) { - res["__verbose"] = result; - } - - if (result.contains("completion_probabilities")) { - res["completion_probabilities"] = json_value(result, "completion_probabilities", json::array()); - } - - return res; -} - -// return value is vector as there is one case where we might need to generate two responses -static std::vector format_partial_response_oaicompat(const json & result, const std::string & completion_id) { - if (!result.contains("model") || !result.contains("oaicompat_token_ctr")) { - return std::vector({result}); - } - - bool first = json_value(result, "oaicompat_token_ctr", 0) == 0; - std::string modelname = json_value(result, "model", std::string(DEFAULT_OAICOMPAT_MODEL)); - - bool stopped_word = json_value(result, "stopped_word", false); - bool stopped_eos = json_value(result, "stopped_eos", false); - bool stopped_limit = json_value(result, "stopped_limit", false); - std::string content = json_value(result, "content", std::string("")); - - std::string finish_reason; - if (stopped_word || stopped_eos) { - finish_reason = "stop"; - } - if (stopped_limit) { - finish_reason = "length"; - } - - std::time_t t = std::time(0); - - json choices; - - if (!finish_reason.empty()) { - choices = json::array({json{{"finish_reason", finish_reason}, - {"index", 0}, - {"delta", json::object()}}}); - } else { - if (first) { - if (content.empty()) { - choices = json::array({json{{"finish_reason", nullptr}, - {"index", 0}, - {"delta", json{{"role", "assistant"}}}}}); - } else { - // We have to send this as two updates to conform to openai behavior - json initial_ret = json{{"choices", json::array({json{ - {"finish_reason", nullptr}, - {"index", 0}, - {"delta", json{ - {"role", "assistant"} - }}}})}, - {"created", t}, - {"id", completion_id}, - {"model", modelname}, - {"object", "chat.completion.chunk"}}; - - json second_ret = json{ - {"choices", json::array({json{{"finish_reason", nullptr}, - {"index", 0}, - {"delta", json{ - {"content", content}}} - }})}, - {"created", t}, - {"id", completion_id}, - {"model", modelname}, - {"object", "chat.completion.chunk"}}; - - return std::vector({initial_ret, second_ret}); - } - } else { - // Some idiosyncrasy in task processing logic makes several trailing calls - // with empty content, we ignore these at the calee site. - if (content.empty()) { - return std::vector({json::object()}); - } - - choices = json::array({json{ - {"finish_reason", nullptr}, - {"index", 0}, - {"delta", - json{ - {"content", content}, - }}, - }}); - } - } - - json ret = json { - {"choices", choices}, - {"created", t}, - {"id", completion_id}, - {"model", modelname}, - {"object", "chat.completion.chunk"} - }; - if (!finish_reason.empty()) { - int num_tokens_predicted = json_value(result, "tokens_predicted", 0); - int num_prompt_tokens = json_value(result, "tokens_evaluated", 0); - ret.push_back({"usage", json { - {"completion_tokens", num_tokens_predicted}, - {"prompt_tokens", num_prompt_tokens}, - {"total_tokens", num_tokens_predicted + num_prompt_tokens} - }}); - } - - return std::vector({ret}); -} - -static json format_embeddings_response_oaicompat(const json & request, const json & embeddings) { +static json format_embeddings_response_oaicompat(const json & request, const json & embeddings, bool use_base64 = false) { json data = json::array(); + int32_t n_tokens = 0; int i = 0; for (const auto & elem : embeddings) { - data.push_back(json{ - {"embedding", json_value(elem, "embedding", json::array())}, - {"index", i++}, - {"object", "embedding"} - }); + json embedding_obj; + + if (use_base64) { + const auto& vec = json_value(elem, "embedding", json::array()).get>(); + const char* data_ptr = reinterpret_cast(vec.data()); + size_t data_size = vec.size() * sizeof(float); + embedding_obj = { + {"embedding", base64::encode(data_ptr, data_size)}, + {"index", i++}, + {"object", "embedding"}, + {"encoding_format", "base64"} + }; + } else { + embedding_obj = { + {"embedding", json_value(elem, "embedding", json::array())}, + {"index", i++}, + {"object", "embedding"} + }; + } + data.push_back(embedding_obj); + + n_tokens += json_value(elem, "tokens_evaluated", 0); } json res = json { {"model", json_value(request, "model", std::string(DEFAULT_OAICOMPAT_MODEL))}, {"object", "list"}, - {"usage", json { // TODO: fill - {"prompt_tokens", 0}, - {"total_tokens", 0} + {"usage", json { + {"prompt_tokens", n_tokens}, + {"total_tokens", n_tokens} }}, {"data", data} }; @@ -836,20 +685,23 @@ static json format_embeddings_response_oaicompat(const json & request, const jso static json format_response_rerank(const json & request, const json & ranks) { json data = json::array(); + int32_t n_tokens = 0; int i = 0; for (const auto & rank : ranks) { data.push_back(json{ {"index", i++}, {"relevance_score", json_value(rank, "score", 0.0)}, }); + + n_tokens += json_value(rank, "tokens_evaluated", 0); } json res = json { {"model", json_value(request, "model", std::string(DEFAULT_OAICOMPAT_MODEL))}, {"object", "list"}, - {"usage", json { // TODO: fill - {"prompt_tokens", 0}, - {"total_tokens", 0} + {"usage", json { + {"prompt_tokens", n_tokens}, + {"total_tokens", n_tokens} }}, {"results", data} }; @@ -902,42 +754,92 @@ static json format_detokenized_response(const std::string & content) { }; } -static json format_error_response(const std::string & message, const enum error_type type) { - std::string type_str; - int code = 500; - switch (type) { - case ERROR_TYPE_INVALID_REQUEST: - type_str = "invalid_request_error"; - code = 400; - break; - case ERROR_TYPE_AUTHENTICATION: - type_str = "authentication_error"; - code = 401; - break; - case ERROR_TYPE_NOT_FOUND: - type_str = "not_found_error"; - code = 404; - break; - case ERROR_TYPE_SERVER: - type_str = "server_error"; - code = 500; - break; - case ERROR_TYPE_PERMISSION: - type_str = "permission_error"; - code = 403; - break; - case ERROR_TYPE_NOT_SUPPORTED: - type_str = "not_supported_error"; - code = 501; - break; - case ERROR_TYPE_UNAVAILABLE: - type_str = "unavailable_error"; - code = 503; - break; +static json format_logit_bias(const std::vector & logit_bias) { + json data = json::array(); + for (const auto & lb : logit_bias) { + data.push_back(json{ + {"bias", lb.bias}, + {"token", lb.token}, + }); } - return json { - {"code", code}, - {"message", message}, - {"type", type_str}, - }; + return data; +} + +static std::string safe_json_to_str(const json & data) { + return data.dump(-1, ' ', false, json::error_handler_t::replace); +} + +static std::vector get_token_probabilities(llama_context * ctx, int idx) { + std::vector cur; + const auto * logits = llama_get_logits_ith(ctx, idx); + + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_vocab = llama_vocab_n_tokens(vocab); + + cur.resize(n_vocab); + for (llama_token token_id = 0; token_id < n_vocab; token_id++) { + cur[token_id] = llama_token_data{token_id, logits[token_id], 0.0f}; + } + + // sort tokens by logits + std::sort(cur.begin(), cur.end(), [](const llama_token_data & a, const llama_token_data & b) { + return a.logit > b.logit; + }); + + // apply softmax + float max_l = cur[0].logit; + float cum_sum = 0.0f; + for (size_t i = 0; i < cur.size(); ++i) { + float p = expf(cur[i].logit - max_l); + cur[i].p = p; + cum_sum += p; + } + for (size_t i = 0; i < cur.size(); ++i) { + cur[i].p /= cum_sum; + } + + return cur; +} + +static bool are_lora_equal( + const std::vector & l1, + const std::vector & l2) { + if (l1.size() != l2.size()) { + return false; + } + for (size_t i = 0; i < l1.size(); ++i) { + // we don't check lora.path to reduce the time complexity + if (l1[i].scale != l2[i].scale || l1[i].ptr != l2[i].ptr) { + return false; + } + } + return true; +} + +// parse lora config from JSON request, returned a copy of lora_base with updated scale +static std::vector parse_lora_request( + const std::vector & lora_base, + const json & data) { + std::vector lora(lora_base); + int max_idx = lora.size(); + + // clear existing value + for (auto & entry : lora) { + entry.scale = 0.0f; + } + + // set value + for (const auto & entry : data) { + int id = json_value(entry, "id", -1); + float scale = json_value(entry, "scale", 0.0f); + if (0 <= id && id < max_idx) { + lora[id].scale = scale; + } else { + throw std::runtime_error("invalid adapter id"); + } + } + + return lora; } diff --git a/examples/server/webui/index.html b/examples/server/webui/index.html new file mode 100644 index 000000000..2180ef4ad --- /dev/null +++ b/examples/server/webui/index.html @@ -0,0 +1,318 @@ + + + + + + + 🦙 llama.cpp - chat + + + +
+
+ + + +
+ +
+
+

Conversations

+ + + +
+ + +
+ + New conversation +
+
+ {{ conv.messages[0].content }} +
+
+ Conversations are saved to browser's localStorage +
+
+
+ + +
+ +
+ + + +
llama.cpp
+ + +
+
+ + + + +
+
+ +
+ + +
+
+
+ + + +
+
    +
  • + +
    • +
  • + +
    • +
+
+
+
+
+ + +
+
+ + {{ messages.length === 0 ? 'Send a message to start' : '' }} +
+
+ +
+ + +
+ +
+
+ + +
+ + + +
+
+ +
+ + + + +
+

Settings

+
+

Settings below are saved in browser's localStorage

+ + + + + +
+ Other sampler settings +
+ + + + +
+
+ +
+ Penalties settings +
+ +
+
+ +
+ Advanced config +
+
+ + +
+
+ + Show tokens per second +
+ +
+
+
+ + +
+ + + +
+
+ + +
+ + + + + + + + + + + + + diff --git a/examples/server/webui/package-lock.json b/examples/server/webui/package-lock.json new file mode 100644 index 000000000..bbebccbf2 --- /dev/null +++ b/examples/server/webui/package-lock.json @@ -0,0 +1,3309 @@ +{ + "name": "webui", + "version": "0.0.0", + "lockfileVersion": 3, + "requires": true, + "packages": { + "": { + "name": "webui", + "version": "0.0.0", + "dependencies": { + "@sec-ant/readable-stream": "^0.6.0", + "@vscode/markdown-it-katex": "^1.1.1", + "autoprefixer": "^10.4.20", + "daisyui": "^4.12.14", + "highlight.js": "^11.10.0", + "katex": "^0.16.15", + "markdown-it": "^14.1.0", + "postcss": "^8.4.49", + "tailwindcss": "^3.4.15", + "textlinestream": "^1.1.1", + "vite-plugin-singlefile": "^2.0.3", + "vue": "^3.5.13" + }, + "devDependencies": { + "sass-embedded": "^1.83.0", + "vite": "^5.4.10" + } + }, + "node_modules/@alloc/quick-lru": { + "version": "5.2.0", + "resolved": "https://registry.npmjs.org/@alloc/quick-lru/-/quick-lru-5.2.0.tgz", + "integrity": "sha512-UrcABB+4bUrFABwbluTIBErXwvbsU/V7TZWfmbgJfbkwiBuziS9gxdODUyuiecfdGQ85jglMW6juS3+z5TsKLw==", + "license": "MIT", + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/@bufbuild/protobuf": { + "version": "2.2.3", + "resolved": "https://registry.npmjs.org/@bufbuild/protobuf/-/protobuf-2.2.3.tgz", + "integrity": "sha512-tFQoXHJdkEOSwj5tRIZSPNUuXK3RaR7T1nUrPgbYX1pUbvqqaaZAsfo+NXBPsz5rZMSKVFrgK1WL8Q/MSLvprg==", + "devOptional": true, + "license": "(Apache-2.0 AND BSD-3-Clause)" + }, + "node_modules/@esbuild/aix-ppc64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.21.5.tgz", + "integrity": "sha512-1SDgH6ZSPTlggy1yI6+Dbkiz8xzpHJEVAlF/AM1tHPLsf5STom9rwtjE4hKAF20FfXXNTFqEYXyJNWh1GiZedQ==", + "cpu": [ + "ppc64" + ], + "license": "MIT", + "optional": true, + "os": [ + "aix" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/android-arm": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.21.5.tgz", + "integrity": "sha512-vCPvzSjpPHEi1siZdlvAlsPxXl7WbOVUBBAowWug4rJHb68Ox8KualB+1ocNvT5fjv6wpkX6o/iEpbDrf68zcg==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/android-arm64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.21.5.tgz", + "integrity": "sha512-c0uX9VAUBQ7dTDCjq+wdyGLowMdtR/GoC2U5IYk/7D1H1JYC0qseD7+11iMP2mRLN9RcCMRcjC4YMclCzGwS/A==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/android-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.21.5.tgz", + "integrity": "sha512-D7aPRUUNHRBwHxzxRvp856rjUHRFW1SdQATKXH2hqA0kAZb1hKmi02OpYRacl0TxIGz/ZmXWlbZgjwWYaCakTA==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/darwin-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.21.5.tgz", + "integrity": "sha512-se/JjF8NlmKVG4kNIuyWMV/22ZaerB+qaSi5MdrXtd6R08kvs2qCN4C09miupktDitvh8jRFflwGFBQcxZRjbw==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/freebsd-arm64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.21.5.tgz", + "integrity": "sha512-5JcRxxRDUJLX8JXp/wcBCy3pENnCgBR9bN6JsY4OmhfUtIHe3ZW0mawA7+RDAcMLrMIZaf03NlQiX9DGyB8h4g==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/freebsd-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.21.5.tgz", + "integrity": "sha512-J95kNBj1zkbMXtHVH29bBriQygMXqoVQOQYA+ISs0/2l3T9/kj42ow2mpqerRBxDJnmkUDCaQT/dfNXWX/ZZCQ==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-arm": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.21.5.tgz", + "integrity": "sha512-bPb5AHZtbeNGjCKVZ9UGqGwo8EUu4cLq68E95A53KlxAPRmUyYv2D6F0uUI65XisGOL1hBP5mTronbgo+0bFcA==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-arm64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.21.5.tgz", + "integrity": "sha512-ibKvmyYzKsBeX8d8I7MH/TMfWDXBF3db4qM6sy+7re0YXya+K1cem3on9XgdT2EQGMu4hQyZhan7TeQ8XkGp4Q==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-ia32": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.21.5.tgz", + "integrity": "sha512-YvjXDqLRqPDl2dvRODYmmhz4rPeVKYvppfGYKSNGdyZkA01046pLWyRKKI3ax8fbJoK5QbxblURkwK/MWY18Tg==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-loong64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.21.5.tgz", + "integrity": "sha512-uHf1BmMG8qEvzdrzAqg2SIG/02+4/DHB6a9Kbya0XDvwDEKCoC8ZRWI5JJvNdUjtciBGFQ5PuBlpEOXQj+JQSg==", + "cpu": [ + "loong64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-mips64el": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.21.5.tgz", + "integrity": "sha512-IajOmO+KJK23bj52dFSNCMsz1QP1DqM6cwLUv3W1QwyxkyIWecfafnI555fvSGqEKwjMXVLokcV5ygHW5b3Jbg==", + "cpu": [ + "mips64el" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-ppc64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.21.5.tgz", + "integrity": "sha512-1hHV/Z4OEfMwpLO8rp7CvlhBDnjsC3CttJXIhBi+5Aj5r+MBvy4egg7wCbe//hSsT+RvDAG7s81tAvpL2XAE4w==", + "cpu": [ + "ppc64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-riscv64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.21.5.tgz", + "integrity": "sha512-2HdXDMd9GMgTGrPWnJzP2ALSokE/0O5HhTUvWIbD3YdjME8JwvSCnNGBnTThKGEB91OZhzrJ4qIIxk/SBmyDDA==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-s390x": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.21.5.tgz", + "integrity": "sha512-zus5sxzqBJD3eXxwvjN1yQkRepANgxE9lgOW2qLnmr8ikMTphkjgXu1HR01K4FJg8h1kEEDAqDcZQtbrRnB41A==", + "cpu": [ + "s390x" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/linux-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.21.5.tgz", + "integrity": "sha512-1rYdTpyv03iycF1+BhzrzQJCdOuAOtaqHTWJZCWvijKD2N5Xu0TtVC8/+1faWqcP9iBCWOmjmhoH94dH82BxPQ==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/netbsd-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.21.5.tgz", + "integrity": "sha512-Woi2MXzXjMULccIwMnLciyZH4nCIMpWQAs049KEeMvOcNADVxo0UBIQPfSmxB3CWKedngg7sWZdLvLczpe0tLg==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "netbsd" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/openbsd-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.21.5.tgz", + "integrity": "sha512-HLNNw99xsvx12lFBUwoT8EVCsSvRNDVxNpjZ7bPn947b8gJPzeHWyNVhFsaerc0n3TsbOINvRP2byTZ5LKezow==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "openbsd" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/sunos-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.21.5.tgz", + "integrity": "sha512-6+gjmFpfy0BHU5Tpptkuh8+uw3mnrvgs+dSPQXQOv3ekbordwnzTVEb4qnIvQcYXq6gzkyTnoZ9dZG+D4garKg==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "sunos" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/win32-arm64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.21.5.tgz", + "integrity": "sha512-Z0gOTd75VvXqyq7nsl93zwahcTROgqvuAcYDUr+vOv8uHhNSKROyU961kgtCD1e95IqPKSQKH7tBTslnS3tA8A==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/win32-ia32": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.21.5.tgz", + "integrity": "sha512-SWXFF1CL2RVNMaVs+BBClwtfZSvDgtL//G/smwAc5oVK/UPu2Gu9tIaRgFmYFFKrmg3SyAjSrElf0TiJ1v8fYA==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@esbuild/win32-x64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.21.5.tgz", + "integrity": "sha512-tQd/1efJuzPC6rCFwEvLtci/xNFcTZknmXs98FYDfGE4wP9ClFV98nyKrzJKVPMhdDnjzLhdUyMX4PsQAPjwIw==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/@rollup/rollup-android-arm-eabi": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.28.0.tgz", + "integrity": "sha512-wLJuPLT6grGZsy34g4N1yRfYeouklTgPhH1gWXCYspenKYD0s3cR99ZevOGw5BexMNywkbV3UkjADisozBmpPQ==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ] + }, + "node_modules/@rollup/rollup-android-arm64": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm64/-/rollup-android-arm64-4.28.0.tgz", + "integrity": "sha512-eiNkznlo0dLmVG/6wf+Ifi/v78G4d4QxRhuUl+s8EWZpDewgk7PX3ZyECUXU0Zq/Ca+8nU8cQpNC4Xgn2gFNDA==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ] + }, + "node_modules/@rollup/rollup-darwin-x64": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-x64/-/rollup-darwin-x64-4.28.0.tgz", + "integrity": "sha512-8hxgfReVs7k9Js1uAIhS6zq3I+wKQETInnWQtgzt8JfGx51R1N6DRVy3F4o0lQwumbErRz52YqwjfvuwRxGv1w==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ] + }, + "node_modules/@rollup/rollup-freebsd-arm64": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-arm64/-/rollup-freebsd-arm64-4.28.0.tgz", + "integrity": "sha512-lA1zZB3bFx5oxu9fYud4+g1mt+lYXCoch0M0V/xhqLoGatbzVse0wlSQ1UYOWKpuSu3gyN4qEc0Dxf/DII1bhQ==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ] + }, + "node_modules/@rollup/rollup-freebsd-x64": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-x64/-/rollup-freebsd-x64-4.28.0.tgz", + "integrity": "sha512-aI2plavbUDjCQB/sRbeUZWX9qp12GfYkYSJOrdYTL/C5D53bsE2/nBPuoiJKoWp5SN78v2Vr8ZPnB+/VbQ2pFA==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ] + }, + "node_modules/@rollup/rollup-linux-arm-gnueabihf": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-gnueabihf/-/rollup-linux-arm-gnueabihf-4.28.0.tgz", + "integrity": "sha512-WXveUPKtfqtaNvpf0iOb0M6xC64GzUX/OowbqfiCSXTdi/jLlOmH0Ba94/OkiY2yTGTwteo4/dsHRfh5bDCZ+w==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm-musleabihf": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-musleabihf/-/rollup-linux-arm-musleabihf-4.28.0.tgz", + "integrity": "sha512-yLc3O2NtOQR67lI79zsSc7lk31xjwcaocvdD1twL64PK1yNaIqCeWI9L5B4MFPAVGEVjH5k1oWSGuYX1Wutxpg==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm64-gnu": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-gnu/-/rollup-linux-arm64-gnu-4.28.0.tgz", + "integrity": "sha512-+P9G9hjEpHucHRXqesY+3X9hD2wh0iNnJXX/QhS/J5vTdG6VhNYMxJ2rJkQOxRUd17u5mbMLHM7yWGZdAASfcg==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm64-musl": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-musl/-/rollup-linux-arm64-musl-4.28.0.tgz", + "integrity": "sha512-1xsm2rCKSTpKzi5/ypT5wfc+4bOGa/9yI/eaOLW0oMs7qpC542APWhl4A37AENGZ6St6GBMWhCCMM6tXgTIplw==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-powerpc64le-gnu": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-powerpc64le-gnu/-/rollup-linux-powerpc64le-gnu-4.28.0.tgz", + "integrity": "sha512-zgWxMq8neVQeXL+ouSf6S7DoNeo6EPgi1eeqHXVKQxqPy1B2NvTbaOUWPn/7CfMKL7xvhV0/+fq/Z/J69g1WAQ==", + "cpu": [ + "ppc64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-riscv64-gnu": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-gnu/-/rollup-linux-riscv64-gnu-4.28.0.tgz", + "integrity": "sha512-VEdVYacLniRxbRJLNtzwGt5vwS0ycYshofI7cWAfj7Vg5asqj+pt+Q6x4n+AONSZW/kVm+5nklde0qs2EUwU2g==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-s390x-gnu": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-s390x-gnu/-/rollup-linux-s390x-gnu-4.28.0.tgz", + "integrity": "sha512-LQlP5t2hcDJh8HV8RELD9/xlYtEzJkm/aWGsauvdO2ulfl3QYRjqrKW+mGAIWP5kdNCBheqqqYIGElSRCaXfpw==", + "cpu": [ + "s390x" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-x64-gnu": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-gnu/-/rollup-linux-x64-gnu-4.28.0.tgz", + "integrity": "sha512-Nl4KIzteVEKE9BdAvYoTkW19pa7LR/RBrT6F1dJCV/3pbjwDcaOq+edkP0LXuJ9kflW/xOK414X78r+K84+msw==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-x64-musl": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-musl/-/rollup-linux-x64-musl-4.28.0.tgz", + "integrity": "sha512-eKpJr4vBDOi4goT75MvW+0dXcNUqisK4jvibY9vDdlgLx+yekxSm55StsHbxUsRxSTt3JEQvlr3cGDkzcSP8bw==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-win32-arm64-msvc": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-arm64-msvc/-/rollup-win32-arm64-msvc-4.28.0.tgz", + "integrity": "sha512-Vi+WR62xWGsE/Oj+mD0FNAPY2MEox3cfyG0zLpotZdehPFXwz6lypkGs5y38Jd/NVSbOD02aVad6q6QYF7i8Bg==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@rollup/rollup-win32-ia32-msvc": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-ia32-msvc/-/rollup-win32-ia32-msvc-4.28.0.tgz", + "integrity": "sha512-kN/Vpip8emMLn/eOza+4JwqDZBL6MPNpkdaEsgUtW1NYN3DZvZqSQrbKzJcTL6hd8YNmFTn7XGWMwccOcJBL0A==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@rollup/rollup-win32-x64-msvc": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-msvc/-/rollup-win32-x64-msvc-4.28.0.tgz", + "integrity": "sha512-Bvno2/aZT6usSa7lRDL2+hMjVAGjuqaymF1ApZm31JXzniR/hvr14jpU+/z4X6Gt5BPlzosscyJZGUvguXIqeQ==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@sec-ant/readable-stream": { + "version": "0.6.0", + "resolved": "https://registry.npmjs.org/@sec-ant/readable-stream/-/readable-stream-0.6.0.tgz", + "integrity": "sha512-uiBh8DrB5FN35gP6/o8JEhEQ7/ci1jUsOZO/VMUjyvTpjtV54VstOXVj1TvTj/wsT23pfX6butxxh3qufsW3+g==", + "license": "MIT" + }, + "node_modules/@vscode/markdown-it-katex": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/@vscode/markdown-it-katex/-/markdown-it-katex-1.1.1.tgz", + "integrity": "sha512-3KTlbsRBPJQLE2YmLL7K6nunTlU+W9T5+FjfNdWuIUKgxSS6HWLQHaO3L4MkJi7z7MpIPpY+g4N+cWNBPE/MSA==", + "license": "MIT", + "dependencies": { + "katex": "^0.16.4" + } + }, + "node_modules/@vue/compiler-dom": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-dom/-/compiler-dom-3.5.13.tgz", + "integrity": "sha512-ZOJ46sMOKUjO3e94wPdCzQ6P1Lx/vhp2RSvfaab88Ajexs0AHeV0uasYhi99WPaogmBlRHNRuly8xV75cNTMDA==", + "license": "MIT", + "dependencies": { + "@vue/compiler-core": "3.5.13", + "@vue/shared": "3.5.13" + } + }, + "node_modules/@vue/compiler-dom/node_modules/@babel/helper-string-parser": { + "version": "7.25.9", + "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.25.9.tgz", + "integrity": "sha512-4A/SCr/2KLd5jrtOMFzaKjVtAei3+2r/NChoBNoZ3EyP/+GlhoaEGoWOZUmFmoITP7zOJyHIMm+DYRd8o3PvHA==", + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-dom/node_modules/@babel/helper-validator-identifier": { + "version": "7.25.9", + "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.25.9.tgz", + "integrity": "sha512-Ed61U6XJc3CVRfkERJWDz4dJwKe7iLmmJsbOGu9wSloNSFttHV0I8g6UAgb7qnK5ly5bGLPd4oXZlxCdANBOWQ==", + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-dom/node_modules/@babel/parser": { + "version": "7.26.2", + "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.26.2.tgz", + "integrity": "sha512-DWMCZH9WA4Maitz2q21SRKHo9QXZxkDsbNZoVD62gusNtNBBqDg9i7uOhASfTfIGNzW+O+r7+jAlM8dwphcJKQ==", + "license": "MIT", + "dependencies": { + "@babel/types": "^7.26.0" + }, + "bin": { + "parser": "bin/babel-parser.js" + }, + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/@vue/compiler-dom/node_modules/@babel/types": { + "version": "7.26.0", + "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.26.0.tgz", + "integrity": "sha512-Z/yiTPj+lDVnF7lWeKCIJzaIkI0vYO87dMpZ4bg4TDrFe4XXLFWL1TbXU27gBP3QccxV9mZICCrnjnYlJjXHOA==", + "license": "MIT", + "dependencies": { + "@babel/helper-string-parser": "^7.25.9", + "@babel/helper-validator-identifier": "^7.25.9" + }, + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-dom/node_modules/@vue/compiler-core": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-core/-/compiler-core-3.5.13.tgz", + "integrity": "sha512-oOdAkwqUfW1WqpwSYJce06wvt6HljgY3fGeM9NcVA1HaYOij3mZG9Rkysn0OHuyUAGMbEbARIpsG+LPVlBJ5/Q==", + "license": "MIT", + "dependencies": { + "@babel/parser": "^7.25.3", + "@vue/shared": "3.5.13", + "entities": "^4.5.0", + "estree-walker": "^2.0.2", + "source-map-js": "^1.2.0" + } + }, + "node_modules/@vue/compiler-dom/node_modules/estree-walker": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/estree-walker/-/estree-walker-2.0.2.tgz", + "integrity": "sha512-Rfkk/Mp/DL7JVje3u18FxFujQlTNR2q6QfMSMB7AvCBx91NGj/ba3kCfza0f6dVDbw7YlRf/nDrn7pQrCCyQ/w==", + "license": "MIT" + }, + "node_modules/@vue/compiler-dom/node_modules/source-map-js": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz", + "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==", + "license": "BSD-3-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/@vue/compiler-sfc": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-sfc/-/compiler-sfc-3.5.13.tgz", + "integrity": "sha512-6VdaljMpD82w6c2749Zhf5T9u5uLBWKnVue6XWxprDobftnletJ8+oel7sexFfM3qIxNmVE7LSFGTpv6obNyaQ==", + "license": "MIT", + "dependencies": { + "@babel/parser": "^7.25.3", + "@vue/compiler-core": "3.5.13", + "@vue/compiler-dom": "3.5.13", + "@vue/compiler-ssr": "3.5.13", + "@vue/shared": "3.5.13", + "estree-walker": "^2.0.2", + "magic-string": "^0.30.11", + "postcss": "^8.4.48", + "source-map-js": "^1.2.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@babel/helper-string-parser": { + "version": "7.25.9", + "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.25.9.tgz", + "integrity": "sha512-4A/SCr/2KLd5jrtOMFzaKjVtAei3+2r/NChoBNoZ3EyP/+GlhoaEGoWOZUmFmoITP7zOJyHIMm+DYRd8o3PvHA==", + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@babel/helper-validator-identifier": { + "version": "7.25.9", + "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.25.9.tgz", + "integrity": "sha512-Ed61U6XJc3CVRfkERJWDz4dJwKe7iLmmJsbOGu9wSloNSFttHV0I8g6UAgb7qnK5ly5bGLPd4oXZlxCdANBOWQ==", + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@babel/parser": { + "version": "7.26.2", + "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.26.2.tgz", + "integrity": "sha512-DWMCZH9WA4Maitz2q21SRKHo9QXZxkDsbNZoVD62gusNtNBBqDg9i7uOhASfTfIGNzW+O+r7+jAlM8dwphcJKQ==", + "license": "MIT", + "dependencies": { + "@babel/types": "^7.26.0" + }, + "bin": { + "parser": "bin/babel-parser.js" + }, + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@babel/types": { + "version": "7.26.0", + "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.26.0.tgz", + "integrity": "sha512-Z/yiTPj+lDVnF7lWeKCIJzaIkI0vYO87dMpZ4bg4TDrFe4XXLFWL1TbXU27gBP3QccxV9mZICCrnjnYlJjXHOA==", + "license": "MIT", + "dependencies": { + "@babel/helper-string-parser": "^7.25.9", + "@babel/helper-validator-identifier": "^7.25.9" + }, + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@jridgewell/sourcemap-codec": { + "version": "1.5.0", + "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.0.tgz", + "integrity": "sha512-gv3ZRaISU3fjPAgNsriBRqGWQL6quFx04YMPW/zD8XMLsU32mhCCbfbO6KZFLjvYpCZ8zyDEgqsgf+PwPaM7GQ==", + "license": "MIT" + }, + "node_modules/@vue/compiler-sfc/node_modules/@vue/compiler-core": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-core/-/compiler-core-3.5.13.tgz", + "integrity": "sha512-oOdAkwqUfW1WqpwSYJce06wvt6HljgY3fGeM9NcVA1HaYOij3mZG9Rkysn0OHuyUAGMbEbARIpsG+LPVlBJ5/Q==", + "license": "MIT", + "dependencies": { + "@babel/parser": "^7.25.3", + "@vue/shared": "3.5.13", + "entities": "^4.5.0", + "estree-walker": "^2.0.2", + "source-map-js": "^1.2.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/@vue/compiler-ssr": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-ssr/-/compiler-ssr-3.5.13.tgz", + "integrity": "sha512-wMH6vrYHxQl/IybKJagqbquvxpWCuVYpoUJfCqFZwa/JY1GdATAQ+TgVtgrwwMZ0D07QhA99rs/EAAWfvG6KpA==", + "license": "MIT", + "dependencies": { + "@vue/compiler-dom": "3.5.13", + "@vue/shared": "3.5.13" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/estree-walker": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/estree-walker/-/estree-walker-2.0.2.tgz", + "integrity": "sha512-Rfkk/Mp/DL7JVje3u18FxFujQlTNR2q6QfMSMB7AvCBx91NGj/ba3kCfza0f6dVDbw7YlRf/nDrn7pQrCCyQ/w==", + "license": "MIT" + }, + "node_modules/@vue/compiler-sfc/node_modules/magic-string": { + "version": "0.30.14", + "resolved": "https://registry.npmjs.org/magic-string/-/magic-string-0.30.14.tgz", + "integrity": "sha512-5c99P1WKTed11ZC0HMJOj6CDIue6F8ySu+bJL+85q1zBEIY8IklrJ1eiKC2NDRh3Ct3FcvmJPyQHb9erXMTJNw==", + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.5.0" + } + }, + "node_modules/@vue/compiler-sfc/node_modules/source-map-js": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz", + "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==", + "license": "BSD-3-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/@vue/runtime-dom": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/runtime-dom/-/runtime-dom-3.5.13.tgz", + "integrity": "sha512-dLaj94s93NYLqjLiyFzVs9X6dWhTdAlEAciC3Moq7gzAc13VJUdCnjjRurNM6uTLFATRHexHCTu/Xp3eW6yoog==", + "license": "MIT", + "dependencies": { + "@vue/reactivity": "3.5.13", + "@vue/runtime-core": "3.5.13", + "@vue/shared": "3.5.13", + "csstype": "^3.1.3" + } + }, + "node_modules/@vue/runtime-dom/node_modules/@vue/reactivity": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/reactivity/-/reactivity-3.5.13.tgz", + "integrity": "sha512-NaCwtw8o48B9I6L1zl2p41OHo/2Z4wqYGGIK1Khu5T7yxrn+ATOixn/Udn2m+6kZKB/J7cuT9DbWWhRxqixACg==", + "license": "MIT", + "dependencies": { + "@vue/shared": "3.5.13" + } + }, + "node_modules/@vue/runtime-dom/node_modules/@vue/runtime-core": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/runtime-core/-/runtime-core-3.5.13.tgz", + "integrity": "sha512-Fj4YRQ3Az0WTZw1sFe+QDb0aXCerigEpw418pw1HBUKFtnQHWzwojaukAs2X/c9DQz4MQ4bsXTGlcpGxU/RCIw==", + "license": "MIT", + "dependencies": { + "@vue/reactivity": "3.5.13", + "@vue/shared": "3.5.13" + } + }, + "node_modules/@vue/runtime-dom/node_modules/csstype": { + "version": "3.1.3", + "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.1.3.tgz", + "integrity": "sha512-M1uQkMl8rQK/szD0LNhtqxIPLpimGm8sOBwU7lLnCpSbTyY3yeU1Vc7l4KT5zT4s/yOxHH5O7tIuuLOCnLADRw==", + "license": "MIT" + }, + "node_modules/@vue/server-renderer": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/server-renderer/-/server-renderer-3.5.13.tgz", + "integrity": "sha512-wAi4IRJV/2SAW3htkTlB+dHeRmpTiVIK1OGLWV1yeStVSebSQQOwGwIq0D3ZIoBj2C2qpgz5+vX9iEBkTdk5YA==", + "license": "MIT", + "dependencies": { + "@vue/compiler-ssr": "3.5.13", + "@vue/shared": "3.5.13" + }, + "peerDependencies": { + "vue": "3.5.13" + } + }, + "node_modules/@vue/server-renderer/node_modules/@vue/compiler-ssr": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/compiler-ssr/-/compiler-ssr-3.5.13.tgz", + "integrity": "sha512-wMH6vrYHxQl/IybKJagqbquvxpWCuVYpoUJfCqFZwa/JY1GdATAQ+TgVtgrwwMZ0D07QhA99rs/EAAWfvG6KpA==", + "license": "MIT", + "dependencies": { + "@vue/compiler-dom": "3.5.13", + "@vue/shared": "3.5.13" + } + }, + "node_modules/@vue/shared": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/@vue/shared/-/shared-3.5.13.tgz", + "integrity": "sha512-/hnE/qP5ZoGpol0a5mDi45bOd7t3tjYJBjsgCsivow7D48cJeV5l05RD82lPqi7gRiphZM37rnhW1l6ZoCNNnQ==", + "license": "MIT" + }, + "node_modules/arg": { + "version": "5.0.2", + "resolved": "https://registry.npmjs.org/arg/-/arg-5.0.2.tgz", + "integrity": "sha512-PYjyFOLKQ9y57JvQ6QLo8dAgNqswh8M1RMJYdQduT6xbWSgK36P/Z/v+p888pM69jMMfS8Xd8F6I1kQ/I9HUGg==", + "license": "MIT" + }, + "node_modules/argparse": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/argparse/-/argparse-2.0.1.tgz", + "integrity": "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q==", + "license": "Python-2.0" + }, + "node_modules/autoprefixer": { + "version": "10.4.20", + "resolved": "https://registry.npmjs.org/autoprefixer/-/autoprefixer-10.4.20.tgz", + "integrity": "sha512-XY25y5xSv/wEoqzDyXXME4AFfkZI0P23z6Fs3YgymDnKJkCGOnkL0iTxCa85UTqaSgfcqyf3UA6+c7wUvx/16g==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/autoprefixer" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "browserslist": "^4.23.3", + "caniuse-lite": "^1.0.30001646", + "fraction.js": "^4.3.7", + "normalize-range": "^0.1.2", + "picocolors": "^1.0.1", + "postcss-value-parser": "^4.2.0" + }, + "bin": { + "autoprefixer": "bin/autoprefixer" + }, + "engines": { + "node": "^10 || ^12 || >=14" + }, + "peerDependencies": { + "postcss": "^8.1.0" + } + }, + "node_modules/browserslist": { + "version": "4.24.2", + "resolved": "https://registry.npmjs.org/browserslist/-/browserslist-4.24.2.tgz", + "integrity": "sha512-ZIc+Q62revdMcqC6aChtW4jz3My3klmCO1fEmINZY/8J3EpBg5/A/D0AKmBveUh6pgoeycoMkVMko84tuYS+Gg==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/browserslist" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/browserslist" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "caniuse-lite": "^1.0.30001669", + "electron-to-chromium": "^1.5.41", + "node-releases": "^2.0.18", + "update-browserslist-db": "^1.1.1" + }, + "bin": { + "browserslist": "cli.js" + }, + "engines": { + "node": "^6 || ^7 || ^8 || ^9 || ^10 || ^11 || ^12 || >=13.7" + } + }, + "node_modules/browserslist/node_modules/electron-to-chromium": { + "version": "1.5.67", + "resolved": "https://registry.npmjs.org/electron-to-chromium/-/electron-to-chromium-1.5.67.tgz", + "integrity": "sha512-nz88NNBsD7kQSAGGJyp8hS6xSPtWwqNogA0mjtc2nUYeEf3nURK9qpV18TuBdDmEDgVWotS8Wkzf+V52dSQ/LQ==", + "license": "ISC" + }, + "node_modules/browserslist/node_modules/escalade": { + "version": "3.2.0", + "resolved": "https://registry.npmjs.org/escalade/-/escalade-3.2.0.tgz", + "integrity": "sha512-WUj2qlxaQtO4g6Pq5c29GTcWGDyd8itL8zTlipgECz3JesAiiOKotd8JU6otB3PACgG6xkJUyVhboMS+bje/jA==", + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/browserslist/node_modules/node-releases": { + "version": "2.0.18", + "resolved": "https://registry.npmjs.org/node-releases/-/node-releases-2.0.18.tgz", + "integrity": "sha512-d9VeXT4SJ7ZeOqGX6R5EM022wpL+eWPooLI+5UpWn2jCT1aosUQEhQP214x33Wkwx3JQMvIm+tIoVOdodFS40g==", + "license": "MIT" + }, + "node_modules/browserslist/node_modules/update-browserslist-db": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/update-browserslist-db/-/update-browserslist-db-1.1.1.tgz", + "integrity": "sha512-R8UzCaa9Az+38REPiJ1tXlImTJXlVfgHZsglwBD/k6nj76ctsH1E3q4doGrukiLQd3sGQYu56r5+lo5r94l29A==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/browserslist" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/browserslist" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "escalade": "^3.2.0", + "picocolors": "^1.1.0" + }, + "bin": { + "update-browserslist-db": "cli.js" + }, + "peerDependencies": { + "browserslist": ">= 4.21.0" + } + }, + "node_modules/buffer-builder": { + "version": "0.2.0", + "resolved": "https://registry.npmjs.org/buffer-builder/-/buffer-builder-0.2.0.tgz", + "integrity": "sha512-7VPMEPuYznPSoR21NE1zvd2Xna6c/CloiZCfcMXR1Jny6PjX0N4Nsa38zcBFo/FMK+BlA+FLKbJCQ0i2yxp+Xg==", + "devOptional": true, + "license": "MIT/X11" + }, + "node_modules/caniuse-lite": { + "version": "1.0.30001684", + "resolved": "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001684.tgz", + "integrity": "sha512-G1LRwLIQjBQoyq0ZJGqGIJUXzJ8irpbjHLpVRXDvBEScFJ9b17sgK6vlx0GAJFE21okD7zXl08rRRUfq6HdoEQ==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/browserslist" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/caniuse-lite" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "CC-BY-4.0" + }, + "node_modules/chokidar": { + "version": "3.6.0", + "resolved": "https://registry.npmjs.org/chokidar/-/chokidar-3.6.0.tgz", + "integrity": "sha512-7VT13fmjotKpGipCW9JEQAusEPE+Ei8nl6/g4FBAmIm0GOOLMua9NDDo/DWp0ZAxCr3cPq5ZpBqmPAQgDda2Pw==", + "license": "MIT", + "dependencies": { + "anymatch": "~3.1.2", + "braces": "~3.0.2", + "glob-parent": "~5.1.2", + "is-binary-path": "~2.1.0", + "is-glob": "~4.0.1", + "normalize-path": "~3.0.0", + "readdirp": "~3.6.0" + }, + "engines": { + "node": ">= 8.10.0" + }, + "funding": { + "url": "https://paulmillr.com/funding/" + }, + "optionalDependencies": { + "fsevents": "~2.3.2" + } + }, + "node_modules/chokidar/node_modules/anymatch": { + "version": "3.1.3", + "resolved": "https://registry.npmjs.org/anymatch/-/anymatch-3.1.3.tgz", + "integrity": "sha512-KMReFUr0B4t+D+OBkjR3KYqvocp2XaSzO55UcB6mgQMd3KbcE+mWTyvVV7D/zsdEbNnV6acZUutkiHQXvTr1Rw==", + "license": "ISC", + "dependencies": { + "normalize-path": "^3.0.0", + "picomatch": "^2.0.4" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/chokidar/node_modules/binary-extensions": { + "version": "2.3.0", + "resolved": "https://registry.npmjs.org/binary-extensions/-/binary-extensions-2.3.0.tgz", + "integrity": "sha512-Ceh+7ox5qe7LJuLHoY0feh3pHuUDHAcRUeyL2VYghZwfpkNIy/+8Ocg0a3UuSoYzavmylwuLWQOf3hl0jjMMIw==", + "license": "MIT", + "engines": { + "node": ">=8" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/chokidar/node_modules/braces": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.3.tgz", + "integrity": "sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==", + "license": "MIT", + "dependencies": { + "fill-range": "^7.1.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/chokidar/node_modules/fill-range": { + "version": "7.1.1", + "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.1.1.tgz", + "integrity": "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg==", + "license": "MIT", + "dependencies": { + "to-regex-range": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/chokidar/node_modules/glob-parent": { + "version": "5.1.2", + "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-5.1.2.tgz", + "integrity": "sha512-AOIgSQCepiJYwP3ARnGx+5VnTu2HBYdzbGP45eLw1vr3zB3vZLeyed1sC9hnbcOc9/SrMyM5RPQrkGz4aS9Zow==", + "license": "ISC", + "dependencies": { + "is-glob": "^4.0.1" + }, + "engines": { + "node": ">= 6" + } + }, + "node_modules/chokidar/node_modules/is-binary-path": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/is-binary-path/-/is-binary-path-2.1.0.tgz", + "integrity": "sha512-ZMERYes6pDydyuGidse7OsHxtbI7WVeUEozgR/g7rd0xUimYNlvZRE/K2MgZTjWy725IfelLeVcEM97mmtRGXw==", + "license": "MIT", + "dependencies": { + "binary-extensions": "^2.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/chokidar/node_modules/is-number": { + "version": "7.0.0", + "resolved": "https://registry.npmjs.org/is-number/-/is-number-7.0.0.tgz", + "integrity": "sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==", + "license": "MIT", + "engines": { + "node": ">=0.12.0" + } + }, + "node_modules/chokidar/node_modules/picomatch": { + "version": "2.3.1", + "resolved": "https://registry.npmjs.org/picomatch/-/picomatch-2.3.1.tgz", + "integrity": "sha512-JU3teHTNjmE2VCGFzuY8EXzCDVwEqB2a8fsIvwaStHhAWJEeVd1o1QD80CU6+ZdEXXSLbSsuLwJjkCBWqRQUVA==", + "license": "MIT", + "engines": { + "node": ">=8.6" + }, + "funding": { + "url": "https://github.com/sponsors/jonschlinkert" + } + }, + "node_modules/chokidar/node_modules/readdirp": { + "version": "3.6.0", + "resolved": "https://registry.npmjs.org/readdirp/-/readdirp-3.6.0.tgz", + "integrity": "sha512-hOS089on8RduqdbhvQ5Z37A0ESjsqz6qnRcffsMU3495FuTdqSm+7bhJ29JvIOsBDEEnan5DPu9t3To9VRlMzA==", + "license": "MIT", + "dependencies": { + "picomatch": "^2.2.1" + }, + "engines": { + "node": ">=8.10.0" + } + }, + "node_modules/chokidar/node_modules/to-regex-range": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/to-regex-range/-/to-regex-range-5.0.1.tgz", + "integrity": "sha512-65P7iz6X5yEr1cwcgvQxbbIw7Uk3gOy5dIdtZ4rDveLqhrdJP+Li/Hx6tyK0NEb+2GCyneCMJiGqrADCSNk8sQ==", + "license": "MIT", + "dependencies": { + "is-number": "^7.0.0" + }, + "engines": { + "node": ">=8.0" + } + }, + "node_modules/colorjs.io": { + "version": "0.5.2", + "resolved": "https://registry.npmjs.org/colorjs.io/-/colorjs.io-0.5.2.tgz", + "integrity": "sha512-twmVoizEW7ylZSN32OgKdXRmo1qg+wT5/6C3xu5b9QsWzSFAhHLn2xd8ro0diCsKfCj1RdaTP/nrcW+vAoQPIw==", + "devOptional": true, + "license": "MIT" + }, + "node_modules/commander": { + "version": "8.3.0", + "resolved": "https://registry.npmjs.org/commander/-/commander-8.3.0.tgz", + "integrity": "sha512-OkTL9umf+He2DZkUq8f8J9of7yL6RJKI24dVITBmNfZBmri9zYZQrKkuXiKhyfPSu8tUhnVBB1iKXevvnlR4Ww==", + "license": "MIT", + "engines": { + "node": ">= 12" + } + }, + "node_modules/css-selector-tokenizer": { + "version": "0.8.0", + "resolved": "https://registry.npmjs.org/css-selector-tokenizer/-/css-selector-tokenizer-0.8.0.tgz", + "integrity": "sha512-Jd6Ig3/pe62/qe5SBPTN8h8LeUg/pT4lLgtavPf7updwwHpvFzxvOQBHYj2LZDMjUnBzgvIUSjRcf6oT5HzHFg==", + "license": "MIT", + "dependencies": { + "cssesc": "^3.0.0", + "fastparse": "^1.1.2" + } + }, + "node_modules/css-selector-tokenizer/node_modules/cssesc": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/cssesc/-/cssesc-3.0.0.tgz", + "integrity": "sha512-/Tb/JcjK111nNScGob5MNtsntNM1aCNUDipB/TkwZFhyDrrE47SOx/18wF2bbjgc3ZzCSKW1T5nt5EbFoAz/Vg==", + "license": "MIT", + "bin": { + "cssesc": "bin/cssesc" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/css-selector-tokenizer/node_modules/fastparse": { + "version": "1.1.2", + "resolved": "https://registry.npmjs.org/fastparse/-/fastparse-1.1.2.tgz", + "integrity": "sha512-483XLLxTVIwWK3QTrMGRqUfUpoOs/0hbQrl2oz4J0pAcm3A3bu84wxTFqGqkJzewCLdME38xJLJAxBABfQT8sQ==", + "license": "MIT" + }, + "node_modules/culori": { + "version": "3.3.0", + "resolved": "https://registry.npmjs.org/culori/-/culori-3.3.0.tgz", + "integrity": "sha512-pHJg+jbuFsCjz9iclQBqyL3B2HLCBF71BwVNujUYEvCeQMvV97R59MNK3R2+jgJ3a1fcZgI9B3vYgz8lzr/BFQ==", + "license": "MIT", + "engines": { + "node": "^12.20.0 || ^14.13.1 || >=16.0.0" + } + }, + "node_modules/daisyui": { + "version": "4.12.14", + "resolved": "https://registry.npmjs.org/daisyui/-/daisyui-4.12.14.tgz", + "integrity": "sha512-hA27cdBasdwd4/iEjn+aidoCrRroDuo3G5W9NDKaVCJI437Mm/3eSL/2u7MkZ0pt8a+TrYF3aT2pFVemTS3how==", + "license": "MIT", + "dependencies": { + "css-selector-tokenizer": "^0.8", + "culori": "^3", + "picocolors": "^1", + "postcss-js": "^4" + }, + "engines": { + "node": ">=16.9.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/daisyui" + } + }, + "node_modules/didyoumean": { + "version": "1.2.2", + "resolved": "https://registry.npmjs.org/didyoumean/-/didyoumean-1.2.2.tgz", + "integrity": "sha512-gxtyfqMg7GKyhQmb056K7M3xszy/myH8w+B4RT+QXBQsvAOdc3XymqDDPHx1BgPgsdAA5SIifona89YtRATDzw==", + "license": "Apache-2.0" + }, + "node_modules/dlv": { + "version": "1.1.3", + "resolved": "https://registry.npmjs.org/dlv/-/dlv-1.1.3.tgz", + "integrity": "sha512-+HlytyjlPKnIG8XuRG8WvmBP8xs8P71y+SKKS6ZXWoEgLuePxtDoUEiH7WkdePWrQ5JBpE6aoVqfZfJUQkjXwA==", + "license": "MIT" + }, + "node_modules/entities": { + "version": "4.5.0", + "resolved": "https://registry.npmjs.org/entities/-/entities-4.5.0.tgz", + "integrity": "sha512-V0hjH4dGPh9Ao5p0MoRY6BVqtwCjhz6vI5LT8AJ55H+4g9/4vbHx1I54fS0XuclLhDHArPQCiMjDxjaL8fPxhw==", + "license": "BSD-2-Clause", + "engines": { + "node": ">=0.12" + }, + "funding": { + "url": "https://github.com/fb55/entities?sponsor=1" + } + }, + "node_modules/esbuild": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.21.5.tgz", + "integrity": "sha512-mg3OPMV4hXywwpoDxu3Qda5xCKQi+vCTZq8S9J/EpkhB2HzKXq4SNFZE3+NK93JYxc8VMSep+lOUSC/RVKaBqw==", + "hasInstallScript": true, + "license": "MIT", + "bin": { + "esbuild": "bin/esbuild" + }, + "engines": { + "node": ">=12" + }, + "optionalDependencies": { + "@esbuild/aix-ppc64": "0.21.5", + "@esbuild/android-arm": "0.21.5", + "@esbuild/android-arm64": "0.21.5", + "@esbuild/android-x64": "0.21.5", + "@esbuild/darwin-arm64": "0.21.5", + "@esbuild/darwin-x64": "0.21.5", + "@esbuild/freebsd-arm64": "0.21.5", + "@esbuild/freebsd-x64": "0.21.5", + "@esbuild/linux-arm": "0.21.5", + "@esbuild/linux-arm64": "0.21.5", + "@esbuild/linux-ia32": "0.21.5", + "@esbuild/linux-loong64": "0.21.5", + "@esbuild/linux-mips64el": "0.21.5", + "@esbuild/linux-ppc64": "0.21.5", + "@esbuild/linux-riscv64": "0.21.5", + "@esbuild/linux-s390x": "0.21.5", + "@esbuild/linux-x64": "0.21.5", + "@esbuild/netbsd-x64": "0.21.5", + "@esbuild/openbsd-x64": "0.21.5", + "@esbuild/sunos-x64": "0.21.5", + "@esbuild/win32-arm64": "0.21.5", + "@esbuild/win32-ia32": "0.21.5", + "@esbuild/win32-x64": "0.21.5" + } + }, + "node_modules/esbuild/node_modules/@esbuild/darwin-arm64": { + "version": "0.21.5", + "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.21.5.tgz", + "integrity": "sha512-DwqXqZyuk5AiWWf3UfLiRDJ5EDd49zg6O9wclZ7kUMv2WRFr4HKjXp/5t8JZ11QbQfUS6/cRCKGwYhtNAY88kQ==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=12" + } + }, + "node_modules/fast-glob": { + "version": "3.3.2", + "resolved": "https://registry.npmjs.org/fast-glob/-/fast-glob-3.3.2.tgz", + "integrity": "sha512-oX2ruAFQwf/Orj8m737Y5adxDQO0LAB7/S5MnxCdTNDd4p6BsyIVsv9JQsATbTSq8KHRpLwIHbVlUNatxd+1Ow==", + "license": "MIT", + "dependencies": { + "@nodelib/fs.stat": "^2.0.2", + "@nodelib/fs.walk": "^1.2.3", + "glob-parent": "^5.1.2", + "merge2": "^1.3.0", + "micromatch": "^4.0.4" + }, + "engines": { + "node": ">=8.6.0" + } + }, + "node_modules/fast-glob/node_modules/@nodelib/fs.scandir": { + "version": "2.1.5", + "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz", + "integrity": "sha512-vq24Bq3ym5HEQm2NKCr3yXDwjc7vTsEThRDnkp2DK9p1uqLR+DHurm/NOTo0KG7HYHU7eppKZj3MyqYuMBf62g==", + "license": "MIT", + "dependencies": { + "@nodelib/fs.stat": "2.0.5", + "run-parallel": "^1.1.9" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/fast-glob/node_modules/@nodelib/fs.stat": { + "version": "2.0.5", + "resolved": "https://registry.npmjs.org/@nodelib/fs.stat/-/fs.stat-2.0.5.tgz", + "integrity": "sha512-RkhPPp2zrqDAQA/2jNhnztcPAlv64XdhIp7a7454A5ovI7Bukxgt7MX7udwAu3zg1DcpPU0rz3VV1SeaqvY4+A==", + "license": "MIT", + "engines": { + "node": ">= 8" + } + }, + "node_modules/fast-glob/node_modules/@nodelib/fs.walk": { + "version": "1.2.8", + "resolved": "https://registry.npmjs.org/@nodelib/fs.walk/-/fs.walk-1.2.8.tgz", + "integrity": "sha512-oGB+UxlgWcgQkgwo8GcEGwemoTFt3FIO9ababBmaGwXIoBKZ+GTy0pP185beGg7Llih/NSHSV2XAs1lnznocSg==", + "license": "MIT", + "dependencies": { + "@nodelib/fs.scandir": "2.1.5", + "fastq": "^1.6.0" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/fast-glob/node_modules/fastq": { + "version": "1.17.1", + "resolved": "https://registry.npmjs.org/fastq/-/fastq-1.17.1.tgz", + "integrity": "sha512-sRVD3lWVIXWg6By68ZN7vho9a1pQcN/WBFaAAsDDFzlJjvoGx0P8z7V1t72grFJfJhu3YPZBuu25f7Kaw2jN1w==", + "license": "ISC", + "dependencies": { + "reusify": "^1.0.4" + } + }, + "node_modules/fast-glob/node_modules/glob-parent": { + "version": "5.1.2", + "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-5.1.2.tgz", + "integrity": "sha512-AOIgSQCepiJYwP3ARnGx+5VnTu2HBYdzbGP45eLw1vr3zB3vZLeyed1sC9hnbcOc9/SrMyM5RPQrkGz4aS9Zow==", + "license": "ISC", + "dependencies": { + "is-glob": "^4.0.1" + }, + "engines": { + "node": ">= 6" + } + }, + "node_modules/fast-glob/node_modules/merge2": { + "version": "1.4.1", + "resolved": "https://registry.npmjs.org/merge2/-/merge2-1.4.1.tgz", + "integrity": "sha512-8q7VEgMJW4J8tcfVPy8g09NcQwZdbwFEqhe/WZkoIzjn/3TGDwtOCYtXGxA3O8tPzpczCCDgv+P2P5y00ZJOOg==", + "license": "MIT", + "engines": { + "node": ">= 8" + } + }, + "node_modules/fast-glob/node_modules/queue-microtask": { + "version": "1.2.3", + "resolved": "https://registry.npmjs.org/queue-microtask/-/queue-microtask-1.2.3.tgz", + "integrity": "sha512-NuaNSa6flKT5JaSYQzJok04JzTL1CA6aGhv5rfLW3PgqA+M2ChpZQnAC8h8i4ZFkBS8X5RqkDBHA7r4hej3K9A==", + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/feross" + }, + { + "type": "patreon", + "url": "https://www.patreon.com/feross" + }, + { + "type": "consulting", + "url": "https://feross.org/support" + } + ], + "license": "MIT" + }, + "node_modules/fast-glob/node_modules/reusify": { + "version": "1.0.4", + "resolved": "https://registry.npmjs.org/reusify/-/reusify-1.0.4.tgz", + "integrity": "sha512-U9nH88a3fc/ekCF1l0/UP1IosiuIjyTh7hBvXVMHYgVcfGvt897Xguj2UOLDeI5BG2m7/uwyaLVT6fbtCwTyzw==", + "license": "MIT", + "engines": { + "iojs": ">=1.0.0", + "node": ">=0.10.0" + } + }, + "node_modules/fast-glob/node_modules/run-parallel": { + "version": "1.2.0", + "resolved": "https://registry.npmjs.org/run-parallel/-/run-parallel-1.2.0.tgz", + "integrity": "sha512-5l4VyZR86LZ/lDxZTR6jqL8AFE2S0IFLMP26AbjsLVADxHdhB/c0GUsH+y39UfCi3dzz8OlQuPmnaJOMoDHQBA==", + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/feross" + }, + { + "type": "patreon", + "url": "https://www.patreon.com/feross" + }, + { + "type": "consulting", + "url": "https://feross.org/support" + } + ], + "license": "MIT", + "dependencies": { + "queue-microtask": "^1.2.2" + } + }, + "node_modules/fraction.js": { + "version": "4.3.7", + "resolved": "https://registry.npmjs.org/fraction.js/-/fraction.js-4.3.7.tgz", + "integrity": "sha512-ZsDfxO51wGAXREY55a7la9LScWpwv9RxIrYABrlvOFBlH/ShPnrtsXeuUIfXKKOVicNxQ+o8JTbJvjS4M89yew==", + "license": "MIT", + "engines": { + "node": "*" + }, + "funding": { + "type": "patreon", + "url": "https://github.com/sponsors/rawify" + } + }, + "node_modules/fsevents": { + "version": "2.3.3", + "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz", + "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==", + "hasInstallScript": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": "^8.16.0 || ^10.6.0 || >=11.0.0" + } + }, + "node_modules/glob-parent": { + "version": "6.0.2", + "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-6.0.2.tgz", + "integrity": "sha512-XxwI8EOhVQgWp6iDL+3b0r86f4d6AX6zSU55HfB4ydCEuXLXc5FcYeOu+nnGftS4TEju/11rt4KJPTMgbfmv4A==", + "license": "ISC", + "dependencies": { + "is-glob": "^4.0.3" + }, + "engines": { + "node": ">=10.13.0" + } + }, + "node_modules/has-flag": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/has-flag/-/has-flag-4.0.0.tgz", + "integrity": "sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==", + "devOptional": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/highlight.js": { + "version": "11.10.0", + "resolved": "https://registry.npmjs.org/highlight.js/-/highlight.js-11.10.0.tgz", + "integrity": "sha512-SYVnVFswQER+zu1laSya563s+F8VDGt7o35d4utbamowvUNLLMovFqwCLSocpZTz3MgaSRA1IbqRWZv97dtErQ==", + "engines": { + "node": ">=12.0.0" + } + }, + "node_modules/immutable": { + "version": "5.0.3", + "resolved": "https://registry.npmjs.org/immutable/-/immutable-5.0.3.tgz", + "integrity": "sha512-P8IdPQHq3lA1xVeBRi5VPqUm5HDgKnx0Ru51wZz5mjxHr5n3RWhjIpOFU7ybkUxfB+5IToy+OLaHYDBIWsv+uw==", + "devOptional": true, + "license": "MIT" + }, + "node_modules/is-glob": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/is-glob/-/is-glob-4.0.3.tgz", + "integrity": "sha512-xelSayHH36ZgE7ZWhli7pW34hNbNl8Ojv5KVmkJD4hBdD3th8Tfk9vYasLM+mXWOZhFkgZfxhLSnrwRr4elSSg==", + "license": "MIT", + "dependencies": { + "is-extglob": "^2.1.1" + }, + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/is-glob/node_modules/is-extglob": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/is-extglob/-/is-extglob-2.1.1.tgz", + "integrity": "sha512-SbKbANkN603Vi4jEZv49LeVJMn4yGwsbzZworEoyEiutsN3nJYdbO36zfhGJ6QEDpOZIFkDtnq5JRxmvl3jsoQ==", + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/jiti": { + "version": "1.21.6", + "resolved": "https://registry.npmjs.org/jiti/-/jiti-1.21.6.tgz", + "integrity": "sha512-2yTgeWTWzMWkHu6Jp9NKgePDaYHbntiwvYuuJLbbN9vl7DC9DvXKOB2BC3ZZ92D3cvV/aflH0osDfwpHepQ53w==", + "license": "MIT", + "bin": { + "jiti": "bin/jiti.js" + } + }, + "node_modules/katex": { + "version": "0.16.15", + "resolved": "https://registry.npmjs.org/katex/-/katex-0.16.15.tgz", + "integrity": "sha512-yE9YJIEAk2aZ+FL/G8r+UGw0CTUzEA8ZFy6E+8tc3spHUKq3qBnzCkI1CQwGoI9atJhVyFPEypQsTY7mJ1Pi9w==", + "funding": [ + "https://opencollective.com/katex", + "https://github.com/sponsors/katex" + ], + "license": "MIT", + "dependencies": { + "commander": "^8.3.0" + }, + "bin": { + "katex": "cli.js" + } + }, + "node_modules/lilconfig": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/lilconfig/-/lilconfig-2.1.0.tgz", + "integrity": "sha512-utWOt/GHzuUxnLKxB6dk81RoOeoNeHgbrXiuGk4yyF5qlRz+iIVWu56E2fqGHFrXz0QNUhLB/8nKqvRH66JKGQ==", + "license": "MIT", + "engines": { + "node": ">=10" + } + }, + "node_modules/linkify-it": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/linkify-it/-/linkify-it-5.0.0.tgz", + "integrity": "sha512-5aHCbzQRADcdP+ATqnDuhhJ/MRIqDkZX5pyjFHRRysS8vZ5AbqGEoFIb6pYHPZ+L/OC2Lc+xT8uHVVR5CAK/wQ==", + "license": "MIT", + "dependencies": { + "uc.micro": "^2.0.0" + } + }, + "node_modules/markdown-it": { + "version": "14.1.0", + "resolved": "https://registry.npmjs.org/markdown-it/-/markdown-it-14.1.0.tgz", + "integrity": "sha512-a54IwgWPaeBCAAsv13YgmALOF1elABB08FxO9i+r4VFk5Vl4pKokRPeX8u5TCgSsPi6ec1otfLjdOpVcgbpshg==", + "license": "MIT", + "dependencies": { + "argparse": "^2.0.1", + "entities": "^4.4.0", + "linkify-it": "^5.0.0", + "mdurl": "^2.0.0", + "punycode.js": "^2.3.1", + "uc.micro": "^2.1.0" + }, + "bin": { + "markdown-it": "bin/markdown-it.mjs" + } + }, + "node_modules/mdurl": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/mdurl/-/mdurl-2.0.0.tgz", + "integrity": "sha512-Lf+9+2r+Tdp5wXDXC4PcIBjTDtq4UKjCPMQhKIuzpJNW0b96kVqSwW0bT7FhRSfmAiFYgP+SCRvdrDozfh0U5w==", + "license": "MIT" + }, + "node_modules/micromatch": { + "version": "4.0.8", + "resolved": "https://registry.npmjs.org/micromatch/-/micromatch-4.0.8.tgz", + "integrity": "sha512-PXwfBhYu0hBCPw8Dn0E+WDYb7af3dSLVWKi3HGv84IdF4TyFoC0ysxFd0Goxw7nSv4T/PzEJQxsYsEiFCKo2BA==", + "license": "MIT", + "dependencies": { + "braces": "^3.0.3", + "picomatch": "^2.3.1" + }, + "engines": { + "node": ">=8.6" + } + }, + "node_modules/micromatch/node_modules/braces": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.3.tgz", + "integrity": "sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==", + "license": "MIT", + "dependencies": { + "fill-range": "^7.1.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/micromatch/node_modules/fill-range": { + "version": "7.1.1", + "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.1.1.tgz", + "integrity": "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg==", + "license": "MIT", + "dependencies": { + "to-regex-range": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/micromatch/node_modules/is-number": { + "version": "7.0.0", + "resolved": "https://registry.npmjs.org/is-number/-/is-number-7.0.0.tgz", + "integrity": "sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==", + "license": "MIT", + "engines": { + "node": ">=0.12.0" + } + }, + "node_modules/micromatch/node_modules/picomatch": { + "version": "2.3.1", + "resolved": "https://registry.npmjs.org/picomatch/-/picomatch-2.3.1.tgz", + "integrity": "sha512-JU3teHTNjmE2VCGFzuY8EXzCDVwEqB2a8fsIvwaStHhAWJEeVd1o1QD80CU6+ZdEXXSLbSsuLwJjkCBWqRQUVA==", + "license": "MIT", + "engines": { + "node": ">=8.6" + }, + "funding": { + "url": "https://github.com/sponsors/jonschlinkert" + } + }, + "node_modules/micromatch/node_modules/to-regex-range": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/to-regex-range/-/to-regex-range-5.0.1.tgz", + "integrity": "sha512-65P7iz6X5yEr1cwcgvQxbbIw7Uk3gOy5dIdtZ4rDveLqhrdJP+Li/Hx6tyK0NEb+2GCyneCMJiGqrADCSNk8sQ==", + "license": "MIT", + "dependencies": { + "is-number": "^7.0.0" + }, + "engines": { + "node": ">=8.0" + } + }, + "node_modules/normalize-path": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/normalize-path/-/normalize-path-3.0.0.tgz", + "integrity": "sha512-6eZs5Ls3WtCisHWp9S2GUy8dqkpGi4BVSz3GaqiE6ezub0512ESztXUwUB6C6IKbQkY2Pnb/mD4WYojCRwcwLA==", + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/normalize-range": { + "version": "0.1.2", + "resolved": "https://registry.npmjs.org/normalize-range/-/normalize-range-0.1.2.tgz", + "integrity": "sha512-bdok/XvKII3nUpklnV6P2hxtMNrCboOjAcyBuQnWEhO665FwrSNRxU+AqpsyvO6LgGYPspN+lu5CLtw4jPRKNA==", + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/object-hash": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/object-hash/-/object-hash-3.0.0.tgz", + "integrity": "sha512-RSn9F68PjH9HqtltsSnqYC1XXoWe9Bju5+213R98cNGttag9q9yAOTzdbsqvIa7aNm5WffBZFpWYr2aWrklWAw==", + "license": "MIT", + "engines": { + "node": ">= 6" + } + }, + "node_modules/picocolors": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/picocolors/-/picocolors-1.1.1.tgz", + "integrity": "sha512-xceH2snhtb5M9liqDsmEw56le376mTZkEX/jEb/RxNFyegNul7eNslCXP9FDj/Lcu0X8KEyMceP2ntpaHrDEVA==", + "license": "ISC" + }, + "node_modules/postcss": { + "version": "8.4.49", + "resolved": "https://registry.npmjs.org/postcss/-/postcss-8.4.49.tgz", + "integrity": "sha512-OCVPnIObs4N29kxTjzLfUryOkvZEq+pf8jTF0lg8E7uETuWHA+v7j3c/xJmiqpX450191LlmZfUKkXxkTry7nA==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/postcss" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "nanoid": "^3.3.7", + "picocolors": "^1.1.1", + "source-map-js": "^1.2.1" + }, + "engines": { + "node": "^10 || ^12 || >=14" + } + }, + "node_modules/postcss-import": { + "version": "15.1.0", + "resolved": "https://registry.npmjs.org/postcss-import/-/postcss-import-15.1.0.tgz", + "integrity": "sha512-hpr+J05B2FVYUAXHeK1YyI267J/dDDhMU6B6civm8hSY1jYJnBXxzKDKDswzJmtLHryrjhnDjqqp/49t8FALew==", + "license": "MIT", + "dependencies": { + "postcss-value-parser": "^4.0.0", + "read-cache": "^1.0.0", + "resolve": "^1.1.7" + }, + "engines": { + "node": ">=14.0.0" + }, + "peerDependencies": { + "postcss": "^8.0.0" + } + }, + "node_modules/postcss-import/node_modules/pify": { + "version": "2.3.0", + "resolved": "https://registry.npmjs.org/pify/-/pify-2.3.0.tgz", + "integrity": "sha512-udgsAY+fTnvv7kI7aaxbqwWNb0AHiB0qBO89PZKPkoTmGOgdbrHDKD+0B2X4uTfJ/FT1R09r9gTsjUjNJotuog==", + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/postcss-import/node_modules/read-cache": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/read-cache/-/read-cache-1.0.0.tgz", + "integrity": "sha512-Owdv/Ft7IjOgm/i0xvNDZ1LrRANRfew4b2prF3OWMQLxLfu3bS8FVhCsrSCMK4lR56Y9ya+AThoTpDCTxCmpRA==", + "license": "MIT", + "dependencies": { + "pify": "^2.3.0" + } + }, + "node_modules/postcss-js": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/postcss-js/-/postcss-js-4.0.1.tgz", + "integrity": "sha512-dDLF8pEO191hJMtlHFPRa8xsizHaM82MLfNkUHdUtVEV3tgTp5oj+8qbEqYM57SLfc74KSbw//4SeJma2LRVIw==", + "license": "MIT", + "dependencies": { + "camelcase-css": "^2.0.1" + }, + "engines": { + "node": "^12 || ^14 || >= 16" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + "peerDependencies": { + "postcss": "^8.4.21" + } + }, + "node_modules/postcss-js/node_modules/camelcase-css": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/camelcase-css/-/camelcase-css-2.0.1.tgz", + "integrity": "sha512-QOSvevhslijgYwRx6Rv7zKdMF8lbRmx+uQGx2+vDc+KI/eBnsy9kit5aj23AgGu3pa4t9AgwbnXWqS+iOY+2aA==", + "license": "MIT", + "engines": { + "node": ">= 6" + } + }, + "node_modules/postcss-load-config": { + "version": "4.0.2", + "resolved": "https://registry.npmjs.org/postcss-load-config/-/postcss-load-config-4.0.2.tgz", + "integrity": "sha512-bSVhyJGL00wMVoPUzAVAnbEoWyqRxkjv64tUl427SKnPrENtq6hJwUojroMz2VB+Q1edmi4IfrAPpami5VVgMQ==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "lilconfig": "^3.0.0", + "yaml": "^2.3.4" + }, + "engines": { + "node": ">= 14" + }, + "peerDependencies": { + "postcss": ">=8.0.9", + "ts-node": ">=9.0.0" + }, + "peerDependenciesMeta": { + "postcss": { + "optional": true + }, + "ts-node": { + "optional": true + } + } + }, + "node_modules/postcss-load-config/node_modules/lilconfig": { + "version": "3.1.2", + "resolved": "https://registry.npmjs.org/lilconfig/-/lilconfig-3.1.2.tgz", + "integrity": "sha512-eop+wDAvpItUys0FWkHIKeC9ybYrTGbU41U5K7+bttZZeohvnY7M9dZ5kB21GNWiFT2q1OoPTvncPCgSOVO5ow==", + "license": "MIT", + "engines": { + "node": ">=14" + }, + "funding": { + "url": "https://github.com/sponsors/antonk52" + } + }, + "node_modules/postcss-load-config/node_modules/yaml": { + "version": "2.6.1", + "resolved": "https://registry.npmjs.org/yaml/-/yaml-2.6.1.tgz", + "integrity": "sha512-7r0XPzioN/Q9kXBro/XPnA6kznR73DHq+GXh5ON7ZozRO6aMjbmiBuKste2wslTFkC5d1dw0GooOCepZXJ2SAg==", + "license": "ISC", + "bin": { + "yaml": "bin.mjs" + }, + "engines": { + "node": ">= 14" + } + }, + "node_modules/postcss-nested": { + "version": "6.2.0", + "resolved": "https://registry.npmjs.org/postcss-nested/-/postcss-nested-6.2.0.tgz", + "integrity": "sha512-HQbt28KulC5AJzG+cZtj9kvKB93CFCdLvog1WFLf1D+xmMvPGlBstkpTEZfK5+AN9hfJocyBFCNiqyS48bpgzQ==", + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "postcss-selector-parser": "^6.1.1" + }, + "engines": { + "node": ">=12.0" + }, + "peerDependencies": { + "postcss": "^8.2.14" + } + }, + "node_modules/postcss-selector-parser": { + "version": "6.1.2", + "resolved": "https://registry.npmjs.org/postcss-selector-parser/-/postcss-selector-parser-6.1.2.tgz", + "integrity": "sha512-Q8qQfPiZ+THO/3ZrOrO0cJJKfpYCagtMUkXbnEfmgUjwXg6z/WBeOyS9APBBPCTSiDV+s4SwQGu8yFsiMRIudg==", + "license": "MIT", + "dependencies": { + "cssesc": "^3.0.0", + "util-deprecate": "^1.0.2" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/postcss-selector-parser/node_modules/cssesc": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/cssesc/-/cssesc-3.0.0.tgz", + "integrity": "sha512-/Tb/JcjK111nNScGob5MNtsntNM1aCNUDipB/TkwZFhyDrrE47SOx/18wF2bbjgc3ZzCSKW1T5nt5EbFoAz/Vg==", + "license": "MIT", + "bin": { + "cssesc": "bin/cssesc" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/postcss-selector-parser/node_modules/util-deprecate": { + "version": "1.0.2", + "resolved": "https://registry.npmjs.org/util-deprecate/-/util-deprecate-1.0.2.tgz", + "integrity": "sha512-EPD5q1uXyFxJpCrLnCc1nHnq3gOa6DZBocAIiI2TaSCA7VCJ1UJDMagCzIkXNsUYfD1daK//LTEQ8xiIbrHtcw==", + "license": "MIT" + }, + "node_modules/postcss-value-parser": { + "version": "4.2.0", + "resolved": "https://registry.npmjs.org/postcss-value-parser/-/postcss-value-parser-4.2.0.tgz", + "integrity": "sha512-1NNCs6uurfkVbeXG4S8JFT9t19m45ICnif8zWLd5oPSZ50QnwMfK+H3jv408d4jw/7Bttv5axS5IiHoLaVNHeQ==", + "license": "MIT" + }, + "node_modules/postcss/node_modules/nanoid": { + "version": "3.3.8", + "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-3.3.8.tgz", + "integrity": "sha512-WNLf5Sd8oZxOm+TzppcYk8gVOgP+l58xNy58D0nbUnOxOWRWvlcCV4kUF7ltmI6PsrLl/BgKEyS4mqsGChFN0w==", + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "bin": { + "nanoid": "bin/nanoid.cjs" + }, + "engines": { + "node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1" + } + }, + "node_modules/postcss/node_modules/source-map-js": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz", + "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==", + "license": "BSD-3-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/punycode.js": { + "version": "2.3.1", + "resolved": "https://registry.npmjs.org/punycode.js/-/punycode.js-2.3.1.tgz", + "integrity": "sha512-uxFIHU0YlHYhDQtV4R9J6a52SLx28BCjT+4ieh7IGbgwVJWO+km431c4yRlREUAsAmt/uMjQUyQHNEPf0M39CA==", + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/resolve": { + "version": "1.22.8", + "resolved": "https://registry.npmjs.org/resolve/-/resolve-1.22.8.tgz", + "integrity": "sha512-oKWePCxqpd6FlLvGV1VU0x7bkPmmCNolxzjMf4NczoDnQcIWrAF+cPtZn5i6n+RfD2d9i0tzpKnG6Yk168yIyw==", + "license": "MIT", + "dependencies": { + "is-core-module": "^2.13.0", + "path-parse": "^1.0.7", + "supports-preserve-symlinks-flag": "^1.0.0" + }, + "bin": { + "resolve": "bin/resolve" + }, + "funding": { + "url": "https://github.com/sponsors/ljharb" + } + }, + "node_modules/resolve/node_modules/function-bind": { + "version": "1.1.2", + "resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.2.tgz", + "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==", + "license": "MIT", + "funding": { + "url": "https://github.com/sponsors/ljharb" + } + }, + "node_modules/resolve/node_modules/hasown": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.2.tgz", + "integrity": "sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==", + "license": "MIT", + "dependencies": { + "function-bind": "^1.1.2" + }, + "engines": { + "node": ">= 0.4" + } + }, + "node_modules/resolve/node_modules/is-core-module": { + "version": "2.15.1", + "resolved": "https://registry.npmjs.org/is-core-module/-/is-core-module-2.15.1.tgz", + "integrity": "sha512-z0vtXSwucUJtANQWldhbtbt7BnL0vxiFjIdDLAatwhDYty2bad6s+rijD6Ri4YuYJubLzIJLUidCh09e1djEVQ==", + "license": "MIT", + "dependencies": { + "hasown": "^2.0.2" + }, + "engines": { + "node": ">= 0.4" + }, + "funding": { + "url": "https://github.com/sponsors/ljharb" + } + }, + "node_modules/resolve/node_modules/path-parse": { + "version": "1.0.7", + "resolved": "https://registry.npmjs.org/path-parse/-/path-parse-1.0.7.tgz", + "integrity": "sha512-LDJzPVEEEPR+y48z93A0Ed0yXb8pAByGWo/k5YYdYgpY2/2EsOsksJrq7lOHxryrVOn1ejG6oAp8ahvOIQD8sw==", + "license": "MIT" + }, + "node_modules/resolve/node_modules/supports-preserve-symlinks-flag": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/supports-preserve-symlinks-flag/-/supports-preserve-symlinks-flag-1.0.0.tgz", + "integrity": "sha512-ot0WnXS9fgdkgIcePe6RHNk1WA8+muPa6cSjeR3V8K27q9BB1rTE3R1p7Hv0z1ZyAc8s6Vvv8DIyWf681MAt0w==", + "license": "MIT", + "engines": { + "node": ">= 0.4" + }, + "funding": { + "url": "https://github.com/sponsors/ljharb" + } + }, + "node_modules/rollup": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/rollup/-/rollup-4.28.0.tgz", + "integrity": "sha512-G9GOrmgWHBma4YfCcX8PjH0qhXSdH8B4HDE2o4/jaxj93S4DPCIDoLcXz99eWMji4hB29UFCEd7B2gwGJDR9cQ==", + "license": "MIT", + "dependencies": { + "@types/estree": "1.0.6" + }, + "bin": { + "rollup": "dist/bin/rollup" + }, + "engines": { + "node": ">=18.0.0", + "npm": ">=8.0.0" + }, + "optionalDependencies": { + "@rollup/rollup-android-arm-eabi": "4.28.0", + "@rollup/rollup-android-arm64": "4.28.0", + "@rollup/rollup-darwin-arm64": "4.28.0", + "@rollup/rollup-darwin-x64": "4.28.0", + "@rollup/rollup-freebsd-arm64": "4.28.0", + "@rollup/rollup-freebsd-x64": "4.28.0", + "@rollup/rollup-linux-arm-gnueabihf": "4.28.0", + "@rollup/rollup-linux-arm-musleabihf": "4.28.0", + "@rollup/rollup-linux-arm64-gnu": "4.28.0", + "@rollup/rollup-linux-arm64-musl": "4.28.0", + "@rollup/rollup-linux-powerpc64le-gnu": "4.28.0", + "@rollup/rollup-linux-riscv64-gnu": "4.28.0", + "@rollup/rollup-linux-s390x-gnu": "4.28.0", + "@rollup/rollup-linux-x64-gnu": "4.28.0", + "@rollup/rollup-linux-x64-musl": "4.28.0", + "@rollup/rollup-win32-arm64-msvc": "4.28.0", + "@rollup/rollup-win32-ia32-msvc": "4.28.0", + "@rollup/rollup-win32-x64-msvc": "4.28.0", + "fsevents": "~2.3.2" + } + }, + "node_modules/rollup/node_modules/@rollup/rollup-darwin-arm64": { + "version": "4.28.0", + "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-arm64/-/rollup-darwin-arm64-4.28.0.tgz", + "integrity": "sha512-lmKx9yHsppblnLQZOGxdO66gT77bvdBtr/0P+TPOseowE7D9AJoBw8ZDULRasXRWf1Z86/gcOdpBrV6VDUY36Q==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ] + }, + "node_modules/rollup/node_modules/@types/estree": { + "version": "1.0.6", + "resolved": "https://registry.npmjs.org/@types/estree/-/estree-1.0.6.tgz", + "integrity": "sha512-AYnb1nQyY49te+VRAVgmzfcgjYS91mY5P0TKUDCLEM+gNnA+3T6rWITXRLYCpahpqSQbN5cE+gHpnPyXjHWxcw==", + "license": "MIT" + }, + "node_modules/rxjs": { + "version": "7.8.1", + "resolved": "https://registry.npmjs.org/rxjs/-/rxjs-7.8.1.tgz", + "integrity": "sha512-AA3TVj+0A2iuIoQkWEK/tqFjBq2j+6PO6Y0zJcvzLAFhEFIO3HL0vls9hWLncZbAAbK0mar7oZ4V079I/qPMxg==", + "devOptional": true, + "license": "Apache-2.0", + "dependencies": { + "tslib": "^2.1.0" + } + }, + "node_modules/sass-embedded": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded/-/sass-embedded-1.83.0.tgz", + "integrity": "sha512-/8cYZeL39evUqe0o//193na51Q1VWZ61qhxioQvLJwOtWIrX+PgNhCyD8RSuTtmzc4+6+waFZf899bfp/MCUwA==", + "devOptional": true, + "license": "MIT", + "dependencies": { + "@bufbuild/protobuf": "^2.0.0", + "buffer-builder": "^0.2.0", + "colorjs.io": "^0.5.0", + "immutable": "^5.0.2", + "rxjs": "^7.4.0", + "supports-color": "^8.1.1", + "sync-child-process": "^1.0.2", + "varint": "^6.0.0" + }, + "bin": { + "sass": "dist/bin/sass.js" + }, + "engines": { + "node": ">=16.0.0" + }, + "optionalDependencies": { + "sass-embedded-android-arm": "1.83.0", + "sass-embedded-android-arm64": "1.83.0", + "sass-embedded-android-ia32": "1.83.0", + "sass-embedded-android-riscv64": "1.83.0", + "sass-embedded-android-x64": "1.83.0", + "sass-embedded-darwin-arm64": "1.83.0", + "sass-embedded-darwin-x64": "1.83.0", + "sass-embedded-linux-arm": "1.83.0", + "sass-embedded-linux-arm64": "1.83.0", + "sass-embedded-linux-ia32": "1.83.0", + "sass-embedded-linux-musl-arm": "1.83.0", + "sass-embedded-linux-musl-arm64": "1.83.0", + "sass-embedded-linux-musl-ia32": "1.83.0", + "sass-embedded-linux-musl-riscv64": "1.83.0", + "sass-embedded-linux-musl-x64": "1.83.0", + "sass-embedded-linux-riscv64": "1.83.0", + "sass-embedded-linux-x64": "1.83.0", + "sass-embedded-win32-arm64": "1.83.0", + "sass-embedded-win32-ia32": "1.83.0", + "sass-embedded-win32-x64": "1.83.0" + } + }, + "node_modules/sass-embedded-android-arm": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-android-arm/-/sass-embedded-android-arm-1.83.0.tgz", + "integrity": "sha512-uwFSXzJlfbd4Px189xE5l+cxN8+TQpXdQgJec7TIrb4HEY7imabtpYufpVdqUVwT1/uiis5V4+qIEC4Vl5XObQ==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-android-arm64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-android-arm64/-/sass-embedded-android-arm64-1.83.0.tgz", + "integrity": "sha512-GBiCvM4a2rkWBLdYDxI6XYnprfk5U5c81g69RC2X6kqPuzxzx8qTArQ9M6keFK4+iDQ5N9QTwFCr0KbZTn+ZNQ==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-android-ia32": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-android-ia32/-/sass-embedded-android-ia32-1.83.0.tgz", + "integrity": "sha512-5ATPdGo2SICqAhiJl/Z8KQ23zH4sGgobGgux0TnrNtt83uHZ+r+To/ubVJ7xTkZxed+KJZnIpolGD8dQyQqoTg==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-android-riscv64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-android-riscv64/-/sass-embedded-android-riscv64-1.83.0.tgz", + "integrity": "sha512-aveknUOB8GZewOzVn2Uwk+DKcncTR50Q6vtzslNMGbYnxtgQNHzy8A1qVEviNUruex+pHofppeMK4iMPFAbiEQ==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-android-x64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-android-x64/-/sass-embedded-android-x64-1.83.0.tgz", + "integrity": "sha512-WqIay/72ncyf9Ph4vS742J3a73wZihWmzFUwpn1OD6lme1Aj4eWzWIve5IVnlTEJgcZcDHu6ECID9IZgehJKoA==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-darwin-arm64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-darwin-arm64/-/sass-embedded-darwin-arm64-1.83.0.tgz", + "integrity": "sha512-XQl9QqgxFFIPm/CzHhmppse5o9ocxrbaAdC2/DAnlAqvYWBBtgFqPjGoYlej13h9SzfvNoogx+y9r+Ap+e+hYg==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-darwin-x64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-darwin-x64/-/sass-embedded-darwin-x64-1.83.0.tgz", + "integrity": "sha512-ERQ7Tvp1kFOW3ux4VDFIxb7tkYXHYc+zJpcrbs0hzcIO5ilIRU2tIOK1OrNwrFO6Qxyf7AUuBwYKLAtIU/Nz7g==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-arm": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-arm/-/sass-embedded-linux-arm-1.83.0.tgz", + "integrity": "sha512-baG9RYBJxUFmqwDNC9h9ZFElgJoyO3jgHGjzEZ1wHhIS9anpG+zZQvO8bHx3dBpKEImX+DBeLX+CxsFR9n81gQ==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-arm64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-arm64/-/sass-embedded-linux-arm64-1.83.0.tgz", + "integrity": "sha512-syEAVTJt4qhaMLxrSwOWa46zdqHJdnqJkLUK+t9aCr8xqBZLPxSUeIGji76uOehQZ1C+KGFj6n9xstHN6wzOJw==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-ia32": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-ia32/-/sass-embedded-linux-ia32-1.83.0.tgz", + "integrity": "sha512-RRBxQxMpoxu5+XcSSc6QR/o9asEwUzR8AbCS83RaXcdTIHTa/CccQsiAoDDoPlRsMTLqnzs0LKL4CfOsf7zBbA==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-musl-arm": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-arm/-/sass-embedded-linux-musl-arm-1.83.0.tgz", + "integrity": "sha512-Yc7u2TelCfBab+PRob9/MNJFh3EooMiz4urvhejXkihTiKSHGCv5YqDdtWzvyb9tY2Jb7YtYREVuHwfdVn3dTQ==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-musl-arm64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-arm64/-/sass-embedded-linux-musl-arm64-1.83.0.tgz", + "integrity": "sha512-Y7juhPHClUO2H5O+u+StRy6SEAcwZ+hTEk5WJdEmo1Bb1gDtfHvJaWB/iFZJ2tW0W1e865AZeUrC4OcOFjyAQA==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-musl-ia32": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-ia32/-/sass-embedded-linux-musl-ia32-1.83.0.tgz", + "integrity": "sha512-arQeYwGmwXV8byx5G1PtSzZWW1jbkfR5qrIHMEbTFSAvAxpqjgSvCvrHMOFd73FcMxVaYh4BX9LQNbKinkbEdg==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-musl-riscv64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-riscv64/-/sass-embedded-linux-musl-riscv64-1.83.0.tgz", + "integrity": "sha512-E6uzlIWz59rut+Z3XR6mLG915zNzv07ISvj3GUNZENdHM7dF8GQ//ANoIpl5PljMQKp89GnYdvo6kj2gnaBf/g==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-musl-x64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-x64/-/sass-embedded-linux-musl-x64-1.83.0.tgz", + "integrity": "sha512-eAMK6tyGqvqr21r9g8BnR3fQc1rYFj85RGduSQ3xkITZ6jOAnOhuU94N5fwRS852Hpws0lXhET+7JHXgg3U18w==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-riscv64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-riscv64/-/sass-embedded-linux-riscv64-1.83.0.tgz", + "integrity": "sha512-Ojpi78pTv02sy2fUYirRGXHLY3fPnV/bvwuC2i5LwPQw2LpCcFyFTtN0c5h4LJDk9P6wr+/ZB/JXU8tHIOlK+Q==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-linux-x64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-linux-x64/-/sass-embedded-linux-x64-1.83.0.tgz", + "integrity": "sha512-3iLjlXdoPfgZRtX4odhRvka1BQs5mAXqfCtDIQBgh/o0JnGPzJIWWl9bYLpHxK8qb+uyVBxXYgXpI0sCzArBOw==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-win32-arm64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-win32-arm64/-/sass-embedded-win32-arm64-1.83.0.tgz", + "integrity": "sha512-iOHw/8/t2dlTW3lOFwG5eUbiwhEyGWawivlKWJ8lkXH7fjMpVx2VO9zCFAm8RvY9xOHJ9sf1L7g5bx3EnNP9BQ==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-win32-ia32": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-win32-ia32/-/sass-embedded-win32-ia32-1.83.0.tgz", + "integrity": "sha512-2PxNXJ8Pad4geVcTXY4rkyTr5AwbF8nfrCTDv0ulbTvPhzX2mMKEGcBZUXWn5BeHZTBc6whNMfS7d5fQXR9dDQ==", + "cpu": [ + "ia32" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sass-embedded-win32-x64": { + "version": "1.83.0", + "resolved": "https://registry.npmjs.org/sass-embedded-win32-x64/-/sass-embedded-win32-x64-1.83.0.tgz", + "integrity": "sha512-muBXkFngM6eLTNqOV0FQi7Dv9s+YRQ42Yem26mosdan/GmJQc81deto6uDTgrYn+bzFNmiXcOdfm+0MkTWK3OQ==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/sucrase": { + "version": "3.35.0", + "resolved": "https://registry.npmjs.org/sucrase/-/sucrase-3.35.0.tgz", + "integrity": "sha512-8EbVDiu9iN/nESwxeSxDKe0dunta1GOlHufmSSXxMD2z2/tMZpDMpvXQGsc+ajGo8y2uYUmixaSRUc/QPoQ0GA==", + "license": "MIT", + "dependencies": { + "@jridgewell/gen-mapping": "^0.3.2", + "commander": "^4.0.0", + "glob": "^10.3.10", + "lines-and-columns": "^1.1.6", + "mz": "^2.7.0", + "pirates": "^4.0.1", + "ts-interface-checker": "^0.1.9" + }, + "bin": { + "sucrase": "bin/sucrase", + "sucrase-node": "bin/sucrase-node" + }, + "engines": { + "node": ">=16 || 14 >=14.17" + } + }, + "node_modules/sucrase/node_modules/@isaacs/cliui": { + "version": "8.0.2", + "resolved": "https://registry.npmjs.org/@isaacs/cliui/-/cliui-8.0.2.tgz", + "integrity": "sha512-O8jcjabXaleOG9DQ0+ARXWZBTfnP4WNAqzuiJK7ll44AmxGKv/J2M4TPjxjY3znBCfvBXFzucm1twdyFybFqEA==", + "license": "ISC", + "dependencies": { + "string-width": "^5.1.2", + "string-width-cjs": "npm:string-width@^4.2.0", + "strip-ansi": "^7.0.1", + "strip-ansi-cjs": "npm:strip-ansi@^6.0.1", + "wrap-ansi": "^8.1.0", + "wrap-ansi-cjs": "npm:wrap-ansi@^7.0.0" + }, + "engines": { + "node": ">=12" + } + }, + "node_modules/sucrase/node_modules/@jridgewell/gen-mapping": { + "version": "0.3.5", + "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.5.tgz", + "integrity": "sha512-IzL8ZoEDIBRWEzlCcRhOaCupYyN5gdIK+Q6fbFdPDg6HqX6jpkItn7DFIpW9LQzXG6Df9sA7+OKnq0qlz/GaQg==", + "license": "MIT", + "dependencies": { + "@jridgewell/set-array": "^1.2.1", + "@jridgewell/sourcemap-codec": "^1.4.10", + "@jridgewell/trace-mapping": "^0.3.24" + }, + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/sucrase/node_modules/@jridgewell/resolve-uri": { + "version": "3.1.2", + "resolved": "https://registry.npmjs.org/@jridgewell/resolve-uri/-/resolve-uri-3.1.2.tgz", + "integrity": "sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==", + "license": "MIT", + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/sucrase/node_modules/@jridgewell/set-array": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/@jridgewell/set-array/-/set-array-1.2.1.tgz", + "integrity": "sha512-R8gLRTZeyp03ymzP/6Lil/28tGeGEzhx1q2k703KGWRAI1VdvPIXdG70VJc2pAMw3NA6JKL5hhFu1sJX0Mnn/A==", + "license": "MIT", + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/sucrase/node_modules/@jridgewell/sourcemap-codec": { + "version": "1.5.0", + "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.0.tgz", + "integrity": "sha512-gv3ZRaISU3fjPAgNsriBRqGWQL6quFx04YMPW/zD8XMLsU32mhCCbfbO6KZFLjvYpCZ8zyDEgqsgf+PwPaM7GQ==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/@jridgewell/trace-mapping": { + "version": "0.3.25", + "resolved": "https://registry.npmjs.org/@jridgewell/trace-mapping/-/trace-mapping-0.3.25.tgz", + "integrity": "sha512-vNk6aEwybGtawWmy/PzwnGDOjCkLWSD2wqvjGGAgOAwCGWySYXfYoxt00IJkTF+8Lb57DwOb3Aa0o9CApepiYQ==", + "license": "MIT", + "dependencies": { + "@jridgewell/resolve-uri": "^3.1.0", + "@jridgewell/sourcemap-codec": "^1.4.14" + } + }, + "node_modules/sucrase/node_modules/@pkgjs/parseargs": { + "version": "0.11.0", + "resolved": "https://registry.npmjs.org/@pkgjs/parseargs/-/parseargs-0.11.0.tgz", + "integrity": "sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg==", + "license": "MIT", + "optional": true, + "engines": { + "node": ">=14" + } + }, + "node_modules/sucrase/node_modules/ansi-regex": { + "version": "6.1.0", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-6.1.0.tgz", + "integrity": "sha512-7HSX4QQb4CspciLpVFwyRe79O3xsIZDDLER21kERQ71oaPodF8jL725AgJMFAYbooIqolJoRLuM81SpeUkpkvA==", + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/ansi-regex?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/ansi-styles": { + "version": "6.2.1", + "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-6.2.1.tgz", + "integrity": "sha512-bN798gFfQX+viw3R7yrGWRqnrN2oRkEkUjjl4JNn4E8GxxbjtG3FbrEIIY3l8/hrwUwIeCZvi4QuOTP4MErVug==", + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/ansi-styles?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/any-promise": { + "version": "1.3.0", + "resolved": "https://registry.npmjs.org/any-promise/-/any-promise-1.3.0.tgz", + "integrity": "sha512-7UvmKalWRt1wgjL1RrGxoSJW/0QZFIegpeGvZG9kjp8vrRu55XTHbwnqq2GpXm9uLbcuhxm3IqX9OB4MZR1b2A==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/balanced-match": { + "version": "1.0.2", + "resolved": "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.2.tgz", + "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/brace-expansion": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz", + "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==", + "license": "MIT", + "dependencies": { + "balanced-match": "^1.0.0" + } + }, + "node_modules/sucrase/node_modules/color-convert": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/color-convert/-/color-convert-2.0.1.tgz", + "integrity": "sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==", + "license": "MIT", + "dependencies": { + "color-name": "~1.1.4" + }, + "engines": { + "node": ">=7.0.0" + } + }, + "node_modules/sucrase/node_modules/color-name": { + "version": "1.1.4", + "resolved": "https://registry.npmjs.org/color-name/-/color-name-1.1.4.tgz", + "integrity": "sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/commander": { + "version": "4.1.1", + "resolved": "https://registry.npmjs.org/commander/-/commander-4.1.1.tgz", + "integrity": "sha512-NOKm8xhkzAjzFx8B2v5OAHT+u5pRQc2UCa2Vq9jYL/31o2wi9mxBA7LIFs3sV5VSC49z6pEhfbMULvShKj26WA==", + "license": "MIT", + "engines": { + "node": ">= 6" + } + }, + "node_modules/sucrase/node_modules/cross-spawn": { + "version": "7.0.6", + "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz", + "integrity": "sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==", + "license": "MIT", + "dependencies": { + "path-key": "^3.1.0", + "shebang-command": "^2.0.0", + "which": "^2.0.1" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/sucrase/node_modules/eastasianwidth": { + "version": "0.2.0", + "resolved": "https://registry.npmjs.org/eastasianwidth/-/eastasianwidth-0.2.0.tgz", + "integrity": "sha512-I88TYZWc9XiYHRQ4/3c5rjjfgkjhLyW2luGIheGERbNQ6OY7yTybanSpDXZa8y7VUP9YmDcYa+eyq4ca7iLqWA==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/emoji-regex": { + "version": "9.2.2", + "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-9.2.2.tgz", + "integrity": "sha512-L18DaJsXSUk2+42pv8mLs5jJT2hqFkFE4j21wOmgbUqsZ2hL72NsUU785g9RXgo3s0ZNgVl42TiHp3ZtOv/Vyg==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/foreground-child": { + "version": "3.3.0", + "resolved": "https://registry.npmjs.org/foreground-child/-/foreground-child-3.3.0.tgz", + "integrity": "sha512-Ld2g8rrAyMYFXBhEqMz8ZAHBi4J4uS1i/CxGMDnjyFWddMXLVcDp051DZfu+t7+ab7Wv6SMqpWmyFIj5UbfFvg==", + "license": "ISC", + "dependencies": { + "cross-spawn": "^7.0.0", + "signal-exit": "^4.0.1" + }, + "engines": { + "node": ">=14" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/sucrase/node_modules/glob": { + "version": "10.4.5", + "resolved": "https://registry.npmjs.org/glob/-/glob-10.4.5.tgz", + "integrity": "sha512-7Bv8RF0k6xjo7d4A/PxYLbUCfb6c+Vpd2/mB2yRDlew7Jb5hEXiCD9ibfO7wpk8i4sevK6DFny9h7EYbM3/sHg==", + "license": "ISC", + "dependencies": { + "foreground-child": "^3.1.0", + "jackspeak": "^3.1.2", + "minimatch": "^9.0.4", + "minipass": "^7.1.2", + "package-json-from-dist": "^1.0.0", + "path-scurry": "^1.11.1" + }, + "bin": { + "glob": "dist/esm/bin.mjs" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/sucrase/node_modules/is-fullwidth-code-point": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/is-fullwidth-code-point/-/is-fullwidth-code-point-3.0.0.tgz", + "integrity": "sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/isexe": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/isexe/-/isexe-2.0.0.tgz", + "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==", + "license": "ISC" + }, + "node_modules/sucrase/node_modules/jackspeak": { + "version": "3.4.3", + "resolved": "https://registry.npmjs.org/jackspeak/-/jackspeak-3.4.3.tgz", + "integrity": "sha512-OGlZQpz2yfahA/Rd1Y8Cd9SIEsqvXkLVoSw/cgwhnhFMDbsQFeZYoJJ7bIZBS9BcamUW96asq/npPWugM+RQBw==", + "license": "BlueOak-1.0.0", + "dependencies": { + "@isaacs/cliui": "^8.0.2" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + }, + "optionalDependencies": { + "@pkgjs/parseargs": "^0.11.0" + } + }, + "node_modules/sucrase/node_modules/lines-and-columns": { + "version": "1.2.4", + "resolved": "https://registry.npmjs.org/lines-and-columns/-/lines-and-columns-1.2.4.tgz", + "integrity": "sha512-7ylylesZQ/PV29jhEDl3Ufjo6ZX7gCqJr5F7PKrqc93v7fzSymt1BpwEU8nAUXs8qzzvqhbjhK5QZg6Mt/HkBg==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/lru-cache": { + "version": "10.4.3", + "resolved": "https://registry.npmjs.org/lru-cache/-/lru-cache-10.4.3.tgz", + "integrity": "sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ==", + "license": "ISC" + }, + "node_modules/sucrase/node_modules/minimatch": { + "version": "9.0.5", + "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.5.tgz", + "integrity": "sha512-G6T0ZX48xgozx7587koeX9Ys2NYy6Gmv//P89sEte9V9whIapMNF4idKxnW2QtCcLiTWlb/wfCabAtAFWhhBow==", + "license": "ISC", + "dependencies": { + "brace-expansion": "^2.0.1" + }, + "engines": { + "node": ">=16 || 14 >=14.17" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/sucrase/node_modules/minipass": { + "version": "7.1.2", + "resolved": "https://registry.npmjs.org/minipass/-/minipass-7.1.2.tgz", + "integrity": "sha512-qOOzS1cBTWYF4BH8fVePDBOO9iptMnGUEZwNc/cMWnTV2nVLZ7VoNWEPHkYczZA0pdoA7dl6e7FL659nX9S2aw==", + "license": "ISC", + "engines": { + "node": ">=16 || 14 >=14.17" + } + }, + "node_modules/sucrase/node_modules/mz": { + "version": "2.7.0", + "resolved": "https://registry.npmjs.org/mz/-/mz-2.7.0.tgz", + "integrity": "sha512-z81GNO7nnYMEhrGh9LeymoE4+Yr0Wn5McHIZMK5cfQCl+NDX08sCZgUc9/6MHni9IWuFLm1Z3HTCXu2z9fN62Q==", + "license": "MIT", + "dependencies": { + "any-promise": "^1.0.0", + "object-assign": "^4.0.1", + "thenify-all": "^1.0.0" + } + }, + "node_modules/sucrase/node_modules/object-assign": { + "version": "4.1.1", + "resolved": "https://registry.npmjs.org/object-assign/-/object-assign-4.1.1.tgz", + "integrity": "sha512-rJgTQnkUnH1sFw8yT6VSU3zD3sWmu6sZhIseY8VX+GRu3P6F7Fu+JNDoXfklElbLJSnc3FUQHVe4cU5hj+BcUg==", + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/sucrase/node_modules/package-json-from-dist": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/package-json-from-dist/-/package-json-from-dist-1.0.1.tgz", + "integrity": "sha512-UEZIS3/by4OC8vL3P2dTXRETpebLI2NiI5vIrjaD/5UtrkFX/tNbwjTSRAGC/+7CAo2pIcBaRgWmcBBHcsaCIw==", + "license": "BlueOak-1.0.0" + }, + "node_modules/sucrase/node_modules/path-key": { + "version": "3.1.1", + "resolved": "https://registry.npmjs.org/path-key/-/path-key-3.1.1.tgz", + "integrity": "sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/path-scurry": { + "version": "1.11.1", + "resolved": "https://registry.npmjs.org/path-scurry/-/path-scurry-1.11.1.tgz", + "integrity": "sha512-Xa4Nw17FS9ApQFJ9umLiJS4orGjm7ZzwUrwamcGQuHSzDyth9boKDaycYdDcZDuqYATXw4HFXgaqWTctW/v1HA==", + "license": "BlueOak-1.0.0", + "dependencies": { + "lru-cache": "^10.2.0", + "minipass": "^5.0.0 || ^6.0.2 || ^7.0.0" + }, + "engines": { + "node": ">=16 || 14 >=14.18" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/sucrase/node_modules/pirates": { + "version": "4.0.6", + "resolved": "https://registry.npmjs.org/pirates/-/pirates-4.0.6.tgz", + "integrity": "sha512-saLsH7WeYYPiD25LDuLRRY/i+6HaPYr6G1OUlN39otzkSTxKnubR9RTxS3/Kk50s1g2JTgFwWQDQyplC5/SHZg==", + "license": "MIT", + "engines": { + "node": ">= 6" + } + }, + "node_modules/sucrase/node_modules/shebang-command": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/shebang-command/-/shebang-command-2.0.0.tgz", + "integrity": "sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==", + "license": "MIT", + "dependencies": { + "shebang-regex": "^3.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/shebang-regex": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/shebang-regex/-/shebang-regex-3.0.0.tgz", + "integrity": "sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/signal-exit": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/signal-exit/-/signal-exit-4.1.0.tgz", + "integrity": "sha512-bzyZ1e88w9O1iNJbKnOlvYTrWPDl46O1bG0D3XInv+9tkPrxrN8jUUTiFlDkkmKWgn1M6CfIA13SuGqOa9Korw==", + "license": "ISC", + "engines": { + "node": ">=14" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/sucrase/node_modules/string-width": { + "version": "5.1.2", + "resolved": "https://registry.npmjs.org/string-width/-/string-width-5.1.2.tgz", + "integrity": "sha512-HnLOCR3vjcY8beoNLtcjZ5/nxn2afmME6lhrDrebokqMap+XbeW8n9TXpPDOqdGK5qcI3oT0GKTW6wC7EMiVqA==", + "license": "MIT", + "dependencies": { + "eastasianwidth": "^0.2.0", + "emoji-regex": "^9.2.2", + "strip-ansi": "^7.0.1" + }, + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/sucrase/node_modules/string-width-cjs": { + "name": "string-width", + "version": "4.2.3", + "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz", + "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==", + "license": "MIT", + "dependencies": { + "emoji-regex": "^8.0.0", + "is-fullwidth-code-point": "^3.0.0", + "strip-ansi": "^6.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/string-width-cjs/node_modules/ansi-regex": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz", + "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/string-width-cjs/node_modules/emoji-regex": { + "version": "8.0.0", + "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz", + "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/string-width-cjs/node_modules/strip-ansi": { + "version": "6.0.1", + "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", + "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", + "license": "MIT", + "dependencies": { + "ansi-regex": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/strip-ansi": { + "version": "7.1.0", + "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-7.1.0.tgz", + "integrity": "sha512-iq6eVVI64nQQTRYq2KtEg2d2uU7LElhTJwsH4YzIHZshxlgZms/wIc4VoDQTlG/IvVIrBKG06CrZnp0qv7hkcQ==", + "license": "MIT", + "dependencies": { + "ansi-regex": "^6.0.1" + }, + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/strip-ansi?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/strip-ansi-cjs": { + "name": "strip-ansi", + "version": "6.0.1", + "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", + "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", + "license": "MIT", + "dependencies": { + "ansi-regex": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/strip-ansi-cjs/node_modules/ansi-regex": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz", + "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/thenify": { + "version": "3.3.1", + "resolved": "https://registry.npmjs.org/thenify/-/thenify-3.3.1.tgz", + "integrity": "sha512-RVZSIV5IG10Hk3enotrhvz0T9em6cyHBLkH/YAZuKqd8hRkKhSfCGIcP2KUY0EPxndzANBmNllzWPwak+bheSw==", + "license": "MIT", + "dependencies": { + "any-promise": "^1.0.0" + } + }, + "node_modules/sucrase/node_modules/thenify-all": { + "version": "1.6.0", + "resolved": "https://registry.npmjs.org/thenify-all/-/thenify-all-1.6.0.tgz", + "integrity": "sha512-RNxQH/qI8/t3thXJDwcstUO4zeqo64+Uy/+sNVRBx4Xn2OX+OZ9oP+iJnNFqplFra2ZUVeKCSa2oVWi3T4uVmA==", + "license": "MIT", + "dependencies": { + "thenify": ">= 3.1.0 < 4" + }, + "engines": { + "node": ">=0.8" + } + }, + "node_modules/sucrase/node_modules/ts-interface-checker": { + "version": "0.1.13", + "resolved": "https://registry.npmjs.org/ts-interface-checker/-/ts-interface-checker-0.1.13.tgz", + "integrity": "sha512-Y/arvbn+rrz3JCKl9C4kVNfTfSm2/mEp5FSz5EsZSANGPSlQrpRI5M4PKF+mJnE52jOO90PnPSc3Ur3bTQw0gA==", + "license": "Apache-2.0" + }, + "node_modules/sucrase/node_modules/which": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz", + "integrity": "sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==", + "license": "ISC", + "dependencies": { + "isexe": "^2.0.0" + }, + "bin": { + "node-which": "bin/node-which" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi": { + "version": "8.1.0", + "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-8.1.0.tgz", + "integrity": "sha512-si7QWI6zUMq56bESFvagtmzMdGOtoxfR+Sez11Mobfc7tm+VkUckk9bW2UeffTGVUbOksxmSw0AA2gs8g71NCQ==", + "license": "MIT", + "dependencies": { + "ansi-styles": "^6.1.0", + "string-width": "^5.0.1", + "strip-ansi": "^7.0.1" + }, + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/wrap-ansi?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs": { + "name": "wrap-ansi", + "version": "7.0.0", + "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-7.0.0.tgz", + "integrity": "sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==", + "license": "MIT", + "dependencies": { + "ansi-styles": "^4.0.0", + "string-width": "^4.1.0", + "strip-ansi": "^6.0.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/chalk/wrap-ansi?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs/node_modules/ansi-regex": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz", + "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==", + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs/node_modules/ansi-styles": { + "version": "4.3.0", + "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz", + "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==", + "license": "MIT", + "dependencies": { + "color-convert": "^2.0.1" + }, + "engines": { + "node": ">=8" + }, + "funding": { + "url": "https://github.com/chalk/ansi-styles?sponsor=1" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs/node_modules/emoji-regex": { + "version": "8.0.0", + "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz", + "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==", + "license": "MIT" + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs/node_modules/string-width": { + "version": "4.2.3", + "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz", + "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==", + "license": "MIT", + "dependencies": { + "emoji-regex": "^8.0.0", + "is-fullwidth-code-point": "^3.0.0", + "strip-ansi": "^6.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/sucrase/node_modules/wrap-ansi-cjs/node_modules/strip-ansi": { + "version": "6.0.1", + "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz", + "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==", + "license": "MIT", + "dependencies": { + "ansi-regex": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/supports-color": { + "version": "8.1.1", + "resolved": "https://registry.npmjs.org/supports-color/-/supports-color-8.1.1.tgz", + "integrity": "sha512-MpUEN2OodtUzxvKQl72cUF7RQ5EiHsGvSsVG0ia9c5RbWGL2CI4C7EpPS8UTBIplnlzZiNuV56w+FuNxy3ty2Q==", + "devOptional": true, + "license": "MIT", + "dependencies": { + "has-flag": "^4.0.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/chalk/supports-color?sponsor=1" + } + }, + "node_modules/sync-child-process": { + "version": "1.0.2", + "resolved": "https://registry.npmjs.org/sync-child-process/-/sync-child-process-1.0.2.tgz", + "integrity": "sha512-8lD+t2KrrScJ/7KXCSyfhT3/hRq78rC0wBFqNJXv3mZyn6hW2ypM05JmlSvtqRbeq6jqA94oHbxAr2vYsJ8vDA==", + "devOptional": true, + "license": "MIT", + "dependencies": { + "sync-message-port": "^1.0.0" + }, + "engines": { + "node": ">=16.0.0" + } + }, + "node_modules/sync-message-port": { + "version": "1.1.3", + "resolved": "https://registry.npmjs.org/sync-message-port/-/sync-message-port-1.1.3.tgz", + "integrity": "sha512-GTt8rSKje5FilG+wEdfCkOcLL7LWqpMlr2c3LRuKt/YXxcJ52aGSbGBAdI4L3aaqfrBt6y711El53ItyH1NWzg==", + "devOptional": true, + "license": "MIT", + "engines": { + "node": ">=16.0.0" + } + }, + "node_modules/tailwindcss": { + "version": "3.4.15", + "resolved": "https://registry.npmjs.org/tailwindcss/-/tailwindcss-3.4.15.tgz", + "integrity": "sha512-r4MeXnfBmSOuKUWmXe6h2CcyfzJCEk4F0pptO5jlnYSIViUkVmsawj80N5h2lO3gwcmSb4n3PuN+e+GC1Guylw==", + "license": "MIT", + "dependencies": { + "@alloc/quick-lru": "^5.2.0", + "arg": "^5.0.2", + "chokidar": "^3.6.0", + "didyoumean": "^1.2.2", + "dlv": "^1.1.3", + "fast-glob": "^3.3.2", + "glob-parent": "^6.0.2", + "is-glob": "^4.0.3", + "jiti": "^1.21.6", + "lilconfig": "^2.1.0", + "micromatch": "^4.0.8", + "normalize-path": "^3.0.0", + "object-hash": "^3.0.0", + "picocolors": "^1.1.1", + "postcss": "^8.4.47", + "postcss-import": "^15.1.0", + "postcss-js": "^4.0.1", + "postcss-load-config": "^4.0.2", + "postcss-nested": "^6.2.0", + "postcss-selector-parser": "^6.1.2", + "resolve": "^1.22.8", + "sucrase": "^3.35.0" + }, + "bin": { + "tailwind": "lib/cli.js", + "tailwindcss": "lib/cli.js" + }, + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/textlinestream": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/textlinestream/-/textlinestream-1.1.1.tgz", + "integrity": "sha512-iBHbi7BQxrFmwZUQJsT0SjNzlLLsXhvW/kg7EyOMVMBIrlnj/qYofwo1LVLZi+3GbUEo96Iu2eqToI2+lZoAEQ==", + "license": "MIT" + }, + "node_modules/tslib": { + "version": "2.8.1", + "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz", + "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==", + "devOptional": true, + "license": "0BSD" + }, + "node_modules/uc.micro": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/uc.micro/-/uc.micro-2.1.0.tgz", + "integrity": "sha512-ARDJmphmdvUk6Glw7y9DQ2bFkKBHwQHLi2lsaH6PPmz/Ka9sFOBsBluozhDltWmnv9u/cF6Rt87znRTPV+yp/A==", + "license": "MIT" + }, + "node_modules/varint": { + "version": "6.0.0", + "resolved": "https://registry.npmjs.org/varint/-/varint-6.0.0.tgz", + "integrity": "sha512-cXEIW6cfr15lFv563k4GuVuW/fiwjknytD37jIOLSdSWuOI6WnO/oKwmP2FQTU2l01LP8/M5TSAJpzUaGe3uWg==", + "devOptional": true, + "license": "MIT" + }, + "node_modules/vite": { + "version": "5.4.11", + "resolved": "https://registry.npmjs.org/vite/-/vite-5.4.11.tgz", + "integrity": "sha512-c7jFQRklXua0mTzneGW9QVyxFjUgwcihC4bXEtujIo2ouWCe1Ajt/amn2PCxYnhYfd5k09JX3SB7OYWFKYqj8Q==", + "license": "MIT", + "dependencies": { + "esbuild": "^0.21.3", + "postcss": "^8.4.43", + "rollup": "^4.20.0" + }, + "bin": { + "vite": "bin/vite.js" + }, + "engines": { + "node": "^18.0.0 || >=20.0.0" + }, + "funding": { + "url": "https://github.com/vitejs/vite?sponsor=1" + }, + "optionalDependencies": { + "fsevents": "~2.3.3" + }, + "peerDependencies": { + "@types/node": "^18.0.0 || >=20.0.0", + "less": "*", + "lightningcss": "^1.21.0", + "sass": "*", + "sass-embedded": "*", + "stylus": "*", + "sugarss": "*", + "terser": "^5.4.0" + }, + "peerDependenciesMeta": { + "@types/node": { + "optional": true + }, + "less": { + "optional": true + }, + "lightningcss": { + "optional": true + }, + "sass": { + "optional": true + }, + "sass-embedded": { + "optional": true + }, + "stylus": { + "optional": true + }, + "sugarss": { + "optional": true + }, + "terser": { + "optional": true + } + } + }, + "node_modules/vite-plugin-singlefile": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/vite-plugin-singlefile/-/vite-plugin-singlefile-2.0.3.tgz", + "integrity": "sha512-OEBEwdX8nCGPSdtaB1D7rryYnT+YfPTS8ojL1TDyeUF+bWDCTfRriQqw6T0vl9EbKI/KMg7szN3awst6cLrKkA==", + "license": "MIT", + "dependencies": { + "micromatch": "^4.0.8" + }, + "engines": { + "node": ">18.0.0" + }, + "peerDependencies": { + "rollup": "^4.24.3", + "vite": "^5.4.10" + } + }, + "node_modules/vue": { + "version": "3.5.13", + "resolved": "https://registry.npmjs.org/vue/-/vue-3.5.13.tgz", + "integrity": "sha512-wmeiSMxkZCSc+PM2w2VRsOYAZC8GdipNFRTsLSfodVqI9mbejKeXEGr8SckuLnrQPGe3oJN5c3K0vpoU9q/wCQ==", + "license": "MIT", + "dependencies": { + "@vue/compiler-dom": "3.5.13", + "@vue/compiler-sfc": "3.5.13", + "@vue/runtime-dom": "3.5.13", + "@vue/server-renderer": "3.5.13", + "@vue/shared": "3.5.13" + }, + "peerDependencies": { + "typescript": "*" + }, + "peerDependenciesMeta": { + "typescript": { + "optional": true + } + } + } + } +} diff --git a/examples/server/webui/package.json b/examples/server/webui/package.json new file mode 100644 index 000000000..2836cce00 --- /dev/null +++ b/examples/server/webui/package.json @@ -0,0 +1,30 @@ +{ + "name": "webui", + "private": true, + "version": "0.0.0", + "type": "module", + "scripts": { + "dev": "vite", + "build": "vite build", + "preview": "vite preview", + "analyze": "ANALYZE=1 npx vite-bundle-visualizer" + }, + "devDependencies": { + "sass-embedded": "^1.83.0", + "vite": "^5.4.10" + }, + "dependencies": { + "@sec-ant/readable-stream": "^0.6.0", + "@vscode/markdown-it-katex": "^1.1.1", + "autoprefixer": "^10.4.20", + "daisyui": "^4.12.14", + "highlight.js": "^11.10.0", + "katex": "^0.16.15", + "markdown-it": "^14.1.0", + "postcss": "^8.4.49", + "tailwindcss": "^3.4.15", + "textlinestream": "^1.1.1", + "vite-plugin-singlefile": "^2.0.3", + "vue": "^3.5.13" + } +} diff --git a/examples/server/webui/postcss.config.js b/examples/server/webui/postcss.config.js new file mode 100644 index 000000000..2e7af2b7f --- /dev/null +++ b/examples/server/webui/postcss.config.js @@ -0,0 +1,6 @@ +export default { + plugins: { + tailwindcss: {}, + autoprefixer: {}, + }, +} diff --git a/examples/server/webui/public/demo-conversation.json b/examples/server/webui/public/demo-conversation.json new file mode 100644 index 000000000..75ab599dd --- /dev/null +++ b/examples/server/webui/public/demo-conversation.json @@ -0,0 +1,33 @@ +{ + "demo": true, + "id": "conv-1734086746930", + "lastModified": 1734087548943, + "messages": [ + { + "id": 1734086764521, + "role": "user", + "content": "this is a demo conversation, used in dev mode" + }, + { + "id": 1734087548327, + "role": "assistant", + "content": "This is the formula:\n\n$\\frac{e^{x_i}}{\\sum_{j=1}^{n}e^{x_j}}$\n\nGiven an input vector \\(\\mathbf{x} = [x_1, x_2, \\ldots, x_n]\\)\n\n\\[\ny_i = \\frac{e^{x_i}}{\\sum_{j=1}^n e^{x_j}}\n\\]\n\nCode block latex:\n```latex\n\\frac{e^{x_i}}{\\sum_{j=1}^{n}e^{x_j}}\n```\n\nTest dollar sign: $1234 $4567\n\nInvalid latex syntax: $E = mc^$ and $$E = mc^$$", + "timings": { + "prompt_n": 1, + "prompt_ms": 28.923, + "predicted_n": 25, + "predicted_ms": 573.016 + } + }, + { + "id": 1734087548328, + "role": "user", + "content": "this is a demo conversation, used in dev mode" + }, + { + "id": 1734087548329, + "role": "assistant", + "content": "Code block:\n```js\nconsole.log('hello world')\n```\n```sh\nls -la /dev\n```" + } + ] +} diff --git a/examples/server/webui/src/highlight-config.js b/examples/server/webui/src/highlight-config.js new file mode 100644 index 000000000..96c7028f9 --- /dev/null +++ b/examples/server/webui/src/highlight-config.js @@ -0,0 +1,60 @@ +import hljs from 'highlight.js/lib/core'; + +// only import commonly used languages to reduce bundle size + +import python from 'highlight.js/lib/languages/python'; +import javascript from 'highlight.js/lib/languages/javascript'; +import json from 'highlight.js/lib/languages/json'; +import bash from 'highlight.js/lib/languages/bash'; +import yaml from 'highlight.js/lib/languages/yaml'; +import markdown from 'highlight.js/lib/languages/markdown'; +import scss from 'highlight.js/lib/languages/scss'; +import xml from 'highlight.js/lib/languages/xml'; +import ruby from 'highlight.js/lib/languages/ruby'; +import go from 'highlight.js/lib/languages/go'; +import java from 'highlight.js/lib/languages/java'; +import rust from 'highlight.js/lib/languages/rust'; +import scala from 'highlight.js/lib/languages/scala'; +import cpp from 'highlight.js/lib/languages/cpp'; +import csharp from 'highlight.js/lib/languages/csharp'; +import swift from 'highlight.js/lib/languages/swift'; +import dart from 'highlight.js/lib/languages/dart'; +import elixir from 'highlight.js/lib/languages/elixir'; +import kotlin from 'highlight.js/lib/languages/kotlin'; +import lua from 'highlight.js/lib/languages/lua'; +import php from 'highlight.js/lib/languages/php'; +import latex from 'highlight.js/lib/languages/latex'; + +hljs.registerLanguage('python', python); +hljs.registerLanguage('javascript', javascript); +hljs.registerLanguage('json', json); +hljs.registerLanguage('yaml', yaml); +hljs.registerLanguage('markdown', markdown); +hljs.registerLanguage('xml', xml); +hljs.registerLanguage('ruby', ruby); +hljs.registerLanguage('go', go); +hljs.registerLanguage('java', java); +hljs.registerLanguage('rust', rust); +hljs.registerLanguage('scala', scala); +hljs.registerLanguage('csharp', csharp); +hljs.registerLanguage('swift', swift); +hljs.registerLanguage('dart', dart); +hljs.registerLanguage('elixir', elixir); +hljs.registerLanguage('kotlin', kotlin); +hljs.registerLanguage('lua', lua); +hljs.registerLanguage('php', php); +hljs.registerLanguage('latex', latex); + +// reuse some languages to further reduce bundle size + +hljs.registerLanguage('shell', bash); +hljs.registerLanguage('bash', bash); +hljs.registerLanguage('sh', bash); + +hljs.registerLanguage('css', scss); +hljs.registerLanguage('scss', scss); + +hljs.registerLanguage('c', cpp); +hljs.registerLanguage('cpp', cpp); + +export default hljs; diff --git a/examples/server/webui/src/katex-gpt.js b/examples/server/webui/src/katex-gpt.js new file mode 100644 index 000000000..7c7c5e22c --- /dev/null +++ b/examples/server/webui/src/katex-gpt.js @@ -0,0 +1,66 @@ +import katex from 'katex'; + +// Adapted from https://github.com/SchneeHertz/markdown-it-katex-gpt +// MIT license + +const defaultOptions = { + delimiters: [ + { left: '\\[', right: '\\]', display: true }, + { left: '\\(', right: '\\)', display: false }, + ], +}; + +export function renderLatexHTML(content, display = false) { + return katex.renderToString(content, { + throwOnError: false, + output: 'mathml', + displayMode: display, + }); +} + +function escapedBracketRule(options) { + return (state, silent) => { + const max = state.posMax; + const start = state.pos; + + for (const { left, right, display } of options.delimiters) { + + // Check if it starts with the left delimiter + if (!state.src.slice(start).startsWith(left)) continue; + + // Skip the length of the left delimiter + let pos = start + left.length; + + // Find the matching right delimiter + while (pos < max) { + if (state.src.slice(pos).startsWith(right)) { + break; + } + pos++; + } + + // No matching right delimiter found, skip to the next match + if (pos >= max) continue; + + // If not in silent mode, convert LaTeX formula to MathML + if (!silent) { + const content = state.src.slice(start + left.length, pos); + try { + const renderedContent = renderLatexHTML(content, display); + const token = state.push('html_inline', '', 0); + token.content = renderedContent; + } catch (e) { + console.error(e); + } + } + + // Update position, skip the length of the right delimiter + state.pos = pos + right.length; + return true; + } + } +} + +export default function (md, options = defaultOptions) { + md.inline.ruler.after('text', 'escaped_bracket', escapedBracketRule(options)); +} diff --git a/examples/server/webui/src/main.js b/examples/server/webui/src/main.js new file mode 100644 index 000000000..feb741a4e --- /dev/null +++ b/examples/server/webui/src/main.js @@ -0,0 +1,600 @@ +import './styles.scss'; +import { createApp, defineComponent, shallowRef, computed, h } from 'vue/dist/vue.esm-bundler.js'; +import MarkdownIt from 'markdown-it'; +import TextLineStream from 'textlinestream'; + +// math formula rendering +import 'katex/dist/katex.min.css'; +import markdownItKatexGpt from './katex-gpt'; +import markdownItKatexNormal from '@vscode/markdown-it-katex'; + +// code highlighting +import hljs from './highlight-config'; +import daisyuiThemes from 'daisyui/src/theming/themes'; + +// ponyfill for missing ReadableStream asyncIterator on Safari +import { asyncIterator } from '@sec-ant/readable-stream/ponyfill/asyncIterator'; + +const isDev = import.meta.env.MODE === 'development'; + +// utility functions +const isString = (x) => !!x.toLowerCase; +const isBoolean = (x) => x === true || x === false; +const isNumeric = (n) => !isString(n) && !isNaN(n) && !isBoolean(n); +const escapeAttr = (str) => str.replace(/>/g, '>').replace(/"/g, '"'); +const copyStr = (textToCopy) => { + // Navigator clipboard api needs a secure context (https) + if (navigator.clipboard && window.isSecureContext) { + navigator.clipboard.writeText(textToCopy); + } else { + // Use the 'out of viewport hidden text area' trick + const textArea = document.createElement('textarea'); + textArea.value = textToCopy; + // Move textarea out of the viewport so it's not visible + textArea.style.position = 'absolute'; + textArea.style.left = '-999999px'; + document.body.prepend(textArea); + textArea.select(); + document.execCommand('copy'); + } +}; + +// constants +const BASE_URL = isDev + ? (localStorage.getItem('base') || 'https://localhost:8080') // for debugging + : (new URL('.', document.baseURI).href).toString().replace(/\/$/, ''); // for production +console.log({ BASE_URL }); + +const CONFIG_DEFAULT = { + // Note: in order not to introduce breaking changes, please keep the same data type (number, string, etc) if you want to change the default value. Do not use null or undefined for default value. + apiKey: '', + systemMessage: 'You are a helpful assistant.', + showTokensPerSecond: false, + // make sure these default values are in sync with `common.h` + samplers: 'edkypmxt', + temperature: 0.8, + dynatemp_range: 0.0, + dynatemp_exponent: 1.0, + top_k: 40, + top_p: 0.95, + min_p: 0.05, + xtc_probability: 0.0, + xtc_threshold: 0.1, + typical_p: 1.0, + repeat_last_n: 64, + repeat_penalty: 1.0, + presence_penalty: 0.0, + frequency_penalty: 0.0, + dry_multiplier: 0.0, + dry_base: 1.75, + dry_allowed_length: 2, + dry_penalty_last_n: -1, + max_tokens: -1, + custom: '', // custom json-stringified object +}; +const CONFIG_INFO = { + apiKey: 'Set the API Key if you are using --api-key option for the server.', + systemMessage: 'The starting message that defines how model should behave.', + samplers: 'The order at which samplers are applied, in simplified way. Default is "dkypmxt": dry->top_k->typ_p->top_p->min_p->xtc->temperature', + temperature: 'Controls the randomness of the generated text by affecting the probability distribution of the output tokens. Higher = more random, lower = more focused.', + dynatemp_range: 'Addon for the temperature sampler. The added value to the range of dynamic temperature, which adjusts probabilities by entropy of tokens.', + dynatemp_exponent: 'Addon for the temperature sampler. Smoothes out the probability redistribution based on the most probable token.', + top_k: 'Keeps only k top tokens.', + top_p: 'Limits tokens to those that together have a cumulative probability of at least p', + min_p: 'Limits tokens based on the minimum probability for a token to be considered, relative to the probability of the most likely token.', + xtc_probability: 'XTC sampler cuts out top tokens; this parameter controls the chance of cutting tokens at all. 0 disables XTC.', + xtc_threshold: 'XTC sampler cuts out top tokens; this parameter controls the token probability that is required to cut that token.', + typical_p: 'Sorts and limits tokens based on the difference between log-probability and entropy.', + repeat_last_n: 'Last n tokens to consider for penalizing repetition', + repeat_penalty: 'Controls the repetition of token sequences in the generated text', + presence_penalty: 'Limits tokens based on whether they appear in the output or not.', + frequency_penalty: 'Limits tokens based on how often they appear in the output.', + dry_multiplier: 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the DRY sampling multiplier.', + dry_base: 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the DRY sampling base value.', + dry_allowed_length: 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the allowed length for DRY sampling.', + dry_penalty_last_n: 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets DRY penalty for the last n tokens.', + max_tokens: 'The maximum number of token per output.', + custom: '', // custom json-stringified object +}; +// config keys having numeric value (i.e. temperature, top_k, top_p, etc) +const CONFIG_NUMERIC_KEYS = Object.entries(CONFIG_DEFAULT).filter(e => isNumeric(e[1])).map(e => e[0]); +// list of themes supported by daisyui +const THEMES = ['light', 'dark'] + // make sure light & dark are always at the beginning + .concat(Object.keys(daisyuiThemes).filter(t => t !== 'light' && t !== 'dark')); + +// markdown support +const VueMarkdown = defineComponent( + (props) => { + const md = shallowRef(new MarkdownIt({ + breaks: true, + highlight: function (str, lang) { // Add highlight.js + if (lang && hljs.getLanguage(lang)) { + try { + return '
' +
+                   hljs.highlight(str, { language: lang, ignoreIllegals: true }).value +
+                   '
'; + } catch (__) {} + } + return '
' + md.value.utils.escapeHtml(str) + '
'; + } + })); + // support latex with double dollar sign and square brackets + md.value.use(markdownItKatexGpt, { + delimiters: [ + { left: '\\[', right: '\\]', display: true }, + { left: '\\(', right: '\\)', display: false }, + { left: '$$', right: '$$', display: false }, + // do not add single dollar sign here, other wise it will confused with dollar used for money symbol + ], + throwOnError: false, + }); + // support latex with single dollar sign + md.value.use(markdownItKatexNormal, { throwOnError: false }); + // add copy button to code blocks + const origFenchRenderer = md.value.renderer.rules.fence; + md.value.renderer.rules.fence = (tokens, idx, ...args) => { + const content = tokens[idx].content; + const origRendered = origFenchRenderer(tokens, idx, ...args); + return `
+ + ${origRendered} +
`; + }; + window.copyStr = copyStr; + const content = computed(() => md.value.render(props.source)); + return () => h('div', { innerHTML: content.value }); + }, + { props: ['source'] } +); + +// input field to be used by settings modal +const SettingsModalShortInput = defineComponent({ + template: document.getElementById('settings-modal-short-input').innerHTML, + props: { + label: { type: String, required: false }, + configKey: String, + configDefault: Object, + configInfo: Object, + modelValue: [Object, String, Number], + }, +}); + +// message bubble component +const MessageBubble = defineComponent({ + components: { + VueMarkdown + }, + template: document.getElementById('message-bubble').innerHTML, + props: { + config: Object, + msg: Object, + isGenerating: Boolean, + editUserMsgAndRegenerate: Function, + regenerateMsg: Function, + }, + data() { + return { + editingContent: null, + }; + }, + computed: { + timings() { + if (!this.msg.timings) return null; + return { + ...this.msg.timings, + prompt_per_second: this.msg.timings.prompt_n / (this.msg.timings.prompt_ms / 1000), + predicted_per_second: this.msg.timings.predicted_n / (this.msg.timings.predicted_ms / 1000), + }; + } + }, + methods: { + copyMsg() { + copyStr(this.msg.content); + }, + editMsg() { + this.editUserMsgAndRegenerate({ + ...this.msg, + content: this.editingContent, + }); + this.editingContent = null; + }, + }, +}); + +// coversations is stored in localStorage +// format: { [convId]: { id: string, lastModified: number, messages: [...] } } +// convId is a string prefixed with 'conv-' +const StorageUtils = { + // manage conversations + getAllConversations() { + const res = []; + for (const key in localStorage) { + if (key.startsWith('conv-')) { + res.push(JSON.parse(localStorage.getItem(key))); + } + } + res.sort((a, b) => b.lastModified - a.lastModified); + return res; + }, + // can return null if convId does not exist + getOneConversation(convId) { + return JSON.parse(localStorage.getItem(convId) || 'null'); + }, + // if convId does not exist, create one + appendMsg(convId, msg) { + if (msg.content === null) return; + const conv = StorageUtils.getOneConversation(convId) || { + id: convId, + lastModified: Date.now(), + messages: [], + }; + conv.messages.push(msg); + conv.lastModified = Date.now(); + localStorage.setItem(convId, JSON.stringify(conv)); + }, + getNewConvId() { + return `conv-${Date.now()}`; + }, + remove(convId) { + localStorage.removeItem(convId); + }, + filterAndKeepMsgs(convId, predicate) { + const conv = StorageUtils.getOneConversation(convId); + if (!conv) return; + conv.messages = conv.messages.filter(predicate); + conv.lastModified = Date.now(); + localStorage.setItem(convId, JSON.stringify(conv)); + }, + popMsg(convId) { + const conv = StorageUtils.getOneConversation(convId); + if (!conv) return; + const msg = conv.messages.pop(); + conv.lastModified = Date.now(); + if (conv.messages.length === 0) { + StorageUtils.remove(convId); + } else { + localStorage.setItem(convId, JSON.stringify(conv)); + } + return msg; + }, + + // manage config + getConfig() { + const savedVal = JSON.parse(localStorage.getItem('config') || '{}'); + // to prevent breaking changes in the future, we always provide default value for missing keys + return { + ...CONFIG_DEFAULT, + ...savedVal, + }; + }, + setConfig(config) { + localStorage.setItem('config', JSON.stringify(config)); + }, + getTheme() { + return localStorage.getItem('theme') || 'auto'; + }, + setTheme(theme) { + if (theme === 'auto') { + localStorage.removeItem('theme'); + } else { + localStorage.setItem('theme', theme); + } + }, +}; + +// scroll to bottom of chat messages +// if requiresNearBottom is true, only auto-scroll if user is near bottom +const chatScrollToBottom = (requiresNearBottom) => { + const msgListElem = document.getElementById('messages-list'); + const spaceToBottom = msgListElem.scrollHeight - msgListElem.scrollTop - msgListElem.clientHeight; + if (!requiresNearBottom || (spaceToBottom < 100)) { + setTimeout(() => msgListElem.scrollTo({ top: msgListElem.scrollHeight }), 1); + } +}; + +// wrapper for SSE +async function* sendSSEPostRequest(url, fetchOptions) { + const res = await fetch(url, fetchOptions); + const lines = res.body + .pipeThrough(new TextDecoderStream()) + .pipeThrough(new TextLineStream()); + for await (const line of asyncIterator(lines)) { + if (isDev) console.log({line}); + if (line.startsWith('data:') && !line.endsWith('[DONE]')) { + const data = JSON.parse(line.slice(5)); + yield data; + } else if (line.startsWith('error:')) { + const data = JSON.parse(line.slice(6)); + throw new Error(data.message || 'Unknown error'); + } + } +}; + +const mainApp = createApp({ + components: { + VueMarkdown, + SettingsModalShortInput, + MessageBubble, + }, + data() { + return { + conversations: StorageUtils.getAllConversations(), + messages: [], // { id: number, role: 'user' | 'assistant', content: string } + viewingConvId: StorageUtils.getNewConvId(), + inputMsg: '', + isGenerating: false, + pendingMsg: null, // the on-going message from assistant + stopGeneration: () => {}, + selectedTheme: StorageUtils.getTheme(), + config: StorageUtils.getConfig(), + showConfigDialog: false, + // const + themes: THEMES, + configDefault: {...CONFIG_DEFAULT}, + configInfo: {...CONFIG_INFO}, + isDev, + } + }, + computed: {}, + mounted() { + document.getElementById('app').classList.remove('opacity-0'); // show app + // scroll to the bottom when the pending message height is updated + const pendingMsgElem = document.getElementById('pending-msg'); + const resizeObserver = new ResizeObserver(() => { + if (this.isGenerating) chatScrollToBottom(true); + }); + resizeObserver.observe(pendingMsgElem); + this.setSelectedTheme(this.selectedTheme); + }, + watch: { + viewingConvId: function(val, oldVal) { + if (val != oldVal) { + this.fetchMessages(); + chatScrollToBottom(); + this.hideSidebar(); + } + } + }, + methods: { + hideSidebar() { + document.getElementById('toggle-drawer').checked = false; + }, + setSelectedTheme(theme) { + this.selectedTheme = theme; + document.body.setAttribute('data-theme', theme); + document.body.setAttribute('data-color-scheme', daisyuiThemes[theme]?.['color-scheme'] ?? 'auto'); + StorageUtils.setTheme(theme); + }, + newConversation() { + if (this.isGenerating) return; + this.viewingConvId = StorageUtils.getNewConvId(); + }, + setViewingConv(convId) { + if (this.isGenerating) return; + this.viewingConvId = convId; + }, + deleteConv(convId) { + if (this.isGenerating) return; + if (window.confirm('Are you sure to delete this conversation?')) { + StorageUtils.remove(convId); + if (this.viewingConvId === convId) { + this.viewingConvId = StorageUtils.getNewConvId(); + } + this.fetchConversation(); + this.fetchMessages(); + } + }, + downloadConv(convId) { + const conversation = StorageUtils.getOneConversation(convId); + if (!conversation) { + alert('Conversation not found.'); + return; + } + const conversationJson = JSON.stringify(conversation, null, 2); + const blob = new Blob([conversationJson], { type: 'application/json' }); + const url = URL.createObjectURL(blob); + const a = document.createElement('a'); + a.href = url; + a.download = `conversation_${convId}.json`; + document.body.appendChild(a); + a.click(); + document.body.removeChild(a); + URL.revokeObjectURL(url); + }, + async sendMessage() { + if (!this.inputMsg) return; + const currConvId = this.viewingConvId; + + StorageUtils.appendMsg(currConvId, { + id: Date.now(), + role: 'user', + content: this.inputMsg, + }); + this.fetchConversation(); + this.fetchMessages(); + this.inputMsg = ''; + this.generateMessage(currConvId); + chatScrollToBottom(); + }, + async generateMessage(currConvId) { + if (this.isGenerating) return; + this.pendingMsg = { id: Date.now()+1, role: 'assistant', content: null }; + this.isGenerating = true; + + try { + const abortController = new AbortController(); + this.stopGeneration = () => abortController.abort(); + const params = { + messages: [ + { role: 'system', content: this.config.systemMessage }, + ...this.messages, + ], + stream: true, + cache_prompt: true, + samplers: this.config.samplers, + temperature: this.config.temperature, + dynatemp_range: this.config.dynatemp_range, + dynatemp_exponent: this.config.dynatemp_exponent, + top_k: this.config.top_k, + top_p: this.config.top_p, + min_p: this.config.min_p, + typical_p: this.config.typical_p, + xtc_probability: this.config.xtc_probability, + xtc_threshold: this.config.xtc_threshold, + repeat_last_n: this.config.repeat_last_n, + repeat_penalty: this.config.repeat_penalty, + presence_penalty: this.config.presence_penalty, + frequency_penalty: this.config.frequency_penalty, + dry_multiplier: this.config.dry_multiplier, + dry_base: this.config.dry_base, + dry_allowed_length: this.config.dry_allowed_length, + dry_penalty_last_n: this.config.dry_penalty_last_n, + max_tokens: this.config.max_tokens, + timings_per_token: !!this.config.showTokensPerSecond, + ...(this.config.custom.length ? JSON.parse(this.config.custom) : {}), + }; + const chunks = sendSSEPostRequest(`${BASE_URL}/v1/chat/completions`, { + method: 'POST', + headers: { + 'Content-Type': 'application/json', + ...(this.config.apiKey ? {'Authorization': `Bearer ${this.config.apiKey}`} : {}) + }, + body: JSON.stringify(params), + signal: abortController.signal, + }); + for await (const chunk of chunks) { + const stop = chunk.stop; + const addedContent = chunk.choices[0].delta.content; + const lastContent = this.pendingMsg.content || ''; + if (addedContent) { + this.pendingMsg = { + id: this.pendingMsg.id, + role: 'assistant', + content: lastContent + addedContent, + }; + } + const timings = chunk.timings; + if (timings && this.config.showTokensPerSecond) { + // only extract what's really needed, to save some space + this.pendingMsg.timings = { + prompt_n: timings.prompt_n, + prompt_ms: timings.prompt_ms, + predicted_n: timings.predicted_n, + predicted_ms: timings.predicted_ms, + }; + } + } + + StorageUtils.appendMsg(currConvId, this.pendingMsg); + this.fetchConversation(); + this.fetchMessages(); + setTimeout(() => document.getElementById('msg-input').focus(), 1); + } catch (error) { + if (error.name === 'AbortError') { + // user stopped the generation via stopGeneration() function + StorageUtils.appendMsg(currConvId, this.pendingMsg); + this.fetchConversation(); + this.fetchMessages(); + } else { + console.error(error); + alert(error); + // pop last user message + const lastUserMsg = StorageUtils.popMsg(currConvId); + this.inputMsg = lastUserMsg ? lastUserMsg.content : ''; + } + } + + this.pendingMsg = null; + this.isGenerating = false; + this.stopGeneration = () => {}; + this.fetchMessages(); + chatScrollToBottom(); + }, + + // message actions + regenerateMsg(msg) { + if (this.isGenerating) return; + // TODO: somehow keep old history (like how ChatGPT has different "tree"). This can be done by adding "sub-conversations" with "subconv-" prefix, and new message will have a list of subconvIds + const currConvId = this.viewingConvId; + StorageUtils.filterAndKeepMsgs(currConvId, (m) => m.id < msg.id); + this.fetchConversation(); + this.fetchMessages(); + this.generateMessage(currConvId); + }, + editUserMsgAndRegenerate(msg) { + if (this.isGenerating) return; + const currConvId = this.viewingConvId; + const newContent = msg.content; + StorageUtils.filterAndKeepMsgs(currConvId, (m) => m.id < msg.id); + StorageUtils.appendMsg(currConvId, { + id: Date.now(), + role: 'user', + content: newContent, + }); + this.fetchConversation(); + this.fetchMessages(); + this.generateMessage(currConvId); + }, + + // settings dialog methods + closeAndSaveConfigDialog() { + try { + if (this.config.custom.length) JSON.parse(this.config.custom); + } catch (error) { + alert('Invalid JSON for custom config. Please either fix it or leave it empty.'); + return; + } + for (const key of CONFIG_NUMERIC_KEYS) { + if (isNaN(this.config[key]) || this.config[key].toString().trim().length === 0) { + alert(`Invalid number for ${key} (expected an integer or a float)`); + return; + } + this.config[key] = parseFloat(this.config[key]); + } + this.showConfigDialog = false; + StorageUtils.setConfig(this.config); + }, + closeAndDiscardConfigDialog() { + this.showConfigDialog = false; + this.config = StorageUtils.getConfig(); + }, + resetConfigDialog() { + if (window.confirm('Are you sure to reset all settings?')) { + this.config = {...CONFIG_DEFAULT}; + } + }, + + // sync state functions + fetchConversation() { + this.conversations = StorageUtils.getAllConversations(); + }, + fetchMessages() { + this.messages = StorageUtils.getOneConversation(this.viewingConvId)?.messages ?? []; + }, + + // debug functions + async debugImportDemoConv() { + const res = await fetch('/demo-conversation.json'); + const demoConv = await res.json(); + StorageUtils.remove(demoConv.id); + for (const msg of demoConv.messages) { + StorageUtils.appendMsg(demoConv.id, msg); + } + this.fetchConversation(); + } + }, +}); +mainApp.config.errorHandler = alert; +try { + mainApp.mount('#app'); +} catch (err) { + console.error(err); + document.getElementById('app').innerHTML = `
+ Failed to start app. Please try clearing localStorage and try again.
+
+ +
`; +} diff --git a/examples/server/webui/src/styles.scss b/examples/server/webui/src/styles.scss new file mode 100644 index 000000000..34fe2aaf0 --- /dev/null +++ b/examples/server/webui/src/styles.scss @@ -0,0 +1,48 @@ +@use "sass:meta"; + +@tailwind base; +@tailwind components; +@tailwind utilities; + +.markdown { + h1, h2, h3, h4, h5, h6, ul, ol, li { all: revert; } + pre { + @apply whitespace-pre-wrap rounded-lg p-2; + border: 1px solid currentColor; + } + /* TODO: fix markdown table */ +} + +.show-on-hover { + @apply md:opacity-0 md:group-hover:opacity-100; +} +.btn-mini { + @apply cursor-pointer hover:shadow-md; +} +.chat-screen { max-width: 900px; } + +.chat-bubble-base-300 { + --tw-bg-opacity: 1; + --tw-text-opacity: 1; + @apply bg-base-300 text-base-content; +} + +/* Highlight.js */ +[data-color-scheme='light'] { + @include meta.load-css('highlight.js/styles/stackoverflow-light'); +} +[data-color-scheme='dark'] { + @include meta.load-css('highlight.js/styles/stackoverflow-dark'); +} +[data-color-scheme='auto'] { + @media (prefers-color-scheme: light) { + @include meta.load-css('highlight.js/styles/stackoverflow-light'); + } + @media (prefers-color-scheme: dark) { + @include meta.load-css('highlight.js/styles/stackoverflow-dark'); + } +} +.hljs { + background: transparent !important; + padding: 0.5em !important; +} diff --git a/examples/server/webui/tailwind.config.js b/examples/server/webui/tailwind.config.js new file mode 100644 index 000000000..c43066a19 --- /dev/null +++ b/examples/server/webui/tailwind.config.js @@ -0,0 +1,16 @@ +/** @type {import('tailwindcss').Config} */ +export default { + content: [ + "./index.html", + "./src/**/*.{js,ts,jsx,tsx}", + ], + theme: { + extend: {}, + }, + plugins: [ + require('daisyui'), + ], + daisyui: { + themes: ['light', 'dark', 'cupcake', 'bumblebee', 'emerald', 'corporate', 'synthwave', 'retro', 'cyberpunk', 'valentine', 'halloween', 'garden', 'forest', 'aqua', 'lofi', 'pastel', 'fantasy', 'wireframe', 'black', 'luxury', 'dracula', 'cmyk', 'autumn', 'business', 'acid', 'lemonade', 'night', 'coffee', 'winter', 'dim', 'nord', 'sunset'], + } +} diff --git a/examples/server/webui/vite.config.js b/examples/server/webui/vite.config.js new file mode 100644 index 000000000..6619a630d --- /dev/null +++ b/examples/server/webui/vite.config.js @@ -0,0 +1,59 @@ + +import { viteSingleFile } from 'vite-plugin-singlefile'; +import path from 'path'; +import fs from 'fs'; +import zlib from 'zlib'; + +const MAX_BUNDLE_SIZE = 1.5 * 1024 * 1024; // only increase when absolutely necessary + +const GUIDE_FOR_FRONTEND = ` + +`.trim(); + +const BUILD_PLUGINS = [ + viteSingleFile(), + (function llamaCppPlugin() { + let config; + return { + name: 'llamacpp:build', + apply: 'build', + async configResolved(_config) { + config = _config; + }, + writeBundle() { + const outputIndexHtml = path.join(config.build.outDir, 'index.html'); + const content = GUIDE_FOR_FRONTEND + '\n' + fs.readFileSync(outputIndexHtml, 'utf-8'); + const compressed = zlib.gzipSync(Buffer.from(content, 'utf-8'), { level: 9 }); + + // because gzip header contains machine-specific info, we must remove these data from the header + // timestamp + compressed[0x4] = 0; + compressed[0x5] = 0; + compressed[0x6] = 0; + compressed[0x7] = 0; + // OS + compressed[0x9] = 0; + + if (compressed.byteLength > MAX_BUNDLE_SIZE) { + throw new Error( + `Bundle size is too large (${Math.ceil(compressed.byteLength / 1024)} KB).\n` + + `Please reduce the size of the frontend or increase MAX_BUNDLE_SIZE in vite.config.js.\n`, + ); + } + + const targetOutputFile = path.join(config.build.outDir, '../../public/index.html.gz'); + fs.writeFileSync(targetOutputFile, compressed); + } + } + })(), +]; + +/** @type {import('vite').UserConfig} */ +export default { + plugins: process.env.ANALYZE ? [] : BUILD_PLUGINS, +}; diff --git a/examples/simple-chat/CMakeLists.txt b/examples/simple-chat/CMakeLists.txt index 87723533b..567f7fbbb 100644 --- a/examples/simple-chat/CMakeLists.txt +++ b/examples/simple-chat/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-simple-chat) add_executable(${TARGET} simple-chat.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/simple-chat/simple-chat.cpp b/examples/simple-chat/simple-chat.cpp index 5f9973163..e8eda9c22 100644 --- a/examples/simple-chat/simple-chat.cpp +++ b/examples/simple-chat/simple-chat.cpp @@ -62,22 +62,27 @@ int main(int argc, char ** argv) { } }, nullptr); + // load dynamic backends + ggml_backend_load_all(); + // initialize the model llama_model_params model_params = llama_model_default_params(); model_params.n_gpu_layers = ngl; - llama_model * model = llama_load_model_from_file(model_path.c_str(), model_params); + llama_model * model = llama_model_load_from_file(model_path.c_str(), model_params); if (!model) { fprintf(stderr , "%s: error: unable to load model\n" , __func__); return 1; } + const llama_vocab * vocab = llama_model_get_vocab(model); + // initialize the context llama_context_params ctx_params = llama_context_default_params(); ctx_params.n_ctx = n_ctx; ctx_params.n_batch = n_ctx; - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); if (!ctx) { fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__); return 1; @@ -94,9 +99,9 @@ int main(int argc, char ** argv) { std::string response; // tokenize the prompt - const int n_prompt_tokens = -llama_tokenize(model, prompt.c_str(), prompt.size(), NULL, 0, true, true); + const int n_prompt_tokens = -llama_tokenize(vocab, prompt.c_str(), prompt.size(), NULL, 0, true, true); std::vector prompt_tokens(n_prompt_tokens); - if (llama_tokenize(model, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), llama_get_kv_cache_used_cells(ctx) == 0, true) < 0) { + if (llama_tokenize(vocab, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), llama_get_kv_cache_used_cells(ctx) == 0, true) < 0) { GGML_ABORT("failed to tokenize the prompt\n"); } @@ -121,13 +126,13 @@ int main(int argc, char ** argv) { new_token_id = llama_sampler_sample(smpl, ctx, -1); // is it an end of generation? - if (llama_token_is_eog(model, new_token_id)) { + if (llama_vocab_is_eog(vocab, new_token_id)) { break; } // convert the token to a string, print it and add it to the response char buf[256]; - int n = llama_token_to_piece(model, new_token_id, buf, sizeof(buf), 0, true); + int n = llama_token_to_piece(vocab, new_token_id, buf, sizeof(buf), 0, true); if (n < 0) { GGML_ABORT("failed to convert token to piece\n"); } @@ -156,12 +161,14 @@ int main(int argc, char ** argv) { break; } + const char * tmpl = llama_model_chat_template(model); + // add the user input to the message list and format it messages.push_back({"user", strdup(user.c_str())}); - int new_len = llama_chat_apply_template(model, nullptr, messages.data(), messages.size(), true, formatted.data(), formatted.size()); + int new_len = llama_chat_apply_template(tmpl, messages.data(), messages.size(), true, formatted.data(), formatted.size()); if (new_len > (int)formatted.size()) { formatted.resize(new_len); - new_len = llama_chat_apply_template(model, nullptr, messages.data(), messages.size(), true, formatted.data(), formatted.size()); + new_len = llama_chat_apply_template(tmpl, messages.data(), messages.size(), true, formatted.data(), formatted.size()); } if (new_len < 0) { fprintf(stderr, "failed to apply the chat template\n"); @@ -178,7 +185,7 @@ int main(int argc, char ** argv) { // add the response to the messages messages.push_back({"assistant", strdup(response.c_str())}); - prev_len = llama_chat_apply_template(model, nullptr, messages.data(), messages.size(), false, nullptr, 0); + prev_len = llama_chat_apply_template(tmpl, messages.data(), messages.size(), false, nullptr, 0); if (prev_len < 0) { fprintf(stderr, "failed to apply the chat template\n"); return 1; @@ -191,7 +198,7 @@ int main(int argc, char ** argv) { } llama_sampler_free(smpl); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); return 0; } diff --git a/examples/simple/CMakeLists.txt b/examples/simple/CMakeLists.txt index b63afbb8b..104ecabfd 100644 --- a/examples/simple/CMakeLists.txt +++ b/examples/simple/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-simple) add_executable(${TARGET} simple.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/simple/README.md b/examples/simple/README.md index 0ff342535..937008b24 100644 --- a/examples/simple/README.md +++ b/examples/simple/README.md @@ -3,7 +3,7 @@ The purpose of this example is to demonstrate a minimal usage of llama.cpp for generating text with a given prompt. ```bash -./llama-simple -m ./models/llama-7b-v2/ggml-model-f16.gguf -p "Hello my name is" +./llama-simple -m ./models/llama-7b-v2/ggml-model-f16.gguf "Hello my name is" ... diff --git a/examples/simple/simple.cpp b/examples/simple/simple.cpp index 59760fe95..10e79a0a6 100644 --- a/examples/simple/simple.cpp +++ b/examples/simple/simple.cpp @@ -74,12 +74,17 @@ int main(int argc, char ** argv) { } } + // load dynamic backends + + ggml_backend_load_all(); + // initialize the model llama_model_params model_params = llama_model_default_params(); model_params.n_gpu_layers = ngl; - llama_model * model = llama_load_model_from_file(model_path.c_str(), model_params); + llama_model * model = llama_model_load_from_file(model_path.c_str(), model_params); + const llama_vocab * vocab = llama_model_get_vocab(model); if (model == NULL) { fprintf(stderr , "%s: error: unable to load model\n" , __func__); @@ -89,11 +94,11 @@ int main(int argc, char ** argv) { // tokenize the prompt // find the number of tokens in the prompt - const int n_prompt = -llama_tokenize(model, prompt.c_str(), prompt.size(), NULL, 0, true, true); + const int n_prompt = -llama_tokenize(vocab, prompt.c_str(), prompt.size(), NULL, 0, true, true); // allocate space for the tokens and tokenize the prompt std::vector prompt_tokens(n_prompt); - if (llama_tokenize(model, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), true, true) < 0) { + if (llama_tokenize(vocab, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), true, true) < 0) { fprintf(stderr, "%s: error: failed to tokenize the prompt\n", __func__); return 1; } @@ -108,7 +113,7 @@ int main(int argc, char ** argv) { // enable performance counters ctx_params.no_perf = false; - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); if (ctx == NULL) { fprintf(stderr , "%s: error: failed to create the llama_context\n" , __func__); @@ -127,7 +132,7 @@ int main(int argc, char ** argv) { for (auto id : prompt_tokens) { char buf[128]; - int n = llama_token_to_piece(model, id, buf, sizeof(buf), 0, true); + int n = llama_token_to_piece(vocab, id, buf, sizeof(buf), 0, true); if (n < 0) { fprintf(stderr, "%s: error: failed to convert token to piece\n", __func__); return 1; @@ -160,12 +165,12 @@ int main(int argc, char ** argv) { new_token_id = llama_sampler_sample(smpl, ctx, -1); // is it an end of generation? - if (llama_token_is_eog(model, new_token_id)) { + if (llama_vocab_is_eog(vocab, new_token_id)) { break; } char buf[128]; - int n = llama_token_to_piece(model, new_token_id, buf, sizeof(buf), 0, true); + int n = llama_token_to_piece(vocab, new_token_id, buf, sizeof(buf), 0, true); if (n < 0) { fprintf(stderr, "%s: error: failed to convert token to piece\n", __func__); return 1; @@ -195,7 +200,7 @@ int main(int argc, char ** argv) { llama_sampler_free(smpl); llama_free(ctx); - llama_free_model(model); + llama_model_free(model); return 0; } diff --git a/examples/speculative-simple/CMakeLists.txt b/examples/speculative-simple/CMakeLists.txt new file mode 100644 index 000000000..aeaea74fc --- /dev/null +++ b/examples/speculative-simple/CMakeLists.txt @@ -0,0 +1,5 @@ +set(TARGET llama-speculative-simple) +add_executable(${TARGET} speculative-simple.cpp) +install(TARGETS ${TARGET} RUNTIME) +target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/speculative-simple/README.md b/examples/speculative-simple/README.md new file mode 100644 index 000000000..e3a6c6b4a --- /dev/null +++ b/examples/speculative-simple/README.md @@ -0,0 +1,12 @@ +# llama.cpp/examples/speculative-simple + +Demonstration of basic greedy speculative decoding + +```bash +./bin/llama-speculative-simple \ + -m ../models/qwen2.5-32b-coder-instruct/ggml-model-q8_0.gguf \ + -md ../models/qwen2.5-1.5b-coder-instruct/ggml-model-q4_0.gguf \ + -f test.txt -c 0 -ngl 99 --color \ + --sampling-seq k --top-k 1 -fa --temp 0.0 \ + -ngld 99 --draft-max 16 --draft-min 5 --draft-p-min 0.9 +``` diff --git a/examples/speculative-simple/speculative-simple.cpp b/examples/speculative-simple/speculative-simple.cpp new file mode 100644 index 000000000..403ba2dd2 --- /dev/null +++ b/examples/speculative-simple/speculative-simple.cpp @@ -0,0 +1,261 @@ +#include "arg.h" +#include "common.h" +#include "sampling.h" +#include "speculative.h" +#include "log.h" +#include "llama.h" + +#include +#include +#include +#include + +int main(int argc, char ** argv) { + common_params params; + + if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_SPECULATIVE)) { + return 1; + } + + if (params.n_predict < -1) { + LOG_ERR("%s: --n-predict must be >= -1\n", __func__); + return 1; + } + + common_init(); + + if (params.speculative.model.empty()) { + LOG_ERR("%s: --model-draft is required\n", __func__); + return 1; + } + + // init llama.cpp + llama_backend_init(); + llama_numa_init(params.numa); + + llama_model * model_tgt = NULL; + //llama_model * model_dft = NULL; + + llama_context * ctx_tgt = NULL; + llama_context * ctx_dft = NULL; + + // load the target model + common_init_result llama_init_tgt = common_init_from_params(params); + + model_tgt = llama_init_tgt.model.get(); + ctx_tgt = llama_init_tgt.context.get(); + + const llama_vocab * vocab = llama_model_get_vocab(model_tgt); + + // load the draft model + params.devices = params.speculative.devices; + params.model = params.speculative.model; + params.n_ctx = params.speculative.n_ctx; + params.n_batch = params.speculative.n_ctx > 0 ? params.speculative.n_ctx : params.n_batch; + params.n_gpu_layers = params.speculative.n_gpu_layers; + + if (params.speculative.cpuparams.n_threads > 0) { + params.cpuparams.n_threads = params.speculative.cpuparams.n_threads; + } + + params.cpuparams_batch.n_threads = params.speculative.cpuparams_batch.n_threads; + common_init_result llama_init_dft = common_init_from_params(params); + + //model_dft = llama_init_dft.model.get(); + ctx_dft = llama_init_dft.context.get(); + + if (!common_speculative_are_compatible(ctx_tgt, ctx_dft)) { + return 1; + } + + // Tokenize the prompt + std::vector inp; + inp = common_tokenize(ctx_tgt, params.prompt, true, true); + + if (llama_n_ctx(ctx_tgt) < (uint32_t) inp.size()) { + LOG_ERR("%s: the prompt exceeds the context size (%d tokens, ctx %d)\n", __func__, (int) inp.size(), llama_n_ctx(ctx_tgt)); + + return 1; + } + + if (llama_n_batch(ctx_tgt) < (uint32_t) inp.size()) { + LOG_ERR("%s: the prompt exceeds the batch size (%d tokens, batch %d)\n", __func__, (int) inp.size(), llama_n_batch(ctx_tgt)); + + return 1; + } + + LOG("\n\n"); + + for (auto id : inp) { + LOG("%s", common_token_to_piece(ctx_tgt, id).c_str()); + } + + // how many tokens to draft each time + int n_draft = params.speculative.n_max; + int n_draft_min = params.speculative.n_min; + + float p_min = params.speculative.p_min; + + int n_predict = 0; + int n_drafted = 0; + int n_accept = 0; + + // used to determine end of generation + bool has_eos = false; + + // ================================================ + // everything until here is standard initialization + // the relevant stuff for speculative decoding starts here + + const auto t_enc_start = ggml_time_us(); + + // target model sampling context + struct common_sampler * smpl = common_sampler_init(model_tgt, params.sampling); + + // eval the prompt + llama_decode(ctx_tgt, llama_batch_get_one(inp.data(), inp.size() - 1)); + + // note: keep the last token separate! + llama_token id_last = inp.back(); + + // all tokens currently in the target context + llama_tokens prompt_tgt(inp.begin(), inp.end() - 1); + prompt_tgt.reserve(llama_n_ctx(ctx_tgt)); + + int n_past = inp.size() - 1; + + // init the speculator + struct common_speculative_params params_spec; + params_spec.n_draft = n_draft; + params_spec.n_reuse = llama_n_ctx(ctx_dft) - n_draft; + params_spec.p_min = p_min; + + struct common_speculative * spec = common_speculative_init(ctx_dft); + + llama_batch batch_tgt = llama_batch_init(llama_n_batch(ctx_tgt), 0, 1); + + const auto t_enc_end = ggml_time_us(); + + const auto t_dec_start = ggml_time_us(); + + while (true) { + // optionally, generate draft tokens that can be appended to the target batch + // + // this is the most important part of the speculation. the more probable tokens that are provided here + // the better the performance will be. in theory, this computation can be performed asynchronously and even + // offloaded to a remote device. it doesn't even have to be based on an LLM. instead, it can provide tokens + // from a cache or lookup tables. + // + llama_tokens draft = common_speculative_gen_draft(spec, params_spec, prompt_tgt, id_last); + + //LOG_DBG("draft: %s\n", string_from(ctx_dft, draft).c_str()); + + // always have a token to evaluate from before - id_last + common_batch_clear(batch_tgt); + common_batch_add (batch_tgt, id_last, n_past++, { 0 }, true); + + // evaluate the target model on [id_last, draft0, draft1, ..., draftN-1] + { + // do not waste time on small drafts + if (draft.size() < (size_t) n_draft_min) { + draft.clear(); + } + + for (size_t i = 0; i < draft.size(); ++i) { + common_batch_add(batch_tgt, draft[i], n_past + i, { 0 }, true); + } + + //LOG_DBG("target batch: %s\n", string_from(ctx_tgt, batch_tgt).c_str()); + + llama_decode(ctx_tgt, batch_tgt); + } + + // sample from the full target batch and return the accepted tokens based on the target sampler + // + // for each token to be accepted, the sampler would have to sample that same token + // in such cases, instead of decoding the sampled token as we normally do, we simply continue with the + // available logits from the batch and sample the next token until we run out of logits or the sampler + // disagrees with the draft + // + const auto ids = common_sampler_sample_and_accept_n(smpl, ctx_tgt, draft); + + //LOG_DBG("ids: %s\n", string_from(ctx_tgt, ids).c_str()); + + GGML_ASSERT(ids.size() > 0); // there will always be at least one accepted token + + n_past += ids.size() - 1; + n_drafted += draft.size(); // note: we ignore the discarded small drafts + n_accept += ids.size() - 1; + n_predict += ids.size(); + + // process the accepted tokens and update contexts + // + // this is the standard token post-processing that we normally do + // in this case, we do it for a group of accepted tokens at once + // + for (size_t i = 0; i < ids.size(); ++i) { + prompt_tgt.push_back(id_last); + + id_last = ids[i]; + + if (llama_vocab_is_eog(vocab, id_last)) { + has_eos = true; + break; + } + + const std::string token_str = common_token_to_piece(ctx_tgt, id_last); + + if (params.use_color && i + 1 < ids.size()) { + LOG("\u001b[%dm%s\u001b[37m", (36 - 0 % 6), token_str.c_str()); + } else { + LOG("%s", token_str.c_str()); + } + } + + LOG_DBG("accepted %d/%d draft tokens, the last target token is: (%d)\n", (int) ids.size() - 1, (int) draft.size(), id_last); + + { + LOG_DBG("clear kv cache from any extra tokens, n_past = %d\n", n_past); + + llama_kv_cache_seq_rm(ctx_tgt, 0, n_past, -1); + } + + if ((params.n_predict >= 0 && n_predict > params.n_predict) || has_eos) { + break; + } + } + + auto t_dec_end = ggml_time_us(); + + const int n_input = inp.size(); + + LOG("\n\n"); + + LOG_INF("encoded %4d tokens in %8.3f seconds, speed: %8.3f t/s\n", n_input, (t_enc_end - t_enc_start) / 1e6f, inp.size() / ((t_enc_end - t_enc_start) / 1e6f)); + LOG_INF("decoded %4d tokens in %8.3f seconds, speed: %8.3f t/s\n", n_predict, (t_dec_end - t_dec_start) / 1e6f, n_predict / ((t_dec_end - t_dec_start) / 1e6f)); + + LOG_INF("\n"); + LOG_INF("n_draft = %d\n", n_draft); + LOG_INF("n_predict = %d\n", n_predict); + LOG_INF("n_drafted = %d\n", n_drafted); + LOG_INF("n_accept = %d\n", n_accept); + LOG_INF("accept = %.3f%%\n", 100.0f * n_accept / n_drafted); + + LOG_INF("\n"); + LOG_INF("draft:\n\n"); + + llama_perf_context_print(ctx_dft); + + LOG_INF("\n"); + LOG_INF("target:\n\n"); + common_perf_print(ctx_tgt, smpl); + + common_sampler_free(smpl); + common_speculative_free(spec); + + llama_backend_free(); + + LOG("\n\n"); + + return 0; +} diff --git a/examples/speculative/CMakeLists.txt b/examples/speculative/CMakeLists.txt index aa208e7aa..c84196bd9 100644 --- a/examples/speculative/CMakeLists.txt +++ b/examples/speculative/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-speculative) add_executable(${TARGET} speculative.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/speculative/speculative.cpp b/examples/speculative/speculative.cpp index a40e755a2..c7ccea50d 100644 --- a/examples/speculative/speculative.cpp +++ b/examples/speculative/speculative.cpp @@ -12,7 +12,7 @@ #include #include -#define SPEC_VOCAB_MAX_SIZE_DIFFERENCE 100 +#define SPEC_VOCAB_MAX_SIZE_DIFFERENCE 128 #define SPEC_VOCAB_CHECK_START_TOKEN_ID 5 struct seq_draft { @@ -33,7 +33,7 @@ int main(int argc, char ** argv) { common_params params; // needed to get candidate probs even for temp <= 0.0 - params.sparams.n_probs = 128; + params.sampling.n_probs = 128; if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_SPECULATIVE)) { return 1; @@ -46,7 +46,7 @@ int main(int argc, char ** argv) { common_init(); - if (params.model_draft.empty()) { + if (params.speculative.model.empty()) { LOG_ERR("%s: --model-draft is required\n", __func__); return 1; } @@ -55,9 +55,9 @@ int main(int argc, char ** argv) { const int n_seq_dft = params.n_parallel; // probability threshold for splitting a draft branch (only for n_seq_dft > 1) - const float p_split = params.p_split; + const float p_draft_split = params.speculative.p_split; - std::default_random_engine rng(params.sparams.seed == LLAMA_DEFAULT_SEED ? std::random_device()() : params.sparams.seed); + std::default_random_engine rng(params.sampling.seed == LLAMA_DEFAULT_SEED ? std::random_device()() : params.sampling.seed); std::uniform_real_distribution<> u_dist; // init llama.cpp @@ -72,25 +72,31 @@ int main(int argc, char ** argv) { // load the target model common_init_result llama_init_tgt = common_init_from_params(params); - model_tgt = llama_init_tgt.model; - ctx_tgt = llama_init_tgt.context; + + model_tgt = llama_init_tgt.model.get(); + ctx_tgt = llama_init_tgt.context.get(); // load the draft model - params.model = params.model_draft; - params.n_gpu_layers = params.n_gpu_layers_draft; - if (params.draft_cpuparams.n_threads > 0) { - params.cpuparams.n_threads = params.draft_cpuparams.n_threads; + params.devices = params.speculative.devices; + params.model = params.speculative.model; + params.n_gpu_layers = params.speculative.n_gpu_layers; + if (params.speculative.cpuparams.n_threads > 0) { + params.cpuparams.n_threads = params.speculative.cpuparams.n_threads; } - params.cpuparams_batch.n_threads = params.draft_cpuparams_batch.n_threads; + params.cpuparams_batch.n_threads = params.speculative.cpuparams_batch.n_threads; common_init_result llama_init_dft = common_init_from_params(params); - model_dft = llama_init_dft.model; - ctx_dft = llama_init_dft.context; - const bool vocab_type_tgt = llama_vocab_type(model_tgt); + model_dft = llama_init_dft.model.get(); + ctx_dft = llama_init_dft.context.get(); + + const llama_vocab * vocab_tgt = llama_model_get_vocab(model_tgt); + const llama_vocab * vocab_dft = llama_model_get_vocab(model_dft); + + const bool vocab_type_tgt = llama_vocab_type(vocab_tgt); LOG_DBG("vocab_type tgt: %d\n", vocab_type_tgt); - const bool vocab_type_dft = llama_vocab_type(model_dft); + const bool vocab_type_dft = llama_vocab_type(vocab_dft); LOG_DBG("vocab_type dft: %d\n", vocab_type_dft); if (vocab_type_tgt != vocab_type_dft) { @@ -100,18 +106,18 @@ int main(int argc, char ** argv) { } if ( - llama_add_bos_token(model_tgt) != llama_add_bos_token(model_dft) || - llama_add_eos_token(model_tgt) != llama_add_eos_token(model_dft) || - llama_token_bos(model_tgt) != llama_token_bos(model_dft) || - llama_token_eos(model_tgt) != llama_token_eos(model_dft) + llama_vocab_get_add_bos(vocab_tgt) != llama_vocab_get_add_bos(vocab_dft) || + llama_vocab_get_add_eos(vocab_tgt) != llama_vocab_get_add_eos(vocab_dft) || + llama_vocab_bos(vocab_tgt) != llama_vocab_bos(vocab_dft) || + llama_vocab_eos(vocab_tgt) != llama_vocab_eos(vocab_dft) ) { LOG_ERR("%s: draft model special tokens must match target model to use speculation\n", __func__); return 1; } { - const int n_vocab_tgt = llama_n_vocab(model_tgt); - const int n_vocab_dft = llama_n_vocab(model_dft); + const int n_vocab_tgt = llama_vocab_n_tokens(vocab_tgt); + const int n_vocab_dft = llama_vocab_n_tokens(vocab_dft); const int vocab_diff = n_vocab_tgt > n_vocab_dft ? n_vocab_tgt - n_vocab_dft : n_vocab_dft - n_vocab_tgt; @@ -119,13 +125,13 @@ int main(int argc, char ** argv) { if (vocab_diff > SPEC_VOCAB_MAX_SIZE_DIFFERENCE) { LOG_ERR("%s: draft model vocab must closely match target model to use speculation but ", __func__); LOG_ERR("target vocab size %d does not match draft vocab size %d - difference %d, max allowed %d\n", - n_vocab_tgt, llama_n_vocab(model_dft), vocab_diff, SPEC_VOCAB_MAX_SIZE_DIFFERENCE); + n_vocab_tgt, llama_vocab_n_tokens(vocab_dft), vocab_diff, SPEC_VOCAB_MAX_SIZE_DIFFERENCE); return 1; } for (int i = SPEC_VOCAB_CHECK_START_TOKEN_ID; i < std::min(n_vocab_tgt, n_vocab_dft); ++i) { - const char * token_text_tgt = llama_token_get_text(model_tgt, i); - const char * token_text_dft = llama_token_get_text(model_dft, i); + const char * token_text_tgt = llama_vocab_get_text(vocab_tgt, i); + const char * token_text_dft = llama_vocab_get_text(vocab_dft, i); if (std::strcmp(token_text_tgt, token_text_dft) != 0) { LOG_ERR("%s: draft model vocab must match target model to use speculation but ", __func__); LOG_ERR("token %d content differs - target '%s', draft '%s'\n", i, @@ -167,10 +173,10 @@ int main(int argc, char ** argv) { const auto t_enc_end = ggml_time_us(); // the 2 models should have the same vocab - //GGML_ASSERT(n_vocab == llama_n_vocab(model_dft)); + //GGML_ASSERT(n_vocab == llama_vocab_n_tokens(model_dft)); // how many tokens to draft each time - int n_draft = params.n_draft; + int n_draft = params.speculative.n_max; int n_predict = 0; int n_drafted = 0; @@ -183,14 +189,14 @@ int main(int argc, char ** argv) { bool has_eos = false; // target model sampling context (reuse the llama_context's sampling instance) - struct common_sampler * smpl = common_sampler_init(model_tgt, params.sparams); + struct common_sampler * smpl = common_sampler_init(model_tgt, params.sampling); // draft sequence data std::vector drafts(n_seq_dft); for (int s = 0; s < n_seq_dft; ++s) { // allocate llama_sampler for each draft sequence - drafts[s].smpl = common_sampler_init(model_dft, params.sparams); + drafts[s].smpl = common_sampler_init(model_dft, params.sampling); } llama_batch batch_dft = llama_batch_init(llama_n_batch(ctx_dft), 0, 1); @@ -230,7 +236,7 @@ int main(int argc, char ** argv) { // for stochastic sampling, attempt to match the token with the drafted tokens { bool accept = false; - if (params.sparams.temp > 0) { + if (params.sampling.temp > 0) { // stochastic verification common_sampler_sample(smpl, ctx_tgt, drafts[s_keep].i_batch_tgt[i_dft], true); @@ -267,11 +273,12 @@ int main(int argc, char ** argv) { for (size_t i = 0; i < dist_tgt.size; i++) { if (dist_tgt.data[i].id == drafts[s].tokens[i_dft]) { p_tgt = dist_tgt.data[i].p; + break; } + } + for (size_t i = 0; i < dist_dft.size; i++) { if (dist_dft.data[i].id == drafts[s].tokens[i_dft]) { p_dft = dist_dft.data[i].p; - } - if (p_tgt && p_dft) { break; } } @@ -382,7 +389,7 @@ int main(int argc, char ** argv) { } } - if (llama_token_is_eog(model_tgt, token_id)) { + if (llama_vocab_is_eog(vocab_tgt, token_id)) { has_eos = true; } ++n_predict; @@ -493,7 +500,7 @@ int main(int argc, char ** argv) { // attempt to split the branch if the probability is high enough for (int f = 1; f < 8; ++f) { - if (n_seq_cur < n_seq_dft && cur_p->data[f].p > p_split) { + if (n_seq_cur < n_seq_dft && cur_p->data[f].p > p_draft_split) { LOG_DBG("splitting seq %3d into %3d\n", s, n_seq_cur); llama_kv_cache_seq_rm(ctx_dft, n_seq_cur, -1, -1); @@ -629,12 +636,6 @@ int main(int argc, char ** argv) { llama_batch_free(batch_dft); - llama_free(ctx_tgt); - llama_free_model(model_tgt); - - llama_free(ctx_dft); - llama_free_model(model_dft); - llama_backend_free(); LOG("\n\n"); diff --git a/examples/tokenize/CMakeLists.txt b/examples/tokenize/CMakeLists.txt index b704dcae1..1690b53e5 100644 --- a/examples/tokenize/CMakeLists.txt +++ b/examples/tokenize/CMakeLists.txt @@ -2,4 +2,4 @@ set(TARGET llama-tokenize) add_executable(${TARGET} tokenize.cpp) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/tokenize/tokenize.cpp b/examples/tokenize/tokenize.cpp index 12ad54256..7375759eb 100644 --- a/examples/tokenize/tokenize.cpp +++ b/examples/tokenize/tokenize.cpp @@ -31,6 +31,7 @@ static void print_usage_information(const char * argv0) { printf(" -p PROMPT, --prompt PROMPT read prompt from the argument.\n"); printf(" --stdin read prompt from standard input.\n"); printf(" --no-bos do not ever add a BOS token to the prompt, even if normally the model uses a BOS token.\n"); + printf(" --no-escape do not escape input (such as \\n, \\t, etc.).\n"); printf(" --no-parse-special do not parse control tokens.\n"); printf(" --log-disable disable logs. Makes stderr quiet when loading the model.\n"); printf(" --show-count print the total number of tokens.\n"); @@ -198,6 +199,7 @@ int main(int raw_argc, char ** raw_argv) { // variables where to put any arguments we see. bool printing_ids = false; bool no_bos = false; + bool no_escape = false; bool no_parse_special = false; bool disable_logging = false; bool show_token_count = false; @@ -233,6 +235,9 @@ int main(int raw_argc, char ** raw_argv) { else if (arg == "--no-bos") { no_bos = true; } + else if (arg == "--no-escape") { + no_escape = true; + } else if (arg == "--no-parse-special") { no_parse_special = true; } @@ -333,14 +338,16 @@ int main(int raw_argc, char ** raw_argv) { llama_model_params model_params = llama_model_default_params(); model_params.vocab_only = true; - llama_model * model = llama_load_model_from_file(model_path, model_params); + llama_model * model = llama_model_load_from_file(model_path, model_params); if (!model) { fprintf(stderr, "Error: could not load model from file '%s'.\n", model_path); return 1; } + const llama_vocab * vocab = llama_model_get_vocab(model); + llama_context_params ctx_params = llama_context_default_params(); - llama_context * ctx = llama_new_context_with_model(model, ctx_params); + llama_context * ctx = llama_init_from_model(model, ctx_params); if (!ctx) { fprintf(stderr, "Error: could not create context.\n"); return 1; @@ -360,12 +367,17 @@ int main(int raw_argc, char ** raw_argv) { prompt = stdin_buffer.str(); } - const bool model_wants_add_bos = llama_add_bos_token(model); + const bool model_wants_add_bos = llama_vocab_get_add_bos(vocab); const bool add_bos = model_wants_add_bos && !no_bos; const bool parse_special = !no_parse_special; + const bool escape = !no_escape; + + if (escape) { + string_process_escapes(prompt); + } std::vector tokens; - tokens = common_tokenize(model, prompt, add_bos, parse_special); + tokens = common_tokenize(vocab, prompt, add_bos, parse_special); if (printing_ids) { printf("["); @@ -394,11 +406,11 @@ int main(int raw_argc, char ** raw_argv) { } if (show_token_count) { - printf("Total number of tokens: %ld\n", tokens.size()); + printf("Total number of tokens: %zu\n", tokens.size()); } // silence valgrind llama_free(ctx); - llama_free_model(model); + llama_model_free(model); return 0; } diff --git a/examples/tts/CMakeLists.txt b/examples/tts/CMakeLists.txt new file mode 100644 index 000000000..c72bd814c --- /dev/null +++ b/examples/tts/CMakeLists.txt @@ -0,0 +1,5 @@ +set(TARGET llama-tts) +add_executable(${TARGET} tts.cpp) +install(TARGETS ${TARGET} RUNTIME) +target_link_libraries(${TARGET} PRIVATE llama common ${CMAKE_THREAD_LIBS_INIT}) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/examples/tts/README.md b/examples/tts/README.md new file mode 100644 index 000000000..4509763c6 --- /dev/null +++ b/examples/tts/README.md @@ -0,0 +1,117 @@ +# llama.cpp/example/tts +This example demonstrates the Text To Speech feature. It uses a +[model](https://www.outeai.com/blog/outetts-0.2-500m) from +[outeai](https://www.outeai.com/). + +## Quickstart +If you have built llama.cpp with `-DLLAMA_CURL=ON` you can simply run the +following command and the required models will be downloaded automatically: +```console +$ build/bin/llama-tts --tts-oute-default -p "Hello world" && aplay output.wav +``` +For details about the models and how to convert them to the required format +see the following sections. + +### Model conversion +Checkout or download the model that contains the LLM model: +```console +$ pushd models +$ git clone --branch main --single-branch --depth 1 https://huggingface.co/OuteAI/OuteTTS-0.2-500M +$ cd OuteTTS-0.2-500M && git lfs install && git lfs pull +$ popd +``` +Convert the model to .gguf format: +```console +(venv) python convert_hf_to_gguf.py models/OuteTTS-0.2-500M \ + --outfile models/outetts-0.2-0.5B-f16.gguf --outtype f16 +``` +The generated model will be `models/outetts-0.2-0.5B-f16.gguf`. + +We can optionally quantize this to Q8_0 using the following command: +```console +$ build/bin/llama-quantize models/outetts-0.2-0.5B-f16.gguf \ + models/outetts-0.2-0.5B-q8_0.gguf q8_0 +``` +The quantized model will be `models/outetts-0.2-0.5B-q8_0.gguf`. + +Next we do something simlar for the audio decoder. First download or checkout +the model for the voice decoder: +```console +$ pushd models +$ git clone --branch main --single-branch --depth 1 https://huggingface.co/novateur/WavTokenizer-large-speech-75token +$ cd WavTokenizer-large-speech-75token && git lfs install && git lfs pull +$ popd +``` +This model file is PyTorch checkpoint (.ckpt) and we first need to convert it to +huggingface format: +```console +(venv) python examples/tts/convert_pt_to_hf.py \ + models/WavTokenizer-large-speech-75token/wavtokenizer_large_speech_320_24k.ckpt +... +Model has been successfully converted and saved to models/WavTokenizer-large-speech-75token/model.safetensors +Metadata has been saved to models/WavTokenizer-large-speech-75token/index.json +Config has been saved to models/WavTokenizer-large-speech-75tokenconfig.json +``` +Then we can convert the huggingface format to gguf: +```console +(venv) python convert_hf_to_gguf.py models/WavTokenizer-large-speech-75token \ + --outfile models/wavtokenizer-large-75-f16.gguf --outtype f16 +... +INFO:hf-to-gguf:Model successfully exported to models/wavtokenizer-large-75-f16.gguf +``` + +### Running the example + +With both of the models generated, the LLM model and the voice decoder model, +we can run the example: +```console +$ build/bin/llama-tts -m ./models/outetts-0.2-0.5B-q8_0.gguf \ + -mv ./models/wavtokenizer-large-75-f16.gguf \ + -p "Hello world" +... +main: audio written to file 'output.wav' +``` +The output.wav file will contain the audio of the prompt. This can be heard +by playing the file with a media player. On Linux the following command will +play the audio: +```console +$ aplay output.wav +``` + +### Running the example with llama-server +Running this example with `llama-server` is also possible and requires two +server instances to be started. One will serve the LLM model and the other +will serve the voice decoder model. + +The LLM model server can be started with the following command: +```console +$ ./build/bin/llama-server -m ./models/outetts-0.2-0.5B-q8_0.gguf --port 8020 +``` + +And the voice decoder model server can be started using: +```console +./build/bin/llama-server -m ./models/wavtokenizer-large-75-f16.gguf --port 8021 --embeddings --pooling none +``` + +Then we can run [tts-outetts.py](tts-outetts.py) to generate the audio. + +First create a virtual environment for python and install the required +dependencies (this in only required to be done once): +```console +$ python3 -m venv venv +$ source venv/bin/activate +(venv) pip install requests numpy +``` + +And then run the python script using: +```conole +(venv) python ./examples/tts/tts-outetts.py http://localhost:8020 http://localhost:8021 "Hello world" +spectrogram generated: n_codes: 90, n_embd: 1282 +converting to audio ... +audio generated: 28800 samples +audio written to file "output.wav" +``` +And to play the audio we can again use aplay or any other media player: +```console +$ aplay output.wav +``` diff --git a/examples/tts/convert_pt_to_hf.py b/examples/tts/convert_pt_to_hf.py new file mode 100644 index 000000000..8909a65fd --- /dev/null +++ b/examples/tts/convert_pt_to_hf.py @@ -0,0 +1,180 @@ +# convert the https://huggingface.co/novateur/WavTokenizer-large-speech-75token to HF format +# the goal is to be able to reuse the convert_hf_to_gguf.py after that to create a GGUF file with the WavTokenizer decoder +# +# TODO: this script is LLM-generated and probably very inefficient and should be rewritten + +import torch +import json +import os +import sys +import re + +from safetensors.torch import save_file + +# default +model_path = './model.pt'; + +# read from CLI +if len(sys.argv) > 1: + model_path = sys.argv[1] + +# get the directory of the input model +path_dst = os.path.dirname(model_path) + +print(f"Loading model from {model_path}") + +model = torch.load(model_path, map_location='cpu') + +#print(model) + +# print all keys +for key in model.keys(): + print(key) + if key == 'hyper_parameters': + #print(model[key]) + # dump as json pretty + print(json.dumps(model[key], indent=4)) + #if key != 'state_dict' and key != 'optimizer_states': + # print(model[key]) + +# Check if the loaded model is a state_dict or a model instance +if isinstance(model, torch.nn.Module): + state_dict = model.state_dict() +else: + state_dict = model + +# Print the structure of the state_dict to understand its format +print("State dictionary keys:") +for key in state_dict.keys(): + print(key) + +# Ensure the state_dict is flat and contains only torch.Tensor objects +def flatten_state_dict(state_dict, parent_key='', sep='.'): + items = [] + items_new = [] + + for k, v in state_dict.items(): + new_key = f"{parent_key}{sep}{k}" if parent_key else k + if isinstance(v, torch.Tensor): + items.append((new_key, v)) + elif isinstance(v, dict): + items.extend(flatten_state_dict(v, new_key, sep=sep).items()) + return dict(items) + + size_total_mb = 0 + + for key, value in list(items): + # keep only what we need for inference + if not key.startswith('state_dict.feature_extractor.encodec.quantizer.') and \ + not key.startswith('state_dict.backbone.') and \ + not key.startswith('state_dict.head.out'): + print('Skipping key: ', key) + continue + + new_key = key + + new_key = new_key.replace('state_dict.', '') + new_key = new_key.replace('pos_net', 'posnet') + + # check if matches "backbone.posnet.%d.bias" or "backbone.posnet.%d.weight" + if new_key.startswith("backbone.posnet."): + match = re.match(r"backbone\.posnet\.(\d+)\.(bias|weight)", new_key) + if match: + new_key = f"backbone.posnet.{match.group(1)}.norm.{match.group(2)}" + + # "feature_extractor.encodec.quantizer.vq.layers.0._codebook.embed" -> "backbone.embedding.weight" + if new_key == "feature_extractor.encodec.quantizer.vq.layers.0._codebook.embed": + new_key = "backbone.embedding.weight" + + # these are the only rows used + # ref: https://github.com/edwko/OuteTTS/blob/a613e79c489d8256dd657ea9168d78de75895d82/outetts/wav_tokenizer/audio_codec.py#L100 + if new_key.endswith("norm.scale.weight"): + new_key = new_key.replace("norm.scale.weight", "norm.weight") + value = value[0] + + if new_key.endswith("norm.shift.weight"): + new_key = new_key.replace("norm.shift.weight", "norm.bias") + value = value[0] + + if new_key.endswith("gamma"): + new_key = new_key.replace("gamma", "gamma.weight") + + # convert from 1D [768] to 2D [768, 1] so that ggml_add can broadcast the bias + if (new_key.endswith("norm.weight") or new_key.endswith("norm1.weight") or new_key.endswith("norm2.weight") or new_key.endswith(".bias")) and (new_key.startswith("backbone.posnet") or new_key.startswith("backbone.embed.bias")): + value = value.unsqueeze(1) + + if new_key.endswith("dwconv.bias"): + value = value.unsqueeze(1) + + size_mb = value.element_size() * value.nelement() / (1024 * 1024) + print(f"{size_mb:8.2f} MB - {new_key}: {value.shape}") + + size_total_mb += size_mb + + #print(key, '->', new_key, ': ', value) + #print(key, '->', new_key) + + items_new.append((new_key, value)) + + print(f"Total size: {size_total_mb:8.2f} MB") + + return dict(items_new) + +flattened_state_dict = flatten_state_dict(state_dict) + + +# Convert the model to the safetensors format +output_path = path_dst + '/model.safetensors' +save_file(flattened_state_dict, output_path) + +print(f"Model has been successfully converted and saved to {output_path}") + +# Calculate the total size of the .safetensors file +total_size = os.path.getsize(output_path) + +# Create the weight map +weight_map = { + "model.safetensors": ["*"] # Assuming all weights are in one file +} + +# Create metadata for the index.json file +metadata = { + "total_size": total_size, + "weight_map": weight_map +} + +# Save the metadata to index.json +index_path = path_dst + '/index.json' +with open(index_path, 'w') as f: + json.dump(metadata, f, indent=4) + +print(f"Metadata has been saved to {index_path}") + +config = { + "architectures": [ + "WavTokenizerDec" + ], + "hidden_size": 1282, + "n_embd_features": 512, + "n_ff": 2304, + "vocab_size": 4096, + "n_head": 1, + "layer_norm_epsilon": 1e-6, + "group_norm_epsilon": 1e-6, + "group_norm_groups": 32, + "max_position_embeddings": 8192, # ? + "n_layer": 12, + "posnet": { + "n_embd": 768, + "n_layer": 6 + }, + "convnext": { + "n_embd": 768, + "n_layer": 12 + }, +} + +with open(path_dst + '/config.json', 'w') as f: + json.dump(config, f, indent=4) + +print(f"Config has been saved to {path_dst + 'config.json'}") diff --git a/examples/tts/tts-outetts.py b/examples/tts/tts-outetts.py new file mode 100644 index 000000000..3791f9fc3 --- /dev/null +++ b/examples/tts/tts-outetts.py @@ -0,0 +1,299 @@ +import sys +#import json +#import struct +import requests +import re +import struct +import numpy as np +from concurrent.futures import ThreadPoolExecutor + + +def fill_hann_window(size, periodic=True): + if periodic: + return np.hanning(size + 1)[:-1] + return np.hanning(size) + + +def irfft(n_fft, complex_input): + return np.fft.irfft(complex_input, n=n_fft) + + +def fold(buffer, n_out, n_win, n_hop, n_pad): + result = np.zeros(n_out) + n_frames = len(buffer) // n_win + + for i in range(n_frames): + start = i * n_hop + end = start + n_win + result[start:end] += buffer[i * n_win:(i + 1) * n_win] + + return result[n_pad:-n_pad] if n_pad > 0 else result + + +def process_frame(args): + l, n_fft, ST, hann = args + frame = irfft(n_fft, ST[l]) + frame = frame * hann + hann2 = hann * hann + return frame, hann2 + + +def embd_to_audio(embd, n_codes, n_embd, n_thread=4): + embd = np.asarray(embd, dtype=np.float32).reshape(n_codes, n_embd) + + n_fft = 1280 + n_hop = 320 + n_win = 1280 + n_pad = (n_win - n_hop) // 2 + n_out = (n_codes - 1) * n_hop + n_win + + hann = fill_hann_window(n_fft, True) + + E = np.zeros((n_embd, n_codes), dtype=np.float32) + for l in range(n_codes): + for k in range(n_embd): + E[k, l] = embd[l, k] + + half_embd = n_embd // 2 + S = np.zeros((n_codes, half_embd + 1), dtype=np.complex64) + + for k in range(half_embd): + for l in range(n_codes): + mag = E[k, l] + phi = E[k + half_embd, l] + + mag = np.clip(np.exp(mag), 0, 1e2) + S[l, k] = mag * np.exp(1j * phi) + + res = np.zeros(n_codes * n_fft) + hann2_buffer = np.zeros(n_codes * n_fft) + + with ThreadPoolExecutor(max_workers=n_thread) as executor: + args = [(l, n_fft, S, hann) for l in range(n_codes)] + results = list(executor.map(process_frame, args)) + + for l, (frame, hann2) in enumerate(results): + res[l*n_fft:(l+1)*n_fft] = frame + hann2_buffer[l*n_fft:(l+1)*n_fft] = hann2 + + audio = fold(res, n_out, n_win, n_hop, n_pad) + env = fold(hann2_buffer, n_out, n_win, n_hop, n_pad) + + mask = env > 1e-10 + audio[mask] /= env[mask] + + return audio + + +def save_wav(filename, audio_data, sample_rate): + num_channels = 1 + bits_per_sample = 16 + bytes_per_sample = bits_per_sample // 8 + data_size = len(audio_data) * bytes_per_sample + byte_rate = sample_rate * num_channels * bytes_per_sample + block_align = num_channels * bytes_per_sample + chunk_size = 36 + data_size # 36 = size of header minus first 8 bytes + + header = struct.pack( + '<4sI4s4sIHHIIHH4sI', + b'RIFF', + chunk_size, + b'WAVE', + b'fmt ', + 16, # fmt chunk size + 1, # audio format (PCM) + num_channels, + sample_rate, + byte_rate, + block_align, + bits_per_sample, + b'data', + data_size + ) + + audio_data = np.clip(audio_data * 32767, -32768, 32767) + pcm_data = audio_data.astype(np.int16) + + with open(filename, 'wb') as f: + f.write(header) + f.write(pcm_data.tobytes()) + + +def process_text(text: str): + text = re.sub(r'\d+(\.\d+)?', lambda x: x.group(), text.lower()) # TODO this needs to be fixed + text = re.sub(r'[-_/,\.\\]', ' ', text) + text = re.sub(r'[^a-z\s]', '', text) + text = re.sub(r'\s+', ' ', text).strip() + return text.split() + +# usage: +# python tts-outetts.py http://server-llm:port http://server-dec:port "text" + +if len(sys.argv) <= 3: + print("usage: python tts-outetts.py http://server-llm:port http://server-dec:port \"text\"") + exit(1) + +host_llm = sys.argv[1] +host_dec = sys.argv[2] +text = sys.argv[3] + +prefix = """<|im_start|> +<|text_start|>the<|text_sep|>overall<|text_sep|>package<|text_sep|>from<|text_sep|>just<|text_sep|>two<|text_sep|>people<|text_sep|>is<|text_sep|>pretty<|text_sep|>remarkable<|text_sep|>sure<|text_sep|>i<|text_sep|>have<|text_sep|>some<|text_sep|>critiques<|text_sep|>about<|text_sep|>some<|text_sep|>of<|text_sep|>the<|text_sep|>gameplay<|text_sep|>aspects<|text_sep|>but<|text_sep|>its<|text_sep|>still<|text_sep|>really<|text_sep|>enjoyable<|text_sep|>and<|text_sep|>it<|text_sep|>looks<|text_sep|>lovely<|text_sep|>""" + +words = process_text(text) +words = "<|text_sep|>".join([i.strip() for i in words]) +words += "<|text_end|>\n" + +# voice data +# TODO: load from json +#suffix = """<|audio_start|> +#the<|t_0.08|><|code_start|><|257|><|740|><|636|><|913|><|788|><|1703|><|code_end|> +#overall<|t_0.36|><|code_start|><|127|><|201|><|191|><|774|><|700|><|532|><|1056|><|557|><|798|><|298|><|1741|><|747|><|1662|><|1617|><|1702|><|1527|><|368|><|1588|><|1049|><|1008|><|1625|><|747|><|1576|><|728|><|1019|><|1696|><|1765|><|code_end|> +#package<|t_0.56|><|code_start|><|935|><|584|><|1319|><|627|><|1016|><|1491|><|1344|><|1117|><|1526|><|1040|><|239|><|1435|><|951|><|498|><|723|><|1180|><|535|><|789|><|1649|><|1637|><|78|><|465|><|1668|><|901|><|595|><|1675|><|117|><|1009|><|1667|><|320|><|840|><|79|><|507|><|1762|><|1508|><|1228|><|1768|><|802|><|1450|><|1457|><|232|><|639|><|code_end|> +#from<|t_0.19|><|code_start|><|604|><|782|><|1682|><|872|><|1532|><|1600|><|1036|><|1761|><|647|><|1554|><|1371|><|653|><|1595|><|950|><|code_end|> +#just<|t_0.25|><|code_start|><|1782|><|1670|><|317|><|786|><|1748|><|631|><|599|><|1155|><|1364|><|1524|><|36|><|1591|><|889|><|1535|><|541|><|440|><|1532|><|50|><|870|><|code_end|> +#two<|t_0.24|><|code_start|><|1681|><|1510|><|673|><|799|><|805|><|1342|><|330|><|519|><|62|><|640|><|1138|><|565|><|1552|><|1497|><|1552|><|572|><|1715|><|1732|><|code_end|> +#people<|t_0.39|><|code_start|><|593|><|274|><|136|><|740|><|691|><|633|><|1484|><|1061|><|1138|><|1485|><|344|><|428|><|397|><|1562|><|645|><|917|><|1035|><|1449|><|1669|><|487|><|442|><|1484|><|1329|><|1832|><|1704|><|600|><|761|><|653|><|269|><|code_end|> +#is<|t_0.16|><|code_start|><|566|><|583|><|1755|><|646|><|1337|><|709|><|802|><|1008|><|485|><|1583|><|652|><|10|><|code_end|> +#pretty<|t_0.32|><|code_start|><|1818|><|1747|><|692|><|733|><|1010|><|534|><|406|><|1697|><|1053|><|1521|><|1355|><|1274|><|816|><|1398|><|211|><|1218|><|817|><|1472|><|1703|><|686|><|13|><|822|><|445|><|1068|><|code_end|> +#remarkable<|t_0.68|><|code_start|><|230|><|1048|><|1705|><|355|><|706|><|1149|><|1535|><|1787|><|1356|><|1396|><|835|><|1583|><|486|><|1249|><|286|><|937|><|1076|><|1150|><|614|><|42|><|1058|><|705|><|681|><|798|><|934|><|490|><|514|><|1399|><|572|><|1446|><|1703|><|1346|><|1040|><|1426|><|1304|><|664|><|171|><|1530|><|625|><|64|><|1708|><|1830|><|1030|><|443|><|1509|><|1063|><|1605|><|1785|><|721|><|1440|><|923|><|code_end|> +#sure<|t_0.36|><|code_start|><|792|><|1780|><|923|><|1640|><|265|><|261|><|1525|><|567|><|1491|><|1250|><|1730|><|362|><|919|><|1766|><|543|><|1|><|333|><|113|><|970|><|252|><|1606|><|133|><|302|><|1810|><|1046|><|1190|><|1675|><|code_end|> +#i<|t_0.08|><|code_start|><|123|><|439|><|1074|><|705|><|1799|><|637|><|code_end|> +#have<|t_0.16|><|code_start|><|1509|><|599|><|518|><|1170|><|552|><|1029|><|1267|><|864|><|419|><|143|><|1061|><|0|><|code_end|> +#some<|t_0.16|><|code_start|><|619|><|400|><|1270|><|62|><|1370|><|1832|><|917|><|1661|><|167|><|269|><|1366|><|1508|><|code_end|> +#critiques<|t_0.60|><|code_start|><|559|><|584|><|1163|><|1129|><|1313|><|1728|><|721|><|1146|><|1093|><|577|><|928|><|27|><|630|><|1080|><|1346|><|1337|><|320|><|1382|><|1175|><|1682|><|1556|><|990|><|1683|><|860|><|1721|><|110|><|786|><|376|><|1085|><|756|><|1523|><|234|><|1334|><|1506|><|1578|><|659|><|612|><|1108|><|1466|><|1647|><|308|><|1470|><|746|><|556|><|1061|><|code_end|> +#about<|t_0.29|><|code_start|><|26|><|1649|><|545|><|1367|><|1263|><|1728|><|450|><|859|><|1434|><|497|><|1220|><|1285|><|179|><|755|><|1154|><|779|><|179|><|1229|><|1213|><|922|><|1774|><|1408|><|code_end|> +#some<|t_0.23|><|code_start|><|986|><|28|><|1649|><|778|><|858|><|1519|><|1|><|18|><|26|><|1042|><|1174|><|1309|><|1499|><|1712|><|1692|><|1516|><|1574|><|code_end|> +#of<|t_0.07|><|code_start|><|197|><|716|><|1039|><|1662|><|64|><|code_end|> +#the<|t_0.08|><|code_start|><|1811|><|1568|><|569|><|886|><|1025|><|1374|><|code_end|> +#gameplay<|t_0.48|><|code_start|><|1269|><|1092|><|933|><|1362|><|1762|><|1700|><|1675|><|215|><|781|><|1086|><|461|><|838|><|1022|><|759|><|649|><|1416|><|1004|><|551|><|909|><|787|><|343|><|830|><|1391|><|1040|><|1622|><|1779|><|1360|><|1231|><|1187|><|1317|><|76|><|997|><|989|><|978|><|737|><|189|><|code_end|> +#aspects<|t_0.56|><|code_start|><|1423|><|797|><|1316|><|1222|><|147|><|719|><|1347|><|386|><|1390|><|1558|><|154|><|440|><|634|><|592|><|1097|><|1718|><|712|><|763|><|1118|><|1721|><|1311|><|868|><|580|><|362|><|1435|><|868|><|247|><|221|><|886|><|1145|><|1274|><|1284|><|457|><|1043|><|1459|><|1818|><|62|><|599|><|1035|><|62|><|1649|><|778|><|code_end|> +#but<|t_0.20|><|code_start|><|780|><|1825|><|1681|><|1007|><|861|><|710|><|702|><|939|><|1669|><|1491|><|613|><|1739|><|823|><|1469|><|648|><|code_end|> +#its<|t_0.09|><|code_start|><|92|><|688|><|1623|><|962|><|1670|><|527|><|599|><|code_end|> +#still<|t_0.27|><|code_start|><|636|><|10|><|1217|><|344|><|713|><|957|><|823|><|154|><|1649|><|1286|><|508|><|214|><|1760|><|1250|><|456|><|1352|><|1368|><|921|><|615|><|5|><|code_end|> +#really<|t_0.36|><|code_start|><|55|><|420|><|1008|><|1659|><|27|><|644|><|1266|><|617|><|761|><|1712|><|109|><|1465|><|1587|><|503|><|1541|><|619|><|197|><|1019|><|817|><|269|><|377|><|362|><|1381|><|507|><|1488|><|4|><|1695|><|code_end|> +#enjoyable<|t_0.49|><|code_start|><|678|><|501|><|864|><|319|><|288|><|1472|><|1341|><|686|><|562|><|1463|><|619|><|1563|><|471|><|911|><|730|><|1811|><|1006|><|520|><|861|><|1274|><|125|><|1431|><|638|><|621|><|153|><|876|><|1770|><|437|><|987|><|1653|><|1109|><|898|><|1285|><|80|><|593|><|1709|><|843|><|code_end|> +#and<|t_0.15|><|code_start|><|1285|><|987|><|303|><|1037|><|730|><|1164|><|502|><|120|><|1737|><|1655|><|1318|><|code_end|> +#it<|t_0.09|><|code_start|><|848|><|1366|><|395|><|1601|><|1513|><|593|><|1302|><|code_end|> +#looks<|t_0.27|><|code_start|><|1281|><|1266|><|1755|><|572|><|248|><|1751|><|1257|><|695|><|1380|><|457|><|659|><|585|><|1315|><|1105|><|1776|><|736|><|24|><|736|><|654|><|1027|><|code_end|> +#lovely<|t_0.56|><|code_start|><|634|><|596|><|1766|><|1556|><|1306|><|1285|><|1481|><|1721|><|1123|><|438|><|1246|><|1251|><|795|><|659|><|1381|><|1658|><|217|><|1772|><|562|><|952|><|107|><|1129|><|1112|><|467|><|550|><|1079|><|840|><|1615|><|1469|><|1380|><|168|><|917|><|836|><|1827|><|437|><|583|><|67|><|595|><|1087|><|1646|><|1493|><|1677|><|code_end|>""" + +# TODO: tokenization is slow for some reason - here is pre-tokenized input +suffix = [ 151667, 198, 1782, 155780, 151669, 151929, 152412, 152308, 152585, 152460, 153375, 151670, 198, 74455, + 155808, 151669, 151799, 151873, 151863, 152446, 152372, 152204, 152728, 152229, 152470, 151970, 153413, + 152419, 153334, 153289, 153374, 153199, 152040, 153260, 152721, 152680, 153297, 152419, 153248, 152400, + 152691, 153368, 153437, 151670, 198, 1722, 155828, 151669, 152607, 152256, 152991, 152299, 152688, 153163, + 153016, 152789, 153198, 152712, 151911, 153107, 152623, 152170, 152395, 152852, 152207, 152461, 153321, + 153309, 151750, 152137, 153340, 152573, 152267, 153347, 151789, 152681, 153339, 151992, 152512, 151751, + 152179, 153434, 153180, 152900, 153440, 152474, 153122, 153129, 151904, 152311, 151670, 198, 1499, 155791, + 151669, 152276, 152454, 153354, 152544, 153204, 153272, 152708, 153433, 152319, 153226, 153043, 152325, + 153267, 152622, 151670, 198, 4250, 155797, 151669, 153454, 153342, 151989, 152458, 153420, 152303, 152271, + 152827, 153036, 153196, 151708, 153263, 152561, 153207, 152213, 152112, 153204, 151722, 152542, 151670, 198, + 19789, 155796, 151669, 153353, 153182, 152345, 152471, 152477, 153014, 152002, 152191, 151734, 152312, 152810, + 152237, 153224, 153169, 153224, 152244, 153387, 153404, 151670, 198, 16069, 155811, 151669, 152265, 151946, + 151808, 152412, 152363, 152305, 153156, 152733, 152810, 153157, 152016, 152100, 152069, 153234, 152317, + 152589, 152707, 153121, 153341, 152159, 152114, 153156, 153001, 153504, 153376, 152272, 152433, 152325, + 151941, 151670, 198, 285, 155788, 151669, 152238, 152255, 153427, 152318, 153009, 152381, 152474, 152680, + 152157, 153255, 152324, 151682, 151670, 198, 32955, 155804, 151669, 153490, 153419, 152364, 152405, 152682, + 152206, 152078, 153369, 152725, 153193, 153027, 152946, 152488, 153070, 151883, 152890, 152489, 153144, + 153375, 152358, 151685, 152494, 152117, 152740, 151670, 198, 37448, 480, 155840, 151669, 151902, 152720, + 153377, 152027, 152378, 152821, 153207, 153459, 153028, 153068, 152507, 153255, 152158, 152921, 151958, + 152609, 152748, 152822, 152286, 151714, 152730, 152377, 152353, 152470, 152606, 152162, 152186, 153071, + 152244, 153118, 153375, 153018, 152712, 153098, 152976, 152336, 151843, 153202, 152297, 151736, 153380, + 153502, 152702, 152115, 153181, 152735, 153277, 153457, 152393, 153112, 152595, 151670, 198, 19098, 155808, + 151669, 152464, 153452, 152595, 153312, 151937, 151933, 153197, 152239, 153163, 152922, 153402, 152034, + 152591, 153438, 152215, 151673, 152005, 151785, 152642, 151924, 153278, 151805, 151974, 153482, 152718, + 152862, 153347, 151670, 198, 72, 155780, 151669, 151795, 152111, 152746, 152377, 153471, 152309, 151670, 198, + 19016, 155788, 151669, 153181, 152271, 152190, 152842, 152224, 152701, 152939, 152536, 152091, 151815, 152733, + 151672, 151670, 198, 14689, 155788, 151669, 152291, 152072, 152942, 151734, 153042, 153504, 152589, 153333, + 151839, 151941, 153038, 153180, 151670, 198, 36996, 8303, 155832, 151669, 152231, 152256, 152835, 152801, + 152985, 153400, 152393, 152818, 152765, 152249, 152600, 151699, 152302, 152752, 153018, 153009, 151992, + 153054, 152847, 153354, 153228, 152662, 153355, 152532, 153393, 151782, 152458, 152048, 152757, 152428, + 153195, 151906, 153006, 153178, 153250, 152331, 152284, 152780, 153138, 153319, 151980, 153142, 152418, + 152228, 152733, 151670, 198, 9096, 155801, 151669, 151698, 153321, 152217, 153039, 152935, 153400, 152122, + 152531, 153106, 152169, 152892, 152957, 151851, 152427, 152826, 152451, 151851, 152901, 152885, 152594, + 153446, 153080, 151670, 198, 14689, 155795, 151669, 152658, 151700, 153321, 152450, 152530, 153191, 151673, + 151690, 151698, 152714, 152846, 152981, 153171, 153384, 153364, 153188, 153246, 151670, 198, 1055, 155779, + 151669, 151869, 152388, 152711, 153334, 151736, 151670, 198, 1782, 155780, 151669, 153483, 153240, 152241, + 152558, 152697, 153046, 151670, 198, 5804, 1363, 155820, 151669, 152941, 152764, 152605, 153034, 153434, + 153372, 153347, 151887, 152453, 152758, 152133, 152510, 152694, 152431, 152321, 153088, 152676, 152223, + 152581, 152459, 152015, 152502, 153063, 152712, 153294, 153451, 153032, 152903, 152859, 152989, 151748, + 152669, 152661, 152650, 152409, 151861, 151670, 198, 300, 7973, 155828, 151669, 153095, 152469, 152988, + 152894, 151819, 152391, 153019, 152058, 153062, 153230, 151826, 152112, 152306, 152264, 152769, 153390, + 152384, 152435, 152790, 153393, 152983, 152540, 152252, 152034, 153107, 152540, 151919, 151893, 152558, + 152817, 152946, 152956, 152129, 152715, 153131, 153490, 151734, 152271, 152707, 151734, 153321, 152450, + 151670, 198, 8088, 155792, 151669, 152452, 153497, 153353, 152679, 152533, 152382, 152374, 152611, 153341, + 153163, 152285, 153411, 152495, 153141, 152320, 151670, 198, 1199, 155781, 151669, 151764, 152360, 153295, + 152634, 153342, 152199, 152271, 151670, 198, 43366, 155799, 151669, 152308, 151682, 152889, 152016, 152385, + 152629, 152495, 151826, 153321, 152958, 152180, 151886, 153432, 152922, 152128, 153024, 153040, 152593, + 152287, 151677, 151670, 198, 53660, 155808, 151669, 151727, 152092, 152680, 153331, 151699, 152316, 152938, + 152289, 152433, 153384, 151781, 153137, 153259, 152175, 153213, 152291, 151869, 152691, 152489, 151941, + 152049, 152034, 153053, 152179, 153160, 151676, 153367, 151670, 198, 268, 4123, 480, 155821, 151669, 152350, + 152173, 152536, 151991, 151960, 153144, 153013, 152358, 152234, 153135, 152291, 153235, 152143, 152583, + 152402, 153483, 152678, 152192, 152533, 152946, 151797, 153103, 152310, 152293, 151825, 152548, 153442, + 152109, 152659, 153325, 152781, 152570, 152957, 151752, 152265, 153381, 152515, 151670, 198, 437, 155787, + 151669, 152957, 152659, 151975, 152709, 152402, 152836, 152174, 151792, 153409, 153327, 152990, 151670, 198, + 275, 155781, 151669, 152520, 153038, 152067, 153273, 153185, 152265, 152974, 151670, 198, 94273, 155799, + 151669, 152953, 152938, 153427, 152244, 151920, 153423, 152929, 152367, 153052, 152129, 152331, 152257, + 152987, 152777, 153448, 152408, 151696, 152408, 152326, 152699, 151670, 198, 385, 16239, 155828, 151669, + 152306, 152268, 153438, 153228, 152978, 152957, 153153, 153393, 152795, 152110, 152918, 152923, 152467, + 152331, 153053, 153330, 151889, 153444, 152234, 152624, 151779, 152801, 152784, 152139, 152222, 152751, + 152512, 153287, 153141, 153052, 151840, 152589, 152508, 153499, 152109, 152255, 151739, 152267, 152759, + 153318, 153165, 153349, 151670, ] + +response = requests.post( + host_llm + "/completion", + json={ + "prompt": [prefix + words, *suffix], + "n_predict": 1024, + "cache_prompt": True, + "return_tokens": True, + "samplers": ["top_k"], + "top_k": 16, + "seed": 1003, + } +) + +response_json = response.json() + +#print(json.dumps(response_json, indent=4)) +#print(json.dumps(response_json["prompt"], indent=4).replace("\\n", "\n")) +#print(json.dumps(response_json["timings"], indent=4)) +#print(json.dumps(response_json["tokens"], indent=4)) + +codes = response_json["tokens"] + +codes = [t - 151672 for t in codes if t >= 151672 and t <= 155772] + +response = requests.post( + host_dec + "/embeddings", + json={ + "input": [*codes], + } +) + +response_json = response.json() + +#print(json.dumps(response_json, indent=4)) + +# spectrogram +embd = response_json[0]["embedding"] + +n_codes = len(embd) +n_embd = len(embd[0]) + +print('spectrogram generated: n_codes: %d, n_embd: %d' % (n_codes, n_embd)) + +# post-process the spectrogram to convert to audio +print('converting to audio ...') +audio = embd_to_audio(embd, n_codes, n_embd) +print('audio generated: %d samples' % len(audio)) + +filename = "output.wav" +sample_rate = 24000 # sampling rate + +# zero out first 0.25 seconds +audio[:24000 // 4] = 0.0 + +save_wav(filename, audio, sample_rate) +print('audio written to file "%s"' % filename) diff --git a/examples/tts/tts.cpp b/examples/tts/tts.cpp new file mode 100644 index 000000000..5a9161181 --- /dev/null +++ b/examples/tts/tts.cpp @@ -0,0 +1,930 @@ +#include "arg.h" +#include "common.h" +#include "sampling.h" +#include "log.h" +#include "llama.h" + +#define _USE_MATH_DEFINES // For M_PI on MSVC + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +// +// Terminal utils +// + +#define SQR(X) ((X) * (X)) +#define UNCUBE(x) x < 48 ? 0 : x < 115 ? 1 : (x - 35) / 40 + +/** + * Quantizes 24-bit RGB to xterm256 code range [16,256). + */ +static int rgb2xterm256(int r, int g, int b) { + unsigned char cube[] = {0, 0137, 0207, 0257, 0327, 0377}; + int av, ir, ig, ib, il, qr, qg, qb, ql; + av = r * .299 + g * .587 + b * .114 + .5; + ql = (il = av > 238 ? 23 : (av - 3) / 10) * 10 + 8; + qr = cube[(ir = UNCUBE(r))]; + qg = cube[(ig = UNCUBE(g))]; + qb = cube[(ib = UNCUBE(b))]; + if (SQR(qr - r) + SQR(qg - g) + SQR(qb - b) <= + SQR(ql - r) + SQR(ql - g) + SQR(ql - b)) + return ir * 36 + ig * 6 + ib + 020; + return il + 0350; +} + +static std::string set_xterm256_foreground(int r, int g, int b) { + int x = rgb2xterm256(r, g, b); + std::ostringstream oss; + oss << "\033[38;5;" << x << "m"; + return oss.str(); +} + +const std::vector k_colors = { + set_xterm256_foreground(220, 5, 12), + set_xterm256_foreground(232, 96, 28), + set_xterm256_foreground(241, 147, 45), + set_xterm256_foreground(246, 193, 65), + set_xterm256_foreground(247, 240, 86), + set_xterm256_foreground(144, 201, 135), + set_xterm256_foreground( 78, 178, 101), +}; + +static void print_usage(int, char ** argv) { + LOG("\nexample usage:\n"); + LOG("\n %s -m model.gguf -p \"Hello!\"\n", argv[0]); + LOG("\n"); +} + +struct wav_header { + char riff[4] = {'R', 'I', 'F', 'F'}; + uint32_t chunk_size; + char wave[4] = {'W', 'A', 'V', 'E'}; + char fmt[4] = {'f', 'm', 't', ' '}; + uint32_t fmt_chunk_size = 16; + uint16_t audio_format = 1; // PCM + uint16_t num_channels = 1; // Mono + uint32_t sample_rate; + uint32_t byte_rate; + uint16_t block_align; + uint16_t bits_per_sample = 16; + char data[4] = {'d', 'a', 't', 'a'}; + uint32_t data_size; +}; + +static void save_wav16(const std::string & fname, const std::vector & data, int sample_rate) { + std::ofstream file(fname, std::ios::binary); + if (!file) { + LOG_ERR("%s: Failed to open file '%s' for writing", __func__, fname.c_str()); + return; + } + + wav_header header; + header.sample_rate = sample_rate; + header.byte_rate = header.sample_rate * header.num_channels * (header.bits_per_sample / 8); + header.block_align = header.num_channels * (header.bits_per_sample / 8); + header.data_size = data.size() * (header.bits_per_sample / 8); + header.chunk_size = 36 + header.data_size; + + file.write(reinterpret_cast(&header), sizeof(header)); + + for (const auto & sample : data) { + int16_t pcm_sample = static_cast(std::clamp(sample * 32767.0, -32768.0, 32767.0)); + file.write(reinterpret_cast(&pcm_sample), sizeof(pcm_sample)); + } + + file.close(); +} + +static void fill_hann_window(int length, bool periodic, float * output) { + int offset = -1; + if (periodic) { + offset = 0; + } + for (int i = 0; i < length; i++) { + output[i] = 0.5 * (1.0 - cosf((2.0 * M_PI * i) / (length + offset))); + } +} + +// very poor-man fft +static void twiddle(float * real, float * imag, int k, int N) { + float angle = 2 * M_PI * k / N; + *real = cos(angle); + *imag = sin(angle); +} + +static void irfft(int n, const float * inp_cplx, float * out_real) { + int N = n / 2 + 1; + + std::vector real_input(N); + std::vector imag_input(N); + for (int i = 0; i < N; ++i) { + real_input[i] = inp_cplx[2 * i]; + imag_input[i] = inp_cplx[2 * i + 1]; + } + + std::vector real_output(n); + std::vector imag_output(n); + + for (int k = 0; k < n; ++k) { + real_output[k] = 0.0f; + imag_output[k] = 0.0f; + for (int m = 0; m < N; ++m) { + float twiddle_real; + float twiddle_imag; + + twiddle(&twiddle_real, &twiddle_imag, k * m, n); + + real_output[k] += real_input[m] * twiddle_real - imag_input[m] * twiddle_imag; + imag_output[k] += real_input[m] * twiddle_imag + imag_input[m] * twiddle_real; + } + } + + for (int i = 0; i < n; ++i) { + out_real[i] = real_output[i] / N; + } +} + +// +// y = torch.nn.functional.fold( +// data, output_size=(1, output_size), kernel_size=(1, self.win_length), stride=(1, self.hop_length), +// )[:, 0, 0, pad:-pad] +// +// data.shape = torch.Size([1, 1280, 261]) +// output_size = 84480 +// win_length = 1280 +// hop_length = 320 +// pad = 480 +// +static void fold(const std::vector & data, int64_t n_out, int64_t n_win, int64_t n_hop, int64_t n_pad, std::vector & output) { + int64_t output_height = n_out; + int64_t kernel_w = n_win; + int64_t stride_w = n_hop; + int64_t width = n_out; + + output.resize(width, 0.0f); + + int64_t col_idx = 0; + for (int64_t w_col = 0; w_col < width; ++w_col) { + int64_t start = w_col * stride_w - n_pad; + int64_t end = start + kernel_w; + + for (int64_t w_im = start; w_im < end; ++w_im) { + if (w_im >= 0 && w_im < output_height && col_idx < (int64_t) data.size()) { + output[w_im] += data[col_idx]; + } + col_idx++; + } + } + + output.resize(n_out - 2 * n_pad); +} + +// TODO: not optimized at all +static std::vector embd_to_audio( + const float * embd, + const int n_codes, + const int n_embd, + const int n_thread) { + const int n_fft = 1280; + const int n_hop = 320; + const int n_win = 1280; + const int n_pad = (n_win - n_hop)/2; + const int n_out = (n_codes - 1)*n_hop + n_win; + + std::vector hann(n_fft); + + fill_hann_window(hann.size(), true, hann.data()); + + int n_spec = n_embd*n_codes; + + std::vector E (n_spec); + std::vector S (n_spec); + std::vector ST(n_spec); + + for (int l = 0; l < n_codes; ++l) { + for (int k = 0; k < n_embd; ++k) { + E[k*n_codes + l] = embd[l*n_embd + k]; + } + } + + for (int k = 0; k < n_embd/2; ++k) { + for (int l = 0; l < n_codes; ++l) { + float mag = E[(k )*n_codes + l]; + float phi = E[(k + n_embd/2)*n_codes + l]; + + mag = exp(mag); + + if (mag > 1e2) { + mag = 1e2; + } + S[2*(k*n_codes + l) + 0] = mag*cosf(phi); + S[2*(k*n_codes + l) + 1] = mag*sinf(phi); + } + } + + for (int l = 0; l < n_codes; ++l) { + for (int k = 0; k < n_embd/2; ++k) { + ST[l*n_embd + 2*k + 0] = S[2*(k*n_codes + l) + 0]; + ST[l*n_embd + 2*k + 1] = S[2*(k*n_codes + l) + 1]; + } + } + + std::vector res (n_codes*n_fft); + std::vector hann2(n_codes*n_fft); + + std::vector workers(n_thread); + for (int i = 0; i < n_thread; ++i) { + workers[i] = std::thread([&, i]() { + for (int l = i; l < n_codes; l += n_thread) { + irfft(n_fft, ST.data() + l*n_embd, res.data() + l*n_fft); + for (int j = 0; j < n_fft; ++j) { + res [l*n_fft + j] *= hann[j]; + hann2[l*n_fft + j] = hann[j] * hann[j]; + } + } + }); + } + for (int i = 0; i < n_thread; ++i) { + workers[i].join(); + } + + std::vector audio; + std::vector env; + + fold(res, n_out, n_win, n_hop, n_pad, audio); + fold(hann2, n_out, n_win, n_hop, n_pad, env); // TODO: can be done once + + for (size_t i = 0; i < audio.size(); ++i) { + audio[i] /= env[i]; + } + + return audio; +} + +static const std::map ones = { + {0, "zero"}, {1, "one"}, {2, "two"}, {3, "three"}, {4, "four"}, + {5, "five"}, {6, "six"}, {7, "seven"}, {8, "eight"}, {9, "nine"}, + {10, "ten"}, {11, "eleven"}, {12, "twelve"}, {13, "thirteen"}, {14, "fourteen"}, + {15, "fifteen"}, {16, "sixteen"}, {17, "seventeen"}, {18, "eighteen"}, {19, "nineteen"} +}; + +static const std::map tens = { + {2, "twenty"}, {3, "thirty"}, {4, "forty"}, {5, "fifty"}, + {6, "sixty"}, {7, "seventy"}, {8, "eighty"}, {9, "ninety"} +}; + +// Convert a number less than 1000 to words +static std::string convert_less_than_thousand(int num) { + std::string result; + + if (num >= 100) { + result += ones.at(num / 100) + " hundred "; + num %= 100; + } + + if (num >= 20) { + result += tens.at(num / 10); + if (num % 10 > 0) { + result += "-" + ones.at(num % 10); + } + } else if (num > 0) { + result += ones.at(num); + } + + return result; +} + +static std::string number_to_words(const std::string & number_str) { + try { + size_t decimal_pos = number_str.find('.'); + std::string integer_part = number_str.substr(0, decimal_pos); + + int int_number = std::stoi(integer_part); + std::string result; + + if (int_number == 0) { + result = "zero"; + } else { + if (int_number >= 1000000000) { + int billions = int_number / 1000000000; + result += convert_less_than_thousand(billions) + " billion "; + int_number %= 1000000000; + } + + if (int_number >= 1000000) { + int millions = int_number / 1000000; + result += convert_less_than_thousand(millions) + " million "; + int_number %= 1000000; + } + + if (int_number >= 1000) { + int thousands = int_number / 1000; + result += convert_less_than_thousand(thousands) + " thousand "; + int_number %= 1000; + } + + if (int_number > 0) { + result += convert_less_than_thousand(int_number); + } + } + + // Handle decimal part + if (decimal_pos != std::string::npos) { + result += " point"; + std::string decimal_part = number_str.substr(decimal_pos + 1); + for (char digit : decimal_part) { + result += " " + ones.at(digit - '0'); + } + } + + return result; + } catch (const std::exception& e) { + // Skip if fails + return " "; + } +} + +static std::string replace_numbers_with_words(const std::string & input_text) { + std::regex number_pattern(R"(\d+(\.\d+)?)"); + std::string result; + auto it = std::sregex_iterator(input_text.begin(), input_text.end(), number_pattern); + auto end = std::sregex_iterator(); + + size_t last_pos = 0; + for (std::sregex_iterator i = it; i != end; ++i) { + const std::smatch& match = *i; + result.append(input_text, last_pos, match.position() - last_pos); + result.append(number_to_words(match.str())); + last_pos = match.position() + match.length(); + } + result.append(input_text, last_pos); + + return result; +} + +// Based on: https://github.com/edwko/OuteTTS/blob/a613e79c489d8256dd657ea9168d78de75895d82/outetts/version/v1/prompt_processor.py#L39 +static std::string process_text(const std::string & text) { + + // For now I skipped text romanization as I am unsure how to handle + // uroman and MeCab implementations in C++ + // maybe something like https://github.com/anyascii/anyascii/ could work. + // currently only English would be supported in this function + + std::string processed_text = replace_numbers_with_words(text); + + std::transform(processed_text.begin(), processed_text.end(), + processed_text.begin(), ::tolower); + + std::regex special_chars(R"([-_/,\.\\])"); + processed_text = std::regex_replace(processed_text, special_chars, " "); + + std::regex non_alpha(R"([^a-z\s])"); + processed_text = std::regex_replace(processed_text, non_alpha, ""); + + std::regex multiple_spaces(R"(\s+)"); + processed_text = std::regex_replace(processed_text, multiple_spaces, " "); + + processed_text = std::regex_replace(processed_text, std::regex(R"(^\s+|\s+$)"), ""); + + /* + Replace spaces with the separator token same as in line 365 + + for (auto & c : prompt_user) { + if (c == ' ') { + prompt_clean += "<|text_sep|>"; + */ + processed_text = std::regex_replace(processed_text, std::regex(R"(\s)"), "<|text_sep|>"); + + return processed_text; +} + +static void prompt_add(llama_tokens & prompt, llama_token token) { + prompt.push_back(token); +} + +static void prompt_add(llama_tokens & prompt, const llama_tokens & tokens) { + prompt.insert(prompt.end(), tokens.begin(), tokens.end()); +} + +static void prompt_add(llama_tokens & prompt, const llama_vocab * vocab, const std::string & txt, bool add_special, bool parse_special) { + auto tmp = common_tokenize(vocab, txt, add_special, parse_special); + prompt_add(prompt, tmp); +} + +static void prompt_init(llama_tokens & prompt, const llama_vocab * vocab) { + prompt.clear(); + + prompt_add(prompt, vocab, "<|im_start|>\n", true, true); +} + +int main(int argc, char ** argv) { + common_params params; + + params.prompt = ""; + + params.n_predict = 4096; + params.n_batch = 8192; + params.n_ctx = 8192; + + params.sampling.top_k = 4; + params.sampling.samplers = { COMMON_SAMPLER_TYPE_TOP_K, }; + + if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_TTS, print_usage)) { + return 1; + } + + const int n_parallel = params.n_parallel; + const int n_predict = params.n_predict; + + common_init(); + + // init LLM + + llama_backend_init(); + llama_numa_init(params.numa); + + llama_model * model_ttc = NULL; // text-to-codes + llama_model * model_cts = NULL; // codes-to-speech + + llama_context * ctx_ttc = NULL; + llama_context * ctx_cts = NULL; + + common_init_result llama_init_ttc = common_init_from_params(params); + + model_ttc = llama_init_ttc.model.get(); + ctx_ttc = llama_init_ttc.context.get(); + + const llama_vocab * vocab = llama_model_get_vocab(model_ttc); + + // TODO: refactor in a common struct + params.model = params.vocoder.model; + params.model_url = params.vocoder.model_url; + params.hf_repo = params.vocoder.hf_repo; + params.hf_file = params.vocoder.hf_file; + + params.embedding = true; + + common_init_result llama_init_cts = common_init_from_params(params); + + model_cts = llama_init_cts.model.get(); + ctx_cts = llama_init_cts.context.get(); + + std::vector smpl(n_parallel); + for (int i = 0; i < n_parallel; ++i) { + params.sampling.no_perf = (i != 0); + params.sampling.seed = params.sampling.seed + 1; + + smpl[i] = common_sampler_init(model_ttc, params.sampling); + } + + LOG_INF("sampler seed: %u\n", common_sampler_get_seed(smpl[0])); + LOG_INF("sampler params: \n%s\n", params.sampling.print().c_str()); + LOG_INF("sampler chain: %s\n", common_sampler_print(smpl[0]).c_str()); + + LOG_INF("%s: loading done\n", __func__); + + const auto t_main_start = ggml_time_us(); + + std::vector codes; + + // process prompt and generate voice codes + { + LOG_INF("%s: constructing prompt ..\n", __func__); + + std::vector prompt_inp; + + prompt_init(prompt_inp, vocab); + + prompt_add(prompt_inp, vocab, "<|text_start|>the<|text_sep|>overall<|text_sep|>package<|text_sep|>from<|text_sep|>just<|text_sep|>two<|text_sep|>people<|text_sep|>is<|text_sep|>pretty<|text_sep|>remarkable<|text_sep|>sure<|text_sep|>i<|text_sep|>have<|text_sep|>some<|text_sep|>critiques<|text_sep|>about<|text_sep|>some<|text_sep|>of<|text_sep|>the<|text_sep|>gameplay<|text_sep|>aspects<|text_sep|>but<|text_sep|>its<|text_sep|>still<|text_sep|>really<|text_sep|>enjoyable<|text_sep|>and<|text_sep|>it<|text_sep|>looks<|text_sep|>lovely<|text_sep|>", false, true); + + // convert the input text into the necessary format expected by OuteTTS + { + std::string prompt_clean = process_text(params.prompt); + + LOG_INF("%s: prompt: '%s'\n", __func__, prompt_clean.c_str()); + + prompt_add(prompt_inp, vocab, prompt_clean, false, true); + } + + prompt_add(prompt_inp, vocab, "<|text_end|>\n", false, true); + + // disabled to save time on tokenizing each time + // TODO: load voices from the json files +#if 0 + const std::string voice_data = R"(<|audio_start|> +the<|t_0.08|><|code_start|><|257|><|740|><|636|><|913|><|788|><|1703|><|code_end|> +overall<|t_0.36|><|code_start|><|127|><|201|><|191|><|774|><|700|><|532|><|1056|><|557|><|798|><|298|><|1741|><|747|><|1662|><|1617|><|1702|><|1527|><|368|><|1588|><|1049|><|1008|><|1625|><|747|><|1576|><|728|><|1019|><|1696|><|1765|><|code_end|> +package<|t_0.56|><|code_start|><|935|><|584|><|1319|><|627|><|1016|><|1491|><|1344|><|1117|><|1526|><|1040|><|239|><|1435|><|951|><|498|><|723|><|1180|><|535|><|789|><|1649|><|1637|><|78|><|465|><|1668|><|901|><|595|><|1675|><|117|><|1009|><|1667|><|320|><|840|><|79|><|507|><|1762|><|1508|><|1228|><|1768|><|802|><|1450|><|1457|><|232|><|639|><|code_end|> +from<|t_0.19|><|code_start|><|604|><|782|><|1682|><|872|><|1532|><|1600|><|1036|><|1761|><|647|><|1554|><|1371|><|653|><|1595|><|950|><|code_end|> +just<|t_0.25|><|code_start|><|1782|><|1670|><|317|><|786|><|1748|><|631|><|599|><|1155|><|1364|><|1524|><|36|><|1591|><|889|><|1535|><|541|><|440|><|1532|><|50|><|870|><|code_end|> +two<|t_0.24|><|code_start|><|1681|><|1510|><|673|><|799|><|805|><|1342|><|330|><|519|><|62|><|640|><|1138|><|565|><|1552|><|1497|><|1552|><|572|><|1715|><|1732|><|code_end|> +people<|t_0.39|><|code_start|><|593|><|274|><|136|><|740|><|691|><|633|><|1484|><|1061|><|1138|><|1485|><|344|><|428|><|397|><|1562|><|645|><|917|><|1035|><|1449|><|1669|><|487|><|442|><|1484|><|1329|><|1832|><|1704|><|600|><|761|><|653|><|269|><|code_end|> +is<|t_0.16|><|code_start|><|566|><|583|><|1755|><|646|><|1337|><|709|><|802|><|1008|><|485|><|1583|><|652|><|10|><|code_end|> +pretty<|t_0.32|><|code_start|><|1818|><|1747|><|692|><|733|><|1010|><|534|><|406|><|1697|><|1053|><|1521|><|1355|><|1274|><|816|><|1398|><|211|><|1218|><|817|><|1472|><|1703|><|686|><|13|><|822|><|445|><|1068|><|code_end|> +remarkable<|t_0.68|><|code_start|><|230|><|1048|><|1705|><|355|><|706|><|1149|><|1535|><|1787|><|1356|><|1396|><|835|><|1583|><|486|><|1249|><|286|><|937|><|1076|><|1150|><|614|><|42|><|1058|><|705|><|681|><|798|><|934|><|490|><|514|><|1399|><|572|><|1446|><|1703|><|1346|><|1040|><|1426|><|1304|><|664|><|171|><|1530|><|625|><|64|><|1708|><|1830|><|1030|><|443|><|1509|><|1063|><|1605|><|1785|><|721|><|1440|><|923|><|code_end|> +sure<|t_0.36|><|code_start|><|792|><|1780|><|923|><|1640|><|265|><|261|><|1525|><|567|><|1491|><|1250|><|1730|><|362|><|919|><|1766|><|543|><|1|><|333|><|113|><|970|><|252|><|1606|><|133|><|302|><|1810|><|1046|><|1190|><|1675|><|code_end|> +i<|t_0.08|><|code_start|><|123|><|439|><|1074|><|705|><|1799|><|637|><|code_end|> +have<|t_0.16|><|code_start|><|1509|><|599|><|518|><|1170|><|552|><|1029|><|1267|><|864|><|419|><|143|><|1061|><|0|><|code_end|> +some<|t_0.16|><|code_start|><|619|><|400|><|1270|><|62|><|1370|><|1832|><|917|><|1661|><|167|><|269|><|1366|><|1508|><|code_end|> +critiques<|t_0.60|><|code_start|><|559|><|584|><|1163|><|1129|><|1313|><|1728|><|721|><|1146|><|1093|><|577|><|928|><|27|><|630|><|1080|><|1346|><|1337|><|320|><|1382|><|1175|><|1682|><|1556|><|990|><|1683|><|860|><|1721|><|110|><|786|><|376|><|1085|><|756|><|1523|><|234|><|1334|><|1506|><|1578|><|659|><|612|><|1108|><|1466|><|1647|><|308|><|1470|><|746|><|556|><|1061|><|code_end|> +about<|t_0.29|><|code_start|><|26|><|1649|><|545|><|1367|><|1263|><|1728|><|450|><|859|><|1434|><|497|><|1220|><|1285|><|179|><|755|><|1154|><|779|><|179|><|1229|><|1213|><|922|><|1774|><|1408|><|code_end|> +some<|t_0.23|><|code_start|><|986|><|28|><|1649|><|778|><|858|><|1519|><|1|><|18|><|26|><|1042|><|1174|><|1309|><|1499|><|1712|><|1692|><|1516|><|1574|><|code_end|> +of<|t_0.07|><|code_start|><|197|><|716|><|1039|><|1662|><|64|><|code_end|> +the<|t_0.08|><|code_start|><|1811|><|1568|><|569|><|886|><|1025|><|1374|><|code_end|> +gameplay<|t_0.48|><|code_start|><|1269|><|1092|><|933|><|1362|><|1762|><|1700|><|1675|><|215|><|781|><|1086|><|461|><|838|><|1022|><|759|><|649|><|1416|><|1004|><|551|><|909|><|787|><|343|><|830|><|1391|><|1040|><|1622|><|1779|><|1360|><|1231|><|1187|><|1317|><|76|><|997|><|989|><|978|><|737|><|189|><|code_end|> +aspects<|t_0.56|><|code_start|><|1423|><|797|><|1316|><|1222|><|147|><|719|><|1347|><|386|><|1390|><|1558|><|154|><|440|><|634|><|592|><|1097|><|1718|><|712|><|763|><|1118|><|1721|><|1311|><|868|><|580|><|362|><|1435|><|868|><|247|><|221|><|886|><|1145|><|1274|><|1284|><|457|><|1043|><|1459|><|1818|><|62|><|599|><|1035|><|62|><|1649|><|778|><|code_end|> +but<|t_0.20|><|code_start|><|780|><|1825|><|1681|><|1007|><|861|><|710|><|702|><|939|><|1669|><|1491|><|613|><|1739|><|823|><|1469|><|648|><|code_end|> +its<|t_0.09|><|code_start|><|92|><|688|><|1623|><|962|><|1670|><|527|><|599|><|code_end|> +still<|t_0.27|><|code_start|><|636|><|10|><|1217|><|344|><|713|><|957|><|823|><|154|><|1649|><|1286|><|508|><|214|><|1760|><|1250|><|456|><|1352|><|1368|><|921|><|615|><|5|><|code_end|> +really<|t_0.36|><|code_start|><|55|><|420|><|1008|><|1659|><|27|><|644|><|1266|><|617|><|761|><|1712|><|109|><|1465|><|1587|><|503|><|1541|><|619|><|197|><|1019|><|817|><|269|><|377|><|362|><|1381|><|507|><|1488|><|4|><|1695|><|code_end|> +enjoyable<|t_0.49|><|code_start|><|678|><|501|><|864|><|319|><|288|><|1472|><|1341|><|686|><|562|><|1463|><|619|><|1563|><|471|><|911|><|730|><|1811|><|1006|><|520|><|861|><|1274|><|125|><|1431|><|638|><|621|><|153|><|876|><|1770|><|437|><|987|><|1653|><|1109|><|898|><|1285|><|80|><|593|><|1709|><|843|><|code_end|> +and<|t_0.15|><|code_start|><|1285|><|987|><|303|><|1037|><|730|><|1164|><|502|><|120|><|1737|><|1655|><|1318|><|code_end|> +it<|t_0.09|><|code_start|><|848|><|1366|><|395|><|1601|><|1513|><|593|><|1302|><|code_end|> +looks<|t_0.27|><|code_start|><|1281|><|1266|><|1755|><|572|><|248|><|1751|><|1257|><|695|><|1380|><|457|><|659|><|585|><|1315|><|1105|><|1776|><|736|><|24|><|736|><|654|><|1027|><|code_end|> +lovely<|t_0.56|><|code_start|><|634|><|596|><|1766|><|1556|><|1306|><|1285|><|1481|><|1721|><|1123|><|438|><|1246|><|1251|><|795|><|659|><|1381|><|1658|><|217|><|1772|><|562|><|952|><|107|><|1129|><|1112|><|467|><|550|><|1079|><|840|><|1615|><|1469|><|1380|><|168|><|917|><|836|><|1827|><|437|><|583|><|67|><|595|><|1087|><|1646|><|1493|><|1677|><|code_end|>)"; + + auto tmp = common_tokenize(vocab, voice_data, false, true); + printf("\n\n"); + for (int i = 0; i < tmp.size(); ++i) { + printf("%d, ", tmp[i]); + } + printf("\n\n"); +#else + prompt_add(prompt_inp, llama_tokens { + 151667, 198, 1782, 155780, 151669, 151929, 152412, 152308, 152585, + 152460, 153375, 151670, 198, 74455, 155808, 151669, 151799, + 151873, 151863, 152446, 152372, 152204, 152728, 152229, 152470, + 151970, 153413, 152419, 153334, 153289, 153374, 153199, 152040, + 153260, 152721, 152680, 153297, 152419, 153248, 152400, 152691, + 153368, 153437, 151670, 198, 1722, 155828, 151669, 152607, + 152256, 152991, 152299, 152688, 153163, 153016, 152789, 153198, + 152712, 151911, 153107, 152623, 152170, 152395, 152852, 152207, + 152461, 153321, 153309, 151750, 152137, 153340, 152573, 152267, + 153347, 151789, 152681, 153339, 151992, 152512, 151751, 152179, + 153434, 153180, 152900, 153440, 152474, 153122, 153129, 151904, + 152311, 151670, 198, 1499, 155791, 151669, 152276, 152454, + 153354, 152544, 153204, 153272, 152708, 153433, 152319, 153226, + 153043, 152325, 153267, 152622, 151670, 198, 4250, 155797, + 151669, 153454, 153342, 151989, 152458, 153420, 152303, 152271, + 152827, 153036, 153196, 151708, 153263, 152561, 153207, 152213, + 152112, 153204, 151722, 152542, 151670, 198, 19789, 155796, + 151669, 153353, 153182, 152345, 152471, 152477, 153014, 152002, + 152191, 151734, 152312, 152810, 152237, 153224, 153169, 153224, + 152244, 153387, 153404, 151670, 198, 16069, 155811, 151669, + 152265, 151946, 151808, 152412, 152363, 152305, 153156, 152733, + 152810, 153157, 152016, 152100, 152069, 153234, 152317, 152589, + 152707, 153121, 153341, 152159, 152114, 153156, 153001, 153504, + 153376, 152272, 152433, 152325, 151941, 151670, 198, 285, + 155788, 151669, 152238, 152255, 153427, 152318, 153009, 152381, + 152474, 152680, 152157, 153255, 152324, 151682, 151670, 198, + 32955, 155804, 151669, 153490, 153419, 152364, 152405, 152682, + 152206, 152078, 153369, 152725, 153193, 153027, 152946, 152488, + 153070, 151883, 152890, 152489, 153144, 153375, 152358, 151685, + 152494, 152117, 152740, 151670, 198, 37448, 480, 155840, 151669, + 151902, 152720, 153377, 152027, 152378, 152821, 153207, 153459, + 153028, 153068, 152507, 153255, 152158, 152921, 151958, 152609, + 152748, 152822, 152286, 151714, 152730, 152377, 152353, 152470, + 152606, 152162, 152186, 153071, 152244, 153118, 153375, 153018, + 152712, 153098, 152976, 152336, 151843, 153202, 152297, 151736, + 153380, 153502, 152702, 152115, 153181, 152735, 153277, 153457, + 152393, 153112, 152595, 151670, 198, 19098, 155808, 151669, + 152464, 153452, 152595, 153312, 151937, 151933, 153197, 152239, + 153163, 152922, 153402, 152034, 152591, 153438, 152215, 151673, + 152005, 151785, 152642, 151924, 153278, 151805, 151974, 153482, + 152718, 152862, 153347, 151670, 198, 72, 155780, 151669, 151795, + 152111, 152746, 152377, 153471, 152309, 151670, 198, 19016, + 155788, 151669, 153181, 152271, 152190, 152842, 152224, 152701, + 152939, 152536, 152091, 151815, 152733, 151672, 151670, 198, + 14689, 155788, 151669, 152291, 152072, 152942, 151734, 153042, + 153504, 152589, 153333, 151839, 151941, 153038, 153180, 151670, + 198, 36996, 8303, 155832, 151669, 152231, 152256, 152835, + 152801, 152985, 153400, 152393, 152818, 152765, 152249, 152600, + 151699, 152302, 152752, 153018, 153009, 151992, 153054, 152847, + 153354, 153228, 152662, 153355, 152532, 153393, 151782, 152458, + 152048, 152757, 152428, 153195, 151906, 153006, 153178, 153250, + 152331, 152284, 152780, 153138, 153319, 151980, 153142, 152418, + 152228, 152733, 151670, 198, 9096, 155801, 151669, 151698, + 153321, 152217, 153039, 152935, 153400, 152122, 152531, 153106, + 152169, 152892, 152957, 151851, 152427, 152826, 152451, 151851, + 152901, 152885, 152594, 153446, 153080, 151670, 198, 14689, + 155795, 151669, 152658, 151700, 153321, 152450, 152530, 153191, + 151673, 151690, 151698, 152714, 152846, 152981, 153171, 153384, + 153364, 153188, 153246, 151670, 198, 1055, 155779, 151669, + 151869, 152388, 152711, 153334, 151736, 151670, 198, 1782, + 155780, 151669, 153483, 153240, 152241, 152558, 152697, 153046, + 151670, 198, 5804, 1363, 155820, 151669, 152941, 152764, 152605, + 153034, 153434, 153372, 153347, 151887, 152453, 152758, 152133, + 152510, 152694, 152431, 152321, 153088, 152676, 152223, 152581, + 152459, 152015, 152502, 153063, 152712, 153294, 153451, 153032, + 152903, 152859, 152989, 151748, 152669, 152661, 152650, 152409, + 151861, 151670, 198, 300, 7973, 155828, 151669, 153095, 152469, + 152988, 152894, 151819, 152391, 153019, 152058, 153062, 153230, + 151826, 152112, 152306, 152264, 152769, 153390, 152384, 152435, + 152790, 153393, 152983, 152540, 152252, 152034, 153107, 152540, + 151919, 151893, 152558, 152817, 152946, 152956, 152129, 152715, + 153131, 153490, 151734, 152271, 152707, 151734, 153321, 152450, + 151670, 198, 8088, 155792, 151669, 152452, 153497, 153353, + 152679, 152533, 152382, 152374, 152611, 153341, 153163, 152285, + 153411, 152495, 153141, 152320, 151670, 198, 1199, 155781, + 151669, 151764, 152360, 153295, 152634, 153342, 152199, 152271, + 151670, 198, 43366, 155799, 151669, 152308, 151682, 152889, + 152016, 152385, 152629, 152495, 151826, 153321, 152958, 152180, + 151886, 153432, 152922, 152128, 153024, 153040, 152593, 152287, + 151677, 151670, 198, 53660, 155808, 151669, 151727, 152092, + 152680, 153331, 151699, 152316, 152938, 152289, 152433, 153384, + 151781, 153137, 153259, 152175, 153213, 152291, 151869, 152691, + 152489, 151941, 152049, 152034, 153053, 152179, 153160, 151676, + 153367, 151670, 198, 268, 4123, 480, 155821, 151669, 152350, + 152173, 152536, 151991, 151960, 153144, 153013, 152358, 152234, + 153135, 152291, 153235, 152143, 152583, 152402, 153483, 152678, + 152192, 152533, 152946, 151797, 153103, 152310, 152293, 151825, + 152548, 153442, 152109, 152659, 153325, 152781, 152570, 152957, + 151752, 152265, 153381, 152515, 151670, 198, 437, 155787, + 151669, 152957, 152659, 151975, 152709, 152402, 152836, 152174, + 151792, 153409, 153327, 152990, 151670, 198, 275, 155781, + 151669, 152520, 153038, 152067, 153273, 153185, 152265, 152974, + 151670, 198, 94273, 155799, 151669, 152953, 152938, 153427, + 152244, 151920, 153423, 152929, 152367, 153052, 152129, 152331, + 152257, 152987, 152777, 153448, 152408, 151696, 152408, 152326, + 152699, 151670, 198, 385, 16239, 155828, 151669, 152306, 152268, + 153438, 153228, 152978, 152957, 153153, 153393, 152795, 152110, + 152918, 152923, 152467, 152331, 153053, 153330, 151889, 153444, + 152234, 152624, 151779, 152801, 152784, 152139, 152222, 152751, + 152512, 153287, 153141, 153052, 151840, 152589, 152508, 153499, + 152109, 152255, 151739, 152267, 152759, 153318, 153165, 153349, + 151670,}); +#endif + + // print the prompt token-by-token + + LOG("\n"); + + for (auto id : prompt_inp) { + LOG("%s", common_token_to_piece(ctx_ttc, id).c_str()); + } + + LOG_INF("%s: prompt size: %d\n", __func__, (int) prompt_inp.size()); + + LOG("\n"); + + // create a llama_batch + // we use this object to submit token data for decoding + llama_batch batch = llama_batch_init(std::max(prompt_inp.size(), (size_t) n_parallel), 0, n_parallel); + + std::vector seq_ids(n_parallel, 0); + for (int32_t i = 0; i < n_parallel; ++i) { + seq_ids[i] = i; + } + + // evaluate the initial prompt + for (size_t i = 0; i < prompt_inp.size(); ++i) { + common_batch_add(batch, prompt_inp[i], i, seq_ids, false); + } + GGML_ASSERT(batch.n_tokens == (int) prompt_inp.size()); + + // llama_decode will output logits only for the last token of the prompt + batch.logits[batch.n_tokens - 1] = true; + + if (llama_decode(ctx_ttc, batch) != 0) { + LOG_ERR("%s: llama_decode() failed\n", __func__); + return 1; + } + + if (n_parallel > 1) { + LOG_INF("\n\n%s: generating %d sequences ...\n", __func__, n_parallel); + } + + llama_synchronize(ctx_ttc); + + LOG_INF("%s: time for prompt: %.3f ms\n\n", __func__, (ggml_time_us() - t_main_start) / 1000.0f); + + const auto t_dec_start = ggml_time_us(); + + // main loop + + // remember the batch index of the last token for each parallel sequence + // we need this to determine which logits to sample from + std::vector i_batch(n_parallel, batch.n_tokens - 1); + + int n_past = batch.n_tokens; + int n_decode = 0; + + while (n_decode <= n_predict) { + // prepare the next batch + common_batch_clear(batch); + + // sample the next token for each parallel sequence / stream + for (int32_t i = 0; i < n_parallel; ++i) { + if (i_batch[i] < 0) { + // the stream has already finished + continue; + } + + const llama_token new_token_id = common_sampler_sample(smpl[i], ctx_ttc, i_batch[i]); + + common_sampler_accept(smpl[i], new_token_id, true); + + codes.push_back(new_token_id); + + const auto * cands = common_sampler_get_candidates(smpl[i]); + + // is it an end of generation? -> mark the stream as finished + if (llama_vocab_is_eog(vocab, new_token_id) || n_decode == n_predict) { + std::string reason; + if (llama_vocab_is_eog(vocab, new_token_id)) { + reason = "eos"; + } else { + reason = "n_predict"; + } + + i_batch[i] = -1; + + LOG("\n"); + if (n_parallel > 1) { + LOG_CNT("\n"); + LOG_INF("%s: stream %d finished at n_past = %d, reason = '%s'\n", __func__, i, n_past, reason.c_str()); + } + + continue; + } + + { + const float p = cands->data[cands->selected].p; + + const int col = std::max(0, std::min((int) k_colors.size() - 1, (int) ((3*p)*float(k_colors.size())))); + + LOG_CNT("%s%d%s", k_colors[col].c_str(), i, "\033[0m"); + //LOG_CNT("%d", i); + } + + i_batch[i] = batch.n_tokens; + + // push this new token for next evaluation + common_batch_add(batch, new_token_id, n_past, { i }, true); + } + + // all streams are finished + if (batch.n_tokens == 0) { + break; + } + + n_decode += 1; + n_past += 1; + + // evaluate the current batch with the transformer model + if (llama_decode(ctx_ttc, batch)) { + LOG_ERR("%s : failed to eval, return code %d\n", __func__, 1); + return 1; + } + } + + llama_batch_free(batch); + + LOG("\n"); + LOG_INF("%s: time for decoder: %.3f ms\n", __func__, (ggml_time_us() - t_dec_start) / 1000.0f); + } + + common_perf_print(ctx_ttc, smpl[0]); + + //std::vector codes = {198, 88225, 155856, 151669, 152205, + // 153064, 152537, 153421, 153209, 152524, 151689, 152993, 152438, 152695, + // 153091, 152945, 152829, 152534, 152934, 153020, 151997, 152263, 153010, + // 153146, 152399, 153208, 152496, 151793, 152848, 152263, 152571, 153286, + // 152227, 153300, 152934, 152263, 153208, 152263, 152965, 152430, 152296, + // 153146, 152920, 152376, 152556, 153363, 151775, 152044, 152972, 152690, + // 153379, 152368, 152233, 153422, 152490, 151996, 152022, 151694, 152061, + // 153238, 152539, 153356, 152640, 153021, 153123, 151962, 153094, 151670, + // 198, 20339, 13189, 155824, 151669, 152070, 152007, 152910, 151683, + // 152000, 152373, 152760, 152046, 151735, 152334, 152394, 153073, 152908, + // 151856, 151953, 153247, 153293, 151903, 153480, 153168, 152478, 153359, + // 153429, 151905, 151678, 152567, 152411, 152165, 152556, 153075, 153424, + // 151993, 152999, 153078, 152151, 152088, 153389, 152484, 151874, 151670, + // 198, 285, 155784, 151669, 152226, 152126, 152638, 153215, 151729, + // 152959, 153479, 153059, 151838, 151670, 198, 1782, 155783, 151669, + // 153288, 153055, 153314, 152497, 152962, 152741, 152076, 153253, 151670, + // 198, 471, 16488, 155825, 151669, 152060, 152916, 151893, 153469, 152501, + // 152080, 152743, 151932, 153161, 152096, 152761, 152698, 153401, 153242, + // 153336, 152441, 152838, 153467, 152706, 153496, 153310, 152422, 153360, + // 153115, 152763, 151998, 152373, 153450, 152554, 151968, 153323, 152055, + // 152468, 153111, 153358, 152813, 152010, 151770, 152823, 152960, 151670, + // 198, 22627, 155823, 151669, 152814, 152366, 153484, 152931, 153441, + // 152164, 152877, 152915, 153463, 151692, 152911, 152747, 152776, 151831, + // 153449, 151882, 152975, 152031, 152513, 153150, 152448, 152667, 153133, + // 153189, 152619, 153466, 152054, 152106, 153119, 152277, 152439, 153109, + // 152997, 152141, 153154, 153256, 153311, 151922, 151670, 198, 1055, + // 155781, 151669, 152633, 151850, 153060, 153270, 152560, 153348, 152729, + // 151670, 198, 25312, 155803, 151669, 152521, 153403, 152561, 153337, + // 153383, 152199, 153493, 153326, 151830, 152254, 152248, 152349, 152153, + // 153007, 151823, 153037, 152575, 152457, 152406, 152592, 153116, 153365, + // 153456, 151670, 198, 88225, 155817, 151669, 153271, 151925, 152218, + // 152418, 152253, 153140, 151903, 153151, 152626, 152338, 152647, 153464, + // 152785, 152768, 151711, 152037, 152033, 151804, 152216, 151701, 151855, + // 152348, 152995, 152955, 152905, 152342, 152340, 153391, 153453, 152418, + // 153415, 151990, 153083, 152884, 151670, 198, 151668, 198, 151645}; + + { + const std::string inp_txt = common_detokenize(ctx_ttc, codes, true); + + LOG("\n"); + LOG_INF("codes: '%s'\n", inp_txt.c_str()); + LOG_INF("%s: codes size: %d\n", __func__, (int) codes.size()); + } + + // remove all non-audio tokens (i.e. < 151672 || > 155772) + codes.erase(std::remove_if(codes.begin(), codes.end(), [](llama_token t) { return t < 151672 || t > 155772; }), codes.end()); + + { + const std::string inp_txt = common_detokenize(ctx_ttc, codes, true); + LOG_INF("codes audio: '%s'\n", inp_txt.c_str()); + LOG_INF("%s: codes audio size: %d\n", __func__, (int) codes.size()); + } + + for (auto & token : codes) { + token -= 151672; + } + + const auto t_voc_start = ggml_time_us(); + + const int n_codes = codes.size(); + + llama_batch batch = llama_batch_init(n_codes, 0, 1); + + for (size_t i = 0; i < codes.size(); ++i) { + common_batch_add(batch, codes[i], i, { 0 }, true); // TODO: all logits? + } + GGML_ASSERT(batch.n_tokens == n_codes); + + if (llama_decode(ctx_cts, batch) != 0) { + LOG_ERR("%s: llama_decode() failed\n", __func__); + return 1; + } + + llama_synchronize(ctx_cts); + + LOG_INF("%s: time for vocoder: %.3f ms\n", __func__, (ggml_time_us() - t_voc_start) / 1000.0f); + + const auto t_spec_start = ggml_time_us(); + +#if 1 + // spectral operations + const int n_embd = llama_model_n_embd(model_cts); + const float * embd = llama_get_embeddings(ctx_cts); + + auto audio = embd_to_audio(embd, n_codes, n_embd, params.cpuparams.n_threads); + +#else + // read the spectrogram from a file for debugging purposes + std::vector audio; + { + std::ifstream fin("out.bin", std::ios::binary); + if (!fin) { + LOG_ERR("%s: failed to open file '%s'\n", __func__, "out.bin"); + return 1; + } + + std::vector embd; + + int n_codes; + int n_embd; + + fin.read(reinterpret_cast(&n_codes), sizeof(int)); + fin.read(reinterpret_cast(&n_embd), sizeof(int)); + + embd.resize(n_codes * n_embd); + fin.read(reinterpret_cast(embd.data()), n_codes * n_embd * sizeof(float)); + fin.close(); + + LOG_INF("%s: n_codes: %d, n_embd: %d\n", __func__, n_codes, n_embd); + + audio = embd_to_audio(embd.data(), n_codes, n_embd, params.cpuparams.n_threads); + } +#endif + + const std::string fname = "output.wav"; + + const int n_sr = 24000; // sampling rate + + // zero out first 0.25 seconds + for (int i = 0; i < 24000/4; ++i) { + audio[i] = 0.0f; + } + + LOG_INF("%s: time for spectral ops: %.3f ms\n", __func__, (ggml_time_us() - t_spec_start) / 1000.0f); + LOG_INF("%s: total time: %.3f ms\n", __func__, (ggml_time_us() - t_main_start) / 1000.0f); + + save_wav16(fname, audio, n_sr); + + LOG_INF("%s: audio written to file '%s'\n", __func__, fname.c_str()); + + llama_backend_free(); + + return 0; +} diff --git a/flake.lock b/flake.lock index c170c4952..d114f4422 100644 --- a/flake.lock +++ b/flake.lock @@ -20,11 +20,11 @@ }, "nixpkgs": { "locked": { - "lastModified": 1730200266, - "narHash": "sha256-l253w0XMT8nWHGXuXqyiIC/bMvh1VRszGXgdpQlfhvU=", + "lastModified": 1732014248, + "narHash": "sha256-y/MEyuJ5oBWrWAic/14LaIr/u5E0wRVzyYsouYY3W6w=", "owner": "NixOS", "repo": "nixpkgs", - "rev": "807e9154dcb16384b1b765ebe9cd2bba2ac287fd", + "rev": "23e89b7da85c3640bbc2173fe04f4bd114342367", "type": "github" }, "original": { diff --git a/ggml/CMakeLists.txt b/ggml/CMakeLists.txt index 6866a25d3..185079aa4 100644 --- a/ggml/CMakeLists.txt +++ b/ggml/CMakeLists.txt @@ -32,7 +32,15 @@ else() endif() endif() +# remove the lib prefix on win32 mingw +if (WIN32) + set(CMAKE_STATIC_LIBRARY_PREFIX "") + set(CMAKE_SHARED_LIBRARY_PREFIX "") + set(CMAKE_SHARED_MODULE_PREFIX "") +endif() + option(BUILD_SHARED_LIBS "ggml: build shared libraries" ${BUILD_SHARED_LIBS_DEFAULT}) +option(GGML_BACKEND_DL "ggml: build backends as dynamic libraries (requires BUILD_SHARED_LIBS)" OFF) # # option list @@ -66,10 +74,10 @@ if (NOT GGML_CUDA_GRAPHS_DEFAULT) endif() # general -option(GGML_STATIC "ggml: static link libraries" OFF) -option(GGML_NATIVE "ggml: enable -march=native flag" ${GGML_NATIVE_DEFAULT}) -option(GGML_LTO "ggml: enable link time optimization" OFF) -option(GGML_CCACHE "ggml: use ccache if available" ON) +option(GGML_STATIC "ggml: static link libraries" OFF) +option(GGML_NATIVE "ggml: optimize the build for the current system" ${GGML_NATIVE_DEFAULT}) +option(GGML_LTO "ggml: enable link time optimization" OFF) +option(GGML_CCACHE "ggml: use ccache if available" ON) # debug option(GGML_ALL_WARNINGS "ggml: enable all compiler warnings" ON) @@ -91,31 +99,39 @@ else() set(INS_ENB ON) endif() -option(GGML_CPU_HBM "ggml: use memkind for CPU HBM" OFF) - -option(GGML_AVX "ggml: enable AVX" ${INS_ENB}) -option(GGML_AVX2 "ggml: enable AVX2" ${INS_ENB}) -option(GGML_AVX512 "ggml: enable AVX512" OFF) -option(GGML_AVX512_VBMI "ggml: enable AVX512-VBMI" OFF) -option(GGML_AVX512_VNNI "ggml: enable AVX512-VNNI" OFF) -option(GGML_AVX512_BF16 "ggml: enable AVX512-BF16" OFF) -option(GGML_AMX_TILE "ggml: enable AMX-TILE" OFF) -option(GGML_AMX_INT8 "ggml: enable AMX-INT8" OFF) -option(GGML_AMX_BF16 "ggml: enable AMX-BF16" OFF) -option(GGML_FMA "ggml: enable FMA" ${INS_ENB}) +option(GGML_CPU_HBM "ggml: use memkind for CPU HBM" OFF) +option(GGML_CPU_AARCH64 "ggml: use runtime weight conversion of Q4_0 to Q4_X_X" ON) +option(GGML_AVX "ggml: enable AVX" ${INS_ENB}) +option(GGML_AVX_VNNI "ggml: enable AVX-VNNI" OFF) +option(GGML_AVX2 "ggml: enable AVX2" ${INS_ENB}) +option(GGML_AVX512 "ggml: enable AVX512F" OFF) +option(GGML_AVX512_VBMI "ggml: enable AVX512-VBMI" OFF) +option(GGML_AVX512_VNNI "ggml: enable AVX512-VNNI" OFF) +option(GGML_AVX512_BF16 "ggml: enable AVX512-BF16" OFF) if (NOT MSVC) - option(GGML_F16C "ggml: enable F16C" ${INS_ENB}) # in MSVC F16C is implied with AVX2/AVX512 + # in MSVC F16C and FMA is implied with AVX2/AVX512 + option(GGML_FMA "ggml: enable FMA" ${INS_ENB}) + option(GGML_F16C "ggml: enable F16C" ${INS_ENB}) + # MSVC does not seem to support AMX + option(GGML_AMX_TILE "ggml: enable AMX-TILE" OFF) + option(GGML_AMX_INT8 "ggml: enable AMX-INT8" OFF) + option(GGML_AMX_BF16 "ggml: enable AMX-BF16" OFF) endif() -option(GGML_LASX "ggml: enable lasx" ON) -option(GGML_LSX "ggml: enable lsx" ON) -option(GGML_SVE "ggml: enable SVE" OFF) +option(GGML_LASX "ggml: enable lasx" ON) +option(GGML_LSX "ggml: enable lsx" ON) +option(GGML_RVV "ggml: enable rvv" ON) + +option(GGML_CPU_ALL_VARIANTS "ggml: build all variants of the CPU backend (requires GGML_BACKEND_DL)" OFF) +set(GGML_CPU_ARM_ARCH "" CACHE STRING "ggml: CPU architecture for ARM") + if (WIN32) - set(GGML_WIN_VER "0x602" CACHE STRING "ggml: Windows Version") + set(GGML_WIN_VER "0x602" CACHE STRING "ggml: Windows version") endif() # ggml core set(GGML_SCHED_MAX_COPIES "4" CACHE STRING "ggml: max input copies for pipeline parallelism") +option(GGML_CPU "ggml: enable CPU backend" ON) # 3rd party libs / backends option(GGML_ACCELERATE "ggml: enable Accelerate framework" ON) @@ -126,14 +142,9 @@ option(GGML_LLAMAFILE "ggml: use LLAMAFILE" option(GGML_CUDA "ggml: use CUDA" OFF) option(GGML_MUSA "ggml: use MUSA" OFF) -option(GGML_CUDA_FORCE_DMMV "ggml: use dmmv instead of mmvq CUDA kernels" OFF) option(GGML_CUDA_FORCE_MMQ "ggml: use mmq kernels instead of cuBLAS" OFF) option(GGML_CUDA_FORCE_CUBLAS "ggml: always use cuBLAS instead of mmq kernels" OFF) -set (GGML_CUDA_DMMV_X "32" CACHE STRING "ggml: x stride for dmmv CUDA kernels") -set (GGML_CUDA_MMV_Y "1" CACHE STRING "ggml: y block size for mmv CUDA kernels") option(GGML_CUDA_F16 "ggml: use 16 bit floats for some calculations" OFF) -set (GGML_CUDA_KQUANTS_ITER "2" CACHE STRING - "ggml: iters./thread per block for Q2_K/Q6_K") set (GGML_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING "ggml: max. batch size for using peer access") option(GGML_CUDA_NO_PEER_COPY "ggml: do not use peer to peer copies" OFF) @@ -141,7 +152,7 @@ option(GGML_CUDA_NO_VMM "ggml: do not try to use CUDA VMM" option(GGML_CUDA_FA_ALL_QUANTS "ggml: compile all quants for FlashAttention" OFF) option(GGML_CUDA_GRAPHS "ggml: use CUDA graphs (llama.cpp only)" ${GGML_CUDA_GRAPHS_DEFAULT}) -option(GGML_HIPBLAS "ggml: use hipBLAS" OFF) +option(GGML_HIP "ggml: use HIP" OFF) option(GGML_HIP_UMA "ggml: use HIP unified memory architecture" OFF) option(GGML_VULKAN "ggml: use Vulkan" OFF) option(GGML_VULKAN_CHECK_RESULTS "ggml: run Vulkan op checks" OFF) @@ -153,6 +164,7 @@ option(GGML_VULKAN_VALIDATE "ggml: enable Vulkan validation" option(GGML_VULKAN_RUN_TESTS "ggml: run Vulkan tests" OFF) option(GGML_KOMPUTE "ggml: use Kompute" OFF) option(GGML_METAL "ggml: use Metal" ${GGML_METAL_DEFAULT}) +option(GGML_METAL_USE_BF16 "ggml: use bfloat if available" OFF) option(GGML_METAL_NDEBUG "ggml: disable Metal debugging" OFF) option(GGML_METAL_SHADER_DEBUG "ggml: compile Metal with -fno-fast-math" OFF) option(GGML_METAL_EMBED_LIBRARY "ggml: embed Metal library" ${GGML_METAL}) @@ -161,11 +173,20 @@ set (GGML_METAL_MACOSX_VERSION_MIN "" CACHE STRING set (GGML_METAL_STD "" CACHE STRING "ggml: metal standard version (-std flag)") option(GGML_OPENMP "ggml: use OpenMP" ON) option(GGML_RPC "ggml: use RPC" OFF) -option(GGML_AMX "ggml: use AMX" OFF) option(GGML_SYCL "ggml: use SYCL" OFF) option(GGML_SYCL_F16 "ggml: use 16 bit floats for sycl calculations" OFF) set (GGML_SYCL_TARGET "INTEL" CACHE STRING "ggml: sycl target device") +set (GGML_SYCL_DEVICE_ARCH "" CACHE STRING + "ggml: sycl device architecture") + +option(GGML_OPENCL "ggml: use OpenCL" OFF) +option(GGML_OPENCL_PROFILING "ggml: use OpenCL profiling (increases overhead)" OFF) +option(GGML_OPENCL_EMBED_KERNELS "ggml: embed kernels" ON) +option(GGML_OPENCL_USE_ADRENO_KERNELS "ggml: use optimized kernels for Adreno" ON) + +# toolchain for vulkan-shaders-gen +set (GGML_VULKAN_SHADERS_GEN_TOOLCHAIN "" CACHE FILEPATH "ggml: toolchain file for vulkan-shaders-gen") # extra artifacts option(GGML_BUILD_TESTS "ggml: build tests" ${GGML_STANDALONE}) @@ -178,11 +199,7 @@ option(GGML_BUILD_EXAMPLES "ggml: build examples" ${GGML_STANDALONE}) set(CMAKE_C_STANDARD 11) set(CMAKE_C_STANDARD_REQUIRED true) -if (GGML_SYCL) - set(CMAKE_CXX_STANDARD 17) -else() - set(CMAKE_CXX_STANDARD 11) -endif() +set(CMAKE_CXX_STANDARD 17) set(CMAKE_CXX_STANDARD_REQUIRED true) set(THREADS_PREFER_PTHREAD_FLAG ON) @@ -225,38 +242,19 @@ set(GGML_PUBLIC_HEADERS include/ggml-cann.h include/ggml-cuda.h include/ggml-kompute.h + include/ggml-opt.h include/ggml-metal.h include/ggml-rpc.h include/ggml-sycl.h - include/ggml-vulkan.h) + include/ggml-vulkan.h + include/gguf.h) set_target_properties(ggml PROPERTIES PUBLIC_HEADER "${GGML_PUBLIC_HEADERS}") #if (GGML_METAL) # set_target_properties(ggml PROPERTIES RESOURCE "${CMAKE_CURRENT_SOURCE_DIR}/src/ggml-metal.metal") #endif() -install(TARGETS ggml PUBLIC_HEADER) - -if (BUILD_SHARED_LIBS) - install(TARGETS ggml LIBRARY) -endif() - -if (GGML_METAL) - install( - FILES src/ggml-metal.metal - PERMISSIONS - OWNER_READ - OWNER_WRITE - GROUP_READ - WORLD_READ - DESTINATION ${CMAKE_INSTALL_BINDIR}) - - if (NOT GGML_METAL_EMBED_LIBRARY) - install( - FILES ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib - DESTINATION ${CMAKE_INSTALL_BINDIR} - ) - endif() -endif() +install(TARGETS ggml LIBRARY PUBLIC_HEADER) +install(TARGETS ggml-base LIBRARY) if (GGML_STANDALONE) configure_file(${CMAKE_CURRENT_SOURCE_DIR}/ggml.pc.in diff --git a/ggml/include/ggml-amx.h b/ggml/include/ggml-amx.h deleted file mode 100644 index 22b3f70f4..000000000 --- a/ggml/include/ggml-amx.h +++ /dev/null @@ -1,25 +0,0 @@ -#pragma once - -#include "ggml.h" -#include "ggml-backend.h" - - -#ifdef __cplusplus -extern "C" { -#endif - -// buffer_type API -GGML_API ggml_backend_buffer_type_t ggml_backend_amx_buffer_type(void); - -GGML_API bool ggml_backend_is_amx(ggml_backend_t backend); - -// backend API -GGML_API ggml_backend_t ggml_backend_amx_init(void); - -GGML_API void ggml_backend_amx_set_n_threads(ggml_backend_t backend_amx, int n_threads); - -GGML_API ggml_backend_reg_t ggml_backend_amx_reg(void); - -#ifdef __cplusplus -} -#endif diff --git a/ggml/include/ggml-backend.h b/ggml/include/ggml-backend.h index 125413d1b..7221a0830 100644 --- a/ggml/include/ggml-backend.h +++ b/ggml/include/ggml-backend.h @@ -3,6 +3,20 @@ #include "ggml.h" #include "ggml-alloc.h" +#ifdef GGML_BACKEND_SHARED +# if defined(_WIN32) && !defined(__MINGW32__) +# ifdef GGML_BACKEND_BUILD +# define GGML_BACKEND_API __declspec(dllexport) extern +# else +# define GGML_BACKEND_API __declspec(dllimport) extern +# endif +# else +# define GGML_BACKEND_API __attribute__ ((visibility ("default"))) extern +# endif +#else +# define GGML_BACKEND_API extern +#endif + #ifdef __cplusplus extern "C" { #endif @@ -72,7 +86,7 @@ extern "C" { GGML_API void ggml_backend_tensor_set_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); GGML_API void ggml_backend_tensor_get_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); - // "offset" refers to the offset of the tensor data for setting/getting data + // "offset" refers to the offset in tensor->data for setting/getting data GGML_API void ggml_backend_tensor_set( struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); GGML_API void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); GGML_API void ggml_backend_tensor_memset( struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size); @@ -176,6 +190,14 @@ extern "C" { typedef void (*ggml_backend_set_n_threads_t)(ggml_backend_t backend, int n_threads); // Get additional buffer types provided by the device (returns a NULL-terminated array) typedef ggml_backend_buffer_type_t * (*ggml_backend_dev_get_extra_bufts_t)(ggml_backend_dev_t device); + // Set the abort callback for the backend + typedef void (*ggml_backend_set_abort_callback_t)(ggml_backend_t backend, ggml_abort_callback abort_callback, void * abort_callback_data); + // Get a list of feature flags supported by the backend (returns a NULL-terminated array) + struct ggml_backend_feature { + const char * name; + const char * value; + }; + typedef struct ggml_backend_feature * (*ggml_backend_get_features_t)(ggml_backend_reg_t reg); // // Backend registry @@ -200,6 +222,14 @@ extern "C" { // = ggml_backend_dev_init(ggml_backend_dev_by_type(GPU) OR ggml_backend_dev_by_type(CPU), NULL) GGML_API ggml_backend_t ggml_backend_init_best(void); + // Load a backend from a dynamic library and register it + GGML_API ggml_backend_reg_t ggml_backend_load(const char * path); + // Unload a backend if loaded dynamically and unregister it + GGML_API void ggml_backend_unload(ggml_backend_reg_t reg); + // Load all known backends from dynamic libraries + GGML_API void ggml_backend_load_all(void); + GGML_API void ggml_backend_load_all_from_path(const char * dir_path); + // // Backend scheduler // @@ -228,14 +258,20 @@ extern "C" { ggml_backend_sched_reserve(sched, reserve_graph); // compute - graph = build_graph(sched); - ggml_backend_sched_graph_compute(sched, graph); + graph = build_graph(sched); // the graph and its tensors are single-use in terms of allocation, multi-use in terms of computation + for (int i = 0; i < 10; ++i) { + ggml_backend_sched_graph_compute(sched, graph); // on the first iteration the graph is allocated automatically + } // if there are graph inputs: - ggml_backend_sched_reset(sched); - ggml_backend_sched_alloc_graph(sched, graph); - ggml_backend_tensor_set(input_tensor, ...); - ggml_backend_sched_graph_compute(sched, graph); + graph = build_graph(sched); // get a new graph that is not allocated (the metadata for the old graph is freed once ggml_free is called) + ggml_backend_sched_reset(sched); // clear the allocation of the previous graph + ggml_backend_sched_alloc_graph(sched, graph); // explicitly allocate the new graph but do not execute it + ggml_backend_tensor_set(input_tensor, ...); // copy data to the newly allocated graph tensors + ggml_backend_sched_graph_compute(sched, graph); // execute the graph + + // as an alternative to the above it is also possible to assign the inputs to a dedicated context and + // allocate them statically via ggml_backend_alloc_ctx_tensors } */ @@ -250,7 +286,7 @@ extern "C" { // typedef bool (*ggml_backend_sched_eval_callback)(struct ggml_tensor * t, bool ask, void * user_data); - // Initialize a backend scheduler + // Initialize a backend scheduler, backends with low index are given priority over backends with high index GGML_API ggml_backend_sched_t ggml_backend_sched_new(ggml_backend_t * backends, ggml_backend_buffer_type_t * bufts, int n_backends, size_t graph_size, bool parallel); GGML_API void ggml_backend_sched_free(ggml_backend_sched_t sched); @@ -275,7 +311,9 @@ extern "C" { GGML_API enum ggml_status ggml_backend_sched_graph_compute_async(ggml_backend_sched_t sched, struct ggml_cgraph * graph); GGML_API void ggml_backend_sched_synchronize(ggml_backend_sched_t sched); - // Reset all assignments and allocators - must be called before changing the node backends + // Reset all assignments and allocators - must be called before changing the node backends or allocating a new graph. + // This in effect deallocates all tensors that were previously allocated and leaves them with dangling pointers. + // The correct way to use this API is to discard the deallocated tensors and create new ones. GGML_API void ggml_backend_sched_reset(ggml_backend_sched_t sched); // Set a callback to be called for each resulting node during graph compute diff --git a/ggml/include/ggml-blas.h b/ggml/include/ggml-blas.h index 25b2e637f..87a81b363 100644 --- a/ggml/include/ggml-blas.h +++ b/ggml/include/ggml-blas.h @@ -9,15 +9,15 @@ extern "C" { #endif // backend API -GGML_API ggml_backend_t ggml_backend_blas_init(void); +GGML_BACKEND_API ggml_backend_t ggml_backend_blas_init(void); -GGML_API bool ggml_backend_is_blas(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_blas(ggml_backend_t backend); // number of threads used for conversion to float // for openblas and blis, this will also set the number of threads used for blas operations -GGML_API void ggml_backend_blas_set_n_threads(ggml_backend_t backend_blas, int n_threads); +GGML_BACKEND_API void ggml_backend_blas_set_n_threads(ggml_backend_t backend_blas, int n_threads); -GGML_API ggml_backend_reg_t ggml_backend_blas_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_blas_reg(void); #ifdef __cplusplus diff --git a/ggml/include/ggml-cann.h b/ggml/include/ggml-cann.h index 528975493..b469e228d 100644 --- a/ggml/include/ggml-cann.h +++ b/ggml/include/ggml-cann.h @@ -34,7 +34,7 @@ extern "C" { */ #define GGML_CANN_MAX_DEVICES 16 -GGML_API ggml_backend_reg_t ggml_backend_cann_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_cann_reg(void); /** * @brief Initializes the CANN backend for a specified device. @@ -46,7 +46,7 @@ GGML_API ggml_backend_reg_t ggml_backend_cann_reg(void); * @param device The index of the device to initialize. * @return A pointer to the initialized backend instance, or nullptr on failure. */ -GGML_API ggml_backend_t ggml_backend_cann_init(int32_t device); +GGML_BACKEND_API ggml_backend_t ggml_backend_cann_init(int32_t device); /** * @brief Checks if a given backend is a CANN backend. @@ -57,7 +57,7 @@ GGML_API ggml_backend_t ggml_backend_cann_init(int32_t device); * @param backend The backend instance to check. * @return True if the backend is a CANN backend, false otherwise. */ -GGML_API bool ggml_backend_is_cann(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_cann(ggml_backend_t backend); /** * @brief Retrieves the CANN buffer type for a specified device. @@ -69,7 +69,7 @@ GGML_API bool ggml_backend_is_cann(ggml_backend_t backend); * @return A pointer to the buffer type interface for the specified device, or * nullptr if the device index is out of range. */ -GGML_API ggml_backend_buffer_type_t +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_cann_buffer_type(int32_t device); /** @@ -80,14 +80,14 @@ ggml_backend_cann_buffer_type(int32_t device); * * @return The number of CANN devices available. */ -GGML_API int32_t ggml_backend_cann_get_device_count(void); +GGML_BACKEND_API int32_t ggml_backend_cann_get_device_count(void); /** * @brief pinned host buffer for use with the CPU backend for faster copies between CPU and NPU. * * @return A pointer to the host buffer type interface. */ -GGML_API ggml_backend_buffer_type_t ggml_backend_cann_host_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_cann_host_buffer_type(void); /** * @brief Retrieves the description of a specific CANN device. @@ -99,7 +99,7 @@ GGML_API ggml_backend_buffer_type_t ggml_backend_cann_host_buffer_type(void); * @param description Pointer to a buffer where the description will be written. * @param description_size Size of the description buffer. */ -GGML_API void ggml_backend_cann_get_device_description( +GGML_BACKEND_API void ggml_backend_cann_get_device_description( int32_t device, char* description, size_t description_size); /** @@ -114,7 +114,7 @@ GGML_API void ggml_backend_cann_get_device_description( * @param total Pointer to a variable where the total memory size will be * stored. */ -GGML_API void ggml_backend_cann_get_device_memory(int32_t device, +GGML_BACKEND_API void ggml_backend_cann_get_device_memory(int32_t device, size_t* free, size_t* total); diff --git a/ggml/include/ggml-cpp.h b/ggml/include/ggml-cpp.h index 219361af4..a12342c25 100644 --- a/ggml/include/ggml-cpp.h +++ b/ggml/include/ggml-cpp.h @@ -7,6 +7,7 @@ #include "ggml.h" #include "ggml-alloc.h" #include "ggml-backend.h" +#include "gguf.h" #include // Smart pointers for ggml types diff --git a/ggml/include/ggml-cpu.h b/ggml/include/ggml-cpu.h index 7f1ee7573..3aa71badb 100644 --- a/ggml/include/ggml-cpu.h +++ b/ggml/include/ggml-cpu.h @@ -7,29 +7,6 @@ extern "C" { #endif - // Scheduling priorities - enum ggml_sched_priority { - GGML_SCHED_PRIO_NORMAL, - GGML_SCHED_PRIO_MEDIUM, - GGML_SCHED_PRIO_HIGH, - GGML_SCHED_PRIO_REALTIME - }; - - // Threadpool params - // Use ggml_threadpool_params_default() or ggml_threadpool_params_init() to populate the defaults - struct ggml_threadpool_params { - bool cpumask[GGML_MAX_N_THREADS]; // mask of cpu cores (all-zeros means use default affinity settings) - int n_threads; // number of threads - enum ggml_sched_priority prio; // thread priority - uint32_t poll; // polling level (0 - no polling, 100 - aggressive polling) - bool strict_cpu; // strict cpu placement - bool paused; // start in paused state - }; - - struct ggml_threadpool; // forward declaration, see ggml.c - - typedef struct ggml_threadpool * ggml_threadpool_t; - // the compute plan that needs to be prepared for ggml_graph_compute() // since https://github.com/ggerganov/ggml/issues/287 struct ggml_cplan { @@ -54,96 +31,104 @@ extern "C" { GGML_NUMA_STRATEGY_COUNT }; - GGML_API void ggml_numa_init(enum ggml_numa_strategy numa); // call once for better performance on NUMA systems - GGML_API bool ggml_is_numa(void); // true if init detected that system has >1 NUMA node + GGML_BACKEND_API void ggml_numa_init(enum ggml_numa_strategy numa); // call once for better performance on NUMA systems + GGML_BACKEND_API bool ggml_is_numa(void); // true if init detected that system has >1 NUMA node - GGML_API struct ggml_tensor * ggml_new_i32(struct ggml_context * ctx, int32_t value); - GGML_API struct ggml_tensor * ggml_new_f32(struct ggml_context * ctx, float value); + GGML_BACKEND_API struct ggml_tensor * ggml_new_i32(struct ggml_context * ctx, int32_t value); + GGML_BACKEND_API struct ggml_tensor * ggml_new_f32(struct ggml_context * ctx, float value); - GGML_API struct ggml_tensor * ggml_set_i32 (struct ggml_tensor * tensor, int32_t value); - GGML_API struct ggml_tensor * ggml_set_f32 (struct ggml_tensor * tensor, float value); + GGML_BACKEND_API struct ggml_tensor * ggml_set_i32 (struct ggml_tensor * tensor, int32_t value); + GGML_BACKEND_API struct ggml_tensor * ggml_set_f32 (struct ggml_tensor * tensor, float value); - GGML_API int32_t ggml_get_i32_1d(const struct ggml_tensor * tensor, int i); - GGML_API void ggml_set_i32_1d(const struct ggml_tensor * tensor, int i, int32_t value); + GGML_BACKEND_API int32_t ggml_get_i32_1d(const struct ggml_tensor * tensor, int i); + GGML_BACKEND_API void ggml_set_i32_1d(const struct ggml_tensor * tensor, int i, int32_t value); - GGML_API int32_t ggml_get_i32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3); - GGML_API void ggml_set_i32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3, int32_t value); + GGML_BACKEND_API int32_t ggml_get_i32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3); + GGML_BACKEND_API void ggml_set_i32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3, int32_t value); - GGML_API float ggml_get_f32_1d(const struct ggml_tensor * tensor, int i); - GGML_API void ggml_set_f32_1d(const struct ggml_tensor * tensor, int i, float value); + GGML_BACKEND_API float ggml_get_f32_1d(const struct ggml_tensor * tensor, int i); + GGML_BACKEND_API void ggml_set_f32_1d(const struct ggml_tensor * tensor, int i, float value); - GGML_API float ggml_get_f32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3); - GGML_API void ggml_set_f32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3, float value); + GGML_BACKEND_API float ggml_get_f32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3); + GGML_BACKEND_API void ggml_set_f32_nd(const struct ggml_tensor * tensor, int i0, int i1, int i2, int i3, float value); - GGML_API struct ggml_threadpool_params ggml_threadpool_params_default(int n_threads); - GGML_API void ggml_threadpool_params_init (struct ggml_threadpool_params * p, int n_threads); - GGML_API bool ggml_threadpool_params_match (const struct ggml_threadpool_params * p0, const struct ggml_threadpool_params * p1); - GGML_API struct ggml_threadpool * ggml_threadpool_new (struct ggml_threadpool_params * params); - GGML_API void ggml_threadpool_free (struct ggml_threadpool * threadpool); - GGML_API int ggml_threadpool_get_n_threads(struct ggml_threadpool * threadpool); - GGML_API void ggml_threadpool_pause (struct ggml_threadpool * threadpool); - GGML_API void ggml_threadpool_resume (struct ggml_threadpool * threadpool); + GGML_BACKEND_API struct ggml_threadpool * ggml_threadpool_new (struct ggml_threadpool_params * params); + GGML_BACKEND_API void ggml_threadpool_free (struct ggml_threadpool * threadpool); + GGML_BACKEND_API int ggml_threadpool_get_n_threads (struct ggml_threadpool * threadpool); + GGML_BACKEND_API void ggml_threadpool_pause (struct ggml_threadpool * threadpool); + GGML_BACKEND_API void ggml_threadpool_resume (struct ggml_threadpool * threadpool); // ggml_graph_plan() has to be called before ggml_graph_compute() // when plan.work_size > 0, caller must allocate memory for plan.work_data - GGML_API struct ggml_cplan ggml_graph_plan( + GGML_BACKEND_API struct ggml_cplan ggml_graph_plan( const struct ggml_cgraph * cgraph, int n_threads, /* = GGML_DEFAULT_N_THREADS */ struct ggml_threadpool * threadpool /* = NULL */ ); - GGML_API enum ggml_status ggml_graph_compute(struct ggml_cgraph * cgraph, struct ggml_cplan * cplan); + GGML_BACKEND_API enum ggml_status ggml_graph_compute(struct ggml_cgraph * cgraph, struct ggml_cplan * cplan); // same as ggml_graph_compute() but the work data is allocated as a part of the context // note: the drawback of this API is that you must have ensured that the context has enough memory for the work data - GGML_API enum ggml_status ggml_graph_compute_with_ctx(struct ggml_context * ctx, struct ggml_cgraph * cgraph, int n_threads); + GGML_BACKEND_API enum ggml_status ggml_graph_compute_with_ctx(struct ggml_context * ctx, struct ggml_cgraph * cgraph, int n_threads); - // TODO: move to backend interface - GGML_API int ggml_cpu_has_neon (void); - GGML_API int ggml_cpu_has_sve (void); - GGML_API int ggml_cpu_has_matmul_int8(void); - // get the sve vector length in bytes - GGML_API int ggml_cpu_get_sve_cnt(void); + // + // system info + // + + // x86 + GGML_BACKEND_API int ggml_cpu_has_sse3 (void); + GGML_BACKEND_API int ggml_cpu_has_ssse3 (void); + GGML_BACKEND_API int ggml_cpu_has_avx (void); + GGML_BACKEND_API int ggml_cpu_has_avx_vnni (void); + GGML_BACKEND_API int ggml_cpu_has_avx2 (void); + GGML_BACKEND_API int ggml_cpu_has_f16c (void); + GGML_BACKEND_API int ggml_cpu_has_fma (void); + GGML_BACKEND_API int ggml_cpu_has_avx512 (void); + GGML_BACKEND_API int ggml_cpu_has_avx512_vbmi(void); + GGML_BACKEND_API int ggml_cpu_has_avx512_vnni(void); + GGML_BACKEND_API int ggml_cpu_has_avx512_bf16(void); + GGML_BACKEND_API int ggml_cpu_has_amx_int8 (void); + // ARM + GGML_BACKEND_API int ggml_cpu_has_neon (void); + GGML_BACKEND_API int ggml_cpu_has_arm_fma (void); + GGML_BACKEND_API int ggml_cpu_has_fp16_va (void); + GGML_BACKEND_API int ggml_cpu_has_dotprod (void); + GGML_BACKEND_API int ggml_cpu_has_matmul_int8(void); + GGML_BACKEND_API int ggml_cpu_has_sve (void); + GGML_BACKEND_API int ggml_cpu_get_sve_cnt (void); // sve vector length in bytes + // other + GGML_BACKEND_API int ggml_cpu_has_riscv_v (void); + GGML_BACKEND_API int ggml_cpu_has_vsx (void); + GGML_BACKEND_API int ggml_cpu_has_wasm_simd (void); + GGML_BACKEND_API int ggml_cpu_has_llamafile (void); // Internal types and functions exposed for tests and benchmarks - typedef void (*ggml_from_float_to_mat_t) - (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t nr, int64_t k, int64_t bs); typedef void (*ggml_vec_dot_t) (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT x, size_t bx, const void * GGML_RESTRICT y, size_t by, int nrc); - typedef void (*ggml_gemv_t) (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT x, - const void * GGML_RESTRICT y, int nr, int nc); - typedef void (*ggml_gemm_t) (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT x, - const void * GGML_RESTRICT y, int nr, int nc); struct ggml_type_traits_cpu { - ggml_from_float_to_mat_t from_float_to_mat; + ggml_from_float_t from_float; ggml_vec_dot_t vec_dot; enum ggml_type vec_dot_type; int64_t nrows; // number of rows to process simultaneously - int64_t ncols; // number of columns to process simultaneously - ggml_gemv_t gemv; - ggml_gemm_t gemm; }; - GGML_API const struct ggml_type_traits_cpu * ggml_get_type_traits_cpu(enum ggml_type type); + GGML_BACKEND_API const struct ggml_type_traits_cpu * ggml_get_type_traits_cpu(enum ggml_type type); - GGML_API void ggml_cpu_init(void); + GGML_BACKEND_API void ggml_cpu_init(void); // // CPU backend // - GGML_API ggml_backend_t ggml_backend_cpu_init(void); + GGML_BACKEND_API ggml_backend_t ggml_backend_cpu_init(void); - GGML_API bool ggml_backend_is_cpu (ggml_backend_t backend); - GGML_API void ggml_backend_cpu_set_n_threads (ggml_backend_t backend_cpu, int n_threads); - GGML_API void ggml_backend_cpu_set_threadpool (ggml_backend_t backend_cpu, ggml_threadpool_t threadpool); - GGML_API void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data); + GGML_BACKEND_API bool ggml_backend_is_cpu (ggml_backend_t backend); + GGML_BACKEND_API void ggml_backend_cpu_set_n_threads (ggml_backend_t backend_cpu, int n_threads); + GGML_BACKEND_API void ggml_backend_cpu_set_threadpool (ggml_backend_t backend_cpu, ggml_threadpool_t threadpool); + GGML_BACKEND_API void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data); - GGML_API ggml_backend_reg_t ggml_backend_cpu_reg(void); - -#ifdef GGML_USE_CPU_HBM - GGML_API ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void); -#endif + GGML_BACKEND_API ggml_backend_reg_t ggml_backend_cpu_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml-cuda.h b/ggml/include/ggml-cuda.h index 305d0b636..22ad2c009 100644 --- a/ggml/include/ggml-cuda.h +++ b/ggml/include/ggml-cuda.h @@ -7,7 +7,7 @@ extern "C" { #endif -#ifdef GGML_USE_HIPBLAS +#ifdef GGML_USE_HIP #define GGML_CUDA_NAME "ROCm" #define GGML_CUBLAS_NAME "hipBLAS" #elif defined(GGML_USE_MUSA) @@ -20,27 +20,27 @@ extern "C" { #define GGML_CUDA_MAX_DEVICES 16 // backend API -GGML_API ggml_backend_t ggml_backend_cuda_init(int device); +GGML_BACKEND_API ggml_backend_t ggml_backend_cuda_init(int device); -GGML_API bool ggml_backend_is_cuda(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_cuda(ggml_backend_t backend); // device buffer -GGML_API ggml_backend_buffer_type_t ggml_backend_cuda_buffer_type(int device); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_cuda_buffer_type(int device); // split tensor buffer that splits matrices by rows across multiple devices -GGML_API ggml_backend_buffer_type_t ggml_backend_cuda_split_buffer_type(int main_device, const float * tensor_split); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_cuda_split_buffer_type(int main_device, const float * tensor_split); // pinned host buffer for use with the CPU backend for faster copies between CPU and GPU -GGML_API ggml_backend_buffer_type_t ggml_backend_cuda_host_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_cuda_host_buffer_type(void); -GGML_API int ggml_backend_cuda_get_device_count(void); -GGML_API void ggml_backend_cuda_get_device_description(int device, char * description, size_t description_size); -GGML_API void ggml_backend_cuda_get_device_memory(int device, size_t * free, size_t * total); +GGML_BACKEND_API int ggml_backend_cuda_get_device_count(void); +GGML_BACKEND_API void ggml_backend_cuda_get_device_description(int device, char * description, size_t description_size); +GGML_BACKEND_API void ggml_backend_cuda_get_device_memory(int device, size_t * free, size_t * total); -GGML_API bool ggml_backend_cuda_register_host_buffer(void * buffer, size_t size); -GGML_API void ggml_backend_cuda_unregister_host_buffer(void * buffer); +GGML_BACKEND_API bool ggml_backend_cuda_register_host_buffer(void * buffer, size_t size); +GGML_BACKEND_API void ggml_backend_cuda_unregister_host_buffer(void * buffer); -GGML_API ggml_backend_reg_t ggml_backend_cuda_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_cuda_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml-kompute.h b/ggml/include/ggml-kompute.h index c0c43521b..154aa56a7 100644 --- a/ggml/include/ggml-kompute.h +++ b/ggml/include/ggml-kompute.h @@ -37,13 +37,13 @@ struct ggml_vk_device ggml_vk_current_device(void); // forward declaration typedef struct ggml_backend * ggml_backend_t; -GGML_API ggml_backend_t ggml_backend_kompute_init(int device); +GGML_BACKEND_API ggml_backend_t ggml_backend_kompute_init(int device); -GGML_API bool ggml_backend_is_kompute(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_kompute(ggml_backend_t backend); -GGML_API ggml_backend_buffer_type_t ggml_backend_kompute_buffer_type(int device); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_kompute_buffer_type(int device); -GGML_API ggml_backend_reg_t ggml_backend_kompute_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_kompute_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml-metal.h b/ggml/include/ggml-metal.h index b8d3f678b..669c1f84a 100644 --- a/ggml/include/ggml-metal.h +++ b/ggml/include/ggml-metal.h @@ -39,27 +39,27 @@ extern "C" { // user-code should use only these functions // -GGML_API ggml_backend_t ggml_backend_metal_init(void); +GGML_BACKEND_API ggml_backend_t ggml_backend_metal_init(void); -GGML_API bool ggml_backend_is_metal(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_metal(ggml_backend_t backend); GGML_DEPRECATED( - GGML_API ggml_backend_buffer_t ggml_backend_metal_buffer_from_ptr(void * data, size_t size, size_t max_size), + GGML_BACKEND_API ggml_backend_buffer_t ggml_backend_metal_buffer_from_ptr(void * data, size_t size, size_t max_size), "obsoleted by the new device interface - https://github.com/ggerganov/llama.cpp/pull/9713"); -GGML_API void ggml_backend_metal_set_abort_callback(ggml_backend_t backend, ggml_abort_callback abort_callback, void * user_data); +GGML_BACKEND_API void ggml_backend_metal_set_abort_callback(ggml_backend_t backend, ggml_abort_callback abort_callback, void * user_data); -GGML_API ggml_backend_buffer_type_t ggml_backend_metal_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_metal_buffer_type(void); // helper to check if the device supports a specific family // ideally, the user code should be doing these checks // ref: https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf -GGML_API bool ggml_backend_metal_supports_family(ggml_backend_t backend, int family); +GGML_BACKEND_API bool ggml_backend_metal_supports_family(ggml_backend_t backend, int family); // capture all command buffers committed the next time `ggml_backend_graph_compute` is called -GGML_API void ggml_backend_metal_capture_next_compute(ggml_backend_t backend); +GGML_BACKEND_API void ggml_backend_metal_capture_next_compute(ggml_backend_t backend); -GGML_API ggml_backend_reg_t ggml_backend_metal_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_metal_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml-opencl.h b/ggml/include/ggml-opencl.h new file mode 100644 index 000000000..6b6177135 --- /dev/null +++ b/ggml/include/ggml-opencl.h @@ -0,0 +1,26 @@ +#ifndef GGML_OPENCL_H +#define GGML_OPENCL_H + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef __cplusplus +extern "C" { +#endif + +// +// backend API +// +GGML_BACKEND_API ggml_backend_t ggml_backend_opencl_init(void); +GGML_BACKEND_API bool ggml_backend_is_opencl(ggml_backend_t backend); + +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_opencl_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_opencl_host_buffer_type(void); + +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_opencl_reg(void); + +#ifdef __cplusplus +} +#endif + +#endif // GGML_OPENCL_H diff --git a/ggml/include/ggml-opt.h b/ggml/include/ggml-opt.h new file mode 100644 index 000000000..eb5eab9de --- /dev/null +++ b/ggml/include/ggml-opt.h @@ -0,0 +1,216 @@ +// This file contains functionality for training models using GGML. +// It is not strictly needed vs. just vanilla GGML but it provides a more high-level interface for common needs such as datasets. +// At the bottom of this file especially there are relatively high-level functions that are suitable use or adaptation in user code. +// +// Module maintainer: Johannes Gäßler (@JohannesGaessler, johannesg@5d6.de) + +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#include + +#ifdef __cplusplus +extern "C" { +#endif + + struct ggml_opt_dataset; + struct ggml_opt_context; + struct ggml_opt_result; + + typedef struct ggml_opt_dataset * ggml_opt_dataset_t; + typedef struct ggml_opt_context * ggml_opt_context_t; + typedef struct ggml_opt_result * ggml_opt_result_t; + + // ====== Loss ====== + + // built-in loss types, i.e. the built-in quantities minimized by the optimizer + // custom loss types can be defined via mean or sum which simply reduce the outputs for all datapoints to a single value + enum ggml_opt_loss_type { + GGML_OPT_LOSS_TYPE_MEAN, + GGML_OPT_LOSS_TYPE_SUM, + GGML_OPT_LOSS_TYPE_CROSS_ENTROPY, + GGML_OPT_LOSS_TYPE_MEAN_SQUARED_ERROR, + }; + + // ====== Dataset ====== + + GGML_API ggml_opt_dataset_t ggml_opt_dataset_init( + int64_t ne_datapoint, // number of elements per datapoint + int64_t ne_label, // number of elements per label + int64_t ndata, // total number of datapoints/labels + int64_t ndata_shard); // number of datapoints/labels per shard (unit at which the dataset is shuffled/copied) + GGML_API void ggml_opt_dataset_free(ggml_opt_dataset_t dataset); + + // get underlying tensors that store the data + GGML_API struct ggml_tensor * ggml_opt_dataset_data (ggml_opt_dataset_t dataset); // shape = [ne_datapoint, ndata] + GGML_API struct ggml_tensor * ggml_opt_dataset_labels(ggml_opt_dataset_t dataset); // shape = [nd_label, ndata] + + // shuffle idata first datapoints from dataset with RNG from opt_ctx, shuffle all datapoints if idata is negative + GGML_API void ggml_opt_dataset_shuffle(ggml_opt_context_t opt_ctx, ggml_opt_dataset_t dataset, int64_t idata); + + // get batch at position ibatch from dataset and copy the data to data_batch and labels_batch + GGML_API void ggml_opt_dataset_get_batch( + ggml_opt_dataset_t dataset, + struct ggml_tensor * data_batch, // shape = [ne_datapoint, ndata_batch] + struct ggml_tensor * labels_batch, // shape = [ne_label, ndata_batch] + int64_t ibatch); + + // ====== Model / Context ====== + + enum ggml_opt_build_type { + GGML_OPT_BUILD_TYPE_FORWARD, + GGML_OPT_BUILD_TYPE_GRAD, + GGML_OPT_BUILD_TYPE_OPT, + }; + + // parameters that control which optimizer is used and how said optimizer tries to find the minimal loss + struct ggml_opt_optimizer_params { + // AdamW optimizer parameters + struct { + float alpha; // learning rate + float beta1; + float beta2; + float eps; // epsilon for numerical stability + float wd; // weight decay for AdamW, use 0.0f to disable + } adamw; + }; + + // callback to calculate optimizer parameters prior to a backward pass + // userdata can be used to pass arbitrary data + typedef struct ggml_opt_optimizer_params (*ggml_opt_get_optimizer_params)(void * userdata); + + // returns the default optimizer params (constant) + // userdata is not used + GGML_API struct ggml_opt_optimizer_params ggml_opt_get_default_optimizer_params(void * userdata); + + // parameters for initializing a new optimization context + struct ggml_opt_params { + ggml_backend_sched_t backend_sched; // defines which backends are used to construct the compute graphs + + struct ggml_context * ctx_compute; // created in user code, holds non-static tensors + + // the forward graph is defined by inputs and outputs + // those tensors and all tensors inbetween are not intended to be reusable between multiple optimization contexts + struct ggml_tensor * inputs; + struct ggml_tensor * outputs; + + enum ggml_opt_loss_type loss_type; + enum ggml_opt_build_type build_type; + + int32_t opt_period; // after how many gradient accumulation steps an optimizer step should be done + + ggml_opt_get_optimizer_params get_opt_pars; // callback for calculating optimizer parameters + void * get_opt_pars_ud; // userdata for calculating optimizer parameters + }; + + // get parameters for an optimization context with defaults set where possible + // parameters for which no sensible defaults exist are supplied as arguments to this function + GGML_API ggml_opt_params ggml_opt_default_params( + ggml_backend_sched_t backend_sched, + struct ggml_context * ctx_compute, + struct ggml_tensor * inputs, + struct ggml_tensor * outputs, + enum ggml_opt_loss_type loss_type); + + GGML_API ggml_opt_context_t ggml_opt_init(struct ggml_opt_params params); + GGML_API void ggml_opt_free(ggml_opt_context_t opt_ctx); + + // set gradients to zero, initilize loss, and optionally reset the optimizer + GGML_API void ggml_opt_reset(ggml_opt_context_t opt_ctx, bool optimizer); + + // get underlying tensors that store data + GGML_API struct ggml_tensor * ggml_opt_inputs( ggml_opt_context_t opt_ctx); // forward graph input tensor + GGML_API struct ggml_tensor * ggml_opt_outputs( ggml_opt_context_t opt_ctx); // forward graph output tensor + GGML_API struct ggml_tensor * ggml_opt_labels( ggml_opt_context_t opt_ctx); // labels to compare outputs against + GGML_API struct ggml_tensor * ggml_opt_loss( ggml_opt_context_t opt_ctx); // scalar tensor that contains the loss + GGML_API struct ggml_tensor * ggml_opt_pred( ggml_opt_context_t opt_ctx); // predictions made by outputs + GGML_API struct ggml_tensor * ggml_opt_ncorrect(ggml_opt_context_t opt_ctx); // number of matching predictions between outputs and labels + + GGML_API struct ggml_tensor * ggml_opt_grad_acc(ggml_opt_context_t opt_ctx, struct ggml_tensor * node); + + // ====== Optimization Result ====== + + GGML_API ggml_opt_result_t ggml_opt_result_init(); + GGML_API void ggml_opt_result_free(ggml_opt_result_t result); + GGML_API void ggml_opt_result_reset(ggml_opt_result_t result); + + // get data from result, uncertainties are optional and can be ignored by passing NULL + GGML_API void ggml_opt_result_ndata( ggml_opt_result_t result, int64_t * ndata); // writes 1 value, number of datapoints + GGML_API void ggml_opt_result_loss( ggml_opt_result_t result, double * loss, double * unc); // writes 1 value + GGML_API void ggml_opt_result_pred( ggml_opt_result_t result, int32_t * pred); // writes ndata values + GGML_API void ggml_opt_result_accuracy(ggml_opt_result_t result, double * accuracy, double * unc); // writes 1 value + + // ====== Computation ====== + + // do forward pass, increment result if not NULL + GGML_API void ggml_opt_forward(ggml_opt_context_t opt_ctx, ggml_opt_result_t result); + + // do forward pass, increment result if not NULL, do backward pass + GGML_API void ggml_opt_forward_backward(ggml_opt_context_t opt_ctx, ggml_opt_result_t result); + + // ############################################################################ + // ## The high-level functions start here. They do not depend on any private ## + // ## functions or structs and can be copied to and adapted for user code. ## + // ############################################################################ + + // ====== Intended Usage ====== + // + // 1. Select the appropriate loss for your problem. + // 2. Create a dataset and set the data for the "data" tensor. Also set the "labels" tensor if your loss needs them. + // Setting the shard size to 1 will be fine, it's the granularity with which data is shuffled/loaded (bigger values are faster). + // 3. Create a GGML graph for your model with no_alloc == true. Use two separate contexts for the tensors. + // The first context should contain the model parameters and inputs and be allocated statically in user code. + // The second context should contain all other tensors and will be (re)allocated automatically. + // Due to this automated allocation the data of the second context is not defined when accessed in user code. + // Note that the second dimension of the inputs/outputs are interpreted as the number of datapoints in those tensors. + // 4. Call ggml_opt_fit. If you need more control you can use ggml_opt_epoch instead. + + // signature for a callback while evaluating opt_ctx on dataset, called after an evaluation + typedef void (*ggml_opt_epoch_callback)( + bool train, // true after training evaluation, false after validation evaluation + ggml_opt_context_t opt_ctx, + ggml_opt_dataset_t dataset, + ggml_opt_result_t result, // result associated with the dataset subsection + int64_t ibatch, // number of batches that have been evaluated so far + int64_t ibatch_max, // total number of batches in this dataset subsection + int64_t t_start_us); // time at which the evaluation on the dataset subsection was started + + // do training on front of dataset, do evaluation only on back of dataset + GGML_API void ggml_opt_epoch( + ggml_opt_context_t opt_ctx, + ggml_opt_dataset_t dataset, + ggml_opt_result_t result_train, // result to increment during training, ignored if NULL + ggml_opt_result_t result_eval, // result to increment during evaluation, ignored if NULL + int64_t idata_split, // data index at which to split training and evaluation + ggml_opt_epoch_callback callback_train, + ggml_opt_epoch_callback callback_eval); + + // callback that prints a progress bar on stderr + GGML_API void ggml_opt_epoch_callback_progress_bar( + bool train, + ggml_opt_context_t opt_ctx, + ggml_opt_dataset_t dataset, + ggml_opt_result_t result, + int64_t ibatch, + int64_t ibatch_max, + int64_t t_start_us); + + // fit model defined by inputs and outputs to dataset + GGML_API void ggml_opt_fit( + ggml_backend_sched_t backend_sched, // backend scheduler for constructing the compute graphs + ggml_context * ctx_compute, // context with temporarily allocated tensors to calculate the outputs + ggml_tensor * inputs, // input tensor with shape [ne_datapoint, ndata_batch] + ggml_tensor * outputs, // output tensor, must have shape [ne_label, ndata_batch] if labels are used + ggml_opt_dataset_t dataset, // dataset with data and optionally also labels + enum ggml_opt_loss_type loss_type, // loss to minimize + ggml_opt_get_optimizer_params get_opt_pars, // callback to get optimizer params, userdata is pointer to epoch (of type int64_t) + int64_t nepoch, // how many times the dataset should be iterated over + int64_t nbatch_logical, // datapoints optimizer step, must be a multiple of ndata_batch in inputs/outputs + float val_split, // fraction of the dataset to use for validation, must be in [0.0f, 1.0f) + bool silent); // whether or not info prints to stderr should be suppressed + +#ifdef __cplusplus +} +#endif diff --git a/ggml/include/ggml-rpc.h b/ggml/include/ggml-rpc.h index d57967368..ade6c3b0e 100644 --- a/ggml/include/ggml-rpc.h +++ b/ggml/include/ggml-rpc.h @@ -10,18 +10,18 @@ extern "C" { #define GGML_RPC_MAX_SERVERS 16 // backend API -GGML_API ggml_backend_t ggml_backend_rpc_init(const char * endpoint); -GGML_API bool ggml_backend_is_rpc(ggml_backend_t backend); +GGML_BACKEND_API ggml_backend_t ggml_backend_rpc_init(const char * endpoint); +GGML_BACKEND_API bool ggml_backend_is_rpc(ggml_backend_t backend); -GGML_API ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint); -GGML_API void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total); +GGML_BACKEND_API void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total); -GGML_API void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint, size_t free_mem, size_t total_mem); +GGML_BACKEND_API void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint, size_t free_mem, size_t total_mem); -GGML_API ggml_backend_reg_t ggml_backend_rpc_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_rpc_reg(void); -GGML_API ggml_backend_dev_t ggml_backend_rpc_add_device(const char * endpoint); +GGML_BACKEND_API ggml_backend_dev_t ggml_backend_rpc_add_device(const char * endpoint); #ifdef __cplusplus } diff --git a/ggml/include/ggml-sycl.h b/ggml/include/ggml-sycl.h index af521f599..5ce349a88 100644 --- a/ggml/include/ggml-sycl.h +++ b/ggml/include/ggml-sycl.h @@ -17,32 +17,32 @@ extern "C" { #endif // backend API -GGML_API ggml_backend_t ggml_backend_sycl_init(int device); +GGML_BACKEND_API ggml_backend_t ggml_backend_sycl_init(int device); -GGML_API bool ggml_backend_is_sycl(ggml_backend_t backend); +GGML_BACKEND_API bool ggml_backend_is_sycl(ggml_backend_t backend); // devide buffer -GGML_API ggml_backend_buffer_type_t ggml_backend_sycl_buffer_type(int device); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_sycl_buffer_type(int device); // split tensor buffer that splits matrices by rows across multiple devices -GGML_API ggml_backend_buffer_type_t ggml_backend_sycl_split_buffer_type(const float * tensor_split); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_sycl_split_buffer_type(const float * tensor_split); // pinned host buffer for use with the CPU backend for faster copies between CPU and GPU -GGML_API ggml_backend_buffer_type_t ggml_backend_sycl_host_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_sycl_host_buffer_type(void); -GGML_API void ggml_backend_sycl_print_sycl_devices(void); -GGML_API void ggml_backend_sycl_get_gpu_list(int *id_list, int max_len); -GGML_API void ggml_backend_sycl_get_device_description(int device, +GGML_BACKEND_API void ggml_backend_sycl_print_sycl_devices(void); +GGML_BACKEND_API void ggml_backend_sycl_get_gpu_list(int *id_list, int max_len); +GGML_BACKEND_API void ggml_backend_sycl_get_device_description(int device, char *description, size_t description_size); -GGML_API int ggml_backend_sycl_get_device_count(); -GGML_API void ggml_backend_sycl_get_device_memory(int device, size_t *free, size_t *total); +GGML_BACKEND_API int ggml_backend_sycl_get_device_count(); +GGML_BACKEND_API void ggml_backend_sycl_get_device_memory(int device, size_t *free, size_t *total); // SYCL doesn't support registering host memory, keep here for reference -// GGML_API bool ggml_backend_sycl_register_host_buffer(void * buffer, size_t size); -// GGML_API void ggml_backend_sycl_unregister_host_buffer(void * buffer); +// GGML_BACKEND_API bool ggml_backend_sycl_register_host_buffer(void * buffer, size_t size); +// GGML_BACKEND_API void ggml_backend_sycl_unregister_host_buffer(void * buffer); -GGML_API ggml_backend_reg_t ggml_backend_sycl_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_sycl_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml-vulkan.h b/ggml/include/ggml-vulkan.h index c03bbfe5e..53cdba072 100644 --- a/ggml/include/ggml-vulkan.h +++ b/ggml/include/ggml-vulkan.h @@ -10,21 +10,21 @@ extern "C" { #define GGML_VK_NAME "Vulkan" #define GGML_VK_MAX_DEVICES 16 -GGML_API void ggml_vk_instance_init(void); +GGML_BACKEND_API void ggml_vk_instance_init(void); // backend API -GGML_API ggml_backend_t ggml_backend_vk_init(size_t dev_num); +GGML_BACKEND_API ggml_backend_t ggml_backend_vk_init(size_t dev_num); -GGML_API bool ggml_backend_is_vk(ggml_backend_t backend); -GGML_API int ggml_backend_vk_get_device_count(void); -GGML_API void ggml_backend_vk_get_device_description(int device, char * description, size_t description_size); -GGML_API void ggml_backend_vk_get_device_memory(int device, size_t * free, size_t * total); +GGML_BACKEND_API bool ggml_backend_is_vk(ggml_backend_t backend); +GGML_BACKEND_API int ggml_backend_vk_get_device_count(void); +GGML_BACKEND_API void ggml_backend_vk_get_device_description(int device, char * description, size_t description_size); +GGML_BACKEND_API void ggml_backend_vk_get_device_memory(int device, size_t * free, size_t * total); -GGML_API ggml_backend_buffer_type_t ggml_backend_vk_buffer_type(size_t dev_num); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_vk_buffer_type(size_t dev_num); // pinned host buffer for use with the CPU backend for faster copies between CPU and GPU -GGML_API ggml_backend_buffer_type_t ggml_backend_vk_host_buffer_type(void); +GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_vk_host_buffer_type(void); -GGML_API ggml_backend_reg_t ggml_backend_vk_reg(void); +GGML_BACKEND_API ggml_backend_reg_t ggml_backend_vk_reg(void); #ifdef __cplusplus } diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h index 0d143d2fe..a9c051cd5 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -176,15 +176,15 @@ #ifdef GGML_SHARED # if defined(_WIN32) && !defined(__MINGW32__) # ifdef GGML_BUILD -# define GGML_API __declspec(dllexport) +# define GGML_API __declspec(dllexport) extern # else -# define GGML_API __declspec(dllimport) +# define GGML_API __declspec(dllimport) extern # endif # else -# define GGML_API __attribute__ ((visibility ("default"))) +# define GGML_API __attribute__ ((visibility ("default"))) extern # endif #else -# define GGML_API +# define GGML_API extern #endif // TODO: support for clang @@ -237,13 +237,9 @@ #define GGML_EXIT_SUCCESS 0 #define GGML_EXIT_ABORTED 1 -#define GGML_ROPE_TYPE_NEOX 2 - -#define GGUF_MAGIC "GGUF" - -#define GGUF_VERSION 3 - -#define GGUF_DEFAULT_ALIGNMENT 32 +#define GGML_ROPE_TYPE_NEOX 2 +#define GGML_ROPE_TYPE_MROPE 8 +#define GGML_ROPE_TYPE_VISION 24 #define GGML_UNUSED(x) (void)(x) @@ -384,12 +380,15 @@ extern "C" { GGML_TYPE_F64 = 28, GGML_TYPE_IQ1_M = 29, GGML_TYPE_BF16 = 30, - GGML_TYPE_Q4_0_4_4 = 31, - GGML_TYPE_Q4_0_4_8 = 32, - GGML_TYPE_Q4_0_8_8 = 33, + // GGML_TYPE_Q4_0_4_4 = 31, support has been removed from gguf files + // GGML_TYPE_Q4_0_4_8 = 32, + // GGML_TYPE_Q4_0_8_8 = 33, GGML_TYPE_TQ1_0 = 34, GGML_TYPE_TQ2_0 = 35, - GGML_TYPE_COUNT, + // GGML_TYPE_IQ4_NL_4_4 = 36, + // GGML_TYPE_IQ4_NL_4_8 = 37, + // GGML_TYPE_IQ4_NL_8_8 = 38, + GGML_TYPE_COUNT = 39, }; // precision @@ -398,12 +397,6 @@ extern "C" { GGML_PREC_F32, }; - enum ggml_backend_type { - GGML_BACKEND_TYPE_CPU = 0, - GGML_BACKEND_TYPE_GPU = 10, - GGML_BACKEND_TYPE_GPU_SPLIT = 20, - }; - // model file types enum ggml_ftype { GGML_FTYPE_UNKNOWN = -1, @@ -430,9 +423,6 @@ extern "C" { GGML_FTYPE_MOSTLY_IQ4_XS = 22, // except 1d tensors GGML_FTYPE_MOSTLY_IQ1_M = 23, // except 1d tensors GGML_FTYPE_MOSTLY_BF16 = 24, // except 1d tensors - GGML_FTYPE_MOSTLY_Q4_0_4_4 = 25, // except 1d tensors - GGML_FTYPE_MOSTLY_Q4_0_4_8 = 26, // except 1d tensors - GGML_FTYPE_MOSTLY_Q4_0_8_8 = 27, // except 1d tensors }; // available tensor operations: @@ -496,6 +486,7 @@ extern "C" { GGML_OP_POOL_2D_BACK, GGML_OP_UPSCALE, // nearest interpolate GGML_OP_PAD, + GGML_OP_PAD_REFLECT_1D, GGML_OP_ARANGE, GGML_OP_TIMESTEP_EMBEDDING, GGML_OP_ARGSORT, @@ -510,6 +501,7 @@ extern "C" { GGML_OP_GET_REL_POS, GGML_OP_ADD_REL_POS, GGML_OP_RWKV_WKV6, + GGML_OP_GATED_LINEAR_ATTN, GGML_OP_UNARY, @@ -584,8 +576,6 @@ extern "C" { struct ggml_tensor { enum ggml_type type; - GGML_DEPRECATED(enum ggml_backend_type backend, "use the buffer type to find the storage location of the tensor"); - struct ggml_backend_buffer * buffer; int64_t ne[GGML_MAX_DIMS]; // number of elements @@ -602,7 +592,6 @@ extern "C" { int32_t flags; - struct ggml_tensor * grad; struct ggml_tensor * src[GGML_MAX_SRC]; // source tensor and offset for views @@ -615,7 +604,7 @@ extern "C" { void * extra; // extra things e.g. for ggml-cuda.cu - // char padding[4]; + char padding[8]; }; static const size_t GGML_TENSOR_SIZE = sizeof(struct ggml_tensor); @@ -1443,6 +1432,22 @@ extern "C" { float beta_fast, float beta_slow); + GGML_API struct ggml_tensor * ggml_rope_multi( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + struct ggml_tensor * c, + int n_dims, + int sections[4], + int mode, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow); + // in-place, returns view(a) GGML_API struct ggml_tensor * ggml_rope_ext_inplace( struct ggml_context * ctx, @@ -1490,12 +1495,12 @@ extern "C" { "use ggml_rope_ext_inplace instead"); // compute correction dims for YaRN RoPE scaling - void ggml_rope_yarn_corr_dims( + GGML_API void ggml_rope_yarn_corr_dims( int n_dims, int n_ctx_orig, float freq_base, float beta_fast, float beta_slow, float dims[2]); // rotary position embedding backward, i.e compute dx from dy // a - dy - GGML_API struct ggml_tensor * ggml_rope_back( + GGML_API struct ggml_tensor * ggml_rope_ext_back( struct ggml_context * ctx, struct ggml_tensor * a, // gradients of ggml_rope result struct ggml_tensor * b, // positions @@ -1510,6 +1515,23 @@ extern "C" { float beta_fast, float beta_slow); + GGML_API struct ggml_tensor * ggml_rope_multi_back( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + struct ggml_tensor * c, + int n_dims, + int sections[4], + int mode, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow); + + // clamp // in-place, returns view(a) GGML_API struct ggml_tensor * ggml_clamp( @@ -1546,17 +1568,6 @@ extern "C" { int d1, // dilation dimension 1 bool is_2D); - GGML_API struct ggml_tensor * ggml_conv_depthwise_2d( - struct ggml_context * ctx, - struct ggml_tensor * a, // convolution kernel - struct ggml_tensor * b, // data - int s0, // stride dimension 0 - int s1, // stride dimension 1 - int p0, // padding dimension 0 - int p1, // padding dimension 1 - int d0, // dilation dimension 0 - int d1); // dilation dimension 1 - GGML_API struct ggml_tensor * ggml_conv_1d( struct ggml_context * ctx, struct ggml_tensor * a, // convolution kernel @@ -1574,6 +1585,23 @@ extern "C" { int s, // stride int d); // dilation + // depthwise + // TODO: this is very likely wrong for some cases! - needs more testing + GGML_API struct ggml_tensor * ggml_conv_1d_dw( + struct ggml_context * ctx, + struct ggml_tensor * a, // convolution kernel + struct ggml_tensor * b, // data + int s0, // stride + int p0, // padding + int d0); // dilation + + GGML_API struct ggml_tensor * ggml_conv_1d_dw_ph( + struct ggml_context * ctx, + struct ggml_tensor * a, // convolution kernel + struct ggml_tensor * b, // data + int s0, // stride + int d0); // dilation + GGML_API struct ggml_tensor * ggml_conv_transpose_1d( struct ggml_context * ctx, struct ggml_tensor * a, // convolution kernel @@ -1593,7 +1621,6 @@ extern "C" { int d0, // dilation dimension 0 int d1); // dilation dimension 1 - // kernel size is a->ne[0] x a->ne[1] // stride is equal to kernel size // padding is zero @@ -1620,6 +1647,18 @@ extern "C" { struct ggml_tensor * a, struct ggml_tensor * b); + // depthwise + GGML_API struct ggml_tensor * ggml_conv_2d_dw( + struct ggml_context * ctx, + struct ggml_tensor * a, // convolution kernel + struct ggml_tensor * b, // data + int s0, // stride dimension 0 + int s1, // stride dimension 1 + int p0, // padding dimension 0 + int p1, // padding dimension 1 + int d0, // dilation dimension 0 + int d1); // dilation dimension 1 + GGML_API struct ggml_tensor * ggml_conv_transpose_2d_p0( struct ggml_context * ctx, struct ggml_tensor * a, @@ -1693,6 +1732,13 @@ extern "C" { int p2, int p3); + // pad each dimension with reflection: [a, b, c, d] -> [b, a, b, c, d, c] + GGML_API struct ggml_tensor * ggml_pad_reflect_1d( + struct ggml_context * ctx, + struct ggml_tensor * a, + int p0, + int p1); + // Ref: https://github.com/CompVis/stable-diffusion/blob/main/ldm/modules/diffusionmodules/util.py#L151 // timesteps: [N,] // return: [N, dim] @@ -1746,6 +1792,9 @@ extern "C" { struct ggml_tensor * a, enum ggml_prec prec); + GGML_API enum ggml_prec ggml_flash_attn_ext_get_prec( + const struct ggml_tensor * a); + // TODO: needs to be adapted to ggml_flash_attn_ext GGML_API struct ggml_tensor * ggml_flash_attn_back( struct ggml_context * ctx, @@ -1828,6 +1877,15 @@ extern "C" { struct ggml_tensor * td, struct ggml_tensor * state); + GGML_API struct ggml_tensor * ggml_gated_linear_attn( + struct ggml_context * ctx, + struct ggml_tensor * k, + struct ggml_tensor * v, + struct ggml_tensor * q, + struct ggml_tensor * g, + struct ggml_tensor * state, + float scale); + // custom operators typedef void (*ggml_unary_op_f32_t) (const int, float *, const float *); @@ -1982,28 +2040,20 @@ extern "C" { struct ggml_context * ctx, struct ggml_tensor * a, struct ggml_tensor * grad, - float alpha, - float beta1, - float beta2, - float eps, - float wd); // weight decay + struct ggml_tensor * m, + struct ggml_tensor * v, + struct ggml_tensor * adamw_params); // parameters such a the learning rate // // automatic differentiation // - GGML_API void ggml_build_forward_expand (struct ggml_cgraph * cgraph, struct ggml_tensor * tensor); - GGML_API void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * gf, struct ggml_cgraph * gb, bool accumulate); - - GGML_API void ggml_build_opt_adamw( - struct ggml_context * ctx, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - float alpha, - float beta1, - float beta2, - float eps, - float wd); // weight decay + GGML_API void ggml_build_forward_expand(struct ggml_cgraph * cgraph, struct ggml_tensor * tensor); + GGML_API void ggml_build_backward_expand( + struct ggml_context * ctx_static, // context for static gradients (loss + gradient accumulation) + struct ggml_context * ctx_compute, // context for gradient computation + struct ggml_cgraph * cgraph, + bool accumulate); // whether or not gradients should be accumulated, requires static allocation of tensors in ctx_static // graph allocation in a context GGML_API struct ggml_cgraph * ggml_new_graph (struct ggml_context * ctx); // size = GGML_DEFAULT_GRAPH_SIZE, grads = false @@ -2023,7 +2073,9 @@ extern "C" { GGML_API size_t ggml_graph_overhead(void); GGML_API size_t ggml_graph_overhead_custom(size_t size, bool grads); - GGML_API struct ggml_tensor * ggml_graph_get_tensor(struct ggml_cgraph * cgraph, const char * name); + GGML_API struct ggml_tensor * ggml_graph_get_tensor (const struct ggml_cgraph * cgraph, const char * name); + GGML_API struct ggml_tensor * ggml_graph_get_grad (const struct ggml_cgraph * cgraph, const struct ggml_tensor * node); + GGML_API struct ggml_tensor * ggml_graph_get_grad_acc(const struct ggml_cgraph * cgraph, const struct ggml_tensor * node); GGML_API void ggml_graph_export(const struct ggml_cgraph * cgraph, const char * fname); GGML_API struct ggml_cgraph * ggml_graph_import(const char * fname, struct ggml_context ** ctx_data, struct ggml_context ** ctx_eval); @@ -2034,198 +2086,15 @@ extern "C" { // dump the graph into a file using the dot format GGML_API void ggml_graph_dump_dot(const struct ggml_cgraph * gb, const struct ggml_cgraph * gf, const char * filename); - // build gradient checkpointing backward graph gb for gf using provided checkpoints - // gb_tmp will contain original backward graph with rewritten backward process nodes, - // but without the second forward pass nodes. - GGML_API void ggml_build_backward_gradient_checkpointing( - struct ggml_context * ctx, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - struct ggml_cgraph * gb_tmp, - struct ggml_tensor * * checkpoints, - int n_checkpoints); - // - // optimization - // - - // optimization methods - enum ggml_opt_type { - GGML_OPT_TYPE_ADAM, - GGML_OPT_TYPE_LBFGS, - }; - - // linesearch methods - enum ggml_linesearch { - GGML_LINESEARCH_DEFAULT = 1, - - GGML_LINESEARCH_BACKTRACKING_ARMIJO = 0, - GGML_LINESEARCH_BACKTRACKING_WOLFE = 1, - GGML_LINESEARCH_BACKTRACKING_STRONG_WOLFE = 2, - }; - - // optimization return values - enum ggml_opt_result { - GGML_OPT_RESULT_OK = 0, - GGML_OPT_RESULT_DID_NOT_CONVERGE, - GGML_OPT_RESULT_NO_CONTEXT, - GGML_OPT_RESULT_INVALID_WOLFE, - GGML_OPT_RESULT_FAIL, - GGML_OPT_RESULT_CANCEL, - - GGML_LINESEARCH_FAIL = -128, - GGML_LINESEARCH_MINIMUM_STEP, - GGML_LINESEARCH_MAXIMUM_STEP, - GGML_LINESEARCH_MAXIMUM_ITERATIONS, - GGML_LINESEARCH_INVALID_PARAMETERS, - }; - - typedef void (*ggml_opt_callback)(void * data, int accum_step, float * sched, bool * cancel); + // TODO these functions were sandwiched in the old optimization interface, is there a better place for them? typedef void (*ggml_log_callback)(enum ggml_log_level level, const char * text, void * user_data); // Set callback for all future logging events. // If this is not called, or NULL is supplied, everything is output on stderr. GGML_API void ggml_log_set(ggml_log_callback log_callback, void * user_data); - // optimization parameters - // - // see ggml.c (ggml_opt_default_params) for default values - // - struct ggml_opt_params { - enum ggml_opt_type type; - - size_t graph_size; - - int n_threads; - - // delta-based convergence test - // - // if past == 0 - disabled - // if past > 0: - // stop if |f(x) - f(x_past)| < delta * max(1, |f(x)|) - // - int past; - float delta; - - // maximum number of iterations without improvement - // - // if 0 - disabled - // if > 0: - // assume convergence if no cost improvement in this number of iterations - // - int max_no_improvement; - - bool print_forward_graph; - bool print_backward_graph; - - int n_gradient_accumulation; - - // ADAM parameters - struct { - int n_iter; - - float sched; // schedule multiplier (fixed, decay or warmup) - float decay; // weight decay for AdamW, use 0.0f to disable - int decay_min_ndim; // minimum number of tensor dimension to apply weight decay - float alpha; // learning rate - float beta1; - float beta2; - float eps; // epsilon for numerical stability - float eps_f; // epsilon for convergence test - float eps_g; // epsilon for convergence test - float gclip; // gradient clipping - } adam; - - // LBFGS parameters - struct { - int m; // number of corrections to approximate the inv. Hessian - int n_iter; - int max_linesearch; - - float eps; // convergence tolerance - float ftol; // line search tolerance - float wolfe; - float min_step; - float max_step; - - enum ggml_linesearch linesearch; - } lbfgs; - }; - - struct ggml_opt_context { - struct ggml_context * ctx; - struct ggml_opt_params params; - - int iter; - int64_t nx; // number of parameter elements - - bool just_initialized; - - float loss_before; - float loss_after; - - struct { - struct ggml_tensor * g; // current gradient - struct ggml_tensor * m; // first moment - struct ggml_tensor * v; // second moment - struct ggml_tensor * pf; // past function values - float fx_best; - float fx_prev; - int n_no_improvement; - } adam; - - struct { - struct ggml_tensor * x; // current parameters - struct ggml_tensor * xp; // previous parameters - struct ggml_tensor * g; // current gradient - struct ggml_tensor * gp; // previous gradient - struct ggml_tensor * d; // search direction - struct ggml_tensor * pf; // past function values - struct ggml_tensor * lmal; // the L-BFGS memory alpha - struct ggml_tensor * lmys; // the L-BFGS memory ys - struct ggml_tensor * lms; // the L-BFGS memory s - struct ggml_tensor * lmy; // the L-BFGS memory y - float fx_best; - float step; - int j; - int k; - int end; - int n_no_improvement; - } lbfgs; - }; - GGML_API struct ggml_tensor * ggml_set_zero(struct ggml_tensor * tensor); - GGML_API struct ggml_opt_params ggml_opt_default_params(enum ggml_opt_type type); - - // optimize the function defined by the tensor f - GGML_API enum ggml_opt_result ggml_opt( - struct ggml_context * ctx, - struct ggml_opt_params params, - struct ggml_tensor * f); - - // initialize optimizer context - GGML_API void ggml_opt_init( - struct ggml_context * ctx, - struct ggml_opt_context * opt, - struct ggml_opt_params params, - int64_t nx); - - // continue optimizing the function defined by the tensor f - GGML_API enum ggml_opt_result ggml_opt_resume( - struct ggml_context * ctx, - struct ggml_opt_context * opt, - struct ggml_tensor * f); - - // continue optimizing the function defined by the tensor f - GGML_API enum ggml_opt_result ggml_opt_resume_g( - struct ggml_context * ctx, - struct ggml_opt_context * opt, - struct ggml_tensor * f, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - ggml_opt_callback callback, - void * callback_data); - // // quantization // @@ -2255,169 +2124,19 @@ extern "C" { int64_t n_per_row, const float * imatrix); - // - // gguf - // - - enum gguf_type { - GGUF_TYPE_UINT8 = 0, - GGUF_TYPE_INT8 = 1, - GGUF_TYPE_UINT16 = 2, - GGUF_TYPE_INT16 = 3, - GGUF_TYPE_UINT32 = 4, - GGUF_TYPE_INT32 = 5, - GGUF_TYPE_FLOAT32 = 6, - GGUF_TYPE_BOOL = 7, - GGUF_TYPE_STRING = 8, - GGUF_TYPE_ARRAY = 9, - GGUF_TYPE_UINT64 = 10, - GGUF_TYPE_INT64 = 11, - GGUF_TYPE_FLOAT64 = 12, - GGUF_TYPE_COUNT, // marks the end of the enum - }; - - struct gguf_context; - - struct gguf_init_params { - bool no_alloc; - - // if not NULL, create a ggml_context and allocate the tensor data in it - struct ggml_context ** ctx; - }; - - GGML_API struct gguf_context * gguf_init_empty(void); - GGML_API struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_params params); - //GGML_API struct gguf_context * gguf_init_from_buffer(..); - - GGML_API void gguf_free(struct gguf_context * ctx); - - GGML_API const char * gguf_type_name(enum gguf_type type); - - GGML_API int gguf_get_version (const struct gguf_context * ctx); - GGML_API size_t gguf_get_alignment (const struct gguf_context * ctx); - GGML_API size_t gguf_get_data_offset(const struct gguf_context * ctx); - GGML_API void * gguf_get_data (const struct gguf_context * ctx); - - GGML_API int gguf_get_n_kv(const struct gguf_context * ctx); - GGML_API int gguf_find_key(const struct gguf_context * ctx, const char * key); - GGML_API const char * gguf_get_key (const struct gguf_context * ctx, int key_id); - - GGML_API enum gguf_type gguf_get_kv_type (const struct gguf_context * ctx, int key_id); - GGML_API enum gguf_type gguf_get_arr_type(const struct gguf_context * ctx, int key_id); - - // will abort if the wrong type is used for the key - GGML_API uint8_t gguf_get_val_u8 (const struct gguf_context * ctx, int key_id); - GGML_API int8_t gguf_get_val_i8 (const struct gguf_context * ctx, int key_id); - GGML_API uint16_t gguf_get_val_u16 (const struct gguf_context * ctx, int key_id); - GGML_API int16_t gguf_get_val_i16 (const struct gguf_context * ctx, int key_id); - GGML_API uint32_t gguf_get_val_u32 (const struct gguf_context * ctx, int key_id); - GGML_API int32_t gguf_get_val_i32 (const struct gguf_context * ctx, int key_id); - GGML_API float gguf_get_val_f32 (const struct gguf_context * ctx, int key_id); - GGML_API uint64_t gguf_get_val_u64 (const struct gguf_context * ctx, int key_id); - GGML_API int64_t gguf_get_val_i64 (const struct gguf_context * ctx, int key_id); - GGML_API double gguf_get_val_f64 (const struct gguf_context * ctx, int key_id); - GGML_API bool gguf_get_val_bool(const struct gguf_context * ctx, int key_id); - GGML_API const char * gguf_get_val_str (const struct gguf_context * ctx, int key_id); - GGML_API const void * gguf_get_val_data(const struct gguf_context * ctx, int key_id); - GGML_API int gguf_get_arr_n (const struct gguf_context * ctx, int key_id); - GGML_API const void * gguf_get_arr_data(const struct gguf_context * ctx, int key_id); - GGML_API const char * gguf_get_arr_str (const struct gguf_context * ctx, int key_id, int i); - - GGML_API int gguf_get_n_tensors (const struct gguf_context * ctx); - GGML_API int gguf_find_tensor (const struct gguf_context * ctx, const char * name); - GGML_API size_t gguf_get_tensor_offset(const struct gguf_context * ctx, int i); - GGML_API char * gguf_get_tensor_name (const struct gguf_context * ctx, int i); - GGML_API enum ggml_type gguf_get_tensor_type (const struct gguf_context * ctx, int i); - - // removes key if it exists - GGML_API void gguf_remove_key(struct gguf_context * ctx, const char * key); - - // overrides existing values or adds a new one - GGML_API void gguf_set_val_u8 (struct gguf_context * ctx, const char * key, uint8_t val); - GGML_API void gguf_set_val_i8 (struct gguf_context * ctx, const char * key, int8_t val); - GGML_API void gguf_set_val_u16 (struct gguf_context * ctx, const char * key, uint16_t val); - GGML_API void gguf_set_val_i16 (struct gguf_context * ctx, const char * key, int16_t val); - GGML_API void gguf_set_val_u32 (struct gguf_context * ctx, const char * key, uint32_t val); - GGML_API void gguf_set_val_i32 (struct gguf_context * ctx, const char * key, int32_t val); - GGML_API void gguf_set_val_f32 (struct gguf_context * ctx, const char * key, float val); - GGML_API void gguf_set_val_u64 (struct gguf_context * ctx, const char * key, uint64_t val); - GGML_API void gguf_set_val_i64 (struct gguf_context * ctx, const char * key, int64_t val); - GGML_API void gguf_set_val_f64 (struct gguf_context * ctx, const char * key, double val); - GGML_API void gguf_set_val_bool(struct gguf_context * ctx, const char * key, bool val); - GGML_API void gguf_set_val_str (struct gguf_context * ctx, const char * key, const char * val); - GGML_API void gguf_set_arr_data(struct gguf_context * ctx, const char * key, enum gguf_type type, const void * data, int n); - GGML_API void gguf_set_arr_str (struct gguf_context * ctx, const char * key, const char ** data, int n); - - // set or add KV pairs from another context - GGML_API void gguf_set_kv(struct gguf_context * ctx, struct gguf_context * src); - - // manage tensor info - GGML_API void gguf_add_tensor(struct gguf_context * ctx, const struct ggml_tensor * tensor); - GGML_API void gguf_set_tensor_type(struct gguf_context * ctx, const char * name, enum ggml_type type); - GGML_API void gguf_set_tensor_data(struct gguf_context * ctx, const char * name, const void * data, size_t size); - - // writing gguf files can be done in 2 ways: - // - // - write the entire gguf_context to a binary file in a single pass: - // - // gguf_write_to_file(ctx, fname); - // - // - first prepare a file with a placeholder for the meta data, write the tensor data, then write the meta data: - // - // FILE * f = fopen(fname, "wb"); - // fseek(f, gguf_get_meta_size(ctx), SEEK_SET); - // fwrite(f, ...); - // void * data = gguf_meta_get_meta_data(ctx); - // fseek(f, 0, SEEK_SET); - // fwrite(f, data, gguf_get_meta_size(ctx)); - // free(data); - // fclose(f); - // - - // write the entire context to a binary file - GGML_API void gguf_write_to_file(const struct gguf_context * ctx, const char * fname, bool only_meta); - - // get the size in bytes of the meta data (header, kv pairs, tensor info) including padding - GGML_API size_t gguf_get_meta_size(const struct gguf_context * ctx); - GGML_API void gguf_get_meta_data(const struct gguf_context * ctx, void * data); - - // - // system info - // - - GGML_API int ggml_cpu_has_avx (void); - GGML_API int ggml_cpu_has_avx_vnni (void); - GGML_API int ggml_cpu_has_avx2 (void); - GGML_API int ggml_cpu_has_avx512 (void); - GGML_API int ggml_cpu_has_avx512_vbmi(void); - GGML_API int ggml_cpu_has_avx512_vnni(void); - GGML_API int ggml_cpu_has_avx512_bf16(void); - GGML_API int ggml_cpu_has_amx_int8 (void); - GGML_API int ggml_cpu_has_fma (void); - GGML_API int ggml_cpu_has_arm_fma (void); - GGML_API int ggml_cpu_has_metal (void); - GGML_API int ggml_cpu_has_f16c (void); - GGML_API int ggml_cpu_has_fp16_va (void); - GGML_API int ggml_cpu_has_wasm_simd (void); - GGML_API int ggml_cpu_has_blas (void); - GGML_API int ggml_cpu_has_cuda (void); - GGML_API int ggml_cpu_has_vulkan (void); - GGML_API int ggml_cpu_has_kompute (void); - GGML_API int ggml_cpu_has_gpublas (void); - GGML_API int ggml_cpu_has_sse3 (void); - GGML_API int ggml_cpu_has_ssse3 (void); - GGML_API int ggml_cpu_has_riscv_v (void); - GGML_API int ggml_cpu_has_sycl (void); - GGML_API int ggml_cpu_has_rpc (void); - GGML_API int ggml_cpu_has_vsx (void); - GGML_API int ggml_cpu_has_cann (void); - GGML_API int ggml_cpu_has_llamafile (void); - -#ifdef __cplusplus -// restrict not standard in C++ -#define GGML_RESTRICT +#ifdef __cplusplus + // restrict not standard in C++ +# if defined(__GNUC__) +# define GGML_RESTRICT __restrict__ +# elif defined(__clang__) +# define GGML_RESTRICT __restrict +# elif defined(_MSC_VER) +# define GGML_RESTRICT __restrict +# else +# define GGML_RESTRICT +# endif #else -#define GGML_RESTRICT restrict +# define GGML_RESTRICT restrict #endif typedef void (*ggml_to_float_t) (const void * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); typedef void (*ggml_from_float_t)(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); @@ -2429,12 +2148,42 @@ extern "C" { size_t type_size; bool is_quantized; ggml_to_float_t to_float; - ggml_from_float_t from_float; ggml_from_float_t from_float_ref; }; GGML_API const struct ggml_type_traits * ggml_get_type_traits(enum ggml_type type); + // ggml threadpool + // TODO: currently, only a few functions are in the base ggml API, while the rest are in the CPU backend + // the goal should be to create an API that other backends can use move everything to the ggml base + + // scheduling priorities + enum ggml_sched_priority { + GGML_SCHED_PRIO_NORMAL, + GGML_SCHED_PRIO_MEDIUM, + GGML_SCHED_PRIO_HIGH, + GGML_SCHED_PRIO_REALTIME + }; + + // threadpool params + // Use ggml_threadpool_params_default() or ggml_threadpool_params_init() to populate the defaults + struct ggml_threadpool_params { + bool cpumask[GGML_MAX_N_THREADS]; // mask of cpu cores (all-zeros means use default affinity settings) + int n_threads; // number of threads + enum ggml_sched_priority prio; // thread priority + uint32_t poll; // polling level (0 - no polling, 100 - aggressive polling) + bool strict_cpu; // strict cpu placement + bool paused; // start in paused state + }; + + struct ggml_threadpool; // forward declaration, see ggml.c + + typedef struct ggml_threadpool * ggml_threadpool_t; + + GGML_API struct ggml_threadpool_params ggml_threadpool_params_default(int n_threads); + GGML_API void ggml_threadpool_params_init (struct ggml_threadpool_params * p, int n_threads); + GGML_API bool ggml_threadpool_params_match (const struct ggml_threadpool_params * p0, const struct ggml_threadpool_params * p1); + #ifdef __cplusplus } #endif diff --git a/ggml/include/gguf.h b/ggml/include/gguf.h new file mode 100644 index 000000000..79ee20206 --- /dev/null +++ b/ggml/include/gguf.h @@ -0,0 +1,202 @@ +// This file contains functionality related to "GGUF" files, the binary file format used by ggml. +// GGUF files have the following structure: +// +// 1. File magic "GGUF" (4 bytes). +// 2. File version (uint32_t). +// 3. Number of ggml tensors in file (int64_t). +// 4. Number of key-value-pairs in file (int64_t). +// 5. For each KV pair: +// 1. The key (string). +// 2. The value type (gguf_type). +// 3a. If the value type is GGUF_TYPE_ARRAY: +// 1. The type of the array (gguf_type). +// 2. The number of elements in the array (uint64_t). +// 3. The binary representation of each element in the array. +// 3b. Otherwise: +// 1. The binary representation of the value. +// 6. For each ggml tensor: +// 1. The tensor name (string). +// 2. The number of dimensions of the tensor (uint32_t). +// 3. For each dimension: +// 1. The size of the tensor in the dimension (int64_t). +// 4. The tensor data type (ggml_type). +// 5. The tensor data offset in the tensor data binary blob (uint64_t). +// 7. The tensor data binary blob (optional, aligned). +// +// Strings are serialized as the string length (uint64_t) followed by the C string without the null terminator. +// All enums are stored as int32_t. +// All bool values are stored as int8_t. +// If the special key "general.alignment" (uint32_t) is defined it is used for alignment, +// otherwise GGUF_DEFAULT_ALIGNMENT is used. +// +// Module maintainer: Johannes Gäßler (@JohannesGaessler, johannesg@5d6.de) + +#pragma once + +#include "ggml.h" + +#include +#include + +#define GGUF_MAGIC "GGUF" +#define GGUF_VERSION 3 + +#define GGUF_KEY_GENERAL_ALIGNMENT "general.alignment" + +#define GGUF_DEFAULT_ALIGNMENT 32 + +#ifdef __cplusplus +extern "C" { +#endif + + // types that can be stored as GGUF KV data + enum gguf_type { + GGUF_TYPE_UINT8 = 0, + GGUF_TYPE_INT8 = 1, + GGUF_TYPE_UINT16 = 2, + GGUF_TYPE_INT16 = 3, + GGUF_TYPE_UINT32 = 4, + GGUF_TYPE_INT32 = 5, + GGUF_TYPE_FLOAT32 = 6, + GGUF_TYPE_BOOL = 7, + GGUF_TYPE_STRING = 8, + GGUF_TYPE_ARRAY = 9, + GGUF_TYPE_UINT64 = 10, + GGUF_TYPE_INT64 = 11, + GGUF_TYPE_FLOAT64 = 12, + GGUF_TYPE_COUNT, // marks the end of the enum + }; + + struct gguf_context; + + struct gguf_init_params { + bool no_alloc; + + // if not NULL, create a ggml_context and allocate the tensor data in it + struct ggml_context ** ctx; + }; + + GGML_API struct gguf_context * gguf_init_empty(void); + GGML_API struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_params params); + //GGML_API struct gguf_context * gguf_init_from_buffer(..); + + GGML_API void gguf_free(struct gguf_context * ctx); + + GGML_API const char * gguf_type_name(enum gguf_type type); + + GGML_API uint32_t gguf_get_version (const struct gguf_context * ctx); + GGML_API size_t gguf_get_alignment (const struct gguf_context * ctx); + GGML_API size_t gguf_get_data_offset(const struct gguf_context * ctx); + + GGML_API int64_t gguf_get_n_kv(const struct gguf_context * ctx); + GGML_API int64_t gguf_find_key(const struct gguf_context * ctx, const char * key); // returns -1 if key is not found + GGML_API const char * gguf_get_key (const struct gguf_context * ctx, int64_t key_id); + + GGML_API enum gguf_type gguf_get_kv_type (const struct gguf_context * ctx, int64_t key_id); + GGML_API enum gguf_type gguf_get_arr_type(const struct gguf_context * ctx, int64_t key_id); + + // will abort if the wrong type is used for the key + GGML_API uint8_t gguf_get_val_u8 (const struct gguf_context * ctx, int64_t key_id); + GGML_API int8_t gguf_get_val_i8 (const struct gguf_context * ctx, int64_t key_id); + GGML_API uint16_t gguf_get_val_u16 (const struct gguf_context * ctx, int64_t key_id); + GGML_API int16_t gguf_get_val_i16 (const struct gguf_context * ctx, int64_t key_id); + GGML_API uint32_t gguf_get_val_u32 (const struct gguf_context * ctx, int64_t key_id); + GGML_API int32_t gguf_get_val_i32 (const struct gguf_context * ctx, int64_t key_id); + GGML_API float gguf_get_val_f32 (const struct gguf_context * ctx, int64_t key_id); + GGML_API uint64_t gguf_get_val_u64 (const struct gguf_context * ctx, int64_t key_id); + GGML_API int64_t gguf_get_val_i64 (const struct gguf_context * ctx, int64_t key_id); + GGML_API double gguf_get_val_f64 (const struct gguf_context * ctx, int64_t key_id); + GGML_API bool gguf_get_val_bool(const struct gguf_context * ctx, int64_t key_id); + GGML_API const char * gguf_get_val_str (const struct gguf_context * ctx, int64_t key_id); + GGML_API const void * gguf_get_val_data(const struct gguf_context * ctx, int64_t key_id); + GGML_API size_t gguf_get_arr_n (const struct gguf_context * ctx, int64_t key_id); + + // get raw pointer to the first element of the array with the given key_id + // for bool arrays, note that they are always stored as int8 on all platforms (usually this makes no difference) + GGML_API const void * gguf_get_arr_data(const struct gguf_context * ctx, int64_t key_id); + + // get ith C string from array with given key_id + GGML_API const char * gguf_get_arr_str (const struct gguf_context * ctx, int64_t key_id, size_t i); + + GGML_API int64_t gguf_get_n_tensors (const struct gguf_context * ctx); + GGML_API int64_t gguf_find_tensor (const struct gguf_context * ctx, const char * name); // returns -1 if the tensor is not found + GGML_API size_t gguf_get_tensor_offset(const struct gguf_context * ctx, int64_t tensor_id); + GGML_API const char * gguf_get_tensor_name (const struct gguf_context * ctx, int64_t tensor_id); + GGML_API enum ggml_type gguf_get_tensor_type (const struct gguf_context * ctx, int64_t tensor_id); + GGML_API size_t gguf_get_tensor_size (const struct gguf_context * ctx, int64_t tensor_id); + + // removes key if it exists, returns id that the key had prior to removal (-1 if it didn't exist) + GGML_API int64_t gguf_remove_key(struct gguf_context * ctx, const char * key); + + // overrides an existing KV pair or adds a new one, the new KV pair is always at the back + GGML_API void gguf_set_val_u8 (struct gguf_context * ctx, const char * key, uint8_t val); + GGML_API void gguf_set_val_i8 (struct gguf_context * ctx, const char * key, int8_t val); + GGML_API void gguf_set_val_u16 (struct gguf_context * ctx, const char * key, uint16_t val); + GGML_API void gguf_set_val_i16 (struct gguf_context * ctx, const char * key, int16_t val); + GGML_API void gguf_set_val_u32 (struct gguf_context * ctx, const char * key, uint32_t val); + GGML_API void gguf_set_val_i32 (struct gguf_context * ctx, const char * key, int32_t val); + GGML_API void gguf_set_val_f32 (struct gguf_context * ctx, const char * key, float val); + GGML_API void gguf_set_val_u64 (struct gguf_context * ctx, const char * key, uint64_t val); + GGML_API void gguf_set_val_i64 (struct gguf_context * ctx, const char * key, int64_t val); + GGML_API void gguf_set_val_f64 (struct gguf_context * ctx, const char * key, double val); + GGML_API void gguf_set_val_bool(struct gguf_context * ctx, const char * key, bool val); + GGML_API void gguf_set_val_str (struct gguf_context * ctx, const char * key, const char * val); + + // creates a new array with n elements of the given type and copies the corresponding number of bytes from data + GGML_API void gguf_set_arr_data(struct gguf_context * ctx, const char * key, enum gguf_type type, const void * data, size_t n); + + // creates a new array with n strings and copies the corresponding strings from data + GGML_API void gguf_set_arr_str (struct gguf_context * ctx, const char * key, const char ** data, size_t n); + + // set or add KV pairs from another context + GGML_API void gguf_set_kv(struct gguf_context * ctx, const struct gguf_context * src); + + // add tensor to GGUF context, tensor name must be unique + GGML_API void gguf_add_tensor(struct gguf_context * ctx, const struct ggml_tensor * tensor); + + // after changing a tensor's type, the offsets of all tensors with higher indices are immediately recalculated + // in such a way that the tensor data remains as one contiguous block (except for padding) + GGML_API void gguf_set_tensor_type(struct gguf_context * ctx, const char * name, enum ggml_type type); + + // assumes that at least gguf_get_tensor_size bytes can be read from data + GGML_API void gguf_set_tensor_data(struct gguf_context * ctx, const char * name, const void * data); + + // writing gguf files can be done in 3 ways: + // + // - write the entire gguf_context to a binary file in a single pass: + // + // gguf_write_to_file(ctx, fname, /*only_meta =*/ false); + // + // - write only the meta data to a file, then re-open the file and append the tensor data: + // + // gguf_write_to_file(ctx, fname, /*only_meta =*/ true); + // FILE * f = fopen(fname, "ab"); + // fwrite(f, ...); // write tensor data + // fclose(f); + // + // - first prepare a file with a placeholder for the meta data, write the tensor data, then write the meta data: + // + // FILE * f = fopen(fname, "wb"); + // const size_t size_meta = gguf_get_meta_size(ctx); + // fseek(f, size_meta, SEEK_SET); + // fwrite(f, ...); // write tensor data + // void * data = malloc(size_meta); + // gguf_get_meta_data(ctx, data); + // rewind(f); + // fwrite(data, 1, data, f); + // free(data); + // fclose(f); + // + + // write the entire context to a binary file + GGML_API bool gguf_write_to_file(const struct gguf_context * ctx, const char * fname, bool only_meta); + + // get the size in bytes of the meta data (header, kv pairs, tensor info) including padding + GGML_API size_t gguf_get_meta_size(const struct gguf_context * ctx); + + // writes the meta data to pointer "data" + GGML_API void gguf_get_meta_data(const struct gguf_context * ctx, void * data); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/CMakeLists.txt b/ggml/src/CMakeLists.txt index 34b81bd7f..ae1cd2337 100644 --- a/ggml/src/CMakeLists.txt +++ b/ggml/src/CMakeLists.txt @@ -1,7 +1,5 @@ include(CheckCXXCompilerFlag) -unset(GGML_CDEF_PUBLIC) - add_compile_definitions(GGML_SCHED_MAX_COPIES=${GGML_SCHED_MAX_COPIES}) # enable libstdc++ assertions for debug builds @@ -26,926 +24,7 @@ if (NOT MSVC) endif() endif() -unset(GGML_EXTRA_LIBS_PRIVATE) -unset(GGML_EXTRA_LIBS_PUBLIC) - -if (APPLE AND GGML_ACCELERATE) - find_library(ACCELERATE_FRAMEWORK Accelerate) - if (ACCELERATE_FRAMEWORK) - message(STATUS "Accelerate framework found") - - add_compile_definitions(GGML_USE_ACCELERATE) - add_compile_definitions(ACCELERATE_NEW_LAPACK) - add_compile_definitions(ACCELERATE_LAPACK_ILP64) - - list(APPEND GGML_EXTRA_LIBS_PRIVATE ${ACCELERATE_FRAMEWORK}) - else() - message(WARNING "Accelerate framework not found") - endif() -endif() - -if (GGML_METAL) - find_library(FOUNDATION_LIBRARY Foundation REQUIRED) - find_library(METAL_FRAMEWORK Metal REQUIRED) - find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) - - message(STATUS "Metal framework found") - set(GGML_HEADERS_METAL ../include/ggml-metal.h) - set(GGML_SOURCES_METAL ggml-metal.m) - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_METAL) - if (GGML_METAL_NDEBUG) - add_compile_definitions(GGML_METAL_NDEBUG) - endif() - - # copy ggml-common.h and ggml-metal.metal to bin directory - configure_file(ggml-common.h ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h COPYONLY) - configure_file(ggml-metal.metal ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal COPYONLY) - - if (GGML_METAL_EMBED_LIBRARY) - enable_language(ASM) - - add_compile_definitions(GGML_METAL_EMBED_LIBRARY) - - set(METALLIB_COMMON "${CMAKE_CURRENT_SOURCE_DIR}/ggml-common.h") - set(METALLIB_SOURCE "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal.metal") - - file(MAKE_DIRECTORY "${CMAKE_BINARY_DIR}/autogenerated") - - # merge ggml-common.h and ggml-metal.metal into a single file - set(METALLIB_EMBED_ASM "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.s") - set(METALLIB_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.metal") - - add_custom_command( - OUTPUT ${METALLIB_EMBED_ASM} - COMMAND echo "Embedding Metal library" - COMMAND sed -e '/\#include \"ggml-common.h\"/r ${METALLIB_COMMON}' -e '/\#include \"ggml-common.h\"/d' < ${METALLIB_SOURCE} > ${METALLIB_SOURCE_EMBED} - COMMAND echo ".section __DATA,__ggml_metallib" > ${METALLIB_EMBED_ASM} - COMMAND echo ".globl _ggml_metallib_start" >> ${METALLIB_EMBED_ASM} - COMMAND echo "_ggml_metallib_start:" >> ${METALLIB_EMBED_ASM} - COMMAND echo ".incbin \\\"${METALLIB_SOURCE_EMBED}\\\"" >> ${METALLIB_EMBED_ASM} - COMMAND echo ".globl _ggml_metallib_end" >> ${METALLIB_EMBED_ASM} - COMMAND echo "_ggml_metallib_end:" >> ${METALLIB_EMBED_ASM} - DEPENDS ggml-metal.metal ggml-common.h - COMMENT "Generate assembly for embedded Metal library" - ) - - list(APPEND GGML_SOURCES_METAL ${METALLIB_EMBED_ASM}) - else() - if (GGML_METAL_SHADER_DEBUG) - # custom command to do the following: - # xcrun -sdk macosx metal -fno-fast-math -c ggml-metal.metal -o ggml-metal.air - # xcrun -sdk macosx metallib ggml-metal.air -o default.metallib - # - # note: this is the only way I found to disable fast-math in Metal. it's ugly, but at least it works - # disabling fast math is needed in order to pass tests/test-backend-ops - # note: adding -fno-inline fixes the tests when using MTL_SHADER_VALIDATION=1 - # note: unfortunately, we have to call it default.metallib instead of ggml.metallib - # ref: https://github.com/ggerganov/whisper.cpp/issues/1720 - set(XC_FLAGS -fno-fast-math -fno-inline -g) - else() - set(XC_FLAGS -O3) - endif() - - # Append macOS metal versioning flags - if (GGML_METAL_MACOSX_VERSION_MIN) - message(STATUS "Adding -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN} flag to metal compilation") - list (APPEND XC_FLAGS -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN}) - endif() - - if (GGML_METAL_STD) - message(STATUS "Adding -std=${GGML_METAL_STD} flag to metal compilation") - list (APPEND XC_FLAGS -std=${GGML_METAL_STD}) - endif() - - add_custom_command( - OUTPUT ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib - COMMAND xcrun -sdk macosx metal ${XC_FLAGS} -c ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal -o ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air - COMMAND xcrun -sdk macosx metallib ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air -o ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib - COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air - COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h - COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal - DEPENDS ggml-metal.metal ggml-common.h - COMMENT "Compiling Metal kernels" - ) - - add_custom_target( - ggml-metal ALL - DEPENDS ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib - ) - endif() # GGML_METAL_EMBED_LIBRARY - - list(APPEND GGML_EXTRA_LIBS_PRIVATE - ${FOUNDATION_LIBRARY} - ${METAL_FRAMEWORK} - ${METALKIT_FRAMEWORK} - ) -endif() - -if (GGML_MUSA) - set(CMAKE_C_COMPILER clang) - set(CMAKE_C_EXTENSIONS OFF) - set(CMAKE_CXX_COMPILER clang++) - set(CMAKE_CXX_EXTENSIONS OFF) - - set(GGML_CUDA ON) - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_MUSA) -endif() - -if (GGML_OPENMP) - find_package(OpenMP) - if (OpenMP_FOUND) - message(STATUS "OpenMP found") - - add_compile_definitions(GGML_USE_OPENMP) - - list(APPEND GGML_EXTRA_LIBS_PRIVATE OpenMP::OpenMP_C OpenMP::OpenMP_CXX) - - if (GGML_MUSA) - list(APPEND GGML_EXTRA_INCLUDES "/usr/lib/llvm-14/lib/clang/14.0.0/include") - list(APPEND GGML_EXTRA_LIBS_PRIVATE "/usr/lib/llvm-14/lib/libomp.so") - endif() - else() - message(WARNING "OpenMP not found") - endif() -endif() - -if (GGML_BLAS) - if (GGML_STATIC) - set(BLA_STATIC ON) - endif() - #if (CMAKE_VERSION VERSION_GREATER_EQUAL 3.22) - # set(BLA_SIZEOF_INTEGER 8) - #endif() - - set(BLA_VENDOR ${GGML_BLAS_VENDOR}) - find_package(BLAS) - - if (BLAS_FOUND) - message(STATUS "BLAS found, Libraries: ${BLAS_LIBRARIES}") - - if (("${BLAS_INCLUDE_DIRS}" STREQUAL "") AND NOT (${GGML_BLAS_VENDOR} MATCHES "Apple")) - # BLAS_INCLUDE_DIRS is missing in FindBLAS.cmake. - # see https://gitlab.kitware.com/cmake/cmake/-/issues/20268 - find_package(PkgConfig REQUIRED) - if (${GGML_BLAS_VENDOR} MATCHES "Generic") - pkg_check_modules(DepBLAS blas) - elseif (${GGML_BLAS_VENDOR} MATCHES "OpenBLAS") - # As of openblas v0.3.22, the 64-bit is named openblas64.pc - pkg_check_modules(DepBLAS openblas64) - if (NOT DepBLAS_FOUND) - pkg_check_modules(DepBLAS openblas) - endif() - elseif (${GGML_BLAS_VENDOR} MATCHES "FLAME") - add_compile_definitions(GGML_BLAS_USE_BLIS) - pkg_check_modules(DepBLAS blis) - elseif (${GGML_BLAS_VENDOR} MATCHES "ATLAS") - pkg_check_modules(DepBLAS blas-atlas) - elseif (${GGML_BLAS_VENDOR} MATCHES "FlexiBLAS") - pkg_check_modules(DepBLAS flexiblas_api) - elseif (${GGML_BLAS_VENDOR} MATCHES "Intel") - add_compile_definitions(GGML_BLAS_USE_MKL) - # all Intel* libraries share the same include path - pkg_check_modules(DepBLAS mkl-sdl) - elseif (${GGML_BLAS_VENDOR} MATCHES "NVHPC") - # this doesn't provide pkg-config - # suggest to assign BLAS_INCLUDE_DIRS on your own - if ("${NVHPC_VERSION}" STREQUAL "") - message(WARNING "Better to set NVHPC_VERSION") - else() - set(DepBLAS_FOUND ON) - set(DepBLAS_INCLUDE_DIRS "/opt/nvidia/hpc_sdk/${CMAKE_SYSTEM_NAME}_${CMAKE_SYSTEM_PROCESSOR}/${NVHPC_VERSION}/math_libs/include") - endif() - endif() - if (DepBLAS_FOUND) - set(BLAS_INCLUDE_DIRS ${DepBLAS_INCLUDE_DIRS}) - else() - message(WARNING "BLAS_INCLUDE_DIRS neither been provided nor been automatically" - " detected by pkgconfig, trying to find cblas.h from possible paths...") - find_path(BLAS_INCLUDE_DIRS - NAMES cblas.h - HINTS - /usr/include - /usr/local/include - /usr/include/openblas - /opt/homebrew/opt/openblas/include - /usr/local/opt/openblas/include - /usr/include/x86_64-linux-gnu/openblas/include - ) - endif() - endif() - - message(STATUS "BLAS found, Includes: ${BLAS_INCLUDE_DIRS}") - - add_compile_options(${BLAS_LINKER_FLAGS}) - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_BLAS) - - if (${BLAS_INCLUDE_DIRS} MATCHES "mkl" AND (${GGML_BLAS_VENDOR} MATCHES "Generic" OR ${GGML_BLAS_VENDOR} MATCHES "Intel")) - add_compile_definitions(GGML_BLAS_USE_MKL) - endif() - - set(GGML_HEADERS_BLAS ../include/ggml-blas.h) - set(GGML_SOURCES_BLAS ggml-blas.cpp) - - list(APPEND GGML_EXTRA_LIBS_PRIVATE ${BLAS_LIBRARIES}) - list(APPEND GGML_EXTRA_INCLUDES ${BLAS_INCLUDE_DIRS}) - else() - message(WARNING "BLAS not found, please refer to " - "https://cmake.org/cmake/help/latest/module/FindBLAS.html#blas-lapack-vendors" - " to set correct GGML_BLAS_VENDOR") - endif() -endif() - -if (GGML_LLAMAFILE) - message(STATUS "Using llamafile") - - add_compile_definitions(GGML_USE_LLAMAFILE) - - set(GGML_HEADERS_LLAMAFILE llamafile/sgemm.h) - set(GGML_SOURCES_LLAMAFILE llamafile/sgemm.cpp) -endif() - -if (GGML_AMX) - if (CMAKE_COMPILER_IS_GNUCC AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER 11.0) - else() - set(GGML_AMX OFF) - message(WARNING "AMX requires gcc version > 11.0. Turning off GGML_AMX.") - endif() - - if (GGML_AMX) - message(STATUS "Using AMX") - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_AMX) - - file(GLOB GGML_HEADERS_AMX "ggml-amx/*.h") - list(APPEND GGML_HEADERS_AMX "../include/ggml-amx.h") - - file(GLOB GGML_SOURCES_AMX "ggml-amx/*.cpp") - list(APPEND GGML_SOURCES_AMX "ggml-amx.cpp") - endif() -endif() - -if (GGML_CUDA) - cmake_minimum_required(VERSION 3.18) # for CMAKE_CUDA_ARCHITECTURES - - if (GGML_MUSA) - list(APPEND CMAKE_MODULE_PATH "/usr/local/musa/cmake/") - find_package(MUSAToolkit) - set(CUDAToolkit_FOUND ${MUSAToolkit_FOUND}) - else() - find_package(CUDAToolkit) - endif() - - if (CUDAToolkit_FOUND) - message(STATUS "CUDA found") - - if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES) - # 52 == lowest CUDA 12 standard - # 60 == FP16 CUDA intrinsics - # 61 == integer CUDA intrinsics - # 70 == compute capability at which unrolling a loop in mul_mat_q kernels is faster - if (GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16) - set(CMAKE_CUDA_ARCHITECTURES "60;61;70;75") - else() - set(CMAKE_CUDA_ARCHITECTURES "52;61;70;75") - #set(CMAKE_CUDA_ARCHITECTURES "OFF") # use this to compile much faster, but only F16 models work - endif() - endif() - message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}") - - if (GGML_MUSA) - set(CMAKE_CUDA_COMPILER ${MUSAToolkit_MCC_EXECUTABLE}) - else() - enable_language(CUDA) - endif() - - file(GLOB GGML_HEADERS_CUDA "ggml-cuda/*.cuh") - list(APPEND GGML_HEADERS_CUDA "../include/ggml-cuda.h") - - file(GLOB GGML_SOURCES_CUDA "ggml-cuda/*.cu") - list(APPEND GGML_SOURCES_CUDA "ggml-cuda.cu") - file(GLOB SRCS "ggml-cuda/template-instances/fattn-wmma*.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/mmq*.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - - if (GGML_CUDA_FA_ALL_QUANTS) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) - else() - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*f16-f16.cu") - list(APPEND GGML_SOURCES_CUDA ${SRCS}) - endif() - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_CUDA) - - add_compile_definitions(GGML_CUDA_DMMV_X=${GGML_CUDA_DMMV_X}) - add_compile_definitions(GGML_CUDA_MMV_Y=${GGML_CUDA_MMV_Y}) - add_compile_definitions(K_QUANTS_PER_ITERATION=${GGML_CUDA_KQUANTS_ITER}) - add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${GGML_CUDA_PEER_MAX_BATCH_SIZE}) - - if (GGML_CUDA_GRAPHS) - add_compile_definitions(GGML_CUDA_USE_GRAPHS) - endif() - - if (GGML_CUDA_FORCE_DMMV) - add_compile_definitions(GGML_CUDA_FORCE_DMMV) - endif() - - if (GGML_CUDA_FORCE_MMQ) - add_compile_definitions(GGML_CUDA_FORCE_MMQ) - endif() - - if (GGML_CUDA_FORCE_CUBLAS) - add_compile_definitions(GGML_CUDA_FORCE_CUBLAS) - endif() - - if (GGML_CUDA_NO_VMM) - add_compile_definitions(GGML_CUDA_NO_VMM) - endif() - - if (DEFINED GGML_CUDA_DMMV_Y) - add_compile_definitions(GGML_CUDA_MMV_Y=${GGML_CUDA_DMMV_Y}) # for backwards compatibility - endif() - - if (GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16) - add_compile_definitions(GGML_CUDA_F16) - endif() - - if (GGML_CUDA_NO_PEER_COPY) - add_compile_definitions(GGML_CUDA_NO_PEER_COPY) - endif() - - if (GGML_MUSA) - set_source_files_properties(${GGML_SOURCES_CUDA} PROPERTIES LANGUAGE CXX) - foreach(SOURCE ${GGML_SOURCES_CUDA}) - set_property(SOURCE ${SOURCE} PROPERTY COMPILE_FLAGS "-x musa -mtgpu --cuda-gpu-arch=mp_21 --cuda-gpu-arch=mp_22") - endforeach() - endif() - - if (GGML_STATIC) - if (WIN32) - # As of 12.3.1 CUDA Toolkit for Windows does not offer a static cublas library - list(APPEND GGML_EXTRA_LIBS_PRIVATE CUDA::cudart_static CUDA::cublas CUDA::cublasLt) - else () - if (GGML_MUSA) - list(APPEND GGML_EXTRA_LIBS_PRIVATE MUSA::musart_static MUSA::mublas_static) - else() - list(APPEND GGML_EXTRA_LIBS_PRIVATE CUDA::cudart_static CUDA::cublas_static CUDA::cublasLt_static) - endif() - endif() - else() - if (GGML_MUSA) - list(APPEND GGML_EXTRA_LIBS_PRIVATE MUSA::musart MUSA::mublas) - else() - list(APPEND GGML_EXTRA_LIBS_PRIVATE CUDA::cudart CUDA::cublas CUDA::cublasLt) - endif() - endif() - - if (GGML_CUDA_NO_VMM) - # No VMM requested, no need to link directly with the cuda driver lib (libcuda.so) - else() - if (GGML_MUSA) - list(APPEND GGML_EXTRA_LIBS_PRIVATE MUSA::musa_driver) # required by muDeviceGetAttribute(), muMemGetAllocationGranularity(...), ... - else() - list(APPEND GGML_EXTRA_LIBS_PRIVATE CUDA::cuda_driver) # required by cuDeviceGetAttribute(), cuMemGetAllocationGranularity(...), ... - endif() - endif() - else() - message(WARNING "CUDA not found") - endif() -endif() - -if (GGML_HIPBLAS) - if (NOT EXISTS $ENV{ROCM_PATH}) - if (NOT EXISTS /opt/rocm) - set(ROCM_PATH /usr) - else() - set(ROCM_PATH /opt/rocm) - endif() - else() - set(ROCM_PATH $ENV{ROCM_PATH}) - endif() - - list(APPEND CMAKE_PREFIX_PATH ${ROCM_PATH}) - list(APPEND CMAKE_PREFIX_PATH "${ROCM_PATH}/lib64/cmake") - - # CMake on Windows doesn't support the HIP language yet - if (WIN32) - set(CXX_IS_HIPCC TRUE) - else() - string(REGEX MATCH "hipcc(\.bat)?$" CXX_IS_HIPCC "${CMAKE_CXX_COMPILER}") - endif() - - if (CXX_IS_HIPCC) - if (LINUX) - if (NOT ${CMAKE_CXX_COMPILER_ID} MATCHES "Clang") - message(WARNING "Only LLVM is supported for HIP, hint: CXX=/opt/rocm/llvm/bin/clang++") - endif() - - message(WARNING "Setting hipcc as the C++ compiler is legacy behavior." - " Prefer setting the HIP compiler directly. See README for details.") - endif() - else() - # Forward AMDGPU_TARGETS to CMAKE_HIP_ARCHITECTURES. - if (AMDGPU_TARGETS AND NOT CMAKE_HIP_ARCHITECTURES) - set(CMAKE_HIP_ARCHITECTURES ${AMDGPU_TARGETS}) - endif() - cmake_minimum_required(VERSION 3.21) - enable_language(HIP) - endif() - - find_package(hip REQUIRED) - find_package(hipblas REQUIRED) - find_package(rocblas REQUIRED) - - message(STATUS "HIP and hipBLAS found") - - file(GLOB GGML_HEADERS_ROCM "ggml-cuda/*.cuh") - list(APPEND GGML_HEADERS_ROCM "../include/ggml-cuda.h") - - file(GLOB GGML_SOURCES_ROCM "ggml-cuda/*.cu") - list(APPEND GGML_SOURCES_ROCM "ggml-cuda.cu") - file(GLOB SRCS "ggml-cuda/template-instances/fattn-wmma*.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/mmq*.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - - if (GGML_CUDA_FA_ALL_QUANTS) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) - else() - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*f16-f16.cu") - list(APPEND GGML_SOURCES_ROCM ${SRCS}) - endif() - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_CUDA) - - add_compile_definitions(GGML_USE_HIPBLAS) - add_compile_definitions(GGML_CUDA_DMMV_X=${GGML_CUDA_DMMV_X}) - add_compile_definitions(GGML_CUDA_MMV_Y=${GGML_CUDA_MMV_Y}) - add_compile_definitions(K_QUANTS_PER_ITERATION=${GGML_CUDA_KQUANTS_ITER}) - - if (GGML_HIP_UMA) - add_compile_definitions(GGML_HIP_UMA) - endif() - - if (GGML_CUDA_FORCE_DMMV) - add_compile_definitions(GGML_CUDA_FORCE_DMMV) - endif() - - if (GGML_CUDA_FORCE_MMQ) - add_compile_definitions(GGML_CUDA_FORCE_MMQ) - endif() - - if (GGML_CUDA_FORCE_CUBLAS) - add_compile_definitions(GGML_CUDA_FORCE_CUBLAS) - endif() - - if (GGML_CUDA_NO_PEER_COPY) - add_compile_definitions(GGML_CUDA_NO_PEER_COPY) - endif() - - if (CXX_IS_HIPCC) - set_source_files_properties(${GGML_SOURCES_ROCM} PROPERTIES LANGUAGE CXX) - list(APPEND GGML_EXTRA_LIBS_PRIVATE hip::device) - else() - set_source_files_properties(${GGML_SOURCES_ROCM} PROPERTIES LANGUAGE HIP) - endif() - - if (GGML_STATIC) - message(FATAL_ERROR "Static linking not supported for HIP/ROCm") - endif() - - list(APPEND GGML_EXTRA_LIBS_PUBLIC hip::host roc::rocblas roc::hipblas) -endif() - -if (GGML_SYCL) - if (NOT GGML_SYCL_TARGET MATCHES "^(INTEL|NVIDIA|AMD)$") - message(FATAL_ERROR "Invalid backend chosen, supported options are INTEL, NVIDIA, or AMD") - endif() - - check_cxx_compiler_flag("-fsycl" SUPPORTS_SYCL) - - if (DEFINED ENV{ONEAPI_ROOT}) - message(STATUS "Using oneAPI Release SYCL compiler (icpx).") - elseif(SUPPORTS_SYCL) - message(WARNING "Using open-source SYCL compiler (clang++). Didn't detect ENV {ONEAPI_ROOT}. - If you expected the oneAPI Release compiler, please install oneAPI & source it, like: - source /opt/intel/oneapi/setvars.sh") - else() - message(FATAL_ERROR, "C++ compiler lacks SYCL support.") - endif() - message(STATUS "SYCL found") - #todo: AOT - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_SYCL) - - if (GGML_SYCL_F16) - if (GGML_SYCL_TARGET STREQUAL "AMD") - message(WARNING "AMD target does not entirely support FP16 in the SYCL backend.") - endif() - add_compile_definitions(GGML_SYCL_F16) - endif() - - if (GGML_CUDA_FORCE_MMQ) - add_compile_definitions(GGML_SYCL_FORCE_MMQ) - endif() - - set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-narrowing -fsycl") - - if (GGML_SYCL_TARGET STREQUAL "NVIDIA") - add_compile_definitions(GGML_SYCL_WARP_SIZE=32) - elseif (GGML_SYCL_TARGET STREQUAL "AMD") - # INFO: Allowed Sub_group_sizes are not consistent through all - # hip targets. For example, 64 is used for certain models, but the backend - # does not support it. - # Target archs tested working: gfx1030, gfx1031, (Only tested sub_group_size = 32) - add_compile_definitions(GGML_SYCL_WARP_SIZE=32) - else() - add_compile_definitions(GGML_SYCL_WARP_SIZE=16) - endif() - - file(GLOB GGML_HEADERS_SYCL "ggml-sycl/*.hpp") - list(APPEND GGML_HEADERS_SYCL "../include/ggml-sycl.h") - - file(GLOB GGML_SOURCES_SYCL "ggml-sycl/*.cpp") - list(APPEND GGML_SOURCES_SYCL "ggml-sycl.cpp") - - find_package(DNNL) - message("-- DNNL found:" ${DNNL_FOUND}) - - if (GGML_SYCL_TARGET STREQUAL "INTEL") - add_compile_definitions(GGML_SYCL_DNNL=${DNNL_FOUND}) - else() - add_compile_definitions(GGML_SYCL_DNNL=0) - endif() - - if (${DNNL_FOUND} AND GGML_SYCL_TARGET STREQUAL "INTEL") - list(APPEND GGML_EXTRA_LIBS_PRIVATE DNNL::dnnl) - endif() - - if (WIN32) - find_package(IntelSYCL REQUIRED) - find_package(MKL REQUIRED) - list(APPEND GGML_EXTRA_LIBS_PRIVATE IntelSYCL::SYCL_CXX MKL::MKL MKL::MKL_SYCL) - else() - if (GGML_SYCL_TARGET STREQUAL "INTEL") - list(APPEND GGML_EXTRA_LIBS_PRIVATE sycl OpenCL mkl_core pthread m dl mkl_sycl_blas mkl_intel_ilp64 mkl_tbb_thread) - elseif (GGML_SYCL_TARGET STREQUAL "NVIDIA") - set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsycl-targets=nvptx64-nvidia-cuda") - list(APPEND GGML_EXTRA_LIBS_PRIVATE sycl pthread m dl onemkl) - elseif (GGML_SYCL_TARGET STREQUAL "AMD") - if (GGML_SYCL_HIP_TARGET STREQUAL "") - message(ERROR "Can't enable SYCL hip backend, GGML_SYCL_HIP_TARGET has not been set.") - endif() - set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsycl-targets=amdgcn-amd-amdhsa -Xsycl-target-backend --offload-arch=${GGML_SYCL_HIP_TARGET}") - list(APPEND GGML_EXTRA_LIBS_PRIVATE sycl pthread m dl onemkl) - endif() - endif() -endif() - -if (GGML_RPC) - message(STATUS "RPC found") - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_RPC) - - if (WIN32) - list(APPEND GGML_EXTRA_LIBS_PRIVATE ws2_32) - endif() - - set(GGML_HEADERS_RPC ../include/ggml-rpc.h) - set(GGML_SOURCES_RPC ggml-rpc.cpp) -endif() - -if (GGML_VULKAN) - find_package(Vulkan COMPONENTS glslc REQUIRED) - - if (Vulkan_FOUND) - message(STATUS "Vulkan found") - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_VULKAN) - - # Workaround to the "can't dereference invalidated vector iterator" bug in clang-cl debug build - # Posssibly relevant: https://stackoverflow.com/questions/74748276/visual-studio-no-displays-the-correct-length-of-stdvector - if (MSVC AND CMAKE_CXX_COMPILER_ID STREQUAL "Clang") - add_compile_definitions(_ITERATOR_DEBUG_LEVEL=0) - endif() - - if (GGML_VULKAN_CHECK_RESULTS) - add_compile_definitions(GGML_VULKAN_CHECK_RESULTS) - endif() - - if (GGML_VULKAN_DEBUG) - add_compile_definitions(GGML_VULKAN_DEBUG) - endif() - - if (GGML_VULKAN_MEMORY_DEBUG) - add_compile_definitions(GGML_VULKAN_MEMORY_DEBUG) - endif() - - if (GGML_VULKAN_SHADER_DEBUG_INFO) - add_compile_definitions(GGML_VULKAN_SHADER_DEBUG_INFO) - endif() - - if (GGML_VULKAN_PERF) - add_compile_definitions(GGML_VULKAN_PERF) - endif() - - if (GGML_VULKAN_VALIDATE) - add_compile_definitions(GGML_VULKAN_VALIDATE) - endif() - - if (GGML_VULKAN_RUN_TESTS) - add_compile_definitions(GGML_VULKAN_RUN_TESTS) - endif() - - add_subdirectory(vulkan-shaders) - - set (_ggml_vk_genshaders_cmd vulkan-shaders-gen) - set (_ggml_vk_header ${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.hpp) - set (_ggml_vk_source ${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.cpp) - set (_ggml_vk_input_dir ${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders) - set (_ggml_vk_output_dir ${CMAKE_CURRENT_BINARY_DIR}/vulkan-shaders.spv) - - file(GLOB _ggml_vk_shader_deps "${_ggml_vk_input_dir}/*.comp") - - add_custom_command( - OUTPUT ${_ggml_vk_header} - ${_ggml_vk_source} - - COMMAND ${_ggml_vk_genshaders_cmd} - --glslc ${Vulkan_GLSLC_EXECUTABLE} - --input-dir ${_ggml_vk_input_dir} - --output-dir ${_ggml_vk_output_dir} - --target-hpp ${_ggml_vk_header} - --target-cpp ${_ggml_vk_source} - --no-clean - - DEPENDS ${_ggml_vk_shader_deps} - COMMENT "Generate vulkan shaders" - ) - - set(GGML_HEADERS_VULKAN ${CMAKE_CURRENT_SOURCE_DIR}/../include/ggml-vulkan.h ${_ggml_vk_header}) - set(GGML_SOURCES_VULKAN ggml-vulkan.cpp ${_ggml_vk_source}) - - list(APPEND GGML_EXTRA_LIBS_PRIVATE Vulkan::Vulkan) - list(APPEND GGML_EXTRA_INCLUDES ${CMAKE_CURRENT_BINARY_DIR}) - else() - message(WARNING "Vulkan not found") - endif() -endif() - -if (GGML_KOMPUTE) - add_compile_definitions(VULKAN_HPP_DISPATCH_LOADER_DYNAMIC=1) - - find_package(Vulkan COMPONENTS glslc REQUIRED) - find_program(glslc_executable NAMES glslc HINTS Vulkan::glslc) - - if (NOT glslc_executable) - message(FATAL_ERROR "glslc not found") - endif() - - function(compile_shader) - set(options) - set(oneValueArgs) - set(multiValueArgs SOURCES) - cmake_parse_arguments(compile_shader "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN}) - foreach(source ${compile_shader_SOURCES}) - get_filename_component(filename ${source} NAME) - set(spv_file ${filename}.spv) - add_custom_command( - OUTPUT ${spv_file} - DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/${source} - ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/common.comp - ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_getrows.comp - ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_mul_mv_q_n_pre.comp - ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_mul_mv_q_n.comp - COMMAND ${glslc_executable} --target-env=vulkan1.2 -o ${spv_file} ${CMAKE_CURRENT_SOURCE_DIR}/${source} - COMMENT "Compiling ${source} to ${spv_file}" - ) - - get_filename_component(RAW_FILE_NAME ${spv_file} NAME) - set(FILE_NAME "shader${RAW_FILE_NAME}") - string(REPLACE ".comp.spv" ".h" HEADER_FILE ${FILE_NAME}) - string(TOUPPER ${HEADER_FILE} HEADER_FILE_DEFINE) - string(REPLACE "." "_" HEADER_FILE_DEFINE "${HEADER_FILE_DEFINE}") - set(OUTPUT_HEADER_FILE "${HEADER_FILE}") - message(STATUS "${HEADER_FILE} generating ${HEADER_FILE_DEFINE}") - if(CMAKE_GENERATOR MATCHES "Visual Studio") - add_custom_command( - OUTPUT ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "/*THIS FILE HAS BEEN AUTOMATICALLY GENERATED - DO NOT EDIT*/" > ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#ifndef ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "namespace kp {" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "namespace shader_data {" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_BINARY_DIR}/bin/$/xxd -i ${RAW_FILE_NAME} >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "}}" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#endif // define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - DEPENDS ${spv_file} xxd - COMMENT "Converting to hpp: ${FILE_NAME} ${CMAKE_BINARY_DIR}/bin/$/xxd" - ) - else() - add_custom_command( - OUTPUT ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "/*THIS FILE HAS BEEN AUTOMATICALLY GENERATED - DO NOT EDIT*/" > ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#ifndef ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "namespace kp {" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "namespace shader_data {" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_BINARY_DIR}/bin/xxd -i ${RAW_FILE_NAME} >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo "}}" >> ${OUTPUT_HEADER_FILE} - COMMAND ${CMAKE_COMMAND} -E echo \"\#endif // define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} - DEPENDS ${spv_file} xxd - COMMENT "Converting to hpp: ${FILE_NAME} ${CMAKE_BINARY_DIR}/bin/xxd" - ) - endif() - endforeach() - endfunction() - - if (EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/kompute/CMakeLists.txt") - message(STATUS "Kompute found") - set(KOMPUTE_OPT_LOG_LEVEL Error CACHE STRING "Kompute log level") - add_subdirectory(kompute) - - # Compile our shaders - compile_shader(SOURCES - kompute-shaders/op_scale.comp - kompute-shaders/op_scale_8.comp - kompute-shaders/op_add.comp - kompute-shaders/op_addrow.comp - kompute-shaders/op_mul.comp - kompute-shaders/op_silu.comp - kompute-shaders/op_relu.comp - kompute-shaders/op_gelu.comp - kompute-shaders/op_softmax.comp - kompute-shaders/op_norm.comp - kompute-shaders/op_rmsnorm.comp - kompute-shaders/op_diagmask.comp - kompute-shaders/op_mul_mat_mat_f32.comp - kompute-shaders/op_mul_mat_f16.comp - kompute-shaders/op_mul_mat_q8_0.comp - kompute-shaders/op_mul_mat_q4_0.comp - kompute-shaders/op_mul_mat_q4_1.comp - kompute-shaders/op_mul_mat_q4_k.comp - kompute-shaders/op_mul_mat_q6_k.comp - kompute-shaders/op_getrows_f32.comp - kompute-shaders/op_getrows_f16.comp - kompute-shaders/op_getrows_q4_0.comp - kompute-shaders/op_getrows_q4_1.comp - kompute-shaders/op_getrows_q6_k.comp - kompute-shaders/op_rope_f16.comp - kompute-shaders/op_rope_f32.comp - kompute-shaders/op_cpy_f16_f16.comp - kompute-shaders/op_cpy_f16_f32.comp - kompute-shaders/op_cpy_f32_f16.comp - kompute-shaders/op_cpy_f32_f32.comp - ) - - # Create a custom target for our generated shaders - add_custom_target(generated_shaders DEPENDS - shaderop_scale.h - shaderop_scale_8.h - shaderop_add.h - shaderop_addrow.h - shaderop_mul.h - shaderop_silu.h - shaderop_relu.h - shaderop_gelu.h - shaderop_softmax.h - shaderop_norm.h - shaderop_rmsnorm.h - shaderop_diagmask.h - shaderop_mul_mat_mat_f32.h - shaderop_mul_mat_f16.h - shaderop_mul_mat_q8_0.h - shaderop_mul_mat_q4_0.h - shaderop_mul_mat_q4_1.h - shaderop_mul_mat_q4_k.h - shaderop_mul_mat_q6_k.h - shaderop_getrows_f32.h - shaderop_getrows_f16.h - shaderop_getrows_q4_0.h - shaderop_getrows_q4_1.h - shaderop_getrows_q6_k.h - shaderop_rope_f16.h - shaderop_rope_f32.h - shaderop_cpy_f16_f16.h - shaderop_cpy_f16_f32.h - shaderop_cpy_f32_f16.h - shaderop_cpy_f32_f32.h - ) - - # Create a custom command that depends on the generated_shaders - add_custom_command( - OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp - COMMAND ${CMAKE_COMMAND} -E touch ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp - DEPENDS generated_shaders - COMMENT "Ensuring shaders are generated before compiling ggml-kompute.cpp" - ) - - # Add the stamp to the main sources to ensure dependency tracking - set(GGML_SOURCES_KOMPUTE ggml-kompute.cpp ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp) - set(GGML_HEADERS_KOMPUTE ../include/ggml-kompute.h ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp) - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_KOMPUTE) - - list(APPEND GGML_EXTRA_LIBS_PRIVATE kompute) - list(APPEND GGML_EXTRA_INCLUDES ${CMAKE_CURRENT_BINARY_DIR}) - else() - message(WARNING "Kompute not found") - endif() -endif() - -if (GGML_CPU_HBM) - find_library(memkind memkind REQUIRED) - - message(STATUS "Using memkind for CPU HBM") - - add_compile_definitions(GGML_USE_CPU_HBM) - - target_link_libraries(ggml PUBLIC memkind) -endif() - -if (GGML_CANN) - if ("cann${CANN_INSTALL_DIR}" STREQUAL "cann" AND DEFINED ENV{ASCEND_TOOLKIT_HOME}) - set(CANN_INSTALL_DIR $ENV{ASCEND_TOOLKIT_HOME}) - message(STATUS "CANN: updated CANN_INSTALL_DIR from ASCEND_TOOLKIT_HOME=$ENV{ASCEND_TOOLKIT_HOME}") - endif() - - if (CANN_INSTALL_DIR) - # Only Support Linux. - if (GGML_CANN) - if (NOT UNIX) - set(GGML_CANN OFF) - message(WARNING "CANN: CANN toolkit supports unix but not ${CMAKE_SYSTEM_NAME}. Turning off GGML_CANN") - endif() - endif() - - # Supported platforms: x86-64, arm64 - if (GGML_CANN) - if (CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64") - elseif (CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "amd64") - else() - set(GGML_CANN OFF) - message(WARNING "CANN: CANN toolkit supports x86-64 and arm64 but not ${CMAKE_SYSTEM_PROCESSOR}. Turning off GGML_CANN") - endif() - endif() - - # Set header and libs - if(GGML_CANN) - set(CANN_INCLUDE_DIRS - ${CANN_INSTALL_DIR}/include - ${CANN_INSTALL_DIR}/include/aclnn - ${CANN_INSTALL_DIR}/acllib/include - ) - - add_subdirectory(ggml-cann/kernels) - list(APPEND CANN_LIBRARIES - ascendcl - nnopbase - opapi - acl_op_compiler - ascendc_kernels - ) - - set(GGML_HEADERS_CANN "../include/ggml-cann.h") - file(GLOB GGML_SOURCES_CANN "ggml-cann/*.cpp") - list(APPEND GGML_SOURCES_CANN "ggml-cann.cpp") - - message(STATUS "CANN: CANN_INCLUDE_DIRS = ${CANN_INCLUDE_DIRS}") - message(STATUS "CANN: CANN_LIBRARIES = ${CANN_LIBRARIES}") - - list(APPEND GGML_EXTRA_LIBS_PRIVATE ${CANN_LIBRARIES} ) - list(APPEND GGML_EXTRA_INCLUDES ${CANN_INCLUDE_DIRS}) - list(APPEND GGML_EXTRA_LIBDIRS ${CANN_INSTALL_DIR}/lib64) - - list(APPEND GGML_CDEF_PUBLIC GGML_USE_CANN) - endif() - else() - set(GGML_CANN OFF) - message(WARNING "CANN: Can't find CANN_INSTALL_DIR, do you forget to source set_var.sh. Turning off GGML_CANN") - endif() - - if(NOT GGML_CANN) - message(WARNING "CANN: GGML_CANN is turned OFF, see above for details.") - endif() -endif() - -function(get_flags CCID CCVER) +function(ggml_get_flags CCID CCVER) set(C_FLAGS "") set(CXX_FLAGS "") @@ -963,11 +42,6 @@ function(get_flags CCID CCVER) set(C_FLAGS -Wdouble-promotion) set(CXX_FLAGS -Wno-array-bounds) - if (NOT GGML_MUSA) - if (CCVER VERSION_GREATER_EQUAL 7.1.0) - list(APPEND CXX_FLAGS -Wno-format-truncation) - endif() - endif() if (CCVER VERSION_GREATER_EQUAL 8.1.0) list(APPEND CXX_FLAGS -Wextra-semi) endif() @@ -996,7 +70,7 @@ if (GGML_ALL_WARNINGS) list(APPEND C_FLAGS ${WARNING_FLAGS}) list(APPEND CXX_FLAGS ${WARNING_FLAGS}) - get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}) + ggml_get_flags(${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}) add_compile_options("$<$:${C_FLAGS};${GF_C_FLAGS}>" "$<$:${CXX_FLAGS};${GF_CXX_FLAGS}>") @@ -1007,54 +81,6 @@ if (GGML_ALL_WARNINGS) endif() endif() -set(CUDA_CXX_FLAGS "") - -if (GGML_CUDA) - set(CUDA_FLAGS -use_fast_math) - - if (GGML_FATAL_WARNINGS) - list(APPEND CUDA_FLAGS -Werror all-warnings) - endif() - - if (GGML_ALL_WARNINGS AND NOT MSVC) - set(NVCC_CMD ${CMAKE_CUDA_COMPILER} .c) - if (NOT CMAKE_CUDA_HOST_COMPILER STREQUAL "") - list(APPEND NVCC_CMD -ccbin ${CMAKE_CUDA_HOST_COMPILER}) - endif() - - execute_process( - COMMAND ${NVCC_CMD} -Xcompiler --version - OUTPUT_VARIABLE CUDA_CCFULLVER - ERROR_QUIET - ) - - if (NOT CUDA_CCFULLVER MATCHES clang) - set(CUDA_CCID "GNU") - execute_process( - COMMAND ${NVCC_CMD} -Xcompiler "-dumpfullversion -dumpversion" - OUTPUT_VARIABLE CUDA_CCVER - ERROR_QUIET - ) - else() - if (CUDA_CCFULLVER MATCHES Apple) - set(CUDA_CCID "AppleClang") - else() - set(CUDA_CCID "Clang") - endif() - string(REGEX REPLACE "^.* version ([0-9.]*).*$" "\\1" CUDA_CCVER ${CUDA_CCFULLVER}) - endif() - - message("-- CUDA host compiler is ${CUDA_CCID} ${CUDA_CCVER}") - - get_flags(${CUDA_CCID} ${CUDA_CCVER}) - list(APPEND CUDA_CXX_FLAGS ${CXX_FLAGS} ${GF_CXX_FLAGS}) # This is passed to -Xcompiler later - endif() - - if (NOT MSVC) - list(APPEND CUDA_CXX_FLAGS -Wno-pedantic) - endif() -endif() - if (GGML_LTO) include(CheckIPOSupported) check_ipo_supported(RESULT result OUTPUT output) @@ -1112,189 +138,6 @@ if (NOT MSVC) endif() endif() -set(ARCH_FLAGS "") - -if (CMAKE_OSX_ARCHITECTURES STREQUAL "arm64" OR - CMAKE_GENERATOR_PLATFORM_LWR STREQUAL "arm64" OR - (NOT CMAKE_OSX_ARCHITECTURES AND - NOT CMAKE_GENERATOR_PLATFORM_LWR AND - CMAKE_SYSTEM_PROCESSOR MATCHES "^(aarch64|arm.*|ARM64)$")) - - message(STATUS "ARM detected") - - if (MSVC) - add_compile_definitions(__aarch64__) # MSVC defines _M_ARM64 instead - add_compile_definitions(__ARM_NEON) - add_compile_definitions(__ARM_FEATURE_FMA) - - set(CMAKE_REQUIRED_FLAGS_PREV ${CMAKE_REQUIRED_FLAGS}) - string(JOIN " " CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS} "/arch:armv8.2") - - check_cxx_source_compiles("#include \nint main() { int8x16_t _a, _b; int32x4_t _s = vdotq_s32(_s, _a, _b); return 0; }" GGML_COMPILER_SUPPORT_DOTPROD) - if (GGML_COMPILER_SUPPORT_DOTPROD) - add_compile_definitions(__ARM_FEATURE_DOTPROD) - endif () - - check_cxx_source_compiles("#include \nint main() { int8x16_t _a, _b; int32x4_t _s = vmlaq_f32(_s, _a, _b); return 0; }" GGML_COMPILER_SUPPORT_MATMUL_INT8) - - if (GGML_COMPILER_SUPPORT_MATMUL_INT8) - add_compile_definitions(__ARM_FEATURE_MATMUL_INT8) - endif () - - check_cxx_source_compiles("#include \nint main() { float16_t _a; float16x8_t _s = vdupq_n_f16(_a); return 0; }" GGML_COMPILER_SUPPORT_FP16_VECTOR_ARITHMETIC) - if (GGML_COMPILER_SUPPORT_FP16_VECTOR_ARITHMETIC) - add_compile_definitions(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) - endif () - - set(CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS_PREV}) - else() - check_cxx_compiler_flag(-mfp16-format=ieee COMPILER_SUPPORTS_FP16_FORMAT_I3E) - if (NOT "${COMPILER_SUPPORTS_FP16_FORMAT_I3E}" STREQUAL "") - list(APPEND ARCH_FLAGS -mfp16-format=ieee) - endif() - if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv6") - # Raspberry Pi 1, Zero - list(APPEND ARCH_FLAGS -mfpu=neon-fp-armv8 -mno-unaligned-access) - endif() - if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv7") - if ("${CMAKE_SYSTEM_NAME}" STREQUAL "Android") - # Android armeabi-v7a - list(APPEND ARCH_FLAGS -mfpu=neon-vfpv4 -mno-unaligned-access -funsafe-math-optimizations) - else() - # Raspberry Pi 2 - list(APPEND ARCH_FLAGS -mfpu=neon-fp-armv8 -mno-unaligned-access -funsafe-math-optimizations) - endif() - endif() - if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv8") - # Android arm64-v8a - # Raspberry Pi 3, 4, Zero 2 (32-bit) - list(APPEND ARCH_FLAGS -mno-unaligned-access) - endif() - if (GGML_SVE) - list(APPEND ARCH_FLAGS -march=armv8.6-a+sve) - endif() - endif() -elseif (CMAKE_OSX_ARCHITECTURES STREQUAL "x86_64" OR CMAKE_GENERATOR_PLATFORM_LWR MATCHES "^(x86_64|i686|amd64|x64|win32)$" OR - (NOT CMAKE_OSX_ARCHITECTURES AND NOT CMAKE_GENERATOR_PLATFORM_LWR AND - CMAKE_SYSTEM_PROCESSOR MATCHES "^(x86_64|i686|AMD64)$")) - message(STATUS "x86 detected") - if (MSVC) - # instruction set detection for MSVC only - if (GGML_NATIVE) - # TODO: improve, should not reference files from the parent folder - include(../cmake/FindSIMD.cmake) - endif () - if (GGML_AVX512) - list(APPEND ARCH_FLAGS /arch:AVX512) - # MSVC has no compile-time flags enabling specific - # AVX512 extensions, neither it defines the - # macros corresponding to the extensions. - # Do it manually. - if (GGML_AVX512_VBMI) - add_compile_definitions($<$:__AVX512VBMI__>) - add_compile_definitions($<$:__AVX512VBMI__>) - endif() - if (GGML_AVX512_VNNI) - add_compile_definitions($<$:__AVX512VNNI__>) - add_compile_definitions($<$:__AVX512VNNI__>) - endif() - if (GGML_AVX512_BF16) - add_compile_definitions($<$:__AVX512BF16__>) - add_compile_definitions($<$:__AVX512BF16__>) - endif() - if (GGML_AMX_TILE) - add_compile_definitions($<$:__AMX_TILE__>) - add_compile_definitions($<$:__AMX_TILE__>) - endif() - if (GGML_AMX_INT8) - add_compile_definitions($<$:__AMX_INT8__>) - add_compile_definitions($<$:__AMX_INT8__>) - endif() - if (GGML_AMX_BF16) - add_compile_definitions($<$:__AMX_BF16__>) - add_compile_definitions($<$:__AMX_BF16__>) - endif() - elseif (GGML_AVX2) - list(APPEND ARCH_FLAGS /arch:AVX2) - elseif (GGML_AVX) - list(APPEND ARCH_FLAGS /arch:AVX) - endif() - else() - if (GGML_NATIVE) - list(APPEND ARCH_FLAGS -march=native) - endif() - if (GGML_F16C) - list(APPEND ARCH_FLAGS -mf16c) - endif() - if (GGML_FMA) - list(APPEND ARCH_FLAGS -mfma) - endif() - if (GGML_AVX) - list(APPEND ARCH_FLAGS -mavx) - endif() - if (GGML_AVX2) - list(APPEND ARCH_FLAGS -mavx2) - endif() - if (GGML_AVX512) - list(APPEND ARCH_FLAGS -mavx512f) - list(APPEND ARCH_FLAGS -mavx512dq) - list(APPEND ARCH_FLAGS -mavx512bw) - endif() - if (GGML_AVX512_VBMI) - list(APPEND ARCH_FLAGS -mavx512vbmi) - endif() - if (GGML_AVX512_VNNI) - list(APPEND ARCH_FLAGS -mavx512vnni) - endif() - if (GGML_AVX512_BF16) - list(APPEND ARCH_FLAGS -mavx512bf16) - endif() - if (GGML_AMX_TILE) - list(APPEND ARCH_FLAGS -mamx-tile) - endif() - if (GGML_AMX_INT8) - list(APPEND ARCH_FLAGS -mamx-int8) - endif() - if (GGML_AMX_BF16) - list(APPEND ARCH_FLAGS -mamx-bf16) - endif() - endif() -elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64") - message(STATUS "PowerPC detected") - if (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64le") - list(APPEND ARCH_FLAGS -mcpu=powerpc64le) - else() - list(APPEND ARCH_FLAGS -mcpu=native -mtune=native) - #TODO: Add targets for Power8/Power9 (Altivec/VSX) and Power10(MMA) and query for big endian systems (ppc64/le/be) - endif() -elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "loongarch64") - message(STATUS "loongarch64 detected") - - list(APPEND ARCH_FLAGS -march=loongarch64) - if (GGML_LASX) - list(APPEND ARCH_FLAGS -mlasx) - endif() - if (GGML_LSX) - list(APPEND ARCH_FLAGS -mlsx) - endif() -else() - message(STATUS "Unknown architecture") -endif() - -add_compile_options("$<$:${ARCH_FLAGS}>") -add_compile_options("$<$:${ARCH_FLAGS}>") - -if (GGML_CUDA) - list(APPEND CUDA_CXX_FLAGS ${ARCH_FLAGS}) - list(JOIN CUDA_CXX_FLAGS " " CUDA_CXX_FLAGS_JOINED) # pass host compiler flags as a single argument - - if (NOT CUDA_CXX_FLAGS_JOINED STREQUAL "") - list(APPEND CUDA_FLAGS -Xcompiler ${CUDA_CXX_FLAGS_JOINED}) - endif() - - add_compile_options("$<$:${CUDA_FLAGS}>") -endif() - if (MINGW) # Target Windows 8 for PrefetchVirtualMemory add_compile_definitions(_WIN32_WINNT=${GGML_WIN_VER}) @@ -1308,14 +151,14 @@ endif() # CLOCK_MONOTONIC came in POSIX.1-2001 / SUSv3 as optional # posix_memalign came in POSIX.1-2001 / SUSv3 # M_PI is an XSI extension since POSIX.1-2001 / SUSv3, came in XPG1 (1985) -add_compile_definitions(_XOPEN_SOURCE=600) # Somehow in OpenBSD whenever POSIX conformance is specified # some string functions rely on locale_t availability, # which was introduced in POSIX.1-2008, forcing us to go higher if (CMAKE_SYSTEM_NAME MATCHES "OpenBSD") - remove_definitions(-D_XOPEN_SOURCE=600) add_compile_definitions(_XOPEN_SOURCE=700) +else() + add_compile_definitions(_XOPEN_SOURCE=600) endif() # Data types, macros and functions related to controlling CPU affinity and @@ -1351,74 +194,147 @@ endif() if (WIN32) add_compile_definitions(_CRT_SECURE_NO_WARNINGS) - - if (BUILD_SHARED_LIBS) - # TODO: should not use this - set(CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS ON) - endif() endif() -# -# libraries -# - # ggml -add_library(ggml +if (GGML_BACKEND_DL AND NOT BUILD_SHARED_LIBS) + message(FATAL_ERROR "GGML_BACKEND_DL requires BUILD_SHARED_LIBS") +endif() + +add_library(ggml-base ../include/ggml.h - ../include/ggml-cpu.h ../include/ggml-alloc.h ../include/ggml-backend.h ../include/ggml-cpp.h + ../include/ggml-opt.h + ../include/gguf.h ggml.c - ggml-cpu.c ggml-alloc.c ggml-backend.cpp + ggml-opt.cpp + ggml-threading.cpp + ggml-threading.h ggml-quants.c ggml-quants.h - ${GGML_SOURCES_CUDA} ${GGML_HEADERS_CUDA} - ${GGML_SOURCES_METAL} ${GGML_HEADERS_METAL} - ${GGML_SOURCES_RPC} ${GGML_HEADERS_RPC} - ${GGML_SOURCES_EXTRA} ${GGML_HEADERS_EXTRA} - ${GGML_SOURCES_SYCL} ${GGML_HEADERS_SYCL} - ${GGML_SOURCES_KOMPUTE} ${GGML_HEADERS_KOMPUTE} - ${GGML_SOURCES_VULKAN} ${GGML_HEADERS_VULKAN} - ${GGML_SOURCES_ROCM} ${GGML_HEADERS_ROCM} - ${GGML_SOURCES_BLAS} ${GGML_HEADERS_BLAS} - ${GGML_SOURCES_LLAMAFILE} ${GGML_HEADERS_LLAMAFILE} - ${GGML_SOURCES_AMX} ${GGML_HEADERS_AMX} - ${GGML_SOURCES_CANN} ${GGML_HEADERS_CANN} - ggml-aarch64.c ggml-aarch64.h - ) + gguf.cpp) -if (EMSCRIPTEN) - set_target_properties(ggml PROPERTIES COMPILE_FLAGS "-msimd128") +target_include_directories(ggml-base PRIVATE .) + +add_library(ggml + ggml-backend-reg.cpp) + +target_link_libraries(ggml PUBLIC ggml-base) + +if (CMAKE_SYSTEM_NAME MATCHES "Linux") + target_link_libraries(ggml PRIVATE dl) endif() -target_compile_definitions(ggml PUBLIC ${GGML_CDEF_PUBLIC}) -target_include_directories(ggml PUBLIC $ $) -target_include_directories(ggml PRIVATE . ${GGML_EXTRA_INCLUDES}) -target_link_directories (ggml PRIVATE ${GGML_EXTRA_LIBDIRS}) -target_compile_features (ggml PRIVATE c_std_11) # don't bump +function(ggml_add_backend_library backend) + if (GGML_BACKEND_DL) + add_library(${backend} MODULE ${ARGN}) + # write the shared library to the output directory + set_target_properties(${backend} PROPERTIES LIBRARY_OUTPUT_DIRECTORY ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}) + target_compile_definitions(${backend} PRIVATE GGML_BACKEND_DL) + add_dependencies(ggml ${backend}) + else() + add_library(${backend} ${ARGN}) + target_link_libraries(ggml PUBLIC ${backend}) + install(TARGETS ${backend} LIBRARY) + endif() -list(APPEND GGML_EXTRA_LIBS_PRIVATE Threads::Threads) + target_link_libraries(${backend} PRIVATE ggml-base) + target_include_directories(${backend} PRIVATE ..) + + if (${BUILD_SHARED_LIBS}) + target_compile_definitions(${backend} PRIVATE GGML_BACKEND_BUILD) + target_compile_definitions(${backend} PUBLIC GGML_BACKEND_SHARED) + endif() +endfunction() + +function(ggml_add_backend backend) + string(TOUPPER "GGML_${backend}" backend_id) + if (${backend_id}) + string(TOLOWER "ggml-${backend}" backend_target) + add_subdirectory(${backend_target}) + message(STATUS "Including ${backend} backend") + if (NOT GGML_BACKEND_DL) + string(TOUPPER "GGML_USE_${backend}" backend_use) + target_compile_definitions(ggml PUBLIC ${backend_use}) + endif() + endif() +endfunction() + +function(ggml_add_cpu_backend_variant tag_name) + set(GGML_CPU_TAG_NAME ${tag_name}) + # other: OPENMP LLAMAFILE CPU_HBM + foreach (feat NATIVE + AVX AVX2 AVX_VNNI FMA F16C + AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16 + AMX_TILE AMX_INT8 AMX_BF16) + set(GGML_${feat} OFF) + endforeach() + + foreach (feat ${ARGN}) + set(GGML_${feat} ON) + endforeach() + + ggml_add_cpu_backend_variant_impl(${tag_name}) +endfunction() + +ggml_add_backend(CPU) + +if (GGML_CPU_ALL_VARIANTS) + if (NOT GGML_BACKEND_DL) + message(FATAL_ERROR "GGML_CPU_ALL_VARIANTS requires GGML_BACKEND_DL") + endif() + ggml_add_cpu_backend_variant(sandybridge AVX) + ggml_add_cpu_backend_variant(haswell AVX F16C AVX2 FMA) + ggml_add_cpu_backend_variant(skylakex AVX F16C AVX2 FMA AVX512) + ggml_add_cpu_backend_variant(icelake AVX F16C AVX2 FMA AVX512 AVX512_VBMI AVX512_VNNI) + ggml_add_cpu_backend_variant(alderlake AVX F16C AVX2 FMA AVX_VNNI) + if (NOT MSVC) + # MSVC doesn't support AMX + ggml_add_cpu_backend_variant(sapphirerapids AVX F16C AVX2 FMA AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16 AMX_TILE AMX_INT8) + endif() +else () + ggml_add_cpu_backend_variant_impl("") +endif() + +ggml_add_backend(BLAS) +ggml_add_backend(CANN) +ggml_add_backend(CUDA) +ggml_add_backend(HIP) +ggml_add_backend(Kompute) +ggml_add_backend(METAL) +ggml_add_backend(MUSA) +ggml_add_backend(RPC) +ggml_add_backend(SYCL) +ggml_add_backend(Vulkan) +ggml_add_backend(OpenCL) + +foreach (target ggml-base ggml) + target_include_directories(${target} PUBLIC $ $) + target_compile_features (${target} PRIVATE c_std_11 cxx_std_17) # don't bump +endforeach() + +target_link_libraries(ggml-base PRIVATE Threads::Threads) find_library(MATH_LIBRARY m) if (MATH_LIBRARY) if (NOT WIN32 OR NOT DEFINED ENV{ONEAPI_ROOT}) - list(APPEND GGML_EXTRA_LIBS_PRIVATE m) + target_link_libraries(ggml-base PRIVATE m) endif() endif() if (CMAKE_SYSTEM_NAME MATCHES "Android") - list(APPEND GGML_EXTRA_LIBS_PRIVATE dl) # Must be linked explicitly + target_link_libraries(ggml-base PRIVATE dl) endif() -list(REMOVE_DUPLICATES GGML_EXTRA_LIBS_PRIVATE) -list(REMOVE_DUPLICATES GGML_EXTRA_LIBS_PUBLIC) -target_link_libraries(ggml PRIVATE ${GGML_EXTRA_LIBS_PRIVATE} PUBLIC ${GGML_EXTRA_LIBS_PUBLIC}) - if (BUILD_SHARED_LIBS) - set_target_properties(ggml PROPERTIES POSITION_INDEPENDENT_CODE ON) - target_compile_definitions(ggml PRIVATE GGML_SHARED GGML_BUILD) + foreach (target ggml-base ggml) + set_target_properties(${target} PROPERTIES POSITION_INDEPENDENT_CODE ON) + target_compile_definitions(${target} PRIVATE GGML_BUILD) + target_compile_definitions(${target} PUBLIC GGML_SHARED) + endforeach() endif() diff --git a/ggml/src/ggml-aarch64.h b/ggml/src/ggml-aarch64.h deleted file mode 100644 index 517babaf1..000000000 --- a/ggml/src/ggml-aarch64.h +++ /dev/null @@ -1,39 +0,0 @@ -// SPDX-FileCopyrightText: Copyright 2024 Arm Ltd. -#pragma once - -#define GGML_COMMON_DECL_C -#include "ggml-common.h" - -#include "ggml.h" - -// GGML internal header - -#ifdef __cplusplus -extern "C" { -#endif - -// Quantization -void quantize_q8_0_4x4(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_q8_0_4x8(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); - -void quantize_mat_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t nrows, int64_t n_per_row, int64_t blck_size_interleave); - -// Quantization utilizing an importance matrix (a.k.a. "Activation aWare Quantization") -size_t quantize_q4_0_4x4(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q4_0_4x8(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q4_0_8x8(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); - -// GEMV -void ggml_gemv_q4_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); -void ggml_gemv_q4_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); -void ggml_gemv_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); - -// GEMM -void ggml_gemm_q4_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); -void ggml_gemm_q4_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); -void ggml_gemm_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc); - -#ifdef __cplusplus -} -#endif - diff --git a/ggml/src/ggml-alloc.c b/ggml/src/ggml-alloc.c index 041de9e3e..8dc8226ac 100644 --- a/ggml/src/ggml-alloc.c +++ b/ggml/src/ggml-alloc.c @@ -466,18 +466,12 @@ static bool ggml_gallocr_is_own(ggml_gallocr_t galloc, struct ggml_tensor * t) { return ggml_gallocr_hash_get(galloc, t)->allocated; } -static void ggml_gallocr_set_node_offset(ggml_gallocr_t galloc, struct ggml_tensor * node, int buffer_id, size_t offset) { - struct hash_node * hn = ggml_gallocr_hash_get(galloc, node); - hn->buffer_id = buffer_id; - hn->offset = offset; - hn->allocated = true; -} - static bool ggml_gallocr_is_allocated(ggml_gallocr_t galloc, struct ggml_tensor * t) { return t->data != NULL || ggml_gallocr_hash_get(galloc, t)->allocated; } static void ggml_gallocr_allocate_node(ggml_gallocr_t galloc, struct ggml_tensor * node, int buffer_id) { + GGML_ASSERT(buffer_id >= 0); struct hash_node * hn = ggml_gallocr_hash_get(galloc, node); if (!ggml_gallocr_is_allocated(galloc, node) && !ggml_is_view(node)) { @@ -540,7 +534,6 @@ static void ggml_gallocr_allocate_node(ggml_gallocr_t galloc, struct ggml_tensor size_t offset = ggml_dyn_tallocr_alloc(alloc, size, node); hn->buffer_id = buffer_id; hn->offset = offset; - return; } } @@ -816,7 +809,11 @@ static void ggml_gallocr_init_tensor(ggml_gallocr_t galloc, struct ggml_tensor * } static bool ggml_gallocr_node_needs_realloc(ggml_gallocr_t galloc, struct ggml_tensor * node, struct tensor_alloc * talloc) { - size_t node_size = (node->data || node->view_src) ? 0 : ggml_backend_buft_get_alloc_size(galloc->bufts[talloc->buffer_id], node); + size_t node_size = 0; + if (!node->data && !node->view_src) { + GGML_ASSERT(talloc->buffer_id >= 0); // prevent segfault when misusing the API + node_size = ggml_backend_buft_get_alloc_size(galloc->bufts[talloc->buffer_id], node); + } return talloc->size_max >= node_size; } diff --git a/ggml/src/ggml-amx.cpp b/ggml/src/ggml-amx.cpp deleted file mode 100644 index 144dc9d8a..000000000 --- a/ggml/src/ggml-amx.cpp +++ /dev/null @@ -1,436 +0,0 @@ -#include "ggml-amx.h" -#include "ggml-amx/common.h" -#include "ggml-amx/mmq.h" -#include "ggml-backend-impl.h" -#include "ggml-impl.h" - -#if defined(__gnu_linux__) -#include -#include -#endif - -#include -#include -#include - -#if defined(__AMX_INT8__) - -// AMX buffer interface -static void ggml_backend_amx_buffer_free_buffer(ggml_backend_buffer_t buffer) { - free(buffer->context); -} - -static void * ggml_backend_amx_buffer_get_base(ggml_backend_buffer_t buffer) { - return (void *)(buffer->context); -} - -static void ggml_backend_amx_buffer_memset_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { - memset((char *)tensor->data + offset, value, size); - - GGML_UNUSED(buffer); -} - -static void ggml_backend_amx_buffer_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - if (qtype_has_amx_kernels(tensor->type)) { - ggml_backend_amx_convert_weight(tensor, data, offset, size); - } else { - memcpy((char *)tensor->data + offset, data, size); - } - - GGML_UNUSED(buffer); -} - -static void ggml_backend_amx_buffer_get_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - GGML_ASSERT(!qtype_has_amx_kernels(tensor->type)); - memcpy(data, (const char *)tensor->data + offset, size); - - GGML_UNUSED(buffer); -} - -static bool ggml_backend_amx_buffer_cpy_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst) { - if (ggml_backend_buffer_is_host(src->buffer)) { - if (qtype_has_amx_kernels(src->type)) { - ggml_backend_amx_convert_weight(dst, src->data, 0, ggml_backend_amx_get_alloc_size(dst)); - } else { - memcpy(dst->data, src->data, ggml_nbytes(src)); - } - return true; - } - return false; - - GGML_UNUSED(buffer); -} - -static void ggml_backend_amx_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { - memset(buffer->context, value, buffer->size); -} - -static ggml_backend_buffer_i ggml_backend_amx_buffer_interface = { - /* .free_buffer = */ ggml_backend_amx_buffer_free_buffer, - /* .get_base = */ ggml_backend_amx_buffer_get_base, - /* .init_tensor = */ NULL, // no initialization required - /* .memset_tensor = */ ggml_backend_amx_buffer_memset_tensor, - /* .set_tensor = */ ggml_backend_amx_buffer_set_tensor, - /* .get_tensor = */ ggml_backend_amx_buffer_get_tensor, - /* .cpy_tensor = */ ggml_backend_amx_buffer_cpy_tensor, - /* .clear = */ ggml_backend_amx_buffer_clear, - /* .reset = */ NULL, -}; - -static const char * ggml_backend_amx_buffer_type_get_name(ggml_backend_buffer_type_t buft) { - return "AMX"; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_t ggml_backend_amx_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { - void * data = aligned_alloc(TENSOR_ALIGNMENT, size); - if (data == NULL) { - fprintf(stderr, "%s: failed to allocate buffer of size %zu\n", __func__, size); - return NULL; - } - - return ggml_backend_buffer_init(buft, ggml_backend_amx_buffer_interface, data, size); -} - -static size_t ggml_backend_amx_buffer_type_get_alignment(ggml_backend_buffer_type_t buft) { - return TENSOR_ALIGNMENT; - - GGML_UNUSED(buft); -} - -static size_t ggml_backend_amx_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor* tensor) { - return ggml_backend_amx_get_alloc_size(tensor); - - GGML_UNUSED(buft); -} - -static bool ggml_backend_amx_buffer_type_is_host(ggml_backend_buffer_type_t buft) { - return false; - - GGML_UNUSED(buft); -} - -ggml_backend_buffer_type_t ggml_backend_amx_buffer_type() { - static struct ggml_backend_buffer_type ggml_backend_buffer_type_amx = { - /* .iface = */ { - /* .get_name = */ ggml_backend_amx_buffer_type_get_name, - /* .alloc_buffer = */ ggml_backend_amx_buffer_type_alloc_buffer, - /* .get_alignment = */ ggml_backend_amx_buffer_type_get_alignment, - /* .get_max_size = */ NULL, // defaults to SIZE_MAX - /* .get_alloc_size = */ ggml_backend_amx_buffer_type_get_alloc_size, - /* .is_host = */ ggml_backend_amx_buffer_type_is_host, - }, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_amx_reg(), 0), - /* .context = */ NULL, - }; - - return &ggml_backend_buffer_type_amx; -} - -// backend interface - -static const char * ggml_backend_amx_name(ggml_backend_t backend) { - return "AMX"; - - GGML_UNUSED(backend); -} - -static void ggml_backend_amx_free(ggml_backend_t backend) { - ggml_backend_amx_context * ctx = (ggml_backend_amx_context *)backend->context; - delete ctx; - delete backend; -} - -static enum ggml_status ggml_backend_amx_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - ggml_backend_amx_context * ctx = (ggml_backend_amx_context *)backend->context; - - for (int i = 0; i < cgraph->n_nodes; i++) { - struct ggml_tensor * node = cgraph->nodes[i]; - - switch (node->op) { - case GGML_OP_MUL_MAT: - ggml_backend_amx_mul_mat(ctx, node); - break; - - case GGML_OP_NONE: - case GGML_OP_RESHAPE: - case GGML_OP_VIEW: - case GGML_OP_PERMUTE: - case GGML_OP_TRANSPOSE: - break; - - default: - fprintf(stderr, "%s: unsupported op %s\n", __func__, ggml_op_desc(node)); - GGML_ASSERT(false); - } - } - - return GGML_STATUS_SUCCESS; - - GGML_UNUSED(backend); -} - -static struct ggml_backend_i ggml_backend_amx_i = { - /* .get_name = */ ggml_backend_amx_name, - /* .free = */ ggml_backend_amx_free, - /* .set_tensor_async = */ NULL, - /* .get_tensor_async = */ NULL, - /* .cpy_tensor_async = */ NULL, - /* .synchronize = */ NULL, - /* .graph_plan_create = */ NULL, - /* .graph_plan_free = */ NULL, - /* .graph_plan_update = */ NULL, - /* .graph_plan_compute = */ NULL, - /* .graph_compute = */ ggml_backend_amx_graph_compute, - /* .event_record = */ NULL, - /* .event_wait = */ NULL, -}; - -static ggml_guid_t ggml_backend_amx_guid() { - static ggml_guid guid = { 0x13, 0xb8, 0xa4, 0xc4, 0xba, 0xfe, 0x51, 0x67, 0x87, 0x44, 0x55, 0x15, 0xb2, 0x35, 0x62, 0x3e }; - return &guid; -} - -#define ARCH_GET_XCOMP_PERM 0x1022 -#define ARCH_REQ_XCOMP_PERM 0x1023 -#define XFEATURE_XTILECFG 17 -#define XFEATURE_XTILEDATA 18 - -static bool ggml_amx_init() { -#if defined(__gnu_linux__) - if (syscall(SYS_arch_prctl, ARCH_REQ_XCOMP_PERM, XFEATURE_XTILEDATA)) { - fprintf(stderr, "AMX is not ready to be used!\n"); - return false; - } - return true; -#elif defined(_WIN32) - return true; -#endif -} - -ggml_backend_t ggml_backend_amx_init() { - - // invoke a Linux system call to request access to AMX features - ggml_amx_init(); - - // backend context - ggml_backend_amx_context * ctx = new ggml_backend_amx_context; - - // ggml amx backend - ggml_backend_t backend = new ggml_backend { - /* .guid = */ ggml_backend_amx_guid(), - /* .interface = */ ggml_backend_amx_i, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_amx_reg(), 0), - /* .context = */ ctx, - }; - - return backend; -} - -bool ggml_backend_is_amx(ggml_backend_t backend) { - return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_amx_guid()); -} - -void ggml_backend_amx_set_n_threads(ggml_backend_t backend_amx, int n_threads) { - GGML_ASSERT(ggml_backend_is_amx(backend_amx)); - - ggml_backend_amx_context * ctx = (ggml_backend_amx_context *)backend_amx->context; - ctx->n_threads = n_threads; -} - -// device interface - -static const char * ggml_backend_amx_device_get_name(ggml_backend_dev_t dev) { - return "AMX"; - - GGML_UNUSED(dev); -} - -static const char * ggml_backend_amx_device_get_description(ggml_backend_dev_t dev) { - return "Intel Advanced Matrix Extensions"; - - GGML_UNUSED(dev); -} - -static void ggml_backend_amx_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { - // TODO - *free = 0; - *total = 0; - - GGML_UNUSED(dev); -} - -static enum ggml_backend_dev_type ggml_backend_amx_device_get_type(ggml_backend_dev_t dev) { - return GGML_BACKEND_DEVICE_TYPE_ACCEL; - - GGML_UNUSED(dev); -} - -static void ggml_backend_amx_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { - props->name = ggml_backend_amx_device_get_name(dev); - props->description = ggml_backend_amx_device_get_description(dev); - props->type = ggml_backend_amx_device_get_type(dev); - ggml_backend_amx_device_get_memory(dev, &props->memory_free, &props->memory_total); - - // `buffer_from_host_ptr` is intended to be used in mmap, when memory layout unchanged - props->caps = { - /* .async = */ false, - /* .host_buffer = */ false, - /* .buffer_from_host_ptr = */ false, - /* .events = */ false, - }; -} - -static ggml_backend_t ggml_backend_amx_device_init(ggml_backend_dev_t dev, const char * params) { - return ggml_backend_amx_init(); - - GGML_UNUSED(dev); - GGML_UNUSED(params); -} - -static ggml_backend_buffer_type_t ggml_backend_amx_device_get_buffer_type(ggml_backend_dev_t dev) { - return ggml_backend_amx_buffer_type(); - - GGML_UNUSED(dev); -} - -static bool ggml_backend_amx_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { - - // handle only 2d gemm for now - auto is_contiguous_2d = [](const struct ggml_tensor * t) { - return ggml_is_contiguous(t) && t->ne[3] == 1 && t->ne[2] == 1; - }; - - switch (op->op) { - case GGML_OP_NONE: - case GGML_OP_RESHAPE: - case GGML_OP_VIEW: - case GGML_OP_PERMUTE: - case GGML_OP_TRANSPOSE: - return true; - - case GGML_OP_MUL_MAT: { - const struct ggml_tensor * src0 = op->src[0]; - const struct ggml_tensor * src1 = op->src[1]; - - const enum ggml_type type = src0->type; - const int64_t ne0 = op->ne[0]; - - bool is_training = src0->grad || src1->grad; - - // amx kernels enables for Q4_0, Q4_1, Q8_0, F16 - // Q4_K, Q5_K, Q6_K, IQ4_XS enabled for QK_K = 256 - bool has_amx_kernels = qtype_has_amx_kernels(type) || (type == GGML_TYPE_F16); - - bool can_use_amx = - is_contiguous_2d(src0) && // src0 must be contiguous - is_contiguous_2d(src1) && // src1 must be contiguous - !is_training && // inference only - src1->type == GGML_TYPE_F32 && // src1 must be float32 - has_amx_kernels && // with amx kernel impls - ne0 % (TILE_N * 2) == 0; // out_features is 32x - - return can_use_amx; - } - default: - return false; - } - - GGML_UNUSED(dev); -} - -static bool ggml_backend_amx_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { - return buft->iface.get_name == ggml_backend_amx_buffer_type_get_name; - - GGML_UNUSED(dev); -} - -static const struct ggml_backend_device_i ggml_backend_amx_device_i = { - /* .get_name = */ ggml_backend_amx_device_get_name, - /* .get_description = */ ggml_backend_amx_device_get_description, - /* .get_memory = */ ggml_backend_amx_device_get_memory, - /* .get_type = */ ggml_backend_amx_device_get_type, - /* .get_props = */ ggml_backend_amx_device_get_props, - /* .init_backend = */ ggml_backend_amx_device_init, - /* .get_buffer_type = */ ggml_backend_amx_device_get_buffer_type, - /* .get_host_buffer_type = */ NULL, - /* .buffer_from_host_ptr = */ NULL, - /* .supports_op = */ ggml_backend_amx_device_supports_op, - /* .supports_buft = */ ggml_backend_amx_device_supports_buft, - /* .offload_op = */ NULL, - /* .event_new = */ NULL, - /* .event_free = */ NULL, - /* .event_synchronize = */ NULL, -}; - -// backend reg interface - -static const char * ggml_backend_amx_reg_get_name(ggml_backend_reg_t reg) { - return "AMX"; - - GGML_UNUSED(reg); -} - -static size_t ggml_backend_amx_reg_get_device_count(ggml_backend_reg_t reg) { - return 1; - - GGML_UNUSED(reg); -} - -static ggml_backend_dev_t ggml_backend_amx_reg_get_device(ggml_backend_reg_t reg, size_t index) { - GGML_ASSERT(index == 0); - - static ggml_backend_device ggml_backend_amx_device = { - /* .iface = */ ggml_backend_amx_device_i, - /* .reg = */ reg, - /* .context = */ nullptr, - }; - - return &ggml_backend_amx_device; - - GGML_UNUSED(reg); - GGML_UNUSED(index); -} - -static void * ggml_backend_amx_get_proc_address(ggml_backend_reg_t reg, const char * name) { - if (std::strcmp(name, "ggml_backend_set_n_threads") == 0) { - return (void *)ggml_backend_amx_set_n_threads; - } - return NULL; - - GGML_UNUSED(reg); - GGML_UNUSED(name); -} - -static const struct ggml_backend_reg_i ggml_backend_amx_reg_i = { - /* .get_name = */ ggml_backend_amx_reg_get_name, - /* .get_device_count = */ ggml_backend_amx_reg_get_device_count, - /* .get_device = */ ggml_backend_amx_reg_get_device, - /* .get_proc_address = */ ggml_backend_amx_get_proc_address, -}; - -ggml_backend_reg_t ggml_backend_amx_reg(void) { - static struct ggml_backend_reg ggml_backend_amx_reg = { - /* .iface = */ ggml_backend_amx_reg_i, - /* .context = */ NULL, - }; - - return &ggml_backend_amx_reg; -} - -#else // if defined(__AMX_INT8__) - -ggml_backend_t ggml_backend_amx_init(void) { - fprintf(stderr, "GGML is not compiled with AMX support!\n"); - return ggml_backend_t{}; -} - -void ggml_backend_amx_set_n_threads(ggml_backend_t backend_amx, int n_threads) { - fprintf(stderr, "GGML is not compiled with AMX support!\n"); - - GGML_UNUSED(backend_amx); - GGML_UNUSED(n_threads); -} - -#endif diff --git a/ggml/src/ggml-backend-impl.h b/ggml/src/ggml-backend-impl.h index fa8d5b7fb..36d72e95f 100644 --- a/ggml/src/ggml-backend-impl.h +++ b/ggml/src/ggml-backend-impl.h @@ -8,6 +8,8 @@ extern "C" { #endif + #define GGML_BACKEND_API_VERSION 1 + // // Backend buffer type // @@ -63,20 +65,20 @@ extern "C" { enum ggml_backend_buffer_usage usage; }; - ggml_backend_buffer_t ggml_backend_buffer_init( + GGML_API ggml_backend_buffer_t ggml_backend_buffer_init( ggml_backend_buffer_type_t buft, struct ggml_backend_buffer_i iface, void * context, size_t size); // do not use directly, use ggml_backend_tensor_copy instead - bool ggml_backend_buffer_copy_tensor(const struct ggml_tensor * src, struct ggml_tensor * dst); + GGML_API bool ggml_backend_buffer_copy_tensor(const struct ggml_tensor * src, struct ggml_tensor * dst); // multi-buffer // buffer that contains a collection of buffers - ggml_backend_buffer_t ggml_backend_multi_buffer_alloc_buffer(ggml_backend_buffer_t * buffers, size_t n_buffers); - bool ggml_backend_buffer_is_multi_buffer(ggml_backend_buffer_t buffer); - void ggml_backend_multi_buffer_set_usage(ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage); + GGML_API ggml_backend_buffer_t ggml_backend_multi_buffer_alloc_buffer(ggml_backend_buffer_t * buffers, size_t n_buffers); + GGML_API bool ggml_backend_buffer_is_multi_buffer(ggml_backend_buffer_t buffer); + GGML_API void ggml_backend_multi_buffer_set_usage(ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage); // // Backend (stream) @@ -199,17 +201,55 @@ extern "C" { }; struct ggml_backend_reg { - // int api_version; // TODO: for dynamic loading + int api_version; // initialize to GGML_BACKEND_API_VERSION struct ggml_backend_reg_i iface; void * context; }; - // Internal backend registry API - void ggml_backend_register(ggml_backend_reg_t reg); - void ggml_backend_device_register(ggml_backend_dev_t device); - // TODO: backends can be loaded as a dynamic library, in which case it needs to export this function - // typedef ggml_backend_register_t * (*ggml_backend_init)(void); + GGML_API void ggml_backend_register(ggml_backend_reg_t reg); + GGML_API void ggml_backend_device_register(ggml_backend_dev_t device); + + // Add backend dynamic loading support to the backend + + // Initialize the backend + typedef ggml_backend_reg_t (*ggml_backend_init_t)(void); + // Optional: obtain a score for the backend based on the system configuration + // Higher scores are preferred, 0 means the backend is not supported in the current system + typedef int (*ggml_backend_score_t)(void); + +#ifdef GGML_BACKEND_DL +# ifdef __cplusplus +# define GGML_BACKEND_DL_IMPL(reg_fn) \ + extern "C" { \ + GGML_BACKEND_API ggml_backend_reg_t ggml_backend_init(void); \ + } \ + ggml_backend_reg_t ggml_backend_init(void) { \ + return reg_fn(); \ + } +# define GGML_BACKEND_DL_SCORE_IMPL(score_fn) \ + extern "C" { \ + GGML_BACKEND_API int ggml_backend_score(void); \ + } \ + int ggml_backend_score(void) { \ + return score_fn(); \ + } +# else +# define GGML_BACKEND_DL_IMPL(reg_fn) \ + GGML_BACKEND_API ggml_backend_reg_t ggml_backend_init(void); \ + ggml_backend_reg_t ggml_backend_init(void) { \ + return reg_fn(); \ + } +# define GGML_BACKEND_DL_SCORE_IMPL(score_fn) \ + GGML_BACKEND_API int ggml_backend_score(void); \ + int ggml_backend_score(void) { \ + return score_fn(); \ + } +# endif +#else +# define GGML_BACKEND_DL_IMPL(reg_fn) +# define GGML_BACKEND_DL_SCORE_IMPL(score_fn) +#endif #ifdef __cplusplus } diff --git a/ggml/src/ggml-backend-reg.cpp b/ggml/src/ggml-backend-reg.cpp new file mode 100644 index 000000000..955ed505f --- /dev/null +++ b/ggml/src/ggml-backend-reg.cpp @@ -0,0 +1,582 @@ +#include "ggml-backend-impl.h" +#include "ggml-backend.h" +#include "ggml-impl.h" +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef _WIN32 +# define WIN32_LEAN_AND_MEAN +# ifndef NOMINMAX +# define NOMINMAX +# endif +# include +#elif defined(__APPLE__) +# include +# include +#else +# include +# include +#endif + +// Backend registry +#ifdef GGML_USE_CPU +#include "ggml-cpu.h" +#endif + +#ifdef GGML_USE_CUDA +#include "ggml-cuda.h" +#endif + +#ifdef GGML_USE_METAL +#include "ggml-metal.h" +#endif + +#ifdef GGML_USE_SYCL +#include "ggml-sycl.h" +#endif + +#ifdef GGML_USE_VULKAN +#include "ggml-vulkan.h" +#endif + +#ifdef GGML_USE_OPENCL +#include "ggml-opencl.h" +#endif + +#ifdef GGML_USE_BLAS +#include "ggml-blas.h" +#endif + +#ifdef GGML_USE_RPC +#include "ggml-rpc.h" +#endif + +#ifdef GGML_USE_CANN +#include "ggml-cann.h" +#endif + +#ifdef GGML_USE_KOMPUTE +#include "ggml-kompute.h" +#endif + +// disable C++17 deprecation warning for std::codecvt_utf8 +#if defined(__clang__) +# pragma clang diagnostic push +# pragma clang diagnostic ignored "-Wdeprecated-declarations" +#endif + +static std::wstring utf8_to_utf16(const std::string & str) { + std::wstring_convert> converter; + return converter.from_bytes(str); +} + +static std::string utf16_to_utf8(const std::wstring & str) { + std::wstring_convert> converter; + return converter.to_bytes(str); +} + +#if defined(__clang__) +# pragma clang diagnostic pop +#endif + +#ifdef _WIN32 + +using dl_handle = std::remove_pointer_t; + +struct dl_handle_deleter { + void operator()(HMODULE handle) { + FreeLibrary(handle); + } +}; + +static dl_handle * dl_load_library(const std::wstring & path) { + // suppress error dialogs for missing DLLs + DWORD old_mode = SetErrorMode(SEM_FAILCRITICALERRORS); + SetErrorMode(old_mode | SEM_FAILCRITICALERRORS); + + HMODULE handle = LoadLibraryW(path.c_str()); + + SetErrorMode(old_mode); + + return handle; +} + +static void * dl_get_sym(dl_handle * handle, const char * name) { + DWORD old_mode = SetErrorMode(SEM_FAILCRITICALERRORS); + SetErrorMode(old_mode | SEM_FAILCRITICALERRORS); + + void * p = (void *) GetProcAddress(handle, name); + + SetErrorMode(old_mode); + + return p; +} + +#else + +using dl_handle = void; + +struct dl_handle_deleter { + void operator()(void * handle) { + dlclose(handle); + } +}; + +static void * dl_load_library(const std::wstring & path) { + dl_handle * handle = dlopen(utf16_to_utf8(path).c_str(), RTLD_NOW | RTLD_LOCAL); + + return handle; +} + +static void * dl_get_sym(dl_handle * handle, const char * name) { + return dlsym(handle, name); +} + +#endif + +using dl_handle_ptr = std::unique_ptr; + +struct ggml_backend_reg_entry { + ggml_backend_reg_t reg; + dl_handle_ptr handle; +}; + +struct ggml_backend_registry { + std::vector backends; + std::vector devices; + + ggml_backend_registry() { +#ifdef GGML_USE_CUDA + register_backend(ggml_backend_cuda_reg()); +#endif +#ifdef GGML_USE_METAL + register_backend(ggml_backend_metal_reg()); +#endif +#ifdef GGML_USE_SYCL + register_backend(ggml_backend_sycl_reg()); +#endif +#ifdef GGML_USE_VULKAN + register_backend(ggml_backend_vk_reg()); +#endif +#ifdef GGML_USE_OPENCL + register_backend(ggml_backend_opencl_reg()); +#endif +#ifdef GGML_USE_CANN + register_backend(ggml_backend_cann_reg()); +#endif +#ifdef GGML_USE_BLAS + register_backend(ggml_backend_blas_reg()); +#endif +#ifdef GGML_USE_RPC + register_backend(ggml_backend_rpc_reg()); +#endif +#ifdef GGML_USE_KOMPUTE + register_backend(ggml_backend_kompute_reg()); +#endif +#ifdef GGML_USE_CPU + register_backend(ggml_backend_cpu_reg()); +#endif + } + + ~ggml_backend_registry() { + // FIXME: backends cannot be safely unloaded without a function to destroy all the backend resources, + // since backend threads may still be running and accessing resources from the dynamic library + for (auto & entry : backends) { + if (entry.handle) { + entry.handle.release(); // NOLINT + } + } + } + + void register_backend(ggml_backend_reg_t reg, dl_handle_ptr handle = nullptr) { + if (!reg) { + return; + } + +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: registered backend %s (%zu devices)\n", + __func__, ggml_backend_reg_name(reg), ggml_backend_reg_dev_count(reg)); +#endif + backends.push_back({ reg, std::move(handle) }); + for (size_t i = 0; i < ggml_backend_reg_dev_count(reg); i++) { + register_device(ggml_backend_reg_dev_get(reg, i)); + } + } + + void register_device(ggml_backend_dev_t device) { +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: registered device %s (%s)\n", __func__, ggml_backend_dev_name(device), ggml_backend_dev_description(device)); +#endif + devices.push_back(device); + } + + ggml_backend_reg_t load_backend(const std::wstring & path, bool silent) { + dl_handle_ptr handle { dl_load_library(path) }; + if (!handle) { + if (!silent) { + GGML_LOG_ERROR("%s: failed to load %s\n", __func__, utf16_to_utf8(path).c_str()); + } + return nullptr; + } + + auto score_fn = (ggml_backend_score_t) dl_get_sym(handle.get(), "ggml_backend_score"); + if (score_fn && score_fn() == 0) { + if (!silent) { + GGML_LOG_INFO("%s: backend %s is not supported on this system\n", __func__, utf16_to_utf8(path).c_str()); + } + return nullptr; + } + + auto backend_init_fn = (ggml_backend_init_t) dl_get_sym(handle.get(), "ggml_backend_init"); + if (!backend_init_fn) { + if (!silent) { + GGML_LOG_ERROR("%s: failed to find ggml_backend_init in %s\n", __func__, utf16_to_utf8(path).c_str()); + } + return nullptr; + } + + ggml_backend_reg_t reg = backend_init_fn(); + if (!reg || reg->api_version != GGML_BACKEND_API_VERSION) { + if (!silent) { + if (!reg) { + GGML_LOG_ERROR("%s: failed to initialize backend from %s: ggml_backend_init returned NULL\n", __func__, utf16_to_utf8(path).c_str()); + } else { + GGML_LOG_ERROR("%s: failed to initialize backend from %s: incompatible API version (backend: %d, current: %d)\n", + __func__, utf16_to_utf8(path).c_str(), reg->api_version, GGML_BACKEND_API_VERSION); + } + } + return nullptr; + } + + GGML_LOG_INFO("%s: loaded %s backend from %s\n", __func__, ggml_backend_reg_name(reg), utf16_to_utf8(path).c_str()); + + register_backend(reg, std::move(handle)); + + return reg; + } + + void unload_backend(ggml_backend_reg_t reg, bool silent) { + auto it = std::find_if(backends.begin(), backends.end(), + [reg](const ggml_backend_reg_entry & entry) { return entry.reg == reg; }); + + if (it == backends.end()) { + if (!silent) { + GGML_LOG_ERROR("%s: backend not found\n", __func__); + } + return; + } + + if (!silent) { + GGML_LOG_DEBUG("%s: unloading %s backend\n", __func__, ggml_backend_reg_name(reg)); + } + + // remove devices + devices.erase( + std::remove_if(devices.begin(), devices.end(), + [reg](ggml_backend_dev_t dev) { return ggml_backend_dev_backend_reg(dev) == reg; }), + devices.end()); + + // remove backend + backends.erase(it); + } +}; + +static ggml_backend_registry & get_reg() { + static ggml_backend_registry reg; + return reg; +} + +// Internal API +void ggml_backend_register(ggml_backend_reg_t reg) { + get_reg().register_backend(reg); +} + +void ggml_backend_device_register(ggml_backend_dev_t device) { + get_reg().register_device(device); +} + +// Backend (reg) enumeration +static bool striequals(const char * a, const char * b) { + for (; *a && *b; a++, b++) { + if (std::tolower(*a) != std::tolower(*b)) { + return false; + } + } + return *a == *b; +} + +size_t ggml_backend_reg_count() { + return get_reg().backends.size(); +} + +ggml_backend_reg_t ggml_backend_reg_get(size_t index) { + GGML_ASSERT(index < ggml_backend_reg_count()); + return get_reg().backends[index].reg; +} + +ggml_backend_reg_t ggml_backend_reg_by_name(const char * name) { + for (size_t i = 0; i < ggml_backend_reg_count(); i++) { + ggml_backend_reg_t reg = ggml_backend_reg_get(i); + if (striequals(ggml_backend_reg_name(reg), name)) { + return reg; + } + } + return nullptr; +} + +// Device enumeration +size_t ggml_backend_dev_count() { + return get_reg().devices.size(); +} + +ggml_backend_dev_t ggml_backend_dev_get(size_t index) { + GGML_ASSERT(index < ggml_backend_dev_count()); + return get_reg().devices[index]; +} + +ggml_backend_dev_t ggml_backend_dev_by_name(const char * name) { + for (size_t i = 0; i < ggml_backend_dev_count(); i++) { + ggml_backend_dev_t dev = ggml_backend_dev_get(i); + if (striequals(ggml_backend_dev_name(dev), name)) { + return dev; + } + } + return nullptr; +} + +ggml_backend_dev_t ggml_backend_dev_by_type(enum ggml_backend_dev_type type) { + for (size_t i = 0; i < ggml_backend_dev_count(); i++) { + ggml_backend_dev_t dev = ggml_backend_dev_get(i); + if (ggml_backend_dev_type(dev) == type) { + return dev; + } + } + return nullptr; +} + +// Convenience functions +ggml_backend_t ggml_backend_init_by_name(const char * name, const char * params) { + ggml_backend_dev_t dev = ggml_backend_dev_by_name(name); + if (!dev) { + return nullptr; + } + return ggml_backend_dev_init(dev, params); +} + +ggml_backend_t ggml_backend_init_by_type(enum ggml_backend_dev_type type, const char * params) { + ggml_backend_dev_t dev = ggml_backend_dev_by_type(type); + if (!dev) { + return nullptr; + } + return ggml_backend_dev_init(dev, params); +} + +ggml_backend_t ggml_backend_init_best(void) { + ggml_backend_dev_t dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_GPU); + if (!dev) { + dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + } + if (!dev) { + return nullptr; + } + return ggml_backend_dev_init(dev, nullptr); +} + +// Dynamic loading +ggml_backend_reg_t ggml_backend_load(const char * path) { + return get_reg().load_backend(utf8_to_utf16(path), false); +} + +void ggml_backend_unload(ggml_backend_reg_t reg) { + get_reg().unload_backend(reg, true); +} + +static std::wstring get_executable_path() { +#if defined(__APPLE__) + // get executable path + std::vector path; + uint32_t size; + while (true) { + size = path.size(); + if (_NSGetExecutablePath(path.data(), &size) == 0) { + break; + } + path.resize(size); + } + std::string base_path(path.data(), size); + // remove executable name + auto last_slash = base_path.find_last_of('/'); + if (last_slash != std::string::npos) { + base_path = base_path.substr(0, last_slash); + } + return utf8_to_utf16(base_path + "/"); +#elif defined(__linux__) || defined(__FreeBSD__) + std::string base_path = "."; + std::vector path(1024); + while (true) { + // get executable path +# if defined(__linux__) + ssize_t len = readlink("/proc/self/exe", path.data(), path.size()); +# elif defined(__FreeBSD__) + ssize_t len = readlink("/proc/curproc/file", path.data(), path.size()); +# endif + if (len == -1) { + break; + } + if (len < (ssize_t) path.size()) { + base_path = std::string(path.data(), len); + // remove executable name + auto last_slash = base_path.find_last_of('/'); + if (last_slash != std::string::npos) { + base_path = base_path.substr(0, last_slash); + } + break; + } + path.resize(path.size() * 2); + } + + return utf8_to_utf16(base_path + "/"); +#elif defined(_WIN32) + std::vector path(MAX_PATH); + DWORD len = GetModuleFileNameW(NULL, path.data(), path.size()); + if (len == 0) { + return {}; + } + std::wstring base_path(path.data(), len); + // remove executable name + auto last_slash = base_path.find_last_of('\\'); + if (last_slash != std::string::npos) { + base_path = base_path.substr(0, last_slash); + } + return base_path + L"\\"; +#else + return {}; +#endif +} + +static std::wstring backend_filename_prefix() { +#ifdef _WIN32 + return L"ggml-"; +#else + return L"libggml-"; +#endif +} + +static std::wstring backend_filename_suffix() { +#ifdef _WIN32 + return L".dll"; +#else + return L".so"; +#endif +} + +static std::wstring path_separator() { +#ifdef _WIN32 + return L"\\"; +#else + return L"/"; +#endif +} + +static ggml_backend_reg_t ggml_backend_load_best(const char * name, bool silent, const char * user_search_path) { + // enumerate all the files that match [lib]ggml-name-*.[so|dll] in the search paths + // TODO: search system paths + std::wstring file_prefix = backend_filename_prefix() + utf8_to_utf16(name) + L"-"; + std::vector search_paths; + if (user_search_path == nullptr) { + search_paths.push_back(L"." + path_separator()); + search_paths.push_back(get_executable_path()); + } else { + search_paths.push_back(utf8_to_utf16(user_search_path) + path_separator()); + } + + int best_score = 0; + std::wstring best_path; + + namespace fs = std::filesystem; + for (const auto & search_path : search_paths) { + if (!fs::exists(search_path)) { + continue; + } + fs::directory_iterator dir_it(search_path, fs::directory_options::skip_permission_denied); + for (const auto & entry : dir_it) { + if (entry.is_regular_file()) { + std::wstring filename = entry.path().filename().wstring(); + std::wstring ext = entry.path().extension().wstring(); + if (filename.find(file_prefix) == 0 && ext == backend_filename_suffix()) { + dl_handle_ptr handle { dl_load_library(entry.path().wstring()) }; + if (!handle && !silent) { + GGML_LOG_ERROR("%s: failed to load %s\n", __func__, utf16_to_utf8(entry.path().wstring()).c_str()); + } + if (handle) { + auto score_fn = (ggml_backend_score_t) dl_get_sym(handle.get(), "ggml_backend_score"); + if (score_fn) { + int s = score_fn(); +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: %s score: %d\n", __func__, utf16_to_utf8(entry.path().wstring()).c_str(), s); +#endif + if (s > best_score) { + best_score = s; + best_path = entry.path().wstring(); + } + } else { + if (!silent) { + GGML_LOG_INFO("%s: failed to find ggml_backend_score in %s\n", __func__, utf16_to_utf8(entry.path().wstring()).c_str()); + } + } + } + } + } + } + } + + if (best_score == 0) { + // try to load the base backend + for (const auto & search_path : search_paths) { + std::wstring path = search_path + backend_filename_prefix() + utf8_to_utf16(name) + backend_filename_suffix(); + if (fs::exists(path)) { + return get_reg().load_backend(path, silent); + } + } + return nullptr; + } + + return get_reg().load_backend(best_path, silent); +} + +void ggml_backend_load_all() { + ggml_backend_load_all_from_path(nullptr); +} + +void ggml_backend_load_all_from_path(const char * dir_path) { +#ifdef NDEBUG + bool silent = true; +#else + bool silent = false; +#endif + + ggml_backend_load_best("blas", silent, dir_path); + ggml_backend_load_best("cann", silent, dir_path); + ggml_backend_load_best("cuda", silent, dir_path); + ggml_backend_load_best("hip", silent, dir_path); + ggml_backend_load_best("kompute", silent, dir_path); + ggml_backend_load_best("metal", silent, dir_path); + ggml_backend_load_best("rpc", silent, dir_path); + ggml_backend_load_best("sycl", silent, dir_path); + ggml_backend_load_best("vulkan", silent, dir_path); + ggml_backend_load_best("opencl", silent, dir_path); + ggml_backend_load_best("musa", silent, dir_path); + ggml_backend_load_best("cpu", silent, dir_path); + // check the environment variable GGML_BACKEND_PATH to load an out-of-tree backend + const char * backend_path = std::getenv("GGML_BACKEND_PATH"); + if (backend_path) { + ggml_backend_load(backend_path); + } +} diff --git a/ggml/src/ggml-backend.cpp b/ggml/src/ggml-backend.cpp index 0b8ebac53..dba7be33b 100644 --- a/ggml/src/ggml-backend.cpp +++ b/ggml/src/ggml-backend.cpp @@ -252,6 +252,7 @@ void ggml_backend_tensor_get_async(ggml_backend_t backend, const struct ggml_ten } void ggml_backend_tensor_set(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + GGML_ASSERT(tensor); ggml_backend_buffer_t buf = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; if (size == 0) { @@ -266,6 +267,7 @@ void ggml_backend_tensor_set(struct ggml_tensor * tensor, const void * data, siz } void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + GGML_ASSERT(tensor); ggml_backend_buffer_t buf = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; if (size == 0) { @@ -279,7 +281,7 @@ void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, siz buf->iface.get_tensor(buf, tensor, data, offset, size); } -GGML_API void ggml_backend_tensor_memset(struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { +void ggml_backend_tensor_memset(struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { ggml_backend_buffer_t buf = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; if (size == 0) { @@ -525,197 +527,6 @@ void * ggml_backend_reg_get_proc_address(ggml_backend_reg_t reg, const char * na return reg->iface.get_proc_address(reg, name); } -// Backend registry - -#ifdef GGML_USE_CUDA -#include "ggml-cuda.h" -#endif - -#ifdef GGML_USE_METAL -#include "ggml-metal.h" -#endif - -#ifdef GGML_USE_SYCL -#include "ggml-sycl.h" -#endif - -#ifdef GGML_USE_VULKAN -#include "ggml-vulkan.h" -#endif - -#ifdef GGML_USE_BLAS -#include "ggml-blas.h" -#endif - -#ifdef GGML_USE_RPC -#include "ggml-rpc.h" -#endif - -#ifndef __AMX_INT8__ -#undef GGML_USE_AMX -#endif - -#ifdef GGML_USE_AMX -# include "ggml-amx.h" -#endif - -#ifdef GGML_USE_CANN -#include "ggml-cann.h" -#endif - -#ifdef GGML_USE_KOMPUTE -#include "ggml-kompute.h" -#endif - -#include "ggml-cpu.h" - -struct ggml_backend_registry { - std::vector backends; - std::vector devices; - - ggml_backend_registry() { -#ifdef GGML_USE_CUDA - register_backend(ggml_backend_cuda_reg()); -#endif -#ifdef GGML_USE_METAL - register_backend(ggml_backend_metal_reg()); -#endif -#ifdef GGML_USE_SYCL - register_backend(ggml_backend_sycl_reg()); -#endif -#ifdef GGML_USE_VULKAN - register_backend(ggml_backend_vk_reg()); -#endif -#ifdef GGML_USE_CANN - register_backend(ggml_backend_cann_reg()); -#endif -#ifdef GGML_USE_BLAS - register_backend(ggml_backend_blas_reg()); -#endif -#ifdef GGML_USE_RPC - register_backend(ggml_backend_rpc_reg()); -#endif -#ifdef GGML_USE_AMX - register_backend(ggml_backend_amx_reg()); -#endif -#ifdef GGML_USE_KOMPUTE - register_backend(ggml_backend_kompute_reg()); -#endif - - register_backend(ggml_backend_cpu_reg()); - } - - void register_backend(ggml_backend_reg_t reg) { -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: registered backend %s (%zu devices)\n", - __func__, ggml_backend_reg_name(reg), ggml_backend_reg_dev_count(reg)); -#endif - backends.push_back(reg); - for (size_t i = 0; i < ggml_backend_reg_dev_count(reg); i++) { - register_device(ggml_backend_reg_dev_get(reg, i)); - } - } - - void register_device(ggml_backend_dev_t device) { -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: registered device %s (%s)\n", __func__, ggml_backend_dev_name(device), ggml_backend_dev_description(device)); -#endif - devices.push_back(device); - } -}; - -static ggml_backend_registry & get_reg() { - static ggml_backend_registry reg; - return reg; -} - -// Internal API -void ggml_backend_register(ggml_backend_reg_t reg) { - get_reg().register_backend(reg); -} - -void ggml_backend_device_register(ggml_backend_dev_t device) { - get_reg().register_device(device); -} - -// Backend (reg) enumeration -size_t ggml_backend_reg_count() { - return get_reg().backends.size(); -} - -ggml_backend_reg_t ggml_backend_reg_get(size_t index) { - GGML_ASSERT(index < ggml_backend_reg_count()); - return get_reg().backends[index]; -} - -ggml_backend_reg_t ggml_backend_reg_by_name(const char * name) { - for (size_t i = 0; i < ggml_backend_reg_count(); i++) { - ggml_backend_reg_t reg = ggml_backend_reg_get(i); - if (strcmp(ggml_backend_reg_name(reg), name) == 0) { - return reg; - } - } - return NULL; -} - -// Device enumeration -size_t ggml_backend_dev_count() { - return get_reg().devices.size(); -} - -ggml_backend_dev_t ggml_backend_dev_get(size_t index) { - GGML_ASSERT(index < ggml_backend_dev_count()); - return get_reg().devices[index]; -} - -ggml_backend_dev_t ggml_backend_dev_by_name(const char * name) { - for (size_t i = 0; i < ggml_backend_dev_count(); i++) { - ggml_backend_dev_t dev = ggml_backend_dev_get(i); - if (strcmp(ggml_backend_dev_name(dev), name) == 0) { - return dev; - } - } - return NULL; -} - -ggml_backend_dev_t ggml_backend_dev_by_type(enum ggml_backend_dev_type type) { - for (size_t i = 0; i < ggml_backend_dev_count(); i++) { - ggml_backend_dev_t dev = ggml_backend_dev_get(i); - if (ggml_backend_dev_type(dev) == type) { - return dev; - } - } - return NULL; -} - -// Convenience functions -ggml_backend_t ggml_backend_init_by_name(const char * name, const char * params) { - ggml_backend_dev_t dev = ggml_backend_dev_by_name(name); - if (!dev) { - return NULL; - } - return ggml_backend_dev_init(dev, params); -} - -ggml_backend_t ggml_backend_init_by_type(enum ggml_backend_dev_type type, const char * params) { - ggml_backend_dev_t dev = ggml_backend_dev_by_type(type); - if (!dev) { - return NULL; - } - return ggml_backend_dev_init(dev, params); -} - -ggml_backend_t ggml_backend_init_best(void) { - ggml_backend_dev_t dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_GPU); - if (!dev) { - dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); - } - if (!dev) { - return NULL; - } - return ggml_backend_dev_init(dev, NULL); -} - // multi-buffer buffer struct ggml_backend_multi_buffer_context { @@ -880,7 +691,7 @@ static int ggml_backend_sched_backend_id(ggml_backend_sched_t sched, ggml_backen } static int ggml_backend_sched_backend_from_buffer(ggml_backend_sched_t sched, const struct ggml_tensor * tensor, const struct ggml_tensor * op) { - ggml_backend_buffer_t buffer = tensor->buffer; + ggml_backend_buffer_t buffer = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; if (buffer == NULL) { return -1; } @@ -913,8 +724,6 @@ static char causes[GGML_DEFAULT_GRAPH_SIZE*16 + GGML_SCHED_MAX_SPLITS_DEBUG*GGML // returns the backend that should be used for the node based on the current locations static int ggml_backend_sched_backend_id_from_cur(ggml_backend_sched_t sched, struct ggml_tensor * tensor) { - // TODO: use supports_op to check if the backend supports the op - // assign pre-allocated nodes to their backend int cur_backend_id = ggml_backend_sched_backend_from_buffer(sched, tensor, tensor); if (cur_backend_id != -1) { @@ -933,7 +742,8 @@ static int ggml_backend_sched_backend_id_from_cur(ggml_backend_sched_t sched, st if (tensor->buffer || (tensor->view_src && tensor->view_src->buffer)) { // since the tensor is pre-allocated, it cannot be moved to another backend - GGML_ABORT("pre-allocated tensor in a backend that cannot run the operation"); + ggml_backend_buffer_t buffer = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; + GGML_ABORT("pre-allocated tensor (%s) in a buffer (%s) that cannot run the operation (%s)", tensor->name, ggml_backend_buffer_name(buffer), ggml_op_name(tensor->op)); } // graph input @@ -954,7 +764,7 @@ static int ggml_backend_sched_backend_id_from_cur(ggml_backend_sched_t sched, st if (tensor->op != GGML_OP_ROPE && src->buffer != NULL && src->buffer->usage == GGML_BACKEND_BUFFER_USAGE_WEIGHTS) { int src_backend_id = ggml_backend_sched_backend_from_buffer(sched, src, tensor); // check if a backend with higher prio wants to offload the op - if (src_backend_id == sched->n_backends - 1) { + if (src_backend_id == sched->n_backends - 1 && ggml_backend_buffer_is_host(src->buffer)) { for (int b = 0; b < src_backend_id; b++) { if (ggml_backend_supports_op(sched->backends[b], tensor) && ggml_backend_offload_op(sched->backends[b], tensor)) { SET_CAUSE(tensor, "1.off"); @@ -985,9 +795,12 @@ static void ggml_backend_sched_print_assignments(ggml_backend_sched_t sched, str for (int i = 0; i < graph->n_nodes; i++) { if (cur_split < sched->n_splits && i == sched->splits[cur_split].i_start) { ggml_backend_t split_backend = sched->backends[sched->splits[cur_split].backend_id]; - GGML_LOG_DEBUG("\n## SPLIT #%d: %s # %d inputs: ", cur_split, ggml_backend_name(split_backend), + GGML_LOG_DEBUG("\n## SPLIT #%d: %s # %d inputs", cur_split, ggml_backend_name(split_backend), sched->splits[cur_split].n_inputs); for (int j = 0; j < sched->splits[cur_split].n_inputs; j++) { + if (j == 0) { + GGML_LOG_DEBUG(": "); + } GGML_LOG_DEBUG("[%s (%5.5s)] ", sched->splits[cur_split].inputs[j]->name, fmt_size(ggml_nbytes(sched->splits[cur_split].inputs[j]))); } @@ -1640,7 +1453,7 @@ ggml_backend_sched_t ggml_backend_sched_new( bool parallel) { GGML_ASSERT(n_backends > 0); GGML_ASSERT(n_backends <= GGML_SCHED_MAX_BACKENDS); - GGML_ASSERT(ggml_backend_is_cpu(backends[n_backends - 1])); // last backend must be CPU + GGML_ASSERT(ggml_backend_dev_type(ggml_backend_get_device(backends[n_backends - 1])) == GGML_BACKEND_DEVICE_TYPE_CPU); struct ggml_backend_sched * sched = (ggml_backend_sched *) calloc(1, sizeof(struct ggml_backend_sched)); @@ -1729,12 +1542,13 @@ bool ggml_backend_sched_reserve(ggml_backend_sched_t sched, struct ggml_cgraph * ggml_backend_sched_split_graph(sched, measure_graph); + ggml_backend_sched_synchronize(sched); + if (!ggml_gallocr_reserve_n(sched->galloc, &sched->graph, sched->node_backend_ids, sched->leaf_backend_ids)) { return false; } ggml_backend_sched_reset(sched); - ggml_backend_sched_synchronize(sched); return true; } @@ -2036,17 +1850,6 @@ bool ggml_backend_compare_graph_backend(ggml_backend_t backend1, ggml_backend_t return true; } - - -#include "ggml-backend.h" -#include "ggml-backend-impl.h" -#include "ggml-cpu.h" -#include "ggml-impl.h" -#include -#include - -// ggml-backend interface - // CPU backend - buffer static void * ggml_backend_cpu_buffer_get_base(ggml_backend_buffer_t buffer) { @@ -2120,7 +1923,9 @@ static const struct ggml_backend_buffer_i ggml_backend_cpu_buffer_from_ptr_i = { /* .reset = */ NULL, }; -// CPU backend - buffer type +// CPU backend buffer type + +// this buffer type is defined here to make it available to all backends static const char * ggml_backend_cpu_buffer_type_get_name(ggml_backend_buffer_type_t buft) { return "CPU"; @@ -2161,7 +1966,7 @@ ggml_backend_buffer_type_t ggml_backend_cpu_buffer_type(void) { /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, }, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), + /* .device = */ NULL, // FIXME ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), /* .context = */ NULL, }; @@ -2184,478 +1989,14 @@ static ggml_backend_buffer_type_t ggml_backend_cpu_buffer_from_ptr_type(void) { /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, }, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), + /* .device = */ NULL, // FIXME ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), /* .context = */ NULL, }; return &ggml_backend_cpu_buffer_type; } -#ifdef GGML_USE_CPU_HBM - -// buffer type HBM - -#include - -static const char * ggml_backend_cpu_hbm_buffer_type_get_name(ggml_backend_buffer_type_t buft) { - return "CPU_HBM"; - - GGML_UNUSED(buft); -} - -static void ggml_backend_cpu_hbm_buffer_free_buffer(ggml_backend_buffer_t buffer) { - hbw_free(buffer->context); -} - -static ggml_backend_buffer_t ggml_backend_cpu_hbm_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { - void * ptr; - int result = hbw_posix_memalign(&ptr, ggml_backend_cpu_buffer_type_get_alignment(buft), size); - if (result != 0) { - GGML_LOG_ERROR("failed to allocate HBM buffer of size %zu\n", size); - return NULL; - } - - ggml_backend_buffer_t buffer = ggml_backend_cpu_buffer_from_ptr(ptr, size); - buffer->buft = buft; - buffer->iface.free_buffer = ggml_backend_cpu_hbm_buffer_free_buffer; - - return buffer; -} - -ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void) { - static struct ggml_backend_buffer_type ggml_backend_cpu_buffer_type_hbm = { - /* .iface = */ { - /* .get_name = */ ggml_backend_cpu_hbm_buffer_type_get_name, - /* .alloc_buffer = */ ggml_backend_cpu_hbm_buffer_type_alloc_buffer, - /* .get_alignment = */ ggml_backend_cpu_buffer_type_get_alignment, - /* .get_max_size = */ NULL, // defaults to SIZE_MAX - /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes - /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, - }, - /* .context = */ NULL, - }; - - return &ggml_backend_cpu_buffer_type_hbm; -} -#endif - -static ggml_backend_buffer_type_t * ggml_backend_cpu_get_extra_bufts(ggml_backend_dev_t device) { - static ggml_backend_buffer_type_t bufts[] = { -#ifdef GGML_USE_CPU_HBM - ggml_backend_cpu_hbm_buffer_type(), -#endif - NULL - }; - - return bufts; - - GGML_UNUSED(device); -} - -// CPU backend - backend (stream) - -struct ggml_backend_cpu_context { - int n_threads; - ggml_threadpool_t threadpool; - - uint8_t * work_data; - size_t work_size; - - ggml_abort_callback abort_callback; - void * abort_callback_data; -}; - -static const char * ggml_backend_cpu_get_name(ggml_backend_t backend) { - return "CPU"; - - GGML_UNUSED(backend); -} - -static void ggml_backend_cpu_free(ggml_backend_t backend) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; - delete[] cpu_ctx->work_data; - delete cpu_ctx; - delete backend; -} - -struct ggml_backend_plan_cpu { - struct ggml_cplan cplan; - struct ggml_cgraph cgraph; -}; - -static ggml_backend_graph_plan_t ggml_backend_cpu_graph_plan_create(ggml_backend_t backend, const struct ggml_cgraph * cgraph) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; - - struct ggml_backend_plan_cpu * cpu_plan = new ggml_backend_plan_cpu; - - cpu_plan->cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads, cpu_ctx->threadpool); - cpu_plan->cgraph = *cgraph; // FIXME: deep copy - - if (cpu_plan->cplan.work_size > 0) { - cpu_plan->cplan.work_data = new uint8_t[cpu_plan->cplan.work_size]; - if (cpu_plan->cplan.work_data == NULL) { - delete cpu_plan; - return NULL; - } - } - - cpu_plan->cplan.abort_callback = cpu_ctx->abort_callback; - cpu_plan->cplan.abort_callback_data = cpu_ctx->abort_callback_data; - - return cpu_plan; -} - -static void ggml_backend_cpu_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { - struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; - - delete[] cpu_plan->cplan.work_data; - delete cpu_plan; - - GGML_UNUSED(backend); -} - -static enum ggml_status ggml_backend_cpu_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { - struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; - - return ggml_graph_compute(&cpu_plan->cgraph, &cpu_plan->cplan); - - GGML_UNUSED(backend); -} - -static enum ggml_status ggml_backend_cpu_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; - - struct ggml_cplan cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads, cpu_ctx->threadpool); - - if (cpu_ctx->work_size < cplan.work_size) { - delete[] cpu_ctx->work_data; - cpu_ctx->work_data = new uint8_t[cplan.work_size]; - if (cpu_ctx->work_data == NULL) { - cpu_ctx->work_size = 0; - return GGML_STATUS_ALLOC_FAILED; - } - cpu_ctx->work_size = cplan.work_size; - } - cplan.work_data = (uint8_t *)cpu_ctx->work_data; - - cplan.abort_callback = cpu_ctx->abort_callback; - cplan.abort_callback_data = cpu_ctx->abort_callback_data; - - return ggml_graph_compute(cgraph, &cplan); -} - -static const struct ggml_backend_i ggml_backend_cpu_i = { - /* .get_name = */ ggml_backend_cpu_get_name, - /* .free = */ ggml_backend_cpu_free, - /* .set_tensor_async = */ NULL, - /* .get_tensor_async = */ NULL, - /* .cpy_tensor_async = */ NULL, - /* .synchronize = */ NULL, - /* .graph_plan_create = */ ggml_backend_cpu_graph_plan_create, - /* .graph_plan_free = */ ggml_backend_cpu_graph_plan_free, - /* .graph_plan_update = */ NULL, - /* .graph_plan_compute = */ ggml_backend_cpu_graph_plan_compute, - /* .graph_compute = */ ggml_backend_cpu_graph_compute, - /* .event_record = */ NULL, - /* .event_wait = */ NULL, -}; - -static ggml_guid_t ggml_backend_cpu_guid(void) { - static ggml_guid guid = { 0xaa, 0x67, 0xc7, 0x43, 0x96, 0xe6, 0xa3, 0x8a, 0xe3, 0xaf, 0xea, 0x92, 0x36, 0xbc, 0xfc, 0x89 }; - return &guid; -} - -ggml_backend_t ggml_backend_cpu_init(void) { - // initialize CPU backend now to avoid slowing the first graph computation - ggml_cpu_init(); - - struct ggml_backend_cpu_context * ctx = new ggml_backend_cpu_context; - if (ctx == NULL) { - return NULL; - } - - ctx->n_threads = GGML_DEFAULT_N_THREADS; - ctx->threadpool = NULL; - ctx->work_data = NULL; - ctx->work_size = 0; - ctx->abort_callback = NULL; - ctx->abort_callback_data = NULL; - - ggml_backend_t cpu_backend = new ggml_backend { - /* .guid = */ ggml_backend_cpu_guid(), - /* .interface = */ ggml_backend_cpu_i, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), - /* .context = */ ctx, - }; - - if (cpu_backend == NULL) { - delete ctx; - return NULL; - } - - return cpu_backend; -} - -bool ggml_backend_is_cpu(ggml_backend_t backend) { - return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_cpu_guid()); -} - -void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads) { - GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); - - struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; - ctx->n_threads = n_threads; -} - -void ggml_backend_cpu_set_threadpool(ggml_backend_t backend_cpu, ggml_threadpool_t threadpool) { - GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); - - struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; - - if (ctx->threadpool && ctx->threadpool != threadpool) { - // already had a different threadpool, pause/suspend it before switching - ggml_threadpool_pause(ctx->threadpool); - } - ctx->threadpool = threadpool; -} - -void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data) { - GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); - - struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; - ctx->abort_callback = abort_callback; - ctx->abort_callback_data = abort_callback_data; -} - ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(void * ptr, size_t size) { GGML_ASSERT((uintptr_t)ptr % TENSOR_ALIGNMENT == 0 && "buffer pointer must be aligned"); return ggml_backend_buffer_init(ggml_backend_cpu_buffer_from_ptr_type(), ggml_backend_cpu_buffer_from_ptr_i, ptr, size); } - -// CPU backend - device - -struct ggml_backend_cpu_device_context { - std::string description = "CPU"; - - ggml_backend_cpu_device_context() { -#ifdef __APPLE__ - size_t len = 0; - if (!sysctlbyname("machdep.cpu.brand_string", NULL, &len, NULL, 0)) { - description.resize(len); - sysctlbyname("machdep.cpu.brand_string", &description[0], &len, NULL, 0); // NOLINT - } -#elif defined(__linux__) - FILE * f = fopen("/proc/cpuinfo", "r"); - if (f) { - char buf[1024]; - while (fgets(buf, sizeof(buf), f)) { - if (strncmp(buf, "model name", 10) == 0) { - char * p = strchr(buf, ':'); - if (p) { - p++; - while (std::isspace(*p)) { - p++; - } - while (std::isspace(p[strlen(p) - 1])) { - p[strlen(p) - 1] = '\0'; - } - description = p; - break; - } - } - } - fclose(f); - } -#elif defined(_WIN32) - HKEY hKey; - if (RegOpenKeyEx(HKEY_LOCAL_MACHINE, - TEXT("HARDWARE\\DESCRIPTION\\System\\CentralProcessor\\0"), - 0, - KEY_READ, - &hKey) == ERROR_SUCCESS) { - DWORD cpu_brand_size = 0; - if (RegQueryValueExA(hKey, - TEXT("ProcessorNameString"), - NULL, - NULL, - NULL, - &cpu_brand_size) == ERROR_SUCCESS) { - description.resize(cpu_brand_size); - if (RegQueryValueExA(hKey, - TEXT("ProcessorNameString"), - NULL, - NULL, - (LPBYTE)&description[0], // NOLINT - &cpu_brand_size) == ERROR_SUCCESS) { - if (description.find('\0') != std::string::npos) { - description.resize(description.find('\0')); - } - } - } - RegCloseKey(hKey); - } -#endif - } -}; - -static const char * ggml_backend_cpu_device_get_name(ggml_backend_dev_t dev) { - return "CPU"; - - GGML_UNUSED(dev); -} - -static const char * ggml_backend_cpu_device_get_description(ggml_backend_dev_t dev) { - struct ggml_backend_cpu_device_context * ctx = (struct ggml_backend_cpu_device_context *)dev->context; - - return ctx->description.c_str(); -} - -static void ggml_backend_cpu_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { - // TODO - *free = 0; - *total = 0; - - GGML_UNUSED(dev); -} - -static enum ggml_backend_dev_type ggml_backend_cpu_device_get_type(ggml_backend_dev_t dev) { - return GGML_BACKEND_DEVICE_TYPE_CPU; - - GGML_UNUSED(dev); -} - -static void ggml_backend_cpu_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { - props->name = ggml_backend_cpu_device_get_name(dev); - props->description = ggml_backend_cpu_device_get_description(dev); - props->type = ggml_backend_cpu_device_get_type(dev); - ggml_backend_cpu_device_get_memory(dev, &props->memory_free, &props->memory_total); - props->caps = { - /* .async = */ false, - /* .host_buffer = */ false, - /* .buffer_from_host_ptr = */ true, - /* .events = */ false, - }; -} - -static ggml_backend_t ggml_backend_cpu_device_init_backend(ggml_backend_dev_t dev, const char * params) { - return ggml_backend_cpu_init(); - - GGML_UNUSED(dev); - GGML_UNUSED(params); -} - -static ggml_backend_buffer_type_t ggml_backend_cpu_device_get_buffer_type(ggml_backend_dev_t dev) { - return ggml_backend_cpu_buffer_type(); - - GGML_UNUSED(dev); -} - -static ggml_backend_buffer_t ggml_backend_cpu_device_buffer_from_host_ptr(ggml_backend_dev_t dev, void * ptr, size_t size, size_t max_tensor_size) { - return ggml_backend_cpu_buffer_from_ptr(ptr, size); - - GGML_UNUSED(dev); - GGML_UNUSED(max_tensor_size); -} - -static bool ggml_backend_cpu_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { - switch (op->op) { - case GGML_OP_CPY: - return - op->type != GGML_TYPE_IQ2_XXS && - op->type != GGML_TYPE_IQ2_XS && - op->type != GGML_TYPE_IQ1_S && - op->type != GGML_TYPE_IQ1_M; // missing type_traits.from_float - case GGML_OP_MUL_MAT: - return op->src[1]->type == GGML_TYPE_F32;// FIXME || op->src[1]->type == ggml_get_type_traits(op->src[0]->type)->vec_dot_type; - case GGML_OP_ROPE_BACK: - return op->src[2] == NULL && (op->op_params[2] & 4) == 0; - case GGML_OP_IM2COL_BACK: - return op->src[0]->type == GGML_TYPE_F32 && op->src[1]->type == GGML_TYPE_F32; - case GGML_OP_OUT_PROD: - return (op->src[0]->type == GGML_TYPE_F32 || ggml_is_quantized(op->src[0]->type)) && op->src[1]->type == GGML_TYPE_F32; - default: - return true; - } - - GGML_UNUSED(dev); -} - -static bool ggml_backend_cpu_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { - return ggml_backend_buft_is_host(buft); - - GGML_UNUSED(dev); -} - -static const struct ggml_backend_device_i ggml_backend_cpu_device_i = { - /* .get_name = */ ggml_backend_cpu_device_get_name, - /* .get_description = */ ggml_backend_cpu_device_get_description, - /* .get_memory = */ ggml_backend_cpu_device_get_memory, - /* .get_type = */ ggml_backend_cpu_device_get_type, - /* .get_props = */ ggml_backend_cpu_device_get_props, - /* .init_backend = */ ggml_backend_cpu_device_init_backend, - /* .get_buffer_type = */ ggml_backend_cpu_device_get_buffer_type, - /* .get_host_buffer_type = */ NULL, - /* .buffer_from_host_ptr = */ ggml_backend_cpu_device_buffer_from_host_ptr, - /* .supports_op = */ ggml_backend_cpu_device_supports_op, - /* .supports_buft = */ ggml_backend_cpu_device_supports_buft, - /* .offload_op = */ NULL, - /* .event_new = */ NULL, - /* .event_free = */ NULL, - /* .event_synchronize = */ NULL, -}; - -// CPU backend - backend (reg) - -static const char * ggml_backend_cpu_reg_get_name(ggml_backend_reg_t reg) { - return "CPU"; - - GGML_UNUSED(reg); -} - -static size_t ggml_backend_cpu_reg_get_device_count(ggml_backend_reg_t reg) { - return 1; - - GGML_UNUSED(reg); -} - -static ggml_backend_dev_t ggml_backend_cpu_reg_get_device(ggml_backend_reg_t reg, size_t index) { - GGML_ASSERT(index == 0); - - static ggml_backend_cpu_device_context ctx; - static ggml_backend_device ggml_backend_cpu_device = { - /* .iface = */ ggml_backend_cpu_device_i, - /* .reg = */ reg, - /* .context = */ &ctx, - }; - - return &ggml_backend_cpu_device; -} - -static void * ggml_backend_cpu_get_proc_address(ggml_backend_reg_t reg, const char * name) { - if (strcmp(name, "ggml_backend_set_n_threads") == 0) { - return (void *)ggml_backend_cpu_set_n_threads; - } - if (strcmp(name, "ggml_backend_dev_get_extra_bufts") == 0) { - return (void *)ggml_backend_cpu_get_extra_bufts; - } - - return NULL; - - GGML_UNUSED(reg); -} - -static const struct ggml_backend_reg_i ggml_backend_cpu_reg_i = { - /* .get_name = */ ggml_backend_cpu_reg_get_name, - /* .get_device_count = */ ggml_backend_cpu_reg_get_device_count, - /* .get_device = */ ggml_backend_cpu_reg_get_device, - /* .get_proc_address = */ ggml_backend_cpu_get_proc_address, -}; - -ggml_backend_reg_t ggml_backend_cpu_reg(void) { - static struct ggml_backend_reg ggml_backend_cpu_reg = { - /* .iface = */ ggml_backend_cpu_reg_i, - /* .context = */ NULL, - }; - - return &ggml_backend_cpu_reg; -} diff --git a/ggml/src/ggml-blas/CMakeLists.txt b/ggml/src/ggml-blas/CMakeLists.txt new file mode 100644 index 000000000..0bf3c05d9 --- /dev/null +++ b/ggml/src/ggml-blas/CMakeLists.txt @@ -0,0 +1,87 @@ +if (GGML_STATIC) + set(BLA_STATIC ON) +endif() +#if (CMAKE_VERSION VERSION_GREATER_EQUAL 3.22) +# set(BLA_SIZEOF_INTEGER 8) +#endif() + +set(BLA_VENDOR ${GGML_BLAS_VENDOR}) +find_package(BLAS) + +if (BLAS_FOUND) + message(STATUS "BLAS found, Libraries: ${BLAS_LIBRARIES}") + + ggml_add_backend_library(ggml-blas + ggml-blas.cpp + ) + + if (${GGML_BLAS_VENDOR} MATCHES "Apple") + add_compile_definitions(ACCELERATE_NEW_LAPACK) + add_compile_definitions(ACCELERATE_LAPACK_ILP64) + add_compile_definitions(GGML_BLAS_USE_ACCELERATE) + elseif ("${BLAS_INCLUDE_DIRS}" STREQUAL "") + # BLAS_INCLUDE_DIRS is missing in FindBLAS.cmake. + # see https://gitlab.kitware.com/cmake/cmake/-/issues/20268 + find_package(PkgConfig REQUIRED) + if (${GGML_BLAS_VENDOR} MATCHES "Generic") + pkg_check_modules(DepBLAS blas) + elseif (${GGML_BLAS_VENDOR} MATCHES "OpenBLAS") + # As of openblas v0.3.22, the 64-bit is named openblas64.pc + pkg_check_modules(DepBLAS openblas64) + if (NOT DepBLAS_FOUND) + pkg_check_modules(DepBLAS openblas) + endif() + elseif (${GGML_BLAS_VENDOR} MATCHES "FLAME") + add_compile_definitions(GGML_BLAS_USE_BLIS) + pkg_check_modules(DepBLAS blis) + elseif (${GGML_BLAS_VENDOR} MATCHES "ATLAS") + pkg_check_modules(DepBLAS blas-atlas) + elseif (${GGML_BLAS_VENDOR} MATCHES "FlexiBLAS") + pkg_check_modules(DepBLAS flexiblas_api) + elseif (${GGML_BLAS_VENDOR} MATCHES "Intel") + add_compile_definitions(GGML_BLAS_USE_MKL) + # all Intel* libraries share the same include path + pkg_check_modules(DepBLAS mkl-sdl) + elseif (${GGML_BLAS_VENDOR} MATCHES "NVHPC") + # this doesn't provide pkg-config + # suggest to assign BLAS_INCLUDE_DIRS on your own + if ("${NVHPC_VERSION}" STREQUAL "") + message(WARNING "Better to set NVHPC_VERSION") + else() + set(DepBLAS_FOUND ON) + set(DepBLAS_INCLUDE_DIRS "/opt/nvidia/hpc_sdk/${CMAKE_SYSTEM_NAME}_${CMAKE_SYSTEM_PROCESSOR}/${NVHPC_VERSION}/math_libs/include") + endif() + endif() + if (DepBLAS_FOUND) + set(BLAS_INCLUDE_DIRS ${DepBLAS_INCLUDE_DIRS}) + else() + message(WARNING "BLAS_INCLUDE_DIRS neither been provided nor been automatically" + " detected by pkgconfig, trying to find cblas.h from possible paths...") + find_path(BLAS_INCLUDE_DIRS + NAMES cblas.h + HINTS + /usr/include + /usr/local/include + /usr/include/openblas + /opt/homebrew/opt/openblas/include + /usr/local/opt/openblas/include + /usr/include/x86_64-linux-gnu/openblas/include + ) + endif() + endif() + + message(STATUS "BLAS found, Includes: ${BLAS_INCLUDE_DIRS}") + + target_compile_options(ggml-blas PRIVATE ${BLAS_LINKER_FLAGS}) + + if (${BLAS_INCLUDE_DIRS} MATCHES "mkl" AND (${GGML_BLAS_VENDOR} MATCHES "Generic" OR ${GGML_BLAS_VENDOR} MATCHES "Intel")) + add_compile_definitions(GGML_BLAS_USE_MKL) + endif() + + target_link_libraries (ggml-blas PRIVATE ${BLAS_LIBRARIES}) + target_include_directories(ggml-blas PRIVATE ${BLAS_INCLUDE_DIRS}) +else() + message(ERROR "BLAS not found, please refer to " + "https://cmake.org/cmake/help/latest/module/FindBLAS.html#blas-lapack-vendors" + " to set correct GGML_BLAS_VENDOR") +endif() diff --git a/ggml/src/ggml-blas.cpp b/ggml/src/ggml-blas/ggml-blas.cpp similarity index 98% rename from ggml/src/ggml-blas.cpp rename to ggml/src/ggml-blas/ggml-blas.cpp index 8d96220b9..ec158dfac 100644 --- a/ggml/src/ggml-blas.cpp +++ b/ggml/src/ggml-blas/ggml-blas.cpp @@ -6,7 +6,7 @@ #include #include -#if defined(GGML_USE_ACCELERATE) +#if defined(GGML_BLAS_USE_ACCELERATE) # include #elif defined(GGML_BLAS_USE_MKL) # include @@ -320,7 +320,7 @@ static const char * ggml_backend_blas_device_get_name(ggml_backend_dev_t dev) { } static const char * ggml_backend_blas_device_get_description(ggml_backend_dev_t dev) { - #if defined(GGML_USE_ACCELERATE) + #if defined(GGML_BLAS_USE_ACCELERATE) return "Accelerate"; #elif defined(GGML_BLAS_USE_MKL) return "MKL"; @@ -506,9 +506,12 @@ static const struct ggml_backend_reg_i ggml_backend_blas_reg_i = { ggml_backend_reg_t ggml_backend_blas_reg(void) { static struct ggml_backend_reg ggml_backend_blas_reg = { - /* .iface = */ ggml_backend_blas_reg_i, - /* .context = */ NULL, + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_blas_reg_i, + /* .context = */ NULL, }; return &ggml_backend_blas_reg; } + +GGML_BACKEND_DL_IMPL(ggml_backend_blas_reg) diff --git a/ggml/src/ggml-cann/CMakeLists.txt b/ggml/src/ggml-cann/CMakeLists.txt new file mode 100644 index 000000000..05cf06bfa --- /dev/null +++ b/ggml/src/ggml-cann/CMakeLists.txt @@ -0,0 +1,76 @@ +if ("cann${CANN_INSTALL_DIR}" STREQUAL "cann" AND DEFINED ENV{ASCEND_TOOLKIT_HOME}) + set(CANN_INSTALL_DIR $ENV{ASCEND_TOOLKIT_HOME}) + message(STATUS "CANN: updated CANN_INSTALL_DIR from ASCEND_TOOLKIT_HOME=$ENV{ASCEND_TOOLKIT_HOME}") +endif() + +# Auto-detech Soc type and Soc version, if detect failed, will abort build +set(SOC_VERSION "") +function(detect_ascend_soc_type SOC_VERSION) + execute_process( + COMMAND bash -c "npu-smi info|awk -F' ' 'NF > 0 && NR==7 {print $3}'" + OUTPUT_VARIABLE npu_info + RESULT_VARIABLE npu_result + OUTPUT_STRIP_TRAILING_WHITESPACE + ) + if("${npu_info}" STREQUAL "" OR ${npu_result}) + message(FATAL_ERROR "Auto-detech ascend soc type failed, please specify manually or check ascend device working normally.") + endif() + set(${SOC_VERSION} "Ascend${npu_info}" PARENT_SCOPE) +endfunction() + +if(NOT SOC_TYPE) + detect_ascend_soc_type(SOC_VERSION) + set(SOC_TYPE "${SOC_VERSION}") + message(STATUS "CANN: SOC_VERSION auto-detected is:${SOC_VERSION}") +endif() + +string(TOLOWER ${SOC_TYPE} SOC_VERSION) # SOC_VERSION need lower + +# Construct Soc specify compile option: ASCEND_#Soc_Major_SN. Such as ASCEND_910B, ASCEND_310P. +string(REGEX MATCH "[0-9]+[a-zA-Z]" SOC_TYPE_MAJOR_SN "${SOC_VERSION}") +set(SOC_TYPE_COMPILE_OPTION "ASCEND_${SOC_TYPE_MAJOR_SN}") +string(TOUPPER ${SOC_TYPE_COMPILE_OPTION} SOC_TYPE_COMPILE_OPTION) + +if (CANN_INSTALL_DIR) + # Only Support Linux. + if (NOT UNIX) + message(FATAL_ERROR "CANN: CANN toolkit supports unix but not ${CMAKE_SYSTEM_NAME}") + endif() + + # Supported platforms: x86-64, arm64 + if (CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64") + elseif (CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "amd64") + else() + message(FATAL_ERROR "CANN: CANN toolkit supports x86-64 and arm64 but not ${CMAKE_SYSTEM_PROCESSOR}") + endif() + + # Set header and libs + set(CANN_INCLUDE_DIRS + ${CANN_INSTALL_DIR}/include + ${CANN_INSTALL_DIR}/include/aclnn + ${CANN_INSTALL_DIR}/acllib/include + ) + + add_subdirectory(kernels) + list(APPEND CANN_LIBRARIES + ascendcl + nnopbase + opapi + acl_op_compiler + ascendc_kernels + ) + + file(GLOB GGML_SOURCES_CANN "*.cpp") + + ggml_add_backend_library(ggml-cann ${GGML_SOURCES_CANN}) + target_link_libraries(ggml-cann PRIVATE ${CANN_LIBRARIES}) + target_include_directories(ggml-cann PRIVATE ${CANN_INCLUDE_DIRS}) + target_link_directories(ggml-cann PRIVATE ${CANN_INSTALL_DIR}/lib64) + + target_compile_definitions(ggml-cann PRIVATE "-D${SOC_TYPE_COMPILE_OPTION}") + + message(STATUS "CANN: CANN_INCLUDE_DIRS = ${CANN_INCLUDE_DIRS}") + message(STATUS "CANN: CANN_LIBRARIES = ${CANN_LIBRARIES}") +else() + message(FATAL_ERROR "CANN: Can't find CANN_INSTALL_DIR, did you forget to source set_var.sh?") +endif() diff --git a/ggml/src/ggml-cann/aclnn_ops.cpp b/ggml/src/ggml-cann/aclnn_ops.cpp index a4ec8418e..b2d857e1e 100644 --- a/ggml/src/ggml-cann/aclnn_ops.cpp +++ b/ggml/src/ggml-cann/aclnn_ops.cpp @@ -22,11 +22,14 @@ #include "aclnn_ops.h" +#include #include +#include #include #include #include #include +#include #include #include #include @@ -34,6 +37,7 @@ #include #include #include +#include #include #include #include @@ -53,6 +57,7 @@ #include #include +#include "ggml-impl.h" #include "kernels/ascendc_kernels.h" #define GGML_COMMON_DECL_C @@ -241,10 +246,14 @@ void ggml_cann_concat(ggml_backend_cann_context& ctx, ggml_tensor* dst) { aclTensor* acl_src1 = ggml_cann_create_tensor(src1); aclTensor* acl_dst = ggml_cann_create_tensor(dst); - int64_t concat_dim = 1; + const int32_t dim = ggml_get_op_params_i32(dst, 0); + + GGML_ASSERT(dim >= 0 && dim < 4); + int32_t acl_dim = 3 - dim; + aclTensor* tensors[] = {acl_src0, acl_src1}; aclTensorList* tensorList = aclCreateTensorList(tensors, 2); - aclnn_concat(ctx, tensorList, acl_dst, concat_dim); + aclnn_concat(ctx, tensorList, acl_dst, acl_dim); ACL_CHECK(aclDestroyTensorList(tensorList)); ACL_CHECK(aclDestroyTensor(acl_dst)); @@ -1096,9 +1105,9 @@ static aclTensor* aclnn_zero(ggml_backend_cann_context& ctx, void* buffer, } /** - * @brief Creates an ACL tensor initialized with ones using a provided buffer. + * @brief Creates an ACL tensor initialized with value using a provided buffer. * - * This function initializes a tensor with ones using the specified buffer and + * This function initializes a tensor with value using the specified buffer and * tensor parameters. * * @param ctx The context for the CANN backend operations. @@ -1111,12 +1120,12 @@ static aclTensor* aclnn_zero(ggml_backend_cann_context& ctx, void* buffer, * @param type_size The size of each element in the tensor data type. * @param value The value to be used for initializing the tensor (default * is 1.0). - * @return An ACL tensor initialized with ones. + * @return An ACL tensor initialized with value. */ -static aclTensor* aclnn_ones(ggml_backend_cann_context& ctx, void* buffer, - size_t n_bytes, int64_t* ne, int64_t dims, - aclDataType type, size_t type_size, - float value = 1.0f) { +static aclTensor* aclnn_values(ggml_backend_cann_context& ctx, void* buffer, + size_t n_bytes, int64_t* ne, int64_t dims, + aclDataType type, size_t type_size, + float value = 1.0f) { aclTensor* acl_tensor = aclnn_zero(ctx, buffer, n_bytes, ne, dims, type, type_size); float alpha_host = 1.0f; @@ -1158,7 +1167,7 @@ void ggml_cann_rms_norm(ggml_backend_cann_context& ctx, ggml_tensor* dst) { size_t one_tensor_n_bytes = src->ne[0] * ggml_element_size(src); ggml_cann_pool_alloc one_tensor_allocator(ctx.pool(), one_tensor_n_bytes); - aclTensor* acl_gamma = aclnn_ones( + aclTensor* acl_gamma = aclnn_values( ctx, one_tensor_allocator.get(), one_tensor_n_bytes, src->ne, 1, ggml_cann_type_mapping(src->type), ggml_element_size(src)); @@ -1202,9 +1211,9 @@ void ggml_cann_diag_mask(ggml_backend_cann_context& ctx, ggml_tensor* dst, ggml_cann_pool_alloc one_tensor_allocator(ctx.pool(), one_tensor_n_bytes); aclTensor* mask_tensor = - aclnn_ones(ctx, one_tensor_allocator.get(), one_tensor_n_bytes, src->ne, - GGML_MAX_DIMS, ggml_cann_type_mapping(src->type), - ggml_element_size(src), value); + aclnn_values(ctx, one_tensor_allocator.get(), one_tensor_n_bytes, + src->ne, GGML_MAX_DIMS, ggml_cann_type_mapping(src->type), + ggml_element_size(src), value); uint64_t workspaceSize = 0; aclOpExecutor* executor; @@ -1437,10 +1446,6 @@ void ggml_cann_im2col(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ggml_tensor* src0 = dst->src[0]; // kernel ggml_tensor* src1 = dst->src[1]; // input - GGML_ASSERT(src0->type == GGML_TYPE_F16); - GGML_ASSERT(src1->type == GGML_TYPE_F32); - GGML_ASSERT(dst->type == GGML_TYPE_F16 || dst->type == GGML_TYPE_F32); - GGML_TENSOR_BINARY_OP_LOCALS; // aclnnIm2col only works on 2D. set s1, p1, d1 to 1 to perform 2D @@ -1462,9 +1467,6 @@ void ggml_cann_im2col(ggml_backend_cann_context& ctx, ggml_tensor* dst) { const int64_t OH = is_2D ? ne2 : 1; const int64_t OW = ne1; - GGML_ASSERT(nb00 == sizeof(ggml_fp16_t)); - GGML_ASSERT(nb10 == sizeof(float)); - // memory allocated increased to 3x when is_2D == false const int64_t n_bytes_factor = is_2D ? 1 : 3; @@ -1768,6 +1770,92 @@ static void aclnn_sin(ggml_backend_cann_context& ctx, aclTensor* acl_src, ACL_CHECK(aclnnSin(workspaceAddr, workspaceSize, executor, ctx.stream())); } +/** + * @brief Performs element-wise division of tensor1 by tensor2 , multiplies the + result by the scalar value and adds it to self . + * + * Performs element-wise division of tensor1 by tensor2, + * multiplies the result by the scalar value and adds it to self . + * The operation is defined as: + * \f[ + * \text{out}_i = \text{selft}_i + \text{value} \times + \frac{\text{tensor1}_i}{\text{tensor2}_i} + * \f] + + * @param ctx The context for the CANN backend operations. + * @param acl_self The source tensor on which the addcdiv function will be + applied. + * @param tensor1 Numerator tensor. + * @param tensor2 Denominator tensor. + * @param value The value to be used for coefficient. + */ +static void aclnn_inplace_addcdiv(ggml_backend_cann_context& ctx, + aclTensor* acl_self, aclTensor* tensor1, + aclTensor* tensor2, float value) { + uint64_t workspaceSize = 0; + aclOpExecutor* executor; + void* workspaceAddr = nullptr; + aclScalar* acl_value = aclCreateScalar(&value, aclDataType::ACL_FLOAT); + + ACL_CHECK(aclnnInplaceAddcdivGetWorkspaceSize( + acl_self, tensor1, tensor2, acl_value, &workspaceSize, &executor)); + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK(aclnnInplaceAddcdiv(workspaceAddr, workspaceSize, executor, + ctx.stream())); +} + +/** + * @brief Matrix division, optionally in-place. + * + * This function division each element of the source tensor `acl_src` by the + * tensor `acl_other` and stores the result in the destination tensor `acl_dst`. + * If `inplace` is true, `acl_dst` will not be used and the operation is + * performed in-place on `acl_src`. The operation is defined as: \f[ + * \text{dst}_i = \frac{\text{acl_src}_i}{\text{acl_other}_i} + * \f] + * + * @param ctx The context for the CANN backend operations. + * @param acl_src Numerator tensor.. + * @param acl_other Denominator tensor. + * @param acl_dst The destination tensor where the result will be stored if + * `inplace` is false. + * @param inplace Flag indicating whether to perform the operation in-place on + * `acl_src`. + */ +static void aclnn_div_tensor(ggml_backend_cann_context& ctx, aclTensor* acl_src, + aclTensor* acl_other, aclTensor* acl_dst, + bool inplace) { + uint64_t workspaceSize = 0; + aclOpExecutor* executor; + void* workspaceAddr = nullptr; + + if (inplace) { + ACL_CHECK(aclnnInplaceDivGetWorkspaceSize(acl_src, acl_other, + &workspaceSize, &executor)); + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK(aclnnInplaceDiv(workspaceAddr, workspaceSize, executor, + ctx.stream())); + } else { + ACL_CHECK(aclnnDivGetWorkspaceSize(acl_src, acl_other, acl_dst, + &workspaceSize, &executor)); + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK( + aclnnDiv(workspaceAddr, workspaceSize, executor, ctx.stream())); + } +} + void ggml_cann_timestep_embedding(ggml_backend_cann_context& ctx, ggml_tensor* dst) { const ggml_tensor* src = dst->src[0]; @@ -2311,7 +2399,16 @@ void ggml_cann_get_rows(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ctx.stream())); switch (src0->type) { - case GGML_TYPE_F32: + case GGML_TYPE_F32: { +#ifdef ASCEND_310P + // Special operation for get_row_f32 kernel of 310P: clear the + // content of dest data buffer when row is not aligned to 32 bytes + if ((src0->ne[0] % 8) != 0) { + size_t dst_len = src1->ne[0] * src1->ne[1] * src1->ne[2] * + src0->ne[0] * ggml_type_size(GGML_TYPE_F32); + ACL_CHECK(aclrtMemset((char*)dst->data, dst_len, 0, dst_len)); + } +#endif aclrtlaunch_ascendc_get_row_f32( 24, ctx.stream(), src0->data, src1->data, dst->data, ((ggml_tensor*)src0->extra)->ne, @@ -2320,7 +2417,19 @@ void ggml_cann_get_rows(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ((ggml_tensor*)src1->extra)->nb, ((ggml_tensor*)dst->extra)->ne, ((ggml_tensor*)dst->extra)->nb); break; - case GGML_TYPE_F16: + } + case GGML_TYPE_F16: { +#ifdef ASCEND_310P + // Special operation for get_row_f16 kernel of 310P: clear the + // content of dest data buffer when row is not aligned to 32 bytes + if ((src0->ne[0] % 16) != 0) { + size_t dst_len = + src1->ne[0] * src1->ne[1] * src1->ne[2] * src0->ne[0] * + ggml_type_size( + GGML_TYPE_F32); // out is also f32, even input is f16 + ACL_CHECK(aclrtMemset((char*)dst->data, dst_len, 0, dst_len)); + } +#endif aclrtlaunch_ascendc_get_row_f16( 24, ctx.stream(), src0->data, src1->data, dst->data, ((ggml_tensor*)src0->extra)->ne, @@ -2329,6 +2438,7 @@ void ggml_cann_get_rows(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ((ggml_tensor*)src1->extra)->nb, ((ggml_tensor*)dst->extra)->ne, ((ggml_tensor*)dst->extra)->nb); break; + } case GGML_TYPE_Q4_0: aclrtlaunch_ascendc_get_row_q4_0( 24, ctx.stream(), src0->data, src1->data, dst->data, @@ -2407,7 +2517,6 @@ static void aclnn_mat_mul(ggml_backend_cann_context& ctx, aclTensor* acl_input, aclTensor* acl_weight, aclTensor* acl_dst) { int8_t cube_math_type = 1; // ALLOW_FP32_DOWN_PRECISION, when input is // fp32, atlas a2 will transpose it to HFLOAT32. - uint64_t workspaceSize = 0; aclOpExecutor* executor; void* workspaceAddr = nullptr; @@ -2425,6 +2534,81 @@ static void aclnn_mat_mul(ggml_backend_cann_context& ctx, aclTensor* acl_input, aclnnMatmul(workspaceAddr, workspaceSize, executor, ctx.stream())); } +/** + * @brief Performs matrix multiplication of two 2D tensors. + * + * This function computes the matrix multiplication of the input tensor + * `acl_input` and the weight tensor `acl_weight`, and stores the result in the + * destination tensor `acl_dst`. + * The operation is defined as: + * \f[ + * \text {acl_dst}=\text {acl_input@acl_weight} + * \f] + * + * @param ctx The context for the CANN backend operations. + * @param acl_input The input tensor for the matrix multiplication. + * @param acl_weight The weight tensor for the matrix multiplication. + * @param acl_dst The destination tensor where the result of the matrix + * multiplication will be stored. + */ +static void aclnn_mat_mul_2d(ggml_backend_cann_context& ctx, + aclTensor* acl_input, aclTensor* acl_weight, + aclTensor* acl_dst) { + int8_t cube_math_type = 2; + uint64_t workspaceSize = 0; + aclOpExecutor* executor; + void* workspaceAddr = nullptr; + + ACL_CHECK(aclnnMmGetWorkspaceSize(acl_input, acl_weight, acl_dst, + cube_math_type, &workspaceSize, + &executor)); + + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK(aclnnMm(workspaceAddr, workspaceSize, executor, ctx.stream())); +} + +/** + * @brief Performs matrix multiplication of two 3D tensors. + * + * This function computes the matrix multiplication of the input tensor + * `acl_input` and the weight tensor `acl_weight`, and stores the result in the + * destination tensor `acl_dst`. + * The operation is defined as: + * \f[ + * \text {acl_dst}=\text {acl_input@acl_weight} + * \f] + * + * @param ctx The context for the CANN backend operations. + * @param acl_input The input tensor for the matrix multiplication. + * @param acl_weight The weight tensor for the matrix multiplication. + * @param acl_dst The destination tensor where the result of the matrix + * multiplication will be stored. + */ +static void aclnn_mat_mul_3d(ggml_backend_cann_context& ctx, + aclTensor* acl_input, aclTensor* acl_weight, + aclTensor* acl_dst) { + int8_t cube_math_type = 2; + uint64_t workspaceSize = 0; + aclOpExecutor* executor; + void* workspaceAddr = nullptr; + + ACL_CHECK(aclnnBatchMatMulGetWorkspaceSize(acl_input, acl_weight, acl_dst, + cube_math_type, &workspaceSize, + &executor)); + + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK( + aclnnBatchMatMul(workspaceAddr, workspaceSize, executor, ctx.stream())); +} + /** * @brief Performs matrix multiplication with floating-point precision on * tensors using the CANN backend. @@ -2446,20 +2630,39 @@ static void ggml_cann_mat_mul_fp(ggml_backend_cann_context& ctx, // broadcast, when weight ne2 or ne3 is not 1, weight need repeat. BCAST_MUL_MAT_SHAPE(input, weight, dst); - // transpose weight: [1,2,3,4] -> [1,2,4,3] + int64_t n_dims = bcast_dims; + if (bcast_input_ne[3] == bcast_weight_ne[3] && bcast_input_ne[3] == 1) { + if (bcast_input_ne[2] == 1 && bcast_weight_ne[2] == 1) { + n_dims = 2; + } else if (bcast_input_ne[2] == 1) { + n_dims = 3; + } + } + + aclTensor* acl_input_tensor = + ggml_cann_create_tensor(input, bcast_input_ne, bcast_input_nb, n_dims); int64_t transpose_ne[] = {bcast_weight_ne[1], bcast_weight_ne[0], bcast_weight_ne[2], bcast_weight_ne[3], bcast_weight_ne[4], bcast_weight_ne[5]}; size_t transpose_nb[] = {bcast_weight_nb[1], bcast_weight_nb[0], bcast_weight_nb[2], bcast_weight_nb[3], bcast_weight_nb[4], bcast_weight_nb[5]}; - aclTensor* acl_weight_tensor = - ggml_cann_create_tensor(weight, transpose_ne, transpose_nb, bcast_dims); - aclTensor* acl_input_tensor = - ggml_cann_create_tensor(input, BCAST_MUL_MAT_PARAM(input)); - aclTensor* acl_dst = ggml_cann_create_tensor(dst, BCAST_MUL_MAT_PARAM(dst)); - aclnn_mat_mul(ctx, acl_input_tensor, acl_weight_tensor, acl_dst); + ggml_cann_create_tensor(weight, transpose_ne, transpose_nb, n_dims); + aclTensor* acl_dst = + ggml_cann_create_tensor(dst, bcast_dst_ne, bcast_dst_nb, n_dims); + + switch (n_dims) { + case 2: + aclnn_mat_mul_2d(ctx, acl_input_tensor, acl_weight_tensor, acl_dst); + break; + case 3: + aclnn_mat_mul_3d(ctx, acl_input_tensor, acl_weight_tensor, acl_dst); + break; + default: + aclnn_mat_mul(ctx, acl_input_tensor, acl_weight_tensor, acl_dst); + break; + } ACL_CHECK(aclDestroyTensor(acl_weight_tensor)); ACL_CHECK(aclDestroyTensor(acl_input_tensor)); @@ -2480,51 +2683,47 @@ static void ggml_cann_mat_mul_fp(ggml_backend_cann_context& ctx, * multiplication will be stored. */ static void ggml_cann_mul_mat_quant(ggml_backend_cann_context& ctx, - ggml_tensor* dst, - const enum ggml_type type) { + ggml_tensor* dst, + const enum ggml_type type) { ggml_tensor* src0 = dst->src[0]; // weight ggml_tensor* src1 = dst->src[1]; // input - // The shape of the weight is NCHW. Matrix multiplication uses HW dims. HC - // is regarded as batch. weight need transpose. - int64_t weight_ne[] = {src0->ne[1], src0->ne[0]}; + // The shape of the weight is NCHW. + // Matrix multiplication uses HW dims. + // HC is regarded as batch. + // weight need transpose. float weight_elem_size; if (type == GGML_TYPE_Q4_0) { weight_elem_size = float(sizeof(uint8_t)) / 2; - } - else if (type == GGML_TYPE_Q8_0) { + } else if (type == GGML_TYPE_Q8_0) { weight_elem_size = float(sizeof(uint8_t)); - } - else { + } else { GGML_ABORT("Only support Q4_0 and Q8_0 MUL_MAT"); } - float weight_nb[] = {weight_elem_size * src0->ne[0], weight_elem_size}; - - // size of one matrix is element_size * height * width. - size_t weight_stride = weight_elem_size * src0->ne[0] * src0->ne[1]; + float weight_nb[] = {src0->ne[0] * weight_elem_size, weight_elem_size}; + size_t weight_stride = src0->ne[1] * src0->ne[0] * weight_elem_size; size_t weight_size = weight_stride * src0->ne[2] * src0->ne[3]; // scale stored at the end of weight. Also need transpose. - GGML_ASSERT(QK4_0 == QK8_0); - int64_t scale_ne[] = {src0->ne[1], src0->ne[0] / QK8_0}; size_t scale_elem_size = sizeof(uint16_t); size_t scale_nb[] = {src0->ne[0] / QK8_0 * scale_elem_size, scale_elem_size}; - size_t scale_stride = scale_elem_size * src0->ne[0] * src0->ne[1] / QK8_0; + size_t scale_stride = src0->ne[1] * src0->ne[0] / QK8_0 * scale_elem_size; char* scale_offset = (char*)src0->data + weight_size; // input - void* input_buffer; size_t input_elem_size = sizeof(uint16_t); int64_t input_ne[] = {src1->ne[0], src1->ne[1]}; - size_t input_nb[] = {input_elem_size, input_elem_size * src1->ne[0]}; - size_t input_stride = input_elem_size * src1->ne[0] * src1->ne[1]; - + size_t input_nb[] = {input_elem_size, input_ne[0] * input_elem_size}; + size_t input_stride = input_ne[0] * input_ne[1] * input_elem_size; ggml_cann_pool_alloc input_alloctor(ctx.pool()); + void* input_buffer = src1->data; + + // case in if (src1->type != GGML_TYPE_F16) { aclTensor* acl_src1_tensor = ggml_cann_create_tensor(src1); - input_alloctor.alloc(ggml_nelements(src1) * input_elem_size); - input_buffer = input_alloctor.get(); + input_buffer = + input_alloctor.alloc(ggml_nelements(src1) * input_elem_size); int64_t* input_cast_ne = src1->ne; size_t input_cast_nb[GGML_MAX_DIMS]; @@ -2537,85 +2736,136 @@ static void ggml_cann_mul_mat_quant(ggml_backend_cann_context& ctx, input_buffer, ACL_FLOAT16, input_elem_size, input_cast_ne, input_cast_nb, GGML_MAX_DIMS); aclnn_cast(ctx, acl_src1_tensor, acl_input_tensor, ACL_FLOAT16); + ACL_CHECK(aclDestroyTensor(acl_input_tensor)); ACL_CHECK(aclDestroyTensor(acl_src1_tensor)); - } else { - input_buffer = src1->data; } // output size_t output_elem_size = sizeof(uint16_t); - int64_t output_ne[] = {dst->ne[0], dst->ne[1]}; - size_t output_nb[] = {output_elem_size, output_elem_size * dst->ne[0]}; - ggml_cann_pool_alloc output_alloctor( - ctx.pool(), ggml_nelements(dst) * output_elem_size); - void* output_buffer = output_alloctor.get(); - size_t output_stride = output_elem_size * dst->ne[0] * dst->ne[1]; + size_t output_nb[] = {output_elem_size, dst->ne[0] * output_elem_size}; + ggml_cann_pool_alloc output_allocator(ctx.pool()); + void* output_buffer = + output_allocator.alloc(ggml_nelements(dst) * output_elem_size); + size_t output_stride = dst->ne[0] * dst->ne[1] * output_elem_size; // aclnn + int64_t max_elem_size = 65535; + int64_t split_size = (src0->ne[1] / max_elem_size) + 1; + ggml_cann_pool_alloc workspace_allocator(ctx.pool()); + aclOpExecutor* executor = nullptr; uint64_t workspaceSize = 0; - aclOpExecutor* executor; void* workspaceAddr = nullptr; - for (int64_t n1 = 0; n1 < src1->ne[3]; n1++) { for (int64_t c1 = 0; c1 < src1->ne[2]; c1++) { int64_t n0 = n1 / (src1->ne[3] / src0->ne[3]); int64_t c0 = c1 / (src1->ne[2] / src0->ne[2]); - int64_t batch1 = n1 * src1->ne[2] + c1; - int64_t batch0 = n0 * src0->ne[2] + c0; + int64_t batch1 = (n1 * src1->ne[2]) + c1; + int64_t batch0 = (n0 * src0->ne[2]) + c0; aclTensor* acl_input_tensor = ggml_cann_create_tensor( (char*)input_buffer + batch1 * input_stride, ACL_FLOAT16, input_elem_size, input_ne, input_nb, 2); + + // first split + int64_t weight_ne_offset = 0; + int64_t weight_ne[2] = { + max_elem_size > src0->ne[1] ? src0->ne[1] : max_elem_size, + src0->ne[0]}; + int64_t scale_ne_offset = 0; + int64_t scale_ne[2] = {weight_ne[0], weight_ne[1] / QK8_0}; + int64_t output_ne_offset = 0; + int64_t output_ne[2] = {weight_ne[0], dst->ne[1]}; + aclTensor* acl_weight_tensor = ggml_cann_create_tensor( (char*)src0->data + batch0 * weight_stride, ggml_cann_type_mapping(type), weight_elem_size, weight_ne, - weight_nb, 2); + weight_nb, 2, ACL_FORMAT_ND, weight_ne_offset); aclTensor* acl_scale_tensor = ggml_cann_create_tensor( scale_offset + batch0 * scale_stride, ACL_FLOAT16, - scale_elem_size, scale_ne, scale_nb, 2); + scale_elem_size, scale_ne, scale_nb, 2, ACL_FORMAT_ND, + scale_ne_offset); aclTensor* acl_output_tensor = ggml_cann_create_tensor( (char*)output_buffer + batch1 * output_stride, ACL_FLOAT16, - output_elem_size, output_ne, output_nb, 2); + output_elem_size, output_ne, output_nb, 2, ACL_FORMAT_ND, + output_ne_offset); ACL_CHECK(aclnnWeightQuantBatchMatmulV2GetWorkspaceSize( acl_input_tensor, acl_weight_tensor, acl_scale_tensor, nullptr, nullptr, nullptr, nullptr, QK8_0, acl_output_tensor, &workspaceSize, &executor)); - - if (workspaceSize > 0 && workspaceAddr == nullptr) { - ggml_cann_pool_alloc workspace_allocator(ctx.pool(), - workspaceSize); - workspaceAddr = workspace_allocator.get(); + if (workspaceAddr == nullptr) { + workspaceAddr = workspace_allocator.alloc(workspaceSize); } - ACL_CHECK(aclnnWeightQuantBatchMatmulV2( workspaceAddr, workspaceSize, executor, ctx.stream())); - ACL_CHECK(aclDestroyTensor(acl_input_tensor)); ACL_CHECK(aclDestroyTensor(acl_weight_tensor)); ACL_CHECK(aclDestroyTensor(acl_scale_tensor)); ACL_CHECK(aclDestroyTensor(acl_output_tensor)); + + // other splits + for (int64_t split = 1; split < split_size; split++) { + weight_ne_offset += + weight_elem_size * weight_ne[0] * weight_ne[1]; + weight_ne[0] = max_elem_size * (split + 1) > src0->ne[1] + ? src0->ne[1] - (max_elem_size * split) + : max_elem_size; + scale_ne_offset += scale_elem_size * scale_ne[0] * scale_ne[1]; + scale_ne[0] = weight_ne[0]; + output_ne_offset += + output_elem_size * output_ne[0] * output_ne[1]; + output_ne[0] = weight_ne[0]; + + acl_weight_tensor = ggml_cann_create_tensor( + (char*)src0->data + batch0 * weight_stride, + ggml_cann_type_mapping(type), weight_elem_size, weight_ne, + weight_nb, 2, ACL_FORMAT_ND, weight_ne_offset); + acl_scale_tensor = ggml_cann_create_tensor( + scale_offset + batch0 * scale_stride, ACL_FLOAT16, + scale_elem_size, scale_ne, scale_nb, 2, ACL_FORMAT_ND, + scale_ne_offset); + acl_output_tensor = ggml_cann_create_tensor( + (char*)output_buffer + batch1 * output_stride, ACL_FLOAT16, + output_elem_size, output_ne, output_nb, 2, ACL_FORMAT_ND, + output_ne_offset); + + ACL_CHECK(aclnnWeightQuantBatchMatmulV2GetWorkspaceSize( + acl_input_tensor, acl_weight_tensor, acl_scale_tensor, + nullptr, nullptr, nullptr, nullptr, QK8_0, + acl_output_tensor, &workspaceSize, &executor)); + ACL_CHECK(aclnnWeightQuantBatchMatmulV2( + workspaceAddr, workspaceSize, executor, ctx.stream())); + + ACL_CHECK(aclDestroyTensor(acl_weight_tensor)); + ACL_CHECK(aclDestroyTensor(acl_scale_tensor)); + ACL_CHECK(aclDestroyTensor(acl_output_tensor)); + } + + ACL_CHECK(aclDestroyTensor(acl_input_tensor)); } } // cast out - int64_t* output_cast_ne = dst->ne; - size_t output_cast_nb[GGML_MAX_DIMS]; - output_cast_nb[0] = sizeof(uint16_t); - for (int i = 1; i < GGML_MAX_DIMS; i++) { - output_cast_nb[i] = output_cast_nb[i - 1] * output_cast_ne[i - 1]; + if (dst->type != GGML_TYPE_F16) { + int64_t* output_cast_ne = dst->ne; + size_t output_cast_nb[GGML_MAX_DIMS]; + output_cast_nb[0] = sizeof(uint16_t); + for (int i = 1; i < GGML_MAX_DIMS; i++) { + output_cast_nb[i] = output_cast_nb[i - 1] * output_cast_ne[i - 1]; + } + + aclTensor* acl_output_tensor = ggml_cann_create_tensor( + output_buffer, ACL_FLOAT16, output_elem_size, output_cast_ne, + output_cast_nb, GGML_MAX_DIMS); + aclTensor* acl_dst_tensor = ggml_cann_create_tensor(dst); + aclnn_cast(ctx, acl_output_tensor, acl_dst_tensor, + ggml_cann_type_mapping(dst->type)); + + ACL_CHECK(aclDestroyTensor(acl_output_tensor)); + ACL_CHECK(aclDestroyTensor(acl_dst_tensor)); } - - aclTensor* acl_output_tensor = - ggml_cann_create_tensor(output_buffer, ACL_FLOAT16, output_elem_size, - output_cast_ne, output_cast_nb, GGML_MAX_DIMS); - aclTensor* acl_dst_tensor = ggml_cann_create_tensor(dst); - aclnn_cast(ctx, acl_output_tensor, acl_dst_tensor, ACL_FLOAT); - - ACL_CHECK(aclDestroyTensor(acl_output_tensor)); - ACL_CHECK(aclDestroyTensor(acl_dst_tensor)); } void ggml_cann_mul_mat(ggml_backend_cann_context& ctx, ggml_tensor* dst) { @@ -2714,12 +2964,14 @@ static void aclnn_index_fill_tensor(ggml_backend_cann_context& ctx, static void aclnn_cache_init(ggml_backend_cann_context& ctx, ggml_tensor* dst, aclTensor* acl_cos_repeat_tensor, aclTensor* acl_sin_repeat_tensor, - float theta_scale, bool is_neox) { + float theta_scale, float freq_scale, + float attn_factor, bool is_neox) { // int sin/cos cache, cache has different repeat method depond on // @param.is_neox ggml_tensor* src0 = dst->src[0]; // input ggml_tensor* src1 = dst->src[1]; // position + ggml_tensor* src2 = dst->src[2]; // freq_factors // arange, [0,1,...,ne0/2] int64_t arange_length = src0->ne[0] / 2; @@ -2748,11 +3000,26 @@ static void aclnn_cache_init(ggml_backend_cann_context& ctx, ggml_tensor* dst, ggml_cann_pool_alloc theta_scale_allocator(ctx.pool(), arange_length * sizeof(float_t)); void* theta_scale_buffer = theta_scale_allocator.get(); - aclTensor* acl_theta_scale_tensor = aclnn_ones( + aclTensor* acl_theta_scale_tensor = aclnn_values( ctx, theta_scale_buffer, arange_length * sizeof(float_t), arange_ne, GGML_MAX_DIMS, ACL_FLOAT, sizeof(float_t), theta_scale); aclnn_pow_tensor_tensor(ctx, acl_theta_scale_tensor, acl_arange_tensor); + // freq_scale + if (freq_scale != 1) { + aclnn_muls(ctx, acl_theta_scale_tensor, freq_scale, nullptr, true); + } + + // freq_factors + if (src2) { + aclTensor* acl_freq_factors_tensor = ggml_cann_create_tensor( + src2->data, ggml_cann_type_mapping(src2->type), + ggml_type_size(src2->type), arange_ne, arange_nb, GGML_MAX_DIMS); + aclnn_div_tensor(ctx, acl_theta_scale_tensor, acl_freq_factors_tensor, + nullptr, true); + ACL_CHECK(aclDestroyTensor(acl_freq_factors_tensor)); + } + // position GGML_ASSERT(src1->type == GGML_TYPE_I32); int64_t position_length = src1->ne[0]; @@ -2816,6 +3083,12 @@ static void aclnn_cache_init(ggml_backend_cann_context& ctx, ggml_tensor* dst, GGML_MAX_DIMS, ACL_FORMAT_ND); aclnn_cos(ctx, acl_permute_tensor, acl_cos_tensor); + // attn_factor + if (attn_factor != 1) { + aclnn_muls(ctx, acl_sin_tensor, attn_factor, nullptr, true); + aclnn_muls(ctx, acl_cos_tensor, attn_factor, nullptr, true); + } + // repeat if (is_neox) { int64_t repeatsArray[] = {1, 1, 1, 2}; @@ -2841,15 +3114,27 @@ static void aclnn_cache_init(ggml_backend_cann_context& ctx, ggml_tensor* dst, ACL_CHECK(aclDestroyTensor(acl_cos_tensor)); } +#ifdef __cplusplus +extern "C" { +#endif +aclnnStatus aclnnRotaryPositionEmbeddingGetWorkspaceSize( + const aclTensor* x, const aclTensor* cos, const aclTensor* sin, + int64_t mode, const aclTensor* yOut, uint64_t* workspaceSize, + aclOpExecutor** executor); +aclnnStatus aclnnRotaryPositionEmbedding(void* workspace, + uint64_t workspaceSize, + aclOpExecutor* executor, + aclrtStream stream); +#ifdef __cplusplus +} +#endif + void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { // TODO: use ascendc // Only test with LLAMA model. ggml_tensor* src0 = dst->src[0]; // input ggml_tensor* src2 = dst->src[2]; // freq_factors - // TODO: with freq_factors - GGML_ASSERT(src2 == NULL); - // param float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; // const int n_past = ((int32_t *) dst->op_params)[0]; @@ -2867,13 +3152,11 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { memcpy(&beta_fast, (int32_t*)dst->op_params + 9, sizeof(float)); memcpy(&beta_slow, (int32_t*)dst->op_params + 10, sizeof(float)); - GGML_ASSERT(n_dims <= ne0); + // TODO: n_dims <= ne0 + GGML_ASSERT(n_dims == ne0); GGML_ASSERT(n_dims % 2 == 0); - // TODO: ext_factor != 0 GGML_ASSERT(ext_factor == 0); - // TODO: freq_scale != 1 - GGML_ASSERT(freq_scale == 1); const float theta_scale = powf(freq_base, -2.0f / n_dims); @@ -2904,7 +3187,13 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ggml_cann_create_tensor(cos_buffer, ACL_FLOAT, sizeof(float_t), sin_reshape_ne, sin_reshape_nb, GGML_MAX_DIMS); aclnn_cache_init(ctx, dst, acl_cos_reshape_tensor, acl_sin_reshape_tensor, - theta_scale, is_neox); + theta_scale, freq_scale, attn_factor, is_neox); + + aclTensor* acl_src = ggml_cann_create_tensor(src0); + aclTensor* acl_dst = ggml_cann_create_tensor(dst); + +#ifdef ASCEND_310P + // Special ROPE operation for 310P // roll input void* input_roll_buffer; @@ -2947,7 +3236,7 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { for (int i = 1; i < GGML_MAX_DIMS; i++) { minus_one_nb[i] = minus_one_nb[i - 1] * minus_one_ne[i - 1]; } - acl_minus_one_tensor = aclnn_ones( + acl_minus_one_tensor = aclnn_values( ctx, minus_one_scale_buffer, sizeof(float_t) * src0->ne[0], minus_one_ne, GGML_MAX_DIMS, ACL_FLOAT, sizeof(float_t), 1); int64_t dim = 3; @@ -2974,17 +3263,15 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ACL_CHECK(aclDestroyTensor(acl_input_roll_tensor)); ACL_CHECK(aclDestroyTensor(acl_input_tensor)); - // init [-1, -1, -1, 1, 1，1，...] minus_one_scale_buffer = minus_one_scale_allocator.get(); - int64_t minus_one_ne[4] = {src0->ne[0], 1, 1, 1}; size_t minus_one_nb[GGML_MAX_DIMS]; minus_one_nb[0] = sizeof(float_t); for (int i = 1; i < GGML_MAX_DIMS; i++) { minus_one_nb[i] = minus_one_nb[i - 1] * minus_one_ne[i - 1]; } - acl_minus_one_tensor = aclnn_ones( + acl_minus_one_tensor = aclnn_values( ctx, minus_one_scale_buffer, sizeof(float_t) * src0->ne[0], minus_one_ne, GGML_MAX_DIMS, ACL_FLOAT, sizeof(float_t), 1); // -1 * first half @@ -3026,14 +3313,12 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { acl_input_roll_mul_scale_tensor); // output - aclTensor* acl_src0 = ggml_cann_create_tensor(src0); - aclTensor* acl_dst = ggml_cann_create_tensor(dst); void* output_fp32_buffer; if (src0->type == GGML_TYPE_F32) { - aclnn_inplace_mul(ctx, acl_src0, acl_cos_reshape_tensor); + aclnn_inplace_mul(ctx, acl_src, acl_cos_reshape_tensor); aclnn_inplace_mul(ctx, acl_input_roll_mul_scale_tensor, acl_sin_reshape_tensor); - aclnn_add(ctx, acl_src0, acl_input_roll_mul_scale_tensor, acl_dst); + aclnn_add(ctx, acl_src, acl_input_roll_mul_scale_tensor, acl_dst); // TODO: ne0 != n_dims in mode2 } else if (src0->type == GGML_TYPE_F16) { size_t input_fp32_nb[GGML_MAX_DIMS]; @@ -3060,7 +3345,7 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { aclTensor* output_fp32_tensor = ggml_cann_create_tensor( output_fp32_buffer, ACL_FLOAT, sizeof(float_t), dst->ne, input_fp32_nb, GGML_MAX_DIMS); - aclnn_mul(ctx, acl_src0, acl_cos_reshape_tensor, input_fp32_tensor1); + aclnn_mul(ctx, acl_src, acl_cos_reshape_tensor, input_fp32_tensor1); aclnn_mul(ctx, acl_input_roll_mul_scale_tensor, acl_sin_reshape_tensor, input_fp32_tensor2); aclnn_add(ctx, input_fp32_tensor1, input_fp32_tensor2, @@ -3070,13 +3355,73 @@ void ggml_cann_rope(ggml_backend_cann_context& ctx, ggml_tensor* dst) { ACL_CHECK(aclDestroyTensor(input_fp32_tensor1)); ACL_CHECK(aclDestroyTensor(input_fp32_tensor2)); ACL_CHECK(aclDestroyTensor(output_fp32_tensor)); + ACL_CHECK(aclDestroyTensor(acl_sin_reshape_tensor)); + ACL_CHECK(aclDestroyTensor(acl_minus_one_tensor)); + ACL_CHECK(aclDestroyTensor(acl_input_roll_mul_scale_tensor)); + ACL_CHECK(aclDestroyTensor(acl_input_roll_reshape_tensor)); + ACL_CHECK(aclDestroyTensor(acl_src)); + } + return; +#endif + + // src0 == GGML_TYPE_F16 + // TODO: optimization this `if` code + if (src0->type == GGML_TYPE_F16) { + ggml_cann_pool_alloc sin_final_allocator( + ctx.pool(), src0->ne[0] * src0->ne[2] * ggml_type_size(src0->type)); + ggml_cann_pool_alloc cos_final_allocator( + ctx.pool(), src0->ne[0] * src0->ne[2] * ggml_type_size(src0->type)); + void* sin_final_buffer = sin_final_allocator.get(); + void* cos_final_buffer = cos_final_allocator.get(); + + int64_t sin_final_ne[4] = {src0->ne[0], 1, src0->ne[2], 1}; + size_t sin_final_nb[GGML_MAX_DIMS]; + sin_final_nb[0] = ggml_type_size(src0->type); + for (int i = 1; i < GGML_MAX_DIMS; i++) { + sin_final_nb[i] = sin_final_nb[i - 1] * sin_final_ne[i - 1]; + } + aclTensor* acl_sin_final_tensor = ggml_cann_create_tensor( + sin_final_buffer, ggml_cann_type_mapping(src0->type), + ggml_type_size(src0->type), sin_final_ne, sin_final_nb, + GGML_MAX_DIMS); + aclTensor* acl_cos_final_tensor = ggml_cann_create_tensor( + cos_final_buffer, ggml_cann_type_mapping(src0->type), + ggml_type_size(src0->type), sin_final_ne, sin_final_nb, + GGML_MAX_DIMS); + + aclnn_cast(ctx, acl_sin_reshape_tensor, acl_sin_final_tensor, + ggml_cann_type_mapping(src0->type)); + aclnn_cast(ctx, acl_cos_reshape_tensor, acl_cos_final_tensor, + ggml_cann_type_mapping(src0->type)); + ACL_CHECK(aclDestroyTensor(acl_cos_reshape_tensor)); + ACL_CHECK(aclDestroyTensor(acl_sin_reshape_tensor)); + acl_sin_reshape_tensor = acl_sin_final_tensor; + acl_cos_reshape_tensor = acl_cos_final_tensor; } - ACL_CHECK(aclDestroyTensor(acl_sin_reshape_tensor)); + uint64_t workspaceSize = 0; + aclOpExecutor* executor; + + void* workspaceAddr = nullptr; + + int acl_mode = mode; + if (mode == 0) { + acl_mode = 1; + } + + ACL_CHECK(aclnnRotaryPositionEmbeddingGetWorkspaceSize( + acl_src, acl_cos_reshape_tensor, acl_sin_reshape_tensor, acl_mode, + acl_dst, &workspaceSize, &executor)); + if (workspaceSize > 0) { + ggml_cann_pool_alloc workspace_allocator(ctx.pool(), workspaceSize); + workspaceAddr = workspace_allocator.get(); + } + + ACL_CHECK(aclnnRotaryPositionEmbedding(workspaceAddr, workspaceSize, + executor, ctx.stream())); + + ACL_CHECK(aclDestroyTensor(acl_src)); ACL_CHECK(aclDestroyTensor(acl_cos_reshape_tensor)); - ACL_CHECK(aclDestroyTensor(acl_minus_one_tensor)); - ACL_CHECK(aclDestroyTensor(acl_input_roll_mul_scale_tensor)); - ACL_CHECK(aclDestroyTensor(acl_input_roll_reshape_tensor)); - ACL_CHECK(aclDestroyTensor(acl_src0)); + ACL_CHECK(aclDestroyTensor(acl_sin_reshape_tensor)); ACL_CHECK(aclDestroyTensor(acl_dst)); } diff --git a/ggml/src/ggml-cann/common.h b/ggml/src/ggml-cann/common.h index edfa49614..5164cb74e 100644 --- a/ggml/src/ggml-cann/common.h +++ b/ggml/src/ggml-cann/common.h @@ -211,17 +211,20 @@ struct ggml_cann_pool_alloc { struct ggml_backend_cann_context { int32_t device; /**< Device ID. */ std::string name; /**< Name of the device. */ + std::string description; /**< Description of the device. */ aclrtEvent copy_event = nullptr; /**< Event for managing copy operations. */ - aclrtStream streams[GGML_CANN_MAX_STREAMS] = { - {nullptr}}; /**< Array of streams for the device. */ + aclrtStream streams[GGML_CANN_MAX_STREAMS] = {nullptr}; /**< Array of streams for the device. */ /** * @brief Constructor for initializing the context with a given device. * @param device Device ID. */ explicit ggml_backend_cann_context(int device) - : device(device), name("CANN" + std::to_string(device)) {} + : device(device), name("CANN" + std::to_string(device)) { + ggml_cann_set_device(device); + description = aclrtGetSocName(); + } /** * @brief Destructor for cleaning up resources. diff --git a/ggml/src/ggml-cann.cpp b/ggml/src/ggml-cann/ggml-cann.cpp similarity index 95% rename from ggml/src/ggml-cann.cpp rename to ggml/src/ggml-cann/ggml-cann.cpp index 776340881..d410c0244 100644 --- a/ggml/src/ggml-cann.cpp +++ b/ggml/src/ggml-cann/ggml-cann.cpp @@ -122,6 +122,10 @@ static ggml_cann_device_info ggml_cann_init() { ACL_CHECK(aclrtMemGetAllocationGranularity( &prop, ACL_RT_MEM_ALLOC_GRANULARITY_RECOMMENDED, &info.devices[id].vmm_granularity)); + + size_t free, total; + ggml_backend_cann_get_device_memory(id, &free, &total); + info.devices[id].total_vram = free; } // TODO: add more device info later. @@ -208,6 +212,11 @@ struct ggml_cann_pool_leg : public ggml_cann_pool { * @return A pointer to the allocated buffer. */ void* alloc(size_t size, size_t* actual_size) override { + const size_t alignment = 128; + size = GGML_PAD(size, alignment); + if (size == 0) { + size = alignment; + } #ifdef DEBUG_CANN_MALLOC int nnz = 0; size_t max_size = 0; @@ -246,13 +255,11 @@ struct ggml_cann_pool_leg : public ggml_cann_pool { return ptr; } void* ptr; - size_t look_ahead_size = (size_t)(1.05 * size); - look_ahead_size = 256 * ((look_ahead_size + 255) / 256); ggml_cann_set_device(device); ACL_CHECK( - aclrtMalloc(&ptr, look_ahead_size, ACL_MEM_MALLOC_HUGE_FIRST)); - *actual_size = look_ahead_size; - pool_size += look_ahead_size; + aclrtMalloc(&ptr, size, ACL_MEM_MALLOC_HUGE_FIRST)); + *actual_size = size; + pool_size += size; #ifdef DEBUG_CANN_MALLOC GGML_LOG_INFO( "%s[%d]: %d buffers, max_size = %u MB, pool_size = %u MB, " @@ -296,7 +303,7 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { /** * @brief The maximum size of the virtual memory pool (32 GB). */ - static const size_t CANN_POOL_VMM_MAX_SIZE = 1ull << 35; // 32 GB + size_t max_size; /** * @brief The device ID associated with this buffer pool. @@ -341,7 +348,11 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { */ explicit ggml_cann_pool_vmm(int device) : device(device), - granularity(ggml_cann_info().devices[device].vmm_granularity) {} + granularity(ggml_cann_info().devices[device].vmm_granularity) { + auto dev = ggml_cann_info().devices[device]; + granularity = dev.vmm_granularity; + max_size = dev.total_vram; + } /** * @brief Destructor to free all buffers in the virtual memory pool. @@ -370,17 +381,19 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { // round up the allocation size to the alignment to ensure that all // allocations are aligned for all data types const size_t alignment = 128; - size = alignment * ((size + alignment - 1) / alignment); + size = GGML_PAD(size, alignment); + if (size == 0) { + size = alignment; + } size_t avail = pool_size - pool_used; if (size > avail) { // round up to the next multiple of the granularity size_t reserve_size = size - avail; - reserve_size = - granularity * ((reserve_size + granularity - 1) / granularity); + reserve_size = GGML_PAD(reserve_size, granularity); - GGML_ASSERT(pool_size + reserve_size <= CANN_POOL_VMM_MAX_SIZE); + GGML_ASSERT(pool_size + reserve_size <= max_size); // allocate more physical memory aclrtPhysicalMemProp prop = {}; @@ -396,7 +409,7 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { // reserve virtual address space (if not already reserved) if (pool_addr == 0) { ACL_CHECK(aclrtReserveMemAddress( - &pool_addr, CANN_POOL_VMM_MAX_SIZE, 0, NULL, 1)); + &pool_addr, max_size, 0, NULL, 1)); } // map at the end of the pool @@ -409,10 +422,11 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { // add to the pool pool_size += reserve_size; - // GGML_LOG_INFO("cann pool[%d]: size increased to %llu MB ( - // reserved %llu MB)\n", - // device, (unsigned long long) (pool_size/1024/1024), - // (unsigned long long) (reserve_size/1024/1024)); +#ifdef DEBUG_CANN_MALLOC + GGML_LOG_INFO("cann pool[%d]: size increased to %llu MB (reserved %llu MB)\n", + device, (unsigned long long) (pool_size/1024/1024), + (unsigned long long) (reserve_size/1024/1024)); +#endif } GGML_ASSERT(pool_addr != 0); @@ -457,7 +471,6 @@ struct ggml_cann_pool_vmm : public ggml_cann_pool { */ std::unique_ptr ggml_backend_cann_context::new_pool_for_device( int device) { - // return std::unique_ptr(new ggml_cann_pool_leg(device)); return std::unique_ptr(new ggml_cann_pool_vmm(device)); } @@ -1130,10 +1143,10 @@ ggml_backend_cann_buffer_type(int32_t device) { static bool ggml_backend_cann_buffer_type_initialized = false; if (!ggml_backend_cann_buffer_type_initialized) { - for (int32_t i = 0; i < GGML_CANN_MAX_DEVICES; i++) { + for (int32_t i = 0; i < ggml_cann_info().device_count; i++) { ggml_backend_cann_buffer_types[i] = { /* .iface = */ ggml_backend_cann_buffer_type_interface, - /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cann_reg(), device), + /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cann_reg(), i), /* .context = */ new ggml_backend_cann_buffer_type_context{ i, "CANN" + std::to_string(i)}, @@ -1199,10 +1212,15 @@ static void * ggml_cann_host_malloc(size_t size) { return nullptr; } + const size_t alignment = 128; + size = GGML_PAD(size, alignment); + if (size == 0) { + size = alignment; + } + void * hostPtr = nullptr; aclError err = aclrtMallocHost((void **) &hostPtr, size); if (err != ACL_SUCCESS) { - GGML_LOG_WARN("%s: failed to allocate %.2f MiB of pinned memory: %s\n", __func__, size / 1024.0 / 1024.0, aclGetRecentErrMsg()); return nullptr; @@ -1669,12 +1687,14 @@ static bool ggml_backend_cann_supports_op(ggml_backend_dev_t dev, } case GGML_OP_MUL_MAT: { switch (op->src[0]->type) { + case GGML_TYPE_Q8_0: + // Current groupsize should not be greater than k-1 in + // aclnnWeightQuantBatchMatmulV2GetWorkspaceSize + if (op->src[0]->ne[0] <= QK8_0) { + return false; + } case GGML_TYPE_F16: case GGML_TYPE_F32: - case GGML_TYPE_Q8_0: - // TODO: fix me - // Current groupsize should not be greater than k-1 in - // aclnnWeightQuantBatchMatmulV2GetWorkspaceSize(). case GGML_TYPE_Q4_0: return true; default: @@ -1706,9 +1726,50 @@ static bool ggml_backend_cann_supports_op(ggml_backend_dev_t dev, return false; } } + case GGML_OP_CONT: { + // TODO: support GGML_TYPE_BF16 + switch (op->src[0]->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + return true; + default: + return false; + } + } + case GGML_OP_ROPE: { + // TODO: with ops-test v == 1 + float * ext_factor = (float*)((int32_t*)op->op_params + 7); + // TODO: n_dims <= ne0 + if (op->src[0]->ne[0] != op->op_params[1]) { + return false; + } + // TODO: ext_factor != 0 + if (*ext_factor != 0) { + return false; + } + + const int mode = ((const int32_t *) op->op_params)[2]; + if (mode & GGML_ROPE_TYPE_MROPE) { + return false; + } + if (mode & GGML_ROPE_TYPE_VISION) { + return false; + } + + return true; + } + case GGML_OP_UPSCALE: { + // aclnnUpsampleNearest2dGetWorkspaceSize not support + // selfDimN[2]/outDimN[2] or selfDimC[3]/outDimC[3] not equal + if (op->src[0]->ne[2] * op->ne[3] != op->src[0]->ne[3] * op->ne[2]) { + return false; + } + return true; + } + case GGML_OP_IM2COL: + case GGML_OP_CONCAT: case GGML_OP_DUP: case GGML_OP_REPEAT: - case GGML_OP_CONCAT: case GGML_OP_NONE: case GGML_OP_RESHAPE: case GGML_OP_VIEW: @@ -1722,17 +1783,13 @@ static bool ggml_backend_cann_supports_op(ggml_backend_dev_t dev, case GGML_OP_SCALE: case GGML_OP_SQR: case GGML_OP_CLAMP: - case GGML_OP_CONT: case GGML_OP_DIAG_MASK_INF: case GGML_OP_SOFT_MAX: - case GGML_OP_ROPE: - case GGML_OP_IM2COL: case GGML_OP_POOL_2D: case GGML_OP_SUM_ROWS: case GGML_OP_ARGSORT: case GGML_OP_ACC: case GGML_OP_GROUP_NORM: - case GGML_OP_UPSCALE: case GGML_OP_PAD: case GGML_OP_ARANGE: case GGML_OP_TIMESTEP_EMBEDDING: @@ -2041,7 +2098,7 @@ static void * ggml_backend_cann_reg_get_proc_address(ggml_backend_reg_t reg, con static const ggml_backend_reg_i ggml_backend_cann_reg_interface = { /* .get_name = */ ggml_backend_cann_reg_get_name, /* .get_device_count = */ ggml_backend_cann_reg_get_device_count, - /* .get_device_get = */ ggml_backend_cann_reg_get_device, + /* .get_device = */ ggml_backend_cann_reg_get_device, /* .get_proc_address = */ ggml_backend_cann_reg_get_proc_address, }; @@ -2064,16 +2121,17 @@ ggml_backend_reg_t ggml_backend_cann_reg() { dev_ctx->name = GGML_CANN_NAME + std::to_string(i); ggml_cann_set_device(i); ggml_backend_dev_t dev = new ggml_backend_device { - /* .interface = */ ggml_backend_cann_device_interface, - /* .reg = */ ®, - /* .context = */ dev_ctx + /* .iface = */ ggml_backend_cann_device_interface, + /* .reg = */ ®, + /* .context = */ dev_ctx }; ctx->devices.push_back(dev); } reg = ggml_backend_reg { - /* .interface = */ ggml_backend_cann_reg_interface, - /* .context = */ ctx + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_cann_reg_interface, + /* .context = */ ctx }; } @@ -2126,3 +2184,5 @@ void ggml_backend_cann_get_device_memory(int32_t device, size_t* free, ggml_cann_set_device(device); ACL_CHECK(aclrtGetMemInfo(ACL_HBM_MEM, free, total)); } + +GGML_BACKEND_DL_IMPL(ggml_backend_cann_reg) diff --git a/ggml/src/ggml-cann/kernels/CMakeLists.txt b/ggml/src/ggml-cann/kernels/CMakeLists.txt index 5b4fef91b..d687220c3 100644 --- a/ggml/src/ggml-cann/kernels/CMakeLists.txt +++ b/ggml/src/ggml-cann/kernels/CMakeLists.txt @@ -1,7 +1,3 @@ -if (NOT SOC_TYPE) - set (SOC_TYPE "Ascend910B3") -endif() - file(GLOB SRC_FILES get_row_f32.cpp get_row_f16.cpp @@ -13,7 +9,6 @@ file(GLOB SRC_FILES dup.cpp ) -string(TOLOWER ${SOC_TYPE} SOC_VERSION) set(ASCEND_CANN_PACKAGE_PATH ${CANN_INSTALL_DIR}) set(RUN_MODE "npu" CACHE STRING "run mode: npu/sim") @@ -30,4 +25,6 @@ ascendc_library(ascendc_kernels STATIC ${SRC_FILES} ) +message(STATUS "CANN: compile ascend kernels witch SOC_TYPE:${SOC_TYPE}, SOC_VERSION:${SOC_VERSION}, compile macro:-D${SOC_TYPE_COMPILE_OPTION}.") +ascendc_compile_definitions(ascendc_kernels PRIVATE "-D${SOC_TYPE_COMPILE_OPTION}") # ascendc_compile_definitions(ascendc_kernels PRIVATE -DASCENDC_DUMP) diff --git a/ggml/src/ggml-cann/kernels/dup.cpp b/ggml/src/ggml-cann/kernels/dup.cpp index e2c651152..c7ba38d10 100644 --- a/ggml/src/ggml-cann/kernels/dup.cpp +++ b/ggml/src/ggml-cann/kernels/dup.cpp @@ -5,6 +5,7 @@ using namespace AscendC; #define BUFFER_NUM 2 +const int64_t SUPPORTED_MAX_DIM = 65535; // currently the limit of max block dim supportted by dup kernel is 65535template template class DupByRows { @@ -51,24 +52,36 @@ class DupByRows { __aicore__ inline void copy_in() { LocalTensor src_local = src_queue.AllocTensor(); - - DataCopyExtParams dataCopyParams; - dataCopyParams.blockCount = 1; - dataCopyParams.blockLen = num_elem * sizeof(SRC_T); - DataCopyPadExtParams padParams; - DataCopyPad(src_local, src_gm, dataCopyParams, padParams); - + const size_t elem_per_block = 32 / sizeof(SRC_T); + size_t tail = num_elem % elem_per_block; + size_t cpy_elements_len = tail > 0 ? num_elem + 1 : num_elem; + DataCopy(src_local, src_gm, cpy_elements_len); src_queue.EnQue(src_local); } __aicore__ inline void copy_out() { LocalTensor dst_local = dst_queue.DeQue(); - +#ifdef ASCEND_310P + const size_t elem_per_block = 32 / sizeof(DST_T); + size_t tail = num_elem % elem_per_block; + size_t len = num_elem & ~(elem_per_block - 1); + if (len > 0) { + DataCopy(dst_gm, dst_local, len); + } + if(tail != 0) { + for (size_t i = tail; i < elem_per_block; i++) { + dst_local[len + i].SetValue(0, 0); + } + SetAtomicAdd(); + DataCopy(dst_gm[len], dst_local[len], elem_per_block); + SetAtomicNone(); + } +#else DataCopyExtParams dataCopyParams; dataCopyParams.blockCount = 1; dataCopyParams.blockLen = num_elem * sizeof(DST_T); DataCopyPad(dst_gm, dst_local, dataCopyParams); - +#endif dst_queue.FreeTensor(dst_local); } diff --git a/ggml/src/ggml-cann/kernels/get_row_f16.cpp b/ggml/src/ggml-cann/kernels/get_row_f16.cpp index c704b5b2e..416b45104 100644 --- a/ggml/src/ggml-cann/kernels/get_row_f16.cpp +++ b/ggml/src/ggml-cann/kernels/get_row_f16.cpp @@ -14,7 +14,7 @@ class GET_ROW_F16 { int64_t *output_ne_ub, size_t *output_nb_ub) { // TODO, use template for F16/f32 int64_t op_block_num = GetBlockNum(); - int64_t op_block_idx = GetBlockIdx(); + op_block_idx = GetBlockIdx(); for (int i = 0; i < 4; i++) { input_ne[i] = input_ne_ub[i]; @@ -59,32 +59,42 @@ class GET_ROW_F16 { } __aicore__ inline void copy_in(uint32_t offset, size_t len) { + size_t origin_len = len; LocalTensor input_local = input_queue.AllocTensor(); - size_t tail = len % 32; - len = len & ~31; - DataCopy(input_local, input_gm[offset], len); + const size_t elem_per_block = 32 / sizeof(half); + size_t tail = len % elem_per_block; + len = len & ~(elem_per_block - 1); if(tail != 0) { - DataCopyExtParams dataCopyParams; - dataCopyParams.blockCount = 1; - dataCopyParams.blockLen = tail * sizeof(half); - DataCopyPadExtParams padParams; - DataCopyPad(input_local[len], input_gm[offset + len], - dataCopyParams, padParams); + len += elem_per_block; } + DataCopy(input_local, input_gm[offset], len); input_queue.EnQue(input_local); } __aicore__ inline void copy_out(uint32_t offset, size_t len) { LocalTensor output_local = output_queue.DeQue(); - size_t tail = len % 32; - len = len & ~31; - DataCopy(output_gm[offset], output_local, len); + const size_t elem_per_block = 32 / sizeof(float); + size_t tail = len % elem_per_block; + len = len & ~(elem_per_block - 1); + if (len > 0) { + DataCopy(output_gm[offset], output_local, len); + } + if(tail != 0) { +#ifdef ASCEND_310P + for (size_t i = tail; i < elem_per_block; i++) { + output_local[len + i].SetValue(0, 0); + } + SetAtomicAdd(); + DataCopy(output_gm[offset + len], output_local[len], elem_per_block); + SetAtomicNone(); +#else DataCopyExtParams dataCopyParams; dataCopyParams.blockCount = 1; dataCopyParams.blockLen = tail * sizeof(float); DataCopyPad(output_gm[offset + len], output_local[len], dataCopyParams); +#endif } output_queue.FreeTensor(output_local); } @@ -150,6 +160,7 @@ class GET_ROW_F16 { GlobalTensor output_gm; TQue input_queue; TQue output_queue; + int64_t op_block_idx; }; template diff --git a/ggml/src/ggml-cann/kernels/get_row_f32.cpp b/ggml/src/ggml-cann/kernels/get_row_f32.cpp index 9db080af3..02116905b 100644 --- a/ggml/src/ggml-cann/kernels/get_row_f32.cpp +++ b/ggml/src/ggml-cann/kernels/get_row_f32.cpp @@ -13,7 +13,7 @@ class GET_ROW_F32 { int64_t *indices_ne_ub, size_t *indices_nb_ub, int64_t *output_ne_ub, size_t *output_nb_ub) { int64_t op_block_num = GetBlockNum(); - int64_t op_block_idx = GetBlockIdx(); + op_block_idx = GetBlockIdx(); for (int i = 0; i < 4; i++) { input_ne[i] = input_ne_ub[i]; @@ -55,31 +55,40 @@ class GET_ROW_F32 { __aicore__ inline void copy_in(uint32_t offset, size_t len) { LocalTensor input_local = input_queue.AllocTensor(); - size_t tail = len % 32; - len = len & ~31; - DataCopy(input_local, input_gm[offset], len); + const size_t elem_per_block = 32 / sizeof(float); + size_t tail = len % elem_per_block; + len = len & ~(elem_per_block - 1); if(tail != 0) { - DataCopyExtParams dataCopyParams; - dataCopyParams.blockCount = 1; - dataCopyParams.blockLen = tail * sizeof(float); - DataCopyPadExtParams padParams; - DataCopyPad(input_local[len], input_gm[offset + len], - dataCopyParams, padParams); + len += elem_per_block; } + DataCopy(input_local, input_gm[offset], len); input_queue.EnQue(input_local); } __aicore__ inline void copy_out(uint32_t offset, size_t len) { LocalTensor output_local = output_queue.DeQue(); - size_t tail = len % 32; - len = len & ~31; - DataCopy(output_gm[offset], output_local, len); + const size_t elem_per_block = 32 / sizeof(float); + size_t tail = len % elem_per_block; + len = len & ~(elem_per_block - 1); + if (len > 0) { + DataCopy(output_gm[offset], output_local, len); + } + if(tail != 0) { +#ifdef ASCEND_310P + for (size_t i = tail; i < elem_per_block; i++) { + output_local[len + i].SetValue(0, 0); + } + SetAtomicAdd(); + DataCopy(output_gm[offset + len], output_local[len], elem_per_block); + SetAtomicNone(); +#else DataCopyExtParams dataCopyParams; dataCopyParams.blockCount = 1; dataCopyParams.blockLen = tail * sizeof(float); DataCopyPad(output_gm[offset + len], output_local[len], dataCopyParams); +#endif } output_queue.FreeTensor(output_local); } @@ -144,6 +153,7 @@ class GET_ROW_F32 { GlobalTensor output_gm; TQue input_queue; TQue output_queue; + int64_t op_block_idx; }; template diff --git a/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp b/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp index a80bfeec2..4fbe72208 100644 --- a/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp +++ b/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp @@ -2,6 +2,15 @@ // optimize me. Use template to avoid copy code. using namespace AscendC; +#ifdef ASCEND_310P // 310P not support 4bit get row + extern "C" __global__ __aicore__ void ascendc_get_row_q4_0( + GM_ADDR input_gm, GM_ADDR indices_gm, GM_ADDR output_gm, + GM_ADDR input_ne_gm, GM_ADDR indices_ne_gm, GM_ADDR indices_nb_gm, + GM_ADDR output_ne_gm, GM_ADDR output_nb_gm) { + // let following test cases can continue run, here just print error information. Of Cource the test case that call this operator is failed. + printf("Ascend310P not support 4bit get row.\n"); + } +#else #define BUFFER_NUM 2 @@ -191,3 +200,5 @@ extern "C" __global__ __aicore__ void ascendc_get_row_q4_0( indices_nb_ub, output_ne_ub, output_nb_ub); op.calculate(); } + +#endif // #ifdef ASCEND_310P diff --git a/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp b/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp index 8423b3f02..504b43afa 100644 --- a/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp +++ b/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp @@ -1,6 +1,14 @@ #include "kernel_operator.h" using namespace AscendC; +#ifdef ASCEND_310P + extern "C" __global__ __aicore__ void ascendc_quantize_f16_q8_0( + GM_ADDR input_gm, GM_ADDR output_gm, GM_ADDR input_ne_gm, + GM_ADDR input_nb_gm, GM_ADDR output_ne_gm) { + // let following test cases can continue run, here just print error information. Of Cource the test case that call this operator is failed. + printf("Ascend310P not support f16->8bit quantization.\n"); + } +#else #define BUFFER_NUM 2 #define QK8_0 32 @@ -206,3 +214,5 @@ extern "C" __global__ __aicore__ void ascendc_quantize_f16_q8_0( op.init(input_gm, output_gm, input_ne_ub, input_nb_ub, output_ne_ub); op.calculate(); } + +#endif // #ifdef ASCEND_310P diff --git a/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp b/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp index b7c575093..05b0bc1df 100644 --- a/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp +++ b/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp @@ -1,6 +1,14 @@ #include "kernel_operator.h" using namespace AscendC; +#ifdef ASCEND_310P // 310P not support f32->8bit quantization + extern "C" __global__ __aicore__ void ascendc_quantize_f32_q8_0( + GM_ADDR input_gm, GM_ADDR output_gm, GM_ADDR input_ne_gm, + GM_ADDR input_nb_gm, GM_ADDR output_ne_gm) { + // let following test cases can continue run, here just print error information. Of Cource the test case that call this operator is failed. + printf("Ascend310P not support f32->8bit quantization.\n"); + } +#else #define BUFFER_NUM 2 #define QK8_0 32 @@ -204,3 +212,5 @@ extern "C" __global__ __aicore__ void ascendc_quantize_f32_q8_0( op.init(input_gm, output_gm, input_ne_ub, input_nb_ub, output_ne_ub); op.calculate(); } + +#endif // #ifdef ASCEND_310P diff --git a/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp b/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp index 9c8c86b66..1188937b7 100644 --- a/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp +++ b/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp @@ -1,6 +1,21 @@ #include "kernel_operator.h" using namespace AscendC; +#ifdef ASCEND_310P // 310P not support float->4bit quantization + extern "C" __global__ __aicore__ void ascendc_quantize_f32_to_q4_0( + GM_ADDR input_gm, GM_ADDR output_gm, GM_ADDR input_ne_gm, + GM_ADDR input_nb_gm, GM_ADDR output_ne_gm) { + // let following test cases can continue run, here just print error information. Of Cource the test case that call this operator is failed. + printf("Ascend310P not support f32->4bit quantization.\n"); + } + + extern "C" __global__ __aicore__ void ascendc_quantize_f16_to_q4_0( + GM_ADDR input_gm, GM_ADDR output_gm, GM_ADDR input_ne_gm, + GM_ADDR input_nb_gm, GM_ADDR output_ne_gm) { + // let following test cases can continue run, here just print error information. Of Cource the test case that call this operator is failed. + printf("Ascend310P not support f16->4bit quantization.\n"); + } +#else #define BUFFER_NUM 2 #define Group_Size 32 @@ -276,3 +291,5 @@ extern "C" __global__ __aicore__ void ascendc_quantize_f32_to_q4_0( op.init(input_gm, output_gm, input_ne_ub, input_nb_ub, output_ne_ub); op.calculate(); } + +#endif // #ifdef ASCEND_310P diff --git a/ggml/src/ggml-common.h b/ggml/src/ggml-common.h index 050161393..f13fd4dea 100644 --- a/ggml/src/ggml-common.h +++ b/ggml/src/ggml-common.h @@ -6,7 +6,20 @@ typedef uint16_t ggml_half; typedef uint32_t ggml_half2; -#define GGML_COMMON_AGGR +#define GGML_COMMON_AGGR_U +#define GGML_COMMON_AGGR_S + +#define GGML_COMMON_DECL +#elif defined(GGML_COMMON_DECL_CPP) +#include + +typedef uint16_t ggml_half; +typedef uint32_t ggml_half2; + +// std-c++ allow anonymous unions but some compiler warn on it +#define GGML_COMMON_AGGR_U data +// std-c++ do not allow it. +#define GGML_COMMON_AGGR_S data #define GGML_COMMON_DECL #elif defined(GGML_COMMON_DECL_METAL) @@ -15,7 +28,8 @@ typedef uint32_t ggml_half2; typedef half ggml_half; typedef half2 ggml_half2; -#define GGML_COMMON_AGGR +#define GGML_COMMON_AGGR_U +#define GGML_COMMON_AGGR_S #define GGML_COMMON_DECL #elif defined(GGML_COMMON_DECL_CUDA) @@ -29,7 +43,8 @@ typedef half2 ggml_half2; typedef half ggml_half; typedef half2 ggml_half2; -#define GGML_COMMON_AGGR data +#define GGML_COMMON_AGGR_U +#define GGML_COMMON_AGGR_S data #define GGML_COMMON_DECL #elif defined(GGML_COMMON_DECL_HIP) @@ -39,7 +54,8 @@ typedef half2 ggml_half2; typedef half ggml_half; typedef half2 ggml_half2; -#define GGML_COMMON_AGGR data +#define GGML_COMMON_AGGR_U +#define GGML_COMMON_AGGR_S data #define GGML_COMMON_DECL #elif defined(GGML_COMMON_DECL_SYCL) @@ -49,7 +65,8 @@ typedef half2 ggml_half2; typedef sycl::half ggml_half; typedef sycl::half2 ggml_half2; -#define GGML_COMMON_AGGR data +#define GGML_COMMON_AGGR_U +#define GGML_COMMON_AGGR_S data #define GGML_COMMON_DECL #endif @@ -154,9 +171,9 @@ typedef struct { struct { ggml_half d; // delta ggml_half m; // min - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 dm; - }; + } GGML_COMMON_AGGR_U; uint8_t qs[QK4_1 / 2]; // nibbles / quants } block_q4_1; static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_half) + QK4_1 / 2, "wrong q4_1 block size/padding"); @@ -175,9 +192,9 @@ typedef struct { struct { ggml_half d; // delta ggml_half m; // min - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 dm; - }; + } GGML_COMMON_AGGR_U; uint8_t qh[4]; // 5-th bit of quants uint8_t qs[QK5_1 / 2]; // nibbles / quants } block_q5_1; @@ -196,37 +213,13 @@ typedef struct { struct { ggml_half d; // delta ggml_half s; // d * sum(qs[i]) - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 ds; - }; + } GGML_COMMON_AGGR_U; int8_t qs[QK8_1]; // quants } block_q8_1; static_assert(sizeof(block_q8_1) == 2*sizeof(ggml_half) + QK8_1, "wrong q8_1 block size/padding"); -typedef struct { - ggml_half d[4]; // deltas for 4 q4_0 blocks - uint8_t qs[QK4_0 * 2]; // nibbles / quants for 4 q4_0 blocks -} block_q4_0x4; -static_assert(sizeof(block_q4_0x4) == 4 * sizeof(ggml_half) + QK4_0 * 2, "wrong q4_0x4 block size/padding"); - -typedef struct { - ggml_half d[8]; // deltas for 8 q4_0 blocks - uint8_t qs[QK4_0 * 4]; // nibbles / quants for 8 q4_0 blocks -} block_q4_0x8; -static_assert(sizeof(block_q4_0x8) == 8 * sizeof(ggml_half) + QK4_0 * 4, "wrong q4_0x8 block size/padding"); - -typedef struct { - ggml_half d[4]; // deltas for 4 q8_0 blocks - int8_t qs[QK8_0 * 4]; // quants for 4 q8_0 blocks -} block_q8_0x4; -static_assert(sizeof(block_q8_0x4) == 4 * sizeof(ggml_half) + QK8_0 * 4, "wrong q8_0x4 block size/padding"); - -typedef struct { - ggml_half d[8]; // deltas for 8 q8_0 blocks - int8_t qs[QK8_0 * 8]; // quants for 8 q8_0 blocks -} block_q8_0x8; -static_assert(sizeof(block_q8_0x8) == 8 * sizeof(ggml_half) + QK8_0 * 8, "wrong q8_0x8 block size/padding"); - // // Ternary quantization // @@ -261,9 +254,9 @@ typedef struct { struct { ggml_half d; // super-block scale for quantized scales ggml_half dmin; // super-block scale for quantized mins - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 dm; - }; + } GGML_COMMON_AGGR_U; } block_q2_K; static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_half) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding"); @@ -288,9 +281,9 @@ typedef struct { struct { ggml_half d; // super-block scale for quantized scales ggml_half dmin; // super-block scale for quantized mins - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 dm; - }; + } GGML_COMMON_AGGR_U; uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits uint8_t qs[QK_K/2]; // 4--bit quants } block_q4_K; @@ -305,9 +298,9 @@ typedef struct { struct { ggml_half d; // super-block scale for quantized scales ggml_half dmin; // super-block scale for quantized mins - } GGML_COMMON_AGGR; + } GGML_COMMON_AGGR_S; ggml_half2 dm; - }; + } GGML_COMMON_AGGR_U; uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits uint8_t qh[QK_K/8]; // quants, high bit uint8_t qs[QK_K/2]; // quants, low 4 bits @@ -431,6 +424,13 @@ static_assert(sizeof(block_iq4_xs) == sizeof(ggml_half) + sizeof(uint16_t) + QK_ #define GGML_TABLE_BEGIN(type, name, size) static const type name[size] = { #define GGML_TABLE_END() }; +#define GGML_COMMON_IMPL +#elif defined(GGML_COMMON_IMPL_CPP) +#include + +#define GGML_TABLE_BEGIN(type, name, size) static const type name[size] = { +#define GGML_TABLE_END() }; + #define GGML_COMMON_IMPL #elif defined(GGML_COMMON_IMPL_METAL) #include @@ -473,7 +473,7 @@ GGML_TABLE_BEGIN(uint8_t, ksigns_iq2xs, 128) 240, 113, 114, 243, 116, 245, 246, 119, 120, 249, 250, 123, 252, 125, 126, 255, GGML_TABLE_END() -//#if __CUDA_ARCH__ >= MIN_CC_DP4A // lowest compute capability for integer intrinsics +//#if __CUDA_ARCH__ >= GGML_CUDA_CC_DP4A // lowest compute capability for integer intrinsics GGML_TABLE_BEGIN(uint64_t, ksigns64, 128) 0x0000000000000000, 0xff000000000000ff, 0xff0000000000ff00, 0x000000000000ffff, 0xff00000000ff0000, 0x0000000000ff00ff, 0x0000000000ffff00, 0xff00000000ffffff, diff --git a/ggml/src/ggml-cpu/CMakeLists.txt b/ggml/src/ggml-cpu/CMakeLists.txt new file mode 100644 index 000000000..6b3641c42 --- /dev/null +++ b/ggml/src/ggml-cpu/CMakeLists.txt @@ -0,0 +1,346 @@ +function(ggml_add_cpu_backend_variant_impl tag_name) + if (tag_name) + set(GGML_CPU_NAME ggml-cpu-${tag_name}) + else() + set(GGML_CPU_NAME ggml-cpu) + endif() + + ggml_add_backend_library(${GGML_CPU_NAME}) + + list (APPEND GGML_CPU_SOURCES + ggml-cpu/ggml-cpu.c + ggml-cpu/ggml-cpu.cpp + ggml-cpu/ggml-cpu-aarch64.cpp + ggml-cpu/ggml-cpu-aarch64.h + ggml-cpu/ggml-cpu-hbm.cpp + ggml-cpu/ggml-cpu-hbm.h + ggml-cpu/ggml-cpu-quants.c + ggml-cpu/ggml-cpu-quants.h + ggml-cpu/ggml-cpu-traits.cpp + ggml-cpu/ggml-cpu-traits.h + ggml-cpu/amx/amx.cpp + ggml-cpu/amx/amx.h + ggml-cpu/amx/mmq.cpp + ggml-cpu/amx/mmq.h + ggml-cpu/ggml-cpu-impl.h + ) + + target_compile_features(${GGML_CPU_NAME} PRIVATE c_std_11 cxx_std_17) + target_include_directories(${GGML_CPU_NAME} PRIVATE . ggml-cpu) + + if (APPLE AND GGML_ACCELERATE) + find_library(ACCELERATE_FRAMEWORK Accelerate) + if (ACCELERATE_FRAMEWORK) + message(STATUS "Accelerate framework found") + + target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_ACCELERATE) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE ACCELERATE_NEW_LAPACK) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE ACCELERATE_LAPACK_ILP64) + + target_link_libraries(${GGML_CPU_NAME} PRIVATE ${ACCELERATE_FRAMEWORK}) + else() + message(WARNING "Accelerate framework not found") + endif() + endif() + + if (GGML_OPENMP) + find_package(OpenMP) + if (OpenMP_FOUND) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_OPENMP) + + target_link_libraries(${GGML_CPU_NAME} PRIVATE OpenMP::OpenMP_C OpenMP::OpenMP_CXX) + else() + message(WARNING "OpenMP not found") + endif() + endif() + + if (GGML_LLAMAFILE) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_LLAMAFILE) + + list(APPEND GGML_CPU_SOURCES + ggml-cpu/llamafile/sgemm.cpp + ggml-cpu/llamafile/sgemm.h) + endif() + + if (GGML_CPU_HBM) + find_library(memkind memkind REQUIRED) + + message(STATUS "Using memkind for CPU HBM") + + target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_CPU_HBM) + + target_link_libraries(${GGML_CPU_NAME} PUBLIC memkind) + endif() + + if (CMAKE_OSX_ARCHITECTURES STREQUAL "arm64" OR + CMAKE_GENERATOR_PLATFORM_LWR STREQUAL "arm64" OR + (NOT CMAKE_OSX_ARCHITECTURES AND NOT CMAKE_GENERATOR_PLATFORM_LWR AND + CMAKE_SYSTEM_PROCESSOR MATCHES "^(aarch64|arm.*|ARM64)$")) + + message(STATUS "ARM detected") + + if (MSVC AND NOT CMAKE_C_COMPILER_ID STREQUAL "Clang") + message(FATAL_ERROR "MSVC is not supported for ARM, use clang") + else() + check_cxx_compiler_flag(-mfp16-format=ieee GGML_COMPILER_SUPPORTS_FP16_FORMAT_I3E) + if (NOT "${GGML_COMPILER_SUPPORTS_FP16_FORMAT_I3E}" STREQUAL "") + list(APPEND ARCH_FLAGS -mfp16-format=ieee) + endif() + + if (GGML_NATIVE) + # -mcpu=native does not always enable all the features in some compilers, + # so we check for them manually and enable them if available + + execute_process( + COMMAND ${CMAKE_C_COMPILER} -mcpu=native -E -v - + INPUT_FILE "/dev/null" + OUTPUT_QUIET + ERROR_VARIABLE ARM_MCPU + RESULT_VARIABLE ARM_MCPU_RESULT + ) + if (NOT ARM_MCPU_RESULT) + string(REGEX MATCH "-mcpu=[^ ']+" ARM_MCPU_FLAG "${ARM_MCPU}") + endif() + if ("${ARM_MCPU_FLAG}" STREQUAL "") + set(ARM_MCPU_FLAG -mcpu=native) + message(STATUS "ARM -mcpu not found, -mcpu=native will be used") + endif() + + include(CheckCXXSourceRuns) + + function(check_arm_feature tag code) + set(CMAKE_REQUIRED_FLAGS_SAVE ${CMAKE_REQUIRED_FLAGS}) + set(CMAKE_REQUIRED_FLAGS "${ARM_MCPU_FLAG}+${tag}") + check_cxx_source_runs( + "${code}" + GGML_MACHINE_SUPPORTS_${tag} + ) + if (GGML_MACHINE_SUPPORTS_${tag}) + set(ARM_MCPU_FLAG_FIX "${ARM_MCPU_FLAG_FIX}+${tag}" PARENT_SCOPE) + else() + set(ARM_MCPU_FLAG_FIX "${ARM_MCPU_FLAG_FIX}+no${tag}" PARENT_SCOPE) + endif() + set(CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS_SAVE}) + endfunction() + + check_arm_feature(dotprod "#include \nint main() { int8x16_t _a, _b; volatile int32x4_t _s = vdotq_s32(_s, _a, _b); return 0; }") + check_arm_feature(i8mm "#include \nint main() { int8x16_t _a, _b; volatile int32x4_t _s = vmmlaq_s32(_s, _a, _b); return 0; }") + check_arm_feature(sve "#include \nint main() { svfloat32_t _a, _b; volatile svfloat32_t _c = svadd_f32_z(svptrue_b8(), _a, _b); return 0; }") + + list(APPEND ARCH_FLAGS "${ARM_MCPU_FLAG}${ARM_MCPU_FLAG_FIX}") + else() + if (GGML_CPU_ARM_ARCH) + list(APPEND ARCH_FLAGS -march=${GGML_CPU_ARM_ARCH}) + endif() + endif() + + # show enabled features + if (CMAKE_HOST_SYSTEM_NAME STREQUAL "Windows") + set(FEAT_INPUT_FILE "NUL") + else() + set(FEAT_INPUT_FILE "/dev/null") + endif() + + execute_process( + COMMAND ${CMAKE_C_COMPILER} ${ARCH_FLAGS} -dM -E - + INPUT_FILE ${FEAT_INPUT_FILE} + OUTPUT_VARIABLE ARM_FEATURE + RESULT_VARIABLE ARM_FEATURE_RESULT + ) + if (ARM_FEATURE_RESULT) + message(WARNING "Failed to get ARM features") + else() + foreach(feature DOTPROD SVE MATMUL_INT8 FMA FP16_VECTOR_ARITHMETIC) + string(FIND "${ARM_FEATURE}" "__ARM_FEATURE_${feature} 1" feature_pos) + if (NOT ${feature_pos} EQUAL -1) + message(STATUS "ARM feature ${feature} enabled") + endif() + endforeach() + endif() + endif() + elseif (CMAKE_OSX_ARCHITECTURES STREQUAL "x86_64" OR CMAKE_GENERATOR_PLATFORM_LWR MATCHES "^(x86_64|i686|amd64|x64|win32)$" OR + (NOT CMAKE_OSX_ARCHITECTURES AND NOT CMAKE_GENERATOR_PLATFORM_LWR AND + CMAKE_SYSTEM_PROCESSOR MATCHES "^(x86_64|i686|AMD64|amd64)$")) + + message(STATUS "x86 detected") + + if (MSVC) + # instruction set detection for MSVC only + if (GGML_NATIVE) + include(ggml-cpu/cmake/FindSIMD.cmake) + endif () + if (GGML_AVX512) + list(APPEND ARCH_FLAGS /arch:AVX512) + # /arch:AVX512 includes: __AVX512F__, __AVX512CD__, __AVX512BW__, __AVX512DQ__, and __AVX512VL__ + # MSVC has no compile-time flags enabling specific + # AVX512 extensions, neither it defines the + # macros corresponding to the extensions. + # Do it manually. + list(APPEND ARCH_DEFINITIONS GGML_AVX512) + if (GGML_AVX512_VBMI) + list(APPEND ARCH_DEFINITIONS __AVX512VBMI__) + if (CMAKE_C_COMPILER_ID STREQUAL "Clang") + list(APPEND ARCH_FLAGS -mavx512vbmi) + endif() + endif() + if (GGML_AVX512_VNNI) + list(APPEND ARCH_DEFINITIONS __AVX512VNNI__ GGML_AVX512_VNNI) + if (CMAKE_C_COMPILER_ID STREQUAL "Clang") + list(APPEND ARCH_FLAGS -mavx512vnni) + endif() + endif() + if (GGML_AVX512_BF16) + list(APPEND ARCH_DEFINITIONS __AVX512BF16__ GGML_AVX512_BF16) + if (CMAKE_C_COMPILER_ID STREQUAL "Clang") + list(APPEND ARCH_FLAGS -mavx512bf16) + endif() + endif() + if (GGML_AMX_TILE) + list(APPEND ARCH_DEFINITIONS __AMX_TILE__ GGML_AMX_TILE) + endif() + if (GGML_AMX_INT8) + list(APPEND ARCH_DEFINITIONS __AMX_INT8__ GGML_AMX_INT8) + endif() + if (GGML_AMX_BF16) + list(APPEND ARCH_DEFINITIONS __AMX_BF16__ GGML_AMX_BF16) + endif() + elseif (GGML_AVX2) + list(APPEND ARCH_FLAGS /arch:AVX2) + list(APPEND ARCH_DEFINITIONS GGML_AVX2 GGML_FMA GGML_F16C) + elseif (GGML_AVX) + list(APPEND ARCH_FLAGS /arch:AVX) + list(APPEND ARCH_DEFINITIONS GGML_AVX) + else () + list(APPEND ARCH_FLAGS /arch:SSE4.2) + list(APPEND ARCH_DEFINITIONS GGML_SSE42) + endif() + if (GGML_AVX_VNNI) + list(APPEND ARCH_DEFINITIONS __AVXVNNI__ GGML_AVX_VNNI) + endif() + else () + if (GGML_NATIVE) + list(APPEND ARCH_FLAGS -march=native) + else () + list(APPEND ARCH_FLAGS -msse4.2) + list(APPEND ARCH_DEFINITIONS GGML_SSE42) + if (GGML_F16C) + list(APPEND ARCH_FLAGS -mf16c) + list(APPEND ARCH_DEFINITIONS GGML_F16C) + endif() + if (GGML_FMA) + list(APPEND ARCH_FLAGS -mfma) + list(APPEND ARCH_DEFINITIONS GGML_FMA) + endif() + if (GGML_AVX) + list(APPEND ARCH_FLAGS -mavx) + list(APPEND ARCH_DEFINITIONS GGML_AVX) + endif() + if (GGML_AVX2) + list(APPEND ARCH_FLAGS -mavx2) + list(APPEND ARCH_DEFINITIONS GGML_AVX2) + endif() + if (GGML_AVX_VNNI) + list(APPEND ARCH_FLAGS -mavxvnni) + list(APPEND ARCH_DEFINITIONS GGML_AVX_VNNI) + endif() + if (GGML_AVX512) + list(APPEND ARCH_FLAGS -mavx512f) + list(APPEND ARCH_FLAGS -mavx512cd) + list(APPEND ARCH_FLAGS -mavx512vl) + list(APPEND ARCH_FLAGS -mavx512dq) + list(APPEND ARCH_FLAGS -mavx512bw) + list(APPEND ARCH_DEFINITIONS GGML_AVX512) + endif() + if (GGML_AVX512_VBMI) + list(APPEND ARCH_FLAGS -mavx512vbmi) + list(APPEND ARCH_DEFINITIONS GGML_AVX512_VBMI) + endif() + if (GGML_AVX512_VNNI) + list(APPEND ARCH_FLAGS -mavx512vnni) + list(APPEND ARCH_DEFINITIONS GGML_AVX512_VNNI) + endif() + if (GGML_AVX512_BF16) + list(APPEND ARCH_FLAGS -mavx512bf16) + list(APPEND ARCH_DEFINITIONS GGML_AVX512_BF16) + endif() + if (GGML_AMX_TILE) + list(APPEND ARCH_FLAGS -mamx-tile) + list(APPEND ARCH_DEFINITIONS GGML_AMX_TILE) + endif() + if (GGML_AMX_INT8) + list(APPEND ARCH_FLAGS -mamx-int8) + list(APPEND ARCH_DEFINITIONS GGML_AMX_INT8) + endif() + if (GGML_AMX_BF16) + list(APPEND ARCH_FLAGS -mamx-bf16) + list(APPEND ARCH_DEFINITIONS GGML_AMX_BF16) + endif() + endif() + endif() + elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64") + message(STATUS "PowerPC detected") + execute_process(COMMAND bash -c "grep POWER10 /proc/cpuinfo | head -n 1" OUTPUT_VARIABLE POWER10_M) + string(FIND "${POWER10_M}" "POWER10" substring_index) + if (NOT DEFINED substring_index OR "${substring_index}" STREQUAL "") + set(substring_index -1) + endif() + + if (${substring_index} GREATER_EQUAL 0) + list(APPEND ARCH_FLAGS -mcpu=power10) + elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "ppc64le") + list(APPEND ARCH_FLAGS -mcpu=powerpc64le) + else() + list(APPEND ARCH_FLAGS -mcpu=native -mtune=native) + # TODO: Add targets for Power8/Power9 (Altivec/VSX) and Power10(MMA) and query for big endian systems (ppc64/le/be) + endif() + elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "loongarch64") + message(STATUS "loongarch64 detected") + + list(APPEND ARCH_FLAGS -march=loongarch64) + if (GGML_LASX) + list(APPEND ARCH_FLAGS -mlasx) + endif() + if (GGML_LSX) + list(APPEND ARCH_FLAGS -mlsx) + endif() + elseif (${CMAKE_SYSTEM_PROCESSOR} MATCHES "riscv64") + message(STATUS "RISC-V detected") + if (GGML_RVV) + list(APPEND ARCH_FLAGS -march=rv64gcv -mabi=lp64d) + endif() + else() + message(STATUS "Unknown architecture") + endif() + + if (GGML_CPU_AARCH64) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE GGML_USE_CPU_AARCH64) + endif() + + message(STATUS "Adding CPU backend variant ${GGML_CPU_NAME}: ${ARCH_FLAGS} ${ARCH_DEFINITIONS}") + target_sources(${GGML_CPU_NAME} PRIVATE ${GGML_CPU_SOURCES}) + target_compile_options(${GGML_CPU_NAME} PRIVATE ${ARCH_FLAGS}) + target_compile_definitions(${GGML_CPU_NAME} PRIVATE ${ARCH_DEFINITIONS}) + + if (GGML_BACKEND_DL) + if (GGML_NATIVE) + # the feature check relies on ARCH_DEFINITIONS, but it is not set with GGML_NATIVE + message(FATAL_ERROR "GGML_NATIVE is not compatible with GGML_BACKEND_DL, consider using GGML_CPU_ALL_VARIANTS") + endif() + + # The feature detection code is compiled as a separate target so that + # it can be built without the architecture flags + # Since multiple variants of the CPU backend may be included in the same + # build, using set_source_files_properties() to set the arch flags is not possible + set(GGML_CPU_FEATS_NAME ${GGML_CPU_NAME}-feats) + add_library(${GGML_CPU_FEATS_NAME} OBJECT ggml-cpu/cpu-feats-x86.cpp) + target_include_directories(${GGML_CPU_FEATS_NAME} PRIVATE . .. ../include) + target_compile_definitions(${GGML_CPU_FEATS_NAME} PRIVATE ${ARCH_DEFINITIONS}) + target_compile_definitions(${GGML_CPU_FEATS_NAME} PRIVATE GGML_BACKEND_DL GGML_BACKEND_BUILD GGML_BACKEND_SHARED) + set_target_properties(${GGML_CPU_FEATS_NAME} PROPERTIES POSITION_INDEPENDENT_CODE ON) + target_link_libraries(${GGML_CPU_NAME} PRIVATE ${GGML_CPU_FEATS_NAME}) + endif() + + if (EMSCRIPTEN) + set_target_properties(${GGML_CPU_NAME} PROPERTIES COMPILE_FLAGS "-msimd128") + endif() +endfunction() diff --git a/ggml/src/ggml-cpu/amx/amx.cpp b/ggml/src/ggml-cpu/amx/amx.cpp new file mode 100644 index 000000000..5ec5263ce --- /dev/null +++ b/ggml/src/ggml-cpu/amx/amx.cpp @@ -0,0 +1,220 @@ +#include "amx.h" +#include "common.h" +#include "mmq.h" +#include "ggml-backend-impl.h" +#include "ggml-backend.h" +#include "ggml-impl.h" +#include "ggml-cpu.h" +#include "ggml-cpu-traits.h" + +#if defined(__gnu_linux__) +#include +#include +#endif + +#include +#include +#include + +#if defined(__AMX_INT8__) && defined(__AVX512VNNI__) + +// AMX type_trais +namespace ggml::cpu::amx { +class tensor_traits : public ggml::cpu::tensor_traits { + bool work_size(int /* n_threads */, const struct ggml_tensor * op, size_t & size) override { + size = ggml_backend_amx_desired_wsize(op); + return true; + } + + bool compute_forward(struct ggml_compute_params * params, struct ggml_tensor * op) override { + if (op->op == GGML_OP_MUL_MAT) { + ggml_backend_amx_mul_mat(params, op); + return true; + } + return false; + } +}; + +static ggml::cpu::tensor_traits * get_tensor_traits(ggml_backend_buffer_t, struct ggml_tensor *) { + static tensor_traits traits; + return &traits; +} +} // namespace ggml::cpu::amx + +// AMX buffer interface +static void ggml_backend_amx_buffer_free_buffer(ggml_backend_buffer_t buffer) { + free(buffer->context); +} + +static void * ggml_backend_amx_buffer_get_base(ggml_backend_buffer_t buffer) { + return (void *) (buffer->context); +} + +static void ggml_backend_amx_buffer_init_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { + tensor->extra = (void *) ggml::cpu::amx::get_tensor_traits(buffer, tensor); + + GGML_UNUSED(buffer); +} + +static void ggml_backend_amx_buffer_memset_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, + uint8_t value, size_t offset, size_t size) { + memset((char *) tensor->data + offset, value, size); + + GGML_UNUSED(buffer); +} + +static void ggml_backend_amx_buffer_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, + const void * data, size_t offset, size_t size) { + if (qtype_has_amx_kernels(tensor->type)) { + GGML_LOG_DEBUG("%s: amx repack tensor %s of type %s\n", __func__, tensor->name, ggml_type_name(tensor->type)); + ggml_backend_amx_convert_weight(tensor, data, offset, size); + } else { + memcpy((char *) tensor->data + offset, data, size); + } + + GGML_UNUSED(buffer); +} + +/* +// need to figure what we need to do with buffer->extra. +static void ggml_backend_amx_buffer_get_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + GGML_ASSERT(!qtype_has_amx_kernels(tensor->type)); + memcpy(data, (const char *)tensor->data + offset, size); + + GGML_UNUSED(buffer); +} + +static bool ggml_backend_amx_buffer_cpy_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst) { + if (ggml_backend_buffer_is_host(src->buffer)) { + if (qtype_has_amx_kernels(src->type)) { + ggml_backend_amx_convert_weight(dst, src->data, 0, ggml_nbytes(dst)); + } else { + memcpy(dst->data, src->data, ggml_nbytes(src)); + } + return true; + } + return false; + + GGML_UNUSED(buffer); +} +*/ + +static void ggml_backend_amx_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + memset(buffer->context, value, buffer->size); +} + +static ggml_backend_buffer_i ggml_backend_amx_buffer_interface = { + /* .free_buffer = */ ggml_backend_amx_buffer_free_buffer, + /* .get_base = */ ggml_backend_amx_buffer_get_base, + /* .init_tensor = */ ggml_backend_amx_buffer_init_tensor, + /* .memset_tensor = */ ggml_backend_amx_buffer_memset_tensor, + /* .set_tensor = */ ggml_backend_amx_buffer_set_tensor, + /* .get_tensor = */ nullptr, + /* .cpy_tensor = */ nullptr, + /* .clear = */ ggml_backend_amx_buffer_clear, + /* .reset = */ nullptr, +}; + +static const char * ggml_backend_amx_buffer_type_get_name(ggml_backend_buffer_type_t buft) { + return "AMX"; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_t ggml_backend_amx_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + void * data = ggml_aligned_malloc(size); + if (data == NULL) { + fprintf(stderr, "%s: failed to allocate buffer of size %zu\n", __func__, size); + return NULL; + } + + return ggml_backend_buffer_init(buft, ggml_backend_amx_buffer_interface, data, size); +} + +static size_t ggml_backend_amx_buffer_type_get_alignment(ggml_backend_buffer_type_t buft) { + return TENSOR_ALIGNMENT; + + GGML_UNUSED(buft); +} + +namespace ggml::cpu::amx { +class extra_buffer_type : ggml::cpu::extra_buffer_type { + bool supports_op(ggml_backend_dev_t, const struct ggml_tensor * op) override { + // handle only 2d gemm for now + auto is_contiguous_2d = [](const struct ggml_tensor * t) { + return ggml_is_contiguous(t) && t->ne[3] == 1 && t->ne[2] == 1; + }; + + if (op->op == GGML_OP_MUL_MAT && is_contiguous_2d(op->src[0]) && // src0 must be contiguous + is_contiguous_2d(op->src[1]) && // src1 must be contiguous + op->src[0]->buffer && op->src[0]->buffer->buft == ggml_backend_amx_buffer_type() && + op->ne[0] % (TILE_N * 2) == 0 && // out_features is 32x + (qtype_has_amx_kernels(op->src[0]->type) || (op->src[0]->type == GGML_TYPE_F16))) { + // src1 must be host buffer + if (op->src[1]->buffer && !ggml_backend_buft_is_host(op->src[1]->buffer->buft)) { + return false; + } + // src1 must be float32 + if (op->src[1]->type == GGML_TYPE_F32) { + return true; + } + } + return false; + } + + ggml::cpu::tensor_traits * get_tensor_traits(const struct ggml_tensor * op) override { + if (op->op == GGML_OP_MUL_MAT && op->src[0]->buffer && + op->src[0]->buffer->buft == ggml_backend_amx_buffer_type()) { + return (ggml::cpu::tensor_traits *) op->src[0]->extra; + } + + return nullptr; + } +}; +} // namespace ggml::cpu::amx + +static size_t ggml_backend_amx_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + return ggml_backend_amx_get_alloc_size(tensor); + + GGML_UNUSED(buft); +} + +#define ARCH_GET_XCOMP_PERM 0x1022 +#define ARCH_REQ_XCOMP_PERM 0x1023 +#define XFEATURE_XTILECFG 17 +#define XFEATURE_XTILEDATA 18 + +static bool ggml_amx_init() { +#if defined(__gnu_linux__) + if (syscall(SYS_arch_prctl, ARCH_REQ_XCOMP_PERM, XFEATURE_XTILEDATA)) { + fprintf(stderr, "AMX is not ready to be used!\n"); + return false; + } + return true; +#elif defined(_WIN32) + return true; +#endif +} + +ggml_backend_buffer_type_t ggml_backend_amx_buffer_type() { + static struct ggml_backend_buffer_type ggml_backend_buffer_type_amx = { + /* .iface = */ { + /* .get_name = */ ggml_backend_amx_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_amx_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_amx_buffer_type_get_alignment, + /* .get_max_size = */ nullptr, // defaults to SIZE_MAX + /* .get_alloc_size = */ ggml_backend_amx_buffer_type_get_alloc_size, + /* .is_host = */ nullptr, + }, + /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), + /* .context = */ new ggml::cpu::amx::extra_buffer_type(), + }; + + if (!ggml_amx_init()) { + return nullptr; + } + + return &ggml_backend_buffer_type_amx; +} + +#endif // defined(__AMX_INT8__) && defined(__AVX512VNNI__) diff --git a/ggml/src/ggml-cpu/amx/amx.h b/ggml/src/ggml-cpu/amx/amx.h new file mode 100644 index 000000000..5b65d76bd --- /dev/null +++ b/ggml/src/ggml-cpu/amx/amx.h @@ -0,0 +1,8 @@ +#include "ggml-backend.h" +#include "ggml-cpu-impl.h" + +// GGML internal header + +#if defined(__AMX_INT8__) && defined(__AVX512VNNI__) +ggml_backend_buffer_type_t ggml_backend_amx_buffer_type(void); +#endif diff --git a/ggml/src/ggml-amx/common.h b/ggml/src/ggml-cpu/amx/common.h similarity index 69% rename from ggml/src/ggml-amx/common.h rename to ggml/src/ggml-cpu/amx/common.h index 2b6c63527..f392e8985 100644 --- a/ggml/src/ggml-amx/common.h +++ b/ggml/src/ggml-cpu/amx/common.h @@ -1,13 +1,13 @@ #pragma once #include "ggml.h" -#include "ggml-cpu-impl.h" // +#include "ggml-cpu-impl.h" #include #include #include -#if defined(_OPENMP) +#if defined(GGML_USE_OPENMP) #include #endif @@ -56,11 +56,11 @@ inline void balance211(T n, T nth, T ith, T& n_start, T& n_end) { } template -inline void parallel_for(int nth, int n, const func_t& f) { -#if defined(_OPENMP) -#pragma omp parallel num_threads(nth) +inline void parallel_for(int n, const func_t& f) { +#if defined(GGML_USE_OPENMP) +#pragma omp parallel { - //int nth = omp_get_num_threads(); + int nth = omp_get_num_threads(); int ith = omp_get_thread_num(); int tbegin, tend; balance211(n, nth, ith, tbegin, tend); @@ -68,26 +68,24 @@ inline void parallel_for(int nth, int n, const func_t& f) { } #else f(0, n); - - GGML_UNUSED(nth); #endif } +template +inline void parallel_for_ggml(const ggml_compute_params * params, int n, const func_t & f) { + int tbegin, tend; + balance211(n, params->nth, params->ith, tbegin, tend); + f(tbegin, tend); +} + // quantized types that have AMX support inline bool qtype_has_amx_kernels(const enum ggml_type type) { // TODO: fix padding for vnni format return (type == GGML_TYPE_Q4_0) || - (type == GGML_TYPE_Q4_1); - //(type == GGML_TYPE_Q8_0) || - //(type == GGML_TYPE_Q4_K) || - //(type == GGML_TYPE_Q5_K) || - //(type == GGML_TYPE_Q6_K) || - //(type == GGML_TYPE_IQ4_XS); + (type == GGML_TYPE_Q4_1) || + (type == GGML_TYPE_Q8_0) || + (type == GGML_TYPE_Q4_K) || + (type == GGML_TYPE_Q5_K) || + (type == GGML_TYPE_Q6_K) || + (type == GGML_TYPE_IQ4_XS); } - -// ggml backend context -struct ggml_backend_amx_context { - int n_threads = GGML_DEFAULT_N_THREADS; - std::unique_ptr work_data; - size_t work_size = 0; -}; diff --git a/ggml/src/ggml-amx/mmq.cpp b/ggml/src/ggml-cpu/amx/mmq.cpp similarity index 96% rename from ggml/src/ggml-amx/mmq.cpp rename to ggml/src/ggml-cpu/amx/mmq.cpp index 239d15121..0ea91596b 100644 --- a/ggml/src/ggml-amx/mmq.cpp +++ b/ggml/src/ggml-cpu/amx/mmq.cpp @@ -4,8 +4,11 @@ #pragma GCC diagnostic ignored "-Wunused-local-typedefs" #endif +#include "amx.h" #include "mmq.h" #include "ggml-impl.h" +#include "ggml-cpu-impl.h" +#include "ggml-cpu-quants.h" #include "ggml-quants.h" #include #include @@ -15,10 +18,6 @@ #include #endif -#if defined(_OPENMP) -#include -#endif - #if (defined(_WIN32) || defined(_WIN64)) #define RESTRICT __restrict #else @@ -33,7 +32,7 @@ #define ALWAYS_INLINE inline #endif -#if defined(__AMX_INT8__) +#if defined(__AMX_INT8__) && defined(__AVX512VNNI__) namespace { @@ -496,19 +495,19 @@ inline void from_float(const float * x, char * vy, int64_t k); template <> inline void from_float(const float * x, char * vy, int64_t k) { - quantize_row_q8_0(x, vy, k); + quantize_row_q8_0(x, (block_q8_0 *)vy, k); } template <> inline void from_float(const float * x, char * vy, int64_t k) { - quantize_row_q8_1(x, vy, k); + quantize_row_q8_1(x, (block_q8_1 *)vy, k); } template <> inline void from_float(const float * x, char * vy, int64_t k) { #if 1 // TODO: this is reference impl! - quantize_row_q8_K(x, vy, k); + quantize_row_q8_K_ref(x, (block_q8_K *)vy, k); #else quantize_row_q8_K_vnni(x, vy, k); #endif @@ -949,7 +948,7 @@ template> void unpack_B(packed_B_t * RESTRICT tile, const void * RESTRICT packed_B) { GGML_UNUSED(tile); GGML_UNUSED(packed_B); -}; +} template <> void unpack_B(int8_t * RESTRICT tile, const void * RESTRICT packed_B) { @@ -1337,21 +1336,19 @@ struct tinygemm_kernel_avx __m512 vb[COLS]; __m512 vc[ROWS * COLS]; - auto loadc = [&](int idx) { + auto loadc = [&](auto idx) { vc[idx] = _mm512_setzero_ps(); }; Unroll{}(loadc); - auto compute = [&](int idx, int k) { - // TODO: use `constexpr` here to get rid of interger div - // when upgraded to C++17 - const int row = idx / COLS; - const int col = idx % COLS; + auto compute = [&](auto idx, auto k) { + constexpr int row = idx / COLS; + constexpr int col = idx % COLS; - if (col == 0) { + if constexpr (col == 0) { va = _mm512_loadu_ps(A + row * K + k); } - if (row == 0) { + if constexpr (row == 0) { vb[col] = _mm512_cvtph_ps(_mm256_loadu_si256((const __m256i *)(B + col * K + k))); } vc[idx] = _mm512_fmadd_ps(va, vb[col], vc[idx]); @@ -1361,9 +1358,9 @@ struct tinygemm_kernel_avx Unroll{}(compute, k); } - auto storec = [&](int idx) { - const int row = idx / COLS; - const int col = idx % COLS; + auto storec = [&](auto idx) { + constexpr int row = idx / COLS; + constexpr int col = idx % COLS; C[row * ldc + col] = _mm512_reduce_add_ps(vc[idx]); }; Unroll{}(storec); @@ -1381,13 +1378,13 @@ struct tinygemm_kernel_avx #define PACKED_INDEX(n, k, KB, tile_size) (n * KB + k) * tile_size template -void convert_B_packed_format(void * RESTRICT packed_B, const TB * RESTRICT B, int N, int K, int n_threads) { +void convert_B_packed_format(void * RESTRICT packed_B, const TB * RESTRICT B, int N, int K) { const int NB = N / TILE_N; const int KB = K / BLOCK_K; const int TILE_SIZE = get_tile_size(); // parallel on NB should be enough - parallel_for(n_threads, NB, [&](int begin, int end) { + parallel_for(NB, [&](int begin, int end) { for (int n = begin; n < end; ++n) { for (int k = 0; k < KB; ++k) { int n0 = n * TILE_N; @@ -1426,14 +1423,14 @@ struct tinygemm_kernel_vnni{}(loadc); - auto compute = [&](int col, int i) { + auto compute = [&](auto col, auto i) { // load a and compute compensation - if (col == 0) { + if constexpr (col == 0) { const int32_t * a_ptr = reinterpret_cast(A[0 * KB + i].qs); vcomp = _mm512_setzero_si512(); for (int k = 0; k < 8; ++k) { @@ -1465,7 +1462,7 @@ struct tinygemm_kernel_vnni{}(storec); @@ -1489,14 +1486,14 @@ struct tinygemm_kernel_vnni const __m512i lowMask = _mm512_set1_epi8(0xF); - auto loadc = [&](int col) { + auto loadc = [&](auto col) { vc[col] = _mm512_setzero_ps(); }; Unroll{}(loadc); - auto compute = [&](int col, int i) { + auto compute = [&](auto col, auto i) { // load a - if (col == 0) { + if constexpr (col == 0) { const int32_t * a_ptr = reinterpret_cast(A[0 * KB + i].qs); for (int k = 0; k < 8; ++k) { va[k] = _mm512_set1_epi32(a_ptr[k]); @@ -1530,7 +1527,7 @@ struct tinygemm_kernel_vnni } //store to C - auto storec = [&](int col) { + auto storec = [&](auto col) { _mm512_storeu_ps((__m512i*)(C + 0 * ldc + col * 16), vc[col]); }; Unroll{}(storec); @@ -1561,14 +1558,14 @@ struct tinygemm_kernel_vnni(0x80)); - auto loadc = [&](int col) { + auto loadc = [&](auto col) { vc[col] = _mm512_setzero_ps(); }; Unroll{}(loadc); - auto compute = [&](int col, int i) { + auto compute = [&](auto col, auto i) { // load a and add offset 128 - if (col == 0) { + if constexpr (col == 0) { const int32_t * a_ptr = reinterpret_cast(A[0 * KB + i].qs); for (int k = 0; k < 8; ++k) { va[k] = _mm512_set1_epi32(a_ptr[k]); @@ -1601,7 +1598,7 @@ struct tinygemm_kernel_vnni{}(storec); @@ -1633,7 +1630,7 @@ struct tinygemm_kernel_vnni{}(loadc); @@ -1647,9 +1644,9 @@ struct tinygemm_kernel_vnni{}(storec); @@ -1734,15 +1731,15 @@ struct tinygemm_kernel_vnni{}(loadc); // Q5_K and Q4_K shares the same vnni formats, refer to notes above. - auto compute = [&](int col, int i) { + auto compute = [&](auto col, auto i) { // load a - if (col == 0) { + if constexpr (col == 0) { for (int k_group = 0; k_group < QK_K / 32; ++k_group) { va[k_group] = _mm512_castsi256_si512(_mm256_loadu_si256((const __m256i *)(A[0 * KB + i].qs + k_group * 32))); } @@ -1807,7 +1804,7 @@ struct tinygemm_kernel_vnni{}(storec); @@ -1840,13 +1837,13 @@ struct tinygemm_kernel_vnni{}(loadc); - auto compute = [&](int col, int i) { - if (col == 0) { + auto compute = [&](auto col, auto i) { + if constexpr (col == 0) { // load a va[0] = _mm512_loadu_si512((const __m512i *)(A[0 * KB + i].qs + 0)); va[1] = _mm512_loadu_si512((const __m512i *)(A[0 * KB + i].qs + 64)); @@ -1958,13 +1955,13 @@ struct tinygemm_kernel_vnni(0x80)); const __m512i values256 = _mm512_add_epi8(values128, off); - auto loadc = [&](int col) { + auto loadc = [&](auto col) { vc[col] = _mm512_setzero_ps(); }; Unroll{}(loadc); - auto compute = [&](int col, int i) { - if (col == 0) { + auto compute = [&](auto col, auto i) { + if constexpr (col == 0) { // load a va[0] = _mm512_loadu_si512((const __m512i *)(A[0 * KB + i].qs + 0)); va[1] = _mm512_loadu_si512((const __m512i *)(A[0 * KB + i].qs + 64)); @@ -2014,7 +2011,7 @@ struct tinygemm_kernel_vnni{}(storec); @@ -2326,25 +2323,39 @@ size_t ggml_backend_amx_get_alloc_size(const struct ggml_tensor * tensor) { // pack weight to vnni format void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - - size_t alloc_size = ggml_backend_amx_get_alloc_size(tensor); - GGML_ASSERT(alloc_size == size); + GGML_ASSERT(offset == 0 && size == ggml_nbytes(tensor)); // only full tensor conversion is supported for now const enum ggml_type TYPE = tensor->type; const int K = tensor->ne[0]; // ne0: in_features const int N = tensor->ne[1]; // ne1: out_features -#if defined(_OPENMP) - // the buffer ctx is not initialized when .set_tensor is called - int n_threads = omp_get_num_threads(); -#else - int n_threads = 1; -#endif + GGML_DISPATCH_QTYPES(TYPE, [&] { + convert_B_packed_format((void *)((char *)tensor->data + offset), (const type *)data, N, K); + }); +} + +size_t ggml_backend_amx_desired_wsize(const struct ggml_tensor * dst) { + struct ggml_tensor * src0 = dst->src[0]; + + const enum ggml_type TYPE = src0->type; + + const bool is_floating_type = TYPE == GGML_TYPE_F16; + if (is_floating_type) { + return 0; + } + + const int M = dst->ne[1]; + const int K = src0->ne[0]; + + size_t desired_wsize = 0; GGML_DISPATCH_QTYPES(TYPE, [&] { - convert_B_packed_format((void *)((char *)tensor->data + offset), (const type *)data, N, K, n_threads); + const size_t row_size_A = K / blck_size * sizeof(vec_dot_type); + desired_wsize = M * row_size_A; }); + + return desired_wsize; } // NB: mixed dtype gemm with Advanced Matrix Extensions (Intel AMX) @@ -2355,14 +2366,12 @@ void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * d // // the function performs: dst = src1 @ src0.T // -void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor * dst) { +void ggml_backend_amx_mul_mat(const ggml_compute_params * params, struct ggml_tensor * dst) { struct ggml_tensor * src0 = dst->src[0]; struct ggml_tensor * src1 = dst->src[1]; const enum ggml_type TYPE = src0->type; - const int n_threads = ctx->n_threads; - // f16 only has avx512 kernels for now, // amx kernels will be added once 6th gen xeon is released. const bool is_floating_type = TYPE == GGML_TYPE_F16; @@ -2378,7 +2387,7 @@ void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor const int MB = div_up(M, BLOCK_M); const int NB = div_up(N, BLOCK_N); - parallel_for(n_threads, MB * NB, [&](int begin, int end) { + parallel_for_ggml(params, MB * NB, [&](int begin, int end) { GGML_DISPATCH_FLOATING_TYPES(TYPE, [&] { for (int i = begin; i < end; ++i) { int mb = i / NB; @@ -2411,27 +2420,29 @@ void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor } // pointer to work space, used convert A from float to quantized type - void * wdata = nullptr; + void * wdata = params->wdata; //TODO: performance improvement: merge quant A - GGML_DISPATCH_QTYPES(TYPE, [&] { - const size_t row_size_A = K / blck_size * sizeof(vec_dot_type); - const size_t desired_wsize = M * row_size_A; - if (ctx->work_size < desired_wsize) { - ctx->work_data.reset(new char[desired_wsize]); - ctx->work_size = desired_wsize; - } - wdata = ctx->work_data.get(); + if (params->ith == 0) { + GGML_DISPATCH_QTYPES(TYPE, [&] { + const size_t row_size_A = K / blck_size * sizeof(vec_dot_type); + const size_t desired_wsize = M * row_size_A; + if (params->wsize < desired_wsize) { + GGML_ABORT("insufficient work space size"); + } - // Q4_0, Q4_1, Q8_0 handles 1 TILE_K per blck_size - // Q4_K, Q5_K, Q6_K, IQ4_XS handles 8 TILE_K per blck_size - GGML_ASSERT(TILE_K == blck_size || TILE_K * 8 == blck_size); + // Q4_0, Q4_1, Q8_0 handles 1 TILE_K per blck_size + // Q4_K, Q5_K, Q6_K, IQ4_XS handles 8 TILE_K per blck_size + GGML_ASSERT(TILE_K == blck_size || TILE_K * 8 == blck_size); - const float * A_data = static_cast(src1->data); - for (int m = 0; m < M; ++m) { - from_float(A_data + m * K, (char *)wdata + m * row_size_A, K); - } - }); + const float * A_data = static_cast(src1->data); + for (int m = 0; m < M; ++m) { + from_float(A_data + m * K, (char *)wdata + m * row_size_A, K); + } + }); + } + + ggml_barrier(params->threadpool); if (M == 1) { // MB = 1 and handle 8 tiles in each block @@ -2439,7 +2450,7 @@ void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor constexpr int BLOCK_N = TILE_N * kTilesN; const int NB = div_up(N, BLOCK_N); - parallel_for(n_threads, NB, [&](int begin, int end) { + parallel_for_ggml(params, NB, [&](int begin, int end) { GGML_DISPATCH_QTYPES(TYPE, [&] { const int KB = K / blck_size; const int TILE_SIZE = get_tile_size(); @@ -2469,7 +2480,7 @@ void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor const int MB = div_up(M, BLOCK_M); const int NB = div_up(N, BLOCK_N); - parallel_for(n_threads, MB * NB, [&](int begin, int end) { + parallel_for_ggml(params, MB * NB, [&](int begin, int end) { // init tile config for each thread ggml_tile_config_init(); @@ -2497,13 +2508,4 @@ void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor }); } -#else // if defined(__AMX_INT8__) - -void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor * dst) { - fprintf(stderr, "GGML is not compiled with AMX support!\n"); - - GGML_UNUSED(ctx); - GGML_UNUSED(dst); -} - -#endif // if defined(__AMX_INT8__) +#endif // if defined(__AMX_INT8__) && defined(__AVX512VNNI__) diff --git a/ggml/src/ggml-amx/mmq.h b/ggml/src/ggml-cpu/amx/mmq.h similarity index 56% rename from ggml/src/ggml-amx/mmq.h rename to ggml/src/ggml-cpu/amx/mmq.h index cf0920620..baf768477 100644 --- a/ggml/src/ggml-amx/mmq.h +++ b/ggml/src/ggml-cpu/amx/mmq.h @@ -1,17 +1,10 @@ #pragma once #include "common.h" -#include -#ifdef __cplusplus -extern "C" { -#endif +size_t ggml_backend_amx_desired_wsize(const struct ggml_tensor * dst); size_t ggml_backend_amx_get_alloc_size(const struct ggml_tensor * tensor); void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); -void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor * dst); - -#ifdef __cplusplus -} -#endif +void ggml_backend_amx_mul_mat(const struct ggml_compute_params * params, struct ggml_tensor * dst); diff --git a/ggml/cmake/FindSIMD.cmake b/ggml/src/ggml-cpu/cmake/FindSIMD.cmake similarity index 100% rename from ggml/cmake/FindSIMD.cmake rename to ggml/src/ggml-cpu/cmake/FindSIMD.cmake diff --git a/ggml/src/ggml-cpu/cpu-feats-x86.cpp b/ggml/src/ggml-cpu/cpu-feats-x86.cpp new file mode 100644 index 000000000..e8133d411 --- /dev/null +++ b/ggml/src/ggml-cpu/cpu-feats-x86.cpp @@ -0,0 +1,323 @@ +#include "ggml-backend-impl.h" + +#if defined(__x86_64__) || (defined(_MSC_VER) && defined(_M_AMD64)) + +#ifdef _MSC_VER +#include +#endif + +#include +#include +#include +#include +#include + +// ref: https://cdrdv2-public.intel.com/782156/325383-sdm-vol-2abcd.pdf +struct cpuid_x86 { + bool SSE3(void) { return f_1_ecx[0]; } + bool PCLMULQDQ(void) { return f_1_ecx[1]; } + bool MONITOR(void) { return f_1_ecx[3]; } + bool SSSE3(void) { return f_1_ecx[9]; } + bool FMA(void) { return f_1_ecx[12]; } + bool CMPXCHG16B(void) { return f_1_ecx[13]; } + bool SSE41(void) { return f_1_ecx[19]; } + bool SSE42(void) { return f_1_ecx[20]; } + bool MOVBE(void) { return f_1_ecx[22]; } + bool POPCNT(void) { return f_1_ecx[23]; } + bool AES(void) { return f_1_ecx[25]; } + bool XSAVE(void) { return f_1_ecx[26]; } + bool OSXSAVE(void) { return f_1_ecx[27]; } + bool AVX(void) { return f_1_ecx[28]; } + bool F16C(void) { return f_1_ecx[29]; } + bool RDRAND(void) { return f_1_ecx[30]; } + + bool MSR(void) { return f_1_edx[5]; } + bool CX8(void) { return f_1_edx[8]; } + bool SEP(void) { return f_1_edx[11]; } + bool CMOV(void) { return f_1_edx[15]; } + bool CLFSH(void) { return f_1_edx[19]; } + bool MMX(void) { return f_1_edx[23]; } + bool FXSR(void) { return f_1_edx[24]; } + bool SSE(void) { return f_1_edx[25]; } + bool SSE2(void) { return f_1_edx[26]; } + + bool FSGSBASE(void) { return f_7_ebx[0]; } + bool BMI1(void) { return f_7_ebx[3]; } + bool HLE(void) { return is_intel && f_7_ebx[4]; } + bool AVX2(void) { return f_7_ebx[5]; } + bool BMI2(void) { return f_7_ebx[8]; } + bool ERMS(void) { return f_7_ebx[9]; } + bool INVPCID(void) { return f_7_ebx[10]; } + bool RTM(void) { return is_intel && f_7_ebx[11]; } + bool AVX512F(void) { return f_7_ebx[16]; } + bool AVX512DQ(void) { return f_7_ebx[17]; } + bool RDSEED(void) { return f_7_ebx[18]; } + bool ADX(void) { return f_7_ebx[19]; } + bool AVX512PF(void) { return f_7_ebx[26]; } + bool AVX512ER(void) { return f_7_ebx[27]; } + bool AVX512CD(void) { return f_7_ebx[28]; } + bool AVX512BW(void) { return f_7_ebx[30]; } + bool AVX512VL(void) { return f_7_ebx[31]; } + + bool SHA(void) { return f_7_ebx[29]; } + + bool PREFETCHWT1(void) { return f_7_ecx[0]; } + + bool LAHF(void) { return f_81_ecx[0]; } + bool LZCNT(void) { return is_intel && f_81_ecx[5]; } + bool ABM(void) { return is_amd && f_81_ecx[5]; } + bool SSE4a(void) { return is_amd && f_81_ecx[6]; } + bool XOP(void) { return is_amd && f_81_ecx[11]; } + bool TBM(void) { return is_amd && f_81_ecx[21]; } + + bool SYSCALL(void) { return is_intel && f_81_edx[11]; } + bool MMXEXT(void) { return is_amd && f_81_edx[22]; } + bool RDTSCP(void) { return is_intel && f_81_edx[27]; } + bool _3DNOWEXT(void) { return is_amd && f_81_edx[30]; } + bool _3DNOW(void) { return is_amd && f_81_edx[31]; } + + bool AVX512_VBMI(void) { return f_7_ecx[1]; } + bool AVX512_VNNI(void) { return f_7_ecx[11]; } + bool AVX512_FP16(void) { return f_7_edx[23]; } + bool AVX512_BF16(void) { return f_7_1_eax[5]; } + bool AVX_VNNI(void) { return f_7_1_eax[4]; } + + bool AMX_TILE(void) { return f_7_edx[24]; } + bool AMX_INT8(void) { return f_7_edx[25]; } + bool AMX_FP16(void) { return f_7_1_eax[21]; } + bool AMX_BF16(void) { return f_7_edx[22]; } + +#ifdef _MSC_VER + static void cpuid(int cpu_info[4], int eax) { + __cpuid(cpu_info, eax); + } + static void cpuidex(int cpu_info[4], int eax, int ecx) { + __cpuidex(cpu_info, eax, ecx); + } +#else + static void cpuid(int cpu_info[4], int eax) { + __asm__ __volatile__( + "cpuid" + : "=a"(cpu_info[0]), "=b"(cpu_info[1]), "=c"(cpu_info[2]), "=d"(cpu_info[3]) + : "a"(eax), "c"(0)); + } + static void cpuidex(int cpu_info[4], int eax, int ecx) { + __asm__ __volatile__( + "cpuid" + : "=a"(cpu_info[0]), "=b"(cpu_info[1]), "=c"(cpu_info[2]), "=d"(cpu_info[3]) + : "a"(eax), "c"(ecx)); + } +#endif + + cpuid_x86() { + std::array cpui; + std::vector> data; + + // calling __cpuid with 0x0 as the function_id argument + // gets the number of the highest valid function ID. + cpuid(cpui.data(), 0); + int n_ids = cpui[0]; + + for (int i = 0; i <= n_ids; ++i) { + cpuidex(cpui.data(), i, 0); + data.push_back(cpui); + } + + // capture vendor string + char vendor[0x20] = {}; + *reinterpret_cast(vendor) = data[0][1]; + *reinterpret_cast(vendor + 4) = data[0][3]; + *reinterpret_cast(vendor + 8) = data[0][2]; + this->vendor = vendor; + if (this->vendor == "GenuineIntel") { + is_intel = true; + } else if (this->vendor == "AuthenticAMD") { + is_amd = true; + } + + // load bitset with flags for function 0x00000001 + if (n_ids >= 1) { + f_1_ecx = data[1][2]; + f_1_edx = data[1][3]; + } + + // load bitset with flags for function 0x00000007 + if (n_ids >= 7) { + f_7_ebx = data[7][1]; + f_7_ecx = data[7][2]; + f_7_edx = data[7][3]; + cpuidex(cpui.data(), 7, 1); + f_7_1_eax = cpui[0]; + } + + // calling __cpuid with 0x80000000 as the function_id argument + // gets the number of the highest valid extended ID. + cpuid(cpui.data(), 0x80000000); + unsigned int n_ex_ids = cpui[0]; + + std::vector> ext_data; + for (unsigned int i = 0x80000000; i <= n_ex_ids; ++i) { + cpuidex(cpui.data(), i, 0); + ext_data.push_back(cpui); + } + + // load bitset with flags for function 0x80000001 + if (n_ex_ids >= 0x80000001) { + f_81_ecx = ext_data[1][2]; + f_81_edx = ext_data[1][3]; + } + + // interpret CPU brand string if reported + char brand[0x40] = {}; + if (n_ex_ids >= 0x80000004) { + std::memcpy(brand, ext_data[2].data(), sizeof(cpui)); + std::memcpy(brand + 16, ext_data[3].data(), sizeof(cpui)); + std::memcpy(brand + 32, ext_data[4].data(), sizeof(cpui)); + this->brand = brand; + } + } + + bool is_intel = false; + bool is_amd = false; + std::string vendor; + std::string brand; + std::bitset<32> f_1_ecx; + std::bitset<32> f_1_edx; + std::bitset<32> f_7_ebx; + std::bitset<32> f_7_ecx; + std::bitset<32> f_7_edx; + std::bitset<32> f_7_1_eax; + std::bitset<32> f_81_ecx; + std::bitset<32> f_81_edx; +}; + +#if 0 +void test_x86_is() { + cpuid_x86 is; + printf("CPU Vendor: %s\n", is.vendor.c_str()); + printf("Brand: %s\n", is.brand.c_str()); + printf("is_intel: %d\n", is.is_intel); + printf("is_amd: %d\n", is.is_amd); + printf("sse3: %d\n", is.SSE3()); + printf("pclmulqdq: %d\n", is.PCLMULQDQ()); + printf("ssse3: %d\n", is.SSSE3()); + printf("fma: %d\n", is.FMA()); + printf("cmpxchg16b: %d\n", is.CMPXCHG16B()); + printf("sse41: %d\n", is.SSE41()); + printf("sse42: %d\n", is.SSE42()); + printf("movbe: %d\n", is.MOVBE()); + printf("popcnt: %d\n", is.POPCNT()); + printf("aes: %d\n", is.AES()); + printf("xsave: %d\n", is.XSAVE()); + printf("osxsave: %d\n", is.OSXSAVE()); + printf("avx: %d\n", is.AVX()); + printf("f16c: %d\n", is.F16C()); + printf("rdrand: %d\n", is.RDRAND()); + printf("msr: %d\n", is.MSR()); + printf("cx8: %d\n", is.CX8()); + printf("sep: %d\n", is.SEP()); + printf("cmov: %d\n", is.CMOV()); + printf("clflush: %d\n", is.CLFSH()); + printf("mmx: %d\n", is.MMX()); + printf("fxsr: %d\n", is.FXSR()); + printf("sse: %d\n", is.SSE()); + printf("sse2: %d\n", is.SSE2()); + printf("fsgsbase: %d\n", is.FSGSBASE()); + printf("bmi1: %d\n", is.BMI1()); + printf("hle: %d\n", is.HLE()); + printf("avx2: %d\n", is.AVX2()); + printf("bmi2: %d\n", is.BMI2()); + printf("erms: %d\n", is.ERMS()); + printf("invpcid: %d\n", is.INVPCID()); + printf("rtm: %d\n", is.RTM()); + printf("avx512f: %d\n", is.AVX512F()); + printf("rdseed: %d\n", is.RDSEED()); + printf("adx: %d\n", is.ADX()); + printf("avx512pf: %d\n", is.AVX512PF()); + printf("avx512er: %d\n", is.AVX512ER()); + printf("avx512cd: %d\n", is.AVX512CD()); + printf("sha: %d\n", is.SHA()); + printf("prefetchwt1: %d\n", is.PREFETCHWT1()); + printf("lahf: %d\n", is.LAHF()); + printf("lzcnt: %d\n", is.LZCNT()); + printf("abm: %d\n", is.ABM()); + printf("sse4a: %d\n", is.SSE4a()); + printf("xop: %d\n", is.XOP()); + printf("tbm: %d\n", is.TBM()); + printf("syscall: %d\n", is.SYSCALL()); + printf("mmxext: %d\n", is.MMXEXT()); + printf("rdtscp: %d\n", is.RDTSCP()); + printf("3dnowext: %d\n", is._3DNOWEXT()); + printf("3dnow: %d\n", is._3DNOW()); + printf("avx512_vbmi: %d\n", is.AVX512_VBMI()); + printf("avx512_vnni: %d\n", is.AVX512_VNNI()); + printf("avx512_fp16: %d\n", is.AVX512_FP16()); + printf("avx512_bf16: %d\n", is.AVX512_BF16()); + printf("amx_tile: %d\n", is.AMX_TILE()); + printf("amx_int8: %d\n", is.AMX_INT8()); + printf("amx_fp16: %d\n", is.AMX_FP16()); + printf("amx_bf16: %d\n", is.AMX_BF16()); +} +#endif + +static int ggml_backend_cpu_x86_score() { + // FIXME: this does not check for OS support + + int score = 0; + cpuid_x86 is; + +#ifdef GGML_FMA + if (!is.FMA()) { return 0; } + score += 1; +#endif +#ifdef GGML_F16C + if (!is.F16C()) { return 0; } + score += 1<<1; +#endif +#ifdef GGML_SSE42 + if (!is.SSE42()) { return 0; } + score += 1<<2; +#endif +#ifdef GGML_AVX + if (!is.AVX()) { return 0; } + score += 1<<4; +#endif +#ifdef GGML_AVX2 + if (!is.AVX2()) { return 0; } + score += 1<<5; +#endif +#ifdef GGML_AVX_VNNI + if (!is.AVX_VNNI()) { return 0; } + score += 1<<6; +#endif +#ifdef GGML_AVX512 + if (!is.AVX512F()) { return 0; } + if (!is.AVX512CD()) { return 0; } + if (!is.AVX512VL()) { return 0; } + if (!is.AVX512DQ()) { return 0; } + if (!is.AVX512BW()) { return 0; } + score += 1<<7; +#endif +#ifdef GGML_AVX512_VBMI + if (!is.AVX512_VBMI()) { return 0; } + score += 1<<8; +#endif +#ifdef GGML_AVX512_BF16 + if (!is.AVX512_BF16()) { return 0; } + score += 1<<9; +#endif +#ifdef GGML_AVX512_VNNI + if (!is.AVX512_VNNI()) { return 0; } + score += 1<<10; +#endif +#ifdef GGML_AMX_INT8 + if (!is.AMX_INT8()) { return 0; } + score += 1<<11; +#endif + + return score; +} + +GGML_BACKEND_DL_SCORE_IMPL(ggml_backend_cpu_x86_score) + +#endif // defined(__x86_64__) || (defined(_MSC_VER) && defined(_M_AMD64)) diff --git a/ggml/src/ggml-aarch64.c b/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp similarity index 77% rename from ggml/src/ggml-aarch64.c rename to ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp index 81f62ff4f..b311a5b1c 100644 --- a/ggml/src/ggml-aarch64.c +++ b/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp @@ -1,23 +1,56 @@ -// SPDX-FileCopyrightText: Copyright 2024 Arm Limited and/or its affiliates -// SPDX-License-Identifier: MIT -// - -#define GGML_COMMON_IMPL_C +#define GGML_COMMON_IMPL_CPP +#define GGML_COMMON_DECL_CPP #include "ggml-common.h" +#include "ggml-backend-impl.h" #include "ggml-quants.h" #include "ggml-impl.h" #include "ggml-cpu.h" #include "ggml-cpu-impl.h" +#include "ggml-cpu-traits.h" -#include -#include -#include -#include -#include // for qsort -#include // for GGML_ASSERT +#include +#include +#include +#include +#include // for qsort +#include // for GGML_ASSERT -#include "ggml-aarch64.h" +#include "ggml-cpu-aarch64.h" + +// TODO: move to include file? +template constexpr int QK_0() { + if constexpr (K == 4) { + return QK4_0; + } + if constexpr (K == 8) { + return QK8_0; + } + return -1; +} + +template struct block { + ggml_half d[N]; // deltas for N qK_0 blocks + int8_t qs[(QK_0() * N * K) / 8]; // quants for N qK_0 blocks +}; + +// control size +static_assert(sizeof(block<4, 4>) == 4 * sizeof(ggml_half) + QK8_0 * 2, "wrong block<4,4> size/padding"); +static_assert(sizeof(block<4, 8>) == 8 * sizeof(ggml_half) + QK8_0 * 4, "wrong block<4,8> size/padding"); +static_assert(sizeof(block<8, 4>) == 4 * sizeof(ggml_half) + QK8_0 * 4, "wrong block<8,4> size/padding"); +static_assert(sizeof(block<8, 8>) == 8 * sizeof(ggml_half) + QK8_0 * 8, "wrong block<8,8> size/padding"); + +using block_q4_0x4 = block<4, 4>; +using block_q4_0x8 = block<4, 8>; +using block_q8_0x4 = block<8, 4>; +using block_q8_0x8 = block<8, 8>; + +struct block_iq4_nlx4 { + ggml_half d[4]; // deltas for 4 iq4_nl blocks + uint8_t qs[QK4_NL * 2]; // nibbles / quants for 4 iq4_nl blocks +}; + +static_assert(sizeof(block_iq4_nlx4) == 4 * sizeof(ggml_half) + QK4_NL * 2, "wrong iq4_nlx4 block size/padding"); #if defined(__GNUC__) #pragma GCC diagnostic ignored "-Woverlength-strings" @@ -132,7 +165,7 @@ static inline __m512i sum_i16_pairs_int_32x16(const __m512i x) { } static inline __m512i mul_sum_us8_pairs_int32x16(const __m512i ax, const __m512i sy) { -#if defined(__AVXVNNI__) || (defined(__AVX512VNNI__) && defined(__AVX512VL__)) +#if defined(__AVX512VNNI__) const __m512i zero = _mm512_setzero_si512(); return _mm512_dpbusd_epi32(zero, ax, sy); #else @@ -161,9 +194,12 @@ static inline __m256i sum_i16_pairs_int32x8(const __m256i x) { } static inline __m256i mul_sum_us8_pairs_int32x8(const __m256i ax, const __m256i sy) { -#if defined(__AVXVNNI__) || (defined(__AVX512VNNI__) && defined(__AVX512VL__)) +#if defined(__AVX512VNNI__) && defined(__AVX512VL__) const __m256i zero = _mm256_setzero_si256(); return _mm256_dpbusd_epi32(zero, ax, sy); +#elif defined(__AVXVNNI__) + const __m256i zero = _mm256_setzero_si256(); + return _mm256_dpbusd_avx_epi32(zero, ax, sy); #else // Perform multiplication and create 16-bit values const __m256i dot = _mm256_maddubs_epi16(ax, sy); @@ -187,52 +223,14 @@ static inline __m256i mul_sum_i8_pairs_int32x8(const __m256i x, const __m256i y) } #endif -static block_q4_0x4 make_block_q4_0x4(block_q4_0 * in, unsigned int blck_size_interleave, unsigned int xor_mask) { - block_q4_0x4 out; +static const int8_t kvalues_iq4nl[16] = {-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113}; - for (int i = 0; i < 4; i++) { - out.d[i] = in[i].d; - } - - for (int i = 0; i < QK4_0 * 2; i++) { - int src_offset = (i / (4 * blck_size_interleave)) * blck_size_interleave; - int src_id = (i % (4 * blck_size_interleave)) / blck_size_interleave; - src_offset += (i % blck_size_interleave); - - out.qs[i] = in[src_id].qs[src_offset] ^ xor_mask; - } - - return out; -} - -// interleave 8 block_q4_0s in blocks of blck_size_interleave -// returns an interleaved block_q4_0x8 -// in the interleaved block_q4_0x8, place deltas for 8 block_q4_0 blocks -// first, then interleave quants from 8 block_q4_0s in blocks of blck_size_interleave -static block_q4_0x8 make_block_q4_0x8(block_q4_0 * in, unsigned int blck_size_interleave, unsigned int xor_mask) { - block_q4_0x8 out; - - for (int i = 0; i < 8; i++) { - out.d[i] = in[i].d; - } - - for (int i = 0; i < QK4_0 * 4; i++) { - int src_offset = (i / (8 * blck_size_interleave)) * blck_size_interleave; - int src_id = (i % (8 * blck_size_interleave)) / blck_size_interleave; - src_offset += (i % blck_size_interleave); - - out.qs[i] = in[src_id].qs[src_offset] ^ xor_mask; - } - - return out; -} - -void quantize_q8_0_4x4(const float * restrict x, void * restrict vy, int64_t k) { +static void quantize_q8_0_4x4(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) { assert(QK8_0 == 32); assert(k % QK8_0 == 0); const int nb = k / QK8_0; - block_q8_0x4 * restrict y = (block_q8_0x4 *) vy; + block_q8_0x4 * GGML_RESTRICT y = (block_q8_0x4 *) vy; #if defined(__ARM_NEON) float32x4_t srcv[4][8]; @@ -321,12 +319,12 @@ void quantize_q8_0_4x4(const float * restrict x, void * restrict vy, int64_t k) #endif } -void quantize_q8_0_4x8(const float * restrict x, void * restrict vy, int64_t k) { +static void quantize_q8_0_4x8(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t k) { assert(QK8_0 == 32); assert(k % QK8_0 == 0); const int nb = k / QK8_0; - block_q8_0x4 * restrict y = (block_q8_0x4 *) vy; + block_q8_0x4 * GGML_RESTRICT y = (block_q8_0x4 *) vy; #if defined(__ARM_NEON) float32x4_t srcv[4][8]; @@ -536,7 +534,7 @@ void quantize_q8_0_4x8(const float * restrict x, void * restrict vy, int64_t k) #endif } -void quantize_mat_q8_0(const float * restrict x, void * restrict vy, int64_t nrow, int64_t n_per_row, int64_t blck_size_interleave) { +static void quantize_mat_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT vy, int64_t nrow, int64_t n_per_row, int64_t blck_size_interleave) { assert(nrow == 4); UNUSED(nrow); if (blck_size_interleave == 4) { @@ -548,58 +546,7 @@ void quantize_mat_q8_0(const float * restrict x, void * restrict vy, int64_t nro } } -static size_t quantize_q4_0_nr_bl(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, int nrows_interleaved, int blck_size_interleave) { - assert(n_per_row % QK4_0 == 0); - const int nb = n_per_row / QK4_0; - - void * out_ptr = NULL; - if (nrows_interleaved == 8) { - out_ptr = (block_q4_0x8 *) dst; - } - else if (nrows_interleaved == 4) { - out_ptr = (block_q4_0x4 *) dst; - } - assert(nrows_interleaved <= 8); - block_q4_0 dst_tmp[8]; - - for (int b = 0; b < (nrow * n_per_row); b += nrows_interleaved * n_per_row) { - - for (int64_t x = 0; x < nb; x++) { - - for (int i = 0; i < nrows_interleaved; i++ ) { - quantize_row_q4_0_ref(src + b + i * n_per_row + x * QK4_0, (block_q4_0 *) dst_tmp + i, QK4_0); - } - - if (nrows_interleaved == 8) { - *(block_q4_0x8 *) out_ptr = make_block_q4_0x8(dst_tmp, blck_size_interleave, 0x88); - out_ptr = (block_q4_0x8 *) out_ptr + 1; - } - else if (nrows_interleaved == 4) { - *(block_q4_0x4 *) out_ptr = make_block_q4_0x4(dst_tmp, blck_size_interleave, 0x88); - out_ptr = (block_q4_0x4 *) out_ptr + 1; - } - } - } - - return ((nrow * n_per_row) / QK4_0 * sizeof(block_q4_0)); -} - -size_t quantize_q4_0_4x4(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { - UNUSED(quant_weights); - return quantize_q4_0_nr_bl(src, dst, nrow, n_per_row, 4, 4); -} - -size_t quantize_q4_0_4x8(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { - UNUSED(quant_weights); - return quantize_q4_0_nr_bl(src, dst, nrow, n_per_row, 4, 8); -} - -size_t quantize_q4_0_8x8(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { - UNUSED(quant_weights); - return quantize_q4_0_nr_bl(src, dst, nrow, n_per_row, 8, 8); -} - -void ggml_gemv_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemv_q4_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 4; @@ -618,67 +565,47 @@ void ggml_gemv_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * UNUSED(ncols_interleaved); UNUSED(blocklen); -#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) - if (ggml_cpu_has_neon()) { - const void * b_ptr = vx; - const void * a_ptr = vy; - float * res_ptr = s; +#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + const block_q4_0x4 * b_ptr = (const block_q4_0x4 *) vx; - __asm__ __volatile__( - "movi v31.16b, #0x4\n" - "movi v30.16b, #0xf0\n" - "add %x[b_ptr], %x[b_ptr], #0x8\n" - "1:" // Column loop - "add x22, %x[a_ptr], #0x2\n" - "movi v29.16b, #0x0\n" - "mov x21, %x[nb]\n" - "2:" // Block loop - "ldr q28, [%x[b_ptr], #0x0]\n" - "ldr q27, [x22, #0x0]\n" - "movi v26.4s, #0x0\n" - "sub x20, x22, #0x2\n" - "ldr q25, [x22, #0x10]\n" - "ldr q24, [%x[b_ptr], #0x10]\n" - "sub x21, x21, #0x1\n" - "add x22, x22, #0x22\n" - "ldr q23, [%x[b_ptr], #0x20]\n" - "ldr q22, [%x[b_ptr], #0x30]\n" - "ld1r { v21.8h }, [x20]\n" - "ldr q20, [%x[b_ptr], #-0x8]\n" - "sshl v16.16b, v28.16b, v31.16b\n" - "and v28.16b, v28.16b, v30.16b\n" - "sshl v19.16b, v24.16b, v31.16b\n" - "and v24.16b, v24.16b, v30.16b\n" - "add %x[b_ptr], %x[b_ptr], #0x48\n" - "sshl v18.16b, v23.16b, v31.16b\n" - "and v23.16b, v23.16b, v30.16b\n" - ".inst 0x4f9be21a // sdot v26.4s, v16.16b, v27.4b[0]\n" - "sshl v17.16b, v22.16b, v31.16b\n" - "and v22.16b, v22.16b, v30.16b\n" - "fcvtl v21.4s, v21.4h\n" - "fcvtl v16.4s, v20.4h\n" - ".inst 0x4f99e39a // sdot v26.4s, v28.16b, v25.4b[0]\n" - "fmul v16.4s, v16.4s, v21.4s\n" - ".inst 0x4fbbe27a // sdot v26.4s, v19.16b, v27.4b[1]\n" - ".inst 0x4fb9e31a // sdot v26.4s, v24.16b, v25.4b[1]\n" - ".inst 0x4f9bea5a // sdot v26.4s, v18.16b, v27.4b[2]\n" - ".inst 0x4f99eafa // sdot v26.4s, v23.16b, v25.4b[2]\n" - ".inst 0x4fbbea3a // sdot v26.4s, v17.16b, v27.4b[3]\n" - ".inst 0x4fb9eada // sdot v26.4s, v22.16b, v25.4b[3]\n" - "scvtf v26.4s, v26.4s, #0x4\n" - "fmla v29.4s, v26.4s, v16.4s\n" - "cbnz x21, 2b\n" - "sub %x[nc], %x[nc], #0x4\n" - "str q29, [%x[res_ptr], #0x0]\n" - "add %x[res_ptr], %x[res_ptr], #0x10\n" - "cbnz %x[nc], 1b\n" - : [b_ptr] "+&r" (b_ptr), [res_ptr] "+&r" (res_ptr), [nc] "+&r" (nc) - : [a_ptr] "r" (a_ptr), [nb] "r" (nb) - : "memory", "v16", "v17", "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30", "v31", "x20", "x21", "x22" - ); + for (int c = 0; c < nc; c += ncols_interleaved) { + const block_q8_0 * a_ptr = (const block_q8_0 *) vy; + float32x4_t acc = vdupq_n_f32(0); + for (int b = 0; b < nb; b++) { + int8x16_t b0 = vld1q_s8((const int8_t *) b_ptr->qs); + int8x16_t b1 = vld1q_s8((const int8_t *) b_ptr->qs + 16); + int8x16_t b2 = vld1q_s8((const int8_t *) b_ptr->qs + 32); + int8x16_t b3 = vld1q_s8((const int8_t *) b_ptr->qs + 48); + float16x4_t bd = vld1_f16((const __fp16 *) b_ptr->d); + + int8x16_t a0 = vld1q_s8(a_ptr->qs); + int8x16_t a1 = vld1q_s8(a_ptr->qs + qk/2); + float16x4_t ad = vld1_dup_f16((const __fp16 *) &a_ptr->d); + + int32x4_t ret = vdupq_n_s32(0); + + ret = vdotq_laneq_s32(ret, b0 << 4, a0, 0); + ret = vdotq_laneq_s32(ret, b1 << 4, a0, 1); + ret = vdotq_laneq_s32(ret, b2 << 4, a0, 2); + ret = vdotq_laneq_s32(ret, b3 << 4, a0, 3); + + ret = vdotq_laneq_s32(ret, b0 & 0xf0U, a1, 0); + ret = vdotq_laneq_s32(ret, b1 & 0xf0U, a1, 1); + ret = vdotq_laneq_s32(ret, b2 & 0xf0U, a1, 2); + ret = vdotq_laneq_s32(ret, b3 & 0xf0U, a1, 3); + + acc = vfmaq_f32(acc, vcvtq_n_f32_s32(ret, 4), + vmulq_f32(vcvt_f32_f16(ad), vcvt_f32_f16(bd))); + a_ptr++; + b_ptr++; + } + vst1q_f32(s, acc); + s += ncols_interleaved; + } return; } -#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) +#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) float sumf[4]; int sumi; @@ -704,7 +631,7 @@ void ggml_gemv_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * } } -void ggml_gemv_q4_0_4x8_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemv_q4_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 4; @@ -723,72 +650,52 @@ void ggml_gemv_q4_0_4x8_q8_0(int n, float * restrict s, size_t bs, const void * UNUSED(ncols_interleaved); UNUSED(blocklen); -#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_MATMUL_INT8) - if (ggml_cpu_has_neon() && ggml_cpu_has_matmul_int8()) { - const void * b_ptr = vx; - const void * a_ptr = vy; - float * res_ptr = s; +#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + const block_q4_0x4 * b_ptr = (const block_q4_0x4 *) vx; - __asm__ __volatile__( - "movi v2.16b, #0x4\n" - "movi v1.16b, #0xf0\n" - "add %x[b_ptr], %x[b_ptr], #0x8\n" - "1:" // Column loop - "add x23, %x[a_ptr], #0x2\n" - "movi v0.16b, #0x0\n" - "mov x22, %x[nb]\n" - "2:" // Block loop - "ldr q31, [%x[b_ptr], #0x0]\n" - "ldr q30, [%x[b_ptr], #0x10]\n" - "mov x21, x23\n" - "movi v29.4s, #0x0\n" - "ldr q28, [%x[b_ptr], #0x20]\n" - "ldr q27, [%x[b_ptr], #0x30]\n" - "movi v26.4s, #0x0\n" - "sub x20, x23, #0x2\n" - "ld1r { v25.8h }, [x20]\n" - "ldr q24, [%x[b_ptr], #-0x8]\n" - "sub x22, x22, #0x1\n" - "add x23, x23, #0x22\n" - "ld1r { v23.2d }, [x21], #0x8\n" - "sshl v22.16b, v31.16b, v2.16b\n" - "sshl v16.16b, v30.16b, v2.16b\n" - "add %x[b_ptr], %x[b_ptr], #0x48\n" - "ld1r { v21.2d }, [x21], #0x8\n" - "sshl v20.16b, v28.16b, v2.16b\n" - "sshl v19.16b, v27.16b, v2.16b\n" - "ld1r { v18.2d }, [x21], #0x8\n" - "ld1r { v17.2d }, [x21], #0x8\n" - "and v31.16b, v31.16b, v1.16b\n" - "and v30.16b, v30.16b, v1.16b\n" - ".inst 0x4e9796dd // sdot v29.4s, v22.16b, v23.16b\n" - ".inst 0x4e97961a // sdot v26.4s, v16.16b, v23.16b\n" - "and v28.16b, v28.16b, v1.16b\n" - "and v27.16b, v27.16b, v1.16b\n" - "fcvtl v25.4s, v25.4h\n" - "fcvtl v16.4s, v24.4h\n" - ".inst 0x4e95969d // sdot v29.4s, v20.16b, v21.16b\n" - ".inst 0x4e95967a // sdot v26.4s, v19.16b, v21.16b\n" - "fmul v16.4s, v16.4s, v25.4s\n" - ".inst 0x4e9297fd // sdot v29.4s, v31.16b, v18.16b\n" - ".inst 0x4e9297da // sdot v26.4s, v30.16b, v18.16b\n" - ".inst 0x4e91979d // sdot v29.4s, v28.16b, v17.16b\n" - ".inst 0x4e91977a // sdot v26.4s, v27.16b, v17.16b\n" - "addp v29.4s, v29.4s, v26.4s\n" - "scvtf v29.4s, v29.4s, #0x4\n" - "fmla v0.4s, v29.4s, v16.4s\n" - "cbnz x22, 2b\n" - "sub %x[nc], %x[nc], #0x4\n" - "str q0, [%x[res_ptr], #0x0]\n" - "add %x[res_ptr], %x[res_ptr], #0x10\n" - "cbnz %x[nc], 1b\n" - : [b_ptr] "+&r" (b_ptr), [res_ptr] "+&r" (res_ptr), [nc] "+&r" (nc) - : [a_ptr] "r" (a_ptr), [nb] "r" (nb) - : "memory", "v0", "v1", "v2", "v16", "v17", "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30", "v31", "x20", "x21", "x22", "x23" - ); + for (int c = 0; c < nc; c += ncols_interleaved) { + const block_q8_0 * a_ptr = (const block_q8_0 *) vy; + float32x4_t acc = vdupq_n_f32(0); + for (int b = 0; b < nb; b++) { + int8x16_t b0 = vld1q_s8((const int8_t *) b_ptr->qs); + int8x16_t b1 = vld1q_s8((const int8_t *) b_ptr->qs + 16); + int8x16_t b2 = vld1q_s8((const int8_t *) b_ptr->qs + 32); + int8x16_t b3 = vld1q_s8((const int8_t *) b_ptr->qs + 48); + float16x4_t bd = vld1_f16((const __fp16 *) b_ptr->d); + + int8x16_t a0 = (int8x16_t) vld1q_dup_s64((const int64_t *) a_ptr->qs); + int8x16_t a1 = (int8x16_t) vld1q_dup_s64((const int64_t *) a_ptr->qs + 1); + int8x16_t a2 = (int8x16_t) vld1q_dup_s64((const int64_t *) a_ptr->qs + 2); + int8x16_t a3 = (int8x16_t) vld1q_dup_s64((const int64_t *) a_ptr->qs + 3); + float16x4_t ad = vld1_dup_f16((const __fp16 *) &a_ptr->d); + + int32x4_t ret0 = vdupq_n_s32(0); + int32x4_t ret1 = vdupq_n_s32(0); + + ret0 = vdotq_s32(ret0, b0 << 4, a0); + ret1 = vdotq_s32(ret1, b1 << 4, a0); + ret0 = vdotq_s32(ret0, b2 << 4, a1); + ret1 = vdotq_s32(ret1, b3 << 4, a1); + + ret0 = vdotq_s32(ret0, b0 & 0xf0U, a2); + ret1 = vdotq_s32(ret1, b1 & 0xf0U, a2); + ret0 = vdotq_s32(ret0, b2 & 0xf0U, a3); + ret1 = vdotq_s32(ret1, b3 & 0xf0U, a3); + + int32x4_t ret = vpaddq_s32(ret0, ret1); + + acc = vfmaq_f32(acc, vcvtq_n_f32_s32(ret, 4), + vmulq_f32(vcvt_f32_f16(ad), vcvt_f32_f16(bd))); + a_ptr++; + b_ptr++; + } + vst1q_f32(s, acc); + s += ncols_interleaved; + } return; } -#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_MATMUL_INT8) +#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) float sumf[4]; int sumi; @@ -814,7 +721,7 @@ void ggml_gemv_q4_0_4x8_q8_0(int n, float * restrict s, size_t bs, const void * } } -void ggml_gemv_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemv_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 8; @@ -1087,7 +994,103 @@ void ggml_gemv_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * } } -void ggml_gemm_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemv_iq4_nl_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { + const int qk = QK8_0; + const int nb = n / qk; + const int ncols_interleaved = 4; + const int blocklen = 4; + + assert (n % qk == 0); + assert (nc % ncols_interleaved == 0); + + UNUSED(s); + UNUSED(bs); + UNUSED(vx); + UNUSED(vy); + UNUSED(nr); + UNUSED(nc); + UNUSED(nb); + UNUSED(ncols_interleaved); + UNUSED(blocklen); + +#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + const int8x16_t kvalues = vld1q_s8(kvalues_iq4nl); + const block_q8_0 * a_ptr = (const block_q8_0 *) vy; + float * res_ptr = s; + + for (int x = 0; x < nc / ncols_interleaved; x++) { + const block_iq4_nlx4 * b_ptr = (const block_iq4_nlx4 *) vx + (x * nb); + + float32x4_t sumf = vdupq_n_f32(0); + for (int l = 0; l < nb; l++) { + uint8x16_t b_0 = vld1q_u8(b_ptr[l].qs + 0); + uint8x16_t b_1 = vld1q_u8(b_ptr[l].qs + 16); + uint8x16_t b_2 = vld1q_u8(b_ptr[l].qs + 32); + uint8x16_t b_3 = vld1q_u8(b_ptr[l].qs + 48); + + int8x16_t b_0_hi = vqtbl1q_s8(kvalues, b_0 >> 4); + int8x16_t b_0_lo = vqtbl1q_s8(kvalues, b_0 & 0x0F); + int8x16_t b_1_hi = vqtbl1q_s8(kvalues, b_1 >> 4); + int8x16_t b_1_lo = vqtbl1q_s8(kvalues, b_1 & 0x0F); + int8x16_t b_2_hi = vqtbl1q_s8(kvalues, b_2 >> 4); + int8x16_t b_2_lo = vqtbl1q_s8(kvalues, b_2 & 0x0F); + int8x16_t b_3_hi = vqtbl1q_s8(kvalues, b_3 >> 4); + int8x16_t b_3_lo = vqtbl1q_s8(kvalues, b_3 & 0x0F); + + int8x16_t a_0 = vld1q_s8(a_ptr[l].qs + 0); + int8x16_t a_1 = vld1q_s8(a_ptr[l].qs + 16); + + int32x4_t sumi = vdupq_n_s32(0); + sumi = vdotq_laneq_s32(sumi, b_0_lo, a_0, 0); + sumi = vdotq_laneq_s32(sumi, b_0_hi, a_1, 0); + sumi = vdotq_laneq_s32(sumi, b_1_lo, a_0, 1); + sumi = vdotq_laneq_s32(sumi, b_1_hi, a_1, 1); + sumi = vdotq_laneq_s32(sumi, b_2_lo, a_0, 2); + sumi = vdotq_laneq_s32(sumi, b_2_hi, a_1, 2); + sumi = vdotq_laneq_s32(sumi, b_3_lo, a_0, 3); + sumi = vdotq_laneq_s32(sumi, b_3_hi, a_1, 3); + + float32x4_t a_d = vcvt_f32_f16(vld1_dup_f16((const float16_t *)&a_ptr[l].d)); + float32x4_t b_d = vcvt_f32_f16(vld1_f16((const float16_t *)b_ptr[l].d)); + float32x4_t d = a_d * b_d; + + sumf = vmlaq_f32(sumf, d, vcvtq_f32_s32(sumi)); + } + + vst1q_f32(res_ptr + x * 4, sumf); + } + return; + } +#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) + { + float sumf[4]; + int sumi; + + const block_q8_0 * a_ptr = (const block_q8_0 *) vy; + for (int x = 0; x < nc / ncols_interleaved; x++) { + const block_iq4_nlx4 * b_ptr = (const block_iq4_nlx4 *) vx + (x * nb); + + for (int j = 0; j < ncols_interleaved; j++) sumf[j] = 0.0; + for (int l = 0; l < nb; l++) { + for (int k = 0; k < (qk / (2 * blocklen)); k++) { + for (int j = 0; j < ncols_interleaved; j++) { + sumi = 0; + for (int i = 0; i < blocklen; ++i) { + const int v0 = kvalues_iq4nl[b_ptr[l].qs[k * ncols_interleaved * blocklen + j * blocklen + i] & 0x0F]; + const int v1 = kvalues_iq4nl[b_ptr[l].qs[k * ncols_interleaved * blocklen + j * blocklen + i] >> 4]; + sumi += ((v0 * a_ptr[l].qs[k * blocklen + i]) + (v1 * a_ptr[l].qs[k * blocklen + i + qk / 2])); + } + sumf[j] += sumi * GGML_FP16_TO_FP32(b_ptr[l].d[j]) * GGML_FP16_TO_FP32(a_ptr[l].d); + } + } + } + for (int j = 0; j < ncols_interleaved; j++) s[x * ncols_interleaved + j] = sumf[j]; + } + } +} + +static void ggml_gemm_q4_0_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 4; @@ -1108,7 +1111,7 @@ void ggml_gemm_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * UNUSED(blocklen); #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) - if (ggml_cpu_has_neon()) { + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { const void * b_ptr = vx; const void * a_ptr = vy; float * res_ptr = s; @@ -1603,7 +1606,7 @@ void ggml_gemm_q4_0_4x4_q8_0(int n, float * restrict s, size_t bs, const void * } } -void ggml_gemm_q4_0_4x8_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemm_q4_0_4x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 4; @@ -2057,7 +2060,7 @@ void ggml_gemm_q4_0_4x8_q8_0(int n, float * restrict s, size_t bs, const void * } } -void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, const void * restrict vy, int nr, int nc) { +static void ggml_gemm_q4_0_8x8_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { const int qk = QK8_0; const int nb = n / qk; const int ncols_interleaved = 8; @@ -2577,31 +2580,31 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * const __m512i rhs_mat_2367ABEF_3 = _mm512_shuffle_epi8(signextendlutexpanded, _mm512_and_si512(_mm512_srli_epi16(rhs_raw_mat_2367ABEF_1, 4), m4bexpanded)); //B2(24-31) B3(24-31) B6(24-31) B7(24-31) BA(24-31) BB(24-31) BE(24-31) BF(24-31) // Shuffle pattern one - right side input - const __m512i rhs_mat_014589CD_0_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, 136); //B0(0-3) B1(0-3) B0(0-3) B1(0-3) B4(0-3) B5(0-3) B4(0-3) B5(0-3) B8(0-3) B9(0-3) B8(0-3) B9(0-3) BC(0-3) BD(0-3) BC(0-3) BD(0-3) - const __m512i rhs_mat_2367ABEF_0_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, 136); //B2(0-3) B3(0-3) B2(0-3) B3(0-3) B6(0-3) B7(0-3) B6(0-3) B7(0-3) BA(0-3) BB(0-3) BA(0-3) BB(0-3) BE(0-3) BF(0-3) BE(0-3) BF(0-3) + const __m512i rhs_mat_014589CD_0_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, (_MM_PERM_ENUM)136); //B0(0-3) B1(0-3) B0(0-3) B1(0-3) B4(0-3) B5(0-3) B4(0-3) B5(0-3) B8(0-3) B9(0-3) B8(0-3) B9(0-3) BC(0-3) BD(0-3) BC(0-3) BD(0-3) + const __m512i rhs_mat_2367ABEF_0_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, (_MM_PERM_ENUM)136); //B2(0-3) B3(0-3) B2(0-3) B3(0-3) B6(0-3) B7(0-3) B6(0-3) B7(0-3) BA(0-3) BB(0-3) BA(0-3) BB(0-3) BE(0-3) BF(0-3) BE(0-3) BF(0-3) - const __m512i rhs_mat_014589CD_1_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, 136); //B0(8-11) B1(8-11) B0(8-11) B1(8-11) B4(8-11) B5(8-11) B4(8-11) B5(8-11) B8(8-11) B9(8-11) B8(8-11) B9(8-11) BC(8-11) BD(8-11) BC(8-11) BD(8-11) - const __m512i rhs_mat_2367ABEF_1_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, 136); //B2(8-11) B3(8-11) B2(8-11) B3(8-11) B6(8-11) B7(8-11) B6(8-11) B7(8-11) BA(8-11) BB(8-11) BA(8-11) BB(8-11) BE(8-11) BF(8-11) BE(8-11) BF(8-11) + const __m512i rhs_mat_014589CD_1_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, (_MM_PERM_ENUM)136); //B0(8-11) B1(8-11) B0(8-11) B1(8-11) B4(8-11) B5(8-11) B4(8-11) B5(8-11) B8(8-11) B9(8-11) B8(8-11) B9(8-11) BC(8-11) BD(8-11) BC(8-11) BD(8-11) + const __m512i rhs_mat_2367ABEF_1_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, (_MM_PERM_ENUM)136); //B2(8-11) B3(8-11) B2(8-11) B3(8-11) B6(8-11) B7(8-11) B6(8-11) B7(8-11) BA(8-11) BB(8-11) BA(8-11) BB(8-11) BE(8-11) BF(8-11) BE(8-11) BF(8-11) - const __m512i rhs_mat_014589CD_2_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, 136); //B0(16-19) B1(16-19) B0(16-19) B1(16-19) B4(16-19) B5(16-19) B4(16-19) B5(16-19) B8(16-19) B9(16-19) B8(16-19) B9(16-19) BC(16-19) BD(16-19) BC(16-19) BD(16-19) - const __m512i rhs_mat_2367ABEF_2_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, 136); //B2(16-19) B3(16-19) B2(16-19) B3(16-19) B6(16-19) B7(16-19) B6(16-19) B7(16-19) BA(16-19) BB(16-19) BA(16-19) BB(16-19) BE(16-19) BF(16-19) BE(16-19) BF(16-19) + const __m512i rhs_mat_014589CD_2_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, (_MM_PERM_ENUM)136); //B0(16-19) B1(16-19) B0(16-19) B1(16-19) B4(16-19) B5(16-19) B4(16-19) B5(16-19) B8(16-19) B9(16-19) B8(16-19) B9(16-19) BC(16-19) BD(16-19) BC(16-19) BD(16-19) + const __m512i rhs_mat_2367ABEF_2_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, (_MM_PERM_ENUM)136); //B2(16-19) B3(16-19) B2(16-19) B3(16-19) B6(16-19) B7(16-19) B6(16-19) B7(16-19) BA(16-19) BB(16-19) BA(16-19) BB(16-19) BE(16-19) BF(16-19) BE(16-19) BF(16-19) - const __m512i rhs_mat_014589CD_3_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, 136); //B0(24-27) B1(24-27) B0(24-27) B1(24-27) B4(24-27) B5(24-27) B4(24-27) B5(24-27) B8(24-27) B9(24-27) B8(24-27) B9(24-27) BC(24-27) BD(24-27) BC(24-27) BD(24-27) - const __m512i rhs_mat_2367ABEF_3_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, 136); //B2(24-27) B3(24-27) B2(24-27) B3(24-27) B6(24-27) B7(24-27) B6(24-27) B7(24-27) BA(24-27) BB(24-27) BA(24-27) BB(24-27) BE(24-27) BF(24-27) BE(24-27) BF(24-27) + const __m512i rhs_mat_014589CD_3_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, (_MM_PERM_ENUM)136); //B0(24-27) B1(24-27) B0(24-27) B1(24-27) B4(24-27) B5(24-27) B4(24-27) B5(24-27) B8(24-27) B9(24-27) B8(24-27) B9(24-27) BC(24-27) BD(24-27) BC(24-27) BD(24-27) + const __m512i rhs_mat_2367ABEF_3_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, (_MM_PERM_ENUM)136); //B2(24-27) B3(24-27) B2(24-27) B3(24-27) B6(24-27) B7(24-27) B6(24-27) B7(24-27) BA(24-27) BB(24-27) BA(24-27) BB(24-27) BE(24-27) BF(24-27) BE(24-27) BF(24-27) // Shuffle pattern two - right side input - const __m512i rhs_mat_014589CD_0_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, 221); //B0(4-7) B1(4-7) B0(4-7) B1(4-7) B4(4-7) B5(4-7) B4(4-7) B5(4-7) B8(4-7) B9(4-7) B8(4-7) B9(4-7) BC(4-7) BD(4-7) BC(4-7) BD(4-7) - const __m512i rhs_mat_2367ABEF_0_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, 221); //B2(4-7) B3(4-7) B2(4-7) B3(4-7) B6(4-7) B7(4-7) B6(4-7) B7(4-7) BA(4-7) BB(4-7) BA(4-7) BB(4-7) BE(4-7) BF(4-7) BE(4-7) BF(4-7) + const __m512i rhs_mat_014589CD_0_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, (_MM_PERM_ENUM)221); //B0(4-7) B1(4-7) B0(4-7) B1(4-7) B4(4-7) B5(4-7) B4(4-7) B5(4-7) B8(4-7) B9(4-7) B8(4-7) B9(4-7) BC(4-7) BD(4-7) BC(4-7) BD(4-7) + const __m512i rhs_mat_2367ABEF_0_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, (_MM_PERM_ENUM)221); //B2(4-7) B3(4-7) B2(4-7) B3(4-7) B6(4-7) B7(4-7) B6(4-7) B7(4-7) BA(4-7) BB(4-7) BA(4-7) BB(4-7) BE(4-7) BF(4-7) BE(4-7) BF(4-7) - const __m512i rhs_mat_014589CD_1_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, 221); //B0(12-15) B1(12-15) B0(12-15) B1(12-15) B4(12-15) B5(12-15) B4(12-15) B5(12-15) B8(12-15) B9(12-15) B8(12-15) B9(12-15) BC(12-15) BD(12-15) BC(12-15) BD(12-15) - const __m512i rhs_mat_2367ABEF_1_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, 221); //B2(12-15) B3(12-15) B2(12-15) B3(12-15) B6(12-15) B7(12-15) B6(12-15) B7(12-15) BA(12-15) BB(12-15) BA(12-15) BB(12-15) BE(12-15) BF(12-15) BE(12-15) BF(12-15) + const __m512i rhs_mat_014589CD_1_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, (_MM_PERM_ENUM)221); //B0(12-15) B1(12-15) B0(12-15) B1(12-15) B4(12-15) B5(12-15) B4(12-15) B5(12-15) B8(12-15) B9(12-15) B8(12-15) B9(12-15) BC(12-15) BD(12-15) BC(12-15) BD(12-15) + const __m512i rhs_mat_2367ABEF_1_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, (_MM_PERM_ENUM)221); //B2(12-15) B3(12-15) B2(12-15) B3(12-15) B6(12-15) B7(12-15) B6(12-15) B7(12-15) BA(12-15) BB(12-15) BA(12-15) BB(12-15) BE(12-15) BF(12-15) BE(12-15) BF(12-15) - const __m512i rhs_mat_014589CD_2_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, 221); //B0(20-23) B1(20-23) B0(20-23) B1(20-23) B4(20-23) B5(20-23) B4(20-23) B5(20-23) B8(20-23) B9(20-23) B8(20-23) B9(20-23) BC(20-23) BD(20-23) BC(20-23) BD(20-23) - const __m512i rhs_mat_2367ABEF_2_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, 221); //B2(20-23) B3(20-23) B2(20-23) B3(20-23) B6(20-23) B7(20-23) B6(20-23) B7(20-23) BA(20-23) BB(20-23) BA(20-23) BB(20-23) BE(20-23) BF(20-23) BE(20-23) BF(20-23) + const __m512i rhs_mat_014589CD_2_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, (_MM_PERM_ENUM)221); //B0(20-23) B1(20-23) B0(20-23) B1(20-23) B4(20-23) B5(20-23) B4(20-23) B5(20-23) B8(20-23) B9(20-23) B8(20-23) B9(20-23) BC(20-23) BD(20-23) BC(20-23) BD(20-23) + const __m512i rhs_mat_2367ABEF_2_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, (_MM_PERM_ENUM)221); //B2(20-23) B3(20-23) B2(20-23) B3(20-23) B6(20-23) B7(20-23) B6(20-23) B7(20-23) BA(20-23) BB(20-23) BA(20-23) BB(20-23) BE(20-23) BF(20-23) BE(20-23) BF(20-23) - const __m512i rhs_mat_014589CD_3_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, 221); //B0(28-31) B1(28-31) B0(28-31) B1(28-31) B4(28-31) B5(28-31) B4(28-31) B5(28-31) B8(28-31) B9(28-31) B8(28-31) B9(28-31) BC(28-31) BD(28-31) BC(28-31) BD(28-31) - const __m512i rhs_mat_2367ABEF_3_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, 221); //B2(28-31) B3(28-31) B2(28-31) B3(28-31) B6(28-31) B7(28-31) B6(28-31) B7(28-31) BA(28-31) BB(28-31) BA(28-31) BB(28-31) BE(28-31) BF(28-31) BE(28-31) BF(28-31) + const __m512i rhs_mat_014589CD_3_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, (_MM_PERM_ENUM)221); //B0(28-31) B1(28-31) B0(28-31) B1(28-31) B4(28-31) B5(28-31) B4(28-31) B5(28-31) B8(28-31) B9(28-31) B8(28-31) B9(28-31) BC(28-31) BD(28-31) BC(28-31) BD(28-31) + const __m512i rhs_mat_2367ABEF_3_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, (_MM_PERM_ENUM)221); //B2(28-31) B3(28-31) B2(28-31) B3(28-31) B6(28-31) B7(28-31) B6(28-31) B7(28-31) BA(28-31) BB(28-31) BA(28-31) BB(28-31) BE(28-31) BF(28-31) BE(28-31) BF(28-31) // Scale values - Load the weight scale values of two block_q4_0x8 const __m512 col_scale_f32 = GGML_F32Cx8x2_LOAD(b_ptr_0[b].d, b_ptr_1[b].d); @@ -2635,31 +2638,31 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * // Shuffle pattern one - left side input - const __m512i lhs_mat_01_0_sp1 = _mm512_shuffle_epi32(lhs_mat_01_0, 160); //A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) - const __m512i lhs_mat_23_0_sp1 = _mm512_shuffle_epi32(lhs_mat_23_0, 160); //A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) + const __m512i lhs_mat_01_0_sp1 = _mm512_shuffle_epi32(lhs_mat_01_0, (_MM_PERM_ENUM)160); //A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) + const __m512i lhs_mat_23_0_sp1 = _mm512_shuffle_epi32(lhs_mat_23_0, (_MM_PERM_ENUM)160); //A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) - const __m512i lhs_mat_01_1_sp1 = _mm512_shuffle_epi32(lhs_mat_01_1, 160); //A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) - const __m512i lhs_mat_23_1_sp1 = _mm512_shuffle_epi32(lhs_mat_23_1, 160); //A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) + const __m512i lhs_mat_01_1_sp1 = _mm512_shuffle_epi32(lhs_mat_01_1, (_MM_PERM_ENUM)160); //A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) + const __m512i lhs_mat_23_1_sp1 = _mm512_shuffle_epi32(lhs_mat_23_1, (_MM_PERM_ENUM)160); //A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) - const __m512i lhs_mat_01_2_sp1 = _mm512_shuffle_epi32(lhs_mat_01_2, 160); //A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) - const __m512i lhs_mat_23_2_sp1 = _mm512_shuffle_epi32(lhs_mat_23_2, 160); //A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) + const __m512i lhs_mat_01_2_sp1 = _mm512_shuffle_epi32(lhs_mat_01_2, (_MM_PERM_ENUM)160); //A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) + const __m512i lhs_mat_23_2_sp1 = _mm512_shuffle_epi32(lhs_mat_23_2, (_MM_PERM_ENUM)160); //A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) - const __m512i lhs_mat_01_3_sp1 = _mm512_shuffle_epi32(lhs_mat_01_3, 160); //A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) - const __m512i lhs_mat_23_3_sp1 = _mm512_shuffle_epi32(lhs_mat_23_3, 160); //A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) + const __m512i lhs_mat_01_3_sp1 = _mm512_shuffle_epi32(lhs_mat_01_3, (_MM_PERM_ENUM)160); //A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) + const __m512i lhs_mat_23_3_sp1 = _mm512_shuffle_epi32(lhs_mat_23_3, (_MM_PERM_ENUM)160); //A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) // Shuffle pattern two - left side input - const __m512i lhs_mat_01_0_sp2 = _mm512_shuffle_epi32(lhs_mat_01_0, 245); //A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) - const __m512i lhs_mat_23_0_sp2 = _mm512_shuffle_epi32(lhs_mat_23_0, 245); //A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) + const __m512i lhs_mat_01_0_sp2 = _mm512_shuffle_epi32(lhs_mat_01_0, (_MM_PERM_ENUM)245); //A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) + const __m512i lhs_mat_23_0_sp2 = _mm512_shuffle_epi32(lhs_mat_23_0, (_MM_PERM_ENUM)245); //A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) - const __m512i lhs_mat_01_1_sp2 = _mm512_shuffle_epi32(lhs_mat_01_1, 245); //A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) - const __m512i lhs_mat_23_1_sp2 = _mm512_shuffle_epi32(lhs_mat_23_1, 245); //A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) + const __m512i lhs_mat_01_1_sp2 = _mm512_shuffle_epi32(lhs_mat_01_1, (_MM_PERM_ENUM)245); //A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) + const __m512i lhs_mat_23_1_sp2 = _mm512_shuffle_epi32(lhs_mat_23_1, (_MM_PERM_ENUM)245); //A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) - const __m512i lhs_mat_01_2_sp2 = _mm512_shuffle_epi32(lhs_mat_01_2, 245); //A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) - const __m512i lhs_mat_23_2_sp2 = _mm512_shuffle_epi32(lhs_mat_23_2, 245); //A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) + const __m512i lhs_mat_01_2_sp2 = _mm512_shuffle_epi32(lhs_mat_01_2, (_MM_PERM_ENUM)245); //A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) + const __m512i lhs_mat_23_2_sp2 = _mm512_shuffle_epi32(lhs_mat_23_2, (_MM_PERM_ENUM)245); //A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) - const __m512i lhs_mat_01_3_sp2 = _mm512_shuffle_epi32(lhs_mat_01_3, 245); //A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) - const __m512i lhs_mat_23_3_sp2 = _mm512_shuffle_epi32(lhs_mat_23_3, 245); //A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) + const __m512i lhs_mat_01_3_sp2 = _mm512_shuffle_epi32(lhs_mat_01_3, (_MM_PERM_ENUM)245); //A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) + const __m512i lhs_mat_23_3_sp2 = _mm512_shuffle_epi32(lhs_mat_23_3, (_MM_PERM_ENUM)245); //A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) // The values arranged in shuffle patterns are operated with dot product operation within 32 bit lane i.e corresponding bytes and multiplied and added into 32 bit integers within 32 bit lane // Resembles MMLAs into 2x2 matrices in ARM Version @@ -2688,10 +2691,10 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * // Straighten out to make 4 row vectors - __m512i iacc_row_0 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_00, _mm512_shuffle_epi32(iacc_mat_01, 78)); - __m512i iacc_row_1 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_00, 78), iacc_mat_01); - __m512i iacc_row_2 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_10, _mm512_shuffle_epi32(iacc_mat_11, 78)); - __m512i iacc_row_3 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_10, 78), iacc_mat_11); + __m512i iacc_row_0 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_00, _mm512_shuffle_epi32(iacc_mat_01, (_MM_PERM_ENUM)78)); + __m512i iacc_row_1 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_00, (_MM_PERM_ENUM)78), iacc_mat_01); + __m512i iacc_row_2 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_10, _mm512_shuffle_epi32(iacc_mat_11, (_MM_PERM_ENUM)78)); + __m512i iacc_row_3 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_10, (_MM_PERM_ENUM)78), iacc_mat_11); // Load the scale(d) values for all the 4 Q8_0 blocks and repeat it across lanes const __m128i row_scale_f16 = _mm_shuffle_epi32(_mm_maskload_epi32((int const*)(a_ptrs[rp][b].d), loadMask), 68); @@ -2770,31 +2773,31 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * const __m512i rhs_mat_2367ABEF_3 = _mm512_shuffle_epi8(signextendlutexpanded, _mm512_and_si512(_mm512_srli_epi16(rhs_raw_mat_2367ABEF_1, 4), m4bexpanded)); //B2(24-31) B3(24-31) B6(24-31) B7(24-31) BA(24-31) BB(24-31) BE(24-31) BF(24-31) // Shuffle pattern one - right side input - const __m512i rhs_mat_014589CD_0_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, 136); //B0(0-3) B1(0-3) B0(0-3) B1(0-3) B4(0-3) B5(0-3) B4(0-3) B5(0-3) B8(0-3) B9(0-3) B8(0-3) B9(0-3) BC(0-3) BD(0-3) BC(0-3) BD(0-3) - const __m512i rhs_mat_2367ABEF_0_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, 136); //B2(0-3) B3(0-3) B2(0-3) B3(0-3) B6(0-3) B7(0-3) B6(0-3) B7(0-3) BA(0-3) BB(0-3) BA(0-3) BB(0-3) BE(0-3) BF(0-3) BE(0-3) BF(0-3) + const __m512i rhs_mat_014589CD_0_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, (_MM_PERM_ENUM)136); //B0(0-3) B1(0-3) B0(0-3) B1(0-3) B4(0-3) B5(0-3) B4(0-3) B5(0-3) B8(0-3) B9(0-3) B8(0-3) B9(0-3) BC(0-3) BD(0-3) BC(0-3) BD(0-3) + const __m512i rhs_mat_2367ABEF_0_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, (_MM_PERM_ENUM)136); //B2(0-3) B3(0-3) B2(0-3) B3(0-3) B6(0-3) B7(0-3) B6(0-3) B7(0-3) BA(0-3) BB(0-3) BA(0-3) BB(0-3) BE(0-3) BF(0-3) BE(0-3) BF(0-3) - const __m512i rhs_mat_014589CD_1_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, 136); //B0(8-11) B1(8-11) B0(8-11) B1(8-11) B4(8-11) B5(8-11) B4(8-11) B5(8-11) B8(8-11) B9(8-11) B8(8-11) B9(8-11) BC(8-11) BD(8-11) BC(8-11) BD(8-11) - const __m512i rhs_mat_2367ABEF_1_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, 136); //B2(8-11) B3(8-11) B2(8-11) B3(8-11) B6(8-11) B7(8-11) B6(8-11) B7(8-11) BA(8-11) BB(8-11) BA(8-11) BB(8-11) BE(8-11) BF(8-11) BE(8-11) BF(8-11) + const __m512i rhs_mat_014589CD_1_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, (_MM_PERM_ENUM)136); //B0(8-11) B1(8-11) B0(8-11) B1(8-11) B4(8-11) B5(8-11) B4(8-11) B5(8-11) B8(8-11) B9(8-11) B8(8-11) B9(8-11) BC(8-11) BD(8-11) BC(8-11) BD(8-11) + const __m512i rhs_mat_2367ABEF_1_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, (_MM_PERM_ENUM)136); //B2(8-11) B3(8-11) B2(8-11) B3(8-11) B6(8-11) B7(8-11) B6(8-11) B7(8-11) BA(8-11) BB(8-11) BA(8-11) BB(8-11) BE(8-11) BF(8-11) BE(8-11) BF(8-11) - const __m512i rhs_mat_014589CD_2_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, 136); //B0(16-19) B1(16-19) B0(16-19) B1(16-19) B4(16-19) B5(16-19) B4(16-19) B5(16-19) B8(16-19) B9(16-19) B8(16-19) B9(16-19) BC(16-19) BD(16-19) BC(16-19) BD(16-19) - const __m512i rhs_mat_2367ABEF_2_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, 136); //B2(16-19) B3(16-19) B2(16-19) B3(16-19) B6(16-19) B7(16-19) B6(16-19) B7(16-19) BA(16-19) BB(16-19) BA(16-19) BB(16-19) BE(16-19) BF(16-19) BE(16-19) BF(16-19) + const __m512i rhs_mat_014589CD_2_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, (_MM_PERM_ENUM)136); //B0(16-19) B1(16-19) B0(16-19) B1(16-19) B4(16-19) B5(16-19) B4(16-19) B5(16-19) B8(16-19) B9(16-19) B8(16-19) B9(16-19) BC(16-19) BD(16-19) BC(16-19) BD(16-19) + const __m512i rhs_mat_2367ABEF_2_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, (_MM_PERM_ENUM)136); //B2(16-19) B3(16-19) B2(16-19) B3(16-19) B6(16-19) B7(16-19) B6(16-19) B7(16-19) BA(16-19) BB(16-19) BA(16-19) BB(16-19) BE(16-19) BF(16-19) BE(16-19) BF(16-19) - const __m512i rhs_mat_014589CD_3_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, 136); //B0(24-27) B1(24-27) B0(24-27) B1(24-27) B4(24-27) B5(24-27) B4(24-27) B5(24-27) B8(24-27) B9(24-27) B8(24-27) B9(24-27) BC(24-27) BD(24-27) BC(24-27) BD(24-27) - const __m512i rhs_mat_2367ABEF_3_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, 136); //B2(24-27) B3(24-27) B2(24-27) B3(24-27) B6(24-27) B7(24-27) B6(24-27) B7(24-27) BA(24-27) BB(24-27) BA(24-27) BB(24-27) BE(24-27) BF(24-27) BE(24-27) BF(24-27) + const __m512i rhs_mat_014589CD_3_sp1 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, (_MM_PERM_ENUM)136); //B0(24-27) B1(24-27) B0(24-27) B1(24-27) B4(24-27) B5(24-27) B4(24-27) B5(24-27) B8(24-27) B9(24-27) B8(24-27) B9(24-27) BC(24-27) BD(24-27) BC(24-27) BD(24-27) + const __m512i rhs_mat_2367ABEF_3_sp1 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, (_MM_PERM_ENUM)136); //B2(24-27) B3(24-27) B2(24-27) B3(24-27) B6(24-27) B7(24-27) B6(24-27) B7(24-27) BA(24-27) BB(24-27) BA(24-27) BB(24-27) BE(24-27) BF(24-27) BE(24-27) BF(24-27) // Shuffle pattern two - right side input - const __m512i rhs_mat_014589CD_0_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, 221); //B0(4-7) B1(4-7) B0(4-7) B1(4-7) B4(4-7) B5(4-7) B4(4-7) B5(4-7) B8(4-7) B9(4-7) B8(4-7) B9(4-7) BC(4-7) BD(4-7) BC(4-7) BD(4-7) - const __m512i rhs_mat_2367ABEF_0_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, 221); //B2(4-7) B3(4-7) B2(4-7) B3(4-7) B6(4-7) B7(4-7) B6(4-7) B7(4-7) BA(4-7) BB(4-7) BA(4-7) BB(4-7) BE(4-7) BF(4-7) BE(4-7) BF(4-7) + const __m512i rhs_mat_014589CD_0_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_0, (_MM_PERM_ENUM)221); //B0(4-7) B1(4-7) B0(4-7) B1(4-7) B4(4-7) B5(4-7) B4(4-7) B5(4-7) B8(4-7) B9(4-7) B8(4-7) B9(4-7) BC(4-7) BD(4-7) BC(4-7) BD(4-7) + const __m512i rhs_mat_2367ABEF_0_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_0, (_MM_PERM_ENUM)221); //B2(4-7) B3(4-7) B2(4-7) B3(4-7) B6(4-7) B7(4-7) B6(4-7) B7(4-7) BA(4-7) BB(4-7) BA(4-7) BB(4-7) BE(4-7) BF(4-7) BE(4-7) BF(4-7) - const __m512i rhs_mat_014589CD_1_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, 221); //B0(12-15) B1(12-15) B0(12-15) B1(12-15) B4(12-15) B5(12-15) B4(12-15) B5(12-15) B8(12-15) B9(12-15) B8(12-15) B9(12-15) BC(12-15) BD(12-15) BC(12-15) BD(12-15) - const __m512i rhs_mat_2367ABEF_1_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, 221); //B2(12-15) B3(12-15) B2(12-15) B3(12-15) B6(12-15) B7(12-15) B6(12-15) B7(12-15) BA(12-15) BB(12-15) BA(12-15) BB(12-15) BE(12-15) BF(12-15) BE(12-15) BF(12-15) + const __m512i rhs_mat_014589CD_1_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_1, (_MM_PERM_ENUM)221); //B0(12-15) B1(12-15) B0(12-15) B1(12-15) B4(12-15) B5(12-15) B4(12-15) B5(12-15) B8(12-15) B9(12-15) B8(12-15) B9(12-15) BC(12-15) BD(12-15) BC(12-15) BD(12-15) + const __m512i rhs_mat_2367ABEF_1_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_1, (_MM_PERM_ENUM)221); //B2(12-15) B3(12-15) B2(12-15) B3(12-15) B6(12-15) B7(12-15) B6(12-15) B7(12-15) BA(12-15) BB(12-15) BA(12-15) BB(12-15) BE(12-15) BF(12-15) BE(12-15) BF(12-15) - const __m512i rhs_mat_014589CD_2_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, 221); //B0(20-23) B1(20-23) B0(20-23) B1(20-23) B4(20-23) B5(20-23) B4(20-23) B5(20-23) B8(20-23) B9(20-23) B8(20-23) B9(20-23) BC(20-23) BD(20-23) BC(20-23) BD(20-23) - const __m512i rhs_mat_2367ABEF_2_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, 221); //B2(20-23) B3(20-23) B2(20-23) B3(20-23) B6(20-23) B7(20-23) B6(20-23) B7(20-23) BA(20-23) BB(20-23) BA(20-23) BB(20-23) BE(20-23) BF(20-23) BE(20-23) BF(20-23) + const __m512i rhs_mat_014589CD_2_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_2, (_MM_PERM_ENUM)221); //B0(20-23) B1(20-23) B0(20-23) B1(20-23) B4(20-23) B5(20-23) B4(20-23) B5(20-23) B8(20-23) B9(20-23) B8(20-23) B9(20-23) BC(20-23) BD(20-23) BC(20-23) BD(20-23) + const __m512i rhs_mat_2367ABEF_2_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_2, (_MM_PERM_ENUM)221); //B2(20-23) B3(20-23) B2(20-23) B3(20-23) B6(20-23) B7(20-23) B6(20-23) B7(20-23) BA(20-23) BB(20-23) BA(20-23) BB(20-23) BE(20-23) BF(20-23) BE(20-23) BF(20-23) - const __m512i rhs_mat_014589CD_3_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, 221); //B0(28-31) B1(28-31) B0(28-31) B1(28-31) B4(28-31) B5(28-31) B4(28-31) B5(28-31) B8(28-31) B9(28-31) B8(28-31) B9(28-31) BC(28-31) BD(28-31) BC(28-31) BD(28-31) - const __m512i rhs_mat_2367ABEF_3_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, 221); //B2(28-31) B3(28-31) B2(28-31) B3(28-31) B6(28-31) B7(28-31) B6(28-31) B7(28-31) BA(28-31) BB(28-31) BA(28-31) BB(28-31) BE(28-31) BF(28-31) BE(28-31) BF(28-31) + const __m512i rhs_mat_014589CD_3_sp2 = _mm512_shuffle_epi32(rhs_mat_014589CD_3, (_MM_PERM_ENUM)221); //B0(28-31) B1(28-31) B0(28-31) B1(28-31) B4(28-31) B5(28-31) B4(28-31) B5(28-31) B8(28-31) B9(28-31) B8(28-31) B9(28-31) BC(28-31) BD(28-31) BC(28-31) BD(28-31) + const __m512i rhs_mat_2367ABEF_3_sp2 = _mm512_shuffle_epi32(rhs_mat_2367ABEF_3, (_MM_PERM_ENUM)221); //B2(28-31) B3(28-31) B2(28-31) B3(28-31) B6(28-31) B7(28-31) B6(28-31) B7(28-31) BA(28-31) BB(28-31) BA(28-31) BB(28-31) BE(28-31) BF(28-31) BE(28-31) BF(28-31) // Scale values - Load the weight scale values of two block_q4_0x8 @@ -2826,31 +2829,31 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * // Shuffle pattern one - left side input - const __m512i lhs_mat_01_0_sp1 = _mm512_shuffle_epi32(lhs_mat_01_0, 160); //A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) - const __m512i lhs_mat_23_0_sp1 = _mm512_shuffle_epi32(lhs_mat_23_0, 160); //A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) + const __m512i lhs_mat_01_0_sp1 = _mm512_shuffle_epi32(lhs_mat_01_0, (_MM_PERM_ENUM)160); //A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) A0(0-3) A0(0-3) A1(0-3) A1(0-3) + const __m512i lhs_mat_23_0_sp1 = _mm512_shuffle_epi32(lhs_mat_23_0, (_MM_PERM_ENUM)160); //A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) A2(0-3) A2(0-3) A3(0-3) A3(0-3) - const __m512i lhs_mat_01_1_sp1 = _mm512_shuffle_epi32(lhs_mat_01_1, 160); //A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) - const __m512i lhs_mat_23_1_sp1 = _mm512_shuffle_epi32(lhs_mat_23_1, 160); //A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) + const __m512i lhs_mat_01_1_sp1 = _mm512_shuffle_epi32(lhs_mat_01_1, (_MM_PERM_ENUM)160); //A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) A0(8-11) A0(8-11) A1(8-11) A1(8-11) + const __m512i lhs_mat_23_1_sp1 = _mm512_shuffle_epi32(lhs_mat_23_1, (_MM_PERM_ENUM)160); //A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) A2(8-11) A2(8-11) A3(8-11) A3(8-11) - const __m512i lhs_mat_01_2_sp1 = _mm512_shuffle_epi32(lhs_mat_01_2, 160); //A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) - const __m512i lhs_mat_23_2_sp1 = _mm512_shuffle_epi32(lhs_mat_23_2, 160); //A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) + const __m512i lhs_mat_01_2_sp1 = _mm512_shuffle_epi32(lhs_mat_01_2, (_MM_PERM_ENUM)160); //A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) A0(16-19) A0(16-19) A1(16-19) A1(16-19) + const __m512i lhs_mat_23_2_sp1 = _mm512_shuffle_epi32(lhs_mat_23_2, (_MM_PERM_ENUM)160); //A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) A2(16-19) A2(16-19) A3(16-19) A3(16-19) - const __m512i lhs_mat_01_3_sp1 = _mm512_shuffle_epi32(lhs_mat_01_3, 160); //A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) - const __m512i lhs_mat_23_3_sp1 = _mm512_shuffle_epi32(lhs_mat_23_3, 160); //A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) + const __m512i lhs_mat_01_3_sp1 = _mm512_shuffle_epi32(lhs_mat_01_3, (_MM_PERM_ENUM)160); //A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) A0(24-27) A0(24-27) A1(24-27) A1(24-27) + const __m512i lhs_mat_23_3_sp1 = _mm512_shuffle_epi32(lhs_mat_23_3, (_MM_PERM_ENUM)160); //A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) A2(24-27) A2(24-27) A3(24-27) A3(24-27) // Shuffle pattern two - left side input - const __m512i lhs_mat_01_0_sp2 = _mm512_shuffle_epi32(lhs_mat_01_0, 245); //A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) - const __m512i lhs_mat_23_0_sp2 = _mm512_shuffle_epi32(lhs_mat_23_0, 245); //A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) + const __m512i lhs_mat_01_0_sp2 = _mm512_shuffle_epi32(lhs_mat_01_0, (_MM_PERM_ENUM)245); //A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) A0(4-7) A0(4-7) A1(4-7) A1(4-7) + const __m512i lhs_mat_23_0_sp2 = _mm512_shuffle_epi32(lhs_mat_23_0, (_MM_PERM_ENUM)245); //A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) A2(4-7) A2(4-7) A3(4-7) A3(4-7) - const __m512i lhs_mat_01_1_sp2 = _mm512_shuffle_epi32(lhs_mat_01_1, 245); //A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) - const __m512i lhs_mat_23_1_sp2 = _mm512_shuffle_epi32(lhs_mat_23_1, 245); //A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) + const __m512i lhs_mat_01_1_sp2 = _mm512_shuffle_epi32(lhs_mat_01_1, (_MM_PERM_ENUM)245); //A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) A0(12-15) A0(12-15) A1(12-15) A1(12-15) + const __m512i lhs_mat_23_1_sp2 = _mm512_shuffle_epi32(lhs_mat_23_1, (_MM_PERM_ENUM)245); //A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) A2(12-15) A2(12-15) A3(12-15) A3(12-15) - const __m512i lhs_mat_01_2_sp2 = _mm512_shuffle_epi32(lhs_mat_01_2, 245); //A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) - const __m512i lhs_mat_23_2_sp2 = _mm512_shuffle_epi32(lhs_mat_23_2, 245); //A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) + const __m512i lhs_mat_01_2_sp2 = _mm512_shuffle_epi32(lhs_mat_01_2, (_MM_PERM_ENUM)245); //A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) A0(20-23) A0(20-23) A1(20-23) A1(20-23) + const __m512i lhs_mat_23_2_sp2 = _mm512_shuffle_epi32(lhs_mat_23_2, (_MM_PERM_ENUM)245); //A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) A2(20-23) A2(20-23) A3(20-23) A3(20-23) - const __m512i lhs_mat_01_3_sp2 = _mm512_shuffle_epi32(lhs_mat_01_3, 245); //A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) - const __m512i lhs_mat_23_3_sp2 = _mm512_shuffle_epi32(lhs_mat_23_3, 245); //A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) + const __m512i lhs_mat_01_3_sp2 = _mm512_shuffle_epi32(lhs_mat_01_3, (_MM_PERM_ENUM)245); //A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) A0(28-31) A0(28-31) A1(28-31) A1(28-31) + const __m512i lhs_mat_23_3_sp2 = _mm512_shuffle_epi32(lhs_mat_23_3, (_MM_PERM_ENUM)245); //A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) A2(28-31) A2(28-31) A3(28-31) A3(28-31) // The values arranged in shuffle patterns are operated with dot product operation within 32 bit lane i.e corresponding bytes and multiplied and added into 32 bit integers within 32 bit lane // Resembles MMLAs into 2x2 matrices in ARM Version @@ -2879,10 +2882,10 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * // Straighten out to make 4 row vectors - __m512i iacc_row_0 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_00, _mm512_shuffle_epi32(iacc_mat_01, 78)); - __m512i iacc_row_1 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_00, 78), iacc_mat_01); - __m512i iacc_row_2 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_10, _mm512_shuffle_epi32(iacc_mat_11, 78)); - __m512i iacc_row_3 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_10, 78), iacc_mat_11); + __m512i iacc_row_0 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_00, _mm512_shuffle_epi32(iacc_mat_01, (_MM_PERM_ENUM)78)); + __m512i iacc_row_1 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_00, (_MM_PERM_ENUM)78), iacc_mat_01); + __m512i iacc_row_2 = _mm512_mask_blend_epi32(0xCCCC, iacc_mat_10, _mm512_shuffle_epi32(iacc_mat_11, (_MM_PERM_ENUM)78)); + __m512i iacc_row_3 = _mm512_mask_blend_epi32(0xCCCC, _mm512_shuffle_epi32(iacc_mat_10, (_MM_PERM_ENUM)78), iacc_mat_11); // Load the scale(d) values for all the 4 Q8_0 blocks and repeat it across lanes const __m128i row_scale_f16 = _mm_shuffle_epi32(_mm_maskload_epi32((int const*)(a_ptr[b].d), loadMask), 68); @@ -3476,3 +3479,769 @@ void ggml_gemm_q4_0_8x8_q8_0(int n, float * restrict s, size_t bs, const void * } } } + +static void ggml_gemm_iq4_nl_4x4_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy, int nr, int nc) { + const int qk = QK8_0; + const int nb = n / qk; + const int ncols_interleaved = 4; + const int blocklen = 4; + + assert (n % qk == 0); + assert (nr % 4 == 0); + assert (nc % ncols_interleaved == 0); + + UNUSED(s); + UNUSED(bs); + UNUSED(vx); + UNUSED(vy); + UNUSED(nr); + UNUSED(nc); + UNUSED(nb); + UNUSED(ncols_interleaved); + UNUSED(blocklen); + +#if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) && defined(__ARM_FEATURE_DOTPROD) + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + const int8x16_t kvalues = vld1q_s8(kvalues_iq4nl); + + for (int y = 0; y < nr / 4; y++) { + const block_q8_0x4 * a_ptr = (const block_q8_0x4 *) vy + (y * nb); + for (int x = 0; x < nc / ncols_interleaved; x++) { + const block_iq4_nlx4 * b_ptr = (const block_iq4_nlx4 *) vx + (x * nb); + + float32x4_t sumf[4]; + for (int m = 0; m < 4; m++) { + sumf[m] = vdupq_n_f32(0); + } + + for (int l = 0; l < nb; l++) { + float32x4_t a_d = vcvt_f32_f16(vld1_f16((const float16_t *)a_ptr[l].d)); + float32x4_t b_d = vcvt_f32_f16(vld1_f16((const float16_t *)b_ptr[l].d)); + + int32x4_t sumi_0 = vdupq_n_s32(0); + int32x4_t sumi_1 = vdupq_n_s32(0); + int32x4_t sumi_2 = vdupq_n_s32(0); + int32x4_t sumi_3 = vdupq_n_s32(0); + + for (int k = 0; k < 4; k++) { + int8x16_t a_0 = vld1q_s8(a_ptr[l].qs + 16 * k + 0); + int8x16_t a_1 = vld1q_s8(a_ptr[l].qs + 16 * k + 64); + + uint8x16_t b = vld1q_u8(b_ptr[l].qs + 16 * k); + int8x16_t b_hi = vqtbl1q_s8(kvalues, b >> 4); + int8x16_t b_lo = vqtbl1q_s8(kvalues, b & 0xF); + + sumi_0 = vdotq_laneq_s32(sumi_0, b_lo, a_0, 0); + sumi_1 = vdotq_laneq_s32(sumi_1, b_lo, a_0, 1); + sumi_2 = vdotq_laneq_s32(sumi_2, b_lo, a_0, 2); + sumi_3 = vdotq_laneq_s32(sumi_3, b_lo, a_0, 3); + sumi_0 = vdotq_laneq_s32(sumi_0, b_hi, a_1, 0); + sumi_1 = vdotq_laneq_s32(sumi_1, b_hi, a_1, 1); + sumi_2 = vdotq_laneq_s32(sumi_2, b_hi, a_1, 2); + sumi_3 = vdotq_laneq_s32(sumi_3, b_hi, a_1, 3); + } + + sumf[0] = vmlaq_f32(sumf[0], vmulq_laneq_f32(b_d, a_d, 0), vcvtq_f32_s32(sumi_0)); + sumf[1] = vmlaq_f32(sumf[1], vmulq_laneq_f32(b_d, a_d, 1), vcvtq_f32_s32(sumi_1)); + sumf[2] = vmlaq_f32(sumf[2], vmulq_laneq_f32(b_d, a_d, 2), vcvtq_f32_s32(sumi_2)); + sumf[3] = vmlaq_f32(sumf[3], vmulq_laneq_f32(b_d, a_d, 3), vcvtq_f32_s32(sumi_3)); + } + + for (int m = 0; m < 4; m++) { + vst1q_f32(s + (y * 4 + m) * bs + x * 4, sumf[m]); + } + } + } + return; + } +#endif // #if ! ((defined(_MSC_VER)) && ! defined(__clang__)) && defined(__aarch64__) && defined(__ARM_NEON) + { + float sumf[4][4]; + int sumi; + + for (int y = 0; y < nr / 4; y++) { + const block_q8_0x4 * a_ptr = (const block_q8_0x4 *) vy + (y * nb); + for (int x = 0; x < nc / ncols_interleaved; x++) { + const block_iq4_nlx4 * b_ptr = (const block_iq4_nlx4 *) vx + (x * nb); + for (int m = 0; m < 4; m++) { + for (int j = 0; j < ncols_interleaved; j++) sumf[m][j] = 0.0; + } + for (int l = 0; l < nb; l++) { + for (int k = 0; k < (qk / (2 * blocklen)); k++) { + for (int m = 0; m < 4; m++) { + for (int j = 0; j < ncols_interleaved; j++) { + sumi = 0; + for (int i = 0; i < blocklen; ++i) { + const int v0 = kvalues_iq4nl[b_ptr[l].qs[k * ncols_interleaved * blocklen + j * blocklen + i] & 0x0F]; + const int v1 = kvalues_iq4nl[b_ptr[l].qs[k * ncols_interleaved * blocklen + j * blocklen + i] >> 4]; + sumi += ((v0 * a_ptr[l].qs[k * 4 * blocklen + m * blocklen + i]) + + (v1 * a_ptr[l].qs[k * 4 * blocklen + m * blocklen + i + qk / 2 * 4])); + } + sumf[m][j] += sumi * GGML_FP16_TO_FP32(b_ptr[l].d[j]) * GGML_FP16_TO_FP32(a_ptr[l].d[m]); + } + } + } + } + for (int m = 0; m < 4; m++) { + for (int j = 0; j < ncols_interleaved; j++) + s[(y * 4 + m) * bs + x * ncols_interleaved + j] = sumf[m][j]; + } + } + } + } +} + +static block_q4_0x4 make_block_q4_0x4(block_q4_0 * in, unsigned int blck_size_interleave) { + block_q4_0x4 out; + + for (int i = 0; i < 4; i++) { + out.d[i] = in[i].d; + } + + const int end = QK4_0 * 2 / blck_size_interleave; + + if (blck_size_interleave == 8) { + const uint64_t xor_mask = 0x8888888888888888ULL; + for (int i = 0; i < end; ++i) { + int src_id = i % 4; + int src_offset = (i / 4) * blck_size_interleave; + int dst_offset = i * blck_size_interleave; + + uint64_t elems; + // Using memcpy to avoid unaligned memory accesses + memcpy(&elems, &in[src_id].qs[src_offset], sizeof(uint64_t)); + elems ^= xor_mask; + memcpy(&out.qs[dst_offset], &elems, sizeof(uint64_t)); + } + } else if (blck_size_interleave == 4) { + const uint32_t xor_mask = 0x88888888; + for (int i = 0; i < end; ++i) { + int src_id = i % 4; + int src_offset = (i / 4) * blck_size_interleave; + int dst_offset = i * blck_size_interleave; + + uint32_t elems; + memcpy(&elems, &in[src_id].qs[src_offset], sizeof(uint32_t)); + elems ^= xor_mask; + memcpy(&out.qs[dst_offset], &elems, sizeof(uint32_t)); + } + } else { + GGML_ASSERT(false); + } + + return out; +} + +// interleave 8 block_q4_0s in blocks of blck_size_interleave +// returns an interleaved block_q4_0x8 +// in the interleaved block_q4_0x8, place deltas for 8 block_q4_0 blocks +// first, then interleave quants from 8 block_q4_0s in blocks of blck_size_interleave +static block_q4_0x8 make_block_q4_0x8(block_q4_0 * in, unsigned int blck_size_interleave) { + block_q4_0x8 out; + + for (int i = 0; i < 8; i++) { + out.d[i] = in[i].d; + } + + const int end = QK4_0 * 4 / blck_size_interleave; + const uint64_t xor_mask = 0x8888888888888888ULL; + + for (int i = 0; i < end; ++i) { + int src_id = i % 8; + int src_offset = (i / 8) * blck_size_interleave; + int dst_offset = i * blck_size_interleave; + + uint64_t elems; + memcpy(&elems, &in[src_id].qs[src_offset], sizeof(uint64_t)); + elems ^= xor_mask; + memcpy(&out.qs[dst_offset], &elems, sizeof(uint64_t)); + } + + return out; +} + +static int repack_q4_0_to_q4_0_4_bl(struct ggml_tensor * t, int interleave_block, const void * GGML_RESTRICT data, size_t data_size) { + GGML_ASSERT(t->type == GGML_TYPE_Q4_0); + GGML_ASSERT(interleave_block == 4 || interleave_block == 8); + constexpr int nrows_interleaved = 4; + + block_q4_0x4 * dst = (block_q4_0x4 *)t->data; + const block_q4_0 * src = (const block_q4_0 *)data; + block_q4_0 dst_tmp[4]; + int nrow = ggml_nrows(t); + int nblocks = t->ne[0] / QK4_0; + + GGML_ASSERT(data_size == nrow * nblocks * sizeof(block_q4_0)); + + if (t->ne[1] % nrows_interleaved != 0 || t->ne[0] % 8 != 0) { + return -1; + } + + for (int b = 0; b < nrow; b += nrows_interleaved) { + for (int64_t x = 0; x < nblocks; x++) { + for (int i = 0; i < nrows_interleaved; i++) { + dst_tmp[i] = src[x + i * nblocks]; + } + *dst++ = make_block_q4_0x4(dst_tmp, interleave_block); + } + src += nrows_interleaved * nblocks; + } + return 0; + + GGML_UNUSED(data_size); +} + +static int repack_q4_0_to_q4_0_8_bl(struct ggml_tensor * t, int interleave_block, const void * GGML_RESTRICT data, size_t data_size) { + GGML_ASSERT(t->type == GGML_TYPE_Q4_0); + GGML_ASSERT(interleave_block == 8); + constexpr int nrows_interleaved = 8; + + block_q4_0x8 * dst = (block_q4_0x8*)t->data; + const block_q4_0 * src = (const block_q4_0*) data; + block_q4_0 dst_tmp[8]; + int nrow = ggml_nrows(t); + int nblocks = t->ne[0] / QK4_0; + + GGML_ASSERT(data_size == nrow * nblocks * sizeof(block_q4_0)); + + if (t->ne[1] % nrows_interleaved != 0 || t->ne[0] % 8 != 0) { + return -1; + } + + for (int b = 0; b < nrow; b += nrows_interleaved) { + for (int64_t x = 0; x < nblocks; x++) { + for (int i = 0; i < nrows_interleaved; i++ ) { + dst_tmp[i] = src[x + i * nblocks]; + } + *dst++ = make_block_q4_0x8(dst_tmp, interleave_block); + } + src += nrows_interleaved * nblocks; + } + return 0; + + GGML_UNUSED(data_size); +} + +static block_iq4_nlx4 make_block_iq4_nlx4(block_iq4_nl * in, unsigned int blck_size_interleave) { + block_iq4_nlx4 out; + + for (int i = 0; i < 4; i++) { + out.d[i] = in[i].d; + } + + const int end = QK4_NL * 2 / blck_size_interleave; + + // TODO: this branch seems wrong + //if (blck_size_interleave == 8) { + // for (int i = 0; i < end; ++i) { + // int src_id = i % 4; + // int src_offset = (i / 4) * blck_size_interleave; + // int dst_offset = i * blck_size_interleave; + + // // Using memcpy to avoid unaligned memory accesses + // memcpy(&out.qs[dst_offset], &in[src_id].qs[src_offset], sizeof(uint64_t)); + // } + //} else + if (blck_size_interleave == 4) { + for (int i = 0; i < end; ++i) { + int src_id = i % 4; + int src_offset = (i / 4) * blck_size_interleave; + int dst_offset = i * blck_size_interleave; + + memcpy(&out.qs[dst_offset], &in[src_id].qs[src_offset], sizeof(uint32_t)); + } + } else { + GGML_ASSERT(false); + } + + return out; +} + +static int repack_iq4_nl_to_iq4_nl_4_bl(struct ggml_tensor * t, int interleave_block, const void * GGML_RESTRICT data, size_t data_size) { + GGML_ASSERT(t->type == GGML_TYPE_IQ4_NL); + //GGML_ASSERT(interleave_block == 4 || interleave_block == 8); + GGML_ASSERT(interleave_block == 4); + + block_iq4_nlx4 * dst = (block_iq4_nlx4 *)t->data; + const block_iq4_nl * src = (const block_iq4_nl *)data; + block_iq4_nl dst_tmp[4]; + int nrow = ggml_nrows(t); + int nrows_interleaved = 4; + int nblocks = t->ne[0] / QK4_0; + + GGML_ASSERT(data_size == nrow * nblocks * sizeof(block_iq4_nl)); + + if (t->ne[1] % nrows_interleaved != 0 || t->ne[0] % 8 != 0) { + return -1; + } + + for (int b = 0; b < nrow; b += nrows_interleaved) { + for (int64_t x = 0; x < nblocks; x++) { + for (int i = 0; i < nrows_interleaved; i++) { + dst_tmp[i] = src[x + i * nblocks]; + } + *dst++ = make_block_iq4_nlx4(dst_tmp, interleave_block); + } + src += nrows_interleaved * nblocks; + } + return 0; + + GGML_UNUSED(data_size); +} + +namespace ggml::cpu::aarch64 { +// repack +template +int repack(struct ggml_tensor *, const void *, size_t); + +// TODO: generalise. +template <> int repack(struct ggml_tensor * t, const void * data, size_t data_size) { + return repack_q4_0_to_q4_0_4_bl(t, 4, data, data_size); +} + +template <> int repack(struct ggml_tensor * t, const void * data, size_t data_size) { + return repack_q4_0_to_q4_0_4_bl(t, 8, data, data_size); +} + +template <> int repack(struct ggml_tensor * t, const void * data, size_t data_size) { + return repack_q4_0_to_q4_0_8_bl(t, 8, data, data_size); +} + +template <> int repack(struct ggml_tensor * t, const void * data, size_t data_size) { + return repack_iq4_nl_to_iq4_nl_4_bl(t, 4, data, data_size); +} + +// TODO: needs to be revisited +//template <> int repack(struct ggml_tensor * t, const void * data, size_t data_size) { +// return repack_iq4_nl_to_iq4_nl_4_bl(t, 8, data, data_size); +//} + +// gemv +template +void gemv(int, float *, size_t, const void *, const void *, int, int); + +template <> void gemv(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemv_q4_0_4x4_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> void gemv(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemv_q4_0_4x8_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> void gemv(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemv_q4_0_8x8_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> +void gemv(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemv_iq4_nl_4x4_q8_0(n, s, bs, vx, vy, nr, nc); +} + +// gemm +template +void gemm(int, float *, size_t, const void *, const void *, int, int); + +template <> void gemm(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemm_q4_0_4x4_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> void gemm(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemm_q4_0_4x8_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> void gemm(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemm_q4_0_8x8_q8_0(n, s, bs, vx, vy, nr, nc); +} + +template <> +void gemm(int n, float * s, size_t bs, const void * vx, const void * vy, int nr, int nc) { + ggml_gemm_iq4_nl_4x4_q8_0(n, s, bs, vx, vy, nr, nc); +} + +class tensor_traits_base : public ggml::cpu::tensor_traits { + public: + virtual int repack(struct ggml_tensor * t, const void * data, size_t data_size) = 0; +}; + +template class tensor_traits : public tensor_traits_base { + + bool work_size(int /* n_threads */, const struct ggml_tensor * op, size_t & size) override { + // not realy a GGML_TYPE_Q8_0 but same size. + switch (op->op) { + case GGML_OP_MUL_MAT: + size = ggml_row_size(GGML_TYPE_Q8_0, ggml_nelements(op->src[1])); + return true; + case GGML_OP_MUL_MAT_ID: + size = ggml_row_size(GGML_TYPE_Q8_0, ggml_nelements(op->src[1])); + size = GGML_PAD(size, sizeof(int64_t)); // + padding for next bloc. + size += sizeof(int64_t) * (1+op->src[0]->ne[2]) * op->src[1]->ne[2]; + return true; + default: + // GGML_ABORT("fatal error"); + break; + } + return false; + } + + bool compute_forward(struct ggml_compute_params * params, struct ggml_tensor * op) override { + switch (op->op) { + case GGML_OP_MUL_MAT: + forward_mul_mat(params, op); + return true; + case GGML_OP_MUL_MAT_ID: + forward_mul_mat_id(params, op); + return true; + default: + // GGML_ABORT("fatal error"); + break; + } + return false; + } + + void forward_mul_mat(ggml_compute_params * params, ggml_tensor * op) { + const ggml_tensor * src0 = op->src[0]; + const ggml_tensor * src1 = op->src[1]; + ggml_tensor * dst = op; + + GGML_TENSOR_BINARY_OP_LOCALS + + const int ith = params->ith; + const int nth = params->nth; + + GGML_ASSERT(ne0 == ne01); + GGML_ASSERT(ne1 == ne11); + GGML_ASSERT(ne2 == ne12); + GGML_ASSERT(ne3 == ne13); + + // dst cannot be transposed or permuted + GGML_ASSERT(nb0 == sizeof(float)); + GGML_ASSERT(nb0 <= nb1); + GGML_ASSERT(nb1 <= nb2); + GGML_ASSERT(nb2 <= nb3); + + GGML_ASSERT(src1->type == GGML_TYPE_F32); + + GGML_ASSERT(ggml_n_dims(op->src[0]) == 2); + // GGML_ASSERT(ggml_n_dims(op->src[1]) == 2); + + char * wdata = static_cast(params->wdata); + const size_t nbw1 = ggml_row_size(GGML_TYPE_Q8_0, ne10); + + assert(params->wsize >= nbw1 * ne11); + + const ggml_from_float_t from_float = ggml_get_type_traits_cpu(GGML_TYPE_Q8_0)->from_float; + + int64_t i11_processed = 0; + for (int64_t i11 = ith * 4; i11 < ne11 - ne11 % 4; i11 += nth * 4) { + quantize_mat_q8_0((float *) ((char *) src1->data + i11 * nb11), (void *) (wdata + i11 * nbw1), 4, ne10, + INTER_SIZE); + } + i11_processed = ne11 - ne11 % 4; + for (int64_t i11 = i11_processed + ith; i11 < ne11; i11 += nth) { + from_float((float *) ((char *) src1->data + i11 * nb11), (void *) (wdata + i11 * nbw1), ne10); + } + + ggml_barrier(params->threadpool); + + const void * src1_wdata = params->wdata; + const size_t src1_col_stride = ggml_row_size(GGML_TYPE_Q8_0, ne10); + int64_t src0_start = (ith * ne01) / nth; + int64_t src0_end = ((ith + 1) * ne01) / nth; + src0_start = (src0_start % NB_COLS) ? src0_start + NB_COLS - (src0_start % NB_COLS) : src0_start; + src0_end = (src0_end % NB_COLS) ? src0_end + NB_COLS - (src0_end % NB_COLS) : src0_end; + if (src0_start >= src0_end) { + return; + } + + // If there are more than three rows in src1, use gemm; otherwise, use gemv. + if (ne11 > 3) { + gemm(ne00, (float *) ((char *) dst->data) + src0_start, ne01, + (const char *) src0->data + src0_start * nb01, + (const char *) src1_wdata, ne11 - ne11 % 4, src0_end - src0_start); + } + for (int iter = ne11 - ne11 % 4; iter < ne11; iter++) { + gemv(ne00, (float *) ((char *) dst->data + (iter * nb1)) + src0_start, ne01, + (const char *) src0->data + src0_start * nb01, + (const char *) src1_wdata + (src1_col_stride * iter), 1, + src0_end - src0_start); + } + } + + void forward_mul_mat_id(ggml_compute_params * params, ggml_tensor * op) { + const ggml_tensor * src0 = op->src[0]; + const ggml_tensor * src1 = op->src[1]; + const ggml_tensor * ids = op->src[2]; + ggml_tensor * dst = op; + + GGML_TENSOR_BINARY_OP_LOCALS + + const int ith = params->ith; + const int nth = params->nth; + + const ggml_from_float_t from_float = ggml_get_type_traits_cpu(GGML_TYPE_Q8_0)->from_float; + + // we don't support permuted src0 or src1 + GGML_ASSERT(nb00 == ggml_type_size(src0->type)); + GGML_ASSERT(nb10 == ggml_type_size(src1->type)); + + // dst cannot be transposed or permuted + GGML_ASSERT(nb0 == sizeof(float)); + GGML_ASSERT(nb0 <= nb1); + GGML_ASSERT(nb1 <= nb2); + GGML_ASSERT(nb2 <= nb3); + + GGML_ASSERT(ne03 == 1); + GGML_ASSERT(ne13 == 1); + GGML_ASSERT(ne3 == 1); + + GGML_ASSERT(src1->type == GGML_TYPE_F32); + + // row groups + const int n_ids = ids->ne[0]; // n_expert_used + const int n_as = ne02; // n_expert + + const size_t nbw1 = ggml_row_size(GGML_TYPE_Q8_0, ne10); + const size_t nbw2 = nbw1*ne11; + const size_t nbw3 = nbw2*ne12; + + struct mmid_row_mapping { + int32_t i1; + int32_t i2; + }; + + GGML_ASSERT(params->wsize >= (GGML_PAD(nbw3, sizeof(int64_t)) + n_as * sizeof(int64_t) + + n_as * ne12 * sizeof(mmid_row_mapping))); + + auto wdata = (char *) params->wdata; + auto wdata_src1_end = (char *) wdata + GGML_PAD(nbw3, sizeof(int64_t)); + int64_t * matrix_row_counts = (int64_t *) (wdata_src1_end); // [n_as] + struct mmid_row_mapping * matrix_rows = (struct mmid_row_mapping *) (matrix_row_counts + n_as); // [n_as][ne12] + + // src1: float32 => block_q8_0 + for (int64_t i12 = 0; i12 < ne12; ++i12) { + for (int64_t i11 = ith; i11 < ne11; i11 += nth) { + from_float((float *)((char *) src1->data + i12 * nb12 + i11 * nb11), + (void *) (wdata + i12 * nbw2 + i11 * nbw1), + ne10); + } + } + +#define MMID_MATRIX_ROW(row_id, i1) matrix_rows[(row_id) * ne12 + (i1)] + + if (ith == 0) { + // initialize matrix_row_counts + memset(matrix_row_counts, 0, n_as * sizeof(int64_t)); + + // group rows by src0 matrix + for (int32_t iid1 = 0; iid1 < ids->ne[1]; ++iid1) { + for (int32_t id = 0; id < n_ids; ++id) { + const int32_t i02 = + *(const int32_t *) ((const char *) ids->data + iid1 * ids->nb[1] + id * ids->nb[0]); + + GGML_ASSERT(i02 >= 0 && i02 < n_as); + + MMID_MATRIX_ROW(i02, matrix_row_counts[i02]) = { id, iid1 }; + matrix_row_counts[i02] += 1; + } + } + } + + ggml_barrier(params->threadpool); + + // compute each matrix multiplication in sequence + for (int cur_a = 0; cur_a < n_as; ++cur_a) { + const int64_t cne1 = matrix_row_counts[cur_a]; + + if (cne1 == 0) { + continue; + } + + auto src0_cur = (const char *) src0->data + cur_a*nb02; + + //const int64_t nr0 = ne01; // src0 rows + const int64_t nr1 = cne1; // src1 rows + + int64_t src0_cur_start = (ith * ne01) / nth; + int64_t src0_cur_end = ((ith + 1) * ne01) / nth; + src0_cur_start = + (src0_cur_start % NB_COLS) ? src0_cur_start + NB_COLS - (src0_cur_start % NB_COLS) : src0_cur_start; + src0_cur_end = (src0_cur_end % NB_COLS) ? src0_cur_end + NB_COLS - (src0_cur_end % NB_COLS) : src0_cur_end; + + if (src0_cur_start >= src0_cur_end) return; + + for (int ir1 = 0; ir1 < nr1; ir1++) { + struct mmid_row_mapping row_mapping = MMID_MATRIX_ROW(cur_a, ir1); + const int id = row_mapping.i1; // selected expert index + + const int64_t i11 = id % ne11; + const int64_t i12 = row_mapping.i2; // row index in src1 + + const int64_t i1 = id; // selected expert index + const int64_t i2 = i12; // row + + auto src1_col = (const char *) wdata + (i11 * nbw1 + i12 * nbw2); + + gemv( + ne00, (float *)((char *) dst->data + (i1 * nb1 + i2 * nb2)) + src0_cur_start, + ne01, src0_cur + src0_cur_start * nb01, + src1_col, 1, src0_cur_end - src0_cur_start); + } + } +#undef MMID_MATRIX_ROW + } + + int repack(struct ggml_tensor * t, const void * data, size_t data_size) override { + GGML_LOG_DEBUG("%s: repack tensor %s with %s_%dx%d\n", __func__, t->name, ggml_type_name(t->type), + (int) NB_COLS, (int) INTER_SIZE); + return ggml::cpu::aarch64::repack(t, data, data_size); + } +}; + +// instance for Q4 +static const tensor_traits q4_0_4x4_q8_0; +static const tensor_traits q4_0_4x8_q8_0; +static const tensor_traits q4_0_8x8_q8_0; + +// instance for IQ4 +static const tensor_traits iq4_nl_4x4_q8_0; + +} // namespace ggml::cpu::aarch64 + +static const ggml::cpu::tensor_traits * ggml_aarch64_get_optimal_repack_type(const struct ggml_tensor * cur) { + if (cur->type == GGML_TYPE_Q4_0) { + if (ggml_cpu_has_avx2() || (ggml_cpu_has_sve() && ggml_cpu_has_matmul_int8() && ggml_cpu_get_sve_cnt() == QK8_0)) { + if (cur->ne[1] % 8 == 0) { + return &ggml::cpu::aarch64::q4_0_8x8_q8_0; + } + } + if (ggml_cpu_has_neon() && ggml_cpu_has_matmul_int8()) { + if (cur->ne[1] % 4 == 0) { + return &ggml::cpu::aarch64::q4_0_4x8_q8_0; + } + } + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + if (cur->ne[1] % 4 == 0) { + return &ggml::cpu::aarch64::q4_0_4x4_q8_0; + } + } + } else if (cur->type == GGML_TYPE_IQ4_NL) { + if (ggml_cpu_has_neon() && ggml_cpu_has_dotprod()) { + if (cur->ne[1] % 4 == 0) { + return &ggml::cpu::aarch64::iq4_nl_4x4_q8_0; + } + } + } + + return nullptr; +} + +static void ggml_backend_cpu_aarch64_buffer_init_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { + tensor->extra = (void *) const_cast(ggml_aarch64_get_optimal_repack_type(tensor)); + + GGML_UNUSED(buffer); +} + +static void ggml_backend_cpu_aarch64_buffer_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, + const void * data, size_t offset, size_t size) { + GGML_ASSERT(offset == 0); + GGML_ASSERT(size == ggml_nbytes(tensor)); + + auto tensor_traits = (ggml::cpu::aarch64::tensor_traits_base *) tensor->extra; + auto OK = tensor_traits->repack(tensor, data, size); + + GGML_ASSERT(OK == 0); + GGML_UNUSED(buffer); +} + +static const char * ggml_backend_cpu_aarch64_buffer_type_get_name(ggml_backend_buffer_type_t buft) { + return "CPU_AARCH64"; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_t ggml_backend_cpu_aarch64_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + ggml_backend_buffer_t buffer = ggml_backend_buft_alloc_buffer(ggml_backend_cpu_buffer_type(), size); + + if (buffer == nullptr) { + return nullptr; + } + + buffer->buft = buft; + buffer->iface.init_tensor = ggml_backend_cpu_aarch64_buffer_init_tensor; + buffer->iface.set_tensor = ggml_backend_cpu_aarch64_buffer_set_tensor; + buffer->iface.get_tensor = nullptr; + buffer->iface.cpy_tensor = nullptr; + return buffer; +} + +static size_t ggml_backend_cpu_aarch64_buffer_type_get_alignment(ggml_backend_buffer_type_t buft) { + return TENSOR_ALIGNMENT; + + GGML_UNUSED(buft); +} + +namespace ggml::cpu::aarch64 { +class extra_buffer_type : ggml::cpu::extra_buffer_type { + bool supports_op(ggml_backend_dev_t, const struct ggml_tensor * op) override { + if ( op->op == GGML_OP_MUL_MAT && + op->src[0]->buffer && + (ggml_n_dims(op->src[0]) == 2) && + op->src[0]->buffer->buft == ggml_backend_cpu_aarch64_buffer_type() && + ggml_aarch64_get_optimal_repack_type(op->src[0]) + ) { + if (op->src[1]->buffer && !ggml_backend_buft_is_host(op->src[1]->buffer->buft)) { + return false; + } + if (op->src[1]->type == GGML_TYPE_F32) { + return true; + } + //if (op->src[1]->type == GGML_TYPE_Q8_0) { + // return true; + //} + // may be possible if Q8_0 packed... + } else if (op->op == GGML_OP_MUL_MAT_ID + && op->src[0]->buffer + && (ggml_n_dims(op->src[0]) == 3) + && op->src[0]->buffer->buft == ggml_backend_cpu_aarch64_buffer_type() + && ggml_aarch64_get_optimal_repack_type(op->src[0]) + ) { + if (op->src[1]->buffer && !ggml_backend_buft_is_host(op->src[1]->buffer->buft)) { + return false; + } + if (op->src[1]->type == GGML_TYPE_F32) { + return true; + } + //if (op->src[1]->type == GGML_TYPE_Q8_0) { + // return true; + //} + } + return false; + } + + ggml::cpu::tensor_traits * get_tensor_traits(const struct ggml_tensor * op) override { + if (op->op == GGML_OP_MUL_MAT || op->op == GGML_OP_MUL_MAT_ID) { + if (op->src[0]->buffer && op->src[0]->buffer->buft == ggml_backend_cpu_aarch64_buffer_type()) { + return (ggml::cpu::tensor_traits *) op->src[0]->extra; + } + } + return nullptr; + } +}; +} // namespace ggml::cpu::aarch64 + +ggml_backend_buffer_type_t ggml_backend_cpu_aarch64_buffer_type(void) { + static struct ggml_backend_buffer_type ggml_backend_cpu_buffer_type_aarch64 = { + /* .iface = */ { + /* .get_name = */ ggml_backend_cpu_aarch64_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_cpu_aarch64_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_cpu_aarch64_buffer_type_get_alignment, + /* .get_max_size = */ nullptr, // defaults to SIZE_MAX + /* .get_alloc_size = */ nullptr, // defaults to ggml_nbytes + /* .is_host = */ nullptr, + }, + /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), + /* .context = */ new ggml::cpu::aarch64::extra_buffer_type(), + }; + + return &ggml_backend_cpu_buffer_type_aarch64; +} diff --git a/ggml/src/ggml-cpu/ggml-cpu-aarch64.h b/ggml/src/ggml-cpu/ggml-cpu-aarch64.h new file mode 100644 index 000000000..6e84c826b --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-aarch64.h @@ -0,0 +1,8 @@ +#pragma once + +#include "ggml-cpu-traits.h" +#include "ggml.h" + +// GGML internal header + +ggml_backend_buffer_type_t ggml_backend_cpu_aarch64_buffer_type(void); diff --git a/ggml/src/ggml-cpu/ggml-cpu-hbm.cpp b/ggml/src/ggml-cpu/ggml-cpu-hbm.cpp new file mode 100644 index 000000000..fa8dea2af --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-hbm.cpp @@ -0,0 +1,55 @@ +#ifdef GGML_USE_CPU_HBM + +#include "ggml-backend.h" +#include "ggml-backend-impl.h" +#include "ggml-cpu.h" +#include "ggml-impl.h" + +#include "ggml-cpu-hbm.h" + +// buffer type HBM + +#include + +static const char * ggml_backend_cpu_hbm_buffer_type_get_name(ggml_backend_buffer_type_t buft) { + return "CPU_HBM"; + + GGML_UNUSED(buft); +} + +static void ggml_backend_cpu_hbm_buffer_free_buffer(ggml_backend_buffer_t buffer) { + hbw_free(buffer->context); +} + +static ggml_backend_buffer_t ggml_backend_cpu_hbm_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, + size_t size) { + void * ptr; + int result = hbw_posix_memalign(&ptr, ggml_backend_cpu_buffer_type_get_alignment(buft), size); + if (result != 0) { + GGML_LOG_ERROR("failed to allocate HBM buffer of size %zu\n", size); + return NULL; + } + + ggml_backend_buffer_t buffer = ggml_backend_cpu_buffer_from_ptr(ptr, size); + buffer->buft = buft; + buffer->iface.free_buffer = ggml_backend_cpu_hbm_buffer_free_buffer; + + return buffer; +} + +ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void) { + static struct ggml_backend_buffer_type ggml_backend_cpu_buffer_type_hbm = { + /* .iface = */ { + /* .get_name = */ ggml_backend_cpu_hbm_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_cpu_hbm_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_cpu_buffer_type_get_alignment, + /* .get_max_size = */ nullptr, // defaults to SIZE_MAX + /* .get_alloc_size = */ nullptr, // defaults to ggml_nbytes + /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, + }, + /* .context = */ nullptr, + }; + + return &ggml_backend_cpu_buffer_type_hbm; +} +#endif diff --git a/ggml/src/ggml-cpu/ggml-cpu-hbm.h b/ggml/src/ggml-cpu/ggml-cpu-hbm.h new file mode 100644 index 000000000..09a1f09d7 --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-hbm.h @@ -0,0 +1,8 @@ +#pragma once + +#include "ggml-backend.h" +#include "ggml.h" + +// GGML CPU internal header + +ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void); diff --git a/ggml/src/ggml-cpu-impl.h b/ggml/src/ggml-cpu/ggml-cpu-impl.h similarity index 53% rename from ggml/src/ggml-cpu-impl.h rename to ggml/src/ggml-cpu/ggml-cpu-impl.h index 5b45155b0..d71076ad1 100644 --- a/ggml/src/ggml-cpu-impl.h +++ b/ggml/src/ggml-cpu/ggml-cpu-impl.h @@ -15,6 +15,18 @@ extern "C" { #endif +struct ggml_compute_params { + // ith = thread index, nth = number of threads + int ith, nth; + + // work buffer for all threads + size_t wsize; + void * wdata; + + struct ggml_threadpool * threadpool; +}; + + #if defined(_MSC_VER) #define m512bh(p) p @@ -27,80 +39,6 @@ extern "C" { #endif -/** - * Converts brain16 to float32. - * - * The bfloat16 floating point format has the following structure: - * - * ┌sign - * │ - * │ ┌exponent - * │ │ - * │ │ ┌mantissa - * │ │ │ - * │┌──┴───┐┌─┴───┐ - * 0b0000000000000000 brain16 - * - * Since bf16 has the same number of exponent bits as a 32bit float, - * encoding and decoding numbers becomes relatively straightforward. - * - * ┌sign - * │ - * │ ┌exponent - * │ │ - * │ │ ┌mantissa - * │ │ │ - * │┌──┴───┐┌─┴───────────────────┐ - * 0b00000000000000000000000000000000 IEEE binary32 - * - * For comparison, the standard fp16 format has fewer exponent bits. - * - * ┌sign - * │ - * │ ┌exponent - * │ │ - * │ │ ┌mantissa - * │ │ │ - * │┌─┴─┐┌─┴──────┐ - * 0b0000000000000000 IEEE binary16 - * - * @see IEEE 754-2008 - */ -static inline float ggml_compute_bf16_to_fp32(ggml_bf16_t h) { - union { - float f; - uint32_t i; - } u; - u.i = (uint32_t)h.bits << 16; - return u.f; -} - -/** - * Converts float32 to brain16. - * - * This is binary identical with Google Brain float conversion. - * Floats shall round to nearest even, and NANs shall be quiet. - * Subnormals aren't flushed to zero, except perhaps when used. - * This code should vectorize nicely if using modern compilers. - */ -static inline ggml_bf16_t ggml_compute_fp32_to_bf16(float s) { - ggml_bf16_t h; - union { - float f; - uint32_t i; - } u; - u.f = s; - if ((u.i & 0x7fffffff) > 0x7f800000) { /* nan */ - h.bits = (u.i >> 16) | 64; /* force to quiet */ - return h; - } - h.bits = (u.i + (0x7fff + ((u.i >> 16) & 1))) >> 16; - return h; -} - -#define GGML_FP32_TO_BF16(x) ggml_compute_fp32_to_bf16(x) -#define GGML_BF16_TO_FP32(x) ggml_compute_bf16_to_fp32(x) - // __FMA__ and __F16C__ are not defined in MSVC, however they are implied with AVX2/AVX512 #if defined(_MSC_VER) && (defined(__AVX2__) || defined(__AVX512F__)) #ifndef __FMA__ @@ -388,28 +326,6 @@ inline static int32x4_t ggml_vdotq_s32(int32x4_t acc, int8x16_t a, int8x16_t b) #endif // defined(__ARM_NEON) -#if defined(__ARM_NEON) && !defined(_MSC_VER) - -#define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) -#define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) - -#define GGML_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) - -static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { - ggml_fp16_internal_t tmp; - memcpy(&tmp, &h, sizeof(ggml_fp16_t)); - return (float)tmp; -} - -static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { - ggml_fp16_t res; - ggml_fp16_internal_t tmp = f; - memcpy(&res, &tmp, sizeof(ggml_fp16_t)); - return res; -} - -#else - #ifdef __wasm_simd128__ #include #else @@ -462,152 +378,8 @@ static __m256 __lasx_xvreplfr2vr_s(float val) { } #endif -#ifdef __F16C__ - -#ifdef _MSC_VER -#define GGML_COMPUTE_FP16_TO_FP32(x) _mm_cvtss_f32(_mm_cvtph_ps(_mm_cvtsi32_si128(x))) -#define GGML_COMPUTE_FP32_TO_FP16(x) _mm_extract_epi16(_mm_cvtps_ph(_mm_set_ss(x), 0), 0) -#else -#define GGML_COMPUTE_FP16_TO_FP32(x) _cvtsh_ss(x) -#define GGML_COMPUTE_FP32_TO_FP16(x) _cvtss_sh(x, 0) -#endif - -#elif defined(__POWER9_VECTOR__) - -#define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) -#define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) -/* the inline asm below is about 12% faster than the lookup method */ -#define GGML_FP16_TO_FP32(x) GGML_COMPUTE_FP16_TO_FP32(x) -#define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) - -static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { - register float f; - register double d; - __asm__( - "mtfprd %0,%2\n" - "xscvhpdp %0,%0\n" - "frsp %1,%0\n" : - /* temp */ "=d"(d), - /* out */ "=f"(f): - /* in */ "r"(h)); - return f; -} - -static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { - register double d; - register ggml_fp16_t r; - __asm__( /* xscvdphp can work on double or single precision */ - "xscvdphp %0,%2\n" - "mffprd %1,%0\n" : - /* temp */ "=d"(d), - /* out */ "=r"(r): - /* in */ "f"(f)); - return r; -} - -#else - -// FP16 <-> FP32 -// ref: https://github.com/Maratyszcza/FP16 - -static inline float fp32_from_bits(uint32_t w) { - union { - uint32_t as_bits; - float as_value; - } fp32; - fp32.as_bits = w; - return fp32.as_value; -} - -static inline uint32_t fp32_to_bits(float f) { - union { - float as_value; - uint32_t as_bits; - } fp32; - fp32.as_value = f; - return fp32.as_bits; -} - -static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { - const uint32_t w = (uint32_t) h << 16; - const uint32_t sign = w & UINT32_C(0x80000000); - const uint32_t two_w = w + w; - - const uint32_t exp_offset = UINT32_C(0xE0) << 23; -#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) || defined(__GNUC__) && !defined(__STRICT_ANSI__) - const float exp_scale = 0x1.0p-112f; -#else - const float exp_scale = fp32_from_bits(UINT32_C(0x7800000)); -#endif - const float normalized_value = fp32_from_bits((two_w >> 4) + exp_offset) * exp_scale; - - const uint32_t magic_mask = UINT32_C(126) << 23; - const float magic_bias = 0.5f; - const float denormalized_value = fp32_from_bits((two_w >> 17) | magic_mask) - magic_bias; - - const uint32_t denormalized_cutoff = UINT32_C(1) << 27; - const uint32_t result = sign | - (two_w < denormalized_cutoff ? fp32_to_bits(denormalized_value) : fp32_to_bits(normalized_value)); - return fp32_from_bits(result); -} - -static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { -#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) || defined(__GNUC__) && !defined(__STRICT_ANSI__) - const float scale_to_inf = 0x1.0p+112f; - const float scale_to_zero = 0x1.0p-110f; -#else - const float scale_to_inf = fp32_from_bits(UINT32_C(0x77800000)); - const float scale_to_zero = fp32_from_bits(UINT32_C(0x08800000)); -#endif - float base = (fabsf(f) * scale_to_inf) * scale_to_zero; - - const uint32_t w = fp32_to_bits(f); - const uint32_t shl1_w = w + w; - const uint32_t sign = w & UINT32_C(0x80000000); - uint32_t bias = shl1_w & UINT32_C(0xFF000000); - if (bias < UINT32_C(0x71000000)) { - bias = UINT32_C(0x71000000); - } - - base = fp32_from_bits((bias >> 1) + UINT32_C(0x07800000)) + base; - const uint32_t bits = fp32_to_bits(base); - const uint32_t exp_bits = (bits >> 13) & UINT32_C(0x00007C00); - const uint32_t mantissa_bits = bits & UINT32_C(0x00000FFF); - const uint32_t nonsign = exp_bits + mantissa_bits; - return (sign >> 16) | (shl1_w > UINT32_C(0xFF000000) ? UINT16_C(0x7E00) : nonsign); -} - -#define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) -#define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) - -#endif // __F16C__ - -#endif // defined(__ARM_NEON) && (!defined(__MSC_VER) - -#ifdef __ARM_FEATURE_SVE -#include -#endif // __ARM_FEATURE_SVE - -// precomputed f32 table for f16 (256 KB) -// defined in ggml.c, initialized in ggml_init() -extern float ggml_table_f32_f16[1 << 16]; - -// On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32, -// so we define GGML_FP16_TO_FP32 and GGML_FP32_TO_FP16 elsewhere for NEON. -// This is also true for POWER9. -#if !defined(GGML_FP16_TO_FP32) -inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) { - uint16_t s; - memcpy(&s, &f, sizeof(uint16_t)); - return ggml_table_f32_f16[s]; -} - -#define GGML_FP16_TO_FP32(x) ggml_lookup_fp16_to_fp32(x) -#endif - -#if !defined(GGML_FP32_TO_FP16) -#define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) -#endif +// TODO: move to ggml-threading +void ggml_barrier(struct ggml_threadpool * tp); #ifdef __cplusplus } diff --git a/ggml/src/ggml-cpu/ggml-cpu-quants.c b/ggml/src/ggml-cpu/ggml-cpu-quants.c new file mode 100644 index 000000000..8e1472266 --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-quants.c @@ -0,0 +1,10839 @@ +#define GGML_COMMON_IMPL_C +#include "ggml-common.h" + +#include "ggml-quants.h" +#include "ggml-cpu-quants.h" +#include "ggml-impl.h" +#include "ggml-cpu-impl.h" +#include "ggml-cpu.h" + +#include +#include +#include +#include +#include // for qsort +#include // for GGML_ASSERT + +#define GROUP_MAX_EPS 1e-15f +#define GROUP_MAX_EPS_IQ3_XXS 1e-8f +#define GROUP_MAX_EPS_IQ2_S 1e-8f +#define GROUP_MAX_EPS_IQ1_M 1e-7f +#define GROUP_MAX_EPS_IQ1_S 1e-12f + +#if defined(_MSC_VER) +// disable "possible loss of data" to avoid warnings for hundreds of casts +// we should just be careful :) +#pragma warning(disable: 4244 4267) +#endif + +#define UNUSED GGML_UNUSED + +// some compilers don't provide _mm256_set_m128i, e.g. gcc 7 +#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1) + +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) +// multiply int8_t, add results pairwise twice +static inline __m128i mul_sum_i8_pairs(const __m128i x, const __m128i y) { + // Get absolute values of x vectors + const __m128i ax = _mm_sign_epi8(x, x); + // Sign the values of the y vectors + const __m128i sy = _mm_sign_epi8(y, x); + // Perform multiplication and create 16-bit values + const __m128i dot = _mm_maddubs_epi16(ax, sy); + const __m128i ones = _mm_set1_epi16(1); + return _mm_madd_epi16(ones, dot); +} + +#if __AVX__ || __AVX2__ || __AVX512F__ +// horizontally add 8 floats +static inline float hsum_float_8(const __m256 x) { + __m128 res = _mm256_extractf128_ps(x, 1); + res = _mm_add_ps(res, _mm256_castps256_ps128(x)); + res = _mm_add_ps(res, _mm_movehl_ps(res, res)); + res = _mm_add_ss(res, _mm_movehdup_ps(res)); + return _mm_cvtss_f32(res); +} + +// horizontally add 8 int32_t +static inline int hsum_i32_8(const __m256i a) { + const __m128i sum128 = _mm_add_epi32(_mm256_castsi256_si128(a), _mm256_extractf128_si256(a, 1)); + const __m128i hi64 = _mm_unpackhi_epi64(sum128, sum128); + const __m128i sum64 = _mm_add_epi32(hi64, sum128); + const __m128i hi32 = _mm_shuffle_epi32(sum64, _MM_SHUFFLE(2, 3, 0, 1)); + return _mm_cvtsi128_si32(_mm_add_epi32(sum64, hi32)); +} + +// horizontally add 4 int32_t +static inline int hsum_i32_4(const __m128i a) { + const __m128i hi64 = _mm_unpackhi_epi64(a, a); + const __m128i sum64 = _mm_add_epi32(hi64, a); + const __m128i hi32 = _mm_shuffle_epi32(sum64, _MM_SHUFFLE(2, 3, 0, 1)); + return _mm_cvtsi128_si32(_mm_add_epi32(sum64, hi32)); +} + +#if defined(__AVX2__) || defined(__AVX512F__) +// spread 32 bits to 32 bytes { 0x00, 0xFF } +static inline __m256i bytes_from_bits_32(const uint8_t * x) { + uint32_t x32; + memcpy(&x32, x, sizeof(uint32_t)); + const __m256i shuf_mask = _mm256_set_epi64x( + 0x0303030303030303, 0x0202020202020202, + 0x0101010101010101, 0x0000000000000000); + __m256i bytes = _mm256_shuffle_epi8(_mm256_set1_epi32(x32), shuf_mask); + const __m256i bit_mask = _mm256_set1_epi64x(0x7fbfdfeff7fbfdfe); + bytes = _mm256_or_si256(bytes, bit_mask); + return _mm256_cmpeq_epi8(bytes, _mm256_set1_epi64x(-1)); +} + +// Unpack 32 4-bit fields into 32 bytes +// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval +static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) +{ + const __m128i tmp = _mm_loadu_si128((const __m128i *)rsi); + const __m256i bytes = MM256_SET_M128I(_mm_srli_epi16(tmp, 4), tmp); + const __m256i lowMask = _mm256_set1_epi8( 0xF ); + return _mm256_and_si256(lowMask, bytes); +} + +// add int16_t pairwise and return as float vector +static inline __m256 sum_i16_pairs_float(const __m256i x) { + const __m256i ones = _mm256_set1_epi16(1); + const __m256i summed_pairs = _mm256_madd_epi16(ones, x); + return _mm256_cvtepi32_ps(summed_pairs); +} + +static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { +#if defined(__AVX512VNNI__) && defined(__AVX512VL__) + const __m256i zero = _mm256_setzero_si256(); + const __m256i summed_pairs = _mm256_dpbusd_epi32(zero, ax, sy); + return _mm256_cvtepi32_ps(summed_pairs); +#elif defined(__AVXVNNI__) + const __m256i zero = _mm256_setzero_si256(); + const __m256i summed_pairs = _mm256_dpbusd_avx_epi32(zero, ax, sy); + return _mm256_cvtepi32_ps(summed_pairs); +#else + // Perform multiplication and create 16-bit values + const __m256i dot = _mm256_maddubs_epi16(ax, sy); + return sum_i16_pairs_float(dot); +#endif +} + +// multiply int8_t, add results pairwise twice and return as float vector +static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { +#if __AVXVNNIINT8__ + const __m256i zero = _mm256_setzero_si256(); + const __m256i summed_pairs = _mm256_dpbssd_epi32(zero, x, y); + return _mm256_cvtepi32_ps(summed_pairs); +#else + // Get absolute values of x vectors + const __m256i ax = _mm256_sign_epi8(x, x); + // Sign the values of the y vectors + const __m256i sy = _mm256_sign_epi8(y, x); + return mul_sum_us8_pairs_float(ax, sy); +#endif +} + +static inline __m128i packNibbles( __m256i bytes ) +{ + // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh +#if __AVX512F__ + const __m256i bytes_srli_4 = _mm256_srli_epi16(bytes, 4); // 0000_0000_abcd_0000 + bytes = _mm256_or_si256(bytes, bytes_srli_4); // 0000_abcd_abcd_efgh + return _mm256_cvtepi16_epi8(bytes); // abcd_efgh +#else + const __m256i lowByte = _mm256_set1_epi16( 0xFF ); + __m256i high = _mm256_andnot_si256( lowByte, bytes ); + __m256i low = _mm256_and_si256( lowByte, bytes ); + high = _mm256_srli_epi16( high, 4 ); + bytes = _mm256_or_si256( low, high ); + + // Compress uint16_t lanes into bytes + __m128i r0 = _mm256_castsi256_si128( bytes ); + __m128i r1 = _mm256_extracti128_si256( bytes, 1 ); + return _mm_packus_epi16( r0, r1 ); +#endif +} +#elif defined(__AVX__) +static inline __m128i packNibbles( __m128i bytes1, __m128i bytes2 ) +{ + // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh + const __m128i lowByte = _mm_set1_epi16( 0xFF ); + __m128i high = _mm_andnot_si128( lowByte, bytes1 ); + __m128i low = _mm_and_si128( lowByte, bytes1 ); + high = _mm_srli_epi16( high, 4 ); + bytes1 = _mm_or_si128( low, high ); + high = _mm_andnot_si128( lowByte, bytes2 ); + low = _mm_and_si128( lowByte, bytes2 ); + high = _mm_srli_epi16( high, 4 ); + bytes2 = _mm_or_si128( low, high ); + + return _mm_packus_epi16( bytes1, bytes2); +} + +static inline __m128i mul_add_epi8_sse(const __m128i x, const __m128i y) { + const __m128i ax = _mm_sign_epi8(x, x); + const __m128i sy = _mm_sign_epi8(y, x); + return _mm_maddubs_epi16(ax, sy); +} + +// spread 32 bits to 32 bytes { 0x00, 0xFF } +static inline __m256i bytes_from_bits_32(const uint8_t * x) { + uint32_t x32; + memcpy(&x32, x, sizeof(uint32_t)); + const __m128i shuf_maskl = _mm_set_epi64x(0x0101010101010101, 0x0000000000000000); + const __m128i shuf_maskh = _mm_set_epi64x(0x0303030303030303, 0x0202020202020202); + __m128i bytesl = _mm_shuffle_epi8(_mm_set1_epi32(x32), shuf_maskl); + __m128i bytesh = _mm_shuffle_epi8(_mm_set1_epi32(x32), shuf_maskh); + const __m128i bit_mask = _mm_set1_epi64x(0x7fbfdfeff7fbfdfe); + bytesl = _mm_or_si128(bytesl, bit_mask); + bytesh = _mm_or_si128(bytesh, bit_mask); + bytesl = _mm_cmpeq_epi8(bytesl, _mm_set1_epi64x(-1)); + bytesh = _mm_cmpeq_epi8(bytesh, _mm_set1_epi64x(-1)); + return MM256_SET_M128I(bytesh, bytesl); +} + +// Unpack 32 4-bit fields into 32 bytes +// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval +static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) +{ + // Load 16 bytes from memory + __m128i tmpl = _mm_loadu_si128((const __m128i *)rsi); + __m128i tmph = _mm_srli_epi16(tmpl, 4); + const __m128i lowMask = _mm_set1_epi8(0xF); + tmpl = _mm_and_si128(lowMask, tmpl); + tmph = _mm_and_si128(lowMask, tmph); + return MM256_SET_M128I(tmph, tmpl); +} + +// add int16_t pairwise and return as float vector +static inline __m256 sum_i16_pairs_float(const __m128i xh, const __m128i xl) { + const __m128i ones = _mm_set1_epi16(1); + const __m128i summed_pairsl = _mm_madd_epi16(ones, xl); + const __m128i summed_pairsh = _mm_madd_epi16(ones, xh); + const __m256i summed_pairs = MM256_SET_M128I(summed_pairsh, summed_pairsl); + return _mm256_cvtepi32_ps(summed_pairs); +} + +static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { + const __m128i axl = _mm256_castsi256_si128(ax); + const __m128i axh = _mm256_extractf128_si256(ax, 1); + const __m128i syl = _mm256_castsi256_si128(sy); + const __m128i syh = _mm256_extractf128_si256(sy, 1); + // Perform multiplication and create 16-bit values + const __m128i dotl = _mm_maddubs_epi16(axl, syl); + const __m128i doth = _mm_maddubs_epi16(axh, syh); + return sum_i16_pairs_float(doth, dotl); +} + +// multiply int8_t, add results pairwise twice and return as float vector +static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { + const __m128i xl = _mm256_castsi256_si128(x); + const __m128i xh = _mm256_extractf128_si256(x, 1); + const __m128i yl = _mm256_castsi256_si128(y); + const __m128i yh = _mm256_extractf128_si256(y, 1); + // Get absolute values of x vectors + const __m128i axl = _mm_sign_epi8(xl, xl); + const __m128i axh = _mm_sign_epi8(xh, xh); + // Sign the values of the y vectors + const __m128i syl = _mm_sign_epi8(yl, xl); + const __m128i syh = _mm_sign_epi8(yh, xh); + // Perform multiplication and create 16-bit values + const __m128i dotl = _mm_maddubs_epi16(axl, syl); + const __m128i doth = _mm_maddubs_epi16(axh, syh); + return sum_i16_pairs_float(doth, dotl); +} + +// larger version of mul_sum_i8_pairs_float where x and y are each represented by four 128-bit vectors +static inline __m256 mul_sum_i8_quad_float(const __m128i x_1_0, const __m128i x_1_1, const __m128i x_2_0, const __m128i x_2_1, + const __m128i y_1_0, const __m128i y_1_1, const __m128i y_2_0, const __m128i y_2_1) { + const __m128i mone = _mm_set1_epi16(1); + + const __m128i p16_1_0 = mul_add_epi8_sse(x_1_0, y_1_0); + const __m128i p16_1_1 = mul_add_epi8_sse(x_1_1, y_1_1); + const __m128i p16_2_0 = mul_add_epi8_sse(x_2_0, y_2_0); + const __m128i p16_2_1 = mul_add_epi8_sse(x_2_1, y_2_1); + const __m128i p_1_0 = _mm_madd_epi16(p16_1_0, mone); + const __m128i p_1_1 = _mm_madd_epi16(p16_1_1, mone); + const __m128i p_2_0 = _mm_madd_epi16(p16_2_0, mone); + const __m128i p_2_1 = _mm_madd_epi16(p16_2_1, mone); + const __m128i p_1 = _mm_add_epi32(p_1_0, p_1_1); + const __m128i p_2 = _mm_add_epi32(p_2_0, p_2_1); + return _mm256_cvtepi32_ps(MM256_SET_M128I(p_2, p_1)); +} + +// quad fp16 delta calculation +static inline __m256 quad_fp16_delta_float(const float x0, const float y0, const float x1, const float y1) { + // GGML_FP16_TO_FP32 is faster than Intel F16C + return _mm256_set_m128(_mm_set1_ps(GGML_FP16_TO_FP32(x1) * GGML_FP16_TO_FP32(y1)), + _mm_set1_ps(GGML_FP16_TO_FP32(x0) * GGML_FP16_TO_FP32(y0))); +} +#endif +#elif defined(__SSSE3__) +// horizontally add 4x4 floats +static inline float hsum_float_4x4(const __m128 a, const __m128 b, const __m128 c, const __m128 d) { + __m128 res_0 =_mm_hadd_ps(a, b); + __m128 res_1 =_mm_hadd_ps(c, d); + __m128 res =_mm_hadd_ps(res_0, res_1); + res =_mm_hadd_ps(res, res); + res =_mm_hadd_ps(res, res); + + return _mm_cvtss_f32(res); +} +#endif // __AVX__ || __AVX2__ || __AVX512F__ +#endif // defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) + +#if defined(__ARM_NEON) || defined(__wasm_simd128__) || defined(__POWER9_VECTOR__) +#define B1(c,s,n) 0x ## n ## c , 0x ## n ## s +#define B2(c,s,n) B1(c,s,n ## c), B1(c,s,n ## s) +#define B3(c,s,n) B2(c,s,n ## c), B2(c,s,n ## s) +#define B4(c,s,n) B3(c,s,n ## c), B3(c,s,n ## s) +#define B5(c,s,n) B4(c,s,n ## c), B4(c,s,n ## s) +#define B6(c,s,n) B5(c,s,n ## c), B5(c,s,n ## s) +#define B7(c,s,n) B6(c,s,n ## c), B6(c,s,n ## s) +#define B8(c,s ) B7(c,s, c), B7(c,s, s) + +// precomputed tables for expanding 8bits to 8 bytes: +static const uint64_t table_b2b_0[1 << 8] = { B8(00, 10) }; // ( b) << 4 +static const uint64_t table_b2b_1[1 << 8] = { B8(10, 00) }; // (!b) << 4 +#endif + +#if defined(__loongarch_asx) + +#ifdef __clang__ +#define VREGS_PREFIX "$vr" +#define XREGS_PREFIX "$xr" +#else // GCC +#define VREGS_PREFIX "$f" +#define XREGS_PREFIX "$f" +#endif +#define __ALL_REGS "0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31" +// Convert __m128i to __m256i +static inline __m256i ____m256i(__m128i in) { + __m256i out = __lasx_xvldi(0); + __asm__ volatile ( + ".irp i," __ALL_REGS "\n\t" + " .ifc %[out], " XREGS_PREFIX"\\i \n\t" + " .irp j," __ALL_REGS "\n\t" + " .ifc %[in], " VREGS_PREFIX "\\j \n\t" + " xvpermi.q $xr\\i, $xr\\j, 0x20 \n\t" + " .endif \n\t" + " .endr \n\t" + " .endif \n\t" + ".endr \n\t" + : [out] "+f" (out) : [in] "f" (in) + ); + return out; +} +// Convert two __m128i to __m256i +static inline __m256i lasx_set_q(__m128i inhi, __m128i inlo) { + __m256i out; + __asm__ volatile ( + ".irp i," __ALL_REGS "\n\t" + " .ifc %[hi], " VREGS_PREFIX "\\i \n\t" + " .irp j," __ALL_REGS "\n\t" + " .ifc %[lo], " VREGS_PREFIX "\\j \n\t" + " xvpermi.q $xr\\i, $xr\\j, 0x20 \n\t" + " .endif \n\t" + " .endr \n\t" + " .endif \n\t" + ".endr \n\t" + ".ifnc %[out], %[hi] \n\t" + ".irp i," __ALL_REGS "\n\t" + " .ifc %[out], " XREGS_PREFIX "\\i \n\t" + " .irp j," __ALL_REGS "\n\t" + " .ifc %[hi], " VREGS_PREFIX "\\j \n\t" + " xvori.b $xr\\i, $xr\\j, 0 \n\t" + " .endif \n\t" + " .endr \n\t" + " .endif \n\t" + ".endr \n\t" + ".endif \n\t" + : [out] "=f" (out), [hi] "+f" (inhi) + : [lo] "f" (inlo) + ); + return out; +} +// Convert __m256i low part to __m128i +static inline __m128i lasx_extracti128_lo(__m256i in) { + __m128i out; + __asm__ volatile ( + ".ifnc %[out], %[in] \n\t" + ".irp i," __ALL_REGS "\n\t" + " .ifc %[out], " VREGS_PREFIX "\\i \n\t" + " .irp j," __ALL_REGS "\n\t" + " .ifc %[in], " XREGS_PREFIX "\\j \n\t" + " vori.b $vr\\i, $vr\\j, 0 \n\t" + " .endif \n\t" + " .endr \n\t" + " .endif \n\t" + ".endr \n\t" + ".endif \n\t" + : [out] "=f" (out) : [in] "f" (in) + ); + return out; +} +// Convert __m256i high part to __m128i +static inline __m128i lasx_extracti128_hi(__m256i in) { + __m128i out; + __asm__ volatile ( + ".irp i," __ALL_REGS "\n\t" + " .ifc %[out], " VREGS_PREFIX "\\i \n\t" + " .irp j," __ALL_REGS "\n\t" + " .ifc %[in], " XREGS_PREFIX "\\j \n\t" + " xvpermi.q $xr\\i, $xr\\j, 0x11 \n\t" + " .endif \n\t" + " .endr \n\t" + " .endif \n\t" + ".endr \n\t" + : [out] "=f" (out) : [in] "f" (in) + ); + return out; +} + +static __m256i lasx_set_w(int e7, int e6, int e5, int e4, int e3, int e2, int e1, int e0) { + v8i32 __ret = {e0, e1, e2, e3, e4, e5, e6, e7}; + return (__m256i)__ret; +} + +static __m128i lsx_set_w(int32_t a, int32_t b, int32_t c, int32_t d) { + v4i32 __ret = {d, c, b, a}; + return (__m128i)__ret; +} + +static __m256i lasx_set_d(int64_t a, int64_t b, int64_t c, int64_t d) { + v4i64 __ret = {d, c, b, a}; + return (__m256i)__ret; +} + +static __m256i lasx_insertf128( __m128i x, __m128i y) { + return lasx_set_q(x, y); +} + +static __m128i lsx_shuffle_b(__m128i a, __m128i b) { + __m128i mask_f, zero, tmp0, tmp2, mask; + int f = 0x8f; + mask_f = __lsx_vreplgr2vr_b(f); + zero = __lsx_vldi(0); + tmp0 = __lsx_vand_v(b, mask_f); // get mask with low 4 bit and sign bits + tmp0 = __lsx_vori_b(tmp0, 0x10); // make each mask or with 0x10 prepare for positive + mask = __lsx_vsle_b(zero, tmp0); // if mask >= 0, set mask + tmp2 = __lsx_vand_v(tmp0, mask); // maskout the in2 < ones + return __lsx_vshuf_b(a, zero, tmp2); +} + +static __m256i lasx_shuffle_b(__m256i a, __m256i b) { + __m256i mask_f, zero, tmp0, tmp2, mask; + int f = 0x8f; + mask_f = __lasx_xvreplgr2vr_b(f); + zero = __lasx_xvldi(0); + tmp0 = __lasx_xvand_v(b, mask_f); // get mask with low 4 bit and sign bits + tmp0 = __lasx_xvori_b(tmp0, 0x10); // make each mask or with 0x10 prepare for positive + mask = __lasx_xvsle_b(zero, tmp0); // if mask >= 0, set mask + tmp2 = __lasx_xvand_v(tmp0, mask); // maskout the in2 < ones + return __lasx_xvshuf_b(a, zero, tmp2); +} + +static __m256i lasx_extu8_16(__m128i a) { + __m128i zero = __lsx_vldi(0); + __m128i vlo = __lsx_vilvl_b(zero, a); + __m128i vhi = __lsx_vilvh_b(zero, a); + return lasx_set_q(vhi, vlo); +} + +static __m256i lasx_ext8_16(__m128i a) { + __m128i sign = __lsx_vslti_b(a, 0); + __m128i vlo = __lsx_vilvl_b(sign, a); + __m128i vhi = __lsx_vilvh_b(sign, a); + return lasx_set_q(vhi, vlo); +} + +static __m256i lasx_ext16_32(__m128i a) { + __m256i tmp1; + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 0), 0); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 1), 1); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 2), 2); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 3), 3); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 4), 4); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 5), 5); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 6), 6); + tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 7), 7); + return tmp1; +} + +static __m128i lasx_extracti128( __m256i a, int pos) { + __m128i ret; + if( pos == 0) + { + ret = lasx_extracti128_lo(a); + } else { + ret = lasx_extracti128_hi(a); + } + return ret; +} + +static __m128 lasx_extractf128( __m256 a, int pos) { + __m128 ret; + if( pos == 0) + { + ret = (__m128)lasx_extracti128_lo((__m256i)a); + } else { + ret = (__m128)lasx_extracti128_hi((__m256i)a); + } + return ret; +} + +static __m128i lsx_hadd_h(__m128i a, __m128i b) { + __m128i tmp1 = __lsx_vpickev_h(b, a); + __m128i tmp2 = __lsx_vpickod_h(b, a); + return __lsx_vadd_h(tmp1, tmp2); +} + +static __m128i lsx_hadd_w(__m128i a, __m128i b) { + __m128i tmp1 = __lsx_vpickev_w(b, a); + __m128i tmp2 = __lsx_vpickod_w(b, a); + return __lsx_vadd_w(tmp1, tmp2); +} + +static __m128 lsx_hadd_s(__m128 a, __m128 b) { + __m128 tmp1 = (__m128)__lsx_vpickev_w((__m128i)b, (__m128i)a); + __m128 tmp2 = (__m128)__lsx_vpickod_w((__m128i)b, (__m128i)a); + + return __lsx_vfadd_s(tmp1, tmp2); +} + +static __m256i lasx_maddubs_h(__m256i a, __m256i b) { + __m256i tmp1, tmp2; + tmp1 = __lasx_xvmulwev_h_b(a, b); + tmp2 = __lasx_xvmulwod_h_b(a, b); + return __lasx_xvsadd_h(tmp1, tmp2); +} + +static __m256i lasx_madd_h(__m256i a, __m256i b) { + __m256i tmp1, tmp2; + tmp1 = __lasx_xvmulwev_w_h(a, b); + tmp2 = __lasx_xvmulwod_w_h(a, b); + return __lasx_xvadd_w(tmp1, tmp2); +} + +static __m256i lasx_packs_w(__m256i a, __m256i b) { + __m256i tmp, tmp1; + tmp = __lasx_xvsat_w(a, 15); + tmp1 = __lasx_xvsat_w(b, 15); + return __lasx_xvpickev_h(tmp1, tmp); +} + +static __m256i lasx_packs_h(__m256i a, __m256i b) { + __m256i tmp, tmp1; + tmp = __lasx_xvsat_h(a, 7); + tmp1 = __lasx_xvsat_h(b, 7); + return __lasx_xvpickev_b(tmp1, tmp); +} + +static __m128i lsx_packs_w(__m128i a, __m128i b) { + __m128i tmp, tmp1; + tmp = __lsx_vsat_w(a, 15); + tmp1 = __lsx_vsat_w(b, 15); + return __lsx_vpickev_h(tmp1, tmp); +} + +static __m128i lsx_packs_h(__m128i a, __m128i b) { + __m128i tmp, tmp1; + tmp = __lsx_vsat_h(a, 7); + tmp1 = __lsx_vsat_h(b, 7); + return __lsx_vpickev_b(tmp1, tmp); +} + +static __m128i lsx_packus_h(__m128i a, __m128i b) { + __m128i tmp, tmp1; + tmp = __lsx_vsat_hu(a, 7); + tmp1 = __lsx_vsat_hu(b, 7); + return __lsx_vpickev_b(tmp1, tmp); +} + + +static __m128i lsx_maddubs_h(__m128i a, __m128i b) { + __m128i tmp1, tmp2; + tmp1 = __lsx_vmulwev_h_b(a, b); + tmp2 = __lsx_vmulwod_h_b(a, b); + return __lsx_vsadd_h(tmp1, tmp2); +} + +static __m128i lsx_madd_h(__m128i a, __m128i b) { + __m128i tmp1, tmp2; + tmp1 = __lsx_vmulwev_w_h(a, b); + tmp2 = __lsx_vmulwod_w_h(a, b); + return __lsx_vadd_w(tmp1, tmp2); +} + +// multiply int8_t, add results pairwise twice +static inline __m128i mul_sum_i8_pairs(const __m128i x, const __m128i y) { + // Get absolute values of x vectors + const __m128i ax = __lsx_vsigncov_b(x, x); + // Sign the values of the y vectors + const __m128i sy = __lsx_vsigncov_b(x, y); + // Perform multiplication and create 16-bit values + const __m128i dot = lsx_maddubs_h(ax, sy); + const __m128i ones = __lsx_vreplgr2vr_h(1); + return lsx_madd_h(ones, dot); +} + +// horizontally add 8 floats +static inline float hsum_float_8(const __m256 x) { + __m128 res = lasx_extractf128(x, 1); + ft_union tmp; + res = __lsx_vfadd_s(res, lasx_extractf128(x, 0)); + res = __lsx_vfadd_s(res, (__m128)__lsx_vpickod_d((__m128i)res, (__m128i)res)); + res = __lsx_vfadd_s(res, (__m128)__lsx_vinsgr2vr_w(__lsx_vldi(0), __lsx_vpickve2gr_w(res, 1), 0)); + tmp.i = __lsx_vpickve2gr_w(res, 0); + return tmp.f; +} + +// horizontally add 8 int32_t +static inline int hsum_i32_8(const __m256i a) { + + __m256i tmp1 = __lasx_xvpermi_q(a, a, 0x11); + __m256i tmp2 = __lasx_xvpermi_q(a, a, 0x00); + + __m128i tmp1_128 = lasx_extracti128_lo(tmp1); + __m128i tmp2_128 = lasx_extracti128_lo(tmp2); + + __m128i sum128 = __lsx_vadd_w(tmp1_128, tmp2_128); + + __m128i ev = __lsx_vpickev_w(sum128, sum128); + __m128i od = __lsx_vpickod_w(sum128, sum128); + __m128i sum64 = __lsx_vadd_w(ev, od); + + int sum64_1, sum64_2; + sum64_1 = __lsx_vpickve2gr_w(sum64, 0); + sum64_2 = __lsx_vpickve2gr_w(sum64, 1); + + return sum64_1 + sum64_2; +} + +// horizontally add 4 int32_t +static inline int hsum_i32_4(const __m128i a) { + __m128i ev = __lsx_vpickev_w(a, a); + __m128i od = __lsx_vpickod_w(a, a); + __m128i sum64 = __lsx_vadd_w(ev, od); + + int sum64_1, sum64_2; + sum64_1 = __lsx_vpickve2gr_w(sum64, 0); + sum64_2 = __lsx_vpickve2gr_w(sum64, 1); + + return sum64_1 + sum64_2; +} + +// spread 32 bits to 32 bytes { 0x00, 0xFF } +static inline __m256i bytes_from_bits_32(const uint8_t * x) { + + uint32_t x32; + memcpy(&x32, x, sizeof(uint32_t)); + const __m256i shuf_mask = lasx_set_d( + 0x0303030303030303, 0x0202020202020202, + 0x0101010101010101, 0x0000000000000000); + + __m256i bytes = lasx_shuffle_b(__lasx_xvreplgr2vr_w(x32), shuf_mask); + const __m256i bit_mask = __lasx_xvreplgr2vr_d(0x7fbfdfeff7fbfdfe); + bytes = __lasx_xvor_v(bytes, bit_mask); + return __lasx_xvseq_b(bytes, __lasx_xvreplgr2vr_d(-1)); +} + +// Unpack 32 4-bit fields into 32 bytes +// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval +static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) { + const __m128i lo = __lsx_vld((const __m128i *)rsi, 0); + __m128i hi = __lsx_vsrli_h(lo, 4); + return __lasx_xvandi_b(lasx_insertf128(hi, lo), 0xf); +} + +// add int16_t pairwise and return as float vector +static inline __m256 sum_i16_pairs_float(const __m256i x) { + __m256i v = __lasx_xvpackod_h(x, x); + __m256i summed_pairs = __lasx_xvaddwev_w_h(x, v); + return __lasx_xvffint_s_w(summed_pairs); +} + +static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { + // Perform multiplication and create 16-bit values + const __m256i dot = lasx_maddubs_h(ax, sy); + return sum_i16_pairs_float(dot); +} + +// multiply int8_t, add results pairwise twice and return as float vector +static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { + + // Get absolute values of x vectors + const __m256i ax = __lasx_xvsigncov_b(x, x); + // Sign the values of the y vectors + const __m256i sy = __lasx_xvsigncov_b(x, y); + + return mul_sum_us8_pairs_float(ax, sy); +} + +static inline __m128i packNibbles( __m256i bytes ) { + // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh + const __m256i lowByte = __lasx_xvreplgr2vr_h(0xFF); + __m256i high = __lasx_xvandn_v(lowByte, bytes); + __m256i low = __lasx_xvand_v(lowByte, bytes); + high = __lasx_xvsrli_h(high, 4); + bytes = __lasx_xvor_v(low, high); + // Compress uint16_t lanes into bytes + __m128i *r0 = (__m128i *)&bytes; + __m256i tmp_h128 = __lasx_xvpermi_q(bytes, bytes, 0x11); + __m128i *r1 = (__m128i *)&tmp_h128; + + __m128i zero = __lsx_vldi(0); + __m128i tmp, tmp2, tmp3; + + tmp = __lsx_vmax_h(zero, *r0); + tmp2 = __lsx_vsat_hu(tmp, 7); + + tmp = __lsx_vmax_h(zero, *r1); + tmp3 = __lsx_vsat_hu(tmp, 7); + return __lsx_vpickev_b(tmp3, tmp2); +} +#endif //__loongarch_asx + +void quantize_row_q4_0(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q4_0_ref(x, y, k); +} + +void quantize_row_q4_1(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q4_1_ref(x, y, k); +} + +void quantize_row_q5_0(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q5_0_ref(x, y, k); +} + +void quantize_row_q5_1(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q5_1_ref(x, y, k); +} + +void quantize_row_q8_0(const float * restrict x, void * restrict vy, int64_t k) { + assert(QK8_0 == 32); + assert(k % QK8_0 == 0); + const int nb = k / QK8_0; + + block_q8_0 * restrict y = vy; + +#if defined(__ARM_NEON) + for (int i = 0; i < nb; i++) { + float32x4_t srcv [8]; + float32x4_t asrcv[8]; + float32x4_t amaxv[8]; + + for (int j = 0; j < 8; j++) srcv[j] = vld1q_f32(x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = vabsq_f32(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = vmaxq_f32(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = vmaxq_f32(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = vmaxq_f32(amaxv[8*j], amaxv[8*j+4]); + + const float amax = vmaxvq_f32(amaxv[0]); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + for (int j = 0; j < 8; j++) { + const float32x4_t v = vmulq_n_f32(srcv[j], id); + const int32x4_t vi = vcvtnq_s32_f32(v); + + y[i].qs[4*j + 0] = vgetq_lane_s32(vi, 0); + y[i].qs[4*j + 1] = vgetq_lane_s32(vi, 1); + y[i].qs[4*j + 2] = vgetq_lane_s32(vi, 2); + y[i].qs[4*j + 3] = vgetq_lane_s32(vi, 3); + } + } +#elif defined(__wasm_simd128__) + for (int i = 0; i < nb; i++) { + v128_t srcv [8]; + v128_t asrcv[8]; + v128_t amaxv[8]; + + for (int j = 0; j < 8; j++) srcv[j] = wasm_v128_load(x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = wasm_f32x4_abs(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = wasm_f32x4_max(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = wasm_f32x4_max(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = wasm_f32x4_max(amaxv[8*j], amaxv[8*j+4]); + + const float amax = MAX(MAX(wasm_f32x4_extract_lane(amaxv[0], 0), + wasm_f32x4_extract_lane(amaxv[0], 1)), + MAX(wasm_f32x4_extract_lane(amaxv[0], 2), + wasm_f32x4_extract_lane(amaxv[0], 3))); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + for (int j = 0; j < 8; j++) { + const v128_t v = wasm_f32x4_mul(srcv[j], wasm_f32x4_splat(id)); + const v128_t vi = wasm_i32x4_trunc_sat_f32x4(v); + + y[i].qs[4*j + 0] = wasm_i32x4_extract_lane(vi, 0); + y[i].qs[4*j + 1] = wasm_i32x4_extract_lane(vi, 1); + y[i].qs[4*j + 2] = wasm_i32x4_extract_lane(vi, 2); + y[i].qs[4*j + 3] = wasm_i32x4_extract_lane(vi, 3); + } + } +#elif defined(__AVX2__) || defined(__AVX__) + for (int i = 0; i < nb; i++) { + // Load elements into 4 AVX vectors + __m256 v0 = _mm256_loadu_ps( x ); + __m256 v1 = _mm256_loadu_ps( x + 8 ); + __m256 v2 = _mm256_loadu_ps( x + 16 ); + __m256 v3 = _mm256_loadu_ps( x + 24 ); + x += 32; + + // Compute max(abs(e)) for the block + const __m256 signBit = _mm256_set1_ps( -0.0f ); + __m256 maxAbs = _mm256_andnot_ps( signBit, v0 ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v1 ) ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v2 ) ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v3 ) ); + + __m128 max4 = _mm_max_ps( _mm256_extractf128_ps( maxAbs, 1 ), _mm256_castps256_ps128( maxAbs ) ); + max4 = _mm_max_ps( max4, _mm_movehl_ps( max4, max4 ) ); + max4 = _mm_max_ss( max4, _mm_movehdup_ps( max4 ) ); + const float maxScalar = _mm_cvtss_f32( max4 ); + + // Quantize these floats + const float d = maxScalar / 127.f; + y[i].d = GGML_FP32_TO_FP16(d); + const float id = ( maxScalar != 0.0f ) ? 127.f / maxScalar : 0.0f; + const __m256 mul = _mm256_set1_ps( id ); + + // Apply the multiplier + v0 = _mm256_mul_ps( v0, mul ); + v1 = _mm256_mul_ps( v1, mul ); + v2 = _mm256_mul_ps( v2, mul ); + v3 = _mm256_mul_ps( v3, mul ); + + // Round to nearest integer + v0 = _mm256_round_ps( v0, _MM_ROUND_NEAREST ); + v1 = _mm256_round_ps( v1, _MM_ROUND_NEAREST ); + v2 = _mm256_round_ps( v2, _MM_ROUND_NEAREST ); + v3 = _mm256_round_ps( v3, _MM_ROUND_NEAREST ); + + // Convert floats to integers + __m256i i0 = _mm256_cvtps_epi32( v0 ); + __m256i i1 = _mm256_cvtps_epi32( v1 ); + __m256i i2 = _mm256_cvtps_epi32( v2 ); + __m256i i3 = _mm256_cvtps_epi32( v3 ); + +#if defined(__AVX2__) + // Convert int32 to int16 + i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 + i2 = _mm256_packs_epi32( i2, i3 ); // 16, 17, 18, 19, 24, 25, 26, 27, 20, 21, 22, 23, 28, 29, 30, 31 + // Convert int16 to int8 + i0 = _mm256_packs_epi16( i0, i2 ); // 0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27, 4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 + + // We got our precious signed bytes, but the order is now wrong + // These AVX2 pack instructions process 16-byte pieces independently + // The following instruction is fixing the order + const __m256i perm = _mm256_setr_epi32( 0, 4, 1, 5, 2, 6, 3, 7 ); + i0 = _mm256_permutevar8x32_epi32( i0, perm ); + + _mm256_storeu_si256((__m256i *)y[i].qs, i0); +#else + // Since we don't have in AVX some necessary functions, + // we split the registers in half and call AVX2 analogs from SSE + __m128i ni0 = _mm256_castsi256_si128( i0 ); + __m128i ni1 = _mm256_extractf128_si256( i0, 1); + __m128i ni2 = _mm256_castsi256_si128( i1 ); + __m128i ni3 = _mm256_extractf128_si256( i1, 1); + __m128i ni4 = _mm256_castsi256_si128( i2 ); + __m128i ni5 = _mm256_extractf128_si256( i2, 1); + __m128i ni6 = _mm256_castsi256_si128( i3 ); + __m128i ni7 = _mm256_extractf128_si256( i3, 1); + + // Convert int32 to int16 + ni0 = _mm_packs_epi32( ni0, ni1 ); + ni2 = _mm_packs_epi32( ni2, ni3 ); + ni4 = _mm_packs_epi32( ni4, ni5 ); + ni6 = _mm_packs_epi32( ni6, ni7 ); + // Convert int16 to int8 + ni0 = _mm_packs_epi16( ni0, ni2 ); + ni4 = _mm_packs_epi16( ni4, ni6 ); + + _mm_storeu_si128((__m128i *)(y[i].qs + 0), ni0); + _mm_storeu_si128((__m128i *)(y[i].qs + 16), ni4); +#endif + } +#elif defined(__riscv_v_intrinsic) + + size_t vl = __riscv_vsetvl_e32m4(QK8_0); + + for (int i = 0; i < nb; i++) { + // load elements + vfloat32m4_t v_x = __riscv_vle32_v_f32m4(x+i*QK8_0, vl); + + vfloat32m4_t vfabs = __riscv_vfabs_v_f32m4(v_x, vl); + vfloat32m1_t tmp = __riscv_vfmv_v_f_f32m1(0.0f, vl); + vfloat32m1_t vmax = __riscv_vfredmax_vs_f32m4_f32m1(vfabs, tmp, vl); + float amax = __riscv_vfmv_f_s_f32m1_f32(vmax); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + vfloat32m4_t x0 = __riscv_vfmul_vf_f32m4(v_x, id, vl); + + // convert to integer + vint16m2_t vi = __riscv_vfncvt_x_f_w_i16m2(x0, vl); + vint8m1_t vs = __riscv_vncvt_x_x_w_i8m1(vi, vl); + + // store result + __riscv_vse8_v_i8m1(y[i].qs , vs, vl); + } + +#elif defined(__POWER9_VECTOR__) + for (int i = 0; i < nb; i++) { + vector float srcv [8]; + vector float asrcv[8]; + vector float amaxv[8]; + vector signed int vi[8]; + + for (int j = 0; j < 8; j++) srcv[j] = vec_xl(0, x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = vec_abs(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = vec_max(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = vec_max(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = vec_max(amaxv[8*j], amaxv[8*j+4]); + + const float amax = MAX(MAX(vec_extract(amaxv[0], 0), + vec_extract(amaxv[0], 1)), + MAX(vec_extract(amaxv[0], 2), + vec_extract(amaxv[0], 3))); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + const vector float vid = vec_splats(id); + + y[i].d = GGML_FP32_TO_FP16(d); + + for (int j = 0; j < 8; j++) { + const vector float v = vec_round(vec_mul(srcv[j], vid)); + vi[j] = vec_cts(v, 0); + } + vec_xst(vec_pack(vec_pack(vi[0], vi[1]), vec_pack(vi[2], vi[3])), 0, &y[i].qs[0]); + vec_xst(vec_pack(vec_pack(vi[4], vi[5]), vec_pack(vi[6], vi[7])), 16, &y[i].qs[0]); + } + +#elif defined(__loongarch_asx) + for (int i = 0; i < nb; i++) { + ft_union fi; + __m256 v0 = (__m256)__lasx_xvld( x , 0); + __m256 v1 = (__m256)__lasx_xvld( x , 32); + __m256 v2 = (__m256)__lasx_xvld( x , 64); + __m256 v3 = (__m256)__lasx_xvld( x , 96); + x += 32; + + // Compute max(abs(e)) for the block + const __m256 sign_bit = __lasx_xvreplfr2vr_s( -0.0f ); + __m256 max_abs = (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v0 ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v1 ) ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v2 ) ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v3 ) ); + + __m128 max4 = __lsx_vfmax_s( lasx_extractf128( max_abs, 1 ), lasx_extractf128( max_abs , 0) ); + max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vpickod_d((__m128i) max4, (__m128i)max4 ) ); + __m128 tmp = max4; + max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vinsgr2vr_w(tmp, __lsx_vpickve2gr_w( max4, 1 ), 0 )); + fi.i = __lsx_vpickve2gr_w( (__m128i)max4, 0 ); + const float max_scalar = fi.f; + + // Quantize these floats + const float d = max_scalar / 127.f; + y[i].d = GGML_FP32_TO_FP16(d); + const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; + const __m256 mul = (__m256)__lasx_xvreplfr2vr_s( id ); + + // Apply the multiplier + v0 = __lasx_xvfmul_s( v0, mul ); + v1 = __lasx_xvfmul_s( v1, mul ); + v2 = __lasx_xvfmul_s( v2, mul ); + v3 = __lasx_xvfmul_s( v3, mul ); + + // Round to nearest integer + __m256i i0 = __lasx_xvftintrne_w_s( v0 ); + __m256i i1 = __lasx_xvftintrne_w_s( v1 ); + __m256i i2 = __lasx_xvftintrne_w_s( v2 ); + __m256i i3 = __lasx_xvftintrne_w_s( v3 ); + + __m128i ni0 = lasx_extracti128( i0, 0 ); + __m128i ni1 = lasx_extracti128( i0, 1); + __m128i ni2 = lasx_extracti128( i1, 0); + __m128i ni3 = lasx_extracti128( i1, 1); + __m128i ni4 = lasx_extracti128( i2, 0); + __m128i ni5 = lasx_extracti128( i2, 1); + __m128i ni6 = lasx_extracti128( i3, 0); + __m128i ni7 = lasx_extracti128( i3, 1); + + // Convert int32 to int16 + ni0 = lsx_packs_w( ni0, ni1 ); + ni2 = lsx_packs_w( ni2, ni3 ); + ni4 = lsx_packs_w( ni4, ni5 ); + ni6 = lsx_packs_w( ni6, ni7 ); + // Convert int16 to int8 + ni0 = lsx_packs_h( ni0, ni2 ); + ni4 = lsx_packs_h( ni4, ni6 ); + + __lsx_vst(ni0, (__m128i *)(y[i].qs + 0), 0); + __lsx_vst(ni4, (__m128i *)(y[i].qs + 16), 0); + + } +#else + GGML_UNUSED(nb); + // scalar + quantize_row_q8_0_ref(x, y, k); +#endif +} + +void quantize_row_q8_1(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK8_1 == 0); + const int nb = k / QK8_1; + + block_q8_1 * restrict y = vy; + +#if defined(__ARM_NEON) + for (int i = 0; i < nb; i++) { + float32x4_t srcv [8]; + float32x4_t asrcv[8]; + float32x4_t amaxv[8]; + + for (int j = 0; j < 8; j++) srcv[j] = vld1q_f32(x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = vabsq_f32(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = vmaxq_f32(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = vmaxq_f32(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = vmaxq_f32(amaxv[8*j], amaxv[8*j+4]); + + const float amax = vmaxvq_f32(amaxv[0]); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + int32x4_t accv = vdupq_n_s32(0); + + for (int j = 0; j < 8; j++) { + const float32x4_t v = vmulq_n_f32(srcv[j], id); + const int32x4_t vi = vcvtnq_s32_f32(v); + + y[i].qs[4*j + 0] = vgetq_lane_s32(vi, 0); + y[i].qs[4*j + 1] = vgetq_lane_s32(vi, 1); + y[i].qs[4*j + 2] = vgetq_lane_s32(vi, 2); + y[i].qs[4*j + 3] = vgetq_lane_s32(vi, 3); + + accv = vaddq_s32(accv, vi); + } + + y[i].s = GGML_FP32_TO_FP16(d * vaddvq_s32(accv)); + } +#elif defined(__wasm_simd128__) + for (int i = 0; i < nb; i++) { + v128_t srcv [8]; + v128_t asrcv[8]; + v128_t amaxv[8]; + + for (int j = 0; j < 8; j++) srcv[j] = wasm_v128_load(x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = wasm_f32x4_abs(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = wasm_f32x4_max(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = wasm_f32x4_max(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = wasm_f32x4_max(amaxv[8*j], amaxv[8*j+4]); + + const float amax = MAX(MAX(wasm_f32x4_extract_lane(amaxv[0], 0), + wasm_f32x4_extract_lane(amaxv[0], 1)), + MAX(wasm_f32x4_extract_lane(amaxv[0], 2), + wasm_f32x4_extract_lane(amaxv[0], 3))); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + v128_t accv = wasm_i32x4_splat(0); + + for (int j = 0; j < 8; j++) { + const v128_t v = wasm_f32x4_mul(srcv[j], wasm_f32x4_splat(id)); + const v128_t vi = wasm_i32x4_trunc_sat_f32x4(v); + + y[i].qs[4*j + 0] = wasm_i32x4_extract_lane(vi, 0); + y[i].qs[4*j + 1] = wasm_i32x4_extract_lane(vi, 1); + y[i].qs[4*j + 2] = wasm_i32x4_extract_lane(vi, 2); + y[i].qs[4*j + 3] = wasm_i32x4_extract_lane(vi, 3); + + accv = wasm_i32x4_add(accv, vi); + } + + y[i].s = GGML_FP32_TO_FP16( + d * (wasm_i32x4_extract_lane(accv, 0) + + wasm_i32x4_extract_lane(accv, 1) + + wasm_i32x4_extract_lane(accv, 2) + + wasm_i32x4_extract_lane(accv, 3))); + } +#elif defined(__AVX2__) || defined(__AVX__) + for (int i = 0; i < nb; i++) { + // Load elements into 4 AVX vectors + __m256 v0 = _mm256_loadu_ps( x ); + __m256 v1 = _mm256_loadu_ps( x + 8 ); + __m256 v2 = _mm256_loadu_ps( x + 16 ); + __m256 v3 = _mm256_loadu_ps( x + 24 ); + x += 32; + + // Compute max(abs(e)) for the block + const __m256 signBit = _mm256_set1_ps( -0.0f ); + __m256 maxAbs = _mm256_andnot_ps( signBit, v0 ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v1 ) ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v2 ) ); + maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v3 ) ); + + __m128 max4 = _mm_max_ps( _mm256_extractf128_ps( maxAbs, 1 ), _mm256_castps256_ps128( maxAbs ) ); + max4 = _mm_max_ps( max4, _mm_movehl_ps( max4, max4 ) ); + max4 = _mm_max_ss( max4, _mm_movehdup_ps( max4 ) ); + const float max_scalar = _mm_cvtss_f32( max4 ); + + // Quantize these floats + const float d = max_scalar / 127.f; + y[i].d = GGML_FP32_TO_FP16(d); + const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; + const __m256 mul = _mm256_set1_ps( id ); + + // Apply the multiplier + v0 = _mm256_mul_ps( v0, mul ); + v1 = _mm256_mul_ps( v1, mul ); + v2 = _mm256_mul_ps( v2, mul ); + v3 = _mm256_mul_ps( v3, mul ); + + // Round to nearest integer + v0 = _mm256_round_ps( v0, _MM_ROUND_NEAREST ); + v1 = _mm256_round_ps( v1, _MM_ROUND_NEAREST ); + v2 = _mm256_round_ps( v2, _MM_ROUND_NEAREST ); + v3 = _mm256_round_ps( v3, _MM_ROUND_NEAREST ); + + // Convert floats to integers + __m256i i0 = _mm256_cvtps_epi32( v0 ); + __m256i i1 = _mm256_cvtps_epi32( v1 ); + __m256i i2 = _mm256_cvtps_epi32( v2 ); + __m256i i3 = _mm256_cvtps_epi32( v3 ); + +#if defined(__AVX2__) + // Compute the sum of the quants and set y[i].s + y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_8(_mm256_add_epi32(_mm256_add_epi32(i0, i1), _mm256_add_epi32(i2, i3)))); + + // Convert int32 to int16 + i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 + i2 = _mm256_packs_epi32( i2, i3 ); // 16, 17, 18, 19, 24, 25, 26, 27, 20, 21, 22, 23, 28, 29, 30, 31 + // Convert int16 to int8 + i0 = _mm256_packs_epi16( i0, i2 ); // 0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27, 4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 + + // We got our precious signed bytes, but the order is now wrong + // These AVX2 pack instructions process 16-byte pieces independently + // The following instruction is fixing the order + const __m256i perm = _mm256_setr_epi32( 0, 4, 1, 5, 2, 6, 3, 7 ); + i0 = _mm256_permutevar8x32_epi32( i0, perm ); + + _mm256_storeu_si256((__m256i *)y[i].qs, i0); +#else + // Since we don't have in AVX some necessary functions, + // we split the registers in half and call AVX2 analogs from SSE + __m128i ni0 = _mm256_castsi256_si128( i0 ); + __m128i ni1 = _mm256_extractf128_si256( i0, 1); + __m128i ni2 = _mm256_castsi256_si128( i1 ); + __m128i ni3 = _mm256_extractf128_si256( i1, 1); + __m128i ni4 = _mm256_castsi256_si128( i2 ); + __m128i ni5 = _mm256_extractf128_si256( i2, 1); + __m128i ni6 = _mm256_castsi256_si128( i3 ); + __m128i ni7 = _mm256_extractf128_si256( i3, 1); + + // Compute the sum of the quants and set y[i].s + const __m128i s0 = _mm_add_epi32(_mm_add_epi32(ni0, ni1), _mm_add_epi32(ni2, ni3)); + const __m128i s1 = _mm_add_epi32(_mm_add_epi32(ni4, ni5), _mm_add_epi32(ni6, ni7)); + y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_4(_mm_add_epi32(s0, s1))); + + // Convert int32 to int16 + ni0 = _mm_packs_epi32( ni0, ni1 ); + ni2 = _mm_packs_epi32( ni2, ni3 ); + ni4 = _mm_packs_epi32( ni4, ni5 ); + ni6 = _mm_packs_epi32( ni6, ni7 ); + // Convert int16 to int8 + ni0 = _mm_packs_epi16( ni0, ni2 ); + ni4 = _mm_packs_epi16( ni4, ni6 ); + + _mm_storeu_si128((__m128i *)(y[i].qs + 0), ni0); + _mm_storeu_si128((__m128i *)(y[i].qs + 16), ni4); +#endif + } +#elif defined(__riscv_v_intrinsic) + + size_t vl = __riscv_vsetvl_e32m4(QK8_1); + + for (int i = 0; i < nb; i++) { + // load elements + vfloat32m4_t v_x = __riscv_vle32_v_f32m4(x+i*QK8_1, vl); + + vfloat32m4_t vfabs = __riscv_vfabs_v_f32m4(v_x, vl); + vfloat32m1_t tmp = __riscv_vfmv_v_f_f32m1(0.0, vl); + vfloat32m1_t vmax = __riscv_vfredmax_vs_f32m4_f32m1(vfabs, tmp, vl); + float amax = __riscv_vfmv_f_s_f32m1_f32(vmax); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + + y[i].d = GGML_FP32_TO_FP16(d); + + vfloat32m4_t x0 = __riscv_vfmul_vf_f32m4(v_x, id, vl); + + // convert to integer + vint16m2_t vi = __riscv_vfncvt_x_f_w_i16m2(x0, vl); + vint8m1_t vs = __riscv_vncvt_x_x_w_i8m1(vi, vl); + + // store result + __riscv_vse8_v_i8m1(y[i].qs , vs, vl); + + // compute sum for y[i].s + vint16m1_t tmp2 = __riscv_vmv_v_x_i16m1(0, vl); + vint16m1_t vwrs = __riscv_vwredsum_vs_i8m1_i16m1(vs, tmp2, vl); + + // set y[i].s + int sum = __riscv_vmv_x_s_i16m1_i16(vwrs); + y[i].s = GGML_FP32_TO_FP16(sum*d); + } + +#elif defined(__POWER9_VECTOR__) + for (int i = 0; i < nb; i++) { + vector float srcv [8]; + vector float asrcv[8]; + vector float amaxv[8]; + vector signed int vi[8]; + + for (int j = 0; j < 8; j++) srcv[j] = vec_xl(0, x + i*32 + 4*j); + for (int j = 0; j < 8; j++) asrcv[j] = vec_abs(srcv[j]); + + for (int j = 0; j < 4; j++) amaxv[2*j] = vec_max(asrcv[2*j], asrcv[2*j+1]); + for (int j = 0; j < 2; j++) amaxv[4*j] = vec_max(amaxv[4*j], amaxv[4*j+2]); + for (int j = 0; j < 1; j++) amaxv[8*j] = vec_max(amaxv[8*j], amaxv[8*j+4]); + + const float amax = MAX(MAX(vec_extract(amaxv[0], 0), + vec_extract(amaxv[0], 1)), + MAX(vec_extract(amaxv[0], 2), + vec_extract(amaxv[0], 3))); + + const float d = amax / ((1 << 7) - 1); + const float id = d ? 1.0f/d : 0.0f; + const vector float vid = vec_splats(id); + + y[i].d = GGML_FP32_TO_FP16(d); + + vector int accv = vec_splats(0); + + for (int j = 0; j < 8; j++) { + const vector float v = vec_round(vec_mul(srcv[j], vid)); + vi[j] = vec_cts(v, 0); + + accv = vec_add(accv, vi[j]); + } + vec_xst(vec_pack(vec_pack(vi[0], vi[1]), vec_pack(vi[2], vi[3])), 0, &y[i].qs[0]); + vec_xst(vec_pack(vec_pack(vi[4], vi[5]), vec_pack(vi[6], vi[7])), 16, &y[i].qs[0]); + + accv = vec_add(accv, vec_sld(accv, accv, 4)); + accv = vec_add(accv, vec_sld(accv, accv, 8)); + y[i].s = GGML_FP32_TO_FP16(d * vec_extract(accv, 0)); + } + +#elif defined(__loongarch_asx) + for (int i = 0; i < nb; i++) { + ft_union ft; + __m256 v0 = (__m256)__lasx_xvld( x , 0 ); + __m256 v1 = (__m256)__lasx_xvld( x , 32 ); + __m256 v2 = (__m256)__lasx_xvld( x , 64 ); + __m256 v3 = (__m256)__lasx_xvld( x , 96 ); + x += 32; + + // Compute max(abs(e)) for the block + const __m256 sign_bit = __lasx_xvreplfr2vr_s( -0.0f ); + __m256 max_abs = (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v0 ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v1 ) ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v2 ) ); + max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v3 ) ); + + __m128 max4 = __lsx_vfmax_s( lasx_extractf128( max_abs, 1 ), lasx_extractf128( max_abs, 0) ); + max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vpickod_d((__m128i) max4, (__m128i)max4 ) ); + __m128 tmp = max4; + max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vextrins_w((__m128i)tmp, (__m128i)max4, 0x10 )); + ft.i = __lsx_vpickve2gr_w( (__m128i)max4, 0 ); + const float max_scalar = ft.f; + + // Quantize these floats + const float d = max_scalar / 127.f; + y[i].d = GGML_FP32_TO_FP16(d); + const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; + const __m256 mul = __lasx_xvreplfr2vr_s( id ); + + // Apply the multiplier + v0 = __lasx_xvfmul_s( v0, mul ); + v1 = __lasx_xvfmul_s( v1, mul ); + v2 = __lasx_xvfmul_s( v2, mul ); + v3 = __lasx_xvfmul_s( v3, mul ); + + // Round to nearest integer + __m256i i0 = __lasx_xvftintrne_w_s( v0 ); + __m256i i1 = __lasx_xvftintrne_w_s( v1 ); + __m256i i2 = __lasx_xvftintrne_w_s( v2 ); + __m256i i3 = __lasx_xvftintrne_w_s( v3 ); + + __m128i ni0 = lasx_extracti128(i0, 0); + __m128i ni1 = lasx_extracti128( i0, 1); + __m128i ni2 = lasx_extracti128( i1, 0); + __m128i ni3 = lasx_extracti128( i1, 1); + __m128i ni4 = lasx_extracti128( i2, 0 ); + __m128i ni5 = lasx_extracti128( i2, 1); + __m128i ni6 = lasx_extracti128( i3, 0); + __m128i ni7 = lasx_extracti128( i3, 1); + + // Compute the sum of the quants and set y[i].s + const __m128i s0 = __lsx_vadd_w(__lsx_vadd_w(ni0, ni1), __lsx_vadd_w(ni2, ni3)); + const __m128i s1 = __lsx_vadd_w(__lsx_vadd_w(ni4, ni5), __lsx_vadd_w(ni6, ni7)); + y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_4(__lsx_vadd_w(s0, s1))); + + // Convert int32 to int16 + ni0 = lsx_packs_w( ni0, ni1 ); + ni2 = lsx_packs_w( ni2, ni3 ); + ni4 = lsx_packs_w( ni4, ni5 ); + ni6 = lsx_packs_w( ni6, ni7 ); + // Convert int16 to int8 + ni0 = lsx_packs_h( ni0, ni2 ); + ni4 = lsx_packs_h( ni4, ni6 ); + + __lsx_vst(ni0, (__m128i *)(y[i].qs + 0), 0); + __lsx_vst(ni4, (__m128i *)(y[i].qs + 16), 0); + } +#else + GGML_UNUSED(nb); + // scalar + quantize_row_q8_1_ref(x, y, k); +#endif +} + +// +// 2-6 bit quantization in super-blocks +// + +// +// ===================== Helper functions +// +static inline int nearest_int(float fval) { + assert(fabsf(fval) <= 4194303.f); + float val = fval + 12582912.f; + int i; memcpy(&i, &val, sizeof(int)); + return (i & 0x007fffff) - 0x00400000; +} + +static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t * restrict L, int rmse_type, + const float * restrict qw) { + float max = 0; + float amax = 0; + for (int i = 0; i < n; ++i) { + float ax = fabsf(x[i]); + if (ax > amax) { amax = ax; max = x[i]; } + } + if (amax < GROUP_MAX_EPS) { // all zero + for (int i = 0; i < n; ++i) { + L[i] = 0; + } + return 0.f; + } + float iscale = -nmax / max; + if (rmse_type == 0) { + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + L[i] = nmax + MAX(-nmax, MIN(nmax-1, l)); + } + return 1/iscale; + } + bool return_early = false; + if (rmse_type < 0) { + rmse_type = -rmse_type; + return_early = true; + } + float sumlx = 0; + float suml2 = 0; +#ifdef HAVE_BUGGY_APPLE_LINKER + // use 'volatile' to prevent unroll and work around a bug in Apple ld64 1015.7 + for (volatile int i = 0; i < n; ++i) { +#else + for (int i = 0; i < n; ++i) { +#endif + int l = nearest_int(iscale * x[i]); + l = MAX(-nmax, MIN(nmax-1, l)); + L[i] = l + nmax; + float w = qw ? qw[i] : rmse_type == 1 ? x[i] * x[i] : rmse_type == 2 ? 1 : rmse_type == 3 ? fabsf(x[i]) : sqrtf(fabsf(x[i])); + sumlx += w*x[i]*l; + suml2 += w*l*l; + } + float scale = suml2 ? sumlx/suml2 : 0.0f; + if (return_early) return suml2 > 0 ? 0.5f*(scale + 1/iscale) : 1/iscale; + float best = scale * sumlx; + for (int is = -9; is <= 9; ++is) { + if (is == 0) { + continue; + } + iscale = -(nmax + 0.1f*is) / max; + sumlx = suml2 = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + l = MAX(-nmax, MIN(nmax-1, l)); + float w = qw ? qw[i] : rmse_type == 1 ? x[i] * x[i] : rmse_type == 2 ? 1 : rmse_type == 3 ? fabsf(x[i]) : sqrtf(fabsf(x[i])); + sumlx += w*x[i]*l; + suml2 += w*l*l; + } + if (suml2 > 0 && sumlx*sumlx > best*suml2) { + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + L[i] = nmax + MAX(-nmax, MIN(nmax-1, l)); + } + scale = sumlx/suml2; best = scale*sumlx; + } + } + return scale; +} + +static float make_q3_quants(int n, int nmax, const float * restrict x, int8_t * restrict L, bool do_rmse) { + float max = 0; + float amax = 0; + for (int i = 0; i < n; ++i) { + float ax = fabsf(x[i]); + if (ax > amax) { amax = ax; max = x[i]; } + } + if (amax < GROUP_MAX_EPS) { // all zero + for (int i = 0; i < n; ++i) { L[i] = 0; } + return 0.f; + } + float iscale = -nmax / max; + if (do_rmse) { + float sumlx = 0; + float suml2 = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + l = MAX(-nmax, MIN(nmax-1, l)); + L[i] = l; + float w = x[i]*x[i]; + sumlx += w*x[i]*l; + suml2 += w*l*l; + } + for (int itry = 0; itry < 5; ++itry) { + int n_changed = 0; + for (int i = 0; i < n; ++i) { + float w = x[i]*x[i]; + float slx = sumlx - w*x[i]*L[i]; + if (slx > 0) { + float sl2 = suml2 - w*L[i]*L[i]; + int new_l = nearest_int(x[i] * sl2 / slx); + new_l = MAX(-nmax, MIN(nmax-1, new_l)); + if (new_l != L[i]) { + slx += w*x[i]*new_l; + sl2 += w*new_l*new_l; + if (sl2 > 0 && slx*slx*suml2 > sumlx*sumlx*sl2) { + L[i] = new_l; sumlx = slx; suml2 = sl2; + ++n_changed; + } + } + } + } + if (!n_changed) { + break; + } + } + for (int i = 0; i < n; ++i) { + L[i] += nmax; + } + return sumlx / suml2; + } + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + l = MAX(-nmax, MIN(nmax-1, l)); + L[i] = l + nmax; + } + return 1/iscale; +} + +static float make_qkx1_quants(int n, int nmax, const float * restrict x, uint8_t * restrict L, float * restrict the_min, + int ntry, float alpha) { + float min = x[0]; + float max = x[0]; + for (int i = 1; i < n; ++i) { + if (x[i] < min) min = x[i]; + if (x[i] > max) max = x[i]; + } + if (max == min) { + for (int i = 0; i < n; ++i) L[i] = 0; + *the_min = 0; + return 0.f; + } + if (min > 0) min = 0; + float iscale = nmax/(max - min); + float scale = 1/iscale; + for (int itry = 0; itry < ntry; ++itry) { + float sumlx = 0; int suml2 = 0; + bool did_change = false; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale*(x[i] - min)); + l = MAX(0, MIN(nmax, l)); + if (l != L[i]) { + L[i] = l; + did_change = true; + } + sumlx += (x[i] - min)*l; + suml2 += l*l; + } + scale = sumlx/suml2; + float sum = 0; + for (int i = 0; i < n; ++i) { + sum += x[i] - scale*L[i]; + } + min = alpha*min + (1 - alpha)*sum/n; + if (min > 0) min = 0; + iscale = 1/scale; + if (!did_change) break; + } + *the_min = -min; + return scale; +} + +static float make_qkx2_quants(int n, int nmax, const float * restrict x, const float * restrict weights, + uint8_t * restrict L, float * restrict the_min, uint8_t * restrict Laux, + float rmin, float rdelta, int nstep, bool use_mad) { + float min = x[0]; + float max = x[0]; + float sum_w = weights[0]; + float sum_x = sum_w * x[0]; +#ifdef HAVE_BUGGY_APPLE_LINKER + // use 'volatile' to prevent unroll and work around a bug in Apple ld64 1015.7 + for (volatile int i = 1; i < n; ++i) { +#else + for (int i = 1; i < n; ++i) { +#endif + if (x[i] < min) min = x[i]; + if (x[i] > max) max = x[i]; + float w = weights[i]; + sum_w += w; + sum_x += w * x[i]; + } + if (min > 0) min = 0; + if (max == min) { + for (int i = 0; i < n; ++i) L[i] = 0; + *the_min = -min; + return 0.f; + } + float iscale = nmax/(max - min); + float scale = 1/iscale; + float best_mad = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale*(x[i] - min)); + L[i] = MAX(0, MIN(nmax, l)); + float diff = scale * L[i] + min - x[i]; + diff = use_mad ? fabsf(diff) : diff * diff; + float w = weights[i]; + best_mad += w * diff; + } + if (nstep < 1) { + *the_min = -min; + return scale; + } + for (int is = 0; is <= nstep; ++is) { + iscale = (rmin + rdelta*is + nmax)/(max - min); + float sum_l = 0, sum_l2 = 0, sum_xl = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale*(x[i] - min)); + l = MAX(0, MIN(nmax, l)); + Laux[i] = l; + float w = weights[i]; + sum_l += w*l; + sum_l2 += w*l*l; + sum_xl += w*l*x[i]; + } + float D = sum_w * sum_l2 - sum_l * sum_l; + if (D > 0) { + float this_scale = (sum_w * sum_xl - sum_x * sum_l)/D; + float this_min = (sum_l2 * sum_x - sum_l * sum_xl)/D; + if (this_min > 0) { + this_min = 0; + this_scale = sum_xl / sum_l2; + } + float mad = 0; + for (int i = 0; i < n; ++i) { + float diff = this_scale * Laux[i] + this_min - x[i]; + diff = use_mad ? fabsf(diff) : diff * diff; + float w = weights[i]; + mad += w * diff; + } + if (mad < best_mad) { + for (int i = 0; i < n; ++i) { + L[i] = Laux[i]; + } + best_mad = mad; + scale = this_scale; + min = this_min; + } + } + } + *the_min = -min; + return scale; +} + +static inline void get_scale_min_k4(int j, const uint8_t * restrict q, uint8_t * restrict d, uint8_t * restrict m) { + if (j < 4) { + *d = q[j] & 63; *m = q[j + 4] & 63; + } else { + *d = (q[j+4] & 0xF) | ((q[j-4] >> 6) << 4); + *m = (q[j+4] >> 4) | ((q[j-0] >> 6) << 4); + } +} + +//========================- 2-bit (de)-quantization + +void quantize_row_q2_K(const float * restrict x, void * restrict vy, int64_t k) { + quantize_row_q2_K_ref(x, vy, k); +} + +//========================= 3-bit (de)-quantization + +void quantize_row_q3_K(const float * restrict x, void * restrict vy, int64_t k) { + quantize_row_q3_K_ref(x, vy, k); +} + +// ====================== 4-bit (de)-quantization + +void quantize_row_q4_K(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_q4_K * restrict y = vy; + quantize_row_q4_K_ref(x, y, k); +} + +// ====================== 5-bit (de)-quantization + +void quantize_row_q5_K(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_q5_K * restrict y = vy; + quantize_row_q5_K_ref(x, y, k); +} + +// ====================== 6-bit (de)-quantization + +void quantize_row_q6_K(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_q6_K * restrict y = vy; + quantize_row_q6_K_ref(x, y, k); +} + +// ====================== Ternary (de)-quantization (BitNet b1.58 and TriLMs) + +void quantize_row_tq1_0(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_tq1_0 * restrict y = vy; + quantize_row_tq1_0_ref(x, y, k); +} + +void quantize_row_tq2_0(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_tq2_0 * restrict y = vy; + quantize_row_tq2_0_ref(x, y, k); +} + +static const int8_t kvalues_iq4nl[16] = {-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113}; + +//===================================== Q8_K ============================================== + +void quantize_row_q8_K(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q8_K_ref(x, y, k); +} + +//===================================== Dot products ================================= + +// +// Helper functions +// +#if __AVX__ || __AVX2__ || __AVX512F__ + +// shuffles to pick the required scales in dot products +static inline __m256i get_scale_shuffle_q3k(int i) { + static const uint8_t k_shuffle[128] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, + }; + return _mm256_loadu_si256((const __m256i*)k_shuffle + i); +} +static inline __m256i get_scale_shuffle_k4(int i) { + static const uint8_t k_shuffle[256] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, + 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, + 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, + 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, + 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 + }; + return _mm256_loadu_si256((const __m256i*)k_shuffle + i); +} +static inline __m128i get_scale_shuffle(int i) { + static const uint8_t k_shuffle[128] = { + 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, + 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, + 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, + 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, + 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, + 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, + 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, + 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 + }; + return _mm_loadu_si128((const __m128i*)k_shuffle + i); +} +#elif defined(__loongarch_asx) +// shuffles to pick the required scales in dot products +static inline __m256i get_scale_shuffle_q3k(int i) { + static const uint8_t k_shuffle[128] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, + }; + return __lasx_xvld((const __m256i*)k_shuffle + i, 0); +} +static inline __m256i get_scale_shuffle_k4(int i) { + static const uint8_t k_shuffle[256] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, + 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, + 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, + 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, + 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 + }; + return __lasx_xvld((const __m256i*)k_shuffle + i, 0); +} +static inline __m128i get_scale_shuffle(int i) { + static const uint8_t k_shuffle[128] = { + 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, + 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, + 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, + 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, + 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, + 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, + 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, + 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 + }; + return __lsx_vld((const __m128i*)k_shuffle + i, 0); +} +#endif + +void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; + const int nb = n / qk; + + assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q4_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; + +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q4_0 * restrict vx0 = vx; + const block_q4_0 * restrict vx1 = (const block_q4_0 *) ((const uint8_t*)vx + bx); + const block_q8_0 * restrict vy0 = vy; + const block_q8_0 * restrict vy1 = (const block_q8_0 *) ((const uint8_t*)vy + by); + + float32x4_t sumv0 = vdupq_n_f32(0.0f); + + for (int i = 0; i < nb; i++) { + const block_q4_0 * restrict b_x0 = &vx0[i]; + const block_q4_0 * restrict b_x1 = &vx1[i]; + const block_q8_0 * restrict b_y0 = &vy0[i]; + const block_q8_0 * restrict b_y1 = &vy1[i]; + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + const int8x16_t s8b = vdupq_n_s8(0x8); + + const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); + const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // sub 8 + const int8x16_t x0_l = vsubq_s8(v0_0l, s8b); + const int8x16_t x0_h = vsubq_s8(v0_0h, s8b); + const int8x16_t x1_l = vsubq_s8(v0_1l, s8b); + const int8x16_t x1_h = vsubq_s8(v0_1h, s8b); + + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); + + float32_t _scale[4] = { + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d) + }; + float32x4_t scale = vld1q_f32(_scale); + + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + + float32x4_t sumv1 = vextq_f32 (sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); + + vst1_f32(s, vget_low_f32 (sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + + return; + } +#endif + + int ib = 0; + float sumf = 0; + +#if defined(__ARM_FEATURE_SVE) + svfloat32_t sumv0 = svdup_n_f32(0.0f); + svfloat32_t sumv1 = svdup_n_f32(0.0f); + + const int vector_length = ggml_cpu_get_sve_cnt()*8; + + // VLA Implementation using switch case + switch (vector_length) { + case 128: + { + // predicate for activating higher lanes for 4 float32 elements + const svbool_t ph4 = svptrue_pat_b32(SV_VL4); + + for (; ib + 1 < nb; ib += 2) { + const block_q4_0 * restrict x0 = &x[ib + 0]; + const block_q4_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + // load x + const svuint8_t qx0r = svld1rq_u8(svptrue_b8(), x0->qs); + const svuint8_t qx1r = svld1rq_u8(svptrue_b8(), x1->qs); + + // 4-bit -> 8-bit + const svint8_t qx0l = svreinterpret_s8_u8(svand_n_u8_m(svptrue_b8(), qx0r, 0x0F)); + const svint8_t qx0h = svreinterpret_s8_u8(svlsr_n_u8_m(svptrue_b8(), qx0r, 0x04)); + const svint8_t qx1l = svreinterpret_s8_u8(svand_n_u8_m(svptrue_b8(), qx1r, 0x0F)); + const svint8_t qx1h = svreinterpret_s8_u8(svlsr_n_u8_m(svptrue_b8(), qx1r, 0x04)); + + // sub 8 + const svint8_t qx0ls = svsub_n_s8_x(svptrue_b8(), qx0h, 8); + const svint8_t qx0hs = svsub_n_s8_x(svptrue_b8(), qx0l, 8); + const svint8_t qx1ls = svsub_n_s8_x(svptrue_b8(), qx1h, 8); + const svint8_t qx1hs = svsub_n_s8_x(svptrue_b8(), qx1l, 8); + + // load y + const svint8_t qy0h = svld1_s8(svptrue_b8(), y0->qs); + const svint8_t qy0l = svld1_s8(svptrue_b8(), y0->qs + 16); + const svint8_t qy1h = svld1_s8(svptrue_b8(), y1->qs); + const svint8_t qy1l = svld1_s8(svptrue_b8(), y1->qs + 16); + + // dot product + sumv0 = svmla_n_f32_x(ph4, sumv0, svcvt_f32_s32_x(ph4, svadd_x(ph4, + svdot_s32(svdup_n_s32(0), qx0ls, qy0l), + svdot_s32(svdup_n_s32(0), qx0hs, qy0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = svmla_n_f32_x(ph4, sumv1, svcvt_f32_s32_x(ph4, svadd_x(ph4, + svdot_s32(svdup_n_s32(0), qx1ls, qy1l), + svdot_s32(svdup_n_s32(0), qx1hs, qy1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); + } break; + case 256: + { + // predicate for activating higher lanes for 16 int8 elements + const svbool_t ph16 = svptrue_pat_b8(SV_VL16); + // predicate for activating lower lanes for 16 int8 elements + const svbool_t pl16 = svnot_b_z(svptrue_b8(), ph16); + + for (; ib + 1 < nb; ib += 2) { + const block_q4_0 * restrict x0 = &x[ib + 0]; + const block_q4_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + // load x + const svuint8_t qx0r = svld1rq_u8(svptrue_b8(), x0->qs); + const svuint8_t qx1r = svld1rq_u8(svptrue_b8(), x1->qs); + + // 4-bit -> 8-bit + const svint8_t qx0 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx0r, 0x0F), 0x04)); + const svint8_t qx1 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx1r, 0x0F), 0x04)); + + // sub 8 + const svint8_t qx0s = svsub_n_s8_x(svptrue_b8(), qx0, 8); + const svint8_t qx1s = svsub_n_s8_x(svptrue_b8(), qx1, 8); + + // load y + const svint8_t qy0 = svld1_s8(svptrue_b8(), y0->qs); + const svint8_t qy1 = svld1_s8(svptrue_b8(), y1->qs); + + // dot product + sumv0 = svmla_n_f32_x(svptrue_b32(), sumv0, svcvt_f32_s32_x(svptrue_b32(), + svdot_s32(svdup_n_s32(0), qx0s, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = svmla_n_f32_x(svptrue_b32(), sumv1, svcvt_f32_s32_x(svptrue_b32(), + svdot_s32(svdup_n_s32(0), qx1s, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); + } break; + case 512: + { + // predicate for activating higher lanes for 32 int8 elements + const svbool_t ph32 = svptrue_pat_b8(SV_VL32); + + // predicate for activating higher lanes for 16 int8 elements + const svbool_t ph16 = svptrue_pat_b8(SV_VL16); + // predicate for activating lower lanes for 16 int8 elements from first 32 int8 activated lanes + const svbool_t pl16 = svnot_b_z(ph32, ph16); + + for (; ib + 1 < nb; ib += 2) { + const block_q4_0 * restrict x0 = &x[ib + 0]; + const block_q4_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + // load x + const svuint8_t qx0r = svld1rq_u8(ph32, x0->qs); + const svuint8_t qx1r = svld1rq_u8(ph32, x1->qs); + + // 4-bit -> 8-bit + const svint8_t qx0 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx0r, 0x0F), 0x04)); + const svint8_t qx1 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx1r, 0x0F), 0x04)); + + // sub 8 + const svint8_t qx0s = svsub_n_s8_x(ph32, qx0, 8); + const svint8_t qx1s = svsub_n_s8_x(ph32, qx1, 8); + + // load y + const svint8_t qy0 = svld1_s8(ph32, y0->qs); + const svint8_t qy1 = svld1_s8(ph32, y1->qs); + + // dot product + sumv0 = svmla_n_f32_x(ph32, sumv0, svcvt_f32_s32_x(ph32, + svdot_s32(svdup_n_s32(0), qx0s, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = svmla_n_f32_x(ph32, sumv1, svcvt_f32_s32_x(ph32, + svdot_s32(svdup_n_s32(0), qx1s, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = svaddv_f32(ph32, svadd_f32_x(ph32, sumv0, sumv1)); + } break; + default: + assert(false && "Unsupported vector length"); + break; + } + +#elif defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + for (; ib + 1 < nb; ib += 2) { + const block_q4_0 * restrict x0 = &x[ib + 0]; + const block_q4_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + const int8x16_t s8b = vdupq_n_s8(0x8); + + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // sub 8 + const int8x16_t v0_0ls = vsubq_s8(v0_0l, s8b); + const int8x16_t v0_0hs = vsubq_s8(v0_0h, s8b); + const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b); + const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b); + + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + + // dot product into int32x4_t + const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0l), v0_0hs, v1_0h); + const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1l), v0_1hs, v1_1h); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); +#elif defined(__AVX2__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + // Main loop + for (; ib < nb; ++ib) { + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + + // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. + const __m256i off = _mm256_set1_epi8( 8 ); + qx = _mm256_sub_epi8( qx, off ); + + __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + /* Multiply q with scale and accumulate */ + acc = _mm256_fmadd_ps( d, q, acc ); + } + + sumf = hsum_float_8(acc); +#elif defined(__AVX__) + __m256 accum = _mm256_setzero_ps(); + for (; ib + 1 < nb; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)x[ib + 0].qs); + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); + const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs); + const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs + 1); + const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); + const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs + 1); + + const __m128i q4b_1_0 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), q4bits_1), _mm_set1_epi8(8)); + const __m128i q4b_1_1 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(q4bits_1, 4)), _mm_set1_epi8(8)); + const __m128i q4b_2_0 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), q4bits_2), _mm_set1_epi8(8)); + const __m128i q4b_2_1 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(q4bits_2, 4)), _mm_set1_epi8(8)); + + const __m128i p16_1_0 = mul_add_epi8_sse(q4b_1_0, q8b_1_0); + const __m128i p16_1_1 = mul_add_epi8_sse(q4b_1_1, q8b_1_1); + const __m128i p16_2_0 = mul_add_epi8_sse(q4b_2_0, q8b_2_0); + const __m128i p16_2_1 = mul_add_epi8_sse(q4b_2_1, q8b_2_1); + const __m128i p_1 = _mm_add_epi16(p16_1_0, p16_1_1); + const __m128i p_2 = _mm_add_epi16(p16_2_0, p16_2_1); + const __m256 p = sum_i16_pairs_float(p_2, p_1); + + const __m256 deltas = quad_fp16_delta_float(x[ib].d, y[ib].d, x[ib + 1].d, y[ib + 1].d); + accum = _mm256_add_ps(_mm256_mul_ps(deltas, p), accum); + } + + sumf = hsum_float_8(accum); +#elif defined(__SSSE3__) + // set constants + const __m128i lowMask = _mm_set1_epi8(0xF); + const __m128i off = _mm_set1_epi8(8); + + // Initialize accumulator with zeros + __m128 acc_0 = _mm_setzero_ps(); + __m128 acc_1 = _mm_setzero_ps(); + __m128 acc_2 = _mm_setzero_ps(); + __m128 acc_3 = _mm_setzero_ps(); + + for (; ib + 1 < nb; ib += 2) { + _mm_prefetch(&x[ib] + sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[ib] + sizeof(block_q8_0), _MM_HINT_T0); + + // Compute combined scale for the block 0 and 1 + const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); + + const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[ib].qs); + + __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); + __m128i by_0 = _mm_loadu_si128((const __m128i *)y[ib].qs); + bx_0 = _mm_sub_epi8(bx_0, off); + const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); + + __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); + __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[ib].qs + 16)); + bx_1 = _mm_sub_epi8(bx_1, off); + const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); + + _mm_prefetch(&x[ib] + 2 * sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[ib] + 2 * sizeof(block_q8_0), _MM_HINT_T0); + + // Compute combined scale for the block 2 and 3 + const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[ib + 1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) ); + + const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); + + __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); + __m128i by_2 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); + bx_2 = _mm_sub_epi8(bx_2, off); + const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); + + __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); + __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[ib + 1].qs + 16)); + bx_3 = _mm_sub_epi8(bx_3, off); + const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); + + // Convert int32_t to float + __m128 p0 = _mm_cvtepi32_ps(i32_0); + __m128 p1 = _mm_cvtepi32_ps(i32_1); + __m128 p2 = _mm_cvtepi32_ps(i32_2); + __m128 p3 = _mm_cvtepi32_ps(i32_3); + + // Apply the scale + __m128 p0_d = _mm_mul_ps( d_0_1, p0 ); + __m128 p1_d = _mm_mul_ps( d_0_1, p1 ); + __m128 p2_d = _mm_mul_ps( d_2_3, p2 ); + __m128 p3_d = _mm_mul_ps( d_2_3, p3 ); + + // Acummulate + acc_0 = _mm_add_ps(p0_d, acc_0); + acc_1 = _mm_add_ps(p1_d, acc_1); + acc_2 = _mm_add_ps(p2_d, acc_2); + acc_3 = _mm_add_ps(p3_d, acc_3); + } + + sumf = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); +#elif defined(__riscv_v_intrinsic) + size_t vl = __riscv_vsetvl_e8m1(qk/2); + + for (; ib < nb; ++ib) { + // load elements + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); + + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); + + // mask and store lower part of x, and then upper part + vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + + vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + + // subtract offset + vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 8, vl); + vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 8, vl); + + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); + + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); + + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + + sumf += sumi*GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d); + } + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed int v0 = vec_splats((int32_t)0); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + const vector signed char v8 = vec_splats((signed char)0x8); + + vector float vsumf0 = vec_splats(0.0f); + +#pragma GCC unroll 8 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl(16, y[ib].qs); + + vector signed char q4x0 = vec_and(qxs, lowMask); + vector signed char q4x1 = vec_sr(qxs, v4); + + q4x0 = vec_sub(q4x0, v8); + q4x1 = vec_sub(q4x1, v8); + + vector signed short qv0 = vec_add(vec_mule(q4x0, q8y0), vec_mulo(q4x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q4x1, q8y1), vec_mulo(q4x1, q8y1)); + + vector signed int vsumi0 = v0; + + vsumi0 = vec_sum4s(qv0, vsumi0); + vsumi0 = vec_sum4s(qv1, vsumi0); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + } + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + // Initialize accumulator with zeros + __m256 acc = (__m256)__lasx_xvldi(0); + + // Main loop + for (; ib < nb; ++ib) { + /* Compute combined scale for the block */ + const __m256 d = __lasx_xvreplfr2vr_s( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + + // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. + const __m256i off = __lasx_xvreplgr2vr_b( 8 ); + qx = __lasx_xvsub_b( qx, off ); + + __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + /* Multiply q with scale and accumulate */ + acc = __lasx_xvfmadd_s( d, q, acc ); + } + + sumf = hsum_float_8(acc); +#elif defined(__loongarch_sx) + // set constants + const __m128i low_mask = __lsx_vreplgr2vr_b(0xF); + const __m128i off = __lsx_vreplgr2vr_b(8); + + // Initialize accumulator with zeros + __m128 acc_0 = __lsx_vldi(0); + __m128 acc_1 = __lsx_vldi(0); + __m128 acc_2 = __lsx_vldi(0); + __m128 acc_3 = __lsx_vldi(0); + + for (; ib + 1 < nb; ib += 2) { + + // Compute combined scale for the block 0 and 1 + const __m128 d_0_1 = __lsx_vreplgr2vr_w( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); + + const __m128i tmp_0_1 = __lsx_vld((const __m128i *)x[ib].qs, 0); + + __m128i bx_0 = __lsx_vand_v(low_mask, tmp_0_1); + __m128i by_0 = __lsx_vld((const __m128i *)y[ib].qs, 0); + bx_0 = __lsx_vsub_b(bx_0, off); + const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); + + __m128i bx_1 = __lsx_vand_v(low_mask, __lsx_vsrli_d(tmp_0_1, 4)); + __m128i by_1 = __lsx_vld((const __m128i *)(y[ib].qs + 16), 0); + bx_1 = __lsx_vsub_b(bx_1, off); + const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); + + //_mm_prefetch(&x[ib] + 2 * sizeof(block_q4_0), _MM_HINT_T0); + //_mm_prefetch(&y[ib] + 2 * sizeof(block_q8_0), _MM_HINT_T0); + + // Compute combined scale for the block 2 and 3 + const __m128 d_2_3 = __lsx_vreplgr2vr_w( GGML_FP16_TO_FP32(x[ib + 1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) ); + + const __m128i tmp_2_3 = __lsx_vld((const __m128i *)x[ib + 1].qs, 0); + + __m128i bx_2 = __lsx_vand_v(low_mask, tmp_2_3); + __m128i by_2 = __lsx_vld((const __m128i *)y[ib + 1].qs, 0); + bx_2 = __lsx_vsub_b(bx_2, off); + const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); + + __m128i bx_3 = __lsx_vand_v(low_mask, __lsx_vsrli_d(tmp_2_3, 4)); + __m128i by_3 = __lsx_vld((const __m128i *)(y[ib + 1].qs + 16), 0); + bx_3 = __lsx_vsub_b(bx_3, off); + const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); + + // Convert int32_t to float + __m128 p0 = __lsx_vffint_s_w(i32_0); + __m128 p1 = __lsx_vffint_s_w(i32_1); + __m128 p2 = __lsx_vffint_s_w(i32_2); + __m128 p3 = __lsx_vffint_s_w(i32_3); + + // Apply the scale + __m128 p0_d = __lsx_vfmul_s( d_0_1, p0 ); + __m128 p1_d = __lsx_vfmul_s( d_0_1, p1 ); + __m128 p2_d = __lsx_vfmul_s( d_2_3, p2 ); + __m128 p3_d = __lsx_vfmul_s( d_2_3, p3 ); + + // Acummulate + acc_0 = __lsx_vfadd_s(p0_d, acc_0); + acc_1 = __lsx_vfadd_s(p1_d, acc_1); + acc_2 = __lsx_vfadd_s(p2_d, acc_2); + acc_3 = __lsx_vfadd_s(p3_d, acc_3); + } + + sumf = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); +#endif + for (; ib < nb; ++ib) { + int sumi0 = 0; + int sumi1 = 0; + + for (int j = 0; j < qk/2; ++j) { + const int v0 = (x[ib].qs[j] & 0x0F) - 8; + const int v1 = (x[ib].qs[j] >> 4) - 8; + + sumi0 += (v0 * y[ib].qs[j]); + sumi1 += (v1 * y[ib].qs[j + qk/2]); + } + + int sumi = sumi0 + sumi1; + sumf += sumi*GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d); + } + + *s = sumf; +} + +void ggml_vec_dot_q4_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_1; + const int nb = n / qk; + + assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q4_1 * restrict x = vx; + const block_q8_1 * restrict y = vy; + +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q4_1 * restrict vx0 = vx; + const block_q4_1 * restrict vx1 = (const block_q4_1 *) ((const uint8_t*)vx + bx); + const block_q8_1 * restrict vy0 = vy; + const block_q8_1 * restrict vy1 = (const block_q8_1 *) ((const uint8_t*)vy + by); + + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t summs0 = vdupq_n_f32(0.0f); + + for (int i = 0; i < nb; i++) { + const block_q4_1 * restrict b_x0 = &vx0[i]; + const block_q4_1 * restrict b_x1 = &vx1[i]; + const block_q8_1 * restrict b_y0 = &vy0[i]; + const block_q8_1 * restrict b_y1 = &vy1[i]; + + float32_t summs_t[4] = { + GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y0->s), + GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y0->s), + GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y1->s), + GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y1->s) + }; + summs0 = vaddq_f32(summs0, vld1q_f32(summs_t)); + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + + const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); + const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); + + // 4-bit -> 8-bit + const int8x16_t x0_l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t x0_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t x1_l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t x1_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); + + // mmla into int32x4_t + float32_t _scale[4] = { + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d) + }; + float32x4_t scale = vld1q_f32(_scale); + + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + + float32x4_t sumv1 = vextq_f32 (sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); + + sumv2 = vaddq_f32(sumv2, summs0); + + vst1_f32(s, vget_low_f32 (sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + + return; + } +#endif + + int ib = 0; + float sumf = 0; + + // TODO: add WASM SIMD +#if defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + float summs = 0; + + for (; ib + 1 < nb; ib += 2) { + const block_q4_1 * restrict x0 = &x[ib + 0]; + const block_q4_1 * restrict x1 = &x[ib + 1]; + const block_q8_1 * restrict y0 = &y[ib + 0]; + const block_q8_1 * restrict y1 = &y[ib + 1]; + + summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s) + GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + + // dot product into int32x4_t + const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0l, v1_0l), v0_0h, v1_0h); + const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1l, v1_1l), v0_1h, v1_1h); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs; +#elif defined(__AVX2__) || defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + float summs = 0; + + // Main loop + for (; ib < nb; ++ib) { + const float d0 = GGML_FP16_TO_FP32(x[ib].d); + const float d1 = GGML_FP16_TO_FP32(y[ib].d); + + summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); + + const __m256 d0v = _mm256_set1_ps( d0 ); + const __m256 d1v = _mm256_set1_ps( d1 ); + + // Compute combined scales + const __m256 d0d1 = _mm256_mul_ps( d0v, d1v ); + + // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes + const __m256i qx = bytes_from_nibbles_32(x[ib].qs); + const __m256i qy = _mm256_loadu_si256( (const __m256i *)y[ib].qs ); + + const __m256 xy = mul_sum_us8_pairs_float(qx, qy); + + // Accumulate d0*d1*x*y +#if defined(__AVX2__) + acc = _mm256_fmadd_ps( d0d1, xy, acc ); +#else + acc = _mm256_add_ps( _mm256_mul_ps( d0d1, xy ), acc ); +#endif + } + + sumf = hsum_float_8(acc) + summs; +#elif defined(__riscv_v_intrinsic) + size_t vl = __riscv_vsetvl_e8m1(qk/2); + + for (; ib < nb; ++ib) { + // load elements + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); + + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); + + // mask and store lower part of x, and then upper part + vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + + vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); + + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); + + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); + } + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed int v0 = vec_splats((int32_t)0); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + +#pragma GCC unroll 4 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[ib].m)); + vector float vys = {GGML_FP16_TO_FP32(y[ib].s), 0.0f, 0.0f, 0.0f}; + vsumf0 = vec_madd(vxmin, vys, vsumf0); + + vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl(16, y[ib].qs); + + vector unsigned char q4x0 = (vector unsigned char)vec_and(qxs, lowMask); + vector unsigned char q4x1 = (vector unsigned char)vec_sr(qxs, v4); + + vector signed int vsumi0 = v0; + + vsumi0 = vec_msum(q8y0, q4x0, vsumi0); + vsumi0 = vec_msum(q8y1, q4x1, vsumi0); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + } + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + // Initialize accumulator with zeros + __m256 acc = (__m256)__lasx_xvldi(0); + + float summs = 0; + + // Main loop + for (; ib < nb; ++ib) { + const float d0 = GGML_FP16_TO_FP32(x[ib].d); + const float d1 = GGML_FP16_TO_FP32(y[ib].d); + + summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); + + const __m256 d0v = __lasx_xvreplfr2vr_s( d0 ); + const __m256 d1v = __lasx_xvreplfr2vr_s( d1 ); + + // Compute combined scales + const __m256 d0d1 = __lasx_xvfmul_s( d0v, d1v ); + + // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes + const __m256i qx = bytes_from_nibbles_32(x[ib].qs); + const __m256i qy = __lasx_xvld( (const __m256i *)y[ib].qs, 0); + + const __m256 xy = mul_sum_us8_pairs_float(qx, qy); + + // Accumulate d0*d1*x*y + acc = __lasx_xvfmadd_s( d0d1, xy, acc ); + } + + sumf = hsum_float_8(acc) + summs; +#endif + for (; ib < nb; ++ib) { + int sumi0 = 0; + int sumi1 = 0; + + for (int j = 0; j < qk/2; ++j) { + const int v0 = (x[ib].qs[j] & 0x0F); + const int v1 = (x[ib].qs[j] >> 4); + + sumi0 += (v0 * y[ib].qs[j]); + sumi1 += (v1 * y[ib].qs[j + qk/2]); + } + + int sumi = sumi0 + sumi1; + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); + } + + *s = sumf; +} + +void ggml_vec_dot_q5_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; + const int nb = n / qk; + + int ib = 0; + float sumf = 0; + + assert(n % qk == 0); + assert(qk == QK5_0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q5_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; + +#if defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + uint32_t qh0; + uint32_t qh1; + + uint64_t tmp0[4]; + uint64_t tmp1[4]; + + for (; ib + 1 < nb; ib += 2) { + const block_q5_0 * restrict x0 = &x[ib]; + const block_q5_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + + // extract the 5th bit via lookup table ((!b) << 4) + memcpy(&qh0, x0->qh, sizeof(qh0)); + memcpy(&qh1, x1->qh, sizeof(qh1)); + + tmp0[0] = table_b2b_1[(qh0 >> 0) & 0xFF]; + tmp0[1] = table_b2b_1[(qh0 >> 8) & 0xFF]; + tmp0[2] = table_b2b_1[(qh0 >> 16) & 0xFF]; + tmp0[3] = table_b2b_1[(qh0 >> 24) ]; + + tmp1[0] = table_b2b_1[(qh1 >> 0) & 0xFF]; + tmp1[1] = table_b2b_1[(qh1 >> 8) & 0xFF]; + tmp1[2] = table_b2b_1[(qh1 >> 16) & 0xFF]; + tmp1[3] = table_b2b_1[(qh1 >> 24) ]; + + const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); + const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); + const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); + const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); + + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); + + // 4-bit -> 8-bit + int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) + const int8x16_t v0_0lf = vsubq_s8(v0_0l, qhl0); + const int8x16_t v0_0hf = vsubq_s8(v0_0h, qhh0); + const int8x16_t v0_1lf = vsubq_s8(v0_1l, qhl1); + const int8x16_t v0_1hf = vsubq_s8(v0_1h, qhh1); + + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); +#elif defined(__wasm_simd128__) + v128_t sumv = wasm_f32x4_splat(0.0f); + + uint32_t qh; + uint64_t tmp[4]; + + // TODO: check if unrolling this is better + for (; ib < nb; ++ib) { + const block_q5_0 * restrict x0 = &x[ib]; + const block_q8_0 * restrict y0 = &y[ib]; + + const v128_t m4b = wasm_i8x16_splat(0x0F); + + // extract the 5th bit + memcpy(&qh, x0->qh, sizeof(qh)); + + tmp[0] = table_b2b_1[(qh >> 0) & 0xFF]; + tmp[1] = table_b2b_1[(qh >> 8) & 0xFF]; + tmp[2] = table_b2b_1[(qh >> 16) & 0xFF]; + tmp[3] = table_b2b_1[(qh >> 24) ]; + + const v128_t qhl = wasm_v128_load(tmp + 0); + const v128_t qhh = wasm_v128_load(tmp + 2); + + const v128_t v0 = wasm_v128_load(x0->qs); + + // 4-bit -> 8-bit + const v128_t v0l = wasm_v128_and (v0, m4b); + const v128_t v0h = wasm_u8x16_shr(v0, 4); + + // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) + const v128_t v0lf = wasm_i8x16_sub(v0l, qhl); + const v128_t v0hf = wasm_i8x16_sub(v0h, qhh); + + // load y + const v128_t v1l = wasm_v128_load(y0->qs); + const v128_t v1h = wasm_v128_load(y0->qs + 16); + + // int8x16 -> int16x8 + const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); + const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); + const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); + const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); + + const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); + const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); + const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); + const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); + + // dot product + sumv = wasm_f32x4_add(sumv, wasm_f32x4_mul(wasm_f32x4_convert_i32x4( + wasm_i32x4_add( + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), + wasm_i32x4_dot_i16x8(v0lfh, v1lh)), + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), + wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), + wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); + } + + sumf = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + + wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3); +#elif defined(__AVX2__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + // Main loop + for (; ib < nb; ++ib) { + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + __m256i bxhi = bytes_from_bits_32(x[ib].qh); + bxhi = _mm256_andnot_si256(bxhi, _mm256_set1_epi8((char)0xF0)); + qx = _mm256_or_si256(qx, bxhi); + + __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + /* Multiply q with scale and accumulate */ + acc = _mm256_fmadd_ps(d, q, acc); + } + + sumf = hsum_float_8(acc); +#elif defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + __m128i mask = _mm_set1_epi8((char)0xF0); + + // Main loop + for (; ib < nb; ++ib) { + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); + + __m256i bx_0 = bytes_from_nibbles_32(x[ib].qs); + const __m256i bxhi = bytes_from_bits_32(x[ib].qh); + __m128i bxhil = _mm256_castsi256_si128(bxhi); + __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); + bxhil = _mm_andnot_si128(bxhil, mask); + bxhih = _mm_andnot_si128(bxhih, mask); + __m128i bxl = _mm256_castsi256_si128(bx_0); + __m128i bxh = _mm256_extractf128_si256(bx_0, 1); + bxl = _mm_or_si128(bxl, bxhil); + bxh = _mm_or_si128(bxh, bxhih); + bx_0 = MM256_SET_M128I(bxh, bxl); + + const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_i8_pairs_float(bx_0, by_0); + + /* Multiply q with scale and accumulate */ + acc = _mm256_add_ps(_mm256_mul_ps(d, q), acc); + } + + sumf = hsum_float_8(acc); +#elif defined(__riscv_v_intrinsic) + uint32_t qh; + + size_t vl = __riscv_vsetvl_e8m1(qk/2); + + // These temporary registers are for masking and shift operations + vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); + vuint32m2_t vt_2 = __riscv_vsll_vv_u32m2(__riscv_vmv_v_x_u32m2(1, vl), vt_1, vl); + + vuint32m2_t vt_3 = __riscv_vsll_vx_u32m2(vt_2, 16, vl); + vuint32m2_t vt_4 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); + + for (; ib < nb; ++ib) { + memcpy(&qh, x[ib].qh, sizeof(uint32_t)); + + // ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; + vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(vt_2, qh, vl); + vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(xha_0, vt_1, vl); + vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); + + // ((qh & (1u << (j + 16))) >> (j + 12)); + vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(vt_3, qh, vl); + vuint32m2_t xhl_1 = __riscv_vsrl_vv_u32m2(xha_1, vt_4, vl); + + // narrowing + vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xhl_0, vl); + vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); + + vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xhl_1, vl); + vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); + + // load + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); + + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); + + vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + + vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); + vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); + + vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + + vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 16, vl); + vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 16, vl); + + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); + + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); + + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)) * sumi; + } + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector unsigned char v4 = vec_splats((unsigned char)4); + + vector float vsumf0 = vec_splats(0.0f); + +#pragma GCC unroll 4 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector signed long long aux64x2_0 = {(uint64_t)(table_b2b_1[x[ib].qh[0]]), (uint64_t)(table_b2b_1[x[ib].qh[1]])}; + vector signed long long aux64x2_1 = {(uint64_t)(table_b2b_1[x[ib].qh[2]]), (uint64_t)(table_b2b_1[x[ib].qh[3]])}; + + vector signed char qh0 = (vector signed char)aux64x2_0; + vector signed char qh1 = (vector signed char)aux64x2_1; + + vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); + + vector signed char q5x0 = vec_sub(vec_and (qxs, lowMask), qh0); + vector signed char q5x1 = vec_sub(vec_sr(qxs, v4), qh1); + + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl( 16, y[ib].qs); + + vector signed short qv0 = vec_add(vec_mule(q5x0, q8y0), vec_mulo(q5x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q5x1, q8y1), vec_mulo(q5x1, q8y1)); + + qv0 = vec_add(qv0, qv1); + + vector signed int vsumi0 = vec_add(vec_unpackh(qv0), vec_unpackl(qv0)); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + } + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + // Initialize accumulator with zeros + __m256 acc = (__m256)__lasx_xvldi(0); + + // Main loop + for (; ib < nb; ++ib) { + /* Compute combined scale for the block */ + const __m256 d = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); //FIXME + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + __m256i bxhi = bytes_from_bits_32(x[ib].qh); + bxhi = __lasx_xvandn_v(bxhi, __lasx_xvreplgr2vr_b((char)0xF0)); + qx = __lasx_xvor_v(qx, bxhi); + + __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + /* Multiply q with scale and accumulate */ + acc = __lasx_xvfmadd_s(d, q, acc); + } + + sumf = hsum_float_8(acc); +#endif + for (; ib < nb; ++ib) { + uint32_t qh; + memcpy(&qh, x[ib].qh, sizeof(qh)); + + int sumi0 = 0; + int sumi1 = 0; + + for (int j = 0; j < qk/2; ++j) { + const uint8_t xh_0 = ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; + const uint8_t xh_1 = ((qh & (1u << (j + 16))) >> (j + 12)); + + const int32_t x0 = (int8_t)(((x[ib].qs[j] & 0x0F) | xh_0) - 16); + const int32_t x1 = (int8_t)(((x[ib].qs[j] >> 4) | xh_1) - 16); + + sumi0 += (x0 * y[ib].qs[j]); + sumi1 += (x1 * y[ib].qs[j + qk/2]); + } + + int sumi = sumi0 + sumi1; + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)) * sumi; + } + + *s = sumf; +} + +void ggml_vec_dot_q5_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_1; + const int nb = n / qk; + + int ib = 0; + float sumf = 0; + + assert(n % qk == 0); + assert(qk == QK5_1); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q5_1 * restrict x = vx; + const block_q8_1 * restrict y = vy; + +#if defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + float summs0 = 0.0f; + float summs1 = 0.0f; + + uint32_t qh0; + uint32_t qh1; + + uint64_t tmp0[4]; + uint64_t tmp1[4]; + + for (; ib + 1 < nb; ib += 2) { + const block_q5_1 * restrict x0 = &x[ib]; + const block_q5_1 * restrict x1 = &x[ib + 1]; + const block_q8_1 * restrict y0 = &y[ib]; + const block_q8_1 * restrict y1 = &y[ib + 1]; + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + + summs0 += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); + summs1 += GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); + + // extract the 5th bit via lookup table ((b) << 4) + memcpy(&qh0, x0->qh, sizeof(qh0)); + memcpy(&qh1, x1->qh, sizeof(qh1)); + + tmp0[0] = table_b2b_0[(qh0 >> 0) & 0xFF]; + tmp0[1] = table_b2b_0[(qh0 >> 8) & 0xFF]; + tmp0[2] = table_b2b_0[(qh0 >> 16) & 0xFF]; + tmp0[3] = table_b2b_0[(qh0 >> 24) ]; + + tmp1[0] = table_b2b_0[(qh1 >> 0) & 0xFF]; + tmp1[1] = table_b2b_0[(qh1 >> 8) & 0xFF]; + tmp1[2] = table_b2b_0[(qh1 >> 16) & 0xFF]; + tmp1[3] = table_b2b_0[(qh1 >> 24) ]; + + const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); + const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); + const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); + const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); + + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // add high bit + const int8x16_t v0_0lf = vorrq_s8(v0_0l, qhl0); + const int8x16_t v0_0hf = vorrq_s8(v0_0h, qhh0); + const int8x16_t v0_1lf = vorrq_s8(v0_1l, qhl1); + const int8x16_t v0_1hf = vorrq_s8(v0_1h, qhh1); + + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs0 + summs1; +#elif defined(__wasm_simd128__) + v128_t sumv = wasm_f32x4_splat(0.0f); + + float summs = 0.0f; + + uint32_t qh; + uint64_t tmp[4]; + + // TODO: check if unrolling this is better + for (; ib < nb; ++ib) { + const block_q5_1 * restrict x0 = &x[ib]; + const block_q8_1 * restrict y0 = &y[ib]; + + summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); + + const v128_t m4b = wasm_i8x16_splat(0x0F); + + // extract the 5th bit + memcpy(&qh, x0->qh, sizeof(qh)); + + tmp[0] = table_b2b_0[(qh >> 0) & 0xFF]; + tmp[1] = table_b2b_0[(qh >> 8) & 0xFF]; + tmp[2] = table_b2b_0[(qh >> 16) & 0xFF]; + tmp[3] = table_b2b_0[(qh >> 24) ]; + + const v128_t qhl = wasm_v128_load(tmp + 0); + const v128_t qhh = wasm_v128_load(tmp + 2); + + const v128_t v0 = wasm_v128_load(x0->qs); + + // 4-bit -> 8-bit + const v128_t v0l = wasm_v128_and (v0, m4b); + const v128_t v0h = wasm_u8x16_shr(v0, 4); + + // add high bit + const v128_t v0lf = wasm_v128_or(v0l, qhl); + const v128_t v0hf = wasm_v128_or(v0h, qhh); + + // load y + const v128_t v1l = wasm_v128_load(y0->qs); + const v128_t v1h = wasm_v128_load(y0->qs + 16); + + // int8x16 -> int16x8 + const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); + const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); + const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); + const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); + + const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); + const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); + const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); + const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); + + // dot product + sumv = wasm_f32x4_add(sumv, + wasm_f32x4_mul(wasm_f32x4_convert_i32x4(wasm_i32x4_add( + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), + wasm_i32x4_dot_i16x8(v0lfh, v1lh)), + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), + wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), + wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); + } + + sumf = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + + wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3) + summs; +#elif defined(__AVX2__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + float summs = 0.0f; + + // Main loop + for (; ib < nb; ++ib) { + const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d)); + + summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + __m256i bxhi = bytes_from_bits_32(x[ib].qh); + bxhi = _mm256_and_si256(bxhi, _mm256_set1_epi8(0x10)); + qx = _mm256_or_si256(qx, bxhi); + + const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[ib].d)); + const __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_us8_pairs_float(qx, qy); + + acc = _mm256_fmadd_ps(q, _mm256_mul_ps(dx, dy), acc); + } + + sumf = hsum_float_8(acc) + summs; +#elif defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + __m128i mask = _mm_set1_epi8(0x10); + + float summs = 0.0f; + + // Main loop + for (; ib < nb; ++ib) { + const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d)); + + summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); + + __m256i bx_0 = bytes_from_nibbles_32(x[ib].qs); + const __m256i bxhi = bytes_from_bits_32(x[ib].qh); + __m128i bxhil = _mm256_castsi256_si128(bxhi); + __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); + bxhil = _mm_and_si128(bxhil, mask); + bxhih = _mm_and_si128(bxhih, mask); + __m128i bxl = _mm256_castsi256_si128(bx_0); + __m128i bxh = _mm256_extractf128_si256(bx_0, 1); + bxl = _mm_or_si128(bxl, bxhil); + bxh = _mm_or_si128(bxh, bxhih); + bx_0 = MM256_SET_M128I(bxh, bxl); + + const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[ib].d)); + const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_us8_pairs_float(bx_0, by_0); + + acc = _mm256_add_ps(_mm256_mul_ps(q, _mm256_mul_ps(dx, dy)), acc); + } + + sumf = hsum_float_8(acc) + summs; +#elif defined(__riscv_v_intrinsic) + uint32_t qh; + + size_t vl = __riscv_vsetvl_e8m1(qk/2); + + // temporary registers for shift operations + vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); + vuint32m2_t vt_2 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); + + for (; ib < nb; ++ib) { + memcpy(&qh, x[ib].qh, sizeof(uint32_t)); + + // load qh + vuint32m2_t vqh = __riscv_vmv_v_x_u32m2(qh, vl); + + // ((qh >> (j + 0)) << 4) & 0x10; + vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(vqh, vt_1, vl); + vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); + vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(xhl_0, 0x10, vl); + + // ((qh >> (j + 12)) ) & 0x10; + vuint32m2_t xhr_1 = __riscv_vsrl_vv_u32m2(vqh, vt_2, vl); + vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(xhr_1, 0x10, vl); + + // narrowing + vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xha_0, vl); + vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); + + vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xha_1, vl); + vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); + + // load + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); + + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); + + vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + + vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); + vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); + + vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); + + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); + + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); + } + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed int v0 = vec_splats((int32_t)0); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + +#pragma GCC unroll 4 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[ib].m)); + vector float vys = {GGML_FP16_TO_FP32(y[ib].s), 0.f, 0.f, 0.f}; + vsumf0 = vec_madd(vxmin, vys, vsumf0); + + vector unsigned long long aux64x2_0 = {(uint64_t)(table_b2b_0[x[ib].qh[0]]), (uint64_t)(table_b2b_0[x[ib].qh[1]])}; + vector unsigned long long aux64x2_1 = {(uint64_t)(table_b2b_0[x[ib].qh[2]]), (uint64_t)(table_b2b_0[x[ib].qh[3]])}; + + vector signed char qh0 = (vector signed char)aux64x2_0; + vector signed char qh1 = (vector signed char)aux64x2_1; + + vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); + + vector unsigned char q5x0 = (vector unsigned char)vec_or(vec_and(qxs, lowMask), qh0); + vector unsigned char q5x1 = (vector unsigned char)vec_or(vec_sr(qxs, v4), qh1); + + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl( 16, y[ib].qs); + + vector signed int vsumi0 = v0; + + vsumi0 = vec_msum(q8y0, q5x0, vsumi0); + vsumi0 = vec_msum(q8y1, q5x1, vsumi0); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + } + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + // Initialize accumulator with zeros + __m256 acc = (__m256)__lasx_xvldi(0); + + float summs = 0.0f; + + // Main loop + for (; ib < nb; ++ib) { + const __m256 dx = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d)); + + summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); + + __m256i qx = bytes_from_nibbles_32(x[ib].qs); + __m256i bxhi = bytes_from_bits_32(x[ib].qh); + bxhi = __lasx_xvand_v(bxhi, __lasx_xvreplgr2vr_b(0x10)); + qx = __lasx_xvor_v(qx, bxhi); + + const __m256 dy = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib].d)); + const __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); + + const __m256 q = mul_sum_us8_pairs_float(qx, qy); + + acc = __lasx_xvfmadd_s(q, __lasx_xvfmul_s(dx, dy), acc); + } + + sumf = hsum_float_8(acc) + summs; +#endif + for (; ib < nb; ++ib) { + uint32_t qh; + memcpy(&qh, x[ib].qh, sizeof(qh)); + + int sumi0 = 0; + int sumi1 = 0; + + for (int j = 0; j < qk/2; ++j) { + const uint8_t xh_0 = ((qh >> (j + 0)) << 4) & 0x10; + const uint8_t xh_1 = ((qh >> (j + 12)) ) & 0x10; + + const int32_t x0 = (x[ib].qs[j] & 0xF) | xh_0; + const int32_t x1 = (x[ib].qs[j] >> 4) | xh_1; + + sumi0 += (x0 * y[ib].qs[j]); + sumi1 += (x1 * y[ib].qs[j + qk/2]); + } + + int sumi = sumi0 + sumi1; + sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); + } + + *s = sumf; +} + +void ggml_vec_dot_q8_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; + const int nb = n / qk; + + assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q8_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; + +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q8_0 * restrict vx0 = vx; + const block_q8_0 * restrict vx1 = (const block_q8_0 *) ((const uint8_t*)vx + bx); + const block_q8_0 * restrict vy0 = vy; + const block_q8_0 * restrict vy1 = (const block_q8_0 *) ((const uint8_t*)vy + by); + + float32x4_t sumv0 = vdupq_n_f32(0.0f); + + for (int i = 0; i < nb; i++) { + const block_q8_0 * restrict b_x0 = &vx0[i]; + const block_q8_0 * restrict b_y0 = &vy0[i]; + + const block_q8_0 * restrict b_x1 = &vx1[i]; + const block_q8_0 * restrict b_y1 = &vy1[i]; + + const int8x16_t x0_l = vld1q_s8(b_x0->qs); + const int8x16_t x0_h = vld1q_s8(b_x0->qs + 16); + const int8x16_t x1_l = vld1q_s8(b_x1->qs); + const int8x16_t x1_h = vld1q_s8(b_x1->qs + 16); + + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); + + float32_t _scale[4] = { + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d) + }; + float32x4_t scale = vld1q_f32(_scale); + + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + + float32x4_t sumv1 = vextq_f32 (sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); + + vst1_f32(s, vget_low_f32 (sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + + return; + } +#endif + + int ib = 0; + float sumf = 0; + +#if defined(__ARM_FEATURE_SVE) + svfloat32_t sumv0 = svdup_n_f32(0.0f); + svfloat32_t sumv1 = svdup_n_f32(0.0f); + + const int vector_length = ggml_cpu_get_sve_cnt()*8; + + //VLA Implemenation for SVE + switch (vector_length) { + case 128: + { + // predicate for activating lanes for 16 Int8 elements + const svbool_t ph16 = svptrue_pat_b8 (SV_VL16); + const svbool_t pl16 = svptrue_pat_b32(SV_VL4); + + for (; ib + 1 < nb; ib += 2) { + const block_q8_0 * restrict x0 = &x[ib + 0]; + const block_q8_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + // load x + const svint8_t qx0_0 = svld1_s8(ph16, x0->qs); + const svint8_t qx0_1 = svld1_s8(ph16, x0->qs+16); + const svint8_t qx1_0 = svld1_s8(ph16, x1->qs); + const svint8_t qx1_1 = svld1_s8(ph16, x1->qs+16); + + // load y + const svint8_t qy0_0 = svld1_s8(ph16, y0->qs); + const svint8_t qy0_1 = svld1_s8(ph16, y0->qs+16); + const svint8_t qy1_0 = svld1_s8(ph16, y1->qs); + const svint8_t qy1_1 = svld1_s8(ph16, y1->qs+16); + + sumv0 = svmla_n_f32_x(pl16, sumv0, svcvt_f32_s32_x(pl16, svadd_x(pl16, + svdot_s32(svdup_n_s32(0), qx0_0, qy0_0), + svdot_s32(svdup_n_s32(0), qx0_1, qy0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = svmla_n_f32_x(pl16, sumv1, svcvt_f32_s32_x(pl16, svadd_x(pl16, + svdot_s32(svdup_n_s32(0), qx1_0, qy1_0), + svdot_s32(svdup_n_s32(0), qx1_1, qy1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = svaddv_f32(pl16, svadd_f32_x(pl16, sumv0, sumv1)); + } break; + case 256: + { + //printf("sve256"); + for (; ib + 1 < nb; ib += 2) { + const block_q8_0 * restrict x0 = &x[ib + 0]; + const block_q8_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + // load x + const svint8_t qx0 = svld1_s8(svptrue_b8(), x0->qs); + const svint8_t qx1 = svld1_s8(svptrue_b8(), x1->qs); + + // load y + const svint8_t qy0 = svld1_s8(svptrue_b8(), y0->qs); + const svint8_t qy1 = svld1_s8(svptrue_b8(), y1->qs); + + sumv0 = svmla_n_f32_x(svptrue_b32(), sumv0, svcvt_f32_s32_x(svptrue_b32(), + svdot_s32(svdup_n_s32(0), qx0, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = svmla_n_f32_x(svptrue_b32(), sumv1, svcvt_f32_s32_x(svptrue_b32(), + svdot_s32(svdup_n_s32(0), qx1, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); + } break; + case 512: + { + // predicate for activating high 256 bit + const svbool_t ph32 = svptrue_pat_b8(SV_VL32); + // predicate for activating low 256 bit + const svbool_t pl32 = svnot_b_z(svptrue_b8(), ph32); + + // predicate for activating high lanes for 8 float32 elements + const svbool_t ph8 = svptrue_pat_b32(SV_VL8); + // predicate for activating low lanes for 8 float32 elements + const svbool_t pl8 = svnot_b_z(svptrue_b32(), ph8); + + svfloat32_t sumv00 = svdup_n_f32(0.0f); + + for (; ib + 1 < nb; ib += 2) { + const block_q8_0 * restrict x0 = &x[ib + 0]; + const block_q8_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + //load 32 int8_t in first half of vector and put another 32 int8_t in second vector lower bits + // and add them to make one 64 element vector + // load x + const svint8_t qx_32 = svld1_s8(ph32, x0->qs); + svint8_t qx_64 = svld1_s8(pl32, x0->qs + 2); + + qx_64 = svadd_s8_x(svptrue_b8(), qx_32, qx_64); + + // load y + const svint8_t qy_32 = svld1_s8(ph32, y0->qs); + svint8_t qy_64 = svld1_s8(pl32, y0->qs + 2); + + qy_64 = svadd_s8_x(svptrue_b8(), qy_32, qy_64); + + // scale creation + const float32_t deq1 = GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d); + const float32_t deq2 = GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d); + + // duplicate deq1 in first half of vector and deq2 in second half of vector + const svfloat32_t temp = svdup_f32_m(svdup_f32_z(ph8, deq1), pl8, deq2); + + const svfloat32_t sumvt = svcvt_f32_s32_x(svptrue_b32(), svdot_s32(svdup_n_s32(0), qx_64, qy_64)); + + sumv00 = svmla_f32_m(svptrue_b32(), sumv00, sumvt, temp); + } + + sumf = svaddv_f32(svptrue_b32(), sumv00); + break; + } + default: + assert(false && "Unsupported vector length"); + break; + } +#elif defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + for (; ib + 1 < nb; ib += 2) { + const block_q8_0 * restrict x0 = &x[ib + 0]; + const block_q8_0 * restrict x1 = &x[ib + 1]; + const block_q8_0 * restrict y0 = &y[ib + 0]; + const block_q8_0 * restrict y1 = &y[ib + 1]; + + const int8x16_t x0_0 = vld1q_s8(x0->qs); + const int8x16_t x0_1 = vld1q_s8(x0->qs + 16); + const int8x16_t x1_0 = vld1q_s8(x1->qs); + const int8x16_t x1_1 = vld1q_s8(x1->qs + 16); + + // load y + const int8x16_t y0_0 = vld1q_s8(y0->qs); + const int8x16_t y0_1 = vld1q_s8(y0->qs + 16); + const int8x16_t y1_0 = vld1q_s8(y1->qs); + const int8x16_t y1_1 = vld1q_s8(y1->qs + 16); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), x0_0, y0_0), + ggml_vdotq_s32(vdupq_n_s32(0), x0_1, y0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), x1_0, y1_0), + ggml_vdotq_s32(vdupq_n_s32(0), x1_1, y1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); +#elif defined(__AVX2__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + // Main loop + for (; ib < nb; ++ib) { + // Compute combined scale for the block + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); + __m256i qx = _mm256_loadu_si256((const __m256i *)x[ib].qs); + __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + // Multiply q with scale and accumulate + acc = _mm256_fmadd_ps( d, q, acc ); + } + + sumf = hsum_float_8(acc); +#elif defined(__AVX__) + __m256 accum = _mm256_setzero_ps(); + + for (; ib + 1 < nb; ib += 2) { + const __m128i qx_1_0 = _mm_loadu_si128((const __m128i *)x[ib].qs); + const __m128i qx_1_1 = _mm_loadu_si128((const __m128i *)x[ib].qs + 1); + const __m128i qx_2_0 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); + const __m128i qx_2_1 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs + 1); + const __m128i qy_1_0 = _mm_loadu_si128((const __m128i *)y[ib].qs); + const __m128i qy_1_1 = _mm_loadu_si128((const __m128i *)y[ib].qs + 1); + const __m128i qy_2_0 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); + const __m128i qy_2_1 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs + 1); + + const __m256 p = mul_sum_i8_quad_float(qx_1_0, qx_1_1, qx_2_0, qx_2_1, qy_1_0, qy_1_1, qy_2_0, qy_2_1); + const __m256 deltas = quad_fp16_delta_float(x[ib].d, y[ib].d, x[ib + 1].d, y[ib + 1].d); + accum = _mm256_add_ps(_mm256_mul_ps(deltas, p), accum); + } + + sumf = hsum_float_8(accum); +#elif defined(__riscv_v_intrinsic) + size_t vl = __riscv_vsetvl_e8m1(qk); + + for (; ib < nb; ++ib) { + // load elements + vint8m1_t bx_0 = __riscv_vle8_v_i8m1(x[ib].qs, vl); + vint8m1_t by_0 = __riscv_vle8_v_i8m1(y[ib].qs, vl); + + vint16m2_t vw_mul = __riscv_vwmul_vv_i16m2(bx_0, by_0, vl); + + vint32m1_t v_zero = __riscv_vmv_v_x_i32m1(0, vl); + vint32m1_t v_sum = __riscv_vwredsum_vs_i16m2_i32m1(vw_mul, v_zero, vl); + + int sumi = __riscv_vmv_x_s_i32m1_i32(v_sum); + + sumf += sumi*(GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)); + } +#elif defined(__POWER9_VECTOR__) + const vector signed int v0 = vec_splats((int32_t)0); + vector float vsumf0 = vec_splats(0.0f); + +#pragma GCC unroll 8 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector signed char q8x0 = vec_xl( 0, x[ib].qs); + vector signed char q8x1 = vec_xl(16, x[ib].qs); + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl(16, y[ib].qs); + + vector signed short qv0 = vec_mule(q8x0, q8y0); + vector signed short qv1 = vec_mulo(q8x0, q8y0); + vector signed short qv2 = vec_mule(q8x1, q8y1); + vector signed short qv3 = vec_mulo(q8x1, q8y1); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + + vsumi0 = vec_sum4s(qv0, vsumi0); + vsumi1 = vec_sum4s(qv1, vsumi1); + vsumi0 = vec_sum4s(qv2, vsumi0); + vsumi1 = vec_sum4s(qv3, vsumi1); + + vsumi0 = vec_add(vsumi0, vsumi1); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + } + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + // Initialize accumulator with zeros + __m256 acc = (__m256)__lasx_xvldi(0); + + // Main loop + for (; ib < nb; ++ib) { + // Compute combined scale for the block + const __m256 d = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); + __m256i qx = __lasx_xvld((const __m256i *)x[ib].qs, 0); + __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); + + const __m256 q = mul_sum_i8_pairs_float(qx, qy); + + // Multiply q with scale and accumulate + acc = __lasx_xvfmadd_s( d, q, acc ); + } + + sumf = hsum_float_8(acc); +#endif + for (; ib < nb; ++ib) { + int sumi = 0; + + for (int j = 0; j < qk; j++) { + sumi += x[ib].qs[j]*y[ib].qs[j]; + } + + sumf += sumi*(GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)); + } + + *s = sumf; +} + +void ggml_vec_dot_tq1_0_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_tq1_0 * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + float sumf = 0.0f; + + uint8_t k_shift[16] = {1, 1, 1, 1, 3, 3, 3, 3, 9, 9, 9, 9, 27, 27, 27, 27}; + + const uint8x16_t shift = vld1q_u8(k_shift); + + for (int i = 0; i < nb; ++i) { +#if defined(__ARM_FEATURE_DOTPROD) + int32x4_t sumi0 = vdupq_n_s32(0); + int32x4_t sumi1 = vdupq_n_s32(0); +#else + int16x8_t sumi0 = vdupq_n_s16(0); + int16x8_t sumi1 = vdupq_n_s16(0); +#endif + + // first 32 bytes of 5 elements + { + uint8x16_t qx0 = vld1q_u8(x[i].qs + 0); + uint8x16_t qx1 = vld1q_u8(x[i].qs + 16); + uint8x16_t qx2 = vmulq_u8(qx0, vdupq_n_u8(3)); + uint8x16_t qx3 = vmulq_u8(qx1, vdupq_n_u8(3)); + uint8x16_t qx4 = vmulq_u8(qx0, vdupq_n_u8(9)); + uint8x16_t qx5 = vmulq_u8(qx1, vdupq_n_u8(9)); + uint8x16_t qx6 = vmulq_u8(qx0, vdupq_n_u8(27)); + uint8x16_t qx7 = vmulq_u8(qx1, vdupq_n_u8(27)); + uint8x16_t qx8 = vmulq_u8(qx0, vdupq_n_u8(81)); + uint8x16_t qx9 = vmulq_u8(qx1, vdupq_n_u8(81)); + + // multiply by 3 and keep the 2 bits above 8 bits + int8x16_t sqx0 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx0, vshrq_n_u8(qx0, 1)), 6)); + int8x16_t sqx1 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx1, vshrq_n_u8(qx1, 1)), 6)); + int8x16_t sqx2 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx2, vshrq_n_u8(qx2, 1)), 6)); + int8x16_t sqx3 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx3, vshrq_n_u8(qx3, 1)), 6)); + int8x16_t sqx4 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx4, vshrq_n_u8(qx4, 1)), 6)); + int8x16_t sqx5 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx5, vshrq_n_u8(qx5, 1)), 6)); + int8x16_t sqx6 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx6, vshrq_n_u8(qx6, 1)), 6)); + int8x16_t sqx7 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx7, vshrq_n_u8(qx7, 1)), 6)); + int8x16_t sqx8 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx8, vshrq_n_u8(qx8, 1)), 6)); + int8x16_t sqx9 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx9, vshrq_n_u8(qx9, 1)), 6)); + + const int8x16_t qy0 = vld1q_s8(y[i].qs + 0); + const int8x16_t qy1 = vld1q_s8(y[i].qs + 16); + const int8x16_t qy2 = vld1q_s8(y[i].qs + 32); + const int8x16_t qy3 = vld1q_s8(y[i].qs + 48); + const int8x16_t qy4 = vld1q_s8(y[i].qs + 64); + const int8x16_t qy5 = vld1q_s8(y[i].qs + 80); + const int8x16_t qy6 = vld1q_s8(y[i].qs + 96); + const int8x16_t qy7 = vld1q_s8(y[i].qs + 112); + const int8x16_t qy8 = vld1q_s8(y[i].qs + 128); + const int8x16_t qy9 = vld1q_s8(y[i].qs + 144); + +#if defined(__ARM_FEATURE_DOTPROD) + sumi0 = vdotq_s32(sumi0, sqx0, qy0); + sumi1 = vdotq_s32(sumi1, sqx1, qy1); + sumi0 = vdotq_s32(sumi0, sqx2, qy2); + sumi1 = vdotq_s32(sumi1, sqx3, qy3); + sumi0 = vdotq_s32(sumi0, sqx4, qy4); + sumi1 = vdotq_s32(sumi1, sqx5, qy5); + sumi0 = vdotq_s32(sumi0, sqx6, qy6); + sumi1 = vdotq_s32(sumi1, sqx7, qy7); + sumi0 = vdotq_s32(sumi0, sqx8, qy8); + sumi1 = vdotq_s32(sumi1, sqx9, qy9); +#else + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx6), vget_low_s8(qy6)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx6), vget_high_s8(qy6)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx7), vget_low_s8(qy7)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx7), vget_high_s8(qy7)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx8), vget_low_s8(qy8)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx8), vget_high_s8(qy8)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx9), vget_low_s8(qy9)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx9), vget_high_s8(qy9)); +#endif + } + + // last 16 bytes of 5-element, along with the 4 bytes of 4 elements + { + uint8x16_t qx0 = vld1q_u8(x[i].qs + 32); + uint8x16_t qx1 = vmulq_u8(qx0, vdupq_n_u8(3)); + uint8x16_t qx2 = vmulq_u8(qx0, vdupq_n_u8(9)); + uint8x16_t qx3 = vmulq_u8(qx0, vdupq_n_u8(27)); + uint8x16_t qx4 = vmulq_u8(qx0, vdupq_n_u8(81)); + uint32_t qh; + memcpy(&qh, x[i].qh, sizeof(qh)); // potentially unaligned + uint8x16_t qx5 = vreinterpretq_u8_u32(vdupq_n_u32(qh)); + qx5 = vmulq_u8(qx5, shift); + + // multiply by 3 and keep the 2 bits above 8 bits + int8x16_t sqx0 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx0, vshrq_n_u8(qx0, 1)), 6)); + int8x16_t sqx1 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx1, vshrq_n_u8(qx1, 1)), 6)); + int8x16_t sqx2 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx2, vshrq_n_u8(qx2, 1)), 6)); + int8x16_t sqx3 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx3, vshrq_n_u8(qx3, 1)), 6)); + int8x16_t sqx4 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx4, vshrq_n_u8(qx4, 1)), 6)); + int8x16_t sqx5 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx5, vshrq_n_u8(qx5, 1)), 6)); + + const int8x16_t qy0 = vld1q_s8(y[i].qs + 160); + const int8x16_t qy1 = vld1q_s8(y[i].qs + 176); + const int8x16_t qy2 = vld1q_s8(y[i].qs + 192); + const int8x16_t qy3 = vld1q_s8(y[i].qs + 208); + const int8x16_t qy4 = vld1q_s8(y[i].qs + 224); + const int8x16_t qy5 = vld1q_s8(y[i].qs + 240); + +#if defined(__ARM_FEATURE_DOTPROD) + sumi0 = vdotq_s32(sumi0, sqx0, qy0); + sumi1 = vdotq_s32(sumi1, sqx1, qy1); + sumi0 = vdotq_s32(sumi0, sqx2, qy2); + sumi1 = vdotq_s32(sumi1, sqx3, qy3); + sumi0 = vdotq_s32(sumi0, sqx4, qy4); + sumi1 = vdotq_s32(sumi1, sqx5, qy5); +#else + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); +#endif + } + + const int16x8_t ysum0 = vld1q_s16(y[i].bsums); + const int16x8_t ysum1 = vld1q_s16(y[i].bsums + 8); + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + +#if defined(__ARM_FEATURE_DOTPROD) + sumi0 = vaddq_s32(sumi0, sumi1); + sumi0 = vsubq_s32(sumi0, vpaddlq_s16(vaddq_s16(ysum0, ysum1))); + + sumf += d * (float) vaddvq_s32(sumi0); +#else + sumi0 = vaddq_s16(sumi0, sumi1); + sumi0 = vsubq_s16(sumi0, vaddq_s16(ysum0, ysum1)); + + sumf += d * (float) vaddlvq_s16(sumi0); +#endif + } + + *s = sumf; + +#elif defined(__AVX2__) + __m256 sumf = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + // 16-bit sums + __m256i sumi0 = _mm256_setzero_si256(); + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + + // first 32 bytes of 5 elements + { + __m256i qx0 = _mm256_loadu_si256((const __m256i *) (x[i].qs)); + // 8-bit multiplies with shifts, masks and adds + __m256i qx1 = _mm256_add_epi8(qx0, _mm256_add_epi8(qx0, qx0)); // 1 * 3 + __m256i qx2 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx0, 3), _mm256_set1_epi8(-8)), qx0); // 1 * 9 + __m256i qx3 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx1, 3), _mm256_set1_epi8(-8)), qx1); // 3 * 9 + __m256i qx4 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx2, 3), _mm256_set1_epi8(-8)), qx2); // 9 * 9 + + // TODO: can _mm256_mulhi_epu16 be faster even if 16-bits? + + // Cancel the +1 from avg so that it behaves like a halving add + qx0 = _mm256_subs_epu8(qx0, _mm256_set1_epi8(1)); + qx1 = _mm256_subs_epu8(qx1, _mm256_set1_epi8(1)); + qx2 = _mm256_subs_epu8(qx2, _mm256_set1_epi8(1)); + qx3 = _mm256_subs_epu8(qx3, _mm256_set1_epi8(1)); + qx4 = _mm256_subs_epu8(qx4, _mm256_set1_epi8(1)); + // Multiply by 3 and get the top 2 bits + qx0 = _mm256_avg_epu8(qx0, _mm256_avg_epu8(qx0, _mm256_setzero_si256())); + qx1 = _mm256_avg_epu8(qx1, _mm256_avg_epu8(qx1, _mm256_setzero_si256())); + qx2 = _mm256_avg_epu8(qx2, _mm256_avg_epu8(qx2, _mm256_setzero_si256())); + qx3 = _mm256_avg_epu8(qx3, _mm256_avg_epu8(qx3, _mm256_setzero_si256())); + qx4 = _mm256_avg_epu8(qx4, _mm256_avg_epu8(qx4, _mm256_setzero_si256())); + qx0 = _mm256_and_si256(_mm256_srli_epi16(qx0, 6), _mm256_set1_epi8(3)); + qx1 = _mm256_and_si256(_mm256_srli_epi16(qx1, 6), _mm256_set1_epi8(3)); + qx2 = _mm256_and_si256(_mm256_srli_epi16(qx2, 6), _mm256_set1_epi8(3)); + qx3 = _mm256_and_si256(_mm256_srli_epi16(qx3, 6), _mm256_set1_epi8(3)); + qx4 = _mm256_and_si256(_mm256_srli_epi16(qx4, 6), _mm256_set1_epi8(3)); + + const __m256i qy0 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 0)); + const __m256i qy1 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 32)); + const __m256i qy2 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 64)); + const __m256i qy3 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 96)); + const __m256i qy4 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 128)); + + qx0 = _mm256_maddubs_epi16(qx0, qy0); + qx1 = _mm256_maddubs_epi16(qx1, qy1); + qx2 = _mm256_maddubs_epi16(qx2, qy2); + qx3 = _mm256_maddubs_epi16(qx3, qy3); + qx4 = _mm256_maddubs_epi16(qx4, qy4); + + sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(qx0, qx1)); + sumi1 = _mm256_add_epi16(sumi1, _mm256_add_epi16(qx2, qx3)); + sumi2 = _mm256_add_epi16(sumi2, qx4); + } + + // last 16 bytes of 5-element, along with the 4 bytes of 4 elements + { + __m128i qx0 = _mm_loadu_si128((const __m128i *) (x[i].qs + 32)); + uint32_t qh; + memcpy(&qh, x[i].qh, sizeof(qh)); // potentially unaligned + __m256i qx5_l = _mm256_cvtepu8_epi16(_mm_set1_epi32(qh)); + __m128i qx1 = _mm_add_epi8(qx0, _mm_add_epi8(qx0, qx0)); // 1 * 3 + __m128i qx2 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx0, 3), _mm_set1_epi8(-8)), qx0); // 1 * 9 + __m128i qx3 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx1, 3), _mm_set1_epi8(-8)), qx1); // 3 * 9 + __m128i qx4 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx2, 3), _mm_set1_epi8(-8)), qx2); // 9 * 9 + __m256i qx01 = MM256_SET_M128I(qx1, qx0); + __m256i qx23 = MM256_SET_M128I(qx3, qx2); + + // avx2 does not have 8-bit multiplies, so 16-bit it is. + qx5_l = _mm256_mullo_epi16(qx5_l, _mm256_set_epi16(27, 27, 27, 27, 9, 9, 9, 9, 3, 3, 3, 3, 1, 1, 1, 1)); + qx5_l = _mm256_and_si256(qx5_l, _mm256_set1_epi16(0xFF)); + __m128i qx5 = _mm_packus_epi16(_mm256_castsi256_si128(qx5_l), _mm256_extracti128_si256(qx5_l, 1)); + + __m256i qx45 = MM256_SET_M128I(qx5, qx4); + + // Cancel the +1 from avg so that it behaves like a halving add + qx01 = _mm256_subs_epu8(qx01, _mm256_set1_epi8(1)); + qx23 = _mm256_subs_epu8(qx23, _mm256_set1_epi8(1)); + qx45 = _mm256_subs_epu8(qx45, _mm256_set1_epi8(1)); + // Multiply by 3 and get the top 2 bits + qx01 = _mm256_avg_epu8(qx01, _mm256_avg_epu8(qx01, _mm256_setzero_si256())); + qx23 = _mm256_avg_epu8(qx23, _mm256_avg_epu8(qx23, _mm256_setzero_si256())); + qx45 = _mm256_avg_epu8(qx45, _mm256_avg_epu8(qx45, _mm256_setzero_si256())); + qx01 = _mm256_and_si256(_mm256_srli_epi16(qx01, 6), _mm256_set1_epi8(3)); + qx23 = _mm256_and_si256(_mm256_srli_epi16(qx23, 6), _mm256_set1_epi8(3)); + qx45 = _mm256_and_si256(_mm256_srli_epi16(qx45, 6), _mm256_set1_epi8(3)); + + const __m256i qy01 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 160)); + const __m256i qy23 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 192)); + const __m256i qy45 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 224)); + + qx01 = _mm256_maddubs_epi16(qx01, qy01); + qx23 = _mm256_maddubs_epi16(qx23, qy23); + qx45 = _mm256_maddubs_epi16(qx45, qy45); + + sumi0 = _mm256_add_epi16(sumi0, qx01); + sumi1 = _mm256_add_epi16(sumi1, qx23); + sumi2 = _mm256_add_epi16(sumi2, qx45); + } + + const __m256i ysum = _mm256_loadu_si256((const __m256i *) y[i].bsums); + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(x[i].d)); + + sumi0 = _mm256_sub_epi16(sumi0, ysum); + sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(sumi1, sumi2)); + sumi0 = _mm256_madd_epi16(sumi0, _mm256_set1_epi16(1)); + + sumf = _mm256_add_ps(_mm256_mul_ps(_mm256_cvtepi32_ps(sumi0), d), sumf); + } + + *s = hsum_float_8(sumf); + +#else + const uint8_t pow3[6] = {1, 3, 9, 27, 81, 243}; + + float sumf = 0.0f; + + for (int i = 0; i < nb; ++i) { + int sum = 0; + + for (size_t j = 0; j < sizeof(x->qs) - sizeof(x->qs) % 32; j += 32) { + for (size_t l = 0; l < 5; ++l) { + for (size_t m = 0; m < 32; ++m) { + uint8_t q = x[i].qs[j + m] * pow3[l]; + uint16_t xi = ((uint16_t) q * 3) >> 8; + sum += (xi - 1) * y[i].qs[j*5 + l*32 + m]; + } + } + } + for (size_t j = sizeof(x->qs) - sizeof(x->qs) % 32; j < sizeof(x->qs); j += 16) { + for (size_t l = 0; l < 5; ++l) { + for (size_t m = 0; m < 16; ++m) { + uint8_t q = x[i].qs[j + m] * pow3[l]; + uint16_t xi = ((uint16_t) q * 3) >> 8; + sum += (xi - 1) * y[i].qs[j*5 + l*16 + m]; + } + } + } + + for (size_t l = 0; l < 4; ++l) { + for (size_t j = 0; j < sizeof(x->qh); ++j) { + uint8_t q = x[i].qh[j] * pow3[l]; + uint16_t xi = ((uint16_t) q * 3) >> 8; + sum += (xi - 1) * y[i].qs[sizeof(x->qs)*5 + l*sizeof(x->qh) + j]; + } + } + + sumf += (float) sum * (GGML_FP16_TO_FP32(x[i].d) * y[i].d); + } + + *s = sumf; +#endif +} + +void ggml_vec_dot_tq2_0_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_tq2_0 * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + float sumf = 0.0f; + + const uint8x16_t m3 = vdupq_n_u8(3); + + for (int i = 0; i < nb; ++i) { +#if defined(__ARM_FEATURE_DOTPROD) + int32x4_t sumi0 = vdupq_n_s32(0); + int32x4_t sumi1 = vdupq_n_s32(0); +#else + int16x8_t sumi0 = vdupq_n_s16(0); + int16x8_t sumi1 = vdupq_n_s16(0); +#endif + + for (size_t j = 0; j < sizeof(x->qs); j += 32) { + uint8x16_t qx0 = vld1q_u8(x[i].qs + j); + uint8x16_t qx1 = vld1q_u8(x[i].qs + j + 16); + uint8x16_t qx2 = vshrq_n_u8(qx0, 2); + uint8x16_t qx3 = vshrq_n_u8(qx1, 2); + uint8x16_t qx4 = vshrq_n_u8(qx0, 4); + uint8x16_t qx5 = vshrq_n_u8(qx1, 4); + uint8x16_t qx6 = vshrq_n_u8(qx0, 6); + uint8x16_t qx7 = vshrq_n_u8(qx1, 6); + + int8x16_t sqx0 = vreinterpretq_s8_u8(vandq_u8(qx0, m3)); + int8x16_t sqx1 = vreinterpretq_s8_u8(vandq_u8(qx1, m3)); + int8x16_t sqx2 = vreinterpretq_s8_u8(vandq_u8(qx2, m3)); + int8x16_t sqx3 = vreinterpretq_s8_u8(vandq_u8(qx3, m3)); + int8x16_t sqx4 = vreinterpretq_s8_u8(vandq_u8(qx4, m3)); + int8x16_t sqx5 = vreinterpretq_s8_u8(vandq_u8(qx5, m3)); + int8x16_t sqx6 = vreinterpretq_s8_u8(vandq_u8(qx6, m3)); + int8x16_t sqx7 = vreinterpretq_s8_u8(vandq_u8(qx7, m3)); + + const int8x16_t qy0 = vld1q_s8(y[i].qs + j*4 + 0); + const int8x16_t qy1 = vld1q_s8(y[i].qs + j*4 + 16); + const int8x16_t qy2 = vld1q_s8(y[i].qs + j*4 + 32); + const int8x16_t qy3 = vld1q_s8(y[i].qs + j*4 + 48); + const int8x16_t qy4 = vld1q_s8(y[i].qs + j*4 + 64); + const int8x16_t qy5 = vld1q_s8(y[i].qs + j*4 + 80); + const int8x16_t qy6 = vld1q_s8(y[i].qs + j*4 + 96); + const int8x16_t qy7 = vld1q_s8(y[i].qs + j*4 + 112); + +#if defined(__ARM_FEATURE_DOTPROD) + sumi0 = vdotq_s32(sumi0, sqx0, qy0); + sumi1 = vdotq_s32(sumi1, sqx1, qy1); + sumi0 = vdotq_s32(sumi0, sqx2, qy2); + sumi1 = vdotq_s32(sumi1, sqx3, qy3); + sumi0 = vdotq_s32(sumi0, sqx4, qy4); + sumi1 = vdotq_s32(sumi1, sqx5, qy5); + sumi0 = vdotq_s32(sumi0, sqx6, qy6); + sumi1 = vdotq_s32(sumi1, sqx7, qy7); +#else + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx6), vget_low_s8(qy6)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx6), vget_high_s8(qy6)); + sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx7), vget_low_s8(qy7)); + sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx7), vget_high_s8(qy7)); +#endif + } + + const int16x8_t ysum0 = vld1q_s16(y[i].bsums); + const int16x8_t ysum1 = vld1q_s16(y[i].bsums + 8); + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + +#if defined(__ARM_FEATURE_DOTPROD) + sumi0 = vaddq_s32(sumi0, sumi1); + sumi0 = vsubq_s32(sumi0, vpaddlq_s16(vaddq_s16(ysum0, ysum1))); + + sumf += d * (float) vaddvq_s32(sumi0); +#else + sumi0 = vaddq_s16(sumi0, sumi1); + sumi0 = vsubq_s16(sumi0, vaddq_s16(ysum0, ysum1)); + + sumf += d * (float) vaddlvq_s16(sumi0); +#endif + } + + *s = sumf; + +#elif defined(__AVX2__) + __m256 sumf = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + // 16-bit sums, because 256*127 still fits + __m256i sumi0 = _mm256_setzero_si256(); + __m256i sumi1 = _mm256_setzero_si256(); + + for (size_t j = 0; j < sizeof(x->qs); j += 32) { + __m256i qx0 = _mm256_loadu_si256((const __m256i *) (x[i].qs + j)); + __m256i qx1 = _mm256_srli_epi16(qx0, 2); + __m256i qx2 = _mm256_srli_epi16(qx0, 4); + __m256i qx3 = _mm256_srli_epi16(qx0, 6); + + // 0, 1, 2 (should not be 3) + qx0 = _mm256_and_si256(qx0, _mm256_set1_epi8(3)); + qx1 = _mm256_and_si256(qx1, _mm256_set1_epi8(3)); + qx2 = _mm256_and_si256(qx2, _mm256_set1_epi8(3)); + qx3 = _mm256_and_si256(qx3, _mm256_set1_epi8(3)); + + const __m256i qy0 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 0)); + const __m256i qy1 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 32)); + const __m256i qy2 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 64)); + const __m256i qy3 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 96)); + + qx0 = _mm256_maddubs_epi16(qx0, qy0); + qx1 = _mm256_maddubs_epi16(qx1, qy1); + qx2 = _mm256_maddubs_epi16(qx2, qy2); + qx3 = _mm256_maddubs_epi16(qx3, qy3); + + sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(qx0, qx1)); + sumi1 = _mm256_add_epi16(sumi1, _mm256_add_epi16(qx2, qx3)); + } + + const __m256i ysum = _mm256_loadu_si256((const __m256i *) y[i].bsums); + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(x[i].d)); + + sumi0 = _mm256_add_epi16(sumi0, sumi1); + sumi0 = _mm256_sub_epi16(sumi0, ysum); + sumi0 = _mm256_madd_epi16(sumi0, _mm256_set1_epi16(1)); + + sumf = _mm256_add_ps(_mm256_mul_ps(_mm256_cvtepi32_ps(sumi0), d), sumf); + } + + *s = hsum_float_8(sumf); + +#else + float sumf = 0.0f; + + for (int i = 0; i < nb; ++i) { + int32_t sumi = 0; + + for (size_t j = 0; j < sizeof(x->qs); j += 32) { + for (size_t l = 0; l < 4; ++l) { + for (size_t k = 0; k < 32; ++k) { + sumi += y[i].qs[j*4 + l*32 + k] * (((x[i].qs[j + k] >> (l*2)) & 3) - 1); + } + } + } + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + sumf += (float) sumi * d; + } + + *s = sumf; +#endif +} + +void ggml_vec_dot_q2_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q2_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON + const uint8x16_t m3 = vdupq_n_u8(0x3); + const uint8x16_t m4 = vdupq_n_u8(0xF); + + const int32x4_t vzero = vdupq_n_s32(0); + + ggml_int8x16x2_t q2bytes; + uint8_t aux[16]; + + float sum = 0; + + for (int i = 0; i < nb; ++i) { + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + const uint8_t * restrict sc = x[i].scales; + + const uint8x16_t mins_and_scales = vld1q_u8(sc); + const uint8x16_t scales = vandq_u8(mins_and_scales, m4); + vst1q_u8(aux, scales); + + const uint8x16_t mins = vshrq_n_u8(mins_and_scales, 4); + const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); + const ggml_int16x8x2_t mins16 = {{vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(mins))), vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(mins)))}}; + const int32x4_t s0 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[0]), vget_low_s16 (q8sums.val[0])), + vmull_s16(vget_high_s16(mins16.val[0]), vget_high_s16(q8sums.val[0]))); + const int32x4_t s1 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[1]), vget_low_s16 (q8sums.val[1])), + vmull_s16(vget_high_s16(mins16.val[1]), vget_high_s16(q8sums.val[1]))); + sum += dmin * vaddvq_s32(vaddq_s32(s0, s1)); + + int isum = 0; + int is = 0; + +// We use this macro instead of a function call because for some reason +// the code runs 2-3% slower, even if the function is declared inline +#define MULTIPLY_ACCUM_WITH_SCALE(index)\ + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * aux[is+(index)];\ + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * aux[is+1+(index)]; + +#define SHIFT_MULTIPLY_ACCUM_WITH_SCALE(shift, index)\ + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32;\ + q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[0], (shift)), m3));\ + q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[1], (shift)), m3));\ + MULTIPLY_ACCUM_WITH_SCALE((index)); + + for (int j = 0; j < QK_K/128; ++j) { + const ggml_uint8x16x2_t q2bits = ggml_vld1q_u8_x2(q2); q2 += 32; + + ggml_int8x16x2_t q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[0], m3)); + q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[1], m3)); + + MULTIPLY_ACCUM_WITH_SCALE(0); + + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(2, 2); + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(4, 4); + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(6, 6); + + is += 8; + } + + sum += d * isum; + } + + *s = sum; + +#elif defined __AVX2__ + + const __m256i m3 = _mm256_set1_epi8(3); + const __m128i m4 = _mm_set1_epi8(0xF); + + __m256 acc = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); + const __m128i scales8 = _mm_and_si128(mins_and_scales, m4); + const __m128i mins8 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); + const __m256i mins = _mm256_cvtepi8_epi16(mins8); + const __m256i prod = _mm256_madd_epi16(mins, _mm256_loadu_si256((const __m256i*)y[i].bsums)); + + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(prod), acc); + + const __m256i all_scales = _mm256_cvtepi8_epi16(scales8); + const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); + const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); + const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; + + __m256i sumi = _mm256_setzero_si256(); + + for (int j = 0; j < QK_K/128; ++j) { + + const __m256i q2bits = _mm256_loadu_si256((const __m256i*)q2); q2 += 32; + + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + const __m256i q2_0 = _mm256_and_si256(q2bits, m3); + const __m256i q2_1 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 2), m3); + const __m256i q2_2 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 4), m3); + const __m256i q2_3 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 6), m3); + + __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); + __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); + __m256i p2 = _mm256_maddubs_epi16(q2_2, q8_2); + __m256i p3 = _mm256_maddubs_epi16(q2_3, q8_3); + + p0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(0)), p0); + p1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(1)), p1); + p2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(2)), p2); + p3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(3)), p3); + + p0 = _mm256_add_epi32(p0, p1); + p2 = _mm256_add_epi32(p2, p3); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p0, p2)); + } + + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(0x3); + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i m2 = _mm_set1_epi8(0x2); + + __m256 acc = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + // load mins and scales from block_q2_K.scales[QK_K/16] + const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); + const __m128i scales16 = _mm_and_si128(mins_and_scales, m4); + const __m128i mins16 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); + const __m128i mins_0 = _mm_cvtepi8_epi16(mins16); + const __m128i mins_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(mins16, mins16)); + + // summs = y[i].bsums * (x[i].scales >> 4) in 16bits*8*2 to 32bits*4*2 + const __m128i summs_0 = _mm_madd_epi16(mins_0, _mm_loadu_si128((const __m128i*)&y[i].bsums[0])); + const __m128i summs_1 = _mm_madd_epi16(mins_1, _mm_loadu_si128((const __m128i*)&y[i].bsums[8])); + + // sumf += -dmin * summs in 32bits*8 + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(MM256_SET_M128I(summs_1, summs_0))), acc); + + const __m128i scales_0 = _mm_cvtepi8_epi16(scales16); + const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales16, scales16)); + const __m128i scales[2] = { scales_0, scales_1 }; + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + for (int j = 0; j < QK_K/128; ++j) { + + // load Q8 quants int8*16*8 from block_q8_K.qs[QK_K] + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + // load 2bits*16*8 from block_q2_K.qs[QK_K/4] + __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; + const __m128i q2_0 = _mm_and_si128(q2bits, m3); + const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); + const __m128i q2_4 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); + const __m128i q2_6 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; + const __m128i q2_1 = _mm_and_si128(q2bits, m3); + const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); + const __m128i q2_5 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); + const __m128i q2_7 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + + // isuml = q8[l] * ((q2[l] >> shift) & 3) in 8bits*16*8 to 16bits*8*8 + __m128i p0 = _mm_maddubs_epi16(q2_0, q8_0); + __m128i p1 = _mm_maddubs_epi16(q2_1, q8_1); + __m128i p2 = _mm_maddubs_epi16(q2_2, q8_2); + __m128i p3 = _mm_maddubs_epi16(q2_3, q8_3); + __m128i p4 = _mm_maddubs_epi16(q2_4, q8_4); + __m128i p5 = _mm_maddubs_epi16(q2_5, q8_5); + __m128i p6 = _mm_maddubs_epi16(q2_6, q8_6); + __m128i p7 = _mm_maddubs_epi16(q2_7, q8_7); + + // isum += (x[i].scales[is++] & 0xF) * isuml in 16bits*8*8 to 32bits*4*8 + __m128i shuffle = _mm_set1_epi16(0x0100); + p0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p0); + shuffle = _mm_add_epi16(shuffle, m2); + p1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p1); + shuffle = _mm_add_epi16(shuffle, m2); + p2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p2); + shuffle = _mm_add_epi16(shuffle, m2); + p3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p3); + shuffle = _mm_add_epi16(shuffle, m2); + p4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p4); + shuffle = _mm_add_epi16(shuffle, m2); + p5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p5); + shuffle = _mm_add_epi16(shuffle, m2); + p6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p6); + shuffle = _mm_add_epi16(shuffle, m2); + p7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p7); + + p0 = _mm_add_epi32(p0, p1); + p2 = _mm_add_epi32(p2, p3); + p4 = _mm_add_epi32(p4, p5); + p6 = _mm_add_epi32(p6, p7); + + // isum in 32bits*4*2 + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p0, p2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p4, p6)); + } + + // sumf += dall * isum - dmin * summs in 32bits + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dall), _mm256_cvtepi32_ps(sumi)), acc); + } + + *s = hsum_float_8(acc); + +#elif defined __riscv_v_intrinsic + + float sumf = 0; + uint8_t temp_01[32] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * q2 = x[i].qs; + const int8_t * q8 = y[i].qs; + const uint8_t * sc = x[i].scales; + + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + size_t vl = 16; + + vuint8m1_t scales = __riscv_vle8_v_u8m1(sc, vl); + vuint8m1_t aux = __riscv_vand_vx_u8m1(scales, 0x0F, vl); + + vint16m1_t q8sums = __riscv_vle16_v_i16m1(y[i].bsums, vl); + + vuint8mf2_t scales_2 = __riscv_vle8_v_u8mf2(sc, vl); + vuint8mf2_t mins8 = __riscv_vsrl_vx_u8mf2(scales_2, 0x4, vl); + vint16m1_t mins = __riscv_vreinterpret_v_u16m1_i16m1(__riscv_vzext_vf2_u16m1(mins8, vl)); + vint32m2_t prod = __riscv_vwmul_vv_i32m2(q8sums, mins, vl); + vint32m1_t vsums = __riscv_vredsum_vs_i32m2_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + + sumf += dmin * __riscv_vmv_x_s_i32m1_i32(vsums); + + vl = 32; + + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t v_b = __riscv_vle8_v_u8m1(temp_01, vl); + + uint8_t is=0; + int isum=0; + + for (int j = 0; j < QK_K/128; ++j) { + // load Q2 + vuint8m1_t q2_x = __riscv_vle8_v_u8m1(q2, vl); + + vuint8m1_t q2_0 = __riscv_vand_vx_u8m1(q2_x, 0x03, vl); + vuint8m1_t q2_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x2, vl), 0x03 , vl); + vuint8m1_t q2_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x4, vl), 0x03 , vl); + vuint8m1_t q2_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x6, vl), 0x03 , vl); + + // duplicate scale elements for product + vuint8m1_t sc0 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 0+is, vl), vl); + vuint8m1_t sc1 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 2+is, vl), vl); + vuint8m1_t sc2 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 4+is, vl), vl); + vuint8m1_t sc3 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 6+is, vl), vl); + + vint16m2_t p0 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_0, sc0, vl)); + vint16m2_t p1 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_1, sc1, vl)); + vint16m2_t p2 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_2, sc2, vl)); + vint16m2_t p3 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_3, sc3, vl)); + + // load Q8 + vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); + vint8m1_t q8_2 = __riscv_vle8_v_i8m1(q8+64, vl); + vint8m1_t q8_3 = __riscv_vle8_v_i8m1(q8+96, vl); + + vint32m4_t s0 = __riscv_vwmul_vv_i32m4(p0, __riscv_vwcvt_x_x_v_i16m2(q8_0, vl), vl); + vint32m4_t s1 = __riscv_vwmul_vv_i32m4(p1, __riscv_vwcvt_x_x_v_i16m2(q8_1, vl), vl); + vint32m4_t s2 = __riscv_vwmul_vv_i32m4(p2, __riscv_vwcvt_x_x_v_i16m2(q8_2, vl), vl); + vint32m4_t s3 = __riscv_vwmul_vv_i32m4(p3, __riscv_vwcvt_x_x_v_i16m2(q8_3, vl), vl); + + vint32m1_t isum0 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s0, s1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s2, s3, vl), isum0, vl); + + isum += __riscv_vmv_x_s_i32m1_i32(isum1); + + q2+=32; q8+=128; is=8; + + } + + sumf += dall * isum; + + } + + *s = sumf; + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0x3); + const vector signed char lowScaleMask = vec_splats((signed char)0xF); + const vector int v0 = vec_splats((int32_t)0); + const vector unsigned char v2 = vec_splats((unsigned char)0x2); + const vector unsigned char v6 = vec_splats((unsigned char)0x6); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); + vector float vdmin = vec_mul(vxmin, vyd); + + vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); + vector signed short q8ysums1 = vec_xl(16, y[i].bsums); + + vector signed char q2xmins = (vector signed char)vec_xl( 0, x[i].scales); + vector signed char vscales = vec_and(q2xmins, lowScaleMask); + + q2xmins = vec_sr(q2xmins, v4); + vector signed short q2xmins0 = vec_unpackh(q2xmins); + vector signed short q2xmins1 = vec_unpackl(q2xmins); + + vector signed int prod0 = vec_mule(q2xmins0, q8ysums0); + vector signed int prod1 = vec_mulo(q2xmins0, q8ysums0); + vector signed int prod2 = vec_mule(q2xmins1, q8ysums1); + vector signed int prod3 = vec_mulo(q2xmins1, q8ysums1); + + vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); + vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); + vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); + vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + vector signed int vsumi4 = v0; + vector signed int vsumi5 = v0; + vector signed int vsumi6 = v0; + vector signed int vsumi7 = v0; + + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/128; ++j) { + __builtin_prefetch(q2, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q2); + vector signed char qxs1 = (vector signed char)vec_xl(16, q2); + q2 += 32; + + vector unsigned char q2x00 = (vector unsigned char)vec_and(qxs0, lowMask); + vector unsigned char q2x01 = (vector unsigned char)vec_and(vec_sr(qxs0, v2), lowMask); + vector unsigned char q2x02 = (vector unsigned char)vec_and(vec_sr(qxs0, v4), lowMask); + vector unsigned char q2x03 = (vector unsigned char)vec_and(vec_sr(qxs0, v6), lowMask); + vector unsigned char q2x10 = (vector unsigned char)vec_and(qxs1, lowMask); + vector unsigned char q2x11 = (vector unsigned char)vec_and(vec_sr(qxs1, v2), lowMask); + vector unsigned char q2x12 = (vector unsigned char)vec_and(vec_sr(qxs1, v4), lowMask); + vector unsigned char q2x13 = (vector unsigned char)vec_and(vec_sr(qxs1, v6), lowMask); + + vector signed char q8y00 = vec_xl( 0, q8); + vector signed char q8y10 = vec_xl( 16, q8); + vector signed char q8y01 = vec_xl( 32, q8); + vector signed char q8y11 = vec_xl( 48, q8); + vector signed char q8y02 = vec_xl( 64, q8); + vector signed char q8y12 = vec_xl( 80, q8); + vector signed char q8y03 = vec_xl( 96, q8); + vector signed char q8y13 = vec_xl(112, q8); + q8 += 128; + + vector signed int qv0 = vec_msum(q8y00, q2x00, v0); + vector signed int qv1 = vec_msum(q8y01, q2x01, v0); + vector signed int qv2 = vec_msum(q8y02, q2x02, v0); + vector signed int qv3 = vec_msum(q8y03, q2x03, v0); + vector signed int qv4 = vec_msum(q8y10, q2x10, v0); + vector signed int qv5 = vec_msum(q8y11, q2x11, v0); + vector signed int qv6 = vec_msum(q8y12, q2x12, v0); + vector signed int qv7 = vec_msum(q8y13, q2x13, v0); + + vector signed short vscales_07 = vec_unpackh(vscales); + vector signed int vscales_03 = vec_unpackh(vscales_07); + vector signed int vscales_47 = vec_unpackl(vscales_07); + vector signed int vs0 = vec_splat(vscales_03, 0); + vector signed int vs1 = vec_splat(vscales_03, 1); + vector signed int vs2 = vec_splat(vscales_03, 2); + vector signed int vs3 = vec_splat(vscales_03, 3); + vector signed int vs4 = vec_splat(vscales_47, 0); + vector signed int vs5 = vec_splat(vscales_47, 1); + vector signed int vs6 = vec_splat(vscales_47, 2); + vector signed int vs7 = vec_splat(vscales_47, 3); + vscales = vec_sld(vscales, vscales, 8); + + vsumi0 = vec_add(vec_mul(qv0, vs0), vsumi0); + vsumi1 = vec_add(vec_mul(qv1, vs2), vsumi1); + vsumi2 = vec_add(vec_mul(qv2, vs4), vsumi2); + vsumi3 = vec_add(vec_mul(qv3, vs6), vsumi3); + vsumi4 = vec_add(vec_mul(qv4, vs1), vsumi4); + vsumi5 = vec_add(vec_mul(qv5, vs3), vsumi5); + vsumi6 = vec_add(vec_mul(qv6, vs5), vsumi6); + vsumi7 = vec_add(vec_mul(qv7, vs7), vsumi7); + } + + vsumi0 = vec_add(vsumi0, vsumi4); + vsumi1 = vec_add(vsumi1, vsumi5); + vsumi2 = vec_add(vsumi2, vsumi6); + vsumi3 = vec_add(vsumi3, vsumi7); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined __loongarch_asx + + const __m256i m3 = __lasx_xvreplgr2vr_b(3); + const __m128i m4 = __lsx_vreplgr2vr_b(0xF); + + __m256 acc = (__m256)__lasx_xvldi(0); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m128i mins_and_scales = __lsx_vld((const __m128i*)x[i].scales, 0); + const __m128i scales8 = __lsx_vand_v(mins_and_scales, m4); + const __m128i mins8 = __lsx_vand_v(__lsx_vsrli_h(mins_and_scales, 4), m4); + const __m256i mins = lasx_ext8_16(mins8); + const __m256i prod = lasx_madd_h(mins, __lasx_xvld((const __m256i*)y[i].bsums, 0)); + + acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(dmin), __lasx_xvffint_s_w(prod), acc); + + const __m256i all_scales = lasx_ext8_16(scales8); + const __m128i l_scales = lasx_extracti128(all_scales, 0); + const __m128i h_scales = lasx_extracti128(all_scales, 1); + const __m256i scales[2] = {lasx_insertf128(l_scales, l_scales), lasx_insertf128(h_scales, h_scales)}; + + __m256i sumi = __lasx_xvldi(0); + + for (int j = 0; j < QK_K/128; ++j) { + + const __m256i q2bits = __lasx_xvld((const __m256i*)q2, 0); q2 += 32; + + const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + + const __m256i q2_0 = __lasx_xvand_v(q2bits, m3); + const __m256i q2_1 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 2), m3); + const __m256i q2_2 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 4), m3); + const __m256i q2_3 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 6), m3); + + __m256i p0 = lasx_maddubs_h(q2_0, q8_0); + __m256i p1 = lasx_maddubs_h(q2_1, q8_1); + __m256i p2 = lasx_maddubs_h(q2_2, q8_2); + __m256i p3 = lasx_maddubs_h(q2_3, q8_3); + + p0 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(0)), p0); + p1 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(1)), p1); + p2 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(2)), p2); + p3 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(3)), p3); + + p0 = __lasx_xvadd_w(p0, p1); + p2 = __lasx_xvadd_w(p2, p3); + + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p0, p2)); + } + + acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc); + + } + + *s = hsum_float_8(acc); + +#else + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * q2 = x[i].qs; + const int8_t * q8 = y[i].qs; + const uint8_t * sc = x[i].scales; + + int summs = 0; + for (int j = 0; j < 16; ++j) { + summs += y[i].bsums[j] * (sc[j] >> 4); + } + + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + int isum = 0; + int is = 0; + int d; + for (int k = 0; k < QK_K/128; ++k) { + int shift = 0; + for (int j = 0; j < 4; ++j) { + d = sc[is++] & 0xF; + int isuml = 0; + for (int l = 0; l < 16; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); + isum += d * isuml; + d = sc[is++] & 0xF; + isuml = 0; + for (int l = 16; l < 32; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); + isum += d * isuml; + shift += 2; + q8 += 32; + } + q2 += 32; + } + sumf += dall * isum - dmin * summs; + } + *s = sumf; +#endif +} + +void ggml_vec_dot_q3_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const uint32_t kmask1 = 0x03030303; + const uint32_t kmask2 = 0x0f0f0f0f; + + const block_q3_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON + + uint32_t aux[3]; + uint32_t utmp[4]; + + const uint8x16_t m3b = vdupq_n_u8(0x3); + const int32x4_t vzero = vdupq_n_s32(0); + + const uint8x16_t m0 = vdupq_n_u8(1); + const uint8x16_t m1 = vshlq_n_u8(m0, 1); + const uint8x16_t m2 = vshlq_n_u8(m0, 2); + const uint8x16_t m3 = vshlq_n_u8(m0, 3); + const int8_t m32 = 32; + + ggml_int8x16x4_t q3bytes; + + float sum = 0; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict qh = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); + + ggml_uint8x16x4_t q3h; + + int32_t isum = 0; + + // Set up scales + memcpy(aux, x[i].scales, 12); + utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); + utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); + utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); + utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); + + int8_t * scale = (int8_t *)utmp; + for (int j = 0; j < 16; ++j) scale[j] -= m32; + + for (int j = 0; j < QK_K/128; ++j) { + + const ggml_uint8x16x2_t q3bits = ggml_vld1q_u8_x2(q3); q3 += 32; + const ggml_int8x16x4_t q8bytes_1 = ggml_vld1q_s8_x4(q8); q8 += 64; + const ggml_int8x16x4_t q8bytes_2 = ggml_vld1q_s8_x4(q8); q8 += 64; + + q3h.val[0] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[0]), 2); + q3h.val[1] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[1]), 2); + q3h.val[2] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[0]), 1); + q3h.val[3] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[1]), 1); + + q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[0], m3b)), vreinterpretq_s8_u8(q3h.val[0])); + q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[1], m3b)), vreinterpretq_s8_u8(q3h.val[1])); + q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 2), m3b)), vreinterpretq_s8_u8(q3h.val[2])); + q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 2), m3b)), vreinterpretq_s8_u8(q3h.val[3])); + + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_1.val[0])) * scale[0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_1.val[1])) * scale[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_1.val[2])) * scale[2]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_1.val[3])) * scale[3]; + + scale += 4; + + q3h.val[0] = vbicq_u8(m2, qhbits.val[0]); + q3h.val[1] = vbicq_u8(m2, qhbits.val[1]); + q3h.val[2] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[0]), 1); + q3h.val[3] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[1]), 1); + + q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 4), m3b)), vreinterpretq_s8_u8(q3h.val[0])); + q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 4), m3b)), vreinterpretq_s8_u8(q3h.val[1])); + q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 6), m3b)), vreinterpretq_s8_u8(q3h.val[2])); + q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 6), m3b)), vreinterpretq_s8_u8(q3h.val[3])); + + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_2.val[0])) * scale[0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_2.val[1])) * scale[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_2.val[2])) * scale[2]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_2.val[3])) * scale[3]; + + scale += 4; + + if (j == 0) { + qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 4); + qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 4); + } + + } + sum += d * isum; + + } + + *s = sum; + +#elif defined __AVX2__ + + const __m256i m3 = _mm256_set1_epi8(3); + const __m256i mone = _mm256_set1_epi8(1); + const __m128i m32 = _mm_set1_epi8(32); + + __m256 acc = _mm256_setzero_ps(); + + uint32_t aux[3]; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + // Set up scales + memcpy(aux, x[i].scales, 12); + __m128i scales128 = _mm_set_epi32( + ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), + ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), + (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), + (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); + scales128 = _mm_sub_epi8(scales128, m32); + const __m256i all_scales = _mm256_cvtepi8_epi16(scales128); + const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); + const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); + const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; + + // high bit + const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].hmask); + + // integer accumulator + __m256i sumi = _mm256_setzero_si256(); + + int bit = 0; + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + // load low 2 bits + const __m256i q3bits = _mm256_loadu_si256((const __m256i*)q3); q3 += 32; + + // prepare low and high bits + const __m256i q3l_0 = _mm256_and_si256(q3bits, m3); + const __m256i q3h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 2), m3); + const __m256i q3h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_2 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 4), m3); + const __m256i q3h_2 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_3 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 6), m3); + const __m256i q3h_3 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + // load Q8 quants + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); + __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); + __m256i q8s_2 = _mm256_maddubs_epi16(q3h_2, q8_2); + __m256i q8s_3 = _mm256_maddubs_epi16(q3h_3, q8_3); + + __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); + __m256i p16_2 = _mm256_maddubs_epi16(q3l_2, q8_2); + __m256i p16_3 = _mm256_maddubs_epi16(q3l_3, q8_3); + + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + p16_2 = _mm256_sub_epi16(p16_2, q8s_2); + p16_3 = _mm256_sub_epi16(p16_3, q8s_3); + + // multiply with scales + p16_0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); + p16_1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); + p16_2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); + p16_3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); + + // accumulate + p16_0 = _mm256_add_epi32(p16_0, p16_1); + p16_2 = _mm256_add_epi32(p16_2, p16_3); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_2)); + + } + + // multiply with block scale and accumulate + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(3); + const __m128i mone = _mm_set1_epi8(1); + const __m128i m32 = _mm_set1_epi8(32); + const __m128i m2 = _mm_set1_epi8(2); + + __m256 acc = _mm256_setzero_ps(); + + const uint32_t *aux; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + // Set up scales + aux = (const uint32_t *)x[i].scales; + __m128i scales128 = _mm_set_epi32( + ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), + ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), + (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), + (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); + scales128 = _mm_sub_epi8(scales128, m32); + const __m128i scales_0 = _mm_cvtepi8_epi16(scales128); + const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales128, scales128)); + const __m128i scales[2] = { scales_0, scales_1 }; + + // high bit *128*2 from block_q3_K.hmask[QK_K/8] + const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].hmask[0]); + const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].hmask[16]); + + // integer accumulator + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + for (int j = 0; j < QK_K/128; ++j) { + // load low 2 bits *64*2 from block_q3_K.qs[QK_K/4] + const __m128i q3bits_0 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; + const __m128i q3bits_1 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; + + // prepare low and high bits + const int bit = j << 2; + + const __m128i q3l_0 = _mm_and_si128(q3bits_0, m3); + const __m128i q3l_1 = _mm_and_si128(q3bits_1, m3); + const __m128i q3h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit)), bit), 2); + const __m128i q3h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit)), bit), 2); + + const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 2), m3); + const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 2), m3); + const __m128i q3h_2 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+1)), bit+1), 2); + const __m128i q3h_3 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+1)), bit+1), 2); + + const __m128i q3l_4 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 4), m3); + const __m128i q3l_5 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 4), m3); + const __m128i q3h_4 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+2)), bit+2), 2); + const __m128i q3h_5 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+2)), bit+2), 2); + + const __m128i q3l_6 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 6), m3); + const __m128i q3l_7 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 6), m3); + const __m128i q3h_6 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+3)), bit+3), 2); + const __m128i q3h_7 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+3)), bit+3), 2); + + // load Q8 quants from block_q8_K.qs[QK_K] + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, q8_0); + __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, q8_1); + __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, q8_2); + __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, q8_3); + __m128i q8s_4 = _mm_maddubs_epi16(q3h_4, q8_4); + __m128i q8s_5 = _mm_maddubs_epi16(q3h_5, q8_5); + __m128i q8s_6 = _mm_maddubs_epi16(q3h_6, q8_6); + __m128i q8s_7 = _mm_maddubs_epi16(q3h_7, q8_7); + + __m128i p16_0 = _mm_maddubs_epi16(q3l_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q3l_1, q8_1); + __m128i p16_2 = _mm_maddubs_epi16(q3l_2, q8_2); + __m128i p16_3 = _mm_maddubs_epi16(q3l_3, q8_3); + __m128i p16_4 = _mm_maddubs_epi16(q3l_4, q8_4); + __m128i p16_5 = _mm_maddubs_epi16(q3l_5, q8_5); + __m128i p16_6 = _mm_maddubs_epi16(q3l_6, q8_6); + __m128i p16_7 = _mm_maddubs_epi16(q3l_7, q8_7); + + p16_0 = _mm_sub_epi16(p16_0, q8s_0); + p16_1 = _mm_sub_epi16(p16_1, q8s_1); + p16_2 = _mm_sub_epi16(p16_2, q8s_2); + p16_3 = _mm_sub_epi16(p16_3, q8s_3); + p16_4 = _mm_sub_epi16(p16_4, q8s_4); + p16_5 = _mm_sub_epi16(p16_5, q8s_5); + p16_6 = _mm_sub_epi16(p16_6, q8s_6); + p16_7 = _mm_sub_epi16(p16_7, q8s_7); + + // multiply with scales + __m128i shuffle = _mm_set1_epi16(0x0100); + p16_0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_0); + shuffle = _mm_add_epi16(shuffle, m2); + p16_1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_1); + shuffle = _mm_add_epi16(shuffle, m2); + p16_2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_2); + shuffle = _mm_add_epi16(shuffle, m2); + p16_3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_3); + shuffle = _mm_add_epi16(shuffle, m2); + p16_4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_4); + shuffle = _mm_add_epi16(shuffle, m2); + p16_5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_5); + shuffle = _mm_add_epi16(shuffle, m2); + p16_6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_6); + shuffle = _mm_add_epi16(shuffle, m2); + p16_7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_7); + + // accumulate + p16_0 = _mm_add_epi32(p16_0, p16_1); + p16_2 = _mm_add_epi32(p16_2, p16_3); + p16_4 = _mm_add_epi32(p16_4, p16_5); + p16_6 = _mm_add_epi32(p16_6, p16_7); + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_4, p16_6)); + + } + + // multiply with block scale and accumulate + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __riscv_v_intrinsic + + uint32_t aux[3]; + uint32_t utmp[4]; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict qh = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + + memcpy(aux, x[i].scales, 12); + utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); + utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); + utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); + utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); + + int8_t * scale = (int8_t *)utmp; + for (int j = 0; j < 16; ++j) scale[j] -= 32; + + + size_t vl = 32; + uint8_t m = 1; + + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t vqh = __riscv_vle8_v_u8m1(qh, vl); + + int sum_t = 0; + + for (int j = 0; j < QK_K; j += 128) { + + vl = 32; + + // load Q3 + vuint8m1_t q3_x = __riscv_vle8_v_u8m1(q3, vl); + + vint8m1_t q3_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q3_x, 0x03, vl)); + vint8m1_t q3_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x2, vl), 0x03 , vl)); + vint8m1_t q3_2 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x4, vl), 0x03 , vl)); + vint8m1_t q3_3 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x6, vl), 0x03 , vl)); + + // compute mask for subtraction + vuint8m1_t qh_m0 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_0 = __riscv_vmseq_vx_u8m1_b8(qh_m0, 0, vl); + vint8m1_t q3_m0 = __riscv_vsub_vx_i8m1_mu(vmask_0, q3_0, q3_0, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_1 = __riscv_vmseq_vx_u8m1_b8(qh_m1, 0, vl); + vint8m1_t q3_m1 = __riscv_vsub_vx_i8m1_mu(vmask_1, q3_1, q3_1, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_2 = __riscv_vmseq_vx_u8m1_b8(qh_m2, 0, vl); + vint8m1_t q3_m2 = __riscv_vsub_vx_i8m1_mu(vmask_2, q3_2, q3_2, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m3 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_3 = __riscv_vmseq_vx_u8m1_b8(qh_m3, 0, vl); + vint8m1_t q3_m3 = __riscv_vsub_vx_i8m1_mu(vmask_3, q3_3, q3_3, 0x4, vl); + m <<= 1; + + // load Q8 and take product with Q3 + vint16m2_t a0 = __riscv_vwmul_vv_i16m2(q3_m0, __riscv_vle8_v_i8m1(q8, vl), vl); + vint16m2_t a1 = __riscv_vwmul_vv_i16m2(q3_m1, __riscv_vle8_v_i8m1(q8+32, vl), vl); + vint16m2_t a2 = __riscv_vwmul_vv_i16m2(q3_m2, __riscv_vle8_v_i8m1(q8+64, vl), vl); + vint16m2_t a3 = __riscv_vwmul_vv_i16m2(q3_m3, __riscv_vle8_v_i8m1(q8+96, vl), vl); + + vl = 16; + + // retrieve lane to multiply with scale + vint32m2_t aux0_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 0), (scale[0]), vl); + vint32m2_t aux0_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 1), (scale[1]), vl); + vint32m2_t aux1_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 0), (scale[2]), vl); + vint32m2_t aux1_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 1), (scale[3]), vl); + vint32m2_t aux2_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 0), (scale[4]), vl); + vint32m2_t aux2_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 1), (scale[5]), vl); + vint32m2_t aux3_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 0), (scale[6]), vl); + vint32m2_t aux3_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 1), (scale[7]), vl); + + vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux0_0, aux0_1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux1_0, aux1_1, vl), isum0, vl); + vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux2_0, aux2_1, vl), isum1, vl); + vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux3_0, aux3_1, vl), isum2, vl); + + sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); + + q3 += 32; q8 += 128; scale += 8; + + } + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + + sumf += d*sum_t; + + } + + *s = sumf; + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0x3); + const vector signed char lowMask1 = vec_splats((int8_t)0xf); + const vector signed char lowMask2 = vec_splats((int8_t)0x30); + const vector int v0 = vec_splats((int32_t)0); + const vector signed char v1 = vec_splats((signed char)0x1); + const vector unsigned char v2 = vec_splats((unsigned char)0x2); + const vector unsigned char v3 = vec_splats((unsigned char)0x3); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + const vector unsigned char v6 = vec_splats((unsigned char)0x6); + const vector signed char off = vec_splats((signed char)0x20); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + UNUSED(kmask1); + UNUSED(kmask2); + + vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); + vector signed char u1 = vec_and(u0, lowMask1); + vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); + vector signed char u3 = (vector signed char)vec_mergeh((vector signed int)u2, (vector signed int)vec_sr(u2, v2)); + vector signed char u30 = vec_sl(vec_and(u3, lowMask), v4); + vector signed char u31 = vec_and(u3, lowMask2); + + u1 = vec_or(u1, u30); + u2 = vec_or(vec_sr(u0, v4), u31); + + vector signed char vscales = (vector signed char)vec_mergeh((vector signed long long)u1, (vector signed long long)u2); + vector signed char qxhs0 = (vector signed char)vec_xl( 0, x[i].hmask); + vector signed char qxhs1 = (vector signed char)vec_xl(16, x[i].hmask); + + vscales = vec_sub(vscales, off); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + vector signed int vsumi4 = v0; + vector signed int vsumi5 = v0; + vector signed int vsumi6 = v0; + vector signed int vsumi7 = v0; + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/128; ++j) { + __builtin_prefetch(q3, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q3); + vector signed char qxs1 = (vector signed char)vec_xl(16, q3); + q3 += 32; + + //the low 2 bits + vector signed char qxs00 = vec_and(qxs0, lowMask); + vector signed char qxs01 = vec_and(vec_sr(qxs0, v2), lowMask); + vector signed char qxs02 = vec_and(vec_sr(qxs0, v4), lowMask); + vector signed char qxs03 = vec_and(vec_sr(qxs0, v6), lowMask); + vector signed char qxs10 = vec_and(qxs1, lowMask); + vector signed char qxs11 = vec_and(vec_sr(qxs1, v2), lowMask); + vector signed char qxs12 = vec_and(vec_sr(qxs1, v4), lowMask); + vector signed char qxs13 = vec_and(vec_sr(qxs1, v6), lowMask); + + //the 3rd bit + vector signed char qxh00 = vec_sl(vec_andc(v1, qxhs0), v2); + vector signed char qxh01 = vec_sl(vec_andc(v1, vec_sr(qxhs0, (vector unsigned char)v1)), v2); + vector signed char qxh02 = vec_sl(vec_andc(v1, vec_sr(qxhs0, v2)), v2); + vector signed char qxh03 = vec_sl(vec_andc(v1, vec_sr(qxhs0, v3)), v2); + vector signed char qxh10 = vec_sl(vec_andc(v1, qxhs1), v2); + vector signed char qxh11 = vec_sl(vec_andc(v1, vec_sr(qxhs1, (vector unsigned char)v1)), v2); + vector signed char qxh12 = vec_sl(vec_andc(v1, vec_sr(qxhs1, v2)), v2); + vector signed char qxh13 = vec_sl(vec_andc(v1, vec_sr(qxhs1, v3)), v2); + qxhs0 = vec_sr(qxhs0, v4); + qxhs1 = vec_sr(qxhs1, v4); + + vector signed char q3x00 = vec_sub(qxs00, qxh00); + vector signed char q3x01 = vec_sub(qxs01, qxh01); + vector signed char q3x02 = vec_sub(qxs02, qxh02); + vector signed char q3x03 = vec_sub(qxs03, qxh03); + vector signed char q3x10 = vec_sub(qxs10, qxh10); + vector signed char q3x11 = vec_sub(qxs11, qxh11); + vector signed char q3x12 = vec_sub(qxs12, qxh12); + vector signed char q3x13 = vec_sub(qxs13, qxh13); + + vector signed char q8y00 = vec_xl( 0, q8); + vector signed char q8y10 = vec_xl( 16, q8); + vector signed char q8y01 = vec_xl( 32, q8); + vector signed char q8y11 = vec_xl( 48, q8); + vector signed char q8y02 = vec_xl( 64, q8); + vector signed char q8y12 = vec_xl( 80, q8); + vector signed char q8y03 = vec_xl( 96, q8); + vector signed char q8y13 = vec_xl(112, q8); + q8 += 128; + + vector signed short vscales_h = vec_unpackh(vscales); + vector signed short vs0 = vec_splat(vscales_h, 0); + vector signed short vs1 = vec_splat(vscales_h, 1); + vector signed short vs2 = vec_splat(vscales_h, 2); + vector signed short vs3 = vec_splat(vscales_h, 3); + vector signed short vs4 = vec_splat(vscales_h, 4); + vector signed short vs5 = vec_splat(vscales_h, 5); + vector signed short vs6 = vec_splat(vscales_h, 6); + vector signed short vs7 = vec_splat(vscales_h, 7); + vscales = vec_sld(vscales, vscales, 8); + + vector signed short qv00 = vec_add(vec_mule(q3x00, q8y00), vec_mulo(q3x00, q8y00)); + vector signed short qv01 = vec_add(vec_mule(q3x01, q8y01), vec_mulo(q3x01, q8y01)); + vector signed short qv02 = vec_add(vec_mule(q3x02, q8y02), vec_mulo(q3x02, q8y02)); + vector signed short qv03 = vec_add(vec_mule(q3x03, q8y03), vec_mulo(q3x03, q8y03)); + vector signed short qv10 = vec_add(vec_mule(q3x10, q8y10), vec_mulo(q3x10, q8y10)); + vector signed short qv11 = vec_add(vec_mule(q3x11, q8y11), vec_mulo(q3x11, q8y11)); + vector signed short qv12 = vec_add(vec_mule(q3x12, q8y12), vec_mulo(q3x12, q8y12)); + vector signed short qv13 = vec_add(vec_mule(q3x13, q8y13), vec_mulo(q3x13, q8y13)); + + vsumi0 = vec_msum(qv00, vs0, vsumi0); + vsumi1 = vec_msum(qv01, vs2, vsumi1); + vsumi2 = vec_msum(qv02, vs4, vsumi2); + vsumi3 = vec_msum(qv03, vs6, vsumi3); + vsumi4 = vec_msum(qv10, vs1, vsumi4); + vsumi5 = vec_msum(qv11, vs3, vsumi5); + vsumi6 = vec_msum(qv12, vs5, vsumi6); + vsumi7 = vec_msum(qv13, vs7, vsumi7); + } + + vsumi0 = vec_add(vsumi0, vsumi4); + vsumi1 = vec_add(vsumi1, vsumi5); + vsumi2 = vec_add(vsumi2, vsumi6); + vsumi3 = vec_add(vsumi3, vsumi7); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined __loongarch_asx + + const __m256i m3 = __lasx_xvreplgr2vr_b(3); + const __m256i mone = __lasx_xvreplgr2vr_b(1); + const __m128i m32 = __lsx_vreplgr2vr_b(32); + + __m256 acc = (__m256)__lasx_xvldi(0); + + uint32_t aux[3]; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + // Set up scales + memcpy(aux, x[i].scales, 12); + __m128i scales128 = lsx_set_w( + ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), + ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), + (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), + (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); + scales128 = __lsx_vsub_b(scales128, m32); + const __m256i all_scales = lasx_ext8_16(scales128); + const __m128i l_scales = lasx_extracti128(all_scales, 0); + const __m128i h_scales = lasx_extracti128(all_scales, 1); + const __m256i scales[2] = {lasx_insertf128(l_scales, l_scales), lasx_insertf128(h_scales, h_scales)}; + + // high bit + const __m256i hbits = __lasx_xvld((const __m256i*)x[i].hmask, 0); + + // integer accumulator + __m256i sumi = __lasx_xvldi(0); + + int bit = 0; + int is = 0; + __m256i xvbit; + + + for (int j = 0; j < QK_K/128; ++j) { + // load low 2 bits + const __m256i q3bits = __lasx_xvld((const __m256i*)q3, 0); q3 += 32; + + xvbit = __lasx_xvreplgr2vr_h(bit); + // prepare low and high bits + const __m256i q3l_0 = __lasx_xvand_v(q3bits, m3); + const __m256i q3h_0 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); + ++bit; + + xvbit = __lasx_xvreplgr2vr_h(bit); + const __m256i q3l_1 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 2), m3); + const __m256i q3h_1 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); + ++bit; + + xvbit = __lasx_xvreplgr2vr_h(bit); + const __m256i q3l_2 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 4), m3); + const __m256i q3h_2 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); + ++bit; + + xvbit = __lasx_xvreplgr2vr_h(bit); + const __m256i q3l_3 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 6), m3); + const __m256i q3h_3 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); + ++bit; + + // load Q8 quants + const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use lasx_maddubs_h, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + __m256i q8s_0 = lasx_maddubs_h(q3h_0, q8_0); + __m256i q8s_1 = lasx_maddubs_h(q3h_1, q8_1); + __m256i q8s_2 = lasx_maddubs_h(q3h_2, q8_2); + __m256i q8s_3 = lasx_maddubs_h(q3h_3, q8_3); + + __m256i p16_0 = lasx_maddubs_h(q3l_0, q8_0); + __m256i p16_1 = lasx_maddubs_h(q3l_1, q8_1); + __m256i p16_2 = lasx_maddubs_h(q3l_2, q8_2); + __m256i p16_3 = lasx_maddubs_h(q3l_3, q8_3); + + p16_0 = __lasx_xvsub_h(p16_0, q8s_0); + p16_1 = __lasx_xvsub_h(p16_1, q8s_1); + p16_2 = __lasx_xvsub_h(p16_2, q8s_2); + p16_3 = __lasx_xvsub_h(p16_3, q8s_3); + + // multiply with scales + p16_0 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); + p16_1 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); + p16_2 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); + p16_3 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); + + // accumulate + p16_0 = __lasx_xvadd_w(p16_0, p16_1); + p16_2 = __lasx_xvadd_w(p16_2, p16_3); + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_2)); + } + // multiply with block scale and accumulate + acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc);//FIXME + } + + *s = hsum_float_8(acc); + +#else + // scalar version + // This function is written like this so the compiler can manage to vectorize most of it + // Using -Ofast, GCC and clang manage to produce code that is within a factor of 2 or so from the + // manually vectorized version above. Every other version I tried would run at least 4 times slower. + // The ideal situation would be if we could just write the code once, and the compiler would + // automatically produce the best possible set of machine instructions, instead of us having to manually + // write vectorized versions for AVX, ARM_NEON, etc. + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + uint32_t auxs[4]; + const int8_t * scales = (const int8_t*)auxs; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict hm = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + uint8_t m = 1; + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) a[l] = q3[l] & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 2) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 4) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 6) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + q3 += 32; + } + a = aux8; + + memcpy(auxs, x[i].scales, 12); + uint32_t tmp = auxs[2]; + auxs[2] = ((auxs[0] >> 4) & kmask2) | (((tmp >> 4) & kmask1) << 4); + auxs[3] = ((auxs[1] >> 4) & kmask2) | (((tmp >> 6) & kmask1) << 4); + auxs[0] = (auxs[0] & kmask2) | (((tmp >> 0) & kmask1) << 4); + auxs[1] = (auxs[1] & kmask2) | (((tmp >> 2) & kmask1) << 4); + for (int j = 0; j < QK_K/16; ++j) { + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; + +#endif + +} + +void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q4_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + + static const uint32_t kmask1 = 0x3f3f3f3f; + static const uint32_t kmask2 = 0x0f0f0f0f; + static const uint32_t kmask3 = 0x03030303; + + uint32_t utmp[4]; + +#ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + const int32x4_t mzero = vdupq_n_s32(0); + + ggml_int8x16x2_t q4bytes; + ggml_int8x16x2_t q8bytes; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); + + memcpy(utmp, x[i].scales, 12); + + uint32x2_t mins8 = { 0 }; + mins8 = vset_lane_u32(utmp[1] & kmask1, mins8, 0); + mins8 = vset_lane_u32(((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4), mins8, 1); + + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[0] &= kmask1; + + const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(vreinterpret_u8_u32(mins8))); + const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), + vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); + sumf -= dmin * vaddvq_s32(prod); + + const uint8_t * scales = (const uint8_t *)utmp; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + int32_t sumi1 = 0; + int32_t sumi2 = 0; + + for (int j = 0; j < QK_K/64; ++j) { + const ggml_uint8x16x2_t q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; + + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); + q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + sumi1 += vaddvq_s32(p1) * scales[2*j+0]; + + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); + q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); + + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + + sumi2 += vaddvq_s32(p2) * scales[2*j+1]; + } + + sumf += d * (sumi1 + sumi2); + + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + + __m256 acc = _mm256_setzero_ps(); + __m128 acc_m = _mm_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); + const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); + const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); + acc_m = _mm_fmadd_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod), acc_m); + + const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); + const __m256i scales = MM256_SET_M128I(sc128, sc128); + + __m256i sumi = _mm256_setzero_si256(); + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_l = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_h = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4l = _mm256_and_si256(q4bits, m4); + const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); + + const __m256i q8l = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); + p16l = _mm256_madd_epi16(scale_l, p16l); + + const __m256i q8h = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); + p16h = _mm256_madd_epi16(scale_h, p16h); + const __m256i sumj = _mm256_add_epi32(p16l, p16h); + + sumi = _mm256_add_epi32(sumi, sumj); + } + + __m256 vd = _mm256_set1_ps(d); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); + + } + + acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); + acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); + + *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + +#elif defined __AVX__ + + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i m2 = _mm_set1_epi8(0x2); + + __m256 acc = _mm256_setzero_ps(); + __m128 acc_m = _mm_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); + const __m128i scales = _mm_cvtepu8_epi16(utmps); + const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); + + const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); + const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); + const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); + const __m128i prod = _mm_madd_epi16(mins, q8s); + acc_m = _mm_add_ps(_mm_mul_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod)), acc_m); + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + __m128i shuffle = _mm_set1_epi16(0x0100); + for (int j = 0; j < QK_K/64; ++j) { + + const __m128i scale_l = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + const __m128i scale_h = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + + __m128i q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4l_0 = _mm_and_si128(q4bits, m4); + const __m128i q4h_0 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); + q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4l_1 = _mm_and_si128(q4bits, m4); + const __m128i q4h_1 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); + + const __m128i q8l_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16l = _mm_maddubs_epi16(q4l_0, q8l_0); + p16l = _mm_madd_epi16(scale_l, p16l); + sumi_0 = _mm_add_epi32(sumi_0, p16l); + const __m128i q8l_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + p16l = _mm_maddubs_epi16(q4l_1, q8l_1); + p16l = _mm_madd_epi16(scale_l, p16l); + sumi_1 = _mm_add_epi32(sumi_1, p16l); + + const __m128i q8h_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16h = _mm_maddubs_epi16(q4h_0, q8h_0); + p16h = _mm_madd_epi16(scale_h, p16h); + sumi_0 = _mm_add_epi32(sumi_0, p16h); + const __m128i q8h_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + p16h = _mm_maddubs_epi16(q4h_1, q8h_1); + p16h = _mm_madd_epi16(scale_h, p16h); + sumi_1 = _mm_add_epi32(sumi_1, p16h); + + } + + __m256 vd = _mm256_set1_ps(d); + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); + + } + + acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); + acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); + + *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + +#elif defined __riscv_v_intrinsic + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + size_t vl = 8; + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); + vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); + vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); + vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); + vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + + vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + vl = 32; + + int32_t sum_1 = 0; + int32_t sum_2 = 0; + + vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + + for (int j = 0; j < QK_K/64; ++j) { + // load Q4 + vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); + + // load Q8 and multiply it with lower Q4 nibble + vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q4_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); + vint16m2_t qv_0 = __riscv_vwmul_vv_i16m2(q4_0, q8_0, vl); + vint16m1_t vs_0 = __riscv_vredsum_vs_i16m2_i16m1(qv_0, vzero, vl); + + sum_1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[2*j+0]; + + // load Q8 and multiply it with upper Q4 nibble + vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); + vint8m1_t q4_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); + vint16m2_t qv_1 = __riscv_vwmul_vv_i16m2(q4_1, q8_1, vl); + vint16m1_t vs_1 = __riscv_vredsum_vs_i16m2_i16m1(qv_1, vzero, vl); + + sum_2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[2*j+1]; + + q4 += 32; q8 += 64; + + } + + sumf += d*(sum_1 + sum_2); + + } + + *s = sumf; + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed char lowMask1 = vec_splats((int8_t)0x3f); + const vector signed char lowMask2 = vec_splats((int8_t)0x30); + const vector int v0 = vec_splats((int32_t)0); + const vector unsigned char v2 = vec_splats((uint8_t)2); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); + vector float vdmin = vec_mul(vxmin, vyd); + + vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); + vector signed short q8ysums1 = vec_xl(16, y[i].bsums); + + UNUSED(kmask1); + UNUSED(kmask2); + UNUSED(kmask3); + UNUSED(utmp); + + vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); + vector signed char u1 = vec_and(vec_sr(u0, v2), lowMask2); + vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); + vector signed char u3 = vec_sr(u2, v4); + + vector signed char u30 = u1; + vector signed char u31 = (vector signed char)vec_mergeh((vector signed int)vec_and(u2, lowMask), (vector signed int)u3); + + u1 = vec_and(u0, lowMask1); + u2 = vec_or(u30, u31); + + vector signed char utmps = (vector signed char)vec_mergeh((vector signed int)u1, (vector signed int)u2); + + vector signed short vscales = vec_unpackh(utmps); + vector signed short q4xmins = vec_unpackl(utmps); + vector signed short q4xmins0 = vec_mergeh(q4xmins, q4xmins); + vector signed short q4xmins1 = vec_mergel(q4xmins, q4xmins); + + vector signed int prod0 = vec_mule(q4xmins0, q8ysums0); + vector signed int prod1 = vec_mule(q4xmins1, q8ysums1); + vector signed int prod2 = vec_mulo(q4xmins0, q8ysums0); + vector signed int prod3 = vec_mulo(q4xmins1, q8ysums1); + + vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); + vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); + vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); + vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/64; j+=2) { + __builtin_prefetch(q4, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q4); + vector signed char qxs1 = (vector signed char)vec_xl(16, q4); + vector signed char qxs2 = (vector signed char)vec_xl(32, q4); + vector signed char qxs3 = (vector signed char)vec_xl(48, q4); + q4 += 64; + + vector unsigned char q4x00 = (vector unsigned char)vec_and(qxs0, lowMask); + vector unsigned char q4x01 = (vector unsigned char)vec_sr(qxs0, v4); + vector unsigned char q4x10 = (vector unsigned char)vec_and(qxs1, lowMask); + vector unsigned char q4x11 = (vector unsigned char)vec_sr(qxs1, v4); + vector unsigned char q4x20 = (vector unsigned char)vec_and(qxs2, lowMask); + vector unsigned char q4x21 = (vector unsigned char)vec_sr(qxs2, v4); + vector unsigned char q4x30 = (vector unsigned char)vec_and(qxs3, lowMask); + vector unsigned char q4x31 = (vector unsigned char)vec_sr(qxs3, v4); + + vector signed char q8y00 = vec_xl( 0, q8); + vector signed char q8y10 = vec_xl( 16, q8); + vector signed char q8y01 = vec_xl( 32, q8); + vector signed char q8y11 = vec_xl( 48, q8); + vector signed char q8y20 = vec_xl( 64, q8); + vector signed char q8y30 = vec_xl( 80, q8); + vector signed char q8y21 = vec_xl( 96, q8); + vector signed char q8y31 = vec_xl(112, q8); + q8 += 128; + + vector signed int qv00 = vec_msum(q8y00, q4x00, v0); + vector signed int qv01 = vec_msum(q8y01, q4x01, v0); + vector signed int qv10 = vec_msum(q8y10, q4x10, v0); + vector signed int qv11 = vec_msum(q8y11, q4x11, v0); + vector signed int qv20 = vec_msum(q8y20, q4x20, v0); + vector signed int qv21 = vec_msum(q8y21, q4x21, v0); + vector signed int qv30 = vec_msum(q8y30, q4x30, v0); + vector signed int qv31 = vec_msum(q8y31, q4x31, v0); + + vector signed int vscales_h = vec_unpackh(vscales); + vector signed int vs0 = vec_splat(vscales_h, 0); + vector signed int vs1 = vec_splat(vscales_h, 1); + vector signed int vs2 = vec_splat(vscales_h, 2); + vector signed int vs3 = vec_splat(vscales_h, 3); + vscales = vec_sld(vscales, vscales, 8); + + vsumi0 = vec_add(vec_mul(qv00, vs0), vsumi0); + vsumi1 = vec_add(vec_mul(qv01, vs1), vsumi1); + vsumi2 = vec_add(vec_mul(qv20, vs2), vsumi2); + vsumi3 = vec_add(vec_mul(qv21, vs3), vsumi3); + + vsumi0 = vec_add(vec_mul(qv10, vs0), vsumi0); + vsumi1 = vec_add(vec_mul(qv11, vs1), vsumi1); + vsumi2 = vec_add(vec_mul(qv30, vs2), vsumi2); + vsumi3 = vec_add(vec_mul(qv31, vs3), vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined __loongarch_asx + GGML_UNUSED(kmask1); + GGML_UNUSED(kmask2); + GGML_UNUSED(kmask3); + + const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); + + __m256 acc = (__m256)__lasx_xvldi(0); + __m128 acc_m = (__m128)__lsx_vldi(0); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m256i mins_and_scales = lasx_extu8_16(lsx_set_w(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = __lasx_xvld((const __m256i*)y[i].bsums, 0); + const __m128i q8s = lsx_hadd_h(lasx_extracti128(q8sums, 0), lasx_extracti128(q8sums, 1)); + const __m128i prod = lsx_madd_h(lasx_extracti128(mins_and_scales, 1), q8s); + acc_m = __lsx_vfmadd_s(__lsx_vreplfr2vr_s(dmin), __lsx_vffint_s_w(prod), acc_m); + + const __m128i sc128 = lasx_extracti128(mins_and_scales, 0); + const __m256i scales = lasx_insertf128(sc128, sc128); + + __m256i sumi = __lasx_xvldi(0); + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_l = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_h = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q4bits = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; + const __m256i q4l = __lasx_xvand_v(q4bits, m4); + const __m256i q4h = __lasx_xvand_v(__lasx_xvsrli_h(q4bits, 4), m4); + + const __m256i q8l = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + __m256i p16l = lasx_maddubs_h(q4l, q8l); + p16l = lasx_madd_h(scale_l, p16l); + + const __m256i q8h = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + __m256i p16h = lasx_maddubs_h(q4h, q8h); + p16h = lasx_madd_h(scale_h, p16h); + const __m256i sumj = __lasx_xvadd_w(p16l, p16h); + + sumi = __lasx_xvadd_w(sumi, sumj); + } + + __m256 vd = __lasx_xvreplfr2vr_s(d); + acc = __lasx_xvfmadd_s(vd, __lasx_xvffint_s_w(sumi), acc); + + } + + acc_m = __lsx_vfadd_s(acc_m, (__m128)__lsx_vpermi_w((__m128i)acc_m, (__m128i)acc_m, 0xee)); + __m128i tmp1 = __lsx_vinsgr2vr_w(__lsx_vldi(0), __lsx_vpickve2gr_w((__m128i)acc_m, 1), 0); + acc_m = __lsx_vfadd_s(acc_m, (__m128)tmp1); + + + ft_union fi; + fi.i = __lsx_vpickve2gr_w(acc_m, 0); + *s = hsum_float_8(acc) + fi.f ; +#else + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + for (int j = 0; j < QK_K/64; ++j) { + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); + a += 32; + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); + a += 32; q4 += 32; + } + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + int sumi = 0; + for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; + a = aux8; + int is = 0; + for (int j = 0; j < QK_K/32; ++j) { + int32_t scale = scales[is++]; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + sumf -= dmin * sumi; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; +#endif +} + +void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q5_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + + static const uint32_t kmask1 = 0x3f3f3f3f; + static const uint32_t kmask2 = 0x0f0f0f0f; + static const uint32_t kmask3 = 0x03030303; + + uint32_t utmp[4]; + +#ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + const uint8x16_t mone = vdupq_n_u8(1); + const uint8x16_t mtwo = vdupq_n_u8(2); + const int32x4_t mzero = vdupq_n_s32(0); + + ggml_int8x16x4_t q5bytes; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const uint8x8_t mins8 = vld1_u8((const uint8_t*)utmp + 8); + const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(mins8)); + const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), + vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); + int32_t sumi_mins = vaddvq_s32(prod); + + const uint8_t * scales = (const uint8_t *)utmp; + + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); + + ggml_uint8x16x4_t q5h; + + int32_t sumi = 0; + + for (int j = 0; j < QK_K/64; ++j) { + + const ggml_uint8x16x2_t q5bits = ggml_vld1q_u8_x2(q5); q5 += 32; + const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; + + q5h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); + q5h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); + q5h.val[2] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[0]), 3); + q5h.val[3] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[1]), 3); + qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 2); + qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 2); + + q5bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[0], m4b), q5h.val[0])); + q5bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[1], m4b), q5h.val[1])); + q5bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[0], 4), q5h.val[2])); + q5bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[1], 4), q5h.val[3])); + + sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0]), q5bytes.val[1], q8bytes.val[1])) * *scales++; + sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2]), q5bytes.val[3], q8bytes.val[3])) * *scales++; + } + + sumf += d * sumi - dmin * sumi_mins; + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + const __m128i mzero = _mm_setzero_si128(); + const __m256i mone = _mm256_set1_epi8(1); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0.f; + + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); + const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); + const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); + const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); + summs += dmin * _mm_extract_epi32(hsum, 0); + + const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); + const __m256i scales = MM256_SET_M128I(sc128, sc128); + + const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].qh); + __m256i hmask = mone; + + __m256i sumi = _mm256_setzero_si256(); + + int bit = 0; + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_0 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_1 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); q5 += 32; + + const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); + const __m256i q5h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); + const __m256i q5_0 = _mm256_add_epi8(q5l_0, q5h_0); + hmask = _mm256_slli_epi16(hmask, 1); + + const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); + const __m256i q5h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); + const __m256i q5_1 = _mm256_add_epi8(q5l_1, q5h_1); + hmask = _mm256_slli_epi16(hmask, 1); + + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + __m256i p16_0 = _mm256_maddubs_epi16(q5_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q5_1, q8_1); + + p16_0 = _mm256_madd_epi16(scale_0, p16_0); + p16_1 = _mm256_madd_epi16(scale_1, p16_1); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + + } + + __m256 vd = _mm256_set1_ps(d); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); + + } + + *s = hsum_float_8(acc) + summs; + +#elif defined __AVX__ + + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i mzero = _mm_setzero_si128(); + const __m128i mone = _mm_set1_epi8(1); + const __m128i m2 = _mm_set1_epi8(2); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0.f; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); + const __m128i scales = _mm_cvtepu8_epi16(utmps); + const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); + + const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); + const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); + const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); + const __m128i prod = _mm_madd_epi16(mins, q8s); + const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); + summs += dmin * _mm_extract_epi32(hsum, 0); + + const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].qh[0]); + const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].qh[16]); + __m128i hmask = mone; + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + int bit = 0; + + __m128i shuffle = _mm_set1_epi16(0x0100); + for (int j = 0; j < QK_K/64; ++j) { + + const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + + const __m128i q5bits_0 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; + const __m128i q5bits_1 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; + + __m128i q5l_0 = _mm_and_si128(q5bits_0, m4); + __m128i q5l_1 = _mm_and_si128(q5bits_1, m4); + __m128i q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); + __m128i q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); + __m128i q5_0 = _mm_add_epi8(q5l_0, q5h_0); + __m128i q5_1 = _mm_add_epi8(q5l_1, q5h_1); + hmask = _mm_slli_epi16(hmask, 1); + + __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16_0 = _mm_maddubs_epi16(q5_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q5_1, q8_1); + p16_0 = _mm_madd_epi16(scale_0, p16_0); + p16_1 = _mm_madd_epi16(scale_0, p16_1); + + q5l_0 = _mm_and_si128(_mm_srli_epi16(q5bits_0, 4), m4); + q5l_1 = _mm_and_si128(_mm_srli_epi16(q5bits_1, 4), m4); + q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); + q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); + q5_0 = _mm_add_epi8(q5l_0, q5h_0); + q5_1 = _mm_add_epi8(q5l_1, q5h_1); + hmask = _mm_slli_epi16(hmask, 1); + + q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16_2 = _mm_maddubs_epi16(q5_0, q8_0); + __m128i p16_3 = _mm_maddubs_epi16(q5_1, q8_1); + p16_2 = _mm_madd_epi16(scale_1, p16_2); + p16_3 = _mm_madd_epi16(scale_1, p16_3); + + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + + } + + __m256 vd = _mm256_set1_ps(d); + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); + + } + + *s = hsum_float_8(acc) + summs; + +#elif defined __riscv_v_intrinsic + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + float sumf = 0; + float sums = 0.0; + + size_t vl; + + for (int i = 0; i < nb; ++i) { + + vl = 8; + + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict hm = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + + vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); + vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); + vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); + vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); + vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + + vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + + vl = 32; + int32_t aux32 = 0; + int is = 0; + + uint8_t m = 1; + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t vqh = __riscv_vle8_v_u8m1(hm, vl); + + for (int j = 0; j < QK_K/64; ++j) { + // load Q5 and Q8 + vuint8m1_t q5_x = __riscv_vle8_v_u8m1(q5, vl); + vint8m1_t q8_y1 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q8_y2 = __riscv_vle8_v_i8m1(q8+32, vl); + + // compute mask for addition + vint8m1_t q5_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q5_x, 0x0F, vl)); + vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_1 = __riscv_vmsne_vx_u8m1_b8(qh_m1, 0, vl); + vint8m1_t q5_m1 = __riscv_vadd_vx_i8m1_mu(vmask_1, q5_a, q5_a, 16, vl); + m <<= 1; + + vint8m1_t q5_l = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q5_x, 0x04, vl)); + vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_2 = __riscv_vmsne_vx_u8m1_b8(qh_m2, 0, vl); + vint8m1_t q5_m2 = __riscv_vadd_vx_i8m1_mu(vmask_2, q5_l, q5_l, 16, vl); + m <<= 1; + + vint16m2_t v0 = __riscv_vwmul_vv_i16m2(q5_m1, q8_y1, vl); + vint16m2_t v1 = __riscv_vwmul_vv_i16m2(q5_m2, q8_y2, vl); + + vint32m4_t vs1 = __riscv_vwmul_vx_i32m4(v0, scales[is++], vl); + vint32m4_t vs2 = __riscv_vwmul_vx_i32m4(v1, scales[is++], vl); + + vint32m1_t vacc1 = __riscv_vredsum_vs_i32m4_i32m1(vs1, vzero, vl); + vint32m1_t vacc2 = __riscv_vredsum_vs_i32m4_i32m1(vs2, vzero, vl); + + aux32 += __riscv_vmv_x_s_i32m1_i32(vacc1) + __riscv_vmv_x_s_i32m1_i32(vacc2); + q5 += 32; q8 += 64; + + } + + vfloat32m1_t vaux = __riscv_vfmul_vf_f32m1(__riscv_vfmv_v_f_f32m1(aux32, 1), d, 1); + sums += __riscv_vfmv_f_s_f32m1_f32(vaux); + + } + + *s = sumf+sums; + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed char lowMask1 = vec_splats((int8_t)0x3f); + const vector signed char lowMask2 = vec_splats((int8_t)0x30); + const vector int v0 = vec_splats((int32_t)0); + const vector unsigned char v1 = vec_splats((unsigned char)0x1); + const vector unsigned char v2 = vec_splats((unsigned char)0x2); + const vector unsigned char v3 = vec_splats((unsigned char)0x3); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); + vector float vdmin = vec_mul(vxmin, vyd); + + UNUSED(kmask1); + UNUSED(kmask2); + UNUSED(kmask3); + UNUSED(utmp); + + vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); + vector signed char u1 = vec_and(vec_sr(u0, v2), lowMask2); + vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); + vector signed char u3 = vec_sr(u2, v4); + + vector signed char u30 = u1; + vector signed char u31 = (vector signed char)vec_mergeh((vector signed int)vec_and(u2, lowMask), (vector signed int)u3); + + u1 = vec_and(u0, lowMask1); + u2 = vec_or(u30, u31); + + vector signed char utmps = (vector signed char)vec_mergeh((vector signed int)u1, (vector signed int)u2); + + vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); + vector signed short q8ysums1 = vec_xl(16, y[i].bsums); + + vector signed short vscales = vec_unpackh(utmps); + + vector signed short q5xmins = vec_unpackl(utmps); + vector signed short q5xmins0 = vec_mergeh(q5xmins, q5xmins); + vector signed short q5xmins1 = vec_mergel(q5xmins, q5xmins); + + vector signed int prod0 = vec_mule(q5xmins0, q8ysums0); + vector signed int prod1 = vec_mule(q5xmins1, q8ysums1); + vector signed int prod2 = vec_mulo(q5xmins0, q8ysums0); + vector signed int prod3 = vec_mulo(q5xmins1, q8ysums1); + + vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); + vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); + vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); + vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); + + vector signed char qxhs0 = (vector signed char)vec_xl( 0, x[i].qh); + vector signed char qxhs1 = (vector signed char)vec_xl(16, x[i].qh); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/64; ++j) { + __builtin_prefetch(q5, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q5); + vector signed char qxs1 = (vector signed char)vec_xl(16, q5); + q5 += 32; + + vector signed char qxs00 = vec_and(qxs0, lowMask); + vector signed char qxs01 = vec_sr(qxs0, v4); + vector signed char qxs10 = vec_and(qxs1, lowMask); + vector signed char qxs11 = vec_sr(qxs1, v4); + + vector signed char q5h00 = vec_sl(vec_and((vector signed char)v1, qxhs0), v4); + vector signed char q5h01 = vec_sl(vec_and((vector signed char)v2, qxhs0), v3); + vector signed char q5h10 = vec_sl(vec_and((vector signed char)v1, qxhs1), v4); + vector signed char q5h11 = vec_sl(vec_and((vector signed char)v2, qxhs1), v3); + qxhs0 = vec_sr(qxhs0, v2); + qxhs1 = vec_sr(qxhs1, v2); + + vector unsigned char q5x00 = (vector unsigned char)vec_or(q5h00, qxs00); + vector unsigned char q5x01 = (vector unsigned char)vec_or(q5h01, qxs01); + vector unsigned char q5x10 = (vector unsigned char)vec_or(q5h10, qxs10); + vector unsigned char q5x11 = (vector unsigned char)vec_or(q5h11, qxs11); + + vector signed char q8y00 = vec_xl( 0, q8); + vector signed char q8y10 = vec_xl(16, q8); + vector signed char q8y01 = vec_xl(32, q8); + vector signed char q8y11 = vec_xl(48, q8); + q8 += 64; + + vector signed int qv00 = vec_msum(q8y00, q5x00, v0); + vector signed int qv01 = vec_msum(q8y01, q5x01, v0); + vector signed int qv10 = vec_msum(q8y10, q5x10, v0); + vector signed int qv11 = vec_msum(q8y11, q5x11, v0); + + vector signed int vscales_h = vec_unpackh(vscales); + vector signed int vs0 = vec_splat(vscales_h, 0); + vector signed int vs1 = vec_splat(vscales_h, 1); + vscales = vec_sld(vscales, vscales, 12); + + vsumi0 = vec_add(vec_mul(qv00, vs0), vsumi0); + vsumi1 = vec_add(vec_mul(qv10, vs0), vsumi1); + vsumi2 = vec_add(vec_mul(qv01, vs1), vsumi2); + vsumi3 = vec_add(vec_mul(qv11, vs1), vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined __loongarch_asx + GGML_UNUSED(kmask1); + GGML_UNUSED(kmask2); + GGML_UNUSED(kmask3); + + const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); + const __m128i mzero = __lsx_vldi(0); + const __m256i mone = __lasx_xvreplgr2vr_b(1); + + __m256 acc = (__m256)__lasx_xvldi(0); + + float summs = 0.f; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m256i mins_and_scales = lasx_extu8_16(lsx_set_w(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = __lasx_xvld((const __m256i*)y[i].bsums, 0); + const __m128i q8s = lsx_hadd_h(lasx_extracti128(q8sums, 0), lasx_extracti128(q8sums, 1)); + const __m128i prod = lsx_madd_h(lasx_extracti128(mins_and_scales, 1), q8s); + const __m128i hsum = lsx_hadd_w(lsx_hadd_w(prod, mzero), mzero); + summs += dmin * __lsx_vpickve2gr_w(hsum, 0); //TODO check + + const __m128i sc128 = lasx_extracti128(mins_and_scales, 0); + const __m256i scales = lasx_insertf128(sc128, sc128); + + const __m256i hbits = __lasx_xvld((const __m256i*)x[i].qh, 0); + __m256i hmask = mone; + + __m256i sumi = __lasx_xvldi(0); + + int bit = 0; + __m256i xvbit; + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_0 = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_1 = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q5bits = __lasx_xvld((const __m256i*)q5, 0); q5 += 32; + + xvbit = __lasx_xvreplgr2vr_h(bit++); + const __m256i q5l_0 = __lasx_xvand_v(q5bits, m4); + const __m256i q5h_0 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvand_v(hbits, hmask), xvbit), 4); + const __m256i q5_0 = __lasx_xvadd_b(q5l_0, q5h_0); + hmask = __lasx_xvslli_h(hmask, 1); + + xvbit = __lasx_xvreplgr2vr_h(bit++); + const __m256i q5l_1 = __lasx_xvand_v(__lasx_xvsrli_h(q5bits, 4), m4); + const __m256i q5h_1 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvand_v(hbits, hmask), xvbit), 4); + const __m256i q5_1 = __lasx_xvadd_b(q5l_1, q5h_1); + hmask = __lasx_xvslli_h(hmask, 1); + + const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + + __m256i p16_0 = lasx_maddubs_h(q5_0, q8_0); + __m256i p16_1 = lasx_maddubs_h(q5_1, q8_1); + + p16_0 = lasx_madd_h(scale_0, p16_0); + p16_1 = lasx_madd_h(scale_1, p16_1); + + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_1)); + + } + + __m256 vd = __lasx_xvreplfr2vr_s(d); + acc = __lasx_xvfmadd_s(vd, __lasx_xvffint_s_w(sumi), acc); + + } + + *s = hsum_float_8(acc) + summs; + +#else + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict hm = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + uint8_t m = 1; + for (int j = 0; j < QK_K/64; ++j) { + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); + for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); + for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); + a += 32; m <<= 1; + q4 += 32; + } + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + int sumi = 0; + for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; + a = aux8; + int is = 0; + for (int j = 0; j < QK_K/32; ++j) { + int32_t scale = scales[is++]; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + sumf -= dmin * sumi; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; +#endif +} + +void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q6_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON + float sum = 0; + + const uint8x16_t m4b = vdupq_n_u8(0xF); + const int32x4_t vzero = vdupq_n_s32(0); + //const int8x16_t m32s = vdupq_n_s8(32); + + const uint8x16_t mone = vdupq_n_u8(3); + + ggml_int8x16x4_t q6bytes; + ggml_uint8x16x4_t q6h; + + for (int i = 0; i < nb; ++i) { + + const float d_all = GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const int8_t * restrict scale = x[i].scales; + + const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); + const int8x16_t scales = vld1q_s8(scale); + const ggml_int16x8x2_t q6scales = {{vmovl_s8(vget_low_s8(scales)), vmovl_s8(vget_high_s8(scales))}}; + + const int32x4_t prod = vaddq_s32(vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[0]), vget_low_s16 (q6scales.val[0])), + vmull_s16(vget_high_s16(q8sums.val[0]), vget_high_s16(q6scales.val[0]))), + vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[1]), vget_low_s16 (q6scales.val[1])), + vmull_s16(vget_high_s16(q8sums.val[1]), vget_high_s16(q6scales.val[1])))); + int32_t isum_mins = vaddvq_s32(prod); + + int32_t isum = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); qh += 32; + ggml_uint8x16x4_t q6bits = ggml_vld1q_u8_x4(q6); q6 += 64; + ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; + + q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); + q6h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); + uint8x16_t shifted = vshrq_n_u8(qhbits.val[0], 2); + q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 2); + q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + + //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); + //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); + //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])), m32s); + //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])), m32s); + q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])); + q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])); + q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])); + q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])); + + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; + + scale += 4; + + q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; + + shifted = vshrq_n_u8(qhbits.val[0], 4); + q6h.val[0] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 4); + q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[0], 6); + q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 6); + q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + + //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])), m32s); + //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])), m32s); + //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])), m32s); + //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])), m32s); + q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])); + q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])); + q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])); + q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])); + + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; + scale += 4; + } + //sum += isum * d_all * y[i].d; + sum += d_all * y[i].d * (isum - 32 * isum_mins); + + } + *s = sum; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + const __m256i m2 = _mm256_set1_epi8(3); + const __m256i m32s = _mm256_set1_epi8(32); + + __m256 acc = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); + + __m256i sumi = _mm256_setzero_si256(); + + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); + const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); + const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); + const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); + is += 4; + + const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4bits2 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4bitsH = _mm256_loadu_si256((const __m256i*)qh); qh += 32; + + const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(q4bitsH, m2), 4); + const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 2), m2), 4); + const __m256i q4h_2 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 4), m2), 4); + const __m256i q4h_3 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 6), m2), 4); + + const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); + const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(q4bits2, m4), q4h_1); + const __m256i q4_2 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_2); + const __m256i q4_3 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits2, 4), m4), q4h_3); + + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); + __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); + __m256i q8s_2 = _mm256_maddubs_epi16(m32s, q8_2); + __m256i q8s_3 = _mm256_maddubs_epi16(m32s, q8_3); + + __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); + __m256i p16_2 = _mm256_maddubs_epi16(q4_2, q8_2); + __m256i p16_3 = _mm256_maddubs_epi16(q4_3, q8_3); + + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + p16_2 = _mm256_sub_epi16(p16_2, q8s_2); + p16_3 = _mm256_sub_epi16(p16_3, q8s_3); + + p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); + p16_2 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_2), p16_2); + p16_3 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_3), p16_3); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_2, p16_3)); + + } + + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(3); + const __m128i m15 = _mm_set1_epi8(15); + + __m256 acc = _mm256_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + // handle the q6_k -32 offset separately using bsums + const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)y[i].bsums); + const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)y[i].bsums + 1); + const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); + const __m128i scales_16_0 = _mm_cvtepi8_epi16(scales); + const __m128i scales_16_1 = _mm_cvtepi8_epi16(_mm_bsrli_si128(scales, 8)); + const __m128i q8sclsub_0 = _mm_slli_epi32(_mm_madd_epi16(q8sums_0, scales_16_0), 5); + const __m128i q8sclsub_1 = _mm_slli_epi32(_mm_madd_epi16(q8sums_1, scales_16_1), 5); + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + const __m128i q4bitsH_0 = _mm_loadu_si128((const __m128i*)qh); qh += 16; + const __m128i q4bitsH_1 = _mm_loadu_si128((const __m128i*)qh); qh += 16; + + const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, m3), 4); + const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, m3), 4); + const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, _mm_set1_epi8(12)), 2); + const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, _mm_set1_epi8(12)), 2); + const __m128i q4h_4 = _mm_and_si128(q4bitsH_0, _mm_set1_epi8(48)); + const __m128i q4h_5 = _mm_and_si128(q4bitsH_1, _mm_set1_epi8(48)); + const __m128i q4h_6 = _mm_srli_epi16(_mm_and_si128(q4bitsH_0, _mm_set1_epi8(-64)), 2); + const __m128i q4h_7 = _mm_srli_epi16(_mm_and_si128(q4bitsH_1, _mm_set1_epi8(-64)), 2); + + const __m128i q4bits1_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits1_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits2_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits2_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + + const __m128i q4_0 = _mm_or_si128(_mm_and_si128(q4bits1_0, m15), q4h_0); + const __m128i q4_1 = _mm_or_si128(_mm_and_si128(q4bits1_1, m15), q4h_1); + const __m128i q4_2 = _mm_or_si128(_mm_and_si128(q4bits2_0, m15), q4h_2); + const __m128i q4_3 = _mm_or_si128(_mm_and_si128(q4bits2_1, m15), q4h_3); + const __m128i q4_4 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_0, 4), m15), q4h_4); + const __m128i q4_5 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_1, 4), m15), q4h_5); + const __m128i q4_6 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_0, 4), m15), q4h_6); + const __m128i q4_7 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_1, 4), m15), q4h_7); + + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + __m128i p16_0 = _mm_maddubs_epi16(q4_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q4_1, q8_1); + __m128i p16_2 = _mm_maddubs_epi16(q4_2, q8_2); + __m128i p16_3 = _mm_maddubs_epi16(q4_3, q8_3); + __m128i p16_4 = _mm_maddubs_epi16(q4_4, q8_4); + __m128i p16_5 = _mm_maddubs_epi16(q4_5, q8_5); + __m128i p16_6 = _mm_maddubs_epi16(q4_6, q8_6); + __m128i p16_7 = _mm_maddubs_epi16(q4_7, q8_7); + + const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); + const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); + const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); + const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); + is += 4; + + p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_0, 8)), p16_1); + p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); + p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_1, 8)), p16_3); + p16_4 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_2), p16_4); + p16_5 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_2, 8)), p16_5); + p16_6 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_3), p16_6); + p16_7 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_3, 8)), p16_7); + + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_4, p16_6)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_5, p16_7)); + + } + + sumi_0 = _mm_sub_epi32(sumi_0, q8sclsub_0); + sumi_1 = _mm_sub_epi32(sumi_1, q8sclsub_1); + const __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sumi)), acc); + } + + *s = hsum_float_8(acc); + +#elif defined __riscv_v_intrinsic + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const int8_t * restrict scale = x[i].scales; + + size_t vl; + + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + + int sum_t = 0; + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + vl = 32; + + // load qh + vuint8m1_t qh_x = __riscv_vle8_v_u8m1(qh, vl); + + // load Q6 + vuint8m1_t q6_0 = __riscv_vle8_v_u8m1(q6, vl); + vuint8m1_t q6_1 = __riscv_vle8_v_u8m1(q6+32, vl); + + vuint8m1_t q6a_0 = __riscv_vand_vx_u8m1(q6_0, 0x0F, vl); + vuint8m1_t q6a_1 = __riscv_vand_vx_u8m1(q6_1, 0x0F, vl); + vuint8m1_t q6s_0 = __riscv_vsrl_vx_u8m1(q6_0, 0x04, vl); + vuint8m1_t q6s_1 = __riscv_vsrl_vx_u8m1(q6_1, 0x04, vl); + + vuint8m1_t qh_0 = __riscv_vand_vx_u8m1(qh_x, 0x03, vl); + vuint8m1_t qh_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x2, vl), 0x03 , vl); + vuint8m1_t qh_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x4, vl), 0x03 , vl); + vuint8m1_t qh_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x6, vl), 0x03 , vl); + + vuint8m1_t qhi_0 = __riscv_vor_vv_u8m1(q6a_0, __riscv_vsll_vx_u8m1(qh_0, 0x04, vl), vl); + vuint8m1_t qhi_1 = __riscv_vor_vv_u8m1(q6a_1, __riscv_vsll_vx_u8m1(qh_1, 0x04, vl), vl); + vuint8m1_t qhi_2 = __riscv_vor_vv_u8m1(q6s_0, __riscv_vsll_vx_u8m1(qh_2, 0x04, vl), vl); + vuint8m1_t qhi_3 = __riscv_vor_vv_u8m1(q6s_1, __riscv_vsll_vx_u8m1(qh_3, 0x04, vl), vl); + + vint8m1_t a_0 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_0), 32, vl); + vint8m1_t a_1 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_1), 32, vl); + vint8m1_t a_2 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_2), 32, vl); + vint8m1_t a_3 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_3), 32, vl); + + // load Q8 and take product + vint16m2_t va_q_0 = __riscv_vwmul_vv_i16m2(a_0, __riscv_vle8_v_i8m1(q8, vl), vl); + vint16m2_t va_q_1 = __riscv_vwmul_vv_i16m2(a_1, __riscv_vle8_v_i8m1(q8+32, vl), vl); + vint16m2_t va_q_2 = __riscv_vwmul_vv_i16m2(a_2, __riscv_vle8_v_i8m1(q8+64, vl), vl); + vint16m2_t va_q_3 = __riscv_vwmul_vv_i16m2(a_3, __riscv_vle8_v_i8m1(q8+96, vl), vl); + + vl = 16; + + vint32m2_t vaux_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 0), scale[is+0], vl); + vint32m2_t vaux_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 1), scale[is+1], vl); + vint32m2_t vaux_2 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 0), scale[is+2], vl); + vint32m2_t vaux_3 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 1), scale[is+3], vl); + vint32m2_t vaux_4 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 0), scale[is+4], vl); + vint32m2_t vaux_5 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 1), scale[is+5], vl); + vint32m2_t vaux_6 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 0), scale[is+6], vl); + vint32m2_t vaux_7 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 1), scale[is+7], vl); + + vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_0, vaux_1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_2, vaux_3, vl), isum0, vl); + vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_4, vaux_5, vl), isum1, vl); + vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_6, vaux_7, vl), isum2, vl); + + sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); + + q6 += 64; qh += 32; q8 += 128; is=8; + + } + + sumf += d * sum_t; + + } + + *s = sumf; + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector int v0 = vec_splats((int32_t)0); + const vector unsigned char v2 = vec_splats((unsigned char)0x2); + const vector unsigned char v3 = vec_splats((unsigned char)0x3); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + const vector unsigned char v6 = vec_splats((unsigned char)0x6); + const vector signed char off = vec_splats((signed char)0x20); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + vector signed int vsumi4 = v0; + vector signed int vsumi5 = v0; + vector signed int vsumi6 = v0; + vector signed int vsumi7 = v0; + + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict qs = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/128; ++j) { + __builtin_prefetch(q6, 0, 0); + __builtin_prefetch(qh, 0, 0); + __builtin_prefetch(q8, 0, 0); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q6); + vector signed char qxs1 = (vector signed char)vec_xl(16, q6); + vector signed char qxs2 = (vector signed char)vec_xl(32, q6); + vector signed char qxs3 = (vector signed char)vec_xl(48, q6); + q6 += 64; + + vector signed char qxs00 = vec_and(qxs0, lowMask); + vector signed char qxs01 = vec_sr(qxs0, v4); + vector signed char qxs10 = vec_and(qxs1, lowMask); + vector signed char qxs11 = vec_sr(qxs1, v4); + vector signed char qxs20 = vec_and(qxs2, lowMask); + vector signed char qxs21 = vec_sr(qxs2, v4); + vector signed char qxs30 = vec_and(qxs3, lowMask); + vector signed char qxs31 = vec_sr(qxs3, v4); + + vector signed char qxhs0 = (vector signed char)vec_xl( 0, qh); + vector signed char qxhs1 = (vector signed char)vec_xl(16, qh); + qh += 32; + + vector signed char qxh00 = vec_sl(vec_and((vector signed char)v3, qxhs0), v4); + vector signed char qxh01 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v4)), v4); + vector signed char qxh10 = vec_sl(vec_and((vector signed char)v3, qxhs1), v4); + vector signed char qxh11 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v4)), v4); + vector signed char qxh20 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v2)), v4); + vector signed char qxh21 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v6)), v4); + vector signed char qxh30 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v2)), v4); + vector signed char qxh31 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v6)), v4); + + vector signed char q6x00 = vec_sub(vec_or(qxh00, qxs00), off); + vector signed char q6x01 = vec_sub(vec_or(qxh01, qxs01), off); + vector signed char q6x10 = vec_sub(vec_or(qxh10, qxs10), off); + vector signed char q6x11 = vec_sub(vec_or(qxh11, qxs11), off); + vector signed char q6x20 = vec_sub(vec_or(qxh20, qxs20), off); + vector signed char q6x21 = vec_sub(vec_or(qxh21, qxs21), off); + vector signed char q6x30 = vec_sub(vec_or(qxh30, qxs30), off); + vector signed char q6x31 = vec_sub(vec_or(qxh31, qxs31), off); + + vector signed char q8y00 = vec_xl( 0, q8); + vector signed char q8y10 = vec_xl( 16, q8); + vector signed char q8y20 = vec_xl( 32, q8); + vector signed char q8y30 = vec_xl( 48, q8); + vector signed char q8y01 = vec_xl( 64, q8); + vector signed char q8y11 = vec_xl( 80, q8); + vector signed char q8y21 = vec_xl( 96, q8); + vector signed char q8y31 = vec_xl(112, q8); + q8 += 128; + + vector signed short qv00 = vec_add(vec_mule(q6x00, q8y00), vec_mulo(q6x00, q8y00)); + vector signed short qv10 = vec_add(vec_mule(q6x10, q8y10), vec_mulo(q6x10, q8y10)); + vector signed short qv20 = vec_add(vec_mule(q6x20, q8y20), vec_mulo(q6x20, q8y20)); + vector signed short qv30 = vec_add(vec_mule(q6x30, q8y30), vec_mulo(q6x30, q8y30)); + vector signed short qv01 = vec_add(vec_mule(q6x01, q8y01), vec_mulo(q6x01, q8y01)); + vector signed short qv11 = vec_add(vec_mule(q6x11, q8y11), vec_mulo(q6x11, q8y11)); + vector signed short qv21 = vec_add(vec_mule(q6x21, q8y21), vec_mulo(q6x21, q8y21)); + vector signed short qv31 = vec_add(vec_mule(q6x31, q8y31), vec_mulo(q6x31, q8y31)); + + vector signed short vscales = vec_unpackh(vec_xl_len(qs, 8)); + qs += 8; + + vector signed short vs0 = vec_splat(vscales, 0); + vector signed short vs1 = vec_splat(vscales, 1); + vector signed short vs2 = vec_splat(vscales, 2); + vector signed short vs3 = vec_splat(vscales, 3); + vector signed short vs4 = vec_splat(vscales, 4); + vector signed short vs5 = vec_splat(vscales, 5); + vector signed short vs6 = vec_splat(vscales, 6); + vector signed short vs7 = vec_splat(vscales, 7); + + vsumi0 = vec_msum(qv00, vs0, vsumi0); + vsumi1 = vec_msum(qv01, vs4, vsumi1); + vsumi2 = vec_msum(qv10, vs1, vsumi2); + vsumi3 = vec_msum(qv11, vs5, vsumi3); + vsumi4 = vec_msum(qv20, vs2, vsumi4); + vsumi5 = vec_msum(qv21, vs6, vsumi5); + vsumi6 = vec_msum(qv30, vs3, vsumi6); + vsumi7 = vec_msum(qv31, vs7, vsumi7); + } + + vsumi0 = vec_add(vsumi0, vsumi4); + vsumi1 = vec_add(vsumi1, vsumi5); + vsumi2 = vec_add(vsumi2, vsumi6); + vsumi3 = vec_add(vsumi3, vsumi7); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined __loongarch_asx + + const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); + const __m256i m2 = __lasx_xvreplgr2vr_b(3); + const __m256i m32s = __lasx_xvreplgr2vr_b(32); + + __m256 acc = (__m256)__lasx_xvldi(0); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const __m128i scales = __lsx_vld((const __m128i*)x[i].scales, 0); + + __m256i sumi = __lasx_xvldi(0); + + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + const __m128i scale_0 = lsx_shuffle_b(scales, get_scale_shuffle(is + 0)); + const __m128i scale_1 = lsx_shuffle_b(scales, get_scale_shuffle(is + 1)); + const __m128i scale_2 = lsx_shuffle_b(scales, get_scale_shuffle(is + 2)); + const __m128i scale_3 = lsx_shuffle_b(scales, get_scale_shuffle(is + 3)); + is += 4; + + const __m256i q4bits1 = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; + const __m256i q4bits2 = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; + const __m256i q4bitsH = __lasx_xvld((const __m256i*)qh, 0); qh += 32; + + const __m256i q4h_0 = __lasx_xvslli_h(__lasx_xvand_v(q4bitsH, m2), 4); + const __m256i q4h_1 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 2), m2), 4); + const __m256i q4h_2 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 4), m2), 4); + const __m256i q4h_3 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 6), m2), 4); + + const __m256i q4_0 = __lasx_xvor_v(__lasx_xvand_v(q4bits1, m4), q4h_0); + const __m256i q4_1 = __lasx_xvor_v(__lasx_xvand_v(q4bits2, m4), q4h_1); + const __m256i q4_2 = __lasx_xvor_v(__lasx_xvand_v(__lasx_xvsrli_h(q4bits1, 4), m4), q4h_2); + const __m256i q4_3 = __lasx_xvor_v(__lasx_xvand_v(__lasx_xvsrli_h(q4bits2, 4), m4), q4h_3); + + const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + + __m256i q8s_0 = lasx_maddubs_h(m32s, q8_0); + __m256i q8s_1 = lasx_maddubs_h(m32s, q8_1); + __m256i q8s_2 = lasx_maddubs_h(m32s, q8_2); + __m256i q8s_3 = lasx_maddubs_h(m32s, q8_3); + + __m256i p16_0 = lasx_maddubs_h(q4_0, q8_0); + __m256i p16_1 = lasx_maddubs_h(q4_1, q8_1); + __m256i p16_2 = lasx_maddubs_h(q4_2, q8_2); + __m256i p16_3 = lasx_maddubs_h(q4_3, q8_3); + + p16_0 = __lasx_xvsub_h(p16_0, q8s_0); + p16_1 = __lasx_xvsub_h(p16_1, q8s_1); + p16_2 = __lasx_xvsub_h(p16_2, q8s_2); + p16_3 = __lasx_xvsub_h(p16_3, q8s_3); + + p16_0 = lasx_madd_h(lasx_ext8_16(scale_0), p16_0); + p16_1 = lasx_madd_h(lasx_ext8_16(scale_1), p16_1); + p16_2 = lasx_madd_h(lasx_ext8_16(scale_2), p16_2); + p16_3 = lasx_madd_h(lasx_ext8_16(scale_3), p16_3); + + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_1)); + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_2, p16_3)); + } + + acc = __lasx_xvfmadd_s((__m256)__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc); + } + + *s = hsum_float_8(acc); + +#else + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) { + a[l + 0] = (int8_t)((q4[l + 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; + a[l + 32] = (int8_t)((q4[l + 32] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; + a[l + 64] = (int8_t)((q4[l + 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; + a[l + 96] = (int8_t)((q4[l + 32] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; + } + a += 128; + q4 += 64; + qh += 32; + } + a = aux8; + int is = 0; + for (int j = 0; j < QK_K/16; ++j) { + int scale = x[i].scales[is++]; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; +#endif +} + +#if defined (__AVX__) || defined (__AVX2__) || defined (__ARM_NEON) || defined (__POWER9_VECTOR__) || defined(__loongarch_asx) +static const int8_t keven_signs_q2xs[1024] = { + 1, 1, 1, 1, 1, 1, 1, 1, -1, 1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, 1, + 1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, 1, 1, -1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, -1, + 1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, -1, + 1, 1, -1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, 1, + 1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, -1, + 1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, 1, + 1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, 1, + 1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, 1, 1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, -1, + 1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, -1, + 1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, 1, + 1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, 1, + 1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, -1, + 1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, 1, + 1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, -1, + 1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, -1, + 1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, 1, + 1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, -1, + 1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, 1, + 1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, 1, + 1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, -1, + 1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, 1, + 1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, -1, + 1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, -1, + 1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, 1, + 1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, 1, + 1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, -1, + 1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, -1, + 1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, 1, + 1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, -1, + 1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, 1, + 1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, 1, + 1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, 1, 1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, +}; +#endif + +void ggml_vec_dot_iq2_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq2_xxs * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + ggml_int8x16x4_t q2u; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + float sumf1 = 0, sumf2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 0])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 1]))); + q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 2])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 3]))); + q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 8])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 9]))); + q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[10])), vld1_s8((const void *)(iq2xxs_grid + aux8[11]))); + q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); + q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); + q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 7) & 127)))); + q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 21) & 127)))); + q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); + q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); + q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); + q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]), q2u.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]), q2u.val[3], q8b.val[3]); + sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[1] >> 28)); + sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[3] >> 28)); + } + sumf += d*(sumf1 + sumf2); + } + *s = 0.25f * sumf; + +#elif defined(__AVX2__) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + const __m256i q2_1 = _mm256_set_epi64x(iq2xxs_grid[aux8[ 3]], iq2xxs_grid[aux8[ 2]], iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); + const __m256i q2_2 = _mm256_set_epi64x(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]], iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); + const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127], + signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = aux32[1] >> 28; + const uint16_t ls2 = aux32[3] >> 28; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__AVX__) + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + const __m128i q2_1_0 = _mm_set_epi64x(iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); + const __m128i q2_1_1 = _mm_set_epi64x(iq2xxs_grid[aux8[3]], iq2xxs_grid[aux8[2]]); + const __m128i q2_2_0 = _mm_set_epi64x(iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); + const __m128i q2_2_1 = _mm_set_epi64x(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]]); + const __m128i s2_1_0 = _mm_set_epi64x(signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m128i s2_1_1 = _mm_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127]); + const __m128i s2_2_0 = _mm_set_epi64x(signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); + const __m128i s2_2_1 = _mm_set_epi64x(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127]); + const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, s2_1_0); + const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, s2_1_1); + const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, s2_2_0); + const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, s2_2_1); + const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); + const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); + const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); + const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); + const uint16_t ls1 = aux32[1] >> 28; + const uint16_t ls2 = aux32[3] >> 28; + const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); + sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); + sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); + sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); + sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); + } + + accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__POWER9_VECTOR__) + const vector int v0 = vec_splats((int32_t)0); + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/32; j += 2) { + __builtin_prefetch(q2, 0, 1); + __builtin_prefetch(q8, 0, 1); + + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + memcpy(aux32, q2, 4*sizeof(uint32_t)); + q2 += 8; + + vector signed long long aux64x2_0 = {*(const int64_t *)(iq2xxs_grid + aux8[ 0]), *(const int64_t *)(iq2xxs_grid + aux8[ 1])}; + vector signed long long aux64x2_1 = {*(const int64_t *)(iq2xxs_grid + aux8[ 2]), *(const int64_t *)(iq2xxs_grid + aux8[ 3])}; + vector signed long long aux64x2_2 = {*(const int64_t *)(iq2xxs_grid + aux8[ 8]), *(const int64_t *)(iq2xxs_grid + aux8[ 9])}; + vector signed long long aux64x2_3 = {*(const int64_t *)(iq2xxs_grid + aux8[10]), *(const int64_t *)(iq2xxs_grid + aux8[11])}; + + vector signed long long vsigns0 = {*(const int64_t *)(signs64 + ((aux32[1] >> 0) & 127)), *(const int64_t *)(signs64 + ((aux32[1] >> 7) & 127))}; + vector signed long long vsigns1 = {*(const int64_t *)(signs64 + ((aux32[1] >> 14) & 127)), *(const int64_t *)(signs64 + ((aux32[1] >> 21) & 127))}; + vector signed long long vsigns2 = {*(const int64_t *)(signs64 + ((aux32[3] >> 0) & 127)), *(const int64_t *)(signs64 + ((aux32[3] >> 7) & 127))}; + vector signed long long vsigns3 = {*(const int64_t *)(signs64 + ((aux32[3] >> 14) & 127)), *(const int64_t *)(signs64 + ((aux32[3] >> 21) & 127))}; + + vector signed char q2x0 = (vector signed char)vec_mul((vector signed char)vsigns0, (vector signed char)aux64x2_0); + vector signed char q2x1 = (vector signed char)vec_mul((vector signed char)vsigns1, (vector signed char)aux64x2_1); + vector signed char q2x2 = (vector signed char)vec_mul((vector signed char)vsigns2, (vector signed char)aux64x2_2); + vector signed char q2x3 = (vector signed char)vec_mul((vector signed char)vsigns3, (vector signed char)aux64x2_3); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); + + const uint16_t ls0 = aux32[1] >> 28; + const uint16_t ls1 = aux32[3] >> 28; + + vector signed short vscales01 = vec_splats((int16_t)(2*ls0+1)); + vector signed short vscales23 = vec_splats((int16_t)(2*ls1+1)); + + vsumi0 = vec_msum(qv0, vscales01, vsumi0); + vsumi1 = vec_msum(qv1, vscales01, vsumi1); + vsumi2 = vec_msum(qv2, vscales23, vsumi2); + vsumi3 = vec_msum(qv3, vscales23, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = 0.125f * vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + __m256 accumf = (__m256)__lasx_xvldi(0); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + + const __m256i q2_1 = lasx_set_d(iq2xxs_grid[aux8[ 3]], iq2xxs_grid[aux8[ 2]], iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); + const __m256i q2_2 = lasx_set_d(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]], iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); + const __m256i s2_1 = lasx_set_d(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i s2_2 = lasx_set_d(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127], + signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); + const __m256i q8s_1 = __lasx_xvsigncov_b(s2_1, q8_1); + const __m256i q8s_2 = __lasx_xvsigncov_b(s2_2, q8_2); + const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); + const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); + const uint16_t ls1 = aux32[1] >> 28; + const uint16_t ls2 = aux32[3] >> 28; + const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); + const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); + sumi1 = __lasx_xvadd_w(sumi1, p1); + sumi2 = __lasx_xvadd_w(sumi2, p2); + } + + accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); + } + + *s = 0.125f * hsum_float_8(accumf); + +#else + + uint32_t aux32[2]; + const uint8_t * aux8 = (const uint8_t *)aux32; + + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(aux32, q2, 2*sizeof(uint32_t)); + q2 += 4; + const uint32_t ls = 2*(aux32[1] >> 28) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[l]); + const uint8_t signs = ksigns_iq2xs[(aux32[1] >> 7*l) & 127]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls; + } + sumf += d * bsum; + } + *s = 0.125f * sumf; +#endif +} + +void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq2_xs * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + ggml_int8x16x4_t q2u; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; + + int32x4x4_t scales32; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + const uint8x8_t scales8 = vld1_u8(x[i].scales); + const uint8x8_t scales_l = vand_u8(scales8, vdup_n_u8(0xf)); + const uint8x8_t scales_h = vshr_n_u8(scales8, 4); + uint8x16_t scales = vcombine_u8(vzip1_u8(scales_l, scales_h), vzip2_u8(scales_l, scales_h)); + scales = vaddq_u8(vshlq_n_u8(scales, 1), vdupq_n_u8(1)); + const uint16x8_t scales1 = vmovl_u8(vget_low_u8(scales)); + const uint16x8_t scales2 = vmovl_u8(vget_high_u8(scales)); + scales32.val[0] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales1))); + scales32.val[1] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales1))); + scales32.val[2] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales2))); + scales32.val[3] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales2))); + int32x4_t sumi = vdupq_n_s32(0); + for (int ib64 = 0; ib64 < QK_K/64; ++ib64) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[0] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[1] & 511)))); + q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[2] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[3] & 511)))); + q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[4] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[5] & 511)))); + q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[6] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[7] & 511)))); + q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[0] >> 9))), vld1_s8((const void *)(signs64 + (q2[1] >> 9)))); + q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[2] >> 9))), vld1_s8((const void *)(signs64 + (q2[3] >> 9)))); + q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[4] >> 9))), vld1_s8((const void *)(signs64 + (q2[5] >> 9)))); + q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[6] >> 9))), vld1_s8((const void *)(signs64 + (q2[7] >> 9)))); + q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); + q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); + q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); + q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); + const int32x4_t p1 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]); + const int32x4_t p2 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[1], q8b.val[1]); + const int32x4_t p3 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]); + const int32x4_t p4 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[3], q8b.val[3]); + const int32x4_t p = vpaddq_s32(vpaddq_s32(p1, p2), vpaddq_s32(p3, p4)); + sumi = vmlaq_s32(sumi, p, scales32.val[ib64]); + q2 += 8; + } + sumf += d*vaddvq_s32(sumi); + } + *s = 0.125f * sumf; + +#elif defined(__AVX2__) + + const __m256i mone = _mm256_set1_epi8(1); + static const char block_sign_shuffle_mask_1[32] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, + 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + }; + static const char block_sign_shuffle_mask_2[32] = { + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, + 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, + }; + static const uint8_t bit_selector_mask_bytes[32] = { + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m256i bit_selector_mask = _mm256_loadu_si256((const __m256i*)bit_selector_mask_bytes); + const __m256i block_sign_shuffle_1 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_1); + const __m256i block_sign_shuffle_2 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_2); + + static const uint8_t k_bit_helper[32] = { + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + }; + const __m256i bit_helper = _mm256_loadu_si256((const __m256i*)k_bit_helper); + const __m256i m511 = _mm256_set1_epi16(511); + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); + + uint64_t aux64; + + // somewhat hacky, but gives a significant boost in performance + __m256i aux_gindex; + const uint16_t * gindex = (const uint16_t *)&aux_gindex; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(&aux64, x[i].scales, 8); + __m128i stmp = _mm_set1_epi64x(aux64); + stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4)); + const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1); + + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { + + const __m256i q2_data = _mm256_loadu_si256((const __m256i*)q2); q2 += 16; + aux_gindex = _mm256_and_si256(q2_data, m511); + + const __m256i partial_sign_bits = _mm256_srli_epi16(q2_data, 9); + const __m256i partial_sign_bits_upper = _mm256_srli_epi16(q2_data, 13); + const __m256i partial_sign_bits_for_counting = _mm256_xor_si256(partial_sign_bits, partial_sign_bits_upper); + + const __m256i odd_bits = _mm256_shuffle_epi8(bit_helper, partial_sign_bits_for_counting); + const __m256i full_sign_bits = _mm256_or_si256(partial_sign_bits, odd_bits); + + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_4 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + + const __m256i q2_1 = _mm256_set_epi64x(iq2xs_grid[gindex[ 3]], iq2xs_grid[gindex[ 2]], + iq2xs_grid[gindex[ 1]], iq2xs_grid[gindex[ 0]]); + const __m256i q2_2 = _mm256_set_epi64x(iq2xs_grid[gindex[ 7]], iq2xs_grid[gindex[ 6]], + iq2xs_grid[gindex[ 5]], iq2xs_grid[gindex[ 4]]); + const __m256i q2_3 = _mm256_set_epi64x(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]], + iq2xs_grid[gindex[ 9]], iq2xs_grid[gindex[ 8]]); + const __m256i q2_4 = _mm256_set_epi64x(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]], + iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); + + const __m128i full_signs_l = _mm256_castsi256_si128(full_sign_bits); + const __m128i full_signs_h = _mm256_extractf128_si256(full_sign_bits, 1); + const __m256i full_signs_1 = MM256_SET_M128I(full_signs_l, full_signs_l); + const __m256i full_signs_2 = MM256_SET_M128I(full_signs_h, full_signs_h); + + __m256i signs; + signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_1); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, _mm256_or_si256(signs, mone)); + + signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_2); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, _mm256_or_si256(signs, mone)); + + signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_1); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_3 = _mm256_sign_epi8(q8_3, _mm256_or_si256(signs, mone)); + + signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_2); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_4 = _mm256_sign_epi8(q8_4, _mm256_or_si256(signs, mone)); + + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const __m256i dot3 = _mm256_maddubs_epi16(q2_3, q8s_3); + const __m256i dot4 = _mm256_maddubs_epi16(q2_4, q8s_4); + + const __m256i sc1 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0))); + const __m256i sc2 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1))); + const __m256i sc3 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+2))); + const __m256i sc4 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+3))); + + sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot1, sc1)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot2, sc2)); + sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot3, sc3)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot4, sc4)); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__AVX__) + const __m128i mone = _mm_set1_epi8(1); + static const char block_sign_shuffle_mask_1[32] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, + 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + }; + static const char block_sign_shuffle_mask_2[32] = { + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, + 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, + }; + static const uint8_t bit_selector_mask_bytes[32] = { + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m128i bit_selector_mask_0 = _mm_loadu_si128((const __m128i*)bit_selector_mask_bytes); + const __m128i bit_selector_mask_1 = _mm_loadu_si128((const __m128i*)bit_selector_mask_bytes + 1); + const __m128i block_sign_shuffle_1_0 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_1); + const __m128i block_sign_shuffle_1_1 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_1 + 1); + const __m128i block_sign_shuffle_2_0 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_2); + const __m128i block_sign_shuffle_2_1 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_2 + 1); + + static const uint8_t k_bit_helper[32] = { + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + }; + const __m128i bit_helper_0 = _mm_loadu_si128((const __m128i*)k_bit_helper); + const __m128i bit_helper_1 = _mm_loadu_si128((const __m128i*)k_bit_helper + 1); + const __m128i m511 = _mm_set1_epi16(511); + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); + + uint64_t aux64; + + // somewhat hacky, but gives a significant boost in performance + __m256i aux_gindex; + const uint16_t * gindex = (const uint16_t *)&aux_gindex; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(&aux64, x[i].scales, 8); + __m128i stmp = _mm_set1_epi64x(aux64); + stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4)); + const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1); + + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { + + const __m128i q2_data_0 = _mm_loadu_si128((const __m128i*)q2); + const __m128i q2_data_1 = _mm_loadu_si128((const __m128i*)q2 + 1); q2 += 16; + aux_gindex = MM256_SET_M128I(_mm_and_si128(q2_data_1, m511), _mm_and_si128(q2_data_0, m511)); + + const __m128i partial_sign_bits_0 = _mm_srli_epi16(q2_data_0, 9); + const __m128i partial_sign_bits_1 = _mm_srli_epi16(q2_data_1, 9); + const __m128i partial_sign_bits_upper_0 = _mm_srli_epi16(q2_data_0, 13); + const __m128i partial_sign_bits_upper_1 = _mm_srli_epi16(q2_data_1, 13); + const __m128i partial_sign_bits_for_counting_0 = _mm_xor_si128(partial_sign_bits_0, partial_sign_bits_upper_0); + const __m128i partial_sign_bits_for_counting_1 = _mm_xor_si128(partial_sign_bits_1, partial_sign_bits_upper_1); + + const __m128i odd_bits_0 = _mm_shuffle_epi8(bit_helper_0, partial_sign_bits_for_counting_0); + const __m128i odd_bits_1 = _mm_shuffle_epi8(bit_helper_1, partial_sign_bits_for_counting_1); + const __m128i full_sign_bits_0 = _mm_or_si128(partial_sign_bits_0, odd_bits_0); + const __m128i full_sign_bits_1 = _mm_or_si128(partial_sign_bits_1, odd_bits_1); + + const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_3_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_3_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_4_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_4_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + + const __m128i q2_1_0 = _mm_set_epi64x(iq2xs_grid[gindex[1]], iq2xs_grid[gindex[0]]); + const __m128i q2_1_1 = _mm_set_epi64x(iq2xs_grid[gindex[3]], iq2xs_grid[gindex[2]]); + const __m128i q2_2_0 = _mm_set_epi64x(iq2xs_grid[gindex[5]], iq2xs_grid[gindex[4]]); + const __m128i q2_2_1 = _mm_set_epi64x(iq2xs_grid[gindex[7]], iq2xs_grid[gindex[6]]); + const __m128i q2_3_0 = _mm_set_epi64x(iq2xs_grid[gindex[9]], iq2xs_grid[gindex[8]]); + const __m128i q2_3_1 = _mm_set_epi64x(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]]); + const __m128i q2_4_0 = _mm_set_epi64x(iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); + const __m128i q2_4_1 = _mm_set_epi64x(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]]); + + // AVX2 full_signs_1 is full_sign_bits_0 here + // AVX2 full_signs_2 is full_sign_bits_1 here + __m128i signs_0, signs_1; + signs_0 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_1_0); + signs_1 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_1_1); + signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); + signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); + const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, _mm_or_si128(signs_0, mone)); + const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, _mm_or_si128(signs_1, mone)); + + signs_0 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_2_0); + signs_1 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_2_1); + signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); + signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); + const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, _mm_or_si128(signs_0, mone)); + const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, _mm_or_si128(signs_1, mone)); + + signs_0 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_1_0); + signs_1 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_1_1); + signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); + signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); + const __m128i q8s_3_0 = _mm_sign_epi8(q8_3_0, _mm_or_si128(signs_0, mone)); + const __m128i q8s_3_1 = _mm_sign_epi8(q8_3_1, _mm_or_si128(signs_1, mone)); + + signs_0 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_2_0); + signs_1 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_2_1); + signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); + signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); + const __m128i q8s_4_0 = _mm_sign_epi8(q8_4_0, _mm_or_si128(signs_0, mone)); + const __m128i q8s_4_1 = _mm_sign_epi8(q8_4_1, _mm_or_si128(signs_1, mone)); + + const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); + const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); + const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); + const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); + const __m128i dot3_0 = _mm_maddubs_epi16(q2_3_0, q8s_3_0); + const __m128i dot3_1 = _mm_maddubs_epi16(q2_3_1, q8s_3_1); + const __m128i dot4_0 = _mm_maddubs_epi16(q2_4_0, q8s_4_0); + const __m128i dot4_1 = _mm_maddubs_epi16(q2_4_1, q8s_4_1); + + __m128i sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0)); + const __m128i sc1_0 = _mm_cvtepi8_epi16(sc_tmp); + const __m128i sc1_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); + sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1)); + const __m128i sc2_0 = _mm_cvtepi8_epi16(sc_tmp); + const __m128i sc2_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); + sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+2)); + const __m128i sc3_0 = _mm_cvtepi8_epi16(sc_tmp); + const __m128i sc3_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); + sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+3)); + const __m128i sc4_0 = _mm_cvtepi8_epi16(sc_tmp); + const __m128i sc4_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); + + sumi1_0 = _mm_add_epi32(sumi1_0, _mm_madd_epi16(dot1_0, sc1_0)); + sumi1_1 = _mm_add_epi32(sumi1_1, _mm_madd_epi16(dot1_1, sc1_1)); + sumi2_0 = _mm_add_epi32(sumi2_0, _mm_madd_epi16(dot2_0, sc2_0)); + sumi2_1 = _mm_add_epi32(sumi2_1, _mm_madd_epi16(dot2_1, sc2_1)); + sumi1_0 = _mm_add_epi32(sumi1_0, _mm_madd_epi16(dot3_0, sc3_0)); + sumi1_1 = _mm_add_epi32(sumi1_1, _mm_madd_epi16(dot3_1, sc3_1)); + sumi2_0 = _mm_add_epi32(sumi2_0, _mm_madd_epi16(dot4_0, sc4_0)); + sumi2_1 = _mm_add_epi32(sumi2_1, _mm_madd_epi16(dot4_1, sc4_1)); + } + + accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__loongarch_asx) + + const __m256i mone = __lasx_xvreplgr2vr_b(1); + static const char block_sign_shuffle_mask_1[32] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, + 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + }; + static const char block_sign_shuffle_mask_2[32] = { + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, + 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, + }; + static const uint8_t bit_selector_mask_bytes[32] = { + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m256i bit_selector_mask = __lasx_xvld((const __m256i*)bit_selector_mask_bytes, 0); + const __m256i block_sign_shuffle_1 = __lasx_xvld((const __m256i*)block_sign_shuffle_mask_1, 0); + const __m256i block_sign_shuffle_2 = __lasx_xvld((const __m256i*)block_sign_shuffle_mask_2, 0); + + static const uint8_t k_bit_helper[32] = { + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + }; + const __m256i bit_helper = __lasx_xvld((const __m256i*)k_bit_helper, 0); + const __m256i m511 = __lasx_xvreplgr2vr_h(511); + const __m128i m4 = __lsx_vreplgr2vr_b(0xf); + const __m128i m1 = __lsx_vreplgr2vr_b(1); + + uint64_t aux64; + + // somewhat hacky, but gives a significant boost in performance + __m256i aux_gindex; + const uint16_t * gindex = (const uint16_t *)&aux_gindex; + + __m256 accumf = (__m256)__lasx_xvldi(0); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(&aux64, x[i].scales, 8); + __m128i stmp = __lsx_vreplgr2vr_d(aux64); + stmp = __lsx_vilvl_b( __lsx_vand_v(__lsx_vsrli_h(stmp, 4), m4), __lsx_vand_v(stmp, m4)); + const __m128i scales = __lsx_vadd_b(__lsx_vslli_h(stmp, 1), m1); + + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { + + const __m256i q2_data = __lasx_xvld((const __m256i*)q2, 0); q2 += 16; + aux_gindex = __lasx_xvand_v(q2_data, m511); + + const __m256i partial_sign_bits = __lasx_xvsrli_h(q2_data, 9); + const __m256i partial_sign_bits_upper = __lasx_xvsrli_h(q2_data, 13); + const __m256i partial_sign_bits_for_counting = __lasx_xvxor_v(partial_sign_bits, partial_sign_bits_upper); + + const __m256i odd_bits = lasx_shuffle_b(bit_helper, partial_sign_bits_for_counting); + const __m256i full_sign_bits = __lasx_xvor_v(partial_sign_bits, odd_bits); + + const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_3 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_4 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + + const __m256i q2_1 = lasx_set_d(iq2xs_grid[gindex[ 3]], iq2xs_grid[gindex[ 2]], + iq2xs_grid[gindex[ 1]], iq2xs_grid[gindex[ 0]]); + const __m256i q2_2 = lasx_set_d(iq2xs_grid[gindex[ 7]], iq2xs_grid[gindex[ 6]], + iq2xs_grid[gindex[ 5]], iq2xs_grid[gindex[ 4]]); + const __m256i q2_3 = lasx_set_d(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]], + iq2xs_grid[gindex[ 9]], iq2xs_grid[gindex[ 8]]); + const __m256i q2_4 = lasx_set_d(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]], + iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); + + const __m128i full_signs_l = lasx_extracti128(full_sign_bits, 0); + const __m128i full_signs_h = lasx_extracti128(full_sign_bits, 1); + const __m256i full_signs_1 = lasx_insertf128(full_signs_l, full_signs_l); + const __m256i full_signs_2 = lasx_insertf128(full_signs_h, full_signs_h); + + __m256i signs; + signs = lasx_shuffle_b(full_signs_1, block_sign_shuffle_1); + signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_1 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_1); + + signs = lasx_shuffle_b(full_signs_1, block_sign_shuffle_2); + signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_2 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_2); + + signs = lasx_shuffle_b(full_signs_2, block_sign_shuffle_1); + signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_3 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_3); + + signs = lasx_shuffle_b(full_signs_2, block_sign_shuffle_2); + signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_4 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_4); + + const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); + const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); + const __m256i dot3 = lasx_maddubs_h(q2_3, q8s_3); + const __m256i dot4 = lasx_maddubs_h(q2_4, q8s_4); + + const __m256i sc1 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+0))); + const __m256i sc2 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+1))); + const __m256i sc3 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+2))); + const __m256i sc4 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+3))); + + sumi1 = __lasx_xvadd_w(sumi1, lasx_madd_h(dot1, sc1)); + sumi2 = __lasx_xvadd_w(sumi2, lasx_madd_h(dot2, sc2)); + sumi1 = __lasx_xvadd_w(sumi1, lasx_madd_h(dot3, sc3)); + sumi2 = __lasx_xvadd_w(sumi2, lasx_madd_h(dot4, sc4)); + } + + accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); +#elif defined(__POWER9_VECTOR__) + const vector int v0 = vec_splats((int32_t)0); + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint16_t * restrict q2 = x[i].qs; + const uint8_t * restrict sc = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/64; ++j) { + __builtin_prefetch(q2, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed long long aux64x2_0 = {*(const int64_t *)(iq2xs_grid + (q2[0] & 511)), *(const int64_t *)(iq2xs_grid + (q2[1] & 511))}; + vector signed long long aux64x2_1 = {*(const int64_t *)(iq2xs_grid + (q2[2] & 511)), *(const int64_t *)(iq2xs_grid + (q2[3] & 511))}; + vector signed long long aux64x2_2 = {*(const int64_t *)(iq2xs_grid + (q2[4] & 511)), *(const int64_t *)(iq2xs_grid + (q2[5] & 511))}; + vector signed long long aux64x2_3 = {*(const int64_t *)(iq2xs_grid + (q2[6] & 511)), *(const int64_t *)(iq2xs_grid + (q2[7] & 511))}; + + vector signed long long vsigns0 = {*(const int64_t *)(signs64 + ((q2[0] >> 9))), *(const int64_t *)(signs64 + ((q2[1] >> 9)))}; + vector signed long long vsigns1 = {*(const int64_t *)(signs64 + ((q2[2] >> 9))), *(const int64_t *)(signs64 + ((q2[3] >> 9)))}; + vector signed long long vsigns2 = {*(const int64_t *)(signs64 + ((q2[4] >> 9))), *(const int64_t *)(signs64 + ((q2[5] >> 9)))}; + vector signed long long vsigns3 = {*(const int64_t *)(signs64 + ((q2[6] >> 9))), *(const int64_t *)(signs64 + ((q2[7] >> 9)))}; + q2 += 8; + + vector signed char q2x0 = (vector signed char)vec_mul((vector signed char)vsigns0, (vector signed char)aux64x2_0); + vector signed char q2x1 = (vector signed char)vec_mul((vector signed char)vsigns1, (vector signed char)aux64x2_1); + vector signed char q2x2 = (vector signed char)vec_mul((vector signed char)vsigns2, (vector signed char)aux64x2_2); + vector signed char q2x3 = (vector signed char)vec_mul((vector signed char)vsigns3, (vector signed char)aux64x2_3); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); + + const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); + const uint16_t ls1 = (uint16_t)(sc[0] >> 4); + const uint16_t ls2 = (uint16_t)(sc[1] & 0xf); + const uint16_t ls3 = (uint16_t)(sc[1] >> 4); + sc += 2; + + vector signed short vscales0 = vec_splats((int16_t)(2*ls0+1)); + vector signed short vscales1 = vec_splats((int16_t)(2*ls1+1)); + vector signed short vscales2 = vec_splats((int16_t)(2*ls2+1)); + vector signed short vscales3 = vec_splats((int16_t)(2*ls3+1)); + + vsumi0 = vec_msum(qv0, vscales0, vsumi0); + vsumi1 = vec_msum(qv1, vscales1, vsumi1); + vsumi2 = vec_msum(qv2, vscales2, vsumi2); + vsumi3 = vec_msum(qv3, vscales3, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = 0.125f * vec_extract(vsumf0, 0); +#else + + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const uint8_t * restrict sc = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + const uint16_t ls1 = 2*(sc[ib32] & 0xf) + 1; + const uint16_t ls2 = 2*(sc[ib32] >> 4) + 1; + int32_t sumi = 0; + for (int l = 0; l < 2; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); + const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls1; + sumi = 0; + for (int l = 2; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); + const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls2; + q2 += 4; + } + sumf += d * bsum; + } + *s = 0.125f * sumf; +#endif +} + +void ggml_vec_dot_iq2_s_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq2_s * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; + + const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); + const uint8x16_t mask2 = vld1q_u8(k_mask2); + const uint8x16_t m1 = vdupq_n_u8(1); + const int32x4_t vzero = vdupq_n_s32(0); + + uint8x16x2_t vs; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const int8_t * restrict q8 = y[i].qs; + + int sumi1 = 0, sumi2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + q2s.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[0] | ((qh[ib32+0] << 8) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[1] | ((qh[ib32+0] << 6) & 0x300))))); + q2s.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[2] | ((qh[ib32+0] << 4) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[3] | ((qh[ib32+0] << 2) & 0x300))))); + q2s.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[4] | ((qh[ib32+1] << 8) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[5] | ((qh[ib32+1] << 6) & 0x300))))); + q2s.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[6] | ((qh[ib32+1] << 4) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[7] | ((qh[ib32+1] << 2) & 0x300))))); + qs += 8; + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vceqq_u8(vs.val[0], mask2); + vs.val[1] = vceqq_u8(vs.val[1], mask2); + + q2s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[0]); + q2s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[1]); + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vceqq_u8(vs.val[0], mask2); + vs.val[1] = vceqq_u8(vs.val[1], mask2); + + signs += 4; + + q2s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[2]); + q2s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[3]); + + const int32x4_t p1 = ggml_vdotq_s32(vzero, q2s.val[0], q8b.val[0]); + const int32x4_t p2 = ggml_vdotq_s32(vzero, q2s.val[1], q8b.val[1]); + const int32x4_t p3 = ggml_vdotq_s32(vzero, q2s.val[2], q8b.val[2]); + const int32x4_t p4 = ggml_vdotq_s32(vzero, q2s.val[3], q8b.val[3]); + + sumi1 += vaddvq_s32(p1) * (1 + 2*(x[i].scales[ib32+0] & 0xf)); + sumi2 += vaddvq_s32(p2) * (1 + 2*(x[i].scales[ib32+0] >> 4)); + sumi1 += vaddvq_s32(p3) * (1 + 2*(x[i].scales[ib32+1] & 0xf)); + sumi2 += vaddvq_s32(p4) * (1 + 2*(x[i].scales[ib32+1] >> 4)); + } + sumf += d*(sumi1 + sumi2); + } + + *s = 0.125f * sumf; + +#elif defined(__AVX2__) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); + + const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); + const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); + + uint64_t aux64; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const int8_t * restrict q8 = y[i].qs; + + memcpy(&aux64, x[i].scales, 8); + const __m128i scales8 = _mm_add_epi8(_mm_slli_epi16(_mm_and_si128(_mm_set_epi64x(aux64 >> 4, aux64), m4), 1), m1); + const __m256i scales16 = _mm256_cvtepi8_epi16(scales8); // 0 2 4 6 8 10 12 14 1 3 5 7 9 11 13 15 + + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q2_1 = _mm256_set_epi64x(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], + iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)], + iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], + iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); + const __m256i q2_2 = _mm256_set_epi64x(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], + iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)], + iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], + iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); + qs += 8; + + __m256i aux256 = _mm256_set1_epi32(signs[0] | ((uint32_t) signs[1] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); + + aux256 = _mm256_set1_epi32(signs[2] | ((uint32_t) signs[3] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); + + signs += 4; + + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); // blocks 2*ib32+0, 2*ib32+1 + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); // blocks 2*ib32+2, 2*ib32+3 + + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+0))); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+1))); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__AVX__) + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); + + const __m128i mask1_0 = _mm_loadu_si128((const __m128i*)k_mask1); + const __m128i mask1_1 = _mm_loadu_si128((const __m128i*)k_mask1 + 1); + const __m128i mask2_0 = _mm_loadu_si128((const __m128i*)k_mask2); + const __m128i mask2_1 = _mm_loadu_si128((const __m128i*)k_mask2 + 1); + + uint64_t aux64; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const int8_t * restrict q8 = y[i].qs; + + memcpy(&aux64, x[i].scales, 8); + const __m128i scales8 = _mm_add_epi8(_mm_slli_epi16(_mm_and_si128(_mm_set_epi64x(aux64 >> 4, aux64), m4), 1), m1); + const __m128i scales16_0 = _mm_cvtepi8_epi16(scales8); + const __m128i scales16_1 = _mm_cvtepi8_epi16(_mm_srli_si128(scales8, 8)); + + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q2_1_0 = _mm_set_epi64x(iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], + iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); + const __m128i q2_1_1 = _mm_set_epi64x(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], + iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)]); + const __m128i q2_2_0 = _mm_set_epi64x(iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], + iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); + const __m128i q2_2_1 = _mm_set_epi64x(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], + iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)]); + qs += 8; + + __m128i aux128_0 = _mm_set1_epi32(signs[0] | ((uint32_t) signs[1] << 16)); + __m128i aux128_1 = aux128_0; + aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); + aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); + const __m128i s2_1_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); + const __m128i s2_1_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); + const __m128i q8s_1_0 = _mm_sub_epi8(_mm_xor_si128(s2_1_0, q8_1_0), s2_1_0); + const __m128i q8s_1_1 = _mm_sub_epi8(_mm_xor_si128(s2_1_1, q8_1_1), s2_1_1); + + aux128_0 = _mm_set1_epi32(signs[2] | ((uint32_t) signs[3] << 16)); + aux128_1 = aux128_0; + aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); + aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); + const __m128i s2_2_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); + const __m128i s2_2_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); + const __m128i q8s_2_0 = _mm_sub_epi8(_mm_xor_si128(s2_2_0, q8_2_0), s2_2_0); + const __m128i q8s_2_1 = _mm_sub_epi8(_mm_xor_si128(s2_2_1, q8_2_1), s2_2_1); + + signs += 4; + + const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); + const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); + const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); + const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); + + const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_shuffle_epi8(scales16_0, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+0), 0))); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_shuffle_epi8(scales16_1, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+0), 1))); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_shuffle_epi8(scales16_0, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+1), 0))); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_shuffle_epi8(scales16_1, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+1), 1))); + sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); + sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); + sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); + sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); + } + + accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); + +#elif defined(__POWER9_VECTOR__) + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; + + const vector int v0 = vec_splats((int32_t)0); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + const vector unsigned char mask0 = vec_xl( 0, k_mask1); + const vector unsigned char mask1 = vec_xl(16, k_mask1); + const vector signed char mask2 = (vector signed char)vec_xl( 0, k_mask2); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint8_t * restrict q2 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const uint8_t * restrict sc = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + + for (int j = 0; j < QK_K/32; j += 2) { + __builtin_prefetch(q2, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed long long aux64x2_0 = {*(const int64_t *)(iq2s_grid + (q2[0] | ((qh[0] << 8) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[1] | ((qh[0] << 6) & 0x300)))}; + vector signed long long aux64x2_1 = {*(const int64_t *)(iq2s_grid + (q2[2] | ((qh[0] << 4) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[3] | ((qh[0] << 2) & 0x300)))}; + vector signed long long aux64x2_2 = {*(const int64_t *)(iq2s_grid + (q2[4] | ((qh[1] << 8) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[5] | ((qh[1] << 6) & 0x300)))}; + vector signed long long aux64x2_3 = {*(const int64_t *)(iq2s_grid + (q2[6] | ((qh[1] << 4) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[7] | ((qh[1] << 2) & 0x300)))}; + q2 += 8; + qh += 2; + + vector signed char vsigns01 = (vector signed char)vec_splats(*(const uint32_t *)&signs[0]); + vector signed char vsigns23 = (vector signed char)vec_splats(*(const uint32_t *)&signs[2]); + signs += 4; + + vector signed char vsigns0 = vec_perm(vsigns01, vsigns01, mask0); + vector signed char vsigns1 = vec_perm(vsigns01, vsigns01, mask1); + vector signed char vsigns2 = vec_perm(vsigns23, vsigns23, mask0); + vector signed char vsigns3 = vec_perm(vsigns23, vsigns23, mask1); + + vsigns0 = (vector signed char)vec_cmpeq(vec_and(vsigns0, mask2), mask2); + vsigns1 = (vector signed char)vec_cmpeq(vec_and(vsigns1, mask2), mask2); + vsigns2 = (vector signed char)vec_cmpeq(vec_and(vsigns2, mask2), mask2); + vsigns3 = (vector signed char)vec_cmpeq(vec_and(vsigns3, mask2), mask2); + + vector signed char q2x0 = vec_sub(vec_xor(vsigns0, (vector signed char)aux64x2_0), vsigns0); + vector signed char q2x1 = vec_sub(vec_xor(vsigns1, (vector signed char)aux64x2_1), vsigns1); + vector signed char q2x2 = vec_sub(vec_xor(vsigns2, (vector signed char)aux64x2_2), vsigns2); + vector signed char q2x3 = vec_sub(vec_xor(vsigns3, (vector signed char)aux64x2_3), vsigns3); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); + + const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); + const uint16_t ls1 = (uint16_t)(sc[0] >> 4); + const uint16_t ls2 = (uint16_t)(sc[1] & 0xf); + const uint16_t ls3 = (uint16_t)(sc[1] >> 4); + sc += 2; + + vector signed short vscales0 = vec_splats((int16_t)(2*ls0+1)); + vector signed short vscales1 = vec_splats((int16_t)(2*ls1+1)); + vector signed short vscales2 = vec_splats((int16_t)(2*ls2+1)); + vector signed short vscales3 = vec_splats((int16_t)(2*ls3+1)); + + vsumi0 = vec_msum(qv0, vscales0, vsumi0); + vsumi1 = vec_msum(qv1, vscales1, vsumi1); + vsumi2 = vec_msum(qv2, vscales2, vsumi2); + vsumi3 = vec_msum(qv3, vscales3, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = 0.125f * vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + + const __m128i m4 = __lsx_vreplgr2vr_b(0xf); + const __m128i m1 = __lsx_vreplgr2vr_b(1); + + const __m256i mask1 = __lasx_xvld((const __m256i*)k_mask1, 0); + const __m256i mask2 = __lasx_xvld((const __m256i*)k_mask2, 0); + uint64_t aux64; + + __m256 accumf = (__m256)__lasx_xvldi(0); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const int8_t * restrict q8 = y[i].qs; + + __m128i tmp1; + memcpy(&aux64, x[i].scales, 8); + tmp1 = __lsx_vinsgr2vr_d(tmp1, aux64, 0); + tmp1 = __lsx_vinsgr2vr_d(tmp1, aux64 >> 4, 1); + const __m128i scales8 = __lsx_vadd_b(__lsx_vslli_h(__lsx_vand_v(tmp1, m4), 1), m1); + const __m256i scales16 = lasx_ext8_16(scales8); // 0 2 4 6 8 10 12 14 1 3 5 7 9 11 13 15 + + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q2_1 = lasx_set_d(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], + iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)], + iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], + iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); + const __m256i q2_2 = lasx_set_d(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], + iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)], + iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], + iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); + qs += 8; + + __m256i aux256 = __lasx_xvreplgr2vr_w(signs[0] | ((uint32_t) signs[1] << 16)); + aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); + const __m256i s2_1 = __lasx_xvseq_b(aux256, mask2); + const __m256i q8s_1 = __lasx_xvsub_b(__lasx_xvxor_v(s2_1, q8_1), s2_1); + + aux256 = __lasx_xvreplgr2vr_w(signs[2] | ((uint32_t) signs[3] << 16)); + aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); + const __m256i s2_2 = __lasx_xvseq_b(aux256, mask2); + const __m256i q8s_2 = __lasx_xvsub_b(__lasx_xvxor_v(s2_2, q8_2), s2_2); + + signs += 4; + + const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); // blocks 2*ib32+0, 2*ib32+1 + const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); // blocks 2*ib32+2, 2*ib32+3 + + const __m256i p1 = lasx_madd_h(dot1, lasx_shuffle_b(scales16, get_scale_shuffle_k4(ib32+0))); + const __m256i p2 = lasx_madd_h(dot2, lasx_shuffle_b(scales16, get_scale_shuffle_k4(ib32+1))); + sumi1 = __lasx_xvadd_w(sumi1, p1); + sumi2 = __lasx_xvadd_w(sumi2, p2); + } + + accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); + } + + *s = 0.125f * hsum_float_8(accumf); + +#else + + float sumf = 0; + for (int i = 0; i < nb; i++) { + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint8_t * signs = qs + QK_K/8; + + int bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + int ls1 = 1 + 2*(x[i].scales[ib32] & 0xf); + int ls2 = 1 + 2*(x[i].scales[ib32] >> 4); + int sumi1 = 0, sumi2 = 0; + for (int l = 0; l < 2; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); + for (int j = 0; j < 8; ++j) { + sumi1 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + for (int l = 2; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); + for (int j = 0; j < 8; ++j) { + sumi2 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += ls1 * sumi1 + ls2 * sumi2; + qs += 4; + signs += 4; + } + + sumf += d * bsum; + } + + *s = 0.125f * sumf; + +#endif + +} + +void ggml_vec_dot_iq3_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq3_xxs * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + ggml_int8x16x4_t q3s; + ggml_int8x16x4_t q8b; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + float sumf1 = 0, sumf2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + memcpy(aux32, gas, 2*sizeof(uint32_t)); gas += 2*sizeof(uint32_t); + const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3xxs_grid[q3[ 0]], iq3xxs_grid[q3[ 1]], iq3xxs_grid[q3[ 2]], iq3xxs_grid[q3[ 3]]); + const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3xxs_grid[q3[ 4]], iq3xxs_grid[q3[ 5]], iq3xxs_grid[q3[ 6]], iq3xxs_grid[q3[ 7]]); + const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3xxs_grid[q3[ 8]], iq3xxs_grid[q3[ 9]], iq3xxs_grid[q3[10]], iq3xxs_grid[q3[11]]); + const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3xxs_grid[q3[12]], iq3xxs_grid[q3[13]], iq3xxs_grid[q3[14]], iq3xxs_grid[q3[15]]); + q3 += 16; + q3s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 7) & 127)))); + q3s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 21) & 127)))); + q3s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); + q3s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); + q3s.val[0] = vmulq_s8(q3s.val[0], vreinterpretq_s8_u32(aux32x4_0)); + q3s.val[1] = vmulq_s8(q3s.val[1], vreinterpretq_s8_u32(aux32x4_1)); + q3s.val[2] = vmulq_s8(q3s.val[2], vreinterpretq_s8_u32(aux32x4_2)); + q3s.val[3] = vmulq_s8(q3s.val[3], vreinterpretq_s8_u32(aux32x4_3)); + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); + sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[0] >> 28)); + sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[1] >> 28)); + } + sumf += d*(sumf1 + sumf2); + } + *s = 0.5f * sumf; + +#elif defined(__AVX2__) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q2_1 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + const __m256i q2_2 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + memcpy(aux32, gas, 8); gas += 8; + const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127], + signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); + const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = aux32[0] >> 28; + const uint16_t ls2 = aux32[1] >> 28; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.25f * hsum_float_8(accumf); + +#elif defined(__AVX__) + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q2_1_0 = _mm_set_epi32(iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + const __m128i q2_1_1 = _mm_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]]); + q3 += 8; + const __m128i q2_2_0 = _mm_set_epi32(iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + const __m128i q2_2_1 = _mm_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]]); + q3 += 8; + memcpy(aux32, gas, 8); gas += 8; + const __m128i s2_1_0 = _mm_set_epi64x(signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); + const __m128i s2_1_1 = _mm_set_epi64x(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127]); + const __m128i s2_2_0 = _mm_set_epi64x(signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m128i s2_2_1 = _mm_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127]); + const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, s2_1_0); + const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, s2_1_1); + const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, s2_2_0); + const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, s2_2_1); + const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); + const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); + const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); + const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); + const uint16_t ls1 = aux32[0] >> 28; + const uint16_t ls2 = aux32[1] >> 28; + const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); + sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); + sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); + sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); + sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); + } + + accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); + + } + + *s = 0.25f * hsum_float_8(accumf); + +#elif defined(__POWER9_VECTOR__) + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + const vector int v0 = vec_splats((int32_t)0); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + const uint8_t * restrict q3 = x[i].qs; + const uint32_t * restrict signs = (const uint32_t *)(x[i].qs + QK_K/4); + const int8_t * restrict q8 = y[i].qs; + +#pragma GCC unroll 1 + for (int j = 0; j < QK_K/32; j += 2) { + __builtin_prefetch(q3, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector unsigned int aux32x4_0 = {iq3xxs_grid[q3[ 0]], iq3xxs_grid[q3[ 1]], iq3xxs_grid[q3[ 2]], iq3xxs_grid[q3[ 3]]}; + vector unsigned int aux32x4_1 = {iq3xxs_grid[q3[ 4]], iq3xxs_grid[q3[ 5]], iq3xxs_grid[q3[ 6]], iq3xxs_grid[q3[ 7]]}; + vector unsigned int aux32x4_2 = {iq3xxs_grid[q3[ 8]], iq3xxs_grid[q3[ 9]], iq3xxs_grid[q3[10]], iq3xxs_grid[q3[11]]}; + vector unsigned int aux32x4_3 = {iq3xxs_grid[q3[12]], iq3xxs_grid[q3[13]], iq3xxs_grid[q3[14]], iq3xxs_grid[q3[15]]}; + q3 += 16; + + vector unsigned long long aux64x2_0 = {(uint64_t)(signs64[(signs[0] >> 0) & 127]), (uint64_t)(signs64[(signs[0] >> 7) & 127])}; + vector unsigned long long aux64x2_1 = {(uint64_t)(signs64[(signs[0] >> 14) & 127]), (uint64_t)(signs64[(signs[0] >> 21) & 127])}; + vector unsigned long long aux64x2_2 = {(uint64_t)(signs64[(signs[1] >> 0) & 127]), (uint64_t)(signs64[(signs[1] >> 7) & 127])}; + vector unsigned long long aux64x2_3 = {(uint64_t)(signs64[(signs[1] >> 14) & 127]), (uint64_t)(signs64[(signs[1] >> 21) & 127])}; + + vector signed char q3x0 = vec_mul((vector signed char)aux64x2_0, (vector signed char)aux32x4_0); + vector signed char q3x1 = vec_mul((vector signed char)aux64x2_1, (vector signed char)aux32x4_1); + vector signed char q3x2 = vec_mul((vector signed char)aux64x2_2, (vector signed char)aux32x4_2); + vector signed char q3x3 = vec_mul((vector signed char)aux64x2_3, (vector signed char)aux32x4_3); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q3x0, q8y0), vec_mulo(q3x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q3x1, q8y1), vec_mulo(q3x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q3x2, q8y2), vec_mulo(q3x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q3x3, q8y3), vec_mulo(q3x3, q8y3)); + + const uint16_t ls0 = (uint16_t)(signs[0] >> 28); + const uint16_t ls1 = (uint16_t)(signs[1] >> 28); + signs += 2; + + vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); + vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); + + vsumi0 = vec_msum(qv0, vscales01, vsumi0); + vsumi1 = vec_msum(qv1, vscales01, vsumi1); + vsumi2 = vec_msum(qv2, vscales23, vsumi2); + vsumi3 = vec_msum(qv3, vscales23, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = 0.25f * vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + __m256 accumf = (__m256)__lasx_xvldi(0); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q2_1 = lasx_set_w(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + const __m256i q2_2 = lasx_set_w(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + memcpy(aux32, gas, 8); gas += 8; + + const __m256i s2_1 = lasx_set_d(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127], + signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); + const __m256i s2_2 = lasx_set_d(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i q8s_1 = __lasx_xvsigncov_b(s2_1, q8_1); + const __m256i q8s_2 = __lasx_xvsigncov_b(s2_2, q8_2); + const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); + const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); + const uint16_t ls1 = aux32[0] >> 28; + const uint16_t ls2 = aux32[1] >> 28; + + const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); + const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); + sumi1 = __lasx_xvadd_w(sumi1, p1); + sumi2 = __lasx_xvadd_w(sumi2, p2); + } + + accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); + } + + *s = 0.25f * hsum_float_8(accumf); + +#else + + uint32_t aux32; + + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(&aux32, gas, sizeof(uint32_t)); gas += sizeof(uint32_t); + const uint32_t ls = 2*(aux32 >> 28) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3xxs_grid + q3[2*l+0]); + const uint8_t * grid2 = (const uint8_t *)(iq3xxs_grid + q3[2*l+1]); + const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*l) & 127]; + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + q3 += 8; + bsum += sumi * ls; + } + sumf += d * bsum; + } + *s = 0.25f * sumf; +#endif +} + +void ggml_vec_dot_iq3_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq3_s * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined(__ARM_NEON) + + typedef union { + uint16x8_t vec_index; + uint16_t index[8]; + } vec_index_t; + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; + + static const int16_t k_shift[8] = {8, 7, 6, 5, 4, 3, 2, 1}; + + const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); + const uint8x16_t mask2 = vld1q_u8(k_mask2); + + const int16x8_t hshift = vld1q_s16(k_shift); + const uint16x8_t m256 = vdupq_n_u16(256); + const uint8x16_t m1 = vdupq_n_u8(1); + + uint8x16x2_t vs; + ggml_int8x16x4_t q3s; + ggml_int8x16x4_t q8b; + vec_index_t idx; + + uint32_t scales32[2]; + const uint8_t * scales8 = (const uint8_t *)scales32; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(scales32, x[i].scales, 4); + scales32[1] = (((scales32[0] >> 4) & 0x0f0f0f0f) << 1) | 0x01010101; + scales32[0] = ((scales32[0] & 0x0f0f0f0f) << 1) | 0x01010101; + + int sumi1 = 0, sumi2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + + const uint8x16_t idx_l = vld1q_u8(qs); qs += 16; + idx.vec_index = vorrq_u16(vmovl_u8(vget_low_u8 (idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+0]), hshift), m256)); + const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], + iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); + const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], + iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); + idx.vec_index = vorrq_u16(vmovl_u8(vget_high_u8(idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+1]), hshift), m256)); + const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], + iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); + const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], + iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); + + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); + vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); + + q3s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_0)); + q3s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_1)); + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); + vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); + + signs += 4; + + q3s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_2)); + q3s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_3)); + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); + + sumi1 += vaddvq_s32(p1) * scales8[ib32/2+0]; + sumi2 += vaddvq_s32(p2) * scales8[ib32/2+4]; + } + sumf += d*(sumi1 + sumi2); + } + *s = sumf; + +#elif defined(__AVX2__) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); + const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); + + const __m256i idx_shift = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); + const __m256i idx_mask = _mm256_set1_epi32(256); + + typedef union { + __m256i vec[2]; + uint32_t index[16]; + } index_t; + + index_t idx; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i idx_l = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i *)qs)); qs += 16; + idx.vec[0] = _mm256_set1_epi32(qh[ib32+0]); + idx.vec[1] = _mm256_set1_epi32(qh[ib32+1]); + idx.vec[0] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[0], idx_shift), idx_mask); + idx.vec[1] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[1], idx_shift), idx_mask); + idx.vec[0] = _mm256_or_si256(idx.vec[0], _mm256_cvtepi16_epi32(_mm256_castsi256_si128(idx_l))); + idx.vec[1] = _mm256_or_si256(idx.vec[1], _mm256_cvtepi16_epi32(_mm256_extractf128_si256(idx_l, 1))); + + // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange. + //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4); + //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4); + const __m256i q2_1 = _mm256_set_epi32( + iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]], + iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]] + ); + const __m256i q2_2 = _mm256_set_epi32( + iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]], + iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]] + ); + + __m256i aux256 = _mm256_set1_epi32(signs[0] | (signs[1] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); + + aux256 = _mm256_set1_epi32(signs[2] | (signs[3] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); + + signs += 4; + + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; + const uint16_t ls2 = x[i].scales[ib32/2] >> 4; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = hsum_float_8(accumf); + +#elif defined(__AVX__) + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m128i mask1_0 = _mm_loadu_si128((const __m128i*)k_mask1); + const __m128i mask1_1 = _mm_loadu_si128((const __m128i*)k_mask1 + 1); + const __m128i mask2_0 = _mm_loadu_si128((const __m128i*)k_mask2); + const __m128i mask2_1 = _mm_loadu_si128((const __m128i*)k_mask2 + 1); + + const __m128i idx_mul_0 = _mm_set_epi32(32, 64, 128, 256); + const __m128i idx_mul_1 = _mm_set_epi32(2, 4, 8, 16); + const __m128i idx_mask = _mm_set1_epi32(256); + + typedef union { + __m128i vec[4]; + uint32_t index[16]; + } index_t; + + index_t idx; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i qs_tmp = _mm_loadu_si128((const __m128i *)qs); + const __m128i idx_l_0 = _mm_cvtepu8_epi16(qs_tmp); + const __m128i idx_l_1 = _mm_cvtepu8_epi16(_mm_srli_si128(qs_tmp, 8)); qs += 16; + idx.vec[0] = _mm_set1_epi32(qh[ib32+0]); + idx.vec[1] = idx.vec[0]; + idx.vec[2] = _mm_set1_epi32(qh[ib32+1]); + idx.vec[3] = idx.vec[2]; + + idx.vec[0] = _mm_and_si128(_mm_mullo_epi32(idx.vec[0], idx_mul_0), idx_mask); + idx.vec[1] = _mm_and_si128(_mm_mullo_epi32(idx.vec[1], idx_mul_1), idx_mask); + idx.vec[2] = _mm_and_si128(_mm_mullo_epi32(idx.vec[2], idx_mul_0), idx_mask); + idx.vec[3] = _mm_and_si128(_mm_mullo_epi32(idx.vec[3], idx_mul_1), idx_mask); + + idx.vec[0] = _mm_or_si128(idx.vec[0], _mm_cvtepi16_epi32(idx_l_0)); + idx.vec[1] = _mm_or_si128(idx.vec[1], _mm_cvtepi16_epi32(_mm_srli_si128(idx_l_0, 8))); + idx.vec[2] = _mm_or_si128(idx.vec[2], _mm_cvtepi16_epi32(idx_l_1)); + idx.vec[3] = _mm_or_si128(idx.vec[3], _mm_cvtepi16_epi32(_mm_srli_si128(idx_l_1, 8))); + + const __m128i q2_1_0 = _mm_set_epi32(iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]]); + const __m128i q2_1_1 = _mm_set_epi32(iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]]); + const __m128i q2_2_0 = _mm_set_epi32(iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[9]], iq3s_grid[idx.index[8]]); + const __m128i q2_2_1 = _mm_set_epi32(iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]]); + + __m128i aux128_0 = _mm_set1_epi32(signs[0] | (signs[1] << 16)); + __m128i aux128_1 = aux128_0; + aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); + aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); + const __m128i s2_1_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); + const __m128i s2_1_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); + const __m128i q8s_1_0 = _mm_sub_epi8(_mm_xor_si128(s2_1_0, q8_1_0), s2_1_0); + const __m128i q8s_1_1 = _mm_sub_epi8(_mm_xor_si128(s2_1_1, q8_1_1), s2_1_1); + + aux128_0 = _mm_set1_epi32(signs[2] | (signs[3] << 16)); + aux128_1 = aux128_0; + aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); + aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); + const __m128i s2_2_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); + const __m128i s2_2_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); + const __m128i q8s_2_0 = _mm_sub_epi8(_mm_xor_si128(s2_2_0, q8_2_0), s2_2_0); + const __m128i q8s_2_1 = _mm_sub_epi8(_mm_xor_si128(s2_2_1, q8_2_1), s2_2_1); + + signs += 4; + + const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); + const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); + const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); + const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); + const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; + const uint16_t ls2 = x[i].scales[ib32/2] >> 4; + const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); + sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); + sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); + sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); + sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); + } + + accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); + + } + + *s = hsum_float_8(accumf); + +#elif defined(__POWER9_VECTOR__) + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; + + const vector int v0 = vec_splats((int32_t)0); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + const vector unsigned char mask0 = vec_xl( 0, k_mask1); + const vector unsigned char mask1 = vec_xl(16, k_mask1); + const vector signed char mask2 = (vector signed char)vec_xl( 0, k_mask2); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].signs); + const uint8_t * restrict sc = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + for (int j = 0; j < QK_K/32; j += 2) { + __builtin_prefetch(q3, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector unsigned int aux32x4_0 = {iq3s_grid[q3[ 0] | ((qh[0] << 8) & 256)], iq3s_grid[q3[ 1] | ((qh[0] << 7) & 256)], + iq3s_grid[q3[ 2] | ((qh[0] << 6) & 256)], iq3s_grid[q3[ 3] | ((qh[0] << 5) & 256)]}; + vector unsigned int aux32x4_1 = {iq3s_grid[q3[ 4] | ((qh[0] << 4) & 256)], iq3s_grid[q3[ 5] | ((qh[0] << 3) & 256)], + iq3s_grid[q3[ 6] | ((qh[0] << 2) & 256)], iq3s_grid[q3[ 7] | ((qh[0] << 1) & 256)]}; + vector unsigned int aux32x4_2 = {iq3s_grid[q3[ 8] | ((qh[1] << 8) & 256)], iq3s_grid[q3[ 9] | ((qh[1] << 7) & 256)], + iq3s_grid[q3[10] | ((qh[1] << 6) & 256)], iq3s_grid[q3[11] | ((qh[1] << 5) & 256)]}; + vector unsigned int aux32x4_3 = {iq3s_grid[q3[12] | ((qh[1] << 4) & 256)], iq3s_grid[q3[13] | ((qh[1] << 3) & 256)], + iq3s_grid[q3[14] | ((qh[1] << 2) & 256)], iq3s_grid[q3[15] | ((qh[1] << 1) & 256)]}; + q3 += 16; + qh += 2; + + vector signed char vsigns01 = (vector signed char)vec_splats(*(const uint32_t *)&signs[0]); + vector signed char vsigns02 = (vector signed char)vec_splats(*(const uint32_t *)&signs[2]); + signs += 4; + + vector signed char vsigns0 = vec_perm(vsigns01, vsigns01, mask0); + vector signed char vsigns1 = vec_perm(vsigns01, vsigns01, mask1); + vector signed char vsigns2 = vec_perm(vsigns02, vsigns02, mask0); + vector signed char vsigns3 = vec_perm(vsigns02, vsigns02, mask1); + + vsigns0 = (vector signed char)vec_cmpeq(vec_and(vsigns0, mask2), mask2); + vsigns1 = (vector signed char)vec_cmpeq(vec_and(vsigns1, mask2), mask2); + vsigns2 = (vector signed char)vec_cmpeq(vec_and(vsigns2, mask2), mask2); + vsigns3 = (vector signed char)vec_cmpeq(vec_and(vsigns3, mask2), mask2); + + vector signed char q3x0 = vec_sub(vec_xor(vsigns0, (vector signed char)aux32x4_0), vsigns0); + vector signed char q3x1 = vec_sub(vec_xor(vsigns1, (vector signed char)aux32x4_1), vsigns1); + vector signed char q3x2 = vec_sub(vec_xor(vsigns2, (vector signed char)aux32x4_2), vsigns2); + vector signed char q3x3 = vec_sub(vec_xor(vsigns3, (vector signed char)aux32x4_3), vsigns3); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q3x0, q8y0), vec_mulo(q3x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q3x1, q8y1), vec_mulo(q3x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q3x2, q8y2), vec_mulo(q3x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q3x3, q8y3), vec_mulo(q3x3, q8y3)); + + const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); + const uint16_t ls1 = (uint16_t)(sc[0] >> 4); + sc ++; + + vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); + vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); + + vsumi0 = vec_msum(qv0, vscales01, vsumi0); + vsumi1 = vec_msum(qv1, vscales01, vsumi1); + vsumi2 = vec_msum(qv2, vscales23, vsumi2); + vsumi3 = vec_msum(qv3, vscales23, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m256i mask1 = __lasx_xvld((const __m256i*)k_mask1, 0); + const __m256i mask2 = __lasx_xvld((const __m256i*)k_mask2, 0); + + __m256i idx_shift = lasx_set_w(1, 2, 3, 4, 5, 6, 7, 8); + const __m256i idx_mask = __lasx_xvreplgr2vr_w(256); + + typedef union { + __m256i vec[2]; + uint32_t index[16]; + } index_t; + + index_t idx; + + __m256 accumf = (__m256)__lasx_xvldi(0); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i idx_l = lasx_extu8_16(__lsx_vld(qs, 0)); qs += 16; + idx.vec[0] = __lasx_xvreplgr2vr_w(qh[ib32+0]); + idx.vec[1] = __lasx_xvreplgr2vr_w(qh[ib32+1]); + idx.vec[0] = __lasx_xvand_v(__lasx_xvsll_w(idx.vec[0], idx_shift), idx_mask); + idx.vec[1] = __lasx_xvand_v(__lasx_xvsll_w(idx.vec[1], idx_shift), idx_mask); + idx.vec[0] = __lasx_xvor_v(idx.vec[0], lasx_ext16_32(lasx_extracti128(idx_l, 0))); + idx.vec[1] = __lasx_xvor_v(idx.vec[1], lasx_ext16_32(lasx_extracti128(idx_l, 1))); + + // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange. + //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4); + //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4); + const __m256i q2_1 = lasx_set_w( + iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]], + iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]] + ); + const __m256i q2_2 = lasx_set_w( + iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]], + iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]] + ); + + __m256i aux256 = __lasx_xvreplgr2vr_w(signs[0] | (signs[1] << 16)); + aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); + const __m256i s2_1 = __lasx_xvseq_b(aux256, mask2); + const __m256i q8s_1 = __lasx_xvsub_b(__lasx_xvxor_v(s2_1, q8_1), s2_1); + + aux256 = __lasx_xvreplgr2vr_w(signs[2] | (signs[3] << 16)); + aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); + const __m256i s2_2 = __lasx_xvseq_b(aux256, mask2); + const __m256i q8s_2 = __lasx_xvsub_b(__lasx_xvxor_v(s2_2, q8_2), s2_2); + + signs += 4; + + const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); + const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); + const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; + const uint16_t ls2 = x[i].scales[ib32/2] >> 4; + const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); + const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); + sumi1 = __lasx_xvadd_w(sumi1, p1); + sumi2 = __lasx_xvadd_w(sumi2, p2); + } + + accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); + } + + *s = hsum_float_8(accumf); + +#else + + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint8_t * restrict signs = x[i].signs; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const uint32_t ls1 = 2*(x[i].scales[ib32/2] & 0xf) + 1; + const uint32_t ls2 = 2*(x[i].scales[ib32/2] >> 4) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+0] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+0] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + qs += 8; + signs += 4; + bsum += sumi * ls1; + sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+1] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+1] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + qs += 8; + signs += 4; + bsum += sumi * ls2; + } + sumf += d * bsum; + } + *s = sumf; +#endif +} + +#if defined(__AVX2__) +static inline __m256i mul_add_epi8(const __m256i x, const __m256i y) { + const __m256i ax = _mm256_sign_epi8(x, x); + const __m256i sy = _mm256_sign_epi8(y, x); + return _mm256_maddubs_epi16(ax, sy); +} +#elif defined(__loongarch_asx) +static inline __m256i mul_add_epi8(const __m256i x, const __m256i y) { + const __m256i ax = __lasx_xvsigncov_b(x, x); + const __m256i sy = __lasx_xvsigncov_b(x, y); + __m256i tmp1, tmp2, tmp3; + tmp1 = __lasx_xvmulwev_h_bu_b(ax, sy); + tmp2 = __lasx_xvmulwod_h_bu_b(ax, sy); + tmp3 = __lasx_xvadd_h(tmp1, tmp2); + return __lasx_xvsat_h(tmp3, 15); +} +#endif + +void ggml_vec_dot_iq1_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq1_s * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined __ARM_NEON + + ggml_int8x16x4_t q1b; + ggml_int8x16x4_t q8b; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + int sumi1 = 0, sumi2 = 0, sumi3 = 0; + + for (int ib = 0; ib < QK_K/32; ib += 2) { + + q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[ib+0] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[ib+0] << 5) & 0x700))))); + q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[ib+0] << 2) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[ib+0] >> 1) & 0x700))))); + q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[ib+1] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[ib+1] << 5) & 0x700))))); + q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[ib+1] << 2) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[ib+1] >> 1) & 0x700))))); + qs += 8; + + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[0], q8b.val[0]), q1b.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[2], q8b.val[2]), q1b.val[3], q8b.val[3]); + + const int ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + sumi1 += vaddvq_s32(p1) * ls1; + sumi2 += vaddvq_s32(p2) * ls2; + sumi3 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * ls1 * (qh[ib+0] & 0x8000 ? -1 : 1) + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * ls2 * (qh[ib+1] & 0x8000 ? -1 : 1); + + } + + sumf += y[i].d * GGML_FP16_TO_FP32(x[i].d) * (sumi1 + sumi2 + IQ1S_DELTA * sumi3); + } + + *s = sumf; + +#elif defined __AVX2__ + + __m256 accum = _mm256_setzero_ps(); + float accum1 = 0; + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + __m256i sumi = _mm256_setzero_si256(); + int sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m256i q1b_1 = _mm256_set_epi64x(iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)], + iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)]); + const __m256i q1b_2 = _mm256_set_epi64x(iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)], + iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)]); + qs += 8; + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); + const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); + const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(ls1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(ls2)); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p1, p2)); + sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; + } + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + accum = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sumi), accum); + accum1 += d * sumi1; + + } + + *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; + +#elif defined __AVX__ + __m256 accum = _mm256_setzero_ps(); + float accum1 = 0; + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + int sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q1b_1_0 = _mm_set_epi64x(iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)]); + const __m128i q1b_1_1 = _mm_set_epi64x(iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)]); + const __m128i q1b_2_0 = _mm_set_epi64x(iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)]); + const __m128i q1b_2_1 = _mm_set_epi64x(iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)]); + qs += 8; + const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + + const __m128i dot1_0 = mul_add_epi8_sse(q1b_1_0, q8b_1_0); + const __m128i dot1_1 = mul_add_epi8_sse(q1b_1_1, q8b_1_1); + const __m128i dot2_0 = mul_add_epi8_sse(q1b_2_0, q8b_2_0); + const __m128i dot2_1 = mul_add_epi8_sse(q1b_2_1, q8b_2_1); + const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(ls1)); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(ls1)); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(ls2)); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(ls2)); + + sumi1_0 = _mm_add_epi32(sumi1_0, _mm_add_epi32(p1_0, p2_0)); + sumi1_1 = _mm_add_epi32(sumi1_1, _mm_add_epi32(p1_1, p2_1)); + sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; + } + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + accum = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(sumi1_1, sumi1_0))), accum); + accum1 += d * sumi1; + + } + + *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; + +#elif defined(__POWER9_VECTOR__) + const vector unsigned char v0 = vec_splats((unsigned char)0x0); + const vector unsigned short vsign = vec_splats((unsigned short)0x8000); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + for (int i = 0; i < nb; ++i) { + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); + vector float vyd = vec_splats(y[i].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = vec_splats((int32_t)0); + vector signed int vsumi1 = vec_splats((int32_t)0); + vector signed int vsumi2 = vec_splats((int32_t)0); + vector signed int vsumi3 = vec_splats((int32_t)0); + vector signed int vsumi8 = vec_splats((int32_t)0); + + const uint8_t * restrict q1 = x[i].qs; + const uint16_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + const int16_t * restrict qs = y[i].bsums; + + for (int j = 0; j < QK_K/32; j += 2) { + __builtin_prefetch(q1, 0, 1); + __builtin_prefetch(qh, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed long long aux64x2_0 = {*(const int64_t *)(iq1s_grid + (q1[0] | ((qh[0] << 8) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[1] | ((qh[0] << 5) & 0x700)))}; + vector signed long long aux64x2_1 = {*(const int64_t *)(iq1s_grid + (q1[2] | ((qh[0] << 2) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[3] | ((qh[0] >> 1) & 0x700)))}; + vector signed long long aux64x2_2 = {*(const int64_t *)(iq1s_grid + (q1[4] | ((qh[1] << 8) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[5] | ((qh[1] << 5) & 0x700)))}; + vector signed long long aux64x2_3 = {*(const int64_t *)(iq1s_grid + (q1[6] | ((qh[1] << 2) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[7] | ((qh[1] >> 1) & 0x700)))}; + q1 += 8; + + vector signed char q1x0 = (vector signed char)aux64x2_0; + vector signed char q1x1 = (vector signed char)aux64x2_1; + vector signed char q1x2 = (vector signed char)aux64x2_2; + vector signed char q1x3 = (vector signed char)aux64x2_3; + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q1x0, q8y0), vec_mulo(q1x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q1x1, q8y1), vec_mulo(q1x1, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q1x2, q8y2), vec_mulo(q1x2, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q1x3, q8y3), vec_mulo(q1x3, q8y3)); + + const uint16_t ls0 = (uint16_t)((qh[0] >> 12) & 7); + const uint16_t ls1 = (uint16_t)((qh[1] >> 12) & 7); + + vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); + vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); + vector signed short vscales = vec_sld(vscales23, vscales01, 8); + + vsumi0 = vec_msum(qv0, vscales01, vsumi0); + vsumi1 = vec_msum(qv1, vscales01, vsumi1); + vsumi2 = vec_msum(qv2, vscales23, vsumi2); + vsumi3 = vec_msum(qv3, vscales23, vsumi3); + + vector signed short q8ysums = vec_xl_len(qs, 8); + qs += 4; + q8ysums = vec_mergeh(q8ysums, (vector signed short)v0); + + vector signed short qxh = (vector signed short)vec_sld(vec_splats(qh[1]), vec_splats(qh[0]), 8); + qh += 2; + vector __bool short vsel = vec_cmpge(qxh, (vector signed short)v0); + + vector signed short q8ysum = vec_sel((vector signed short)vec_xor((vector unsigned short)q8ysums, vsign), q8ysums, vsel); + + vsumi8 = vec_add(vec_mule(q8ysum, vscales), vsumi8); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + + vsumf0 = vec_madd(vec_ctf(vsumi8, 0), vec_mul(vd, vec_splats(IQ1S_DELTA)), vsumf0); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + __m256 accum = (__m256)__lasx_xvldi(0); + float accum1 = 0; + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + __m256i sumi = __lasx_xvldi(0); + int sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ib += 2) { + __m256i q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)], 0); + q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], 1); + q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)], 2); + q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], 3); + + __m256i q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)], 0); + q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], 1); + q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)], 2); + q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], 3); + + qs += 8; + const __m256i q8b_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + const __m256i q8b_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; + + const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); + const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); + const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + + __m256i tmp1, tmp5, tmp6; + tmp1 = __lasx_xvreplgr2vr_h(ls1); + tmp5 = __lasx_xvmulwev_w_h(dot1, tmp1); + tmp6 = __lasx_xvmulwod_w_h(dot1, tmp1); + const __m256i p1 = __lasx_xvadd_w(tmp5, tmp6); + + tmp1 = __lasx_xvreplgr2vr_h(ls2); + tmp5 = __lasx_xvmulwev_w_h(dot2, tmp1); + tmp6 = __lasx_xvmulwod_w_h(dot2, tmp1); + const __m256i p2 = __lasx_xvadd_w(tmp5, tmp6); + + sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p1, p2)); + sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; + } + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + accum = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), accum); + accum1 += d * sumi1; + } + + *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; + +#else + + float sumf = 0; + for (int i = 0; i < nb; i++) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + int sumi = 0, sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ++ib) { + const int ls = 2*((qh[ib] >> 12) & 7) + 1; + const int delta = qh[ib] & 0x8000 ? -1 : 1; + int lsum = 0; + for (int l = 0; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((qh[ib] >> 3*l) & 7) << 8))); + for (int j = 0; j < 8; ++j) { + lsum += q8[j] * grid[j]; + } + q8 += 8; + } + sumi += ls * lsum; + sumi1 += ls * delta * (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]); + qs += 4; + } + + sumf += GGML_FP16_TO_FP32(x[i].d) * y[i].d * (sumi + IQ1S_DELTA * sumi1); + } + + *s = sumf; + +#endif +} + +void ggml_vec_dot_iq1_m_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_iq1_m * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + + iq1m_scale_t scale; + +#if defined __ARM_NEON + const int32x4_t mask = vdupq_n_s32(0x7); + const int32x4_t mone = vdupq_n_s32(1); + const int32x4_t mzero = vdupq_n_s32(0); + + ggml_int8x16x4_t deltas; + deltas.val[0] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(+1)); + deltas.val[1] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(+1)); + deltas.val[2] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(-1)); + deltas.val[3] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(-1)); + + ggml_int8x16x4_t q1b; + ggml_int8x16x4_t q8b; + + uint32_t aux32; + const uint8_t * aux8 = (const uint8_t *)&aux32; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; + + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); + + int32x4_t sumi1 = mzero; + int32x4_t sumi2 = mzero; + + for (int ib = 0; ib < QK_K/32; ib += 2) { + + q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[0] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[0] << 4) & 0x700))))); + q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[1] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[1] << 4) & 0x700))))); + q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[2] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[2] << 4) & 0x700))))); + q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[3] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[3] << 4) & 0x700))))); + + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + + const int32x4_t p1 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[0], q8b.val[0]), ggml_vdotq_s32(mzero, q1b.val[1], q8b.val[1])); + const int32x4_t p2 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[2], q8b.val[2]), ggml_vdotq_s32(mzero, q1b.val[3], q8b.val[3])); + const int32x4_t p12 = vpaddq_s32(p1, p2); + + const uint32_t * qh32 = (const uint32_t *)qh; // we are 4-byte aligned, so we can do that + aux32 = ((qh32[0] >> 3) & 0x01010101) | ((qh32[0] >> 6) & 0x02020202); + + const int32x4_t p3 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[0]], q8b.val[0]), ggml_vdotq_s32(mzero, deltas.val[aux8[1]], q8b.val[1])); + const int32x4_t p4 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[2]], q8b.val[2]), ggml_vdotq_s32(mzero, deltas.val[aux8[3]], q8b.val[3])); + const int32x4_t p34 = vpaddq_s32(p3, p4); + + int32x4_t scales_4 = ggml_vld1q_u32(sc[ib/2] >> 0, sc[ib/2] >> 3, sc[ib/2] >> 6, sc[ib/2] >> 9); + + scales_4 = vaddq_s32(vshlq_n_s32(vandq_s32(scales_4, mask), 1), mone); + + sumi1 = vmlaq_s32(sumi1, scales_4, p12); + sumi2 = vmlaq_s32(sumi2, scales_4, p34); + + qs += 8; qh += 4; + + } + + sumf += y[i].d * GGML_FP16_TO_FP32(scale.f16) * (vaddvq_s32(sumi1) + IQ1M_DELTA * vaddvq_s32(sumi2)); + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m256i mask = _mm256_set1_epi16(0x7); + const __m256i mone = _mm256_set1_epi16(1); + + __m256 accum1 = _mm256_setzero_ps(); + __m256 accum2 = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; + + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); + + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m256i q1b_1 = _mm256_set_epi64x( + iq1s_grid[qs[3] | (((uint16_t)qh[1] << 4) & 0x700)], iq1s_grid[qs[2] | (((uint16_t)qh[1] << 8) & 0x700)], + iq1s_grid[qs[1] | (((uint16_t)qh[0] << 4) & 0x700)], iq1s_grid[qs[0] | (((uint16_t)qh[0] << 8) & 0x700)] + ); + const __m256i q1b_2 = _mm256_set_epi64x( + iq1s_grid[qs[7] | (((uint16_t)qh[3] << 4) & 0x700)], iq1s_grid[qs[6] | (((uint16_t)qh[3] << 8) & 0x700)], + iq1s_grid[qs[5] | (((uint16_t)qh[2] << 4) & 0x700)], iq1s_grid[qs[4] | (((uint16_t)qh[2] << 8) & 0x700)] + ); + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); + const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); + + const __m256i delta1 = _mm256_set_epi64x(qh[1] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[1] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, + qh[0] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[0] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + const __m256i delta2 = _mm256_set_epi64x(qh[3] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[3] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, + qh[2] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[2] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + + const __m256i dot3 = mul_add_epi8(delta1, q8b_1); + const __m256i dot4 = mul_add_epi8(delta2, q8b_2); + + __m256i scale1 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 3), _mm_set1_epi16(sc[ib/2] >> 0)); + __m256i scale2 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 9), _mm_set1_epi16(sc[ib/2] >> 6)); + + scale1 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale1, mask), 1), mone); + scale2 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale2, mask), 1), mone); + const __m256i p1 = _mm256_madd_epi16(dot1, scale1); + const __m256i p2 = _mm256_madd_epi16(dot2, scale2); + const __m256i p3 = _mm256_madd_epi16(dot3, scale1); + const __m256i p4 = _mm256_madd_epi16(dot4, scale2); + + sumi1 = _mm256_add_epi32(sumi1, _mm256_add_epi32(p1, p2)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_add_epi32(p3, p4)); + + qs += 8; qh += 4; + } + + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(scale.f16)); + + accum1 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi1), accum1); + accum2 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi2), accum2); + } + + *s = hsum_float_8(accum1) + IQ1M_DELTA * hsum_float_8(accum2); + +#elif defined __AVX__ + const __m128i mask = _mm_set1_epi16(0x7); + const __m128i mone = _mm_set1_epi16(1); + + __m256 accum1 = _mm256_setzero_ps(); + __m256 accum2 = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; + + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); + + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q1b_1_0 = _mm_set_epi64x( + iq1s_grid[qs[1] | (((uint16_t)qh[0] << 4) & 0x700)], iq1s_grid[qs[0] | (((uint16_t)qh[0] << 8) & 0x700)]); + const __m128i q1b_1_1 = _mm_set_epi64x( + iq1s_grid[qs[3] | (((uint16_t)qh[1] << 4) & 0x700)], iq1s_grid[qs[2] | (((uint16_t)qh[1] << 8) & 0x700)]); + const __m128i q1b_2_0 = _mm_set_epi64x( + iq1s_grid[qs[5] | (((uint16_t)qh[2] << 4) & 0x700)], iq1s_grid[qs[4] | (((uint16_t)qh[2] << 8) & 0x700)]); + const __m128i q1b_2_1 = _mm_set_epi64x( + iq1s_grid[qs[7] | (((uint16_t)qh[3] << 4) & 0x700)], iq1s_grid[qs[6] | (((uint16_t)qh[3] << 8) & 0x700)]); + const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + + const __m128i dot1_0 = mul_add_epi8_sse(q1b_1_0, q8b_1_0); + const __m128i dot1_1 = mul_add_epi8_sse(q1b_1_1, q8b_1_1); + const __m128i dot2_0 = mul_add_epi8_sse(q1b_2_0, q8b_2_0); + const __m128i dot2_1 = mul_add_epi8_sse(q1b_2_1, q8b_2_1); + + const __m128i delta1_0 = _mm_set_epi64x(qh[0] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[0] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + const __m128i delta1_1 = _mm_set_epi64x(qh[1] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[1] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + const __m128i delta2_0 = _mm_set_epi64x(qh[2] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[2] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + const __m128i delta2_1 = _mm_set_epi64x(qh[3] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[3] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + + const __m128i dot3_0 = mul_add_epi8_sse(delta1_0, q8b_1_0); + const __m128i dot3_1 = mul_add_epi8_sse(delta1_1, q8b_1_1); + const __m128i dot4_0 = mul_add_epi8_sse(delta2_0, q8b_2_0); + const __m128i dot4_1 = mul_add_epi8_sse(delta2_1, q8b_2_1); + + __m128i scale1_0 = _mm_set1_epi16(sc[ib/2] >> 0); + __m128i scale1_1 = _mm_set1_epi16(sc[ib/2] >> 3); + __m128i scale2_0 = _mm_set1_epi16(sc[ib/2] >> 6); + __m128i scale2_1 = _mm_set1_epi16(sc[ib/2] >> 9); + + scale1_0 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale1_0, mask), 1), mone); + scale1_1 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale1_1, mask), 1), mone); + scale2_0 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale2_0, mask), 1), mone); + scale2_1 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale2_1, mask), 1), mone); + const __m128i p1_0 = _mm_madd_epi16(dot1_0, scale1_0); + const __m128i p1_1 = _mm_madd_epi16(dot1_1, scale1_1); + const __m128i p2_0 = _mm_madd_epi16(dot2_0, scale2_0); + const __m128i p2_1 = _mm_madd_epi16(dot2_1, scale2_1); + const __m128i p3_0 = _mm_madd_epi16(dot3_0, scale1_0); + const __m128i p3_1 = _mm_madd_epi16(dot3_1, scale1_1); + const __m128i p4_0 = _mm_madd_epi16(dot4_0, scale2_0); + const __m128i p4_1 = _mm_madd_epi16(dot4_1, scale2_1); + + sumi1_0 = _mm_add_epi32(sumi1_0, _mm_add_epi32(p1_0, p2_0)); + sumi1_1 = _mm_add_epi32(sumi1_1, _mm_add_epi32(p1_1, p2_1)); + sumi2_0 = _mm_add_epi32(sumi2_0, _mm_add_epi32(p3_0, p4_0)); + sumi2_1 = _mm_add_epi32(sumi2_1, _mm_add_epi32(p3_1, p4_1)); + + qs += 8; qh += 4; + } + + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(scale.f16)); + + accum1 = _mm256_add_ps(_mm256_mul_ps(d, _mm256_cvtepi32_ps(MM256_SET_M128I(sumi1_1, sumi1_0))), accum1); + accum2 = _mm256_add_ps(_mm256_mul_ps(d, _mm256_cvtepi32_ps(MM256_SET_M128I(sumi2_1, sumi2_0))), accum2); + } + + *s = hsum_float_8(accum1) + IQ1M_DELTA * hsum_float_8(accum2); + +#else + + int sum1[2], sum2[2], delta[4]; + + float sumf = 0; + for (int i = 0; i < nb; i++) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; + + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); + + int sumi1 = 0, sumi2 = 0; + for (int ib = 0; ib < QK_K/32; ++ib) { + delta[0] = qh[0] & 0x08 ? -1 : 1; + delta[1] = qh[0] & 0x80 ? -1 : 1; + delta[2] = qh[1] & 0x08 ? -1 : 1; + delta[3] = qh[1] & 0x80 ? -1 : 1; + sum1[0] = sum1[1] = sum2[0] = sum2[1] = 0; + for (int l = 0; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((uint16_t)qh[l/2] << (8 - 4*(l%2))) & 0x700))); + int lsum1 = 0, lsum2 = 0; + for (int j = 0; j < 8; ++j) { + lsum1 += q8[j] * grid[j]; + lsum2 += q8[j]; + } + q8 += 8; + sum1[l/2] += lsum1; + sum2[l/2] += lsum2*delta[l]; + } + + const int ls1 = 2*((sc[ib/2] >> (6*(ib%2)+0)) & 0x7) + 1; + const int ls2 = 2*((sc[ib/2] >> (6*(ib%2)+3)) & 0x7) + 1; + + sumi1 += sum1[0] * ls1 + sum1[1] * ls2; + sumi2 += sum2[0] * ls1 + sum2[1] * ls2; + qs += 4; + qh += 2; + } + + sumf += GGML_FP16_TO_FP32(scale.f16) * y[i].d * (sumi1 + IQ1M_DELTA * sumi2); + } + + *s = sumf; + +#endif +} + +void ggml_vec_dot_iq4_nl_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + assert(n % QK4_NL == 0); + static_assert(QK4_NL == QK8_0, "QK4_NL and QK8_0 must be the same"); + + const block_iq4_nl * restrict x = vx; + const block_q8_0 * restrict y = vy; + + const int nb = n / QK4_NL; + + int ib = 0; + float sumf = 0; + +#if defined __ARM_NEON + const int8x16_t values = vld1q_s8(kvalues_iq4nl); + const uint8x16_t m4b = vdupq_n_u8(0x0f); + uint8x16x2_t q4bits; + int8x16x4_t q4b; + int8x16x4_t q8b; + int32x4_t prod_1, prod_2; + + for (; ib + 1 < nb; ib += 2) { + + q4bits.val[0] = vld1q_u8(x[ib + 0].qs); + q4bits.val[1] = vld1q_u8(x[ib + 1].qs); + q8b.val[0] = vld1q_s8(y[ib + 0].qs); + q8b.val[1] = vld1q_s8(y[ib + 0].qs + 16); + q8b.val[2] = vld1q_s8(y[ib + 1].qs); + q8b.val[3] = vld1q_s8(y[ib + 1].qs + 16); + + q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); + q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); + q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); + q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); + + prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); + prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); + + sumf += + GGML_FP16_TO_FP32(x[ib+0].d) * GGML_FP16_TO_FP32(y[ib + 0].d) * vaddvq_s32(prod_1) + + GGML_FP16_TO_FP32(x[ib+1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) * vaddvq_s32(prod_2); + } + +#elif defined __AVX2__ + + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + const __m256i mone = _mm256_set1_epi16(1); + + __m256 accum1 = _mm256_setzero_ps(); + __m256 accum2 = _mm256_setzero_ps(); + for (; ib + 1 < nb; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)x[ib + 0].qs); + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)x[ib + 1].qs); + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)y[ib + 0].qs); + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)y[ib + 1].qs); + const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); + const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const __m256i p_1 = _mm256_madd_epi16(p16_1, mone); + const __m256i p_2 = _mm256_madd_epi16(p16_2, mone); + accum1 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), + _mm256_cvtepi32_ps(p_1), accum1); + accum2 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), + _mm256_cvtepi32_ps(p_2), accum2); + } + + sumf = hsum_float_8(_mm256_add_ps(accum1, accum2)); + +#elif defined __AVX__ + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + + __m256 accum = _mm256_setzero_ps(); + for (; ib + 1 < nb; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)x[ib + 0].qs); + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); + const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs); + const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs + 1); + const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); + const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs + 1); + + const __m128i q4b_1_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)); + const __m128i q4b_1_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)); + const __m128i q4b_2_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)); + const __m128i q4b_2_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)); + + const __m256 p = mul_sum_i8_quad_float(q4b_1_0, q4b_1_1, q4b_2_0, q4b_2_1, q8b_1_0, q8b_1_1, q8b_2_0, q8b_2_1); + const __m256 deltas = quad_fp16_delta_float(x[ib].d, y[ib].d, x[ib + 1].d, y[ib + 1].d); + accum = _mm256_add_ps(_mm256_mul_ps(deltas, p), accum); + } + + sumf = hsum_float_8(accum); + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector signed int v0 = vec_splats((int32_t)0); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + + const vector signed char values = vec_xl( 0, kvalues_iq4nl); + +#pragma GCC unroll 4 + for (; ib < nb; ++ib) { + __builtin_prefetch(x[ib].qs, 0, 1); + __builtin_prefetch(y[ib].qs, 0, 1); + + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); + vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); + vector float vd = vec_mul(vxd, vyd); + + vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); + vector signed char q4x0 = vec_and(qxs, lowMask); + vector signed char q4x1 = vec_sr(qxs, v4); + + q4x0 = vec_perm(values, values, (vector unsigned char)q4x0); + q4x1 = vec_perm(values, values, (vector unsigned char)q4x1); + + vector signed char q8y0 = vec_xl( 0, y[ib].qs); + vector signed char q8y1 = vec_xl(16, y[ib].qs); + + vector signed short qv0 = vec_add(vec_mule(q4x0, q8y0), vec_mulo(q4x0, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q4x1, q8y1), vec_mulo(q4x1, q8y1)); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + + vsumi0 = vec_sum4s(qv0, vsumi0); + vsumi1 = vec_sum4s(qv1, vsumi1); + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + } + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + sumf = vec_extract(vsumf0, 0); + +#elif defined (__loongarch_asx) + + const __m128i values128 = __lsx_vld((const __m128i*)kvalues_iq4nl, 0); + const __m128i m4b = __lsx_vreplgr2vr_b(0x0f); + const __m256i mone = __lasx_xvreplgr2vr_h(1); + + __m256 accum1 = (__m256)__lasx_xvldi(0); + __m256 accum2 = (__m256)__lasx_xvldi(0); + for (; ib + 1 < nb; ib += 2) { + const __m128i q4bits_1 = __lsx_vld((const __m128i*)x[ib + 0].qs, 0); + const __m128i q4bits_2 = __lsx_vld((const __m128i*)x[ib + 1].qs, 0); + const __m256i q8b_1 = __lasx_xvld((const __m256i *)y[ib + 0].qs, 0); + const __m256i q8b_2 = __lasx_xvld((const __m256i *)y[ib + 1].qs, 0); + const __m256i q4b_1 = lasx_insertf128(lsx_shuffle_b(values128, __lsx_vand_v(__lsx_vsrli_h(q4bits_1, 4), m4b)), + lsx_shuffle_b(values128, __lsx_vand_v(q4bits_1, m4b))); + const __m256i q4b_2 = lasx_insertf128(lsx_shuffle_b(values128, __lsx_vand_v(__lsx_vsrli_h(q4bits_2, 4), m4b)), + lsx_shuffle_b(values128, __lsx_vand_v(q4bits_2, m4b))); + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const __m256i p_1 = lasx_madd_h(p16_1, mone); + const __m256i p_2 = lasx_madd_h(p16_2, mone); + accum1 = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), + __lasx_xvffint_s_w(p_1), accum1); + accum2 = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), + __lasx_xvffint_s_w(p_2), accum2); + } + + sumf = hsum_float_8(__lasx_xvfadd_s(accum1, accum2)); + +#endif + for (; ib < nb; ++ib) { + const float d = GGML_FP16_TO_FP32(y[ib].d)*GGML_FP16_TO_FP32(x[ib].d); + int sumi1 = 0, sumi2 = 0; + for (int j = 0; j < QK4_NL/2; ++j) { + sumi1 += y[ib].qs[j+ 0] * kvalues_iq4nl[x[ib].qs[j] & 0xf]; + sumi2 += y[ib].qs[j+QK4_NL/2] * kvalues_iq4nl[x[ib].qs[j] >> 4]; + } + sumf += d * (sumi1 + sumi2); + } + *s = sumf; +} + +void ggml_vec_dot_iq4_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + assert(n % QK_K == 0); + + const block_iq4_xs * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined __ARM_NEON + const int8x16_t values = vld1q_s8(kvalues_iq4nl); + const uint8x16_t m4b = vdupq_n_u8(0x0f); + ggml_uint8x16x2_t q4bits; + ggml_int8x16x4_t q4b; + ggml_int8x16x4_t q8b; + int32x4_t prod_1, prod_2; + + float sumf = 0; + + for (int ibl = 0; ibl < nb; ++ibl) { + + const int8_t * q8 = y[ibl].qs; + const uint8_t * q4 = x[ibl].qs; + uint16_t h = x[ibl].scales_h; + + int sumi1 = 0, sumi2 = 0; + for (int ib = 0; ib < QK_K/64; ++ib) { + + q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + + q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); + q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); + q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); + q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); + + prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); + prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); + + int ls1 = ((x[ibl].scales_l[ib] & 0xf) | ((h << 4) & 0x30)) - 32; + int ls2 = ((x[ibl].scales_l[ib] >> 4) | ((h << 2) & 0x30)) - 32; + h >>= 4; + sumi1 += vaddvq_s32(prod_1) * ls1; + sumi2 += vaddvq_s32(prod_2) * ls2; + + } + + sumf += GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d * (sumi1 + sumi2); + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + + __m256 accum = _mm256_setzero_ps(); + for (int ibl = 0; ibl < nb; ++ibl) { + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + uint16_t sh = x[ibl].scales_h; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)qs); qs += 16; + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)qs); qs += 16; + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); + const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; + const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; + sh >>= 4; + const __m256i p_1 = _mm256_madd_epi16(p16_1, _mm256_set1_epi16(ls1)); + const __m256i p_2 = _mm256_madd_epi16(p16_2, _mm256_set1_epi16(ls2)); + sumi1 = _mm256_add_epi32(p_1, sumi1); + sumi2 = _mm256_add_epi32(p_2, sumi2); + } + accum = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), + _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accum); + } + + *s = hsum_float_8(accum); + +#elif defined __AVX__ + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + + __m256 accum = _mm256_setzero_ps(); + for (int ibl = 0; ibl < nb; ++ibl) { + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + uint16_t sh = x[ibl].scales_h; + __m128i sumi1_0 = _mm_setzero_si128(); + __m128i sumi1_1 = _mm_setzero_si128(); + __m128i sumi2_0 = _mm_setzero_si128(); + __m128i sumi2_1 = _mm_setzero_si128(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)qs); qs += 16; + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)qs); qs += 16; + const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; + const __m128i q4b_1_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)); + const __m128i q4b_1_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)); + const __m128i q4b_2_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)); + const __m128i q4b_2_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)); + const __m128i p16_1_0 = mul_add_epi8_sse(q4b_1_0, q8b_1_0); + const __m128i p16_1_1 = mul_add_epi8_sse(q4b_1_1, q8b_1_1); + const __m128i p16_2_0 = mul_add_epi8_sse(q4b_2_0, q8b_2_0); + const __m128i p16_2_1 = mul_add_epi8_sse(q4b_2_1, q8b_2_1); + const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; + const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; + sh >>= 4; + const __m128i p_1_0 = _mm_madd_epi16(p16_1_0, _mm_set1_epi16(ls1)); + const __m128i p_1_1 = _mm_madd_epi16(p16_1_1, _mm_set1_epi16(ls1)); + const __m128i p_2_0 = _mm_madd_epi16(p16_2_0, _mm_set1_epi16(ls2)); + const __m128i p_2_1 = _mm_madd_epi16(p16_2_1, _mm_set1_epi16(ls2)); + sumi1_0 = _mm_add_epi32(p_1_0, sumi1_0); + sumi1_1 = _mm_add_epi32(p_1_1, sumi1_1); + sumi2_0 = _mm_add_epi32(p_2_0, sumi2_0); + sumi2_1 = _mm_add_epi32(p_2_1, sumi2_1); + } + __m128i sumi12_0 = _mm_add_epi32(sumi1_0, sumi2_0); + __m128i sumi12_1 = _mm_add_epi32(sumi1_1, sumi2_1); + accum = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), + _mm256_cvtepi32_ps(MM256_SET_M128I(sumi12_1, sumi12_0))), accum); + } + + *s = hsum_float_8(accum); + +#elif defined(__POWER9_VECTOR__) + const vector signed char lowMask = vec_splats((signed char)0xF); + const vector int v0 = vec_splats((int32_t)0); + const vector unsigned char v4 = vec_splats((unsigned char)0x4); + + vector float vsumf0 = vec_splats(0.0f); + vector float vsumf1 = vec_splats(0.0f); + vector float vsumf2 = vec_splats(0.0f); + vector float vsumf3 = vec_splats(0.0f); + + const vector signed char values = vec_xl( 0, kvalues_iq4nl); + + for (int ibl = 0; ibl < nb; ++ibl) { + + vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ibl].d)); + vector float vyd = vec_splats(y[ibl].d); + vector float vd = vec_mul(vxd, vyd); + + vector signed int vsumi0 = v0; + vector signed int vsumi1 = v0; + vector signed int vsumi2 = v0; + vector signed int vsumi3 = v0; + + uint16_t h = x[ibl].scales_h; + + const uint8_t * restrict q4 = x[ibl].qs; + const uint8_t * restrict sc = x[ibl].scales_l; + const int8_t * restrict q8 = y[ibl].qs; + + for (int ib = 0; ib < QK_K/64; ib ++ ) { + __builtin_prefetch(q4, 0, 1); + __builtin_prefetch(q8, 0, 1); + + vector signed char qxs0 = (vector signed char)vec_xl( 0, q4); + vector signed char qxs1 = (vector signed char)vec_xl(16, q4); + q4 += 32; + + vector signed char q4x00 = (vector signed char)vec_and(qxs0, lowMask); + vector signed char q4x01 = (vector signed char)vec_sr(qxs0, v4); + vector signed char q4x10 = (vector signed char)vec_and(qxs1, lowMask); + vector signed char q4x11 = (vector signed char)vec_sr(qxs1, v4); + + q4x00 = vec_perm(values, values, (vector unsigned char)q4x00); + q4x01 = vec_perm(values, values, (vector unsigned char)q4x01); + q4x10 = vec_perm(values, values, (vector unsigned char)q4x10); + q4x11 = vec_perm(values, values, (vector unsigned char)q4x11); + + vector signed char q8y0 = vec_xl( 0, q8); + vector signed char q8y1 = vec_xl(16, q8); + vector signed char q8y2 = vec_xl(32, q8); + vector signed char q8y3 = vec_xl(48, q8); + q8 += 64; + + vector signed short qv0 = vec_add(vec_mule(q4x00, q8y0), vec_mulo(q4x00, q8y0)); + vector signed short qv1 = vec_add(vec_mule(q4x01, q8y1), vec_mulo(q4x01, q8y1)); + vector signed short qv2 = vec_add(vec_mule(q4x10, q8y2), vec_mulo(q4x10, q8y2)); + vector signed short qv3 = vec_add(vec_mule(q4x11, q8y3), vec_mulo(q4x11, q8y3)); + + const uint16_t ls0 = (uint16_t)(((sc[0] & 0xf) | ((h << 4) & 0x30)) - 32); + const uint16_t ls1 = (uint16_t)(((sc[0] >> 4) | ((h << 2) & 0x30)) - 32); + h >>= 4; + sc ++; + + vector signed short vscales01 = vec_splats((int16_t)ls0); + vector signed short vscales23 = vec_splats((int16_t)ls1); + + vsumi0 = vec_msum(qv0, vscales01, vsumi0); + vsumi1 = vec_msum(qv1, vscales01, vsumi1); + vsumi2 = vec_msum(qv2, vscales23, vsumi2); + vsumi3 = vec_msum(qv3, vscales23, vsumi3); + } + + vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); + vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); + vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); + vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); + } + + vsumf0 = vec_add(vsumf0, vsumf2); + vsumf1 = vec_add(vsumf1, vsumf3); + + vsumf0 = vec_add(vsumf0, vsumf1); + + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); + vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); + + *s = vec_extract(vsumf0, 0); + +#elif defined(__loongarch_asx) + + const __m128i values128 = __lsx_vld((const __m128i*)kvalues_iq4nl, 0); + const __m128i m4b = __lsx_vreplgr2vr_b(0x0f); + + __m256 accum = (__m256)__lasx_xvldi(0); + __m256i tmp1; + __m128i tmp0, tmp2, tmp3, tmp4, mask_8f, mask; + + mask_8f = __lsx_vreplgr2vr_b(0x8f); + for (int ibl = 0; ibl < nb; ++ibl) { + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + uint16_t sh = x[ibl].scales_h; + __m256i sumi1 = __lasx_xvldi(0); + __m256i sumi2 = __lasx_xvldi(0); + __m128i zero = __lsx_vldi(0); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q4bits_1 = __lsx_vld((const __m128i*)qs, 0); qs += 16; + const __m128i q4bits_2 = __lsx_vld((const __m128i*)qs, 0); qs += 16; + const __m256i q8b_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + const __m256i q8b_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; + tmp2 = __lsx_vand_v(__lsx_vand_v(__lsx_vsrli_h(q4bits_1, 4), m4b), mask_8f); + tmp0 = __lsx_vori_b(tmp2, 0x10); + mask = __lsx_vsle_b(zero, tmp2); + tmp3 = __lsx_vand_v(tmp0, mask); + tmp3 = __lsx_vshuf_b(values128, zero, tmp3); + + tmp2 = __lsx_vand_v(__lsx_vand_v(q4bits_1, m4b), mask_8f); + tmp0 = __lsx_vori_b(tmp2, 0x10); + mask = __lsx_vsle_b(zero, tmp2); + tmp4 = __lsx_vand_v(tmp0, mask); + tmp4 = __lsx_vshuf_b(values128, zero, tmp4); + + const __m256i q4b_1 = lasx_insertf128(tmp3, tmp4); + + tmp2 = __lsx_vand_v(__lsx_vand_v(__lsx_vsrli_h(q4bits_2, 4), m4b), mask_8f); + tmp0 = __lsx_vori_b(tmp2, 0x10); + mask = __lsx_vsle_b(zero, tmp2); + tmp3 = __lsx_vand_v(tmp0, mask); + tmp3 = __lsx_vshuf_b(values128, zero, tmp3); + + tmp2 = __lsx_vand_v(__lsx_vand_v(q4bits_2, m4b), mask_8f); + tmp0 = __lsx_vori_b(tmp2, 0x10); + mask = __lsx_vsle_b(zero, tmp2); + tmp4 = __lsx_vand_v(tmp0, mask); + tmp4 = __lsx_vshuf_b(values128, zero, tmp4); + + const __m256i q4b_2 = lasx_insertf128(tmp3, tmp4); + + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; + const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; + sh >>= 4; + __m256i tmp5, tmp6; + tmp1 = __lasx_xvreplgr2vr_h(ls1); + tmp5 = __lasx_xvmulwev_w_h(p16_1, tmp1); + tmp6 = __lasx_xvmulwod_w_h(p16_1, tmp1); + const __m256i p_1 = __lasx_xvadd_w(tmp5, tmp6); + tmp1 = __lasx_xvreplgr2vr_h(ls2); + tmp5 = __lasx_xvmulwev_w_h(p16_2, tmp1); + tmp6 = __lasx_xvmulwod_w_h(p16_2, tmp1); + const __m256i p_2 = __lasx_xvadd_w(tmp5, tmp6); + sumi1 = __lasx_xvadd_w(p_1, sumi1); + sumi2 = __lasx_xvadd_w(p_2, sumi2); + } + accum = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), + __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accum); + } + + *s = hsum_float_8(accum); + +#else + float sumf = 0; + for (int ibl = 0; ibl < nb; ++ibl) { + const float d4d8 = GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d; + uint16_t h = x[ibl].scales_h; + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + for (int ib = 0; ib < QK_K/32; ib += 2) { + const uint8_t ls1 = (x[ibl].scales_l[ib/2] & 0xf) | ((h << 4) & 0x30); + const uint8_t ls2 = (x[ibl].scales_l[ib/2] >> 4) | ((h << 2) & 0x30); + h >>= 4; + const float d1 = d4d8*(ls1 - 32); + const float d2 = d4d8*(ls2 - 32); + int sumi1 = 0, sumi2 = 0; + for (int j = 0; j < 16; ++j) { + sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; + sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; + } + sumf += d1 * (sumi1 + sumi2); + qs += 16; + q8 += 32; + sumi1 = sumi2 = 0; + for (int j = 0; j < 16; ++j) { + sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; + sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; + } + sumf += d2 * (sumi1 + sumi2); + qs += 16; + q8 += 32; + } + } + *s = sumf; +#endif +} + +// ============================ 4-bit non-linear quants + +void quantize_row_iq4_nl(const float * restrict x, void * restrict y, int64_t k) { + assert(k % QK4_NL == 0); + quantize_row_iq4_nl_ref(x, y, k); +} + +void quantize_row_iq4_xs(const float * restrict x, void * restrict y, int64_t k) { + assert(k % QK_K == 0); + quantize_iq4_xs(x, y, 1, k, NULL); +} diff --git a/ggml/src/ggml-cpu/ggml-cpu-quants.h b/ggml/src/ggml-cpu/ggml-cpu-quants.h new file mode 100644 index 000000000..e33d9d473 --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-quants.h @@ -0,0 +1,63 @@ +#pragma once + +#define GGML_COMMON_DECL_C +#include "ggml-common.h" + +#include "ggml.h" + +// GGML CPU internal header + +#ifdef __cplusplus +extern "C" { +#endif + +// Quantization +void quantize_row_q4_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +void quantize_row_q2_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q3_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q6_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +void quantize_row_tq1_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_tq2_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +void quantize_row_iq4_nl (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_xs (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +// Dot product +void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +void ggml_vec_dot_q2_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q3_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q4_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q6_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +void ggml_vec_dot_tq1_0_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_tq2_0_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +void ggml_vec_dot_iq2_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq2_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq2_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq3_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq1_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq1_m_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq4_nl_q8_0 (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq4_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-cpu/ggml-cpu-traits.cpp b/ggml/src/ggml-cpu/ggml-cpu-traits.cpp new file mode 100644 index 000000000..62a0712da --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-traits.cpp @@ -0,0 +1,36 @@ +#include "ggml-cpu-traits.h" + +#include "ggml-backend-impl.h" +#include "ggml-backend.h" + +namespace ggml::cpu { +tensor_traits::~tensor_traits() {} + +extra_buffer_type::~extra_buffer_type() {} +} // namespace ggml::cpu + +bool ggml_cpu_extra_compute_forward(struct ggml_compute_params * params, struct ggml_tensor * op) { + for (auto extra : ggml_backend_cpu_get_extra_buffers_type()) { + if (extra && extra->context) { + auto buf_extra = (ggml::cpu::extra_buffer_type *) extra->context; + auto tensor_traits = buf_extra->get_tensor_traits(op); + if (tensor_traits && tensor_traits->compute_forward(params, op)) { + return true; + } + } + } + return false; +} + +bool ggml_cpu_extra_work_size(int n_threads, const struct ggml_tensor * op, size_t * size) { + for (auto extra : ggml_backend_cpu_get_extra_buffers_type()) { + if (extra && extra->context) { + auto buf_extra = (ggml::cpu::extra_buffer_type *) extra->context; + auto tensor_traits = buf_extra->get_tensor_traits(op); + if (tensor_traits && tensor_traits->work_size(n_threads, op, *size)) { + return true; + } + } + } + return false; +} diff --git a/ggml/src/ggml-cpu/ggml-cpu-traits.h b/ggml/src/ggml-cpu/ggml-cpu-traits.h new file mode 100644 index 000000000..99a6186b1 --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu-traits.h @@ -0,0 +1,38 @@ +#pragma once +#include "ggml-backend-impl.h" +#include "ggml-cpu-impl.h" +#include "ggml.h" + +#ifdef __cplusplus +# include +extern "C" { +#endif + +// return true if op part of extra "accelerator" +bool ggml_cpu_extra_compute_forward(struct ggml_compute_params * params, struct ggml_tensor * op); +bool ggml_cpu_extra_work_size(int n_threads, const struct ggml_tensor * op, size_t * size); + +#ifdef __cplusplus +} + +namespace ggml::cpu { +// register in tensor->extra +class tensor_traits { + public: + virtual ~tensor_traits(); + virtual bool work_size(int n_threads, const struct ggml_tensor * op, size_t & size) = 0; + virtual bool compute_forward(struct ggml_compute_params * params, struct ggml_tensor * op) = 0; +}; + +class extra_buffer_type { + public: + virtual ~extra_buffer_type(); + virtual bool supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) = 0; + virtual tensor_traits * get_tensor_traits(const struct ggml_tensor * op) = 0; +}; +} // namespace ggml::cpu + +// implemented in ggml-cpu.cpp. +std::vector & ggml_backend_cpu_get_extra_buffers_type(); + +#endif diff --git a/ggml/src/ggml-cpu.c b/ggml/src/ggml-cpu/ggml-cpu.c similarity index 92% rename from ggml/src/ggml-cpu.c rename to ggml/src/ggml-cpu/ggml-cpu.c index de1de18ec..7c2e45f86 100644 --- a/ggml/src/ggml-cpu.c +++ b/ggml/src/ggml-cpu/ggml-cpu.c @@ -1,13 +1,16 @@ #define _CRT_SECURE_NO_DEPRECATE // Disables "unsafe" warnings on Windows #define _USE_MATH_DEFINES // For M_PI on MSVC -#include "ggml-aarch64.h" #include "ggml-backend-impl.h" #include "ggml-backend.h" +#include "ggml-cpu-traits.h" #include "ggml-cpu-impl.h" #include "ggml-cpu.h" #include "ggml-impl.h" #include "ggml-quants.h" +#include "ggml-cpu-quants.h" +#include "ggml-threading.h" +#include "amx/amx.h" #include "ggml.h" #if defined(_MSC_VER) || defined(__MINGW32__) @@ -42,7 +45,7 @@ #endif #ifdef GGML_USE_LLAMAFILE -#include +#include "llamafile/sgemm.h" #endif #if defined(_MSC_VER) @@ -104,16 +107,14 @@ static ggml_fp16_t ggml_table_gelu_f16[1 << 16]; // precomputed quick gelu table for f16 (128 KB) static ggml_fp16_t ggml_table_gelu_quick_f16[1 << 16]; -// precomputed f32 table for f16 (256 KB) (ggml-impl.h) -float ggml_table_f32_f16[1 << 16]; - #if defined(__ARM_ARCH) struct ggml_arm_arch_features_type { int has_neon; + int has_dotprod; int has_i8mm; int has_sve; int sve_cnt; -} ggml_arm_arch_features = {-1, -1, -1, 0}; +} ggml_arm_arch_features = {-1, -1, -1, -1, 0}; #endif @@ -125,8 +126,7 @@ struct ggml_arm_arch_features_type { #endif #include - -#if !defined(__clang__) +#if defined(_MSC_VER) && !defined(__clang__) #define GGML_CACHE_ALIGN __declspec(align(GGML_CACHE_LINE)) typedef volatile LONG atomic_int; @@ -223,10 +223,6 @@ typedef void * thread_ret_t; typedef pthread_t ggml_thread_t; -#ifdef GGML_USE_CPU_HBM -#include -#endif - #if defined(__APPLE__) #include #include @@ -261,11 +257,13 @@ static const struct ggml_type_traits_cpu type_traits_cpu[GGML_TYPE_COUNT] = { .nrows = 1, }, [GGML_TYPE_F16] = { + .from_float = (ggml_from_float_t) ggml_fp32_to_fp16_row, .vec_dot = (ggml_vec_dot_t) ggml_vec_dot_f16, .vec_dot_type = GGML_TYPE_F16, .nrows = 1, }, [GGML_TYPE_Q4_0] = { + .from_float = quantize_row_q4_0, .vec_dot = ggml_vec_dot_q4_0_q8_0, .vec_dot_type = GGML_TYPE_Q8_0, #if defined (__ARM_FEATURE_MATMUL_INT8) @@ -275,6 +273,7 @@ static const struct ggml_type_traits_cpu type_traits_cpu[GGML_TYPE_COUNT] = { #endif }, [GGML_TYPE_Q4_1] = { + .from_float = quantize_row_q4_1, .vec_dot = ggml_vec_dot_q4_1_q8_1, .vec_dot_type = GGML_TYPE_Q8_1, #if defined (__ARM_FEATURE_MATMUL_INT8) @@ -283,28 +282,20 @@ static const struct ggml_type_traits_cpu type_traits_cpu[GGML_TYPE_COUNT] = { .nrows = 1, #endif }, - [4] = { // GGML_TYPE_Q4_2 - .vec_dot = NULL, - .vec_dot_type = GGML_TYPE_COUNT, - .nrows = 1, - }, - [5] = { // GGML_TYPE_Q4_3 - .vec_dot = NULL, - .vec_dot_type = GGML_TYPE_COUNT, - .nrows = 1, - }, [GGML_TYPE_Q5_0] = { + .from_float = quantize_row_q5_0, .vec_dot = ggml_vec_dot_q5_0_q8_0, .vec_dot_type = GGML_TYPE_Q8_0, .nrows = 1, }, [GGML_TYPE_Q5_1] = { + .from_float = quantize_row_q5_1, .vec_dot = ggml_vec_dot_q5_1_q8_1, .vec_dot_type = GGML_TYPE_Q8_1, .nrows = 1, }, [GGML_TYPE_Q8_0] = { - .from_float_to_mat = quantize_mat_q8_0, + .from_float = quantize_row_q8_0, .vec_dot = ggml_vec_dot_q8_0_q8_0, .vec_dot_type = GGML_TYPE_Q8_0, #if defined (__ARM_FEATURE_MATMUL_INT8) @@ -314,114 +305,112 @@ static const struct ggml_type_traits_cpu type_traits_cpu[GGML_TYPE_COUNT] = { #endif }, [GGML_TYPE_Q8_1] = { + .from_float = quantize_row_q8_1, .vec_dot_type = GGML_TYPE_Q8_1, .nrows = 1, }, [GGML_TYPE_Q2_K] = { + .from_float = quantize_row_q2_K, .vec_dot = ggml_vec_dot_q2_K_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_Q3_K] = { + .from_float = quantize_row_q3_K, .vec_dot = ggml_vec_dot_q3_K_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_Q4_K] = { + .from_float = quantize_row_q4_K, .vec_dot = ggml_vec_dot_q4_K_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_Q5_K] = { + .from_float = quantize_row_q5_K, .vec_dot = ggml_vec_dot_q5_K_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_Q6_K] = { + .from_float = quantize_row_q6_K, .vec_dot = ggml_vec_dot_q6_K_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ2_XXS] = { + .from_float = NULL, .vec_dot = ggml_vec_dot_iq2_xxs_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ2_XS] = { + .from_float = NULL, .vec_dot = ggml_vec_dot_iq2_xs_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ3_XXS] = { + // NOTE: from_float for iq3 and iq2_s was removed because these quants require initialization in ggml_quantize_init + //.from_float = quantize_row_iq3_xxs, .vec_dot = ggml_vec_dot_iq3_xxs_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ3_S] = { + //.from_float = quantize_row_iq3_s, .vec_dot = ggml_vec_dot_iq3_s_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ2_S] = { + //.from_float = quantize_row_iq2_s, .vec_dot = ggml_vec_dot_iq2_s_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ1_S] = { + .from_float = NULL, .vec_dot = ggml_vec_dot_iq1_s_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ1_M] = { + .from_float = NULL, .vec_dot = ggml_vec_dot_iq1_m_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_IQ4_NL] = { + .from_float = quantize_row_iq4_nl, .vec_dot = ggml_vec_dot_iq4_nl_q8_0, .vec_dot_type = GGML_TYPE_Q8_0, .nrows = 1, }, [GGML_TYPE_IQ4_XS] = { + .from_float = quantize_row_iq4_xs, .vec_dot = ggml_vec_dot_iq4_xs_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, + [GGML_TYPE_Q8_K] = { + .from_float = quantize_row_q8_K, + }, [GGML_TYPE_BF16] = { + .from_float = (ggml_from_float_t) ggml_fp32_to_bf16_row, .vec_dot = (ggml_vec_dot_t) ggml_vec_dot_bf16, .vec_dot_type = GGML_TYPE_BF16, .nrows = 1, }, - [GGML_TYPE_Q4_0_4_4] = { - .vec_dot = NULL, - .vec_dot_type = GGML_TYPE_Q8_0, - .nrows = 1, - .ncols = 4, - .gemv = ggml_gemv_q4_0_4x4_q8_0, - .gemm = ggml_gemm_q4_0_4x4_q8_0, - }, - [GGML_TYPE_Q4_0_4_8] = { - .vec_dot = NULL, - .vec_dot_type = GGML_TYPE_Q8_0, - .nrows = 1, - .ncols = 4, - .gemv = ggml_gemv_q4_0_4x8_q8_0, - .gemm = ggml_gemm_q4_0_4x8_q8_0, - }, - [GGML_TYPE_Q4_0_8_8] = { - .vec_dot = NULL, - .vec_dot_type = GGML_TYPE_Q8_0, - .nrows = 1, - .ncols = 8, - .gemv = ggml_gemv_q4_0_8x8_q8_0, - .gemm = ggml_gemm_q4_0_8x8_q8_0, - }, [GGML_TYPE_TQ1_0] = { + .from_float = quantize_row_tq1_0, .vec_dot = ggml_vec_dot_tq1_0_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, }, [GGML_TYPE_TQ2_0] = { + .from_float = quantize_row_tq2_0, .vec_dot = ggml_vec_dot_tq2_0_q8_K, .vec_dot_type = GGML_TYPE_Q8_K, .nrows = 1, @@ -465,21 +454,21 @@ const struct ggml_type_traits_cpu * ggml_get_type_traits_cpu(enum ggml_type type #define GGML_F32x4_ADD vaddq_f32 #define GGML_F32x4_MUL vmulq_f32 #define GGML_F32x4_REDUCE_ONE(x) vaddvq_f32(x) -#define GGML_F32x4_REDUCE(res, x) \ -{ \ - int offset = GGML_F32_ARR >> 1; \ - for (int i = 0; i < offset; ++i) { \ - (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ - } \ - offset >>= 1; \ - for (int i = 0; i < offset; ++i) { \ - (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ - } \ - offset >>= 1; \ - for (int i = 0; i < offset; ++i) { \ - (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ - } \ - (res) = GGML_F32x4_REDUCE_ONE((x)[0]); \ +#define GGML_F32x4_REDUCE(res, x) \ +{ \ + int offset = GGML_F32_ARR >> 1; \ + for (int i = 0; i < offset; ++i) { \ + (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ + } \ + offset >>= 1; \ + for (int i = 0; i < offset; ++i) { \ + (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ + } \ + offset >>= 1; \ + for (int i = 0; i < offset; ++i) { \ + (x)[i] = vaddq_f32((x)[i], (x)[offset+i]); \ + } \ + (res) = (ggml_float) GGML_F32x4_REDUCE_ONE((x)[0]); \ } #define GGML_F32_VEC GGML_F32x4 @@ -594,7 +583,7 @@ do { \ for (int i = 0; i < offset; ++i) { \ x[i] = _mm512_add_ps(x[i], x[offset+i]); \ } \ - res = _mm512_reduce_add_ps(x[0]); \ + res = (ggml_float) _mm512_reduce_add_ps(x[0]); \ } while (0) // TODO: is this optimal ? @@ -644,7 +633,7 @@ do { \ for (int i = 0; i < offset; ++i) { \ x[i] = _mm512_add_ps(x[i], x[offset+i]); \ } \ - res = _mm512_reduce_add_ps(x[0]); \ + res = (ggml_float) _mm512_reduce_add_ps(x[0]); \ } while (0) #define GGML_F16_VEC GGML_F32Cx16 @@ -655,8 +644,8 @@ do { \ #define GGML_F16_VEC_FMA GGML_F32Cx16_FMA #define GGML_F16_VEC_ADD GGML_F32Cx16_ADD #define GGML_F16_VEC_MUL GGML_F32Cx16_MUL -#define GGML_F16_VEC_REDUCE GGML_F32Cx16_REDUCE +#define GGML_F16_VEC_REDUCE GGML_F32Cx16_REDUCE #elif defined(__AVX__) #define GGML_SIMD @@ -725,7 +714,7 @@ do { \ #define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((const __m128i *)(x))) #define GGML_F32Cx8_STORE(x, y) _mm_storeu_si128((__m128i *)(x), _mm256_cvtps_ph(y, 0)) #else -static inline __m256 __avx_f32cx8_load(ggml_fp16_t *x) { +static inline __m256 __avx_f32cx8_load(const ggml_fp16_t * x) { float tmp[8]; for (int i = 0; i < 8; i++) { @@ -997,7 +986,7 @@ inline static void __wasm_f16x4_store(ggml_fp16_t * p, v128_t x) { #define GGML_F16_STEP 32 #define GGML_F16_EPR 4 -static inline __m128 __sse_f16x4_load(ggml_fp16_t *x) { +static inline __m128 __sse_f16x4_load(const ggml_fp16_t * x) { float tmp[4]; tmp[0] = GGML_FP16_TO_FP32(x[0]); @@ -1008,7 +997,7 @@ static inline __m128 __sse_f16x4_load(ggml_fp16_t *x) { return _mm_loadu_ps(tmp); } -static inline void __sse_f16x4_store(ggml_fp16_t *x, __m128 y) { +static inline void __sse_f16x4_store(ggml_fp16_t * x, __m128 y) { float arr[4]; _mm_storeu_ps(arr, y); @@ -1148,28 +1137,28 @@ static inline void __lasx_f32cx8_store(ggml_fp16_t * x, __m256 y) { #define GGML_F32x4_FMA(a, b, c) __lsx_vfmadd_s(b, c, a) #define GGML_F32x4_ADD __lsx_vfadd_s #define GGML_F32x4_MUL __lsx_vfmul_s -#define GGML_F32x4_REDUCE(res, x) \ -{ \ - int offset = GGML_F32_ARR >> 1; \ - for (int i = 0; i < offset; ++i) { \ - x[i] = __lsx_vfadd_s(x[i], x[offset+i]); \ - } \ - offset >>= 1; \ - for (int i = 0; i < offset; ++i) { \ - x[i] = __lsx_vfadd_s(x[i], x[offset+i]); \ - } \ - offset >>= 1; \ - for (int i = 0; i < offset; ++i) { \ - x[i] = __lsx_vfadd_s(x[i], x[offset+i]); \ - } \ - __m128i tmp = __lsx_vsrli_d((__m128i)x[0], 32); \ - tmp = (__m128i)__lsx_vfadd_s((__m128)tmp, x[0]); \ - tmp = __lsx_vpickev_w(__lsx_vldi(0), tmp); \ - const __m128 t0 = __lsx_vshuf4i_w(tmp, 0x88); \ - tmp = __lsx_vsrli_d((__m128i)t0, 32); \ - tmp = (__m128i)__lsx_vfadd_s((__m128)tmp, t0); \ - tmp = __lsx_vpickev_w(__lsx_vldi(0), tmp); \ - res = (ggml_float) __lsx_vpickve2gr_w(__lsx_vshuf4i_w(tmp, 0x88), 0); \ +#define GGML_F32x4_REDUCE(res, x) \ +{ \ + int offset = GGML_F32_ARR >> 1; \ + for (int i = 0; i < offset; ++i) { \ + x[i] = __lsx_vfadd_s(x[i], x[offset + i]); \ + } \ + offset >>= 1; \ + for (int i = 0; i < offset; ++i) { \ + x[i] = __lsx_vfadd_s(x[i], x[offset + i]); \ + } \ + offset >>= 1; \ + for (int i = 0; i < offset; ++i) { \ + x[i] = __lsx_vfadd_s(x[i], x[offset + i]); \ + } \ + __m128i tmp = __lsx_vsrli_d((__m128i) x[0], 32); \ + tmp = (__m128i) __lsx_vfadd_s((__m128) tmp, x[0]); \ + tmp = __lsx_vpickev_w(__lsx_vldi(0), tmp); \ + const __m128 t0 = __lsx_vshuf4i_w(tmp, 0x88); \ + tmp = __lsx_vsrli_d((__m128i) t0, 32); \ + tmp = (__m128i) __lsx_vfadd_s((__m128) tmp, t0); \ + tmp = __lsx_vpickev_w(__lsx_vldi(0), tmp); \ + res = (ggml_float) __lsx_vpickve2gr_w(__lsx_vshuf4i_w(tmp, 0x88), 0); \ } #define GGML_F32_VEC GGML_F32x4 @@ -1337,31 +1326,18 @@ struct ggml_compute_state { int ith; }; -struct ggml_compute_params { - // ith = thread index, nth = number of threads - int ith, nth; - - // work buffer for all threads - size_t wsize; - void * wdata; - - struct ggml_threadpool * threadpool; -}; - // // fundamental operations // inline static void ggml_vec_set_i8(const int n, int8_t * x, const int8_t v) { for (int i = 0; i < n; ++i) x[i] = v; } - inline static void ggml_vec_set_i16(const int n, int16_t * x, const int16_t v) { for (int i = 0; i < n; ++i) x[i] = v; } -inline static void ggml_vec_set_i32(const int n, int32_t * x, const int32_t v) { for (int i = 0; i < n; ++i) x[i] = v; } +inline static void ggml_vec_set_i32(const int n, int32_t * x, const int32_t v) { for (int i = 0; i < n; ++i) x[i] = v; } +inline static void ggml_vec_cpy_i32(const int n, int32_t * y, const int32_t * x) { for (int i = 0; i < n; ++i) y[i] = x[i]; } inline static void ggml_vec_set_f16(const int n, ggml_fp16_t * x, const int32_t v) { for (int i = 0; i < n; ++i) x[i] = v; } - inline static void ggml_vec_set_bf16(const int n, ggml_bf16_t * x, const ggml_bf16_t v) { for (int i = 0; i < n; ++i) x[i] = v; } - inline static void ggml_vec_add_f32 (const int n, float * z, const float * x, const float * y) { for (int i = 0; i < n; ++i) z[i] = x[i] + y[i]; } inline static void ggml_vec_add1_f32(const int n, float * z, const float * x, const float v) { for (int i = 0; i < n; ++i) z[i] = x[i] + v; } inline static void ggml_vec_acc_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] += x[i]; } @@ -1449,8 +1425,12 @@ static void ggml_vec_dot_bf16(int n, float * restrict s, size_t bs, ggml_bf16_t sumf += (ggml_float)_mm512_reduce_add_ps(c2); #undef LOAD -#elif defined(__AVX2__) +#elif defined(__AVX2__) || defined(__AVX__) +#if defined(__AVX2__) #define LOAD(p) _mm256_castsi256_ps(_mm256_slli_epi32(_mm256_cvtepu16_epi32(_mm_loadu_si128((const __m128i *)(p))), 16)) +#else +#define LOAD(p) _mm256_castsi256_ps(_mm256_insertf128_si256(_mm256_castsi128_si256(_mm_slli_epi32(_mm_cvtepu16_epi32(_mm_loadu_si128((const __m128i *)(p))), 16)), (_mm_slli_epi32(_mm_cvtepu16_epi32(_mm_bsrli_si128(_mm_loadu_si128((const __m128i *)(p)), 8)), 16)), 1)) +#endif __m256 c1 = _mm256_setzero_ps(); __m256 c2 = _mm256_setzero_ps(); __m256 c3 = _mm256_setzero_ps(); @@ -2250,24 +2230,9 @@ struct ggml_state { struct ggml_numa_nodes numa; }; -// global state static struct ggml_state g_state = {0}; -static atomic_flag g_state_critical = ATOMIC_FLAG_INIT; -// TODO: move to threading file -// critical section via spin lock -void ggml_critical_section_start(void) { - while (atomic_flag_test_and_set(&g_state_critical)) { - // spin - sched_yield(); - } -} - -void ggml_critical_section_end(void) { - atomic_flag_clear(&g_state_critical); -} - -static void ggml_barrier(struct ggml_threadpool * tp) { +void ggml_barrier(struct ggml_threadpool * tp) { int n_threads = atomic_load_explicit(&tp->n_threads_cur, memory_order_relaxed); if (n_threads == 1) { return; @@ -2360,7 +2325,7 @@ void ggml_numa_init(enum ggml_numa_strategy numa_flag) { // figure out which node we're on uint current_cpu; int getcpu_ret = 0; -#if __GLIBC__ > 2 || (__GLIBC__ == 2 && __GLIBC_MINOR__ > 28) || defined(__COSMOPOLITAN__) +#if __GLIBC__ > 2 || (__GLIBC__ == 2 && __GLIBC_MINOR__ > 33) || defined(__COSMOPOLITAN__) getcpu_ret = getcpu(¤t_cpu, &g_state.numa.current_node); #else // old glibc doesn't have a wrapper for this call. Fall back on direct syscall @@ -2421,7 +2386,7 @@ bool ggml_is_numa(void) { #endif #if !defined(HWCAP2_I8MM) -#define HWCAP2_I8MM 0 +#define HWCAP2_I8MM (1 << 13) #endif static void ggml_init_arm_arch_features(void) { @@ -2430,6 +2395,7 @@ static void ggml_init_arm_arch_features(void) { uint32_t hwcap2 = getauxval(AT_HWCAP2); ggml_arm_arch_features.has_neon = !!(hwcap & HWCAP_ASIMD); + ggml_arm_arch_features.has_dotprod = !!(hwcap & HWCAP_ASIMDDP); ggml_arm_arch_features.has_i8mm = !!(hwcap2 & HWCAP2_I8MM); ggml_arm_arch_features.has_sve = !!(hwcap & HWCAP_SVE); @@ -2444,6 +2410,11 @@ static void ggml_init_arm_arch_features(void) { } ggml_arm_arch_features.has_neon = oldp; + if (sysctlbyname("hw.optional.arm.FEAT_DotProd", &oldp, &size, NULL, 0) != 0) { + oldp = 0; + } + ggml_arm_arch_features.has_dotprod = oldp; + if (sysctlbyname("hw.optional.arm.FEAT_I8MM", &oldp, &size, NULL, 0) != 0) { oldp = 0; } @@ -2997,8 +2968,8 @@ static void ggml_compute_forward_dup_f16( id += ne00 * (ne01 - ir1); } } - } else if (ggml_get_type_traits(dst->type)->from_float) { - ggml_from_float_t const quantize_row_q = ggml_get_type_traits(dst->type)->from_float; + } else if (ggml_get_type_traits_cpu(dst->type)->from_float) { + ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float; float * src0_f32 = (float *) params->wdata + (ne00 + CACHE_LINE_SIZE_F32) * ith; size_t id = 0; @@ -3278,8 +3249,8 @@ static void ggml_compute_forward_dup_bf16( id += ne00 * (ne01 - ir1); } } - } else if (ggml_get_type_traits(dst->type)->from_float) { - ggml_from_float_t const quantize_row_q = ggml_get_type_traits(dst->type)->from_float; + } else if (ggml_get_type_traits_cpu(dst->type)->from_float) { + ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float; float * src0_f32 = (float *) params->wdata + (ne00 + CACHE_LINE_SIZE_F32) * ith; size_t id = 0; @@ -3594,8 +3565,8 @@ static void ggml_compute_forward_dup_f32( id += rs * (ne01 - ir1); } } - } else if (ggml_get_type_traits(dst->type)->from_float) { - ggml_from_float_t const quantize_row_q = ggml_get_type_traits(dst->type)->from_float; + } else if (ggml_get_type_traits_cpu(dst->type)->from_float) { + ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float; size_t id = 0; size_t rs = nb0 * (ne00 / ggml_blck_size(dst->type)); @@ -4377,7 +4348,7 @@ static void ggml_compute_forward_add_q_f32( const enum ggml_type type = src0->type; const enum ggml_type dtype = dst->type; ggml_to_float_t const dequantize_row_q = ggml_get_type_traits(type)->to_float; - ggml_from_float_t const quantize_row_q = ggml_get_type_traits(dtype)->from_float; + ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dtype)->from_float; // we don't support permuted src0 or src1 GGML_ASSERT(nb00 == ggml_type_size(type)); @@ -4496,9 +4467,6 @@ static void ggml_compute_forward_add( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: { ggml_compute_forward_add_q_f32(params, dst); } break; @@ -4679,7 +4647,7 @@ static void ggml_compute_forward_add1_q_f32( const enum ggml_type type = src0->type; ggml_to_float_t const dequantize_row_q = ggml_get_type_traits(type)->to_float; - ggml_from_float_t const quantize_row_q = ggml_get_type_traits(type)->from_float; + ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(type)->from_float; // we don't support permuted src0 GGML_ASSERT(nb00 == ggml_type_size(type)); @@ -4876,9 +4844,6 @@ static void ggml_compute_forward_add1( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: { ggml_compute_forward_add1_q_f32(params, dst); } break; @@ -5006,9 +4971,6 @@ static void ggml_compute_forward_acc( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: default: { GGML_ABORT("fatal error"); @@ -7325,6 +7287,7 @@ static void ggml_compute_forward_group_norm( static void ggml_compute_forward_mul_mat_one_chunk( const struct ggml_compute_params * params, struct ggml_tensor * dst, + const enum ggml_type type, const int64_t num_rows_per_vec_dot, const int64_t ir0_start, const int64_t ir0_end, @@ -7336,8 +7299,6 @@ static void ggml_compute_forward_mul_mat_one_chunk( GGML_TENSOR_BINARY_OP_LOCALS - const enum ggml_type type = src0->type; - const bool src1_cont = ggml_is_contiguous(src1); ggml_vec_dot_t const vec_dot = type_traits_cpu[type].vec_dot; @@ -7425,16 +7386,9 @@ static void ggml_compute_forward_mul_mat( const int ith = params->ith; const int nth = params->nth; - const enum ggml_type type = src0->type; - - enum ggml_type const vec_dot_type = type_traits_cpu[type].vec_dot_type; - ggml_from_float_t const from_float = ggml_get_type_traits(vec_dot_type)->from_float; - ggml_from_float_to_mat_t const from_float_to_mat = type_traits_cpu[vec_dot_type].from_float_to_mat; - int64_t const vec_dot_num_rows = type_traits_cpu[type].nrows; - int64_t const matmul_num_cols = type_traits_cpu[type].ncols; - int64_t const blck_size_interleave = ggml_get_type_traits(type)->blck_size_interleave; - ggml_gemv_t const gemv = type_traits_cpu[type].gemv; - ggml_gemm_t const gemm = type_traits_cpu[type].gemm; + enum ggml_type const vec_dot_type = type_traits_cpu[src0->type].vec_dot_type; + ggml_from_float_t const from_float = type_traits_cpu[vec_dot_type].from_float; + int64_t const vec_dot_num_rows = type_traits_cpu[src0->type].nrows; GGML_ASSERT(ne0 == ne01); GGML_ASSERT(ne1 == ne11); @@ -7442,7 +7396,7 @@ static void ggml_compute_forward_mul_mat( GGML_ASSERT(ne3 == ne13); // we don't support permuted src0 or src1 - GGML_ASSERT(nb00 == ggml_type_size(type)); + GGML_ASSERT(nb00 == ggml_type_size(src0->type)); GGML_ASSERT(nb10 == ggml_type_size(src1->type)); // dst cannot be transposed or permuted @@ -7454,6 +7408,7 @@ static void ggml_compute_forward_mul_mat( // nb01 >= nb00 - src0 is not transposed // compute by src0 rows + // TODO: extract to "extra_op" #if GGML_USE_LLAMAFILE // broadcast factors const int64_t r2 = ne12 / ne02; @@ -7464,14 +7419,14 @@ static void ggml_compute_forward_mul_mat( if (src1_cont) { for (int64_t i13 = 0; i13 < ne13; i13++) for (int64_t i12 = 0; i12 < ne12; i12++) - if (!llamafile_sgemm(ne01, ne11, ne00/ggml_blck_size(src0->type), + if (!llamafile_sgemm(params, + ne01, ne11, ne00/ggml_blck_size(src0->type), (const char *)src0->data + i12/r2*nb02 + i13/r3*nb03, nb01/ggml_type_size(src0->type), (const char *)src1->data + i12*nb12 + i13*nb13, nb11/ggml_type_size(src1->type), (char *)dst->data + i12*nb2 + i13*nb3, nb1/ggml_type_size(dst->type), - ith, nth, src0->type, src1->type, dst->type)) @@ -7493,19 +7448,10 @@ UseGgmlGemm1:; for (int64_t i13 = 0; i13 < ne13; ++i13) { for (int64_t i12 = 0; i12 < ne12; ++i12) { - int64_t i11_processed = 0; - if ((ggml_n_dims(src1) == 2) && from_float_to_mat && gemm) { - for (int64_t i11 = ith * 4; i11 < ne11 - ne11 % 4; i11 += nth * 4) { - from_float_to_mat((float *)((char *) src1->data + i13*nb13 + i12*nb12 + i11*nb11), - (void *) (wdata + i13*nbw3 + i12*nbw2 + i11*nbw1), - 4, ne10, blck_size_interleave); - } - i11_processed = ne11 - ne11 % 4; - } - for (int64_t i11 = i11_processed + ith; i11 < ne11; i11 += nth) { + for (int64_t i11 = ith; i11 < ne11; i11 += nth) { from_float((float *)((char *) src1->data + i13*nb13 + i12*nb12 + i11*nb11), - (void *) (wdata + i13*nbw3 + i12*nbw2 + i11*nbw1), - ne10); + (void *) (wdata + i13*nbw3 + i12*nbw2 + i11*nbw1), + ne10); } } } @@ -7525,14 +7471,14 @@ UseGgmlGemm1:; for (int64_t i13 = 0; i13 < ne13; i13++) for (int64_t i12 = 0; i12 < ne12; i12++) - if (!llamafile_sgemm(ne01, ne11, ne00/ggml_blck_size(src0->type), + if (!llamafile_sgemm(params, + ne01, ne11, ne00/ggml_blck_size(src0->type), (const char *)src0->data + i12/r2*nb02 + i13/r3*nb03, nb01/ggml_type_size(src0->type), (const char *)wdata + (i12*ne11 + i13*ne12*ne11)*row_size, row_size/ggml_type_size(vec_dot_type), (char *)dst->data + i12*nb2 + i13*nb3, nb1/ggml_type_size(dst->type), - ith, nth, src0->type, vec_dot_type, dst->type)) @@ -7548,14 +7494,6 @@ UseGgmlGemm2:; // This is the size of the rest of the dimensions of the result const int64_t nr1 = ne1 * ne2 * ne3; - // dot kernels can handle 1 row and col at a time, but mmla kernels can process 2 rows and cols - int64_t num_rows_per_vec_dot = vec_dot_num_rows; - // TODO: currently the mmla kernels support only even numbered rows/cols. - // this check can be removed once they are extended to support odd numbered rows/cols too - if ((nr0 % 2 != 0) || (ne11 % 2 != 0)) { - num_rows_per_vec_dot = 1; - } - // Now select a reasonable chunk size. int chunk_size = 16; @@ -7583,28 +7521,6 @@ UseGgmlGemm2:; const int64_t dr0 = (nr0 + nchunk0 - 1) / nchunk0; const int64_t dr1 = (nr1 + nchunk1 - 1) / nchunk1; - if ((ggml_n_dims(src0) == 2) && gemv) { - const void * src1_wdata = (src1->type == vec_dot_type) ? src1->data : params->wdata; - const size_t src1_col_stride = ggml_is_contiguous(src1) || src1->type != vec_dot_type ? ggml_row_size(vec_dot_type, ne10) : nb11; - int64_t src0_start = (ith * ne01) / nth; - int64_t src0_end = ((ith + 1) * ne01) / nth; - src0_start = (src0_start % matmul_num_cols) ? src0_start + matmul_num_cols - (src0_start % matmul_num_cols): src0_start; - src0_end = (src0_end % matmul_num_cols) ? src0_end + matmul_num_cols - (src0_end % matmul_num_cols): src0_end; - if (src0_start >= src0_end) return; - - // If there are more than three rows in src1, use gemm; otherwise, use gemv. - if (gemm && (ne11 > 3)) { - gemm(ne00, (float *)((char *) dst->data) + src0_start, ne01, (const char *) src0->data + src0_start * nb01, - (const char *) src1_wdata, ne11 - ne11 % 4, src0_end - src0_start); - } - for (int iter = gemm ? ne11 - ne11 % 4 : 0; iter < ne11; iter++) { - gemv(ne00, (float *)((char *) dst->data + (iter * nb1)) + src0_start, ne01, - (const char *) src0->data + src0_start * nb01, (const char *) src1_wdata + (src1_col_stride * iter), 1, - src0_end - src0_start); - } - return; - } - // The first chunk comes from our thread_id, the rest will get auto-assigned. int current_chunk = ith; @@ -7618,7 +7534,16 @@ UseGgmlGemm2:; const int64_t ir1_start = dr1 * ith1; const int64_t ir1_end = MIN(ir1_start + dr1, nr1); - ggml_compute_forward_mul_mat_one_chunk(params, dst, num_rows_per_vec_dot, ir0_start, ir0_end, ir1_start, ir1_end); + // dot kernels can handle 1 row and col at a time, but mmla kernels can process 2 rows and cols + int64_t num_rows_per_vec_dot = vec_dot_num_rows; + + // these checks are needed to avoid crossing dim1 boundaries + // can be optimized, but the logic would become more complicated, so keeping it like this for simplicity + if ((nr0 % 2 != 0) || (ne11 % 2 != 0) || ((ir0_end - ir0_start) % 2 != 0) || ((ir1_end - ir1_start) % 2 != 0)) { + num_rows_per_vec_dot = 1; + } + + ggml_compute_forward_mul_mat_one_chunk(params, dst, src0->type, num_rows_per_vec_dot, ir0_start, ir0_end, ir1_start, ir1_end); if (nth >= nchunk0 * nchunk1) { break; @@ -7649,9 +7574,7 @@ static void ggml_compute_forward_mul_mat_id( ggml_vec_dot_t const vec_dot = type_traits_cpu[type].vec_dot; enum ggml_type const vec_dot_type = type_traits_cpu[type].vec_dot_type; - ggml_from_float_t const from_float = ggml_get_type_traits(vec_dot_type)->from_float; - int64_t const matmul_num_cols = type_traits_cpu[type].ncols; - ggml_gemv_t const gemv = type_traits_cpu[type].gemv; + ggml_from_float_t const from_float = type_traits_cpu[vec_dot_type].from_float; // we don't support permuted src0 or src1 GGML_ASSERT(nb00 == ggml_type_size(type)); @@ -7737,34 +7660,6 @@ static void ggml_compute_forward_mul_mat_id( const int64_t nr0 = ne01; // src0 rows const int64_t nr1 = cne1; // src1 rows - if (((ggml_n_dims(src0) - 1) == 2) && gemv) { - int64_t src0_cur_start = (ith * ne01) / nth; - int64_t src0_cur_end = ((ith + 1) * ne01) / nth; - src0_cur_start = (src0_cur_start % matmul_num_cols) ? src0_cur_start + matmul_num_cols - (src0_cur_start % matmul_num_cols): src0_cur_start; - src0_cur_end = (src0_cur_end % matmul_num_cols) ? src0_cur_end + matmul_num_cols - (src0_cur_end % matmul_num_cols): src0_cur_end; - if (src0_cur_start >= src0_cur_end) return; - - for (int ir1 = 0; ir1 < nr1; ir1++) { - struct mmid_row_mapping row_mapping = MMID_MATRIX_ROW(cur_a, ir1); - const int id = row_mapping.i1; // selected expert index - - const int64_t i11 = id % ne11; - const int64_t i12 = row_mapping.i2; // row index in src1 - - const int64_t i1 = id; // selected expert index - const int64_t i2 = i12; // row - - const char * src1_col = (const char *) wdata + - (src1_cont || src1->type != vec_dot_type - ? (i11 + i12 * ne11) * row_size - : (i11 * nb11 + i12 * nb12)); - - gemv(ne00, (float *)((char *) dst->data + (i1 * nb1 + i2 * nb2)) + src0_cur_start, ne01, - (const char *) src0_cur + src0_cur_start * nb01, src1_col, 1, src0_cur_end - src0_cur_start); - } - continue; - } - // distribute the thread work across the inner or outer loop based on which one is larger const int64_t nth0 = nr0 > nr1 ? nth : 1; // parallelize by src0 rows @@ -8072,9 +7967,6 @@ static void ggml_compute_forward_out_prod( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: { ggml_compute_forward_out_prod_q_f32(params, dst); } break; @@ -8227,6 +8119,77 @@ static void ggml_compute_forward_set_f32( } } +static void ggml_compute_forward_set_i32( + const struct ggml_compute_params * params, + struct ggml_tensor * dst) { + + const struct ggml_tensor * src0 = dst->src[0]; + const struct ggml_tensor * src1 = dst->src[1]; + + GGML_ASSERT(ggml_are_same_shape(src0, dst)); + GGML_ASSERT(ggml_is_contiguous(dst) && ggml_is_contiguous(src0)); + + // view src0 and dst with these strides and data offset inbytes during set + // nb0 is implicitly element_size because src0 and dst are contiguous + size_t nb1 = ((int32_t *) dst->op_params)[0]; + size_t nb2 = ((int32_t *) dst->op_params)[1]; + size_t nb3 = ((int32_t *) dst->op_params)[2]; + size_t offset = ((int32_t *) dst->op_params)[3]; + bool inplace = (bool) ((int32_t *) dst->op_params)[4]; + + if (!inplace) { + if (params->ith == 0) { + // memcpy needs to be synchronized across threads to avoid race conditions. + // => do it in INIT phase + memcpy( + ((char *) dst->data), + ((char *) src0->data), + ggml_nbytes(dst)); + } + ggml_barrier(params->threadpool); + } + + const int ith = params->ith; + const int nth = params->nth; + + const int nr = ggml_nrows(src1); + const int nc = src1->ne[0]; + + GGML_TENSOR_LOCALS(int64_t, ne1, src1, ne) + GGML_TENSOR_LOCALS(size_t, nb1, src1, nb) + + // src0 and dst as viewed during set + const size_t nb0 = ggml_element_size(src0); + + const int im0 = (ne10 == 0 ? 0 : ne10-1); + const int im1 = (ne11 == 0 ? 0 : ne11-1); + const int im2 = (ne12 == 0 ? 0 : ne12-1); + const int im3 = (ne13 == 0 ? 0 : ne13-1); + + GGML_ASSERT(offset + im0*nb0 + im1*nb1 + im2*nb2 + im3*nb3 <= ggml_nbytes(dst)); + + GGML_ASSERT(nb10 == sizeof(int32_t)); + + // rows per thread + const int dr = (nr + nth - 1)/nth; + + // row range for this thread + const int ir0 = dr*ith; + const int ir1 = MIN(ir0 + dr, nr); + + for (int ir = ir0; ir < ir1; ++ir) { + // src0 and dst are viewed with shape of src1 and offset + // => same indices + const int i3 = ir/(ne12*ne11); + const int i2 = (ir - i3*ne12*ne11)/ne11; + const int i1 = (ir - i3*ne12*ne11 - i2*ne11); + + ggml_vec_cpy_i32(nc, + (int32_t *) ((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + offset), + (int32_t *) ((char *) src1->data + i3*nb13 + i2*nb12 + i1*nb11)); + } +} + static void ggml_compute_forward_set( const struct ggml_compute_params * params, struct ggml_tensor * dst) { @@ -8238,6 +8201,10 @@ static void ggml_compute_forward_set( { ggml_compute_forward_set_f32(params, dst); } break; + case GGML_TYPE_I32: + { + ggml_compute_forward_set_i32(params, dst); + } break; case GGML_TYPE_F16: case GGML_TYPE_BF16: case GGML_TYPE_Q4_0: @@ -8262,9 +8229,6 @@ static void ggml_compute_forward_set( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: default: { GGML_ABORT("fatal error"); @@ -8526,9 +8490,6 @@ static void ggml_compute_forward_get_rows( case GGML_TYPE_IQ4_XS: case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: { ggml_compute_forward_get_rows_q(params, dst); } break; @@ -9118,9 +9079,6 @@ static void ggml_compute_forward_clamp( case GGML_TYPE_IQ3_S: case GGML_TYPE_IQ2_S: case GGML_TYPE_Q8_K: - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - case GGML_TYPE_Q4_0_8_8: case GGML_TYPE_I8: case GGML_TYPE_I16: case GGML_TYPE_I32: @@ -9159,12 +9117,6 @@ static void rope_yarn( *sin_theta = sinf(theta) * mscale; } -// Apparently solving `n_rot = 2pi * x * base^((2 * max_pos_emb) / n_dims)` for x, we get -// `corr_dim(n_rot) = n_dims * log(max_pos_emb / (n_rot * 2pi)) / (2 * log(base))` -static float ggml_rope_yarn_corr_dim(int n_dims, int n_ctx_orig, float n_rot, float base) { - return n_dims * logf(n_ctx_orig / (n_rot * 2 * (float)M_PI)) / (2 * logf(base)); -} - static void ggml_rope_cache_init( float theta_base, float freq_scale, const float * freq_factors, float corr_dims[2], int64_t ne0, float ext_factor, float mscale, float * cache, float sin_sign, float theta_scale) { @@ -9181,14 +9133,62 @@ static void ggml_rope_cache_init( } } -void ggml_rope_yarn_corr_dims( - int n_dims, int n_ctx_orig, float freq_base, float beta_fast, float beta_slow, float dims[2] -) { - // start and end correction dims - float start = floorf(ggml_rope_yarn_corr_dim(n_dims, n_ctx_orig, beta_fast, freq_base)); - float end = ceilf(ggml_rope_yarn_corr_dim(n_dims, n_ctx_orig, beta_slow, freq_base)); - dims[0] = MAX(0, start); - dims[1] = MIN(n_dims - 1, end); +static void ggml_mrope_cache_init( + float theta_base_t, float theta_base_h, float theta_base_w, float theta_base_e, int sections[4], bool indep_sects, + float freq_scale, const float * freq_factors, float corr_dims[2], int64_t ne0, float ext_factor, float mscale, + float * cache, float sin_sign, float theta_scale) { + // ref: https://github.com/jquesnelle/yarn/blob/master/scaled_rope/LlamaYaRNScaledRotaryEmbedding.py + float theta_t = theta_base_t; + float theta_h = theta_base_h; + float theta_w = theta_base_w; + float theta_e = theta_base_e; // extra position id for vision encoder + int sect_dims = sections[0] + sections[1] + sections[2] + sections[3]; + int sec_w = sections[1] + sections[0]; + int sec_e = sections[2] + sec_w; + GGML_ASSERT(sect_dims <= ne0); + + for (int64_t i0 = 0; i0 < ne0; i0 += 2) { + const float ff = freq_factors ? freq_factors[i0/2] : 1.0f; + + int sector = (i0 / 2) % sect_dims; + if (indep_sects) { + // compute theta independently for each dim sections + // (i.e. reset corresponding theta when `i0` go from one section to another) + if (sector == 0) { + theta_t = theta_base_t; + } + else if (sector == sections[0]) { + theta_h = theta_base_h;; + } + else if (sector == sec_w) { + theta_w = theta_base_w; + } + else if (sector == sec_e) { + theta_e = theta_base_e; + } + } + + float theta = theta_t; + if (sector >= sections[0] && sector < sec_w) { + theta = theta_h; + } + else if (sector >= sec_w && sector < sec_w + sections[2]) { + theta = theta_w; + } + else if (sector >= sec_w + sections[2]) { + theta = theta_e; + } + + rope_yarn( + theta/ff, freq_scale, corr_dims, i0, ext_factor, mscale, &cache[i0 + 0], &cache[i0 + 1] + ); + cache[i0 + 1] *= sin_sign; + + theta_t *= theta_scale; + theta_w *= theta_scale; + theta_h *= theta_scale; + theta_e *= theta_scale; + } } static void ggml_compute_forward_rope_f32( @@ -9201,6 +9201,7 @@ static void ggml_compute_forward_rope_f32( const struct ggml_tensor * src2 = dst->src[2]; float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + int sections[4]; //const int n_past = ((int32_t *) dst->op_params)[0]; const int n_dims = ((int32_t *) dst->op_params)[1]; @@ -9214,6 +9215,7 @@ static void ggml_compute_forward_rope_f32( memcpy(&attn_factor, (int32_t *) dst->op_params + 8, sizeof(float)); memcpy(&beta_fast, (int32_t *) dst->op_params + 9, sizeof(float)); memcpy(&beta_slow, (int32_t *) dst->op_params + 10, sizeof(float)); + memcpy(§ions, (int32_t *) dst->op_params + 11, sizeof(int)*4); GGML_TENSOR_UNARY_OP_LOCALS @@ -9246,6 +9248,16 @@ static void ggml_compute_forward_rope_f32( ggml_rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow, corr_dims); const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; + const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE; // ggml_rope_multi, multimodal rotary position embedding + const bool is_vision = mode == GGML_ROPE_TYPE_VISION; + + if (is_mrope) { + GGML_ASSERT(sections[0] > 0 || sections[1] > 0 || sections[2] > 0); + } + + if (is_vision) { + GGML_ASSERT(n_dims == ne0/2); + } const float * freq_factors = NULL; if (src2 != NULL) { @@ -9261,18 +9273,63 @@ static void ggml_compute_forward_rope_f32( const int32_t * pos = (const int32_t *) src1->data; - for (int64_t i3 = 0; i3 < ne3; i3++) { - for (int64_t i2 = 0; i2 < ne2; i2++) { - const int64_t p = pos[i2]; + for (int64_t i3 = 0; i3 < ne3; i3++) { // batch + for (int64_t i2 = 0; i2 < ne2; i2++) { // seq-len float * cache = (float *) params->wdata + (ne0 + CACHE_LINE_SIZE_F32)*ith; - ggml_rope_cache_init(p, freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + if (!is_mrope) { + const int64_t p = pos[i2]; + ggml_rope_cache_init(p, freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + } + else { + const int64_t p_t = pos[i2]; + const int64_t p_h = pos[i2 + ne2]; + const int64_t p_w = pos[i2 + ne2 * 2]; + const int64_t p_e = pos[i2 + ne2 * 3]; + ggml_mrope_cache_init( + p_t, p_h, p_w, p_e, sections, is_vision, + freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + } - for (int64_t i1 = 0; i1 < ne1; i1++) { + for (int64_t i1 = 0; i1 < ne1; i1++) { // attn-heads if (ir++ < ir0) continue; if (ir > ir1) break; - if (!is_neox) { + if (is_neox || is_mrope) { + if (is_vision){ + for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + const int64_t ic = i0/2; + + const float cos_theta = cache[i0 + 0]; + const float sin_theta = cache[i0 + 1]; + + const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = src[0]; + const float x1 = src[n_dims]; + + dst_data[0] = x0*cos_theta - x1*sin_theta; + dst_data[n_dims] = x0*sin_theta + x1*cos_theta; + } + } else { + for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + const int64_t ic = i0/2; + + const float cos_theta = cache[i0 + 0]; + const float sin_theta = cache[i0 + 1]; + + const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = src[0]; + const float x1 = src[n_dims/2]; + + dst_data[0] = x0*cos_theta - x1*sin_theta; + dst_data[n_dims/2] = x0*sin_theta + x1*cos_theta; + } + } + } else { for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { const float cos_theta = cache[i0 + 0]; const float sin_theta = cache[i0 + 1]; @@ -9286,8 +9343,10 @@ static void ggml_compute_forward_rope_f32( dst_data[0] = x0*cos_theta - x1*sin_theta; dst_data[1] = x0*sin_theta + x1*cos_theta; } - } else { - for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + } + + if (is_vision) { + for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { const int64_t ic = i0/2; const float cos_theta = cache[i0 + 0]; @@ -9297,19 +9356,20 @@ static void ggml_compute_forward_rope_f32( float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); const float x0 = src[0]; - const float x1 = src[n_dims/2]; + const float x1 = src[n_dims]; - dst_data[0] = x0*cos_theta - x1*sin_theta; - dst_data[n_dims/2] = x0*sin_theta + x1*cos_theta; + dst_data[0] = x0*cos_theta - x1*sin_theta; + dst_data[n_dims] = x0*sin_theta + x1*cos_theta; } - } + } else { + // fill the remain channels with data from src tensor + for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { + const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); - for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { - const float * const src = (float *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); - float * dst_data = (float *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); - - dst_data[0] = src[0]; - dst_data[1] = src[1]; + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } } } } @@ -9327,6 +9387,7 @@ static void ggml_compute_forward_rope_f16( const struct ggml_tensor * src2 = dst->src[2]; float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + int sections[4]; //const int n_past = ((int32_t *) dst->op_params)[0]; const int n_dims = ((int32_t *) dst->op_params)[1]; @@ -9339,6 +9400,8 @@ static void ggml_compute_forward_rope_f16( memcpy(&attn_factor, (int32_t *) dst->op_params + 8, sizeof(float)); memcpy(&beta_fast, (int32_t *) dst->op_params + 9, sizeof(float)); memcpy(&beta_slow, (int32_t *) dst->op_params + 10, sizeof(float)); + memcpy(§ions, (int32_t *) dst->op_params + 11, sizeof(int)*4); + GGML_TENSOR_UNARY_OP_LOCALS @@ -9371,6 +9434,16 @@ static void ggml_compute_forward_rope_f16( ggml_rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow, corr_dims); const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; + const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE; + const bool is_vision = mode == GGML_ROPE_TYPE_VISION; + + if (is_mrope) { + GGML_ASSERT(sections[0] > 0 || sections[1] > 0 || sections[2] > 0); + } + + if (is_vision) { + GGML_ASSERT(n_dims == ne0/2); + } const float * freq_factors = NULL; if (src2 != NULL) { @@ -9388,16 +9461,61 @@ static void ggml_compute_forward_rope_f16( for (int64_t i3 = 0; i3 < ne3; i3++) { for (int64_t i2 = 0; i2 < ne2; i2++) { - const int64_t p = pos[i2]; float * cache = (float *) params->wdata + (ne0 + CACHE_LINE_SIZE_F32)*ith; - ggml_rope_cache_init(p, freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + if (!is_mrope) { + const int64_t p = pos[i2]; + ggml_rope_cache_init(p, freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + } + else { + const int64_t p_t = pos[i2]; + const int64_t p_h = pos[i2 + ne2]; + const int64_t p_w = pos[i2 + ne2 * 2]; + const int64_t p_e = pos[i2 + ne2 * 3]; + ggml_mrope_cache_init( + p_t, p_h, p_w, p_e, sections, is_vision, + freq_scale, freq_factors, corr_dims, ne0, ext_factor, attn_factor, cache, sin_sign, theta_scale); + } for (int64_t i1 = 0; i1 < ne1; i1++) { if (ir++ < ir0) continue; if (ir > ir1) break; - if (!is_neox) { + if (is_neox || is_mrope) { + if (is_vision) { + for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + const int64_t ic = i0/2; + + const float cos_theta = cache[i0 + 0]; + const float sin_theta = cache[i0 + 1]; + + const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = GGML_FP16_TO_FP32(src[0]); + const float x1 = GGML_FP16_TO_FP32(src[n_dims]); + + dst_data[0] = GGML_FP32_TO_FP16(x0*cos_theta - x1*sin_theta); + dst_data[n_dims] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); + } + } else { + for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + const int64_t ic = i0/2; + + const float cos_theta = cache[i0 + 0]; + const float sin_theta = cache[i0 + 1]; + + const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = GGML_FP16_TO_FP32(src[0]); + const float x1 = GGML_FP16_TO_FP32(src[n_dims/2]); + + dst_data[0] = GGML_FP32_TO_FP16(x0*cos_theta - x1*sin_theta); + dst_data[n_dims/2] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); + } + } + } else { for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { const float cos_theta = cache[i0 + 0]; const float sin_theta = cache[i0 + 1]; @@ -9411,8 +9529,10 @@ static void ggml_compute_forward_rope_f16( dst_data[0] = GGML_FP32_TO_FP16(x0*cos_theta - x1*sin_theta); dst_data[1] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); } - } else { - for (int64_t i0 = 0; i0 < n_dims; i0 += 2) { + } + + if (is_vision) { + for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { const int64_t ic = i0/2; const float cos_theta = cache[i0 + 0]; @@ -9422,19 +9542,19 @@ static void ggml_compute_forward_rope_f16( ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); const float x0 = GGML_FP16_TO_FP32(src[0]); - const float x1 = GGML_FP16_TO_FP32(src[n_dims/2]); + const float x1 = GGML_FP16_TO_FP32(src[n_dims]); - dst_data[0] = GGML_FP32_TO_FP16(x0*cos_theta - x1*sin_theta); - dst_data[n_dims/2] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); + dst_data[0] = GGML_FP32_TO_FP16(x0*cos_theta - x1*sin_theta); + dst_data[n_dims] = GGML_FP32_TO_FP16(x0*sin_theta + x1*cos_theta); } - } + } else { + for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { + const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); - for (int64_t i0 = n_dims; i0 < ne0; i0 += 2) { - const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); - ggml_fp16_t * dst_data = (ggml_fp16_t *)((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); - - dst_data[0] = src[0]; - dst_data[1] = src[1]; + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } } } } @@ -10433,6 +10553,40 @@ static void ggml_compute_forward_pad( } } +// ggml_compute_forward_pad_reflect_1d + +static void ggml_compute_forward_pad_reflect_1d( + const struct ggml_compute_params * params, + struct ggml_tensor * dst) { + + const struct ggml_tensor * src0 = dst->src[0]; + + GGML_ASSERT(src0->type == GGML_TYPE_F32); + GGML_ASSERT( dst->type == GGML_TYPE_F32); + + const int ith = params->ith; + const int nth = params->nth; + + const int32_t * opts = (const int32_t *) dst->op_params; + const int p0 = opts[0]; + const int p1 = opts[1]; + + GGML_TENSOR_UNARY_OP_LOCALS + + for (int64_t i3 = 0; i3 < ne3; i3++) { + for (int64_t i2 = 0; i2 < ne2; i2++) { + for (int64_t i1 = ith; i1 < ne1; i1 += nth) { + float * left = (float *) ((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + p0*nb0); + float * right = (float *) ((char *) dst->data + i3*nb3 + i2*nb2 + i1*nb1 + (ne0-p1-1)*nb0); + + ggml_vec_cpy_f32(ne00, left, (float *) ((char *) src0->data + i3*nb03 + i2*nb02 + i1*nb01)); + + for (int i0 = 1; i0 <= p0; i0++) { left[-i0] = left[i0]; } + for (int i0 = 1; i0 <= p1; i0++) { right[i0] = right[-i0]; } + } + } + } +} // ggml_compute_forward_arange @@ -10668,7 +10822,7 @@ static void ggml_compute_forward_flash_attn_ext_f16( const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); enum ggml_type const k_vec_dot_type = type_traits_cpu[k->type].vec_dot_type; - ggml_from_float_t const q_to_vec_dot = ggml_get_type_traits(k_vec_dot_type)->from_float; + ggml_from_float_t const q_to_vec_dot = type_traits_cpu[k_vec_dot_type].from_float; ggml_vec_dot_t const kq_vec_dot = type_traits_cpu[k->type].vec_dot; ggml_to_float_t const v_to_float = ggml_get_type_traits(v->type)->to_float; @@ -11649,9 +11803,9 @@ static void ggml_compute_forward_add_rel_pos( static void ggml_compute_forward_rwkv_wkv6_f32( const struct ggml_compute_params * params, struct ggml_tensor * dst) { - const int64_t T = dst->src[1]->ne[3]; + const int64_t T = dst->src[1]->ne[2]; const int64_t C = dst->ne[0]; - const int64_t HEADS = dst->src[1]->ne[2]; + const int64_t HEADS = dst->src[1]->ne[1]; const int64_t n_seqs = dst->src[5]->ne[1]; const int64_t head_size = C / HEADS; @@ -11846,6 +12000,197 @@ static void ggml_compute_forward_rwkv_wkv6( } } +// ggml_compute_forward_gla + +static void ggml_compute_forward_gla_f32( + const struct ggml_compute_params * params, + struct ggml_tensor * dst) { + const int64_t T = dst->src[1]->ne[2]; + const int64_t C = dst->ne[0]; + const int64_t HEADS = dst->src[1]->ne[1]; + const int64_t n_seqs = dst->src[4]->ne[1]; + const int64_t head_size = C / HEADS; + const float scale = ggml_get_op_params_f32(dst, 0); + + float * dst_data = (float *) dst->data; + float * state = ((float *) dst->data) + C * T; + + const int ith = params->ith; + const int nth = params->nth; + + if (ith >= HEADS) { + return; + } + + const int h_start = (HEADS * ith) / nth; + const int h_end = ((HEADS * (ith + 1)) / nth < HEADS) ? + (HEADS * (ith + 1)) / nth : HEADS; + + float * k = (float *) dst->src[0]->data; + float * v = (float *) dst->src[1]->data; + float * q = (float *) dst->src[2]->data; + float * g = (float *) dst->src[3]->data; + + size_t t_stride = HEADS * head_size; // Same to C + + size_t h_stride = C / HEADS; + GGML_ASSERT(C % HEADS == 0); // C must be divisible by HEADS + size_t h_stride_2d = head_size * head_size; + + if (ith == 0) { + memset(dst_data, 0, T * C * sizeof(float)); + } + ggml_barrier(params->threadpool); + + + #if defined(__AVX__) && !defined(__AVX512F__) + #define GGML_F32X GGML_F32x8 + #define GGML_F32X_SET1 GGML_F32x8_SET1 + #define GGML_F32X_LOAD GGML_F32x8_LOAD + #define GGML_F32X_STORE GGML_F32x8_STORE + #define GGML_F32X_MUL GGML_F32x8_MUL + #define GGML_F32X_FMA GGML_F32x8_FMA + #define GLA_VECTOR_SIZE 8 + #elif defined(__AVX512F__) + #define GGML_F32X GGML_F32x16 + #define GGML_F32X_SET1 GGML_F32x16_SET1 + #define GGML_F32X_LOAD GGML_F32x16_LOAD + #define GGML_F32X_STORE GGML_F32x16_STORE + #define GGML_F32X_MUL GGML_F32x16_MUL + #define GGML_F32X_FMA GGML_F32x16_FMA + #define GLA_VECTOR_SIZE 16 + #elif defined(__ARM_NEON) && defined(__aarch64__) + #define GGML_F32X GGML_F32x4 + #define GGML_F32X_SET1 GGML_F32x4_SET1 + #define GGML_F32X_LOAD GGML_F32x4_LOAD + #define GGML_F32X_STORE GGML_F32x4_STORE + #define GGML_F32X_MUL GGML_F32x4_MUL + #define GGML_F32X_FMA GGML_F32x4_FMA + #define GLA_VECTOR_SIZE 4 + #endif + + #ifdef GLA_VECTOR_SIZE + const int64_t vec_count = head_size / GLA_VECTOR_SIZE; + + for (int64_t t = 0; t < T; t++) { + size_t t_offset = t * t_stride; + size_t state_offset = head_size * C * (t / (T / n_seqs)); + float * state_cur = state + state_offset; + float * state_prev = t % (T / n_seqs) ? state_cur : (float*)dst->src[4]->data + state_offset; + + for (int64_t h = h_start; h < h_end; h++) { + size_t h_offset = h * h_stride; + size_t t_h_offset = t_offset + h_offset; + size_t h_2d_offset = h * h_stride_2d; + + for (int64_t i = 0; i < head_size; i++) { + size_t t_h_i_offset = t_h_offset + i; + size_t h_2d_i_offset = h_2d_offset + i * h_stride; + + float k_val = k[t_h_i_offset]; + float q_val = q[t_h_i_offset] * scale; + float g_val = g[t_h_i_offset]; + + // Broadcast scalar values to vectors + GGML_F32X k_vec = GGML_F32X_SET1(k_val); + GGML_F32X q_vec = GGML_F32X_SET1(q_val); + GGML_F32X g_vec = GGML_F32X_SET1(g_val); + + for (int64_t j = 0; j < vec_count; j++) { + size_t base_j = j * GLA_VECTOR_SIZE; + size_t t_h_j_offset = t_h_offset + base_j; + size_t h_2d_i_j_offset = h_2d_i_offset + base_j; + + // Load x elements at once + GGML_F32X v_vec = GGML_F32X_LOAD(&v[t_h_j_offset]); + GGML_F32X prev_state_vec = GGML_F32X_LOAD(&state_prev[h_2d_i_j_offset]); + GGML_F32X dst_vec = GGML_F32X_LOAD(&dst_data[t_h_j_offset]); + + // Compute kv = v * k + GGML_F32X kv_vec = GGML_F32X_MUL(v_vec, k_vec); + + // Compute temp = prev_state * g + kv + GGML_F32X temp_vec = GGML_F32X_FMA(kv_vec, prev_state_vec, g_vec); + + // Update dst: dst += temp * q + dst_vec = GGML_F32X_FMA(dst_vec, temp_vec, q_vec); + GGML_F32X_STORE(&dst_data[t_h_j_offset], dst_vec); + + // Update state + GGML_F32X_STORE(&state_cur[h_2d_i_j_offset], temp_vec); + } + + // Handle remaining elements, this will not be used. + for (int64_t j = vec_count * GLA_VECTOR_SIZE; j < head_size; j++) { + size_t t_h_j_offset = t_h_offset + j; + size_t h_2d_i_j_offset = h_2d_i_offset + j; + float v_val = v[t_h_j_offset]; + float kv_val = v_val * k_val; + float prev_state_val = state_prev[h_2d_i_j_offset]; + float temp_val = kv_val + prev_state_val * g_val; + dst_data[t_h_j_offset] += temp_val * q_val; + state_cur[h_2d_i_j_offset] = temp_val; + } + } + } + } + + #else + for (int64_t t = 0; t < T; t++) { + size_t t_offset = t * t_stride; + size_t state_offset = head_size * C * (t / (T / n_seqs)); + float * state_cur = state + state_offset; + float * state_prev = t % (T / n_seqs) ? state_cur : (float*)dst->src[4]->data + state_offset; + + for (int64_t h = h_start; h < h_end; h++) { + size_t h_offset = h * h_stride; + size_t t_h_offset = t_offset + h_offset; + size_t h_2d_offset = h * h_stride_2d; + + for (int64_t i = 0; i < head_size; i++) { + size_t t_h_i_offset = t_h_offset + i; + size_t h_2d_i_offset = h_2d_offset + i * h_stride; + + float k_val = k[t_h_i_offset]; + float q_val = q[t_h_i_offset] * scale; + float g_val = g[t_h_i_offset]; + + for (int64_t j = 0; j < head_size; j++) { + size_t t_h_j_offset = t_h_offset + j; + size_t h_2d_i_j_offset = h_2d_i_offset + j; + + float v_val = v[t_h_j_offset]; + float kv_val = v_val * k_val; + float prev_state_val = state_prev[h_2d_i_j_offset]; + float temp_val = prev_state_val * g_val + kv_val; + dst_data[t_h_j_offset] += temp_val * q_val; + state_cur[h_2d_i_j_offset] = temp_val; + } + } + } + } + #endif +} + + +static void ggml_compute_forward_gla( + const struct ggml_compute_params * params, + struct ggml_tensor * dst) { + + const struct ggml_tensor * src0 = dst->src[0]; + + switch (src0->type) { + case GGML_TYPE_F32: + { + ggml_compute_forward_gla_f32(params, dst); + } break; + default: + { + GGML_ABORT("fatal error"); + } + } +} + // ggml_compute_forward_map_unary static void ggml_compute_forward_map_unary_f32( @@ -12220,11 +12565,16 @@ static void ggml_compute_forward_opt_step_adamw_f32( const struct ggml_compute_params * params, struct ggml_tensor * dst) { - const struct ggml_tensor * src0 = dst->src[0]; - const struct ggml_tensor * src0_grad = dst->src[1]; - const struct ggml_tensor * src0_grad_m = dst->src[2]; - const struct ggml_tensor * src0_grad_v = dst->src[3]; + const struct ggml_tensor * src0 = dst->src[0]; + const struct ggml_tensor * src0_grad = dst->src[1]; + const struct ggml_tensor * src0_grad_m = dst->src[2]; + const struct ggml_tensor * src0_grad_v = dst->src[3]; + const struct ggml_tensor * adamw_params = dst->src[4]; + GGML_ASSERT(ggml_are_same_shape(src0, src0_grad)); + GGML_ASSERT(ggml_are_same_shape(src0, src0_grad_m)); + GGML_ASSERT(ggml_are_same_shape(src0, src0_grad_v)); + GGML_ASSERT(ggml_nelements(adamw_params) == 7); const int ith = params->ith; const int nth = params->nth; @@ -12241,16 +12591,14 @@ static void ggml_compute_forward_opt_step_adamw_f32( const int ir0 = dr*ith; const int ir1 = MIN(ir0 + dr, nr); - /* const float gnorm = 1.0f; */ - int64_t iter; memcpy(&iter, &dst->op_params[0], sizeof(int64_t)); - const float alpha = ggml_get_op_params_f32(dst, 2); - const float beta1 = ggml_get_op_params_f32(dst, 3); - const float beta2 = ggml_get_op_params_f32(dst, 4); - const float eps = ggml_get_op_params_f32(dst, 5); - const float wd = ggml_get_op_params_f32(dst, 6); - - const float beta1h = alpha/(1.0f - powf(beta1, iter)); - const float beta2h = 1.0f/(1.0f - powf(beta2, iter)); + const float * adamw_params_ptr = ggml_get_data_f32(adamw_params); + const float alpha = adamw_params_ptr[0]; + const float beta1 = adamw_params_ptr[1]; + const float beta2 = adamw_params_ptr[2]; + const float eps = adamw_params_ptr[3]; + const float wd = adamw_params_ptr[4]; + const float beta1h = adamw_params_ptr[5]; + const float beta2h = adamw_params_ptr[6]; for (int ir = ir0; ir < ir1; ++ir) { const int64_t i03 = ir/(ne02*ne01); @@ -12274,17 +12622,9 @@ static void ggml_compute_forward_opt_step_adamw_f32( // The weight decay is applied independently of the Adam momenta m and v. // This is NOT equivalent to l2 regularization that adds w[i00]*w[i00] to the loss. // See: https://arxiv.org/pdf/1711.05101v3.pdf - w[i00] = w[i00]*(1.0f - alpha*wd) - mh/vh; + w[i00] = w[i00]*(1.0f - alpha*wd) - alpha*mh/vh; } } - - ggml_barrier(params->threadpool); - if (ith != 0) { - return; - } - - iter++; - memcpy(&dst->op_params[0], &iter, sizeof(int64_t)); } static void ggml_compute_forward_opt_step_adamw( @@ -12313,6 +12653,9 @@ static void ggml_compute_forward(struct ggml_compute_params * params, struct ggm return; } + // extra_buffer op? + if (ggml_cpu_extra_compute_forward(params, tensor)) return; + switch (tensor->op) { case GGML_OP_DUP: { @@ -12534,6 +12877,10 @@ static void ggml_compute_forward(struct ggml_compute_params * params, struct ggm { ggml_compute_forward_pad(params, tensor); } break; + case GGML_OP_PAD_REFLECT_1D: + { + ggml_compute_forward_pad_reflect_1d(params, tensor); + } break; case GGML_OP_ARANGE: { ggml_compute_forward_arange(params, tensor); @@ -12593,6 +12940,10 @@ static void ggml_compute_forward(struct ggml_compute_params * params, struct ggm { ggml_compute_forward_rwkv_wkv6(params, tensor); } break; + case GGML_OP_GATED_LINEAR_ATTN: + { + ggml_compute_forward_gla(params, tensor); + } break; case GGML_OP_MAP_UNARY: { ggml_unary_op_f32_t fun; @@ -12876,6 +13227,7 @@ static int ggml_get_n_tasks(struct ggml_tensor * node, int n_threads) { } break; case GGML_OP_UPSCALE: case GGML_OP_PAD: + case GGML_OP_PAD_REFLECT_1D: case GGML_OP_ARANGE: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_ARGSORT: @@ -12890,6 +13242,7 @@ static int ggml_get_n_tasks(struct ggml_tensor * node, int n_threads) { case GGML_OP_WIN_UNPART: case GGML_OP_GET_REL_POS: case GGML_OP_RWKV_WKV6: + case GGML_OP_GATED_LINEAR_ATTN: case GGML_OP_MAP_UNARY: case GGML_OP_MAP_BINARY: case GGML_OP_MAP_CUSTOM1_F32: @@ -12965,7 +13318,7 @@ static thread_ret_t ggml_graph_compute_secondary_thread(void* data); #include "windows.h" // TODO: support > 64 CPUs -bool ggml_thread_apply_affinity(bool * mask) { +static bool ggml_thread_apply_affinity(bool * mask) { HANDLE h = GetCurrentThread(); uint64_t bitmask = 0ULL; @@ -13255,140 +13608,143 @@ struct ggml_cplan ggml_graph_plan( size_t cur = 0; - switch (node->op) { - case GGML_OP_CPY: - case GGML_OP_DUP: - { - if (ggml_is_quantized(node->type) || - // F16 -> BF16 and BF16 -> F16 copies go through intermediate F32 - (node->src[0]->type == GGML_TYPE_F16 && node->src[1] && node->src[1]->type == GGML_TYPE_BF16) || - (node->src[0]->type == GGML_TYPE_BF16 && node->src[1] && node->src[1]->type == GGML_TYPE_F16)) { + if (!ggml_cpu_extra_work_size(n_threads, node, &cur)) { + + switch (node->op) { + case GGML_OP_CPY: + case GGML_OP_DUP: + { + if (ggml_is_quantized(node->type) || + // F16 -> BF16 and BF16 -> F16 copies go through intermediate F32 + (node->src[0]->type == GGML_TYPE_F16 && node->src[1] && node->src[1]->type == GGML_TYPE_BF16) || + (node->src[0]->type == GGML_TYPE_BF16 && node->src[1] && node->src[1]->type == GGML_TYPE_F16)) { + cur = ggml_type_size(GGML_TYPE_F32) * node->ne[0] * n_tasks; + } + } break; + case GGML_OP_ADD: + case GGML_OP_ADD1: + { + if (ggml_is_quantized(node->src[0]->type)) { + cur = ggml_type_size(GGML_TYPE_F32) * node->src[0]->ne[0] * n_tasks; + } + } break; + case GGML_OP_ACC: + { + if (ggml_is_quantized(node->src[0]->type)) { + cur = ggml_type_size(GGML_TYPE_F32) * node->src[1]->ne[0] * n_tasks; + } + } break; + case GGML_OP_COUNT_EQUAL: + { + cur = ggml_type_size(node->type)*n_tasks; + } break; + case GGML_OP_MUL_MAT: + { + const enum ggml_type vec_dot_type = type_traits_cpu[node->src[0]->type].vec_dot_type; + + if (node->src[1]->type != vec_dot_type) { + cur = ggml_row_size(vec_dot_type, ggml_nelements(node->src[1])); + } + } break; + case GGML_OP_MUL_MAT_ID: + { + cur = 0; + const struct ggml_tensor * src0 = node->src[0]; + const struct ggml_tensor * src1 = node->src[1]; + const enum ggml_type vec_dot_type = type_traits_cpu[src0->type].vec_dot_type; + if (src1->type != vec_dot_type) { + cur += ggml_row_size(vec_dot_type, ggml_nelements(src1)); + } + const int n_as = src0->ne[2]; + cur += GGML_PAD(cur, sizeof(int64_t)); // align + cur += n_as * sizeof(int64_t); // matrix_row_counts + cur += n_as * src1->ne[2] * sizeof(int64_t); // matrix_rows + } break; + case GGML_OP_OUT_PROD: + { + if (ggml_is_quantized(node->src[0]->type)) { + cur = ggml_type_size(GGML_TYPE_F32) * node->src[0]->ne[0] * n_tasks; + } + } break; + case GGML_OP_SOFT_MAX: + case GGML_OP_ROPE: + case GGML_OP_ROPE_BACK: + { cur = ggml_type_size(GGML_TYPE_F32) * node->ne[0] * n_tasks; - } - } break; - case GGML_OP_ADD: - case GGML_OP_ADD1: - { - if (ggml_is_quantized(node->src[0]->type)) { - cur = ggml_type_size(GGML_TYPE_F32) * node->src[0]->ne[0] * n_tasks; - } - } break; - case GGML_OP_ACC: - { - if (ggml_is_quantized(node->src[0]->type)) { - cur = ggml_type_size(GGML_TYPE_F32) * node->src[1]->ne[0] * n_tasks; - } - } break; - case GGML_OP_COUNT_EQUAL: - { - cur = ggml_type_size(node->type)*n_tasks; - } break; - case GGML_OP_MUL_MAT: - { - const enum ggml_type vec_dot_type = type_traits_cpu[node->src[0]->type].vec_dot_type; + } break; + case GGML_OP_CONV_TRANSPOSE_1D: + { + GGML_ASSERT(node->src[0]->ne[3] == 1); + GGML_ASSERT(node->src[1]->ne[2] == 1); + GGML_ASSERT(node->src[1]->ne[3] == 1); - if (node->src[1]->type != vec_dot_type) { - cur = ggml_row_size(vec_dot_type, ggml_nelements(node->src[1])); - } - } break; - case GGML_OP_MUL_MAT_ID: - { - cur = 0; - const struct ggml_tensor * src0 = node->src[0]; - const struct ggml_tensor * src1 = node->src[1]; - const enum ggml_type vec_dot_type = type_traits_cpu[src0->type].vec_dot_type; - if (src1->type != vec_dot_type) { - cur += ggml_row_size(vec_dot_type, ggml_nelements(src1)); - } - const int n_as = src0->ne[2]; - cur += GGML_PAD(cur, sizeof(int64_t)); // align - cur += n_as * sizeof(int64_t); // matrix_row_counts - cur += n_as * src1->ne[2] * sizeof(int64_t); // matrix_rows - } break; - case GGML_OP_OUT_PROD: - { - if (ggml_is_quantized(node->src[0]->type)) { - cur = ggml_type_size(GGML_TYPE_F32) * node->src[0]->ne[0] * n_tasks; - } - } break; - case GGML_OP_SOFT_MAX: - case GGML_OP_ROPE: - { - cur = ggml_type_size(GGML_TYPE_F32) * node->ne[0] * n_tasks; - } break; - case GGML_OP_CONV_TRANSPOSE_1D: - { - GGML_ASSERT(node->src[0]->ne[3] == 1); - GGML_ASSERT(node->src[1]->ne[2] == 1); - GGML_ASSERT(node->src[1]->ne[3] == 1); + const int64_t ne00 = node->src[0]->ne[0]; // K + const int64_t ne01 = node->src[0]->ne[1]; // Cout + const int64_t ne02 = node->src[0]->ne[2]; // Cin + const int64_t ne10 = node->src[1]->ne[0]; // L + const int64_t ne11 = node->src[1]->ne[1]; // Cin - const int64_t ne00 = node->src[0]->ne[0]; // K - const int64_t ne01 = node->src[0]->ne[1]; // Cout - const int64_t ne02 = node->src[0]->ne[2]; // Cin + if ((node->src[0]->type == GGML_TYPE_F16 || + node->src[0]->type == GGML_TYPE_BF16) && + node->src[1]->type == GGML_TYPE_F32) { + cur += sizeof(ggml_fp16_t)*ne00*ne01*ne02; + cur += sizeof(ggml_fp16_t)*ne10*ne11; + } else if (node->src[0]->type == GGML_TYPE_F32 && + node->src[1]->type == GGML_TYPE_F32) { + cur += sizeof(float)*ne00*ne01*ne02; + cur += sizeof(float)*ne10*ne11; + } else { + GGML_ABORT("fatal error"); + } + } break; + case GGML_OP_CONV_TRANSPOSE_2D: + { + const int64_t ne00 = node->src[0]->ne[0]; // W + const int64_t ne01 = node->src[0]->ne[1]; // H + const int64_t ne02 = node->src[0]->ne[2]; // Channels Out + const int64_t ne03 = node->src[0]->ne[3]; // Channels In - const int64_t ne10 = node->src[1]->ne[0]; // L - const int64_t ne11 = node->src[1]->ne[1]; // Cin + const int64_t ne10 = node->src[1]->ne[0]; // W + const int64_t ne11 = node->src[1]->ne[1]; // H + const int64_t ne12 = node->src[1]->ne[2]; // Channels In - if ((node->src[0]->type == GGML_TYPE_F16 || - node->src[0]->type == GGML_TYPE_BF16) && - node->src[1]->type == GGML_TYPE_F32) { - cur += sizeof(ggml_fp16_t)*ne00*ne01*ne02; - cur += sizeof(ggml_fp16_t)*ne10*ne11; - } else if (node->src[0]->type == GGML_TYPE_F32 && - node->src[1]->type == GGML_TYPE_F32) { - cur += sizeof(float)*ne00*ne01*ne02; - cur += sizeof(float)*ne10*ne11; - } else { + cur += sizeof(ggml_fp16_t)*ne00*ne01*ne02*ne03; + cur += sizeof(ggml_fp16_t)*ne10*ne11*ne12; + } break; + case GGML_OP_FLASH_ATTN_EXT: + { + const int64_t ne00 = node->src[0]->ne[0]; // D + + cur = 3*sizeof(float)*ne00*n_tasks; // 3x head size/thread + } break; + case GGML_OP_FLASH_ATTN_BACK: + { + const int64_t D = node->src[0]->ne[0]; + const int64_t ne11 = ggml_up(node->src[1]->ne[1], GGML_SOFT_MAX_UNROLL); + const int64_t mxDn = MAX(D, ne11) * 2; // *2 because of S and SM in ggml_compute_forward_flash_attn_back + if (node->src[1]->type == GGML_TYPE_F32) { + cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) + cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 + } else if (node->src[1]->type == GGML_TYPE_F16) { + cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) + cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 + } else if (node->src[1]->type == GGML_TYPE_BF16) { + cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) + cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 + } + } break; + + case GGML_OP_CROSS_ENTROPY_LOSS: + { + cur = ggml_type_size(node->type)*(n_tasks + node->src[0]->ne[0]*n_tasks); + } break; + case GGML_OP_COUNT: + { GGML_ABORT("fatal error"); } - } break; - case GGML_OP_CONV_TRANSPOSE_2D: - { - const int64_t ne00 = node->src[0]->ne[0]; // W - const int64_t ne01 = node->src[0]->ne[1]; // H - const int64_t ne02 = node->src[0]->ne[2]; // Channels Out - const int64_t ne03 = node->src[0]->ne[3]; // Channels In - - const int64_t ne10 = node->src[1]->ne[0]; // W - const int64_t ne11 = node->src[1]->ne[1]; // H - const int64_t ne12 = node->src[1]->ne[2]; // Channels In - - cur += sizeof(ggml_fp16_t)*ne00*ne01*ne02*ne03; - cur += sizeof(ggml_fp16_t)*ne10*ne11*ne12; - } break; - case GGML_OP_FLASH_ATTN_EXT: - { - const int64_t ne00 = node->src[0]->ne[0]; // D - - cur = 3*sizeof(float)*ne00*n_tasks; // 3x head size/thread - } break; - case GGML_OP_FLASH_ATTN_BACK: - { - const int64_t D = node->src[0]->ne[0]; - const int64_t ne11 = ggml_up(node->src[1]->ne[1], GGML_SOFT_MAX_UNROLL); - const int64_t mxDn = MAX(D, ne11) * 2; // *2 because of S and SM in ggml_compute_forward_flash_attn_back - if (node->src[1]->type == GGML_TYPE_F32) { - cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) - cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 - } else if (node->src[1]->type == GGML_TYPE_F16) { - cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) - cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 - } else if (node->src[1]->type == GGML_TYPE_BF16) { - cur = sizeof(float)*mxDn*n_tasks; // TODO: this can become (n_tasks-1) - cur += sizeof(float)*mxDn*n_tasks; // this is overestimated by x2 - } - } break; - - case GGML_OP_CROSS_ENTROPY_LOSS: - { - cur = ggml_type_size(node->type)*(n_tasks + node->src[0]->ne[0]*n_tasks); - } break; - case GGML_OP_COUNT: - { - GGML_ABORT("fatal error"); - } - default: - break; + default: + break; + } } work_size = MAX(work_size, cur); @@ -13587,29 +13943,6 @@ static void ggml_graph_compute_kickoff(struct ggml_threadpool * threadpool, int #endif // GGML_USE_OPENMP -void ggml_threadpool_params_init(struct ggml_threadpool_params * p, int n_threads) { - p->n_threads = n_threads; - p->prio = 0; // default priority (usually means normal or inherited) - p->poll = 50; // hybrid-polling enabled - p->strict_cpu = false; // no strict placement (all threads share same cpumask) - p->paused = false; // threads are ready to go - memset(p->cpumask, 0, GGML_MAX_N_THREADS); // all-zero means use the default affinity (usually inherited) -} - -struct ggml_threadpool_params ggml_threadpool_params_default(int n_threads) { - struct ggml_threadpool_params p; - ggml_threadpool_params_init(&p, n_threads); - return p; -} - -bool ggml_threadpool_params_match(const struct ggml_threadpool_params * p0, const struct ggml_threadpool_params * p1) { - if (p0->n_threads != p1->n_threads ) return false; - if (p0->prio != p1->prio ) return false; - if (p0->poll != p1->poll ) return false; - if (p0->strict_cpu != p1->strict_cpu ) return false; - return memcmp(p0->cpumask, p1->cpumask, GGML_MAX_N_THREADS) == 0; -} - static struct ggml_threadpool * ggml_threadpool_new_impl( struct ggml_threadpool_params * tpp, struct ggml_cgraph * cgraph, @@ -13759,16 +14092,169 @@ enum ggml_status ggml_graph_compute_with_ctx(struct ggml_context * ctx, struct g return ggml_graph_compute(cgraph, &cplan); } + +int ggml_cpu_has_avx(void) { +#if defined(__AVX__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx_vnni(void) { +#if defined(__AVXVNNI__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx2(void) { +#if defined(__AVX2__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx512(void) { +#if defined(__AVX512F__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx512_vbmi(void) { +#if defined(__AVX512VBMI__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx512_vnni(void) { +#if defined(__AVX512VNNI__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_avx512_bf16(void) { +#if defined(__AVX512BF16__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_amx_int8(void) { +#if defined(__AMX_INT8__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_fma(void) { +#if defined(__FMA__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_arm_fma(void) { +#if defined(__ARM_FEATURE_FMA) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_riscv_v(void) { +#if defined(__riscv_v_intrinsic) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_f16c(void) { +#if defined(__F16C__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_fp16_va(void) { +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_wasm_simd(void) { +#if defined(__wasm_simd128__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_llamafile(void) { +#if defined(GGML_USE_LLAMAFILE) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_sse3(void) { +#if defined(__SSE3__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_ssse3(void) { +#if defined(__SSSE3__) + return 1; +#else + return 0; +#endif +} + +int ggml_cpu_has_vsx(void) { +#if defined(__POWER9_VECTOR__) + return 1; +#else + return 0; +#endif +} + int ggml_cpu_has_neon(void) { -#if defined(__ARM_ARCH) +#if defined(__ARM_ARCH) && defined(__ARM_NEON) return ggml_arm_arch_features.has_neon; #else return 0; #endif } +int ggml_cpu_has_dotprod(void) { +#if defined(__ARM_ARCH) && defined(__ARM_FEATURE_DOTPROD) + return ggml_arm_arch_features.has_dotprod; +#else + return 0; +#endif +} + int ggml_cpu_has_sve(void) { -#if defined(__ARM_ARCH) +#if defined(__ARM_ARCH) && defined(__ARM_FEATURE_SVE) return ggml_arm_arch_features.has_sve; #else return 0; @@ -13776,7 +14262,7 @@ int ggml_cpu_has_sve(void) { } int ggml_cpu_has_matmul_int8(void) { -#if defined(__ARM_ARCH) +#if defined(__ARM_ARCH) && defined(__ARM_FEATURE_MATMUL_INT8) return ggml_arm_arch_features.has_i8mm; #else return 0; @@ -13784,7 +14270,7 @@ int ggml_cpu_has_matmul_int8(void) { } int ggml_cpu_get_sve_cnt(void) { -#if defined(__ARM_ARCH) +#if defined(__ARM_ARCH) && defined(__ARM_FEATURE_SVE) return ggml_arm_arch_features.sve_cnt; #else return 0; diff --git a/ggml/src/ggml-cpu/ggml-cpu.cpp b/ggml/src/ggml-cpu/ggml-cpu.cpp new file mode 100644 index 000000000..5c47ceb73 --- /dev/null +++ b/ggml/src/ggml-cpu/ggml-cpu.cpp @@ -0,0 +1,626 @@ +#include "ggml-backend.h" +#include "ggml-backend-impl.h" +#include "ggml-cpu.h" +#include "ggml-cpu-aarch64.h" +#include "ggml-cpu-traits.h" +#include "ggml-impl.h" +#include "amx/amx.h" + +#include +#include +#include + +#ifdef GGML_USE_CPU_HBM +#include "ggml-cpu-hbm.h" +#endif + +#if defined(__APPLE__) +#include +#include +#endif + +#if defined(_WIN32) +#define WIN32_LEAN_AND_MEAN +#ifndef NOMINMAX + #define NOMINMAX +#endif +#include +#endif + +// ggml-backend interface + +std::vector& ggml_backend_cpu_get_extra_buffers_type() { + static std::vector bufts = []() { + std::vector bufts; + +#if defined(__AMX_INT8__) && defined(__AVX512VNNI__) + if (ggml_backend_amx_buffer_type()) { + bufts.push_back(ggml_backend_amx_buffer_type()); + } +#endif + +#ifdef GGML_USE_CPU_AARCH64 + if (ggml_backend_cpu_aarch64_buffer_type()) { + bufts.push_back(ggml_backend_cpu_aarch64_buffer_type()); + } +#endif + + bufts.push_back(NULL); + + return bufts; + }(); + + return bufts; +} + +static ggml_backend_buffer_type_t * ggml_backend_cpu_device_get_extra_buffers_type(ggml_backend_dev_t device) { + return ggml_backend_cpu_get_extra_buffers_type().data(); + + GGML_UNUSED(device); +} + +static bool ggml_backend_cpu_is_extra_buffer_type(ggml_backend_buffer_type_t buft) { + for (auto extra : ggml_backend_cpu_get_extra_buffers_type()) { + if (extra && extra == buft) return true; + } + return false; +} + +// CPU backend - backend (stream) + +struct ggml_backend_cpu_context { + int n_threads; + ggml_threadpool_t threadpool; + + uint8_t * work_data; + size_t work_size; + + ggml_abort_callback abort_callback; + void * abort_callback_data; +}; + +static const char * ggml_backend_cpu_get_name(ggml_backend_t backend) { + return "CPU"; + + GGML_UNUSED(backend); +} + +static void ggml_backend_cpu_free(ggml_backend_t backend) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + delete[] cpu_ctx->work_data; + delete cpu_ctx; + delete backend; +} + +struct ggml_backend_plan_cpu { + struct ggml_cplan cplan; + struct ggml_cgraph cgraph; +}; + +static ggml_backend_graph_plan_t ggml_backend_cpu_graph_plan_create(ggml_backend_t backend, const struct ggml_cgraph * cgraph) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + + struct ggml_backend_plan_cpu * cpu_plan = new ggml_backend_plan_cpu; + + cpu_plan->cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads, cpu_ctx->threadpool); + cpu_plan->cgraph = *cgraph; // FIXME: deep copy + + if (cpu_plan->cplan.work_size > 0) { + cpu_plan->cplan.work_data = new uint8_t[cpu_plan->cplan.work_size]; + if (cpu_plan->cplan.work_data == NULL) { + delete cpu_plan; + return NULL; + } + } + + cpu_plan->cplan.abort_callback = cpu_ctx->abort_callback; + cpu_plan->cplan.abort_callback_data = cpu_ctx->abort_callback_data; + + return cpu_plan; +} + +static void ggml_backend_cpu_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; + + delete[] cpu_plan->cplan.work_data; + delete cpu_plan; + + GGML_UNUSED(backend); +} + +static enum ggml_status ggml_backend_cpu_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; + + return ggml_graph_compute(&cpu_plan->cgraph, &cpu_plan->cplan); + + GGML_UNUSED(backend); +} + +static enum ggml_status ggml_backend_cpu_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + + struct ggml_cplan cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads, cpu_ctx->threadpool); + + if (cpu_ctx->work_size < cplan.work_size) { + delete[] cpu_ctx->work_data; + cpu_ctx->work_data = new uint8_t[cplan.work_size]; + if (cpu_ctx->work_data == NULL) { + cpu_ctx->work_size = 0; + return GGML_STATUS_ALLOC_FAILED; + } + cpu_ctx->work_size = cplan.work_size; + } + cplan.work_data = (uint8_t *)cpu_ctx->work_data; + + cplan.abort_callback = cpu_ctx->abort_callback; + cplan.abort_callback_data = cpu_ctx->abort_callback_data; + + return ggml_graph_compute(cgraph, &cplan); +} + +static const struct ggml_backend_i ggml_backend_cpu_i = { + /* .get_name = */ ggml_backend_cpu_get_name, + /* .free = */ ggml_backend_cpu_free, + /* .set_tensor_async = */ NULL, + /* .get_tensor_async = */ NULL, + /* .cpy_tensor_async = */ NULL, + /* .synchronize = */ NULL, + /* .graph_plan_create = */ ggml_backend_cpu_graph_plan_create, + /* .graph_plan_free = */ ggml_backend_cpu_graph_plan_free, + /* .graph_plan_update = */ NULL, + /* .graph_plan_compute = */ ggml_backend_cpu_graph_plan_compute, + /* .graph_compute = */ ggml_backend_cpu_graph_compute, + /* .event_record = */ NULL, + /* .event_wait = */ NULL, +}; + +static ggml_guid_t ggml_backend_cpu_guid(void) { + static ggml_guid guid = { 0xaa, 0x67, 0xc7, 0x43, 0x96, 0xe6, 0xa3, 0x8a, 0xe3, 0xaf, 0xea, 0x92, 0x36, 0xbc, 0xfc, 0x89 }; + return &guid; +} + +ggml_backend_t ggml_backend_cpu_init(void) { + // initialize CPU backend now to avoid slowing the first graph computation + ggml_cpu_init(); + + struct ggml_backend_cpu_context * ctx = new ggml_backend_cpu_context; + if (ctx == NULL) { + return NULL; + } + + ctx->n_threads = GGML_DEFAULT_N_THREADS; + ctx->threadpool = NULL; + ctx->work_data = NULL; + ctx->work_size = 0; + ctx->abort_callback = NULL; + ctx->abort_callback_data = NULL; + + ggml_backend_t cpu_backend = new ggml_backend { + /* .guid = */ ggml_backend_cpu_guid(), + /* .interface = */ ggml_backend_cpu_i, + /* .device = */ ggml_backend_reg_dev_get(ggml_backend_cpu_reg(), 0), + /* .context = */ ctx, + }; + + if (cpu_backend == NULL) { + delete ctx; + return NULL; + } + + return cpu_backend; +} + +bool ggml_backend_is_cpu(ggml_backend_t backend) { + return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_cpu_guid()); +} + +void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads) { + GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); + + struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; + ctx->n_threads = n_threads; +} + +void ggml_backend_cpu_set_threadpool(ggml_backend_t backend_cpu, ggml_threadpool_t threadpool) { + GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); + + struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; + + if (ctx->threadpool && ctx->threadpool != threadpool) { + // already had a different threadpool, pause/suspend it before switching + ggml_threadpool_pause(ctx->threadpool); + } + ctx->threadpool = threadpool; +} + +void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data) { + GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); + + struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; + ctx->abort_callback = abort_callback; + ctx->abort_callback_data = abort_callback_data; +} + +// CPU backend - device + +struct ggml_backend_cpu_device_context { + std::string description = "CPU"; + + ggml_backend_cpu_device_context() { +#ifdef __APPLE__ + size_t len = 0; + if (!sysctlbyname("machdep.cpu.brand_string", NULL, &len, NULL, 0)) { + description.resize(len); + sysctlbyname("machdep.cpu.brand_string", &description[0], &len, NULL, 0); // NOLINT + } +#elif defined(__linux__) + FILE * f = fopen("/proc/cpuinfo", "r"); + if (f) { + char buf[1024]; + while (fgets(buf, sizeof(buf), f)) { + if (strncmp(buf, "model name", 10) == 0) { + char * p = strchr(buf, ':'); + if (p) { + p++; + while (std::isspace(*p)) { + p++; + } + while (std::isspace(p[strlen(p) - 1])) { + p[strlen(p) - 1] = '\0'; + } + description = p; + break; + } + } + } + fclose(f); + } +#elif defined(_WIN32) + HKEY hKey; + if (RegOpenKeyEx(HKEY_LOCAL_MACHINE, + TEXT("HARDWARE\\DESCRIPTION\\System\\CentralProcessor\\0"), + 0, + KEY_READ, + &hKey) == ERROR_SUCCESS) { + DWORD cpu_brand_size = 0; + if (RegQueryValueExA(hKey, + TEXT("ProcessorNameString"), + NULL, + NULL, + NULL, + &cpu_brand_size) == ERROR_SUCCESS) { + description.resize(cpu_brand_size); + if (RegQueryValueExA(hKey, + TEXT("ProcessorNameString"), + NULL, + NULL, + (LPBYTE)&description[0], // NOLINT + &cpu_brand_size) == ERROR_SUCCESS) { + if (description.find('\0') != std::string::npos) { + description.resize(description.find('\0')); + } + } + } + RegCloseKey(hKey); + } +#endif + } +}; + +static const char * ggml_backend_cpu_device_get_name(ggml_backend_dev_t dev) { + return "CPU"; + + GGML_UNUSED(dev); +} + +static const char * ggml_backend_cpu_device_get_description(ggml_backend_dev_t dev) { + struct ggml_backend_cpu_device_context * ctx = (struct ggml_backend_cpu_device_context *)dev->context; + + return ctx->description.c_str(); +} + +static void ggml_backend_cpu_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { + // TODO + *free = 0; + *total = 0; + + GGML_UNUSED(dev); +} + +static enum ggml_backend_dev_type ggml_backend_cpu_device_get_type(ggml_backend_dev_t dev) { + return GGML_BACKEND_DEVICE_TYPE_CPU; + + GGML_UNUSED(dev); +} + +static void ggml_backend_cpu_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { + props->name = ggml_backend_cpu_device_get_name(dev); + props->description = ggml_backend_cpu_device_get_description(dev); + props->type = ggml_backend_cpu_device_get_type(dev); + ggml_backend_cpu_device_get_memory(dev, &props->memory_free, &props->memory_total); + props->caps = { + /* .async = */ false, + /* .host_buffer = */ false, + /* .buffer_from_host_ptr = */ true, + /* .events = */ false, + }; +} + +static ggml_backend_t ggml_backend_cpu_device_init_backend(ggml_backend_dev_t dev, const char * params) { + return ggml_backend_cpu_init(); + + GGML_UNUSED(dev); + GGML_UNUSED(params); +} + +static ggml_backend_buffer_type_t ggml_backend_cpu_device_get_buffer_type(ggml_backend_dev_t dev) { + return ggml_backend_cpu_buffer_type(); + + GGML_UNUSED(dev); +} + +static ggml_backend_buffer_t ggml_backend_cpu_device_buffer_from_host_ptr(ggml_backend_dev_t dev, void * ptr, size_t size, size_t max_tensor_size) { + return ggml_backend_cpu_buffer_from_ptr(ptr, size); + + GGML_UNUSED(dev); + GGML_UNUSED(max_tensor_size); +} + +static bool ggml_backend_cpu_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { + const struct ggml_tensor * src0 = op->src[0]; + const struct ggml_tensor * src1 = op->src[1]; + + if (op->op == GGML_OP_NONE || op->op == GGML_OP_RESHAPE || op->op == GGML_OP_VIEW || op->op == GGML_OP_PERMUTE || op->op == GGML_OP_TRANSPOSE) { + return true; + } + + // extra_buffer_op? + for (auto extra : ggml_backend_cpu_get_extra_buffers_type()) { + if (extra) { + auto buf_extra = (ggml::cpu::extra_buffer_type*) extra->context; + if (buf_extra && buf_extra->supports_op(dev, op)) { + return true; + } + } + } + + // the other case need host buffer. + for (int i = 0; i < GGML_MAX_SRC; i++) { + if (op->src[i] && op->src[i]->buffer && !ggml_backend_buft_is_host(op->src[i]->buffer->buft)) { + return false; + } + } + + switch (op->op) { + case GGML_OP_CPY: + return + op->type != GGML_TYPE_IQ3_XXS && + op->type != GGML_TYPE_IQ3_S && + op->type != GGML_TYPE_IQ2_XXS && + op->type != GGML_TYPE_IQ2_XS && + op->type != GGML_TYPE_IQ2_S && + op->type != GGML_TYPE_IQ1_S && + op->type != GGML_TYPE_IQ1_M; // missing type_traits.from_float + case GGML_OP_MUL_MAT: + return src1->type == GGML_TYPE_F32 || src1->type == ggml_get_type_traits_cpu(src0->type)->vec_dot_type; + case GGML_OP_IM2COL_BACK: + return src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32; + case GGML_OP_OUT_PROD: + return (src0->type == GGML_TYPE_F32 || ggml_is_quantized(src0->type)) && src1->type == GGML_TYPE_F32; + default: + return true; + } +} + +static bool ggml_backend_cpu_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { + return ggml_backend_buft_is_host(buft) || ggml_backend_cpu_is_extra_buffer_type(buft); + GGML_UNUSED(dev); +} + +static const struct ggml_backend_device_i ggml_backend_cpu_device_i = { + /* .get_name = */ ggml_backend_cpu_device_get_name, + /* .get_description = */ ggml_backend_cpu_device_get_description, + /* .get_memory = */ ggml_backend_cpu_device_get_memory, + /* .get_type = */ ggml_backend_cpu_device_get_type, + /* .get_props = */ ggml_backend_cpu_device_get_props, + /* .init_backend = */ ggml_backend_cpu_device_init_backend, + /* .get_buffer_type = */ ggml_backend_cpu_device_get_buffer_type, + /* .get_host_buffer_type = */ NULL, + /* .buffer_from_host_ptr = */ ggml_backend_cpu_device_buffer_from_host_ptr, + /* .supports_op = */ ggml_backend_cpu_device_supports_op, + /* .supports_buft = */ ggml_backend_cpu_device_supports_buft, + /* .offload_op = */ NULL, + /* .event_new = */ NULL, + /* .event_free = */ NULL, + /* .event_synchronize = */ NULL, +}; + +// CPU backend - backend (reg) + +static const char * ggml_backend_cpu_reg_get_name(ggml_backend_reg_t reg) { + return "CPU"; + + GGML_UNUSED(reg); +} + +static size_t ggml_backend_cpu_reg_get_device_count(ggml_backend_reg_t reg) { + return 1; + + GGML_UNUSED(reg); +} + +static ggml_backend_dev_t ggml_backend_cpu_reg_get_device(ggml_backend_reg_t reg, size_t index) { + GGML_ASSERT(index == 0); + + static ggml_backend_cpu_device_context ctx; + static ggml_backend_device ggml_backend_cpu_device = { + /* .iface = */ ggml_backend_cpu_device_i, + /* .reg = */ reg, + /* .context = */ &ctx, + }; + + return &ggml_backend_cpu_device; +} + +// This is intended to replace the the ggml_cpu_has_* functions when loading the CPU backend dynamically, +// and additionally to allow other backends to expose their own list of features that applications can query using the same API +static ggml_backend_feature * ggml_backend_cpu_get_features(ggml_backend_reg_t reg) { + static std::vector features = []() { + ggml_cpu_init(); + + std::vector features; + if (ggml_cpu_has_sse3()) { + features.push_back({ "SSE3", "1" }); + } + if (ggml_cpu_has_ssse3()) { + features.push_back({ "SSSE3", "1" }); + } + if (ggml_cpu_has_avx()) { + features.push_back({ "AVX", "1" }); + } + if (ggml_cpu_has_avx_vnni()) { + features.push_back({ "AVX_VNNI", "1" }); + } + if (ggml_cpu_has_avx2()) { + features.push_back({ "AVX2", "1" }); + } + if (ggml_cpu_has_f16c()) { + features.push_back({ "F16C", "1" }); + } + if (ggml_cpu_has_fma()) { + features.push_back({ "FMA", "1" }); + } + if (ggml_cpu_has_avx512()) { + features.push_back({ "AVX512", "1" }); + } + if (ggml_cpu_has_avx512_vbmi()) { + features.push_back({ "AVX512_VBMI", "1" }); + } + if (ggml_cpu_has_avx512_vnni()) { + features.push_back({ "AVX512_VNNI", "1" }); + } + if (ggml_cpu_has_avx512_bf16()) { + features.push_back({ "AVX512_BF16", "1" }); + } + if (ggml_cpu_has_amx_int8()) { + features.push_back({ "AMX_INT8", "1" }); + } + if (ggml_cpu_has_neon()) { + features.push_back({ "NEON", "1" }); + } + if (ggml_cpu_has_arm_fma()) { + features.push_back({ "ARM_FMA", "1" }); + } + if (ggml_cpu_has_fp16_va()) { + features.push_back({ "FP16_VA", "1" }); + } + if (ggml_cpu_has_matmul_int8()) { + features.push_back({ "MATMUL_INT8", "1" }); + } + if (ggml_cpu_has_sve()) { + features.push_back({ "SVE", "1" }); + } + if (ggml_cpu_has_dotprod()) { + features.push_back({ "DOTPROD", "1" }); + } + if (ggml_cpu_has_matmul_int8()) { + features.push_back({ "MATMUL_INT8", "1" }); + } + if (ggml_cpu_get_sve_cnt() > 0) { + static std::string sve_cnt = std::to_string(ggml_cpu_get_sve_cnt()); + features.push_back({ "SVE_CNT", sve_cnt.c_str() }); + } + if (ggml_cpu_has_riscv_v()) { + features.push_back({ "RISCV_V", "1" }); + } + if (ggml_cpu_has_vsx()) { + features.push_back({ "VSX", "1" }); + } + if (ggml_cpu_has_wasm_simd()) { + features.push_back({ "WASM_SIMD", "1" }); + } + if (ggml_cpu_has_llamafile()) { + features.push_back({ "LLAMAFILE", "1" }); + } + #ifdef GGML_USE_ACCELERATE + features.push_back({ "ACCELERATE", "1" }); + #endif + #ifdef GGML_USE_CPU_HBM + features.push_back({ "CPU_HBM", "1" }); + #endif + #ifdef GGML_USE_OPENMP + features.push_back({ "OPENMP", "1" }); + #endif + #ifdef GGML_USE_CPU_AARCH64 + features.push_back({ "AARCH64_REPACK", "1" }); + #endif + + features.push_back({ nullptr, nullptr }); + + return features; + }(); + + return features.data(); + + GGML_UNUSED(reg); +} + +static void * ggml_backend_cpu_get_proc_address(ggml_backend_reg_t reg, const char * name) { + if (strcmp(name, "ggml_backend_set_n_threads") == 0) { + ggml_backend_set_n_threads_t fct = ggml_backend_cpu_set_n_threads; + return (void *)fct; + } + if (strcmp(name, "ggml_backend_dev_get_extra_bufts") == 0) { + ggml_backend_dev_get_extra_bufts_t fct = ggml_backend_cpu_device_get_extra_buffers_type; + return (void *)fct; + } + if (strcmp(name, "ggml_backend_get_features") == 0) { + return (void *)ggml_backend_cpu_get_features; + } + if (strcmp(name, "ggml_backend_set_abort_callback") == 0) { + return (void *)ggml_backend_cpu_set_abort_callback; + } + if (strcmp(name, "ggml_backend_cpu_numa_init") == 0) { + return (void *)ggml_numa_init; + } + if (strcmp(name, "ggml_backend_cpu_is_numa") == 0) { + return (void *)ggml_is_numa; + } + + // threadpool - TODO: move to ggml-base + if (strcmp(name, "ggml_threadpool_new") == 0) { + return (void *)ggml_threadpool_new; + } + if (strcmp(name, "ggml_threadpool_free") == 0) { + return (void *)ggml_threadpool_free; + } + if (strcmp(name, "ggml_backend_cpu_set_threadpool") == 0) { + return (void *)ggml_backend_cpu_set_threadpool; + } + + return NULL; + + GGML_UNUSED(reg); +} + +static const struct ggml_backend_reg_i ggml_backend_cpu_reg_i = { + /* .get_name = */ ggml_backend_cpu_reg_get_name, + /* .get_device_count = */ ggml_backend_cpu_reg_get_device_count, + /* .get_device = */ ggml_backend_cpu_reg_get_device, + /* .get_proc_address = */ ggml_backend_cpu_get_proc_address, +}; + +ggml_backend_reg_t ggml_backend_cpu_reg(void) { + // init CPU feature detection + ggml_cpu_init(); + + static struct ggml_backend_reg ggml_backend_cpu_reg = { + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_cpu_reg_i, + /* .context = */ NULL, + }; + + return &ggml_backend_cpu_reg; +} + +GGML_BACKEND_DL_IMPL(ggml_backend_cpu_reg) diff --git a/ggml/src/ggml-cpu/llamafile/sgemm.cpp b/ggml/src/ggml-cpu/llamafile/sgemm.cpp new file mode 100644 index 000000000..c22a66287 --- /dev/null +++ b/ggml/src/ggml-cpu/llamafile/sgemm.cpp @@ -0,0 +1,2597 @@ +// Copyright 2024 Mozilla Foundation +// +// Permission is hereby granted, free of charge, to any person obtaining +// a copy of this software and associated documentation files (the +// "Software"), to deal in the Software without restriction, including +// without limitation the rights to use, copy, modify, merge, publish, +// distribute, sublicense, and/or sell copies of the Software, and to +// permit persons to whom the Software is furnished to do so, subject to +// the following conditions: +// +// The above copyright notice and this permission notice shall be +// included in all copies or substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, +// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS +// BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +// ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN +// CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +// SOFTWARE. + +// +// _ _ ___ _ _ ___ +// | |_(_)_ _ _ _| _ ) | /_\ / __| +// | _| | ' \ || | _ \ |__ / _ \\__ \. +// \__|_|_||_\_, |___/____/_/ \_\___/ +// |__/ +// +// BASIC LINEAR ALGEBRA SUBPROGRAMS +// +// +// This file implements multithreaded CPU matrix multiplication for the +// common contiguous use case C = Aᵀ * B. These kernels are designed to +// have excellent performance[1] for matrices that fit in the CPU cache +// without imposing any overhead such as cache filling or malloc calls. +// +// This implementation does not guarantee any upper bound with rounding +// errors, which grow along with k. Our goal's to maximally exploit the +// hardware for performance, and then use whatever resources remain for +// improving numerical accuracy. +// +// [1] J. Tunney, ‘LLaMA Now Goes Faster on CPUs’, Mar. 2024. [Online]. +// Available: https://justine.lol/matmul/. [Accessed: 29-Mar-2024]. + +#if defined(__GNUC__) +#pragma GCC diagnostic ignored "-Wpedantic" +#pragma GCC diagnostic ignored "-Wignored-attributes" +#endif + +#include "sgemm.h" +#include "ggml-impl.h" +#include "ggml-cpu-impl.h" +#include "ggml-quants.h" + +#include +#include + +#ifdef _MSC_VER +#define NOINLINE __declspec(noinline) +#else +#define NOINLINE __attribute__((__noinline__)) +#endif + +#if defined(__ARM_NEON) || defined(__AVX512F__) +#define VECTOR_REGISTERS 32 +#else +#define VECTOR_REGISTERS 16 +#endif + +#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1) + +namespace { + +inline float unhalf(ggml_fp16_t d) { + return GGML_FP16_TO_FP32(d); +} + +//////////////////////////////////////////////////////////////////////////////////////////////////// +// VECTORIZED ARITHMETIC OPERATIONS + +#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +inline __m128 add(__m128 x, __m128 y) { return _mm_add_ps(x, y); } +inline __m128 sub(__m128 x, __m128 y) { return _mm_sub_ps(x, y); } +inline __m128 mul(__m128 x, __m128 y) { return _mm_mul_ps(x, y); } +#endif // __SSE__ + +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +inline __m256 add(__m256 x, __m256 y) { return _mm256_add_ps(x, y); } +inline __m256 sub(__m256 x, __m256 y) { return _mm256_sub_ps(x, y); } +inline __m256 mul(__m256 x, __m256 y) { return _mm256_mul_ps(x, y); } +#endif // __AVX__ + +#if defined(__AVX512F__) +inline __m512 add(__m512 x, __m512 y) { return _mm512_add_ps(x, y); } +inline __m512 sub(__m512 x, __m512 y) { return _mm512_sub_ps(x, y); } +inline __m512 mul(__m512 x, __m512 y) { return _mm512_mul_ps(x, y); } +#endif // __AVX512F__ + +#if defined(__ARM_NEON) +inline float32x4_t add(float32x4_t x, float32x4_t y) { return vaddq_f32(x, y); } +inline float32x4_t sub(float32x4_t x, float32x4_t y) { return vsubq_f32(x, y); } +inline float32x4_t mul(float32x4_t x, float32x4_t y) { return vmulq_f32(x, y); } +#endif // __ARM_NEON + +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) +inline float16x8_t add(float16x8_t x, float16x8_t y) { return vaddq_f16(x, y); } +inline float16x8_t sub(float16x8_t x, float16x8_t y) { return vsubq_f16(x, y); } +inline float16x8_t mul(float16x8_t x, float16x8_t y) { return vmulq_f16(x, y); } +#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC + +#if defined(__MMA__) +typedef vector unsigned char vec_t; +typedef __vector_quad acc_t; +#endif +//////////////////////////////////////////////////////////////////////////////////////////////////// +// VECTORIZED FUSED MULTIPLY ADD + +/** + * Computes a * b + c. + */ +template +inline U madd(T a, T b, U c) { + return add(mul(a, b), c); +} + +#if defined(__FMA__) +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +template <> +inline __m256 madd(__m256 a, __m256 b, __m256 c) { + return _mm256_fmadd_ps(a, b, c); +} +#endif +#if defined(__AVX512F__) +template <> +inline __m512 madd(__m512 a, __m512 b, __m512 c) { + return _mm512_fmadd_ps(a, b, c); +} +#endif +#if defined(__AVX512BF16__) +template <> +inline __m512 madd(__m512bh a, __m512bh b, __m512 c) { + return _mm512_dpbf16_ps(c, a, b); +} +template <> +inline __m256 madd(__m256bh a, __m256bh b, __m256 c) { + return _mm256_dpbf16_ps(c, a, b); +} +#endif +#endif + +#if defined(__ARM_FEATURE_FMA) +template <> +inline float32x4_t madd(float32x4_t a, float32x4_t b, float32x4_t c) { + return vfmaq_f32(c, b, a); +} +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) +template <> +inline float16x8_t madd(float16x8_t a, float16x8_t b, float16x8_t c) { + return vfmaq_f16(c, b, a); +} +#endif +#endif + +//////////////////////////////////////////////////////////////////////////////////////////////////// +// VECTORIZED HORIZONTAL SUM + +#if defined(__ARM_NEON) +inline float hsum(float32x4_t x) { + return vaddvq_f32(x); +} +#endif // __ARM_NEON + +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) +inline float hsum(float16x8_t x) { + return vaddvq_f32(vaddq_f32(vcvt_f32_f16(vget_low_f16(x)), + vcvt_f32_f16(vget_high_f16(x)))); +} +#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC + +#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +inline float hsum(__m128 x) { +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) + x = _mm_add_ps(x, _mm_movehl_ps(x, x)); + x = _mm_add_ss(x, _mm_movehdup_ps(x)); +#else + __m128 t; + t = _mm_shuffle_ps(x, x, _MM_SHUFFLE(2, 3, 0, 1)); + x = _mm_add_ps(x, t); + t = _mm_movehl_ps(t, x); + x = _mm_add_ss(x, t); +#endif + return _mm_cvtss_f32(x); +} +#endif + +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +inline float hsum(__m256 x) { + return hsum(_mm_add_ps(_mm256_extractf128_ps(x, 1), + _mm256_castps256_ps128(x))); +} +#endif // __AVX__ + +#if defined(__AVX512F__) +inline float hsum(__m512 x) { + return _mm512_reduce_add_ps(x); +} +#endif // __AVX512F__ + +//////////////////////////////////////////////////////////////////////////////////////////////////// +// VECTORIZED MEMORY LOADING + +template T load(const U *); + +#if defined(__ARM_NEON) +template <> inline float32x4_t load(const float *p) { + return vld1q_f32(p); +} +#if !defined(_MSC_VER) +// FIXME: this should check for __ARM_FEATURE_FP16_VECTOR_ARITHMETIC +template <> inline float16x8_t load(const ggml_fp16_t *p) { + return vld1q_f16((const float16_t *)p); +} +template <> inline float32x4_t load(const ggml_fp16_t *p) { + return vcvt_f32_f16(vld1_f16((const float16_t *)p)); +} +#endif // _MSC_VER +#endif // __ARM_NEON + +#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +template <> inline __m128 load(const float *p) { + return _mm_loadu_ps(p); +} +#endif // __SSE__ + +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +template <> inline __m256 load(const float *p) { + return _mm256_loadu_ps(p); +} +#endif // __AVX__ + +#if defined(__AVX2__) || defined(__AVX512F__) +template <> inline __m256 load(const ggml_bf16_t *p) { + return _mm256_castsi256_ps( + _mm256_slli_epi32(_mm256_cvtepu16_epi32(_mm_loadu_si128((const __m128i *)p)), 16)); +} +#endif // __AVX2__ + +#if defined(__F16C__) +template <> inline __m256 load(const ggml_fp16_t *p) { + return _mm256_cvtph_ps(_mm_loadu_si128((const __m128i *)p)); +} +#endif // __F16C__ + +#if defined(__AVX512F__) +template <> inline __m512 load(const float *p) { + return _mm512_loadu_ps(p); +} +template <> inline __m512 load(const ggml_fp16_t *p) { + return _mm512_cvtph_ps(_mm256_loadu_si256((const __m256i *)p)); +} +template <> inline __m512 load(const ggml_bf16_t *p) { + return _mm512_castsi512_ps( + _mm512_slli_epi32(_mm512_cvtepu16_epi32(_mm256_loadu_si256((const __m256i *)p)), 16)); +} +#endif // __AVX512F__ + +#if defined(__AVX512BF16__) +template <> inline __m512bh load(const ggml_bf16_t *p) { + return (__m512bh)_mm512_loadu_ps((const float *)p); +} +template <> inline __m256bh load(const ggml_bf16_t *p) { + return (__m256bh)_mm256_loadu_ps((const float *)p); +} +template <> inline __m512bh load(const float *p) { + return _mm512_cvtne2ps_pbh(_mm512_loadu_ps(p + 16), _mm512_loadu_ps(p)); +} +template <> inline __m256bh load(const float *p) { + return _mm512_cvtneps_pbh(_mm512_loadu_ps(p)); +} +#endif + +//////////////////////////////////////////////////////////////////////////////////////////////////// +// CONSTANTS + +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) +static const int8_t kvalues_iq4nl[16] = {-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113}; +static const __m128i iq4nlt = _mm_loadu_si128((const __m128i *) kvalues_iq4nl); +#endif + +//////////////////////////////////////////////////////////////////////////////////////////////////// +// FLOATING POINT MATRIX MULTIPLICATION + +template +static inline int64_t BLOCK_SIZE(size_t m) { + const int64_t NB_BLOC_M = (m + M - 1) / M; + return (m % NB_BLOC_M == 0) ? m / NB_BLOC_M : (m / NB_BLOC_M) + 1; +} + +static constexpr inline int64_t BLOC_POS(int64_t ib, int64_t ibN, int64_t bloc_size) { + return ib < ibN ? ib * bloc_size : ibN * bloc_size + (ib - ibN) * (bloc_size - 1); +} + +template +class tinyBLAS { + public: + tinyBLAS(const ggml_compute_params * params, int64_t k, + const TA *A, int64_t lda, + const TB *B, int64_t ldb, + TC *C, int64_t ldc) + : params(params), A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc) { + } + + bool matmul(int64_t m, int64_t n) { + if (k % KN != 0) + return false; + // compute RM for only need tile with size RM&RM-1 +#if VECTOR_REGISTERS == 32 + if (m % 16 == 0 && (m/16 >= params->nth)) { + const int64_t SIZE_N = BLOCK_SIZE<6>(n); + mnpack<4, 6, 4>(m, n, SIZE_N, 12); + return true; + } + if (m % 8 == 0 ) { + const int64_t SIZE_N = BLOCK_SIZE<6>(n); + mnpack<4, 6, 2>(m, n, SIZE_N, 12); + return true; + } + if (m % 4 == 0) { + const int64_t SIZE_N = BLOCK_SIZE<6>(n); + mnpack<4, 6, 1>(m, n, SIZE_N, 12); + return true; + } +#else // VECTOR_REGISTERS == 16 + if (m % 16 == 0 && (m/16 >= params->nth)) { + const int64_t SIZE_N = BLOCK_SIZE<3>(n); + mnpack<4, 3, 4>(m, n, SIZE_N, 24); + return true; + } + if (m % 8 == 0 ) { + const int64_t SIZE_N = BLOCK_SIZE<3>(n); + mnpack<4, 3, 2>(m, n, SIZE_N, 24); + return true; + } + if (m % 4 == 0) { + const int64_t SIZE_N = BLOCK_SIZE<3>(n); + mnpack<4, 3, 1>(m, n, SIZE_N, 24); + return true; + } +#endif + return false; + } + + private: + template + inline void mnpack(int64_t m, int64_t n, int64_t SIZE_N, int64_t BN) { + if (SIZE_N == RN) { + return gemm(m, n, BN); + } + if constexpr (RN > 1) { + return mnpack(m, n, SIZE_N, BN); + } else { + GGML_LOG_ERROR("mnpack<%d, %d> bloc size not supported\n", RM, (int)SIZE_N); + GGML_ASSERT(false); // we have miss something. + } + } + + template + inline void gemm_bloc(int64_t ii, int64_t jj) { + D Cv[RN][RM] = {}; + for (int64_t l = 0; l < k; l += KN) { + // help compiler for op order. + if constexpr (RM <= RN) { + V Av[RM]; + for (int64_t i = 0; i < RM; ++i) { + Av[i] = load(A + lda * (ii + i) + l); + } + for (int64_t j = 0; j < RN; ++j) { + V Bv = load(B + ldb * (jj + j) + l); + for (int64_t i = 0; i < RM; ++i) { + Cv[j][i] = madd(Av[i], Bv, Cv[j][i]); + } + } + } else { + V Bv[RN]; + for (int64_t j = 0; j < RN; ++j) { + Bv[j] = load(B + ldb * (jj + j) + l); + } + for (int64_t i = 0; i < RM; ++i) { + V Av = load(A + lda * (ii + i) + l); + for (int64_t j = 0; j < RN; ++j) { + Cv[j][i] = madd(Av, Bv[j], Cv[j][i]); + } + } + } + } + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < RM; ++i) + C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); + } + + template + NOINLINE void gemm(int64_t m, int64_t n, int64_t BN) { + static std::atomic current_chunk; + + GGML_ASSERT(m % (RM * BM) == 0); + const int64_t ytiles = m / (RM * BM); + const int64_t xtiles = (n + RN -1) / RN; + const int64_t jj_RN = (xtiles - (xtiles * RN - n)); + + // "round" bloc_size to "nearest" BN + const int64_t NB_BN = xtiles < BN ? 1 : (xtiles + BN / 2) / BN; + const int64_t SIZE_BN = xtiles % NB_BN == 0 ? xtiles / NB_BN : xtiles / NB_BN + 1; + const int64_t jj_BN = (NB_BN - (NB_BN * SIZE_BN - xtiles)); + const int64_t nb_job = ytiles * NB_BN; + + if (params->ith == 0) { + GGML_ASSERT( jj_BN * SIZE_BN + (NB_BN - jj_BN) * (SIZE_BN - 1) == xtiles); + // Every thread starts at ith, so the first unprocessed chunk is nth. This save a bit of coordination right at the start. + std::atomic_store_explicit(¤t_chunk, (int64_t)params->nth, std::memory_order_relaxed); + } + + ggml_barrier(params->threadpool); + + int64_t job = params->ith; + while (job < nb_job) { + const int64_t ii = (job % ytiles) * RM * BM; + const int64_t jb = job / ytiles; + const int64_t jr0 = BLOC_POS(jb , jj_BN, SIZE_BN); + const int64_t jrN = BLOC_POS(jb+1, jj_BN, SIZE_BN); + + const int64_t jj0 = BLOC_POS(jr0, jj_RN, RN); + const int64_t jj2 = BLOC_POS(jrN, jj_RN, RN); + const int64_t jj1 = jj2 < jj_RN * RN ? jj2 : jj_RN * RN; + + for (int64_t bi = 0; bi < BM * RM; bi += RM) { + int64_t jj = jj0; + for (; jj < jj1; jj += RN) { + gemm_bloc(ii + bi, jj); + } + if constexpr (RN > 1) { + for (; jj < jj2; jj += RN - 1) { + gemm_bloc(ii + bi, jj); + } + } + GGML_ASSERT(jj == jj2); + } + + // next step. + job = std::atomic_fetch_add_explicit(¤t_chunk, (int64_t)1, std::memory_order_relaxed); + } + + ggml_barrier(params->threadpool); + return; + } + + const ggml_compute_params * params; + const TA *const A; + const TB *const B; + TC *const C; + const int64_t k; + const int64_t lda; + const int64_t ldb; + const int64_t ldc; +}; + +////////////////////////////////////////////////////////////////////////////////////////// +// QUANT ZERO MATRIX MULTIPLICATION + +#if defined(__ARM_FEATURE_DOTPROD) +template +class tinyBLAS_Q0_ARM { + public: + tinyBLAS_Q0_ARM(int64_t k, + const TA *A, int64_t lda, + const block_q8_0 *B, int64_t ldb, + float *C, int64_t ldc, + int ith, int nth) + : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { + } + + void matmul(int64_t m, int64_t n) { + mnpack(0, m, 0, n); + } + + private: + NOINLINE void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t mc, nc, mp, np; + switch ((MIN(m - m0, 3) << 4) | MIN(n - n0, 3ll)) { + case 0x33: + mc = 3; + nc = 3; + gemm<3, 3>(m0, m, n0, n); + break; + case 0x32: + mc = 3; + nc = 2; + gemm<3, 2>(m0, m, n0, n); + break; + case 0x23: + mc = 2; + nc = 3; + gemm<2, 3>(m0, m, n0, n); + break; + case 0x22: + mc = 2; + nc = 2; + gemm<2, 2>(m0, m, n0, n); + break; + case 0x31: + mc = 3; + nc = 1; + gemm<3, 1>(m0, m, n0, n); + break; + case 0x13: + mc = 1; + nc = 3; + gemm<1, 3>(m0, m, n0, n); + break; + case 0x21: + mc = 2; + nc = 1; + gemm<2, 1>(m0, m, n0, n); + break; + case 0x12: + mc = 1; + nc = 2; + gemm<1, 2>(m0, m, n0, n); + break; + case 0x11: + mc = 1; + nc = 1; + gemm<1, 1>(m0, m, n0, n); + break; + default: + return; + } + mp = m0 + (m - m0) / mc * mc; + np = n0 + (n - n0) / nc * nc; + mnpack(mp, m, n0, np); + mnpack(m0, m, np, n); + } + + template + NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + float32x4_t Cv[RN][RM] = {}; + for (int64_t l = 0; l < k; ++l) + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < RM; ++i) + Cv[j][i] = vmlaq_n_f32(Cv[j][i], + vcvtq_f32_s32(vdotq_s32( + vdotq_s32(vdupq_n_s32(0), + load_lo(A + lda * (ii + i) + l), + load_lo(B + ldb * (jj + j) + l)), + load_hi(A + lda * (ii + i) + l), + load_hi(B + ldb * (jj + j) + l))), + unhalf(A[lda * (ii + i) + l].d) * + unhalf(B[ldb * (jj + j) + l].d)); + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < RM; ++i) + C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); + } + } + + inline int8x16_t load_lo(const block_q8_0 *b) { + return vld1q_s8(b->qs); + } + + inline int8x16_t load_hi(const block_q8_0 *b) { + return vld1q_s8(b->qs + 16); + } + + inline int8x16_t load_lo(const block_q4_0 *b) { + return vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vld1q_u8(b->qs), + vdupq_n_u8(0x0f))), + vdupq_n_s8(0x8)); + } + + inline int8x16_t load_hi(const block_q4_0 *b) { + return vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(vld1q_u8(b->qs), 4)), + vdupq_n_s8(0x8)); + } + + const TA *const A; + const block_q8_0 *const B; + float *const C; + const int64_t k; + const int64_t lda; + const int64_t ldb; + const int64_t ldc; + const int ith; + const int nth; +}; +#endif // __ARM_FEATURE_DOTPROD + +#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) +template +class tinyBLAS_Q0_AVX { + public: + tinyBLAS_Q0_AVX(int64_t k, + const TA *A, int64_t lda, + const TB *B, int64_t ldb, + TC *C, int64_t ldc, + int ith, int nth) + : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { + } + + void matmul(int64_t m, int64_t n) { + mnpack(0, m, 0, n); + } + + private: + void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t mc, nc, mp, np; + switch ((MIN(m - m0, 4) << 4) | MIN(n - n0, 4)) { +#if VECTOR_REGISTERS == 32 + case 0x44: + mc = 4; + nc = 4; +#if defined(__AVX2__) && defined(__F16C__) + gemm4xN<4>(m0, m, n0, n); +#else + gemm<4, 4>(m0, m, n0, n); +#endif + break; + case 0x43: + mc = 4; + nc = 3; +#if defined(__AVX2__) && defined(__F16C__) + gemm4xN<3>(m0, m, n0, n); +#else + gemm<4, 3>(m0, m, n0, n); +#endif + break; + case 0x34: + mc = 3; + nc = 4; +#if defined(__AVX2__) && defined(__F16C__) + gemmMx4<3>(m0, m, n0, n); +#else + gemm<3, 4>(m0, m, n0, n); +#endif + break; + case 0x33: + mc = 3; + nc = 3; + gemm<3, 3>(m0, m, n0, n); + break; + case 0x42: + mc = 4; + nc = 2; +#if defined(__AVX2__) && defined(__F16C__) + gemm4xN<2>(m0, m, n0, n); +#else + gemm<4, 2>(m0, m, n0, n); +#endif + break; + case 0x24: + mc = 2; + nc = 4; +#if defined(__AVX2__) && defined(__F16C__) + gemmMx4<2>(m0, m, n0, n); +#else + gemm<2, 4>(m0, m, n0, n); +#endif + break; +#else + case 0x44: + case 0x43: + case 0x42: + mc = 4; + nc = 2; +#if defined(__AVX2__) && defined(__F16C__) + gemm4xN<2>(m0, m, n0, n); +#else + gemm<4, 2>(m0, m, n0, n); +#endif + break; + case 0x34: + case 0x24: + mc = 2; + nc = 4; +#if defined(__AVX2__) && defined(__F16C__) + gemmMx4<2>(m0, m, n0, n); +#else + gemm<2, 4>(m0, m, n0, n); +#endif + break; + case 0x33: +#endif + case 0x32: + mc = 3; + nc = 2; + gemm<3, 2>(m0, m, n0, n); + break; + case 0x23: + mc = 2; + nc = 3; + gemm<2, 3>(m0, m, n0, n); + break; + case 0x41: + mc = 4; + nc = 1; +#if defined(__AVX2__) && defined(__F16C__) + gemm4xN<1>(m0, m, n0, n); +#else + gemm<4, 1>(m0, m, n0, n); +#endif + break; + case 0x22: + mc = 2; + nc = 2; + gemm<2, 2>(m0, m, n0, n); + break; + case 0x14: + mc = 1; + nc = 4; +#if defined(__AVX2__) && defined(__F16C__) + gemmMx4<1>(m0, m, n0, n); +#else + gemm<1, 4>(m0, m, n0, n); +#endif + break; + case 0x31: + mc = 3; + nc = 1; + gemm<3, 1>(m0, m, n0, n); + break; + case 0x13: + mc = 1; + nc = 3; + gemm<1, 3>(m0, m, n0, n); + break; + case 0x21: + mc = 2; + nc = 1; + gemm<2, 1>(m0, m, n0, n); + break; + case 0x12: + mc = 1; + nc = 2; + gemm<1, 2>(m0, m, n0, n); + break; + case 0x11: + mc = 1; + nc = 1; + gemm<1, 1>(m0, m, n0, n); + break; + default: + return; + } + mp = m0 + (m - m0) / mc * mc; + np = n0 + (n - n0) / nc * nc; + mnpack(mp, m, n0, np); + mnpack(m0, m, np, n); + } + +#if defined(__AVX2__) && defined(__F16C__) +// Templated functions for gemm of dimensions 4xN + template + NOINLINE void gemm4xN(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / 4; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * 4; + int64_t jj = n0 + job % xtiles * RN; + __m256 Cv[RN][4] = {}; + for (int64_t l = 0; l < k; ++l) { + uint64_t a_delta = ((uint64_t)A[lda * (ii + 3) + l].d << 48) | ((uint64_t)A[lda * (ii + 2) + l].d << 32) | ((uint64_t)A[lda * (ii + 1) + l].d << 16) | (A[lda * (ii + 0) + l].d); + // Convert delta values for four blocks to float values + __m128 da = _mm_cvtph_ps(_mm_set_epi64x(0, a_delta)); + __m256i avec0 = load(A + lda * (ii + 0) + l); + __m256i avec1 = load(A + lda * (ii + 1) + l); + __m256i avec2 = load(A + lda * (ii + 2) + l); + __m256i avec3 = load(A + lda * (ii + 3) + l); + for (int64_t j = 0; j < RN; ++j) { + __m128 db = _mm_set1_ps(unhalf(B[ldb * (jj + j) + l].d)); + // Computation of product of delta values for four blocks and replicate it across 256 bit lane + __m256 dvec = _mm256_castps128_ps256(_mm_mul_ps(da, db)); + dvec = _mm256_permute2f128_ps(dvec ,dvec, 0); + // Computation of dot product and multiplication with appropriate delta value products + Cv[j][0] = madd(_mm256_shuffle_ps(dvec, dvec, 0), + updot(_mm256_sign_epi8(avec0, avec0), + _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec0)), + Cv[j][0]); + Cv[j][1] = madd(_mm256_shuffle_ps(dvec, dvec, 85), + updot(_mm256_sign_epi8(avec1, avec1), + _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec1)), + Cv[j][1]); + Cv[j][2] = madd(_mm256_shuffle_ps(dvec, dvec, 170), + updot(_mm256_sign_epi8(avec2, avec2), + _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec2)), + Cv[j][2]); + Cv[j][3] = madd(_mm256_shuffle_ps(dvec, dvec, 255), + updot(_mm256_sign_epi8(avec3, avec3), + _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec3)), + Cv[j][3]); + } + } + + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < 4; ++i) + C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); + } + } + + // Templated functions for gemm of dimensions Mx4 + template + NOINLINE void gemmMx4(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / 4; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * 4; + __m256 Cv[4][RM] = {}; + for (int64_t l = 0; l < k; ++l) { + uint64_t b_delta = ((uint64_t)B[ldb * (jj + 3) + l].d << 48) | ((uint64_t)B[ldb * (jj + 2) + l].d << 32) | ((uint64_t)B[ldb * (jj + 1) + l].d << 16) | (B[ldb * (jj + 0) + l].d); + // Convert delta values for four blocks to float values + __m128 db = _mm_cvtph_ps(_mm_set_epi64x(0, b_delta)); + __m256i bvec0 = load(B + ldb * (jj + 0) + l); + __m256i bvec1 = load(B + ldb * (jj + 1) + l); + __m256i bvec2 = load(B + ldb * (jj + 2) + l); + __m256i bvec3 = load(B + ldb * (jj + 3) + l); + for (int64_t i = 0; i < RM; ++i) { + __m128 da = _mm_set1_ps(unhalf((A[lda * (ii + i) + l].d))); + // Computation of product of delta values for four blocks and replicate it across 256 bit lane + __m256 dvec = _mm256_castps128_ps256(_mm_mul_ps(da, db)); + dvec = _mm256_permute2f128_ps(dvec ,dvec, 0); + // Computation of dot product and multiplication with appropriate delta value products + Cv[0][i] = madd(_mm256_shuffle_ps(dvec, dvec, 0), + updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), + load(A + lda * (ii + i) + l)), + _mm256_sign_epi8(bvec0, load(A + lda * (ii + i) + l))), + Cv[0][i]); + Cv[1][i] = madd(_mm256_shuffle_ps(dvec, dvec, 85), + updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), + load(A + lda * (ii + i) + l)), + _mm256_sign_epi8(bvec1, load(A + lda * (ii + i) + l))), + Cv[1][i]); + Cv[2][i] = madd(_mm256_shuffle_ps(dvec, dvec, 170), + updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), + load(A + lda * (ii + i) + l)), + _mm256_sign_epi8(bvec2, load(A + lda * (ii + i) + l))), + Cv[2][i]); + Cv[3][i] = madd(_mm256_shuffle_ps(dvec, dvec, 255), + updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), + load(A + lda * (ii + i) + l)), + _mm256_sign_epi8(bvec3, load(A + lda * (ii + i) + l))), + Cv[3][i]); + } + } + for (int64_t j = 0; j < 4; ++j) + for (int64_t i = 0; i < RM; ++i) + C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); + } + } +#endif + + template + NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + __m256 Cv[RN][RM] = {}; + for (int64_t l = 0; l < k; ++l) + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < RM; ++i) { +#if defined(__AVX2__) + __m256 udTmp = updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), + load(A + lda * (ii + i) + l)), + _mm256_sign_epi8(load(B + ldb * (jj + j) + l), + load(A + lda * (ii + i) + l))); +#else + __m128i ali0 = load0(A + lda * (ii + i) + l); + __m128i ali1 = load1(A + lda * (ii + i) + l); + __m128i blj0 = load0(B + ldb * (jj + j) + l); + __m128i blj1 = load1(B + ldb * (jj + j) + l); + + __m128i sepAA0 = _mm_sign_epi8(ali0, ali0); + __m128i sepAA1 = _mm_sign_epi8(ali1, ali1); + __m128i sepBA0 = _mm_sign_epi8(blj0, ali0); + __m128i sepBA1 = _mm_sign_epi8(blj1, ali1); + + // updot + const __m128i oneFill = _mm_set1_epi16(1); + __m128i mad0 = _mm_maddubs_epi16(sepAA0, sepBA0); + __m128i mad1 = _mm_maddubs_epi16(sepAA1, sepBA1); + __m256 udTmp = _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_madd_epi16(oneFill, mad1), _mm_madd_epi16(oneFill, mad0))); +#endif + Cv[j][i] = madd(_mm256_set1_ps(unhalf(A[lda * (ii + i) + l].d) * + unhalf(B[ldb * (jj + j) + l].d)), + udTmp, + Cv[j][i]); + } + for (int64_t j = 0; j < RN; ++j) + for (int64_t i = 0; i < RM; ++i) + C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); + } + } + + inline __m256i load(const block_q8_0 *b) { + return _mm256_loadu_si256((const __m256i *)b->qs); + } + + inline __m128i load0(const block_q8_0 *b) { + return _mm_loadu_si128((const __m128i *)b->qs); + } + + inline __m128i load1(const block_q8_0 *b) { + return _mm_loadu_si128(((const __m128i *)b->qs) + 1); + } + + inline __m256i load(const block_q4_0 *b) { + return _mm256_sub_epi8(denibble(b->qs), _mm256_set1_epi8(8)); + } + + inline __m128i load0(const block_q4_0 *b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + return _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), x), _mm_set1_epi8(8)); + } + + inline __m128i load1(const block_q4_0 *b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + return _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4)), _mm_set1_epi8(8)); + } + + inline __m256i load(const block_q5_0 *b) { + return _mm256_or_si256(denibble(b->qs), bittobyte(b->qh)); + } + + inline __m128i load0(const block_q5_0* b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + uint32_t x32; + memcpy(&x32, b->qh, sizeof(uint32_t)); + __m128i qxl = _mm_and_si128(_mm_set1_epi8(15), x); + __m128i bytesl = _mm_cmpeq_epi8(_mm_set1_epi64x(-1), + _mm_or_si128(_mm_set1_epi64x(0x7fbfdfeff7fbfdfe), + _mm_shuffle_epi8(_mm_set1_epi32(x32), + _mm_set_epi64x(0x0101010101010101, 0x0000000000000000)))); + bytesl = _mm_andnot_si128(bytesl, _mm_set1_epi8((char)0xF0)); + return _mm_or_si128(qxl, bytesl); + } + + inline __m128i load1(const block_q5_0* b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + uint32_t x32; + memcpy(&x32, b->qh, sizeof(uint32_t)); + __m128i qxh = _mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4)); + __m128i bytesh = _mm_cmpeq_epi8(_mm_set1_epi64x(-1), + _mm_or_si128(_mm_set1_epi64x(0x7fbfdfeff7fbfdfe), + _mm_shuffle_epi8(_mm_set1_epi32(x32), + _mm_set_epi64x(0x0303030303030303, 0x0202020202020202)))); + bytesh = _mm_andnot_si128(bytesh, _mm_set1_epi8((char)0xF0)); + return _mm_or_si128(qxh, bytesh); + } + + inline __m256i load(const block_iq4_nl *b) { + return MM256_SET_M128I(load1(b), load0(b)); + } + + inline __m128i load0(const block_iq4_nl *b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + return _mm_shuffle_epi8(iq4nlt, _mm_and_si128(_mm_set1_epi8(15), x)); + } + + inline __m128i load1(const block_iq4_nl *b) { + const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); + return _mm_shuffle_epi8(iq4nlt, _mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4))); + } + + inline __m256 updot(__m256i u, __m256i s) { + __m256i res; +#if defined(__AVX512VNNI__) && defined(__AVX512VL__) + res = _mm256_dpbusd_epi32(_mm256_setzero_si256(), u, s); +#elif defined(__AVXVNNI__) + res = _mm256_dpbusd_avx_epi32(_mm256_setzero_si256(), u, s); +#else + res = _mm256_madd_epi16(_mm256_set1_epi16(1), _mm256_maddubs_epi16(u, s)); +#endif + return _mm256_cvtepi32_ps(res); + } + + static inline __m256i denibble(const uint8_t *p) { + __m128i x = _mm_loadu_si128((const __m128i *)p); + return _mm256_and_si256(_mm256_set1_epi8(15), + _mm256_insertf128_si256(_mm256_castsi128_si256(x), + _mm_srli_epi16(x, 4), 1)); + } + + static inline __m256i bittobyte(const uint8_t *p) { + uint32_t x32; + memcpy(&x32, p, sizeof(uint32_t)); + __m256i bytes = _mm256_cmpeq_epi8(_mm256_set1_epi64x(-1), + _mm256_or_si256(_mm256_set1_epi64x(0x7fbfdfeff7fbfdfe), + _mm256_shuffle_epi8(_mm256_set1_epi32(x32), + _mm256_set_epi64x(0x0303030303030303, 0x0202020202020202, + 0x0101010101010101, 0x0000000000000000)))); + return _mm256_andnot_si256(bytes, _mm256_set1_epi8((char)0xF0)); + } + + const TA *const A; + const TB *const B; + TC *const C; + const int64_t k; + const int64_t lda; + const int64_t ldb; + const int64_t ldc; + const int ith; + const int nth; +}; +#endif // __AVX__ + +//PPC Implementation +#if defined(__MMA__) + +#define SAVE_ACC(ACC, ii, jj) \ + __builtin_mma_disassemble_acc(vec_C, ACC); \ + for (int I = 0; I < 4; I++) { \ + for (int J = 0; J < 4; J++) { \ + *((float*)(C+ii+((jj+J)*ldc)+I)) = *((float*)&vec_C[I]+J); \ + } \ + } \ + +template +class tinyBLAS_Q0_PPC { + public: + tinyBLAS_Q0_PPC(int64_t k, + const TA *A, int64_t lda, + const TB *B, int64_t ldb, + TC *C, int64_t ldc, + int ith, int nth) + : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { + } + + void matmul(int64_t m, int64_t n) { + mnpack(0, m, 0, n); + } + + private: + + template + inline void save_res(int ii, int jj, int idx, vector float* fin_res) { + for (int I = 0; I < RM; I++) { + for (int J = 0; J < RN; J++) { + *((float*)(C+ii+((jj+J)*ldc)+I)) = *((float*)&fin_res[idx+I]+J); + } + } + } + + template + inline void compute(acc_t* ACC, int c_idx, int s_idx, std::array& comparray, vector float* vs, vector float* fin_res) { + vector signed int vec_C[4]; + vector float CA[4] = {0}; + vector float res[4] = {0}; + __builtin_mma_disassemble_acc(vec_C, ACC); + for (int i = 0; i < 4; i++) { + CA[i] = vec_splats((float)(((double)comparray[c_idx+i]) * -128.0)); + res[i] = vec_add(vec_ctf(vec_C[i], 0), CA[i]); + fin_res[s_idx+i] = vec_madd(res[i], vs[s_idx+i], fin_res[s_idx+i]); + } + } + + template + void packNormal(const TA* a, int64_t lda, int rows, int cols, VA* vec, bool flip) { + int64_t i, j; + TA *aoffset = NULL; + VA *vecOffset = NULL; + TA *aoffset1 = NULL, *aoffset2 = NULL, *aoffset3 = NULL, *aoffset4 = NULL; + TA *aoffset5 = NULL, *aoffset6 = NULL, *aoffset7 = NULL, *aoffset8 = NULL; + __vector_pair C1, C2, C3, C4, C5, C6, C7, C8; + VB c1[2] = {0}, c2[2] = {0}, c3[2] = {0}, c4[2]={0}; + VB c5[2] = {0}, c6[2] = {0}, c7[2] = {0}, c8[2]={0}; + VB t1, t2, t3, t4, t5, t6, t7, t8; + vector unsigned char xor_vector; + uint8_t flip_vec = 0x80; + xor_vector = vec_splats(flip_vec); + vector unsigned char swiz1 = {0, 1, 2, 3, 4, 5, 6, 7, 16, 17, 18, 19, 20, 21, 22, 23}; + vector unsigned char swiz2 = {8, 9, 10, 11, 12, 13, 14, 15, 24, 25, 26, 27, 28, 29, 30, 31}; + vector unsigned char swiz3 = {0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27}; + vector unsigned char swiz4 = {4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31}; + + aoffset = const_cast(a); + vecOffset = vec; + j = (rows >> 3); + if (j > 0) { + do { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + aoffset4 = aoffset3 + lda; + aoffset5 = aoffset4 + lda; + aoffset6 = aoffset5 + lda; + aoffset7 = aoffset6 + lda; + aoffset8 = aoffset7 + lda; + aoffset += 8 * lda; + + i = (cols >> 3); + if (i > 0) { + do { + C1 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset1->qs); + C2 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset2->qs); + C3 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset3->qs); + C4 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset4->qs); + C5 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset5->qs); + C6 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset6->qs); + C7 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset7->qs); + C8 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset8->qs); + + __builtin_vsx_disassemble_pair(c1, &C1); + __builtin_vsx_disassemble_pair(c2, &C2); + __builtin_vsx_disassemble_pair(c3, &C3); + __builtin_vsx_disassemble_pair(c4, &C4); + __builtin_vsx_disassemble_pair(c5, &C5); + __builtin_vsx_disassemble_pair(c6, &C6); + __builtin_vsx_disassemble_pair(c7, &C7); + __builtin_vsx_disassemble_pair(c8, &C8); + + t1 = vec_perm(c1[0], c2[0], swiz1); + t2 = vec_perm(c1[0], c2[0], swiz2); + t3 = vec_perm(c3[0], c4[0], swiz1); + t4 = vec_perm(c3[0], c4[0], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset); + vec_xst(t6, 0, vecOffset+16); + vec_xst(t7, 0, vecOffset+32); + vec_xst(t8, 0, vecOffset+48); + + t1 = vec_perm(c1[1], c2[1], swiz1); + t2 = vec_perm(c1[1], c2[1], swiz2); + t3 = vec_perm(c3[1], c4[1], swiz1); + t4 = vec_perm(c3[1], c4[1], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset+64); + vec_xst(t6, 0, vecOffset+80); + vec_xst(t7, 0, vecOffset+96); + vec_xst(t8, 0, vecOffset+112); + + t1 = vec_perm(c5[0], c6[0], swiz1); + t2 = vec_perm(c5[0], c6[0], swiz2); + t3 = vec_perm(c7[0], c8[0], swiz1); + t4 = vec_perm(c7[0], c8[0], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset+128); + vec_xst(t6, 0, vecOffset+144); + vec_xst(t7, 0, vecOffset+160); + vec_xst(t8, 0, vecOffset+176); + + t1 = vec_perm(c5[1], c6[1], swiz1); + t2 = vec_perm(c5[1], c6[1], swiz2); + t3 = vec_perm(c7[1], c8[1], swiz1); + t4 = vec_perm(c7[1], c8[1], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset+192); + vec_xst(t6, 0, vecOffset+208); + vec_xst(t7, 0, vecOffset+224); + vec_xst(t8, 0, vecOffset+240); + + aoffset1 += lda; + aoffset2 += lda; + aoffset3 += lda; + aoffset4 += lda; + aoffset5 += lda; + aoffset6 += lda; + aoffset7 += lda; + aoffset8 += lda; + vecOffset += 256; + i--; + } while(i > 0); + } + j--; + } while(j > 0); + } + + if (rows & 4) { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + aoffset4 = aoffset3 + lda; + aoffset += 4 * lda; + + i = (cols >> 3); + if (i > 0) { + do { + C1 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset1->qs); + C2 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset2->qs); + C3 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset3->qs); + C4 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset4->qs); + + __builtin_vsx_disassemble_pair(c1, &C1); + __builtin_vsx_disassemble_pair(c2, &C2); + __builtin_vsx_disassemble_pair(c3, &C3); + __builtin_vsx_disassemble_pair(c4, &C4); + + t1 = vec_perm(c1[0], c2[0], swiz1); + t2 = vec_perm(c1[0], c2[0], swiz2); + t3 = vec_perm(c3[0], c4[0], swiz1); + t4 = vec_perm(c3[0], c4[0], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset); + vec_xst(t6, 0, vecOffset+16); + vec_xst(t7, 0, vecOffset+32); + vec_xst(t8, 0, vecOffset+48); + + t1 = vec_perm(c1[1], c2[1], swiz1); + t2 = vec_perm(c1[1], c2[1], swiz2); + t3 = vec_perm(c3[1], c4[1], swiz1); + t4 = vec_perm(c3[1], c4[1], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset+64); + vec_xst(t6, 0, vecOffset+80); + vec_xst(t7, 0, vecOffset+96); + vec_xst(t8, 0, vecOffset+112); + + aoffset1 += lda; + aoffset2 += lda; + aoffset3 += lda; + aoffset4 += lda; + vecOffset += 128; + i--; + } while(i > 0); + } + } + if (rows & 3) { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + i = (cols >> 3); + if (i > 0) { + do { + switch(rows) { + case 3: C3 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset3->qs); + __builtin_vsx_disassemble_pair(c3, &C3); + case 2: C2 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset2->qs); + __builtin_vsx_disassemble_pair(c2, &C2); + case 1: C1 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset1->qs); + __builtin_vsx_disassemble_pair(c1, &C1); + break; + } + t1 = vec_perm(c1[0], c2[0], swiz1); + t2 = vec_perm(c1[0], c2[0], swiz2); + t3 = vec_perm(c3[0], c4[0], swiz1); + t4 = vec_perm(c3[0], c4[0], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset); + vec_xst(t6, 0, vecOffset+16); + vec_xst(t7, 0, vecOffset+32); + vec_xst(t8, 0, vecOffset+48); + + t1 = vec_perm(c1[1], c2[1], swiz1); + t2 = vec_perm(c1[1], c2[1], swiz2); + t3 = vec_perm(c3[1], c4[1], swiz1); + t4 = vec_perm(c3[1], c4[1], swiz2); + t5 = vec_perm(t1, t3, swiz3); + t6 = vec_perm(t1, t3, swiz4); + t7 = vec_perm(t2, t4, swiz3); + t8 = vec_perm(t2, t4, swiz4); + if (flip == true) { + t5 = vec_xor(t5, xor_vector); + t6 = vec_xor(t6, xor_vector); + t7 = vec_xor(t7, xor_vector); + t8 = vec_xor(t8, xor_vector); + } + vec_xst(t5, 0, vecOffset+64); + vec_xst(t6, 0, vecOffset+80); + vec_xst(t7, 0, vecOffset+96); + vec_xst(t8, 0, vecOffset+112); + + aoffset1 += lda; + aoffset2 += lda; + aoffset3 += lda; + vecOffset += 128; + i--; + } while(i > 0); + } + } + } + + void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t mc, nc, mp, np; + int m_rem = MIN(m - m0, 8); + int n_rem = MIN(n - n0, 8); + // TO-DO: KERNEL_16x8 and KERNEL_8x16 are having some performance + // issues. After resolving them, below code will be enabled. + /*if (m_rem >= 16 && n_rem >= 8) { + mc = 16; + nc = 8; + gemm<16,8>(m0, m, n0, n); + } else if(m_rem >= 8 && n_rem >= 16) { + mc = 8; + nc = 16; + gemm<8,16>(m0, m, n0, n); + }*/ + if (m_rem >= 8 && n_rem >= 8) { + mc = 8; + nc = 8; + gemm<8,8>(m0, m, n0, n); + } else if (m_rem >= 4 && n_rem >= 8) { + mc = 4; + nc = 8; + gemm<4,8>(m0, m, n0, n); + } else if (m_rem >= 8 && n_rem >= 4) { + mc = 8; + nc = 4; + gemm<8,4>(m0, m, n0, n); + } else if (m_rem >= 4 && n_rem >= 4) { + mc = 4; + nc = 4; + gemm_small<4, 4>(m0, m, n0, n); + } else if ((m_rem < 4) && (n_rem > 4)) { + nc = 4; + switch(m_rem) { + case 1: + mc = 1; + gemm_small<1, 4>(m0, m, n0, n); + break; + case 2: + mc = 2; + gemm_small<2, 4>(m0, m, n0, n); + break; + case 3: + mc = 3; + gemm_small<3, 4>(m0, m, n0, n); + break; + default: + return; + } + } else if ((m_rem > 4) && (n_rem < 4)) { + mc = 4; + switch(n_rem) { + case 1: + nc = 1; + gemm_small<4, 1>(m0, m, n0, n); + break; + case 2: + nc = 2; + gemm_small<4, 2>(m0, m, n0, n); + break; + case 3: + nc = 3; + gemm_small<4, 3>(m0, m, n0, n); + break; + default: + return; + } + } else { + switch((m_rem << 4) | n_rem) { + case 0x43: + mc = 4; + nc = 3; + gemm_small<4, 3>(m0, m, n0, n); + break; + case 0x42: + mc = 4; + nc = 2; + gemm_small<4, 2>(m0, m, n0, n); + break; + case 0x41: + mc = 4; + nc = 1; + gemm_small<4, 1>(m0, m, n0, n); + break; + case 0x34: + mc = 3; + nc = 4; + gemm_small<3, 4>(m0, m, n0, n); + break; + case 0x33: + mc = 3; + nc = 3; + gemm_small<3, 3>(m0, m, n0, n); + break; + case 0x32: + mc = 3; + nc = 2; + gemm_small<3, 2>(m0, m, n0, n); + break; + case 0x31: + mc = 3; + nc = 1; + gemm_small<3, 1>(m0, m, n0, n); + break; + case 0x24: + mc = 2; + nc = 4; + gemm_small<2, 4>(m0, m, n0, n); + break; + case 0x23: + mc = 2; + nc = 3; + gemm_small<2, 3>(m0, m, n0, n); + break; + case 0x22: + mc = 2; + nc = 2; + gemm_small<2, 2>(m0, m, n0, n); + break; + case 0x21: + mc = 2; + nc = 1; + gemm_small<2, 1>(m0, m, n0, n); + break; + case 0x14: + mc = 1; + nc = 4; + gemm_small<1, 4>(m0, m, n0, n); + break; + case 0x13: + mc = 1; + nc = 3; + gemm_small<1, 3>(m0, m, n0, n); + break; + case 0x12: + mc = 1; + nc = 2; + gemm_small<1, 2>(m0, m, n0, n); + break; + case 0x11: + mc = 1; + nc = 1; + gemm_small<1, 1>(m0, m, n0, n); + break; + default: + return; + } + } + mp = m0 + (m - m0) / mc * mc; + np = n0 + (n - n0) / nc * nc; + mnpack(mp, m, n0, np); + mnpack(m0, m, np, n); + } + + void KERNEL_4x8(int64_t ii, int64_t jj) { + vec_t vec_A[8], vec_B[16] = {0}; + acc_t acc_0, acc_1; + std::array comparray; + vector float fin_res[8] = {0}; + vector float vs[8] = {0}; + for (int l = 0; l < k; l++) { + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + packNormal((A+(ii*lda)+l), lda, 4, 8, (int8_t*)vec_A, false); + packNormal((B+(jj*ldb)+l), ldb, 8, 8, (uint8_t*)vec_B, true); + for(int x = 0; x < 8; x++) { + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x], vec_B[x]); + __builtin_mma_xvi8ger4pp(&acc_1, vec_A[x], vec_B[x+8]); + } + for (int I = 0; I<4; I++) { + for (int J = 0; J<4; J++) { + *((float*)&vs[I]+J) = (unhalf((A+((ii+I)*lda)+l)->d) * unhalf((B+((jj+J)*ldb)+l)->d)); + *((float*)&vs[I+4]+J) = (unhalf((A+((ii+I)*lda)+l)->d) * unhalf((B+((jj+J+4)*ldb)+l)->d)); + } + } + auto aoffset = A+(ii*lda)+l; + for (int i = 0; i < 4; i++) { + comparray[i] = 0; + int ca = 0; + const int8_t *at = aoffset->qs; + for (int j = 0; j < 32; j++) + ca += (int)*at++; + comparray[i] = ca; + aoffset += lda; + } + compute<4>(&acc_0, 0, 0, comparray, vs, fin_res); + compute<4>(&acc_1, 0, 4, comparray, vs, fin_res); + } + save_res<4, 4>(ii, jj, 0, fin_res); + save_res<4, 4>(ii, jj+4, 4, fin_res); + } + + void KERNEL_8x4(int64_t ii, int64_t jj) { + vec_t vec_A[16], vec_B[8] = {0}; + acc_t acc_0, acc_1; + std::array comparray; + vector float fin_res[8] = {0}; + vector float vs[8] = {0}; + for (int l = 0; l < k; l++) { + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + packNormal((A+(ii*lda)+l), lda, 8, 8, (int8_t*)vec_A, false); + packNormal((B+(jj*ldb)+l), ldb, 4, 8, (uint8_t*)vec_B, true); + for(int x = 0; x < 8; x++) { + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x], vec_B[x]); + __builtin_mma_xvi8ger4pp(&acc_1, vec_A[x+8], vec_B[x]); + } + for (int I = 0; I<8; I++) { + for (int J = 0; J<4; J++) { + *((float*)&vs[I]+J) = (unhalf((A+((ii+I)*lda)+l)->d) * unhalf((B+((jj+J)*ldb)+l)->d)); + } + } + auto aoffset = A+(ii*lda)+l; + for (int i = 0; i < 8; i++) { + comparray[i] = 0; + int ca = 0; + const int8_t *at = aoffset->qs; + for (int j = 0; j < 32; j++) + ca += (int)*at++; + comparray[i] = ca; + aoffset += lda; + } + compute<8>(&acc_0, 0, 0, comparray, vs, fin_res); + compute<8>(&acc_1, 4, 4, comparray, vs, fin_res); + } + save_res<4, 4>(ii, jj, 0, fin_res); + save_res<4, 4>(ii+4, jj, 4, fin_res); + } + + void KERNEL_8x8(int64_t ii, int64_t jj) { + vec_t vec_A[16], vec_B[16] = {0}; + acc_t acc_0, acc_1, acc_2, acc_3; + std::array comparray; + vector float fin_res[16] = {0}; + vector float vs[16] = {0}; + for (int l = 0; l < k; l++) { + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + __builtin_mma_xxsetaccz(&acc_2); + __builtin_mma_xxsetaccz(&acc_3); + packNormal((A+(ii*lda)+l), lda, 8, 8, (int8_t*)vec_A, false); + packNormal((B+(jj*ldb)+l), ldb, 8, 8, (uint8_t*)vec_B, true); + for(int x = 0; x < 8; x++) { + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x], vec_B[x]); + __builtin_mma_xvi8ger4pp(&acc_1, vec_A[x+8], vec_B[x]); + __builtin_mma_xvi8ger4pp(&acc_2, vec_A[x], vec_B[x+8]); + __builtin_mma_xvi8ger4pp(&acc_3, vec_A[x+8], vec_B[x+8]); + } + for (int I = 0; I<8; I++) { + for (int J = 0; J<4; J++) { + *((float*)&vs[I]+J) = (unhalf((A+((ii+I)*lda)+l)->d) * unhalf((B+((jj+J)*ldb)+l)->d)); + *((float*)&vs[I+8]+J) = (unhalf((A+((ii+I)*lda)+l)->d) * unhalf((B+((jj+J+4)*ldb)+l)->d)); + } + } + auto aoffset = A+(ii*lda)+l; + for (int i = 0; i < 8; i++) { + comparray[i] = 0; + int ca = 0; + const int8_t *at = aoffset->qs; + for (int j = 0; j < 32; j++) + ca += (int)*at++; + comparray[i] = ca; + aoffset += lda; + } + compute<8>(&acc_0, 0, 0, comparray, vs, fin_res); + compute<8>(&acc_1, 4, 4, comparray, vs, fin_res); + compute<8>(&acc_2, 0, 8, comparray, vs, fin_res); + compute<8>(&acc_3, 4, 12, comparray, vs, fin_res); + } + save_res<4, 4>(ii, jj, 0, fin_res); + save_res<4, 4>(ii+4, jj, 4, fin_res); + save_res<4, 4>(ii, jj+4, 8, fin_res); + save_res<4, 4>(ii+4, jj+4, 12, fin_res); + } + + template + void gemm_small(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + vec_t vec_A[8], vec_B[8] = {0}; + vector signed int vec_C[4]; + acc_t acc_0; + + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + std::array comparray; + vector float res[4] = {0}; + vector float fin_res[4] = {0}; + vector float vs[4] = {0}; + vector float CA[4] = {0}; + __builtin_prefetch((A+(ii*lda)+0)->qs, 0, 1); // prefetch first value + __builtin_prefetch((B+(jj*ldb)+0)->qs, 0, 1); // prefetch first value + for (int l = 0; l < k; l++) { + __builtin_prefetch((A+(ii*lda)+(l+1))->qs, 0, 1); // prefetch one loop ahead + __builtin_prefetch((B+(jj*ldb)+(l+1))->qs, 0, 1); // prefetch one loop ahead + __builtin_mma_xxsetaccz(&acc_0); + packNormal((A+(ii*lda)+l), lda, RM, 8, (int8_t*)vec_A, false); + packNormal((B+(jj*ldb)+l), ldb, RN, 8, (uint8_t*)vec_B, true); + for(int x = 0; x < 8; x+=4) { + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x], vec_B[x]); + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x+1], vec_B[x+1]); + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x+2], vec_B[x+2]); + __builtin_mma_xvi8ger4pp(&acc_0, vec_A[x+3], vec_B[x+3]); + } + for (int I = 0; Id) * unhalf((B+((jj+J)*ldb)+l)->d)); + } + } + __builtin_mma_disassemble_acc(vec_C, &acc_0); + auto aoffset = A+(ii*lda)+l; + for (int i = 0; i < RM; i++) { + comparray[i] = 0; + int ca = 0; + const int8_t *at = aoffset->qs; + for (int j = 0; j < 32; j++) + ca += (int)*at++; + comparray[i] = ca; + aoffset += lda; + } + + for (int i = 0; i < RM; i++) { + CA[i] = vec_splats((float)(((double)comparray[i]) * -128.0)); + res[i] = vec_add(vec_ctf(vec_C[i], 0), CA[i]); + fin_res[i] = vec_madd(res[i], vs[i], fin_res[i]); + } + } + save_res(ii, jj, 0, fin_res); + } + } + + template + inline void kernel(int64_t ii, int64_t jj) { + if constexpr(RM == 4 && RN == 8) { + KERNEL_4x8(ii,jj); + } else if constexpr(RM == 8 && RN == 4) { + KERNEL_8x4(ii,jj); + } else if constexpr(RM == 8 && RN == 8) { + KERNEL_8x8(ii,jj); + } else { + static_assert(false, "RN/RM values not supported"); + } + } + + template + NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + kernel(ii, jj); + } + } + + const TA *const A; + const TB *const B; + TC *C; + TA *At; + TB *Bt; + const int64_t k; + const int64_t lda; + const int64_t ldb; + const int64_t ldc; + const int ith; + const int nth; +}; + +template +class tinyBLAS_PPC { + public: + tinyBLAS_PPC(int64_t k, + const TA *A, int64_t lda, + const TB *B, int64_t ldb, + TC *C, int64_t ldc, + int ith, int nth) + : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { + } + + void matmul(int64_t m, int64_t n) { + mnpack(0, m, 0, n); + } + + private: + + void (tinyBLAS_PPC::*kernel)(int64_t, int64_t); + + template + void packTranspose(const TA* a, int64_t lda, int rows, int cols, TA* vec) { + int64_t i, j; + TA *aoffset = NULL, *boffset = NULL; + TA *aoffset1 = NULL, *aoffset2 = NULL, *aoffset3 = NULL, *aoffset4 = NULL; + TA *aoffset5 = NULL, *aoffset6 = NULL, *aoffset7 = NULL, *aoffset8 = NULL; + __vector_pair C1, C2, C3, C4, C5, C6, C7, C8; + VA c1[2] = {0}, c2[2] = {0}, c3[2] = {0}, c4[2] = {0}; + VA c5[2] = {0}, c6[2] = {0}, c7[2] = {0}, c8[2] = {0}; + VA t1, t2, t3, t4, t5, t6, t7, t8; + aoffset = const_cast(a); + boffset = vec; + j = (rows >> 3); + if (j > 0) { + do { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + aoffset4 = aoffset3 + lda; + aoffset5 = aoffset4 + lda; + aoffset6 = aoffset5 + lda; + aoffset7 = aoffset6 + lda; + aoffset8 = aoffset7 + lda; + aoffset += 8 * lda; + i = (cols >> 3); + if (i > 0) { + do { + C1 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset1); + C2 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset2); + C3 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset3); + C4 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset4); + C5 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset5); + C6 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset6); + C7 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset7); + C8 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset8); + __builtin_vsx_disassemble_pair(c1, &C1); + __builtin_vsx_disassemble_pair(c2, &C2); + __builtin_vsx_disassemble_pair(c3, &C3); + __builtin_vsx_disassemble_pair(c4, &C4); + __builtin_vsx_disassemble_pair(c5, &C5); + __builtin_vsx_disassemble_pair(c6, &C6); + __builtin_vsx_disassemble_pair(c7, &C7); + __builtin_vsx_disassemble_pair(c8, &C8); + + t1 = vec_mergeh(c1[0], c2[0]); + t2 = vec_mergeh(c3[0], c4[0]); + t3 = vec_mergeh(c5[0], c6[0]); + t4 = vec_mergeh(c7[0], c8[0]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset); + vec_xst(t6, 0, boffset+4); + vec_xst(t7, 0, boffset+8); + vec_xst(t8, 0, boffset+12); + + t1 = vec_mergel(c1[0], c2[0]); + t2 = vec_mergel(c3[0], c4[0]); + t3 = vec_mergel(c5[0], c6[0]); + t4 = vec_mergel(c7[0], c8[0]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset+16); + vec_xst(t6, 0, boffset+20); + vec_xst(t7, 0, boffset+24); + vec_xst(t8, 0, boffset+28); + + t1 = vec_mergeh(c1[1], c2[1]); + t2 = vec_mergeh(c3[1], c4[1]); + t3 = vec_mergeh(c5[1], c6[1]); + t4 = vec_mergeh(c7[1], c8[1]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset+32); + vec_xst(t6, 0, boffset+36); + vec_xst(t7, 0, boffset+40); + vec_xst(t8, 0, boffset+44); + + t1 = vec_mergel(c1[1], c2[1]); + t2 = vec_mergel(c3[1], c4[1]); + t3 = vec_mergel(c5[1], c6[1]); + t4 = vec_mergel(c7[1], c8[1]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset+48); + vec_xst(t6, 0, boffset+52); + vec_xst(t7, 0, boffset+56); + vec_xst(t8, 0, boffset+60); + + aoffset1 += 8*lda; + aoffset2 += 8*lda; + aoffset3 += 8*lda; + aoffset4 += 8*lda; + boffset += 64; + i--; + } while(i > 0); + } + if (cols & 4) { + c1[0] = vec_xl(0, aoffset1); + c2[0] = vec_xl(0, aoffset2); + c3[0] = vec_xl(0, aoffset3); + c4[0] = vec_xl(0, aoffset4); + c5[0] = vec_xl(0, aoffset5); + c6[0] = vec_xl(0, aoffset6); + c7[0] = vec_xl(0, aoffset7); + c8[0] = vec_xl(0, aoffset8); + + t1 = vec_mergeh(c1[0], c2[0]); + t2 = vec_mergeh(c3[0], c4[0]); + t3 = vec_mergeh(c5[0], c6[0]); + t4 = vec_mergeh(c7[0], c8[0]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset); + vec_xst(t6, 0, boffset+4); + vec_xst(t7, 0, boffset+8); + vec_xst(t8, 0, boffset+12); + + t1 = vec_mergel(c1[0], c2[0]); + t2 = vec_mergel(c3[0], c4[0]); + t3 = vec_mergel(c5[0], c6[0]); + t4 = vec_mergel(c7[0], c8[0]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t3, t4, 0); + t7 = vec_xxpermdi(t1, t2, 3); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset+16); + vec_xst(t6, 0, boffset+20); + vec_xst(t7, 0, boffset+24); + vec_xst(t8, 0, boffset+28); + } + j--; + } while(j > 0); + } + + if (rows & 4) { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + aoffset4 = aoffset3 + lda; + aoffset += 4 * lda; + i = (cols >> 3); + if (i > 0) { + do { + C1 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset1); + C2 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset2); + C3 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset3); + C4 = __builtin_vsx_lxvp(0, (__vector_pair*)aoffset4); + __builtin_vsx_disassemble_pair(c1, &C1); + __builtin_vsx_disassemble_pair(c2, &C2); + __builtin_vsx_disassemble_pair(c3, &C3); + __builtin_vsx_disassemble_pair(c4, &C4); + + t1 = vec_mergeh(c1[0], c2[0]); + t2 = vec_mergeh(c3[0], c4[0]); + t3 = vec_mergel(c1[0], c2[0]); + t4 = vec_mergel(c3[0], c4[0]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t1, t2, 3); + t7 = vec_xxpermdi(t3, t4, 0); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset); + vec_xst(t6, 0, boffset+4); + vec_xst(t7, 0, boffset+8); + vec_xst(t8, 0, boffset+12); + + t1 = vec_mergeh(c1[1], c2[1]); + t2 = vec_mergeh(c3[1], c4[1]); + t3 = vec_mergel(c1[1], c2[1]); + t4 = vec_mergel(c3[1], c4[1]); + t5 = vec_xxpermdi(t1, t2, 0); + t6 = vec_xxpermdi(t1, t2, 3); + t7 = vec_xxpermdi(t3, t4, 0); + t8 = vec_xxpermdi(t3, t4, 3); + vec_xst(t5, 0, boffset+16); + vec_xst(t6, 0, boffset+20); + vec_xst(t7, 0, boffset+24); + vec_xst(t8, 0, boffset+28); + + aoffset1 += 8*lda; + aoffset2 += 8*lda; + aoffset3 += 8*lda; + aoffset4 += 8*lda; + boffset += 32; + i--; + } while(i > 0); + } + + if (cols & 4) { + c1[0] = vec_xl(0, aoffset1); + c2[0] = vec_xl(0, aoffset2); + c3[0] = vec_xl(0, aoffset3); + c4[0] = vec_xl(0, aoffset4); + + t1 = vec_mergeh(c1[0], c2[0]); + t2 = vec_mergeh(c3[0], c4[0]); + t3 = vec_xxpermdi(t1, t2, 0); + t4 = vec_xxpermdi(t1, t2, 3); + vec_xst(t3, 0, boffset); + vec_xst(t4, 0, boffset+4); + + t1 = vec_mergel(c1[0], c2[0]); + t2 = vec_mergel(c3[0], c4[0]); + t3 = vec_xxpermdi(t1, t2, 0); + t4 = vec_xxpermdi(t1, t2, 3); + vec_xst(t3, 0, boffset+8); + vec_xst(t4, 0, boffset+12); + } + } + if (rows & 3) { + aoffset1 = aoffset; + aoffset2 = aoffset1 + lda; + aoffset3 = aoffset2 + lda; + if (cols & 4) { + c1[0] = vec_xl(0, aoffset1); + c2[0] = vec_xl(0, aoffset2); + c3[0] = vec_xl(0, aoffset3); + + t1 = vec_mergeh(c1[0], c2[0]); + t2 = vec_mergeh(c3[0], c4[0]); + t3 = vec_xxpermdi(t1, t2, 0); + t4 = vec_xxpermdi(t1, t2, 3); + vec_xst(t3, 0, boffset); + vec_xst(t4, 0, boffset+4); + + t1 = vec_mergel(c1[0], c2[0]); + t2 = vec_mergel(c3[0], c4[0]); + t3 = vec_xxpermdi(t1, t2, 0); + t4 = vec_xxpermdi(t1, t2, 3); + vec_xst(t3, 0, boffset+8); + vec_xst(t4, 0, boffset+12); + } + } + } + void KERNEL_4x4(int64_t ii, int64_t jj) { + vec_t vec_A[4], vec_B[4], vec_C[4]; + acc_t acc_0; + __builtin_mma_xxsetaccz(&acc_0); + for (int l = 0; l < k; l+=4) { + packTranspose(A+(ii*lda)+l, lda, 4, 4, (TA*)vec_A); + packTranspose(B+(jj*ldb)+l, ldb, 4, 4, (TA*)vec_B); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[0], vec_B[0]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[1], vec_B[1]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[2], vec_B[2]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[3], vec_B[3]); + } + SAVE_ACC(&acc_0, ii, jj); + } + + void KERNEL_4x8(int64_t ii, int64_t jj) { + vec_t vec_A[4], vec_B[8], vec_C[4]; + acc_t acc_0, acc_1; + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + for (int64_t l = 0; l < k; l+=4) { + packTranspose(A+(ii*lda)+l, lda, 4, 4, (TA*)vec_A); + packTranspose(B+(jj*ldb)+l, ldb, 8, 4, (TA*)vec_B); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[0], (vec_t)vec_B[0]); + __builtin_mma_xvf32gerpp(&acc_1, vec_A[0], (vec_t)vec_B[1]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[1], (vec_t)vec_B[2]); + __builtin_mma_xvf32gerpp(&acc_1, vec_A[1], (vec_t)vec_B[3]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[2], (vec_t)vec_B[4]); + __builtin_mma_xvf32gerpp(&acc_1, vec_A[2], (vec_t)vec_B[5]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[3], (vec_t)vec_B[6]); + __builtin_mma_xvf32gerpp(&acc_1, vec_A[3], (vec_t)vec_B[7]); + } + SAVE_ACC(&acc_0, ii, jj); + SAVE_ACC(&acc_1, ii, jj+4); + } + + void KERNEL_8x4(int64_t ii, int64_t jj) { + vec_t vec_A[8], vec_B[4], vec_C[4]; + acc_t acc_0, acc_1; + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + for (int64_t l = 0; l < k; l+=4) { + packTranspose(A+(ii*lda)+l, lda, 8, 4, (TA*)vec_A); + packTranspose(B+(jj*ldb)+l, ldb, 4, 4, (TA*)vec_B); + __builtin_mma_xvf32gerpp(&acc_0, (vec_t)vec_A[0], vec_B[0]); + __builtin_mma_xvf32gerpp(&acc_1, (vec_t)vec_A[1], vec_B[0]); + __builtin_mma_xvf32gerpp(&acc_0, (vec_t)vec_A[2], vec_B[1]); + __builtin_mma_xvf32gerpp(&acc_1, (vec_t)vec_A[3], vec_B[1]); + __builtin_mma_xvf32gerpp(&acc_0, (vec_t)vec_A[4], vec_B[2]); + __builtin_mma_xvf32gerpp(&acc_1, (vec_t)vec_A[5], vec_B[2]); + __builtin_mma_xvf32gerpp(&acc_0, (vec_t)vec_A[6], vec_B[3]); + __builtin_mma_xvf32gerpp(&acc_1, (vec_t)vec_A[7], vec_B[3]); + } + SAVE_ACC(&acc_0, ii, jj); + SAVE_ACC(&acc_1, ii+4, jj); + } + + void KERNEL_8x8(int64_t ii, int64_t jj) { + vec_t vec_A[16], vec_B[16], vec_C[4]; + acc_t acc_0, acc_1, acc_2, acc_3; + __builtin_mma_xxsetaccz(&acc_0); + __builtin_mma_xxsetaccz(&acc_1); + __builtin_mma_xxsetaccz(&acc_2); + __builtin_mma_xxsetaccz(&acc_3); + for (int l = 0; l < k; l+=8) { + packTranspose(A+(ii*lda)+l, lda, 8, 8, (TA*)vec_A); + packTranspose(B+(jj*ldb)+l, ldb, 8, 8, (TA*)vec_B); + for(int x = 0; x < 16; x+=2) { + __builtin_mma_xvf32gerpp(&acc_0, (vec_t)vec_A[x], vec_B[x]); + __builtin_mma_xvf32gerpp(&acc_1, (vec_t)vec_A[x], vec_B[x+1]); + __builtin_mma_xvf32gerpp(&acc_2, (vec_t)vec_A[x+1], vec_B[x]); + __builtin_mma_xvf32gerpp(&acc_3, (vec_t)vec_A[x+1], vec_B[x+1]); + } + } + SAVE_ACC(&acc_0, ii, jj); + SAVE_ACC(&acc_1, ii, jj+4); + SAVE_ACC(&acc_2, ii+4, jj); + SAVE_ACC(&acc_3, ii+4, jj+4); + } + + void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t mc, nc, mp, np; + int m_rem = MIN(m - m0, 16); + int n_rem = MIN(n - n0, 16); + if (m_rem >= 16 && n_rem >= 8) { + mc = 8; + nc = 8; + gemm<8,8>(m0, m, n0, n); + } else if(m_rem >= 8 && n_rem >= 16) { + mc = 8; + nc = 8; + gemm<8,8>(m0, m, n0, n); + } else if (m_rem >= 8 && n_rem >= 8) { + mc = 8; + nc = 8; + gemm<8,8>(m0, m, n0, n); + } else if (m_rem >= 4 && n_rem >= 8) { + mc = 4; + nc = 8; + gemm<4,8>(m0, m, n0, n); + } else if (m_rem >= 8 && n_rem >= 4) { + mc = 8; + nc = 4; + gemm<8,4>(m0, m, n0, n); + } else if (m_rem >= 4 && n_rem >= 4) { + mc = 4; + nc = 4; + gemm<4,4>(m0, m, n0, n); + } else if ((m_rem < 4) && (n_rem > 4)) { + nc = 4; + switch(m_rem) { + case 1: + mc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 2: + mc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 3: + mc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + default: + return; + } + } else if ((m_rem > 4) && (n_rem < 4)) { + mc = 4; + switch(n_rem) { + case 1: + nc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 2: + nc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 3: + nc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + default: + return; + } + } else { + switch((m_rem << 4) | n_rem) { + case 0x43: + mc = 4; + nc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x42: + mc = 4; + nc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x41: + mc = 4; + nc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x34: + mc = 3; + nc = 4; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x33: + mc = 3; + nc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x32: + mc = 3; + nc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x31: + mc = 3; + nc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x24: + mc = 2; + nc = 4; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x23: + mc = 2; + nc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x22: + mc = 2; + nc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x21: + mc = 2; + nc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x14: + mc = 1; + nc = 4; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x13: + mc = 1; + nc = 3; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x12: + mc = 1; + nc = 2; + gemm_small(m0, m, n0, n, mc, nc); + break; + case 0x11: + mc = 1; + nc = 1; + gemm_small(m0, m, n0, n, mc, nc); + break; + default: + return; + } + } + mp = m0 + (m - m0) / mc * mc; + np = n0 + (n - n0) / nc * nc; + mnpack(mp, m, n0, np); + mnpack(m0, m, np, n); + } + + void gemm_small(int64_t m0, int64_t m, int64_t n0, int64_t n, int RM, int RN) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + vec_t vec_C[4]; + acc_t acc_0; + __builtin_mma_xxsetaccz(&acc_0); + vec_t vec_A[4], vec_B[4]; + for (int l=0; l= 4 && RM == 1) { + TA* a = const_cast(A+(ii)*lda+l); + packTranspose(B+(jj*ldb)+l, ldb, 4, 4, (TA*)vec_B); + vec_A[0] = (vec_t)vec_xl(0,a); + vec_A[1] = (vec_t)vec_splats(*((TA*)&vec_A+1)); + vec_A[2] = (vec_t)vec_splats(*((TA*)&vec_A+2)); + vec_A[3] = (vec_t)vec_splats(*((TA*)&vec_A+3)); + } else { + packTranspose(A+(ii*lda)+l, lda, RM, 4, (TA*)vec_A); + packTranspose(B+(jj*ldb)+l, ldb, RN, 4, (TA*)vec_B); + } + __builtin_mma_xvf32gerpp(&acc_0, vec_A[0], vec_B[0]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[1], vec_B[1]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[2], vec_B[2]); + __builtin_mma_xvf32gerpp(&acc_0, vec_A[3], vec_B[3]); + } + __builtin_mma_disassemble_acc(vec_C, &acc_0); + for (int I = 0; I < RM; I++) { + for (int J = 0; J < RN; J++) { + *((TC*)(C+ii+((jj+J)*ldc)+I)) = *((TC*)&vec_C[I]+J); + } + } + } + } + + template + NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { + int64_t ytiles = (m - m0) / RM; + int64_t xtiles = (n - n0) / RN; + int64_t tiles = xtiles * ytiles; + int64_t duty = (tiles + nth - 1) / nth; + int64_t start = duty * ith; + int64_t end = start + duty; + if (RM == 4 && RN == 4) { + kernel = &tinyBLAS_PPC::KERNEL_4x4; + } else if (RM == 4 && RN == 8) { + kernel = &tinyBLAS_PPC::KERNEL_4x8; + } else if (RM == 8 && RN == 4) { + kernel = &tinyBLAS_PPC::KERNEL_8x4; + } else if (RM == 8 && RN == 8) { + kernel = &tinyBLAS_PPC::KERNEL_8x8; + } + if (end > tiles) + end = tiles; + for (int64_t job = start; job < end; ++job) { + int64_t ii = m0 + job / xtiles * RM; + int64_t jj = n0 + job % xtiles * RN; + (this->*kernel)(ii, jj); + } + } + + const TA *const A; + const TB *const B; + TC *C; + TA *At; + TB *Bt; + const int64_t k; + const int64_t lda; + const int64_t ldb; + const int64_t ldc; + const int ith; + const int nth; +}; +#endif +} // namespace + +/** + * Performs optimized matrix multiplication on CPU. + * + * This subroutine may compute C = Aᵀ * B with column major ordering. + * Despite its name, this isn't a generalized implementation. Work is + * only performed when a handwritten kernel is written and available. + * Otherwise the caller should fall back to a general matmul routine. + * + * For example, for single-threaded single-precision GEMM you can say + * + * llamafile_sgemm(m, n, k, A, lda, B, ldb, C, ldc, + * 0, 1, + * GGML_TYPE_F32, GGML_TYPE_F32, GGML_TYPE_F32); + * + * @param m is rows in `A` and `C` + * @param n is cols in `B` and `C` + * @param k is cols in `A` and rows in `B` + * @param A is first input matrix (always transposed) + * @param lda is row stride of `A` + * @param B is second input matrix (never transposed) + * @param ldb is row stride of `B` + * @param C is input/output array of output matrices + * @param ldc is row stride of `C` + * @param ith is thread id (must be less than `nth`) + * @param nth is number of threads (must be greater than zero) + * @param Atype is GGML data type of `A` + * @param Btype is GGML data type of `B` + * @param Ctype is GGML data type of `C` + * @return true if this function was able to service the matmul request + */ +bool llamafile_sgemm(const struct ggml_compute_params * params, int64_t m, int64_t n, int64_t k, + const void *A, int64_t lda, const void *B, int64_t ldb, void *C, + int64_t ldc, int Atype, int Btype, int Ctype) { + + assert(m >= 0); + assert(n >= 0); + assert(k >= 0); + assert(lda >= k); + assert(ldb >= k); + assert(ldc >= m); + assert(params->nth > 0); + assert(params->ith < params->nth); + + // only enable sgemm for prompt processing + if (n < 2) + return false; + + if (Ctype != GGML_TYPE_F32) + return false; + + switch (Atype) { + + case GGML_TYPE_F32: { + if (Btype != GGML_TYPE_F32) + return false; +#if defined(__AVX512F__) + tinyBLAS<16, __m512, __m512, float, float, float> tb{ params, + k, (const float *)A, lda, + (const float *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); +#elif defined(__AVX__) || defined(__AVX2__) + tinyBLAS<8, __m256, __m256, float, float, float> tb{ params, + k, (const float *)A, lda, + (const float *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); +#elif defined(__ARM_NEON) + if (n < 4) + return false; + tinyBLAS<4, float32x4_t, float32x4_t, float, float, float> tb{ params, + k, (const float *)A, lda, + (const float *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); +#elif defined(__MMA__) + if (k % 8) + return false; + tinyBLAS_PPC tb{ + k, (const float *)A, lda, + (const float *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#else + return false; +#endif + } + + case GGML_TYPE_BF16: { +#if defined(__AVX512BF16__) + if (Btype == GGML_TYPE_BF16) { + tinyBLAS<32, __m512, __m512bh, ggml_bf16_t, ggml_bf16_t, float> tb{ params, k, + (const ggml_bf16_t *)A, lda, + (const ggml_bf16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#elif defined(__AVX512F__) + if (Btype == GGML_TYPE_BF16) { + tinyBLAS<16, __m512, __m512, ggml_bf16_t, ggml_bf16_t, float> tb{ params, k, + (const ggml_bf16_t *)A, lda, + (const ggml_bf16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#elif defined(__AVX2__) + if (Btype == GGML_TYPE_BF16) { + tinyBLAS<8, __m256, __m256, ggml_bf16_t, ggml_bf16_t, float> tb{ params, k, + (const ggml_bf16_t *)A, lda, + (const ggml_bf16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#endif + return false; + } + case GGML_TYPE_F16: { +#if defined(__AVX512F__) + if (Btype == GGML_TYPE_F16) { + tinyBLAS<16, __m512, __m512, ggml_fp16_t, ggml_fp16_t, float> tb{ params, k, + (const ggml_fp16_t *)A, lda, + (const ggml_fp16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#elif (defined(__AVX__) || defined(__AVX2__)) && defined(__F16C__) + if (Btype == GGML_TYPE_F16) { + tinyBLAS<8, __m256, __m256, ggml_fp16_t, ggml_fp16_t, float> tb{ params, k, + (const ggml_fp16_t *)A, lda, + (const ggml_fp16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#elif defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) + if (n < 8) + return false; + if (Btype == GGML_TYPE_F16) { + tinyBLAS<8, float16x8_t, float16x8_t, ggml_fp16_t, ggml_fp16_t, float> tb{ params, + k, (const ggml_fp16_t *)A, lda, + (const ggml_fp16_t *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#elif defined(__ARM_NEON) && !defined(_MSC_VER) + if (Btype == GGML_TYPE_F32) { + tinyBLAS<4, float32x4_t, float32x4_t, ggml_fp16_t, float, float> tb{ params, + k, (const ggml_fp16_t *)A, lda, + (const float *)B, ldb, + (float *)C, ldc}; + return tb.matmul(m, n); + } +#endif + return false; + } + + case GGML_TYPE_Q8_0: { + if (Btype != GGML_TYPE_Q8_0) + return false; +#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) + tinyBLAS_Q0_AVX tb{ + k, (const block_q8_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#elif defined(__ARM_FEATURE_DOTPROD) + tinyBLAS_Q0_ARM tb{ + k, (const block_q8_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; + +#elif defined(__MMA__) + if (n < 8 && n != 4) + return false; + if (m < 8 && m != 4) + return false; + tinyBLAS_Q0_PPC tb{ + k, (const block_q8_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; + +#else + return false; +#endif + } + + case GGML_TYPE_Q4_0: { + if (Btype != GGML_TYPE_Q8_0) + return false; +#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) + tinyBLAS_Q0_AVX tb{ + k, (const block_q4_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#elif defined(__ARM_FEATURE_DOTPROD) + tinyBLAS_Q0_ARM tb{ + k, (const block_q4_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#else + return false; +#endif + } + + case GGML_TYPE_Q5_0: { + if (Btype != GGML_TYPE_Q8_0) + return false; +#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) + tinyBLAS_Q0_AVX tb{ + k, (const block_q5_0 *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#else + return false; +#endif + } + + case GGML_TYPE_IQ4_NL: { + if (Btype != GGML_TYPE_Q8_0) + return false; +#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) + tinyBLAS_Q0_AVX tb{ + k, (const block_iq4_nl *)A, lda, + (const block_q8_0 *)B, ldb, + (float *)C, ldc, + params->ith, params->nth}; + tb.matmul(m, n); + return true; +#else + return false; +#endif + } + + default: + return false; + } + + (void)params; + (void)m; + (void)n; + (void)k; + (void)A; + (void)lda; + (void)B; + (void)ldb; + (void)C; + (void)ldc; + (void)Atype; + (void)Btype; + (void)Ctype; +} diff --git a/ggml/src/ggml-cpu/llamafile/sgemm.h b/ggml/src/ggml-cpu/llamafile/sgemm.h new file mode 100644 index 000000000..3d2909515 --- /dev/null +++ b/ggml/src/ggml-cpu/llamafile/sgemm.h @@ -0,0 +1,14 @@ +#pragma once +#include +#include +#ifdef __cplusplus +extern "C" { +#endif + +bool llamafile_sgemm(const struct ggml_compute_params * params, int64_t, int64_t, int64_t, + const void *, int64_t, const void *, int64_t, void *, int64_t, + int, int, int); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-cuda/CMakeLists.txt b/ggml/src/ggml-cuda/CMakeLists.txt new file mode 100644 index 000000000..14761650f --- /dev/null +++ b/ggml/src/ggml-cuda/CMakeLists.txt @@ -0,0 +1,152 @@ +cmake_minimum_required(VERSION 3.18) # for CMAKE_CUDA_ARCHITECTURES + +find_package(CUDAToolkit) + +if (CUDAToolkit_FOUND) + message(STATUS "CUDA Toolkit found") + + if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES) + # native == GPUs available at build time + # 52 == Maxwell, lowest CUDA 12 standard + # 60 == P100, FP16 CUDA intrinsics + # 61 == Pascal, __dp4a instruction (per-byte integer dot product) + # 70 == V100, FP16 tensor cores + # 75 == Turing, int8 tensor cores + if (GGML_NATIVE AND CUDAToolkit_VERSION VERSION_GREATER_EQUAL "11.6" AND CMAKE_VERSION VERSION_GREATER_EQUAL "3.24") + set(CMAKE_CUDA_ARCHITECTURES "native") + elseif(GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16) + set(CMAKE_CUDA_ARCHITECTURES "60;61;70;75") + else() + set(CMAKE_CUDA_ARCHITECTURES "52;61;70;75") + endif() + endif() + message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}") + + enable_language(CUDA) + + file(GLOB GGML_HEADERS_CUDA "*.cuh") + list(APPEND GGML_HEADERS_CUDA "../../include/ggml-cuda.h") + + file(GLOB GGML_SOURCES_CUDA "*.cu") + file(GLOB SRCS "template-instances/fattn-wmma*.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + file(GLOB SRCS "template-instances/mmq*.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + + if (GGML_CUDA_FA_ALL_QUANTS) + file(GLOB SRCS "template-instances/fattn-vec*.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) + else() + file(GLOB SRCS "template-instances/fattn-vec*q4_0-q4_0.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + file(GLOB SRCS "template-instances/fattn-vec*q8_0-q8_0.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + file(GLOB SRCS "template-instances/fattn-vec*f16-f16.cu") + list(APPEND GGML_SOURCES_CUDA ${SRCS}) + endif() + + ggml_add_backend_library(ggml-cuda + ${GGML_HEADERS_CUDA} + ${GGML_SOURCES_CUDA} + ) + + add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${GGML_CUDA_PEER_MAX_BATCH_SIZE}) + + if (GGML_CUDA_GRAPHS) + add_compile_definitions(GGML_CUDA_USE_GRAPHS) + endif() + + if (GGML_CUDA_FORCE_MMQ) + add_compile_definitions(GGML_CUDA_FORCE_MMQ) + endif() + + if (GGML_CUDA_FORCE_CUBLAS) + add_compile_definitions(GGML_CUDA_FORCE_CUBLAS) + endif() + + if (GGML_CUDA_NO_VMM) + add_compile_definitions(GGML_CUDA_NO_VMM) + endif() + + if (GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16) + add_compile_definitions(GGML_CUDA_F16) + endif() + + if (GGML_CUDA_NO_PEER_COPY) + add_compile_definitions(GGML_CUDA_NO_PEER_COPY) + endif() + + if (GGML_STATIC) + if (WIN32) + # As of 12.3.1 CUDA Toolkit for Windows does not offer a static cublas library + target_link_libraries(ggml-cuda PRIVATE CUDA::cudart_static CUDA::cublas CUDA::cublasLt) + else () + target_link_libraries(ggml-cuda PRIVATE CUDA::cudart_static CUDA::cublas_static CUDA::cublasLt_static) + endif() + else() + target_link_libraries(ggml-cuda PRIVATE CUDA::cudart CUDA::cublas CUDA::cublasLt) + endif() + + if (GGML_CUDA_NO_VMM) + # No VMM requested, no need to link directly with the cuda driver lib (libcuda.so) + else() + target_link_libraries(ggml-cuda PRIVATE CUDA::cuda_driver) + endif() + + set(CUDA_CXX_FLAGS "") + + set(CUDA_FLAGS -use_fast_math) + + if (GGML_FATAL_WARNINGS) + list(APPEND CUDA_FLAGS -Werror all-warnings) + endif() + + if (GGML_ALL_WARNINGS AND NOT MSVC) + set(NVCC_CMD ${CMAKE_CUDA_COMPILER} .c) + if (NOT CMAKE_CUDA_HOST_COMPILER STREQUAL "") + list(APPEND NVCC_CMD -ccbin ${CMAKE_CUDA_HOST_COMPILER}) + endif() + + execute_process( + COMMAND ${NVCC_CMD} -Xcompiler --version + OUTPUT_VARIABLE CUDA_CCFULLVER + ERROR_QUIET + ) + + if (NOT CUDA_CCFULLVER MATCHES clang) + set(CUDA_CCID "GNU") + execute_process( + COMMAND ${NVCC_CMD} -Xcompiler "-dumpfullversion -dumpversion" + OUTPUT_VARIABLE CUDA_CCVER + ERROR_QUIET + ) + else() + if (CUDA_CCFULLVER MATCHES Apple) + set(CUDA_CCID "AppleClang") + else() + set(CUDA_CCID "Clang") + endif() + string(REGEX REPLACE "^.* version ([0-9.]*).*$" "\\1" CUDA_CCVER ${CUDA_CCFULLVER}) + endif() + + message("-- CUDA host compiler is ${CUDA_CCID} ${CUDA_CCVER}") + + ggml_get_flags(${CUDA_CCID} ${CUDA_CCVER}) + list(APPEND CUDA_CXX_FLAGS ${CXX_FLAGS} ${GF_CXX_FLAGS}) # This is passed to -Xcompiler later + endif() + + if (NOT MSVC) + list(APPEND CUDA_CXX_FLAGS -Wno-pedantic) + endif() + + list(JOIN CUDA_CXX_FLAGS " " CUDA_CXX_FLAGS_JOINED) # pass host compiler flags as a single argument + + if (NOT CUDA_CXX_FLAGS_JOINED STREQUAL "") + list(APPEND CUDA_FLAGS -Xcompiler ${CUDA_CXX_FLAGS_JOINED}) + endif() + + target_compile_options(ggml-cuda PRIVATE "$<$:${CUDA_FLAGS}>") +else() + message(FATAL_ERROR "CUDA Toolkit not found") +endif() diff --git a/ggml/src/ggml-cuda/argmax.cu b/ggml/src/ggml-cuda/argmax.cu index aab04eca7..5340eedc0 100644 --- a/ggml/src/ggml-cuda/argmax.cu +++ b/ggml/src/ggml-cuda/argmax.cu @@ -1,57 +1,69 @@ -#include "common.cuh" -#include "argmax.cuh" -#include "sum.cuh" - +#include #include -static __global__ void argmax_f32( - const float * x, int32_t * dst, const int64_t ncols, const int64_t nrows) { +#include "argmax.cuh" +#include "common.cuh" +#include "sum.cuh" - int argmax_thread = 0; - const int64_t row0 = (int64_t)blockIdx.x*WARP_SIZE; +static __global__ void argmax_f32(const float * __restrict__ x, int32_t * __restrict__ dst, const int64_t ncols) { + const int64_t row = blockIdx.x; -#pragma unroll - for (int64_t row1 = 0; row1 < WARP_SIZE; ++row1) { - const int64_t row = row0 + row1; + float maxval = -FLT_MAX; + int argmax = -1; + const float * rowx = x + row * ncols; - if (row >= nrows) { - break; + for (int32_t col = threadIdx.x; col < ncols; col += blockDim.x) { + const float val = rowx[col]; + if (val > maxval) { + maxval = val; + argmax = col; } - - float maxval = -FLT_MAX; - int argmax = -1; - - for (int32_t col = threadIdx.x; col < ncols; col += WARP_SIZE) { - const float val = x[row*ncols + col]; - const int bigger = val > maxval; - const int not_bigger = bigger ^ 0x00000001; - - maxval = maxval*not_bigger + val*bigger; - argmax = argmax*not_bigger + col*bigger; - } - -#pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - const float val = __shfl_xor_sync(0xFFFFFFFF, maxval, mask, WARP_SIZE); - const int col = __shfl_xor_sync(0xFFFFFFFF, argmax, mask, WARP_SIZE); - const int bigger = val > maxval; - const int not_bigger = bigger ^ 0x00000001; - - maxval = maxval*not_bigger + val*bigger; - argmax = argmax*not_bigger + col*bigger; - } - - const int store = row1 == threadIdx.x; - argmax_thread += store*argmax; } - const int row = row0 + threadIdx.x; - - if (row >= nrows) { - return; +#pragma unroll + for (int offset = 16; offset > 0; offset >>= 1) { + const float val = __shfl_xor_sync(0xFFFFFFFF, maxval, offset, WARP_SIZE); + const int col = __shfl_xor_sync(0xFFFFFFFF, argmax, offset, WARP_SIZE); + if (val > maxval) { + maxval = val; + argmax = col; + } } - dst[row] = argmax_thread; + const int n_warps = blockDim.x / WARP_SIZE; + const int lane_id = threadIdx.x % WARP_SIZE; + const int warp_id = threadIdx.x / WARP_SIZE; + if (n_warps > 1) { + constexpr int max_warps = 1024 / WARP_SIZE; + __shared__ float shared_maxval[max_warps]; + __shared__ int shared_argmax[max_warps]; + if (lane_id == 0) { + shared_maxval[warp_id] = maxval; + shared_argmax[warp_id] = argmax; + } + + __syncthreads(); + + if (warp_id == 0) { + if (lane_id < n_warps) { + maxval = shared_maxval[lane_id]; + argmax = shared_argmax[lane_id]; + } +#pragma unroll + for (int offset = 16; offset > 0; offset >>= 1) { + const float val = __shfl_xor_sync(0xFFFFFFFF, maxval, offset, WARP_SIZE); + const int col = __shfl_xor_sync(0xFFFFFFFF, argmax, offset, WARP_SIZE); + if (val > maxval) { + maxval = val; + argmax = col; + } + } + } + } + + if (warp_id == 0 && lane_id == 0) { + dst[row] = argmax; + } } void ggml_cuda_argmax(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { @@ -70,10 +82,10 @@ void ggml_cuda_argmax(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { cudaStream_t stream = ctx.stream(); - const int64_t num_blocks = (nrows + WARP_SIZE - 1) / WARP_SIZE; - - const dim3 blocks_dim(WARP_SIZE, 1, 1); + const int64_t num_blocks = nrows; + const int64_t num_threads = std::min(1024, (ne00 + WARP_SIZE - 1) / WARP_SIZE * WARP_SIZE); + const dim3 blocks_dim(num_threads, 1, 1); const dim3 blocks_num(num_blocks, 1, 1); - argmax_f32<<>>(src0_d, dst_d, ne00, nrows); + argmax_f32<<>>(src0_d, dst_d, ne00); } diff --git a/ggml/src/ggml-cuda/common.cuh b/ggml/src/ggml-cuda/common.cuh index dd203fcde..2c0a56226 100644 --- a/ggml/src/ggml-cuda/common.cuh +++ b/ggml/src/ggml-cuda/common.cuh @@ -6,7 +6,7 @@ #include #include -#if defined(GGML_USE_HIPBLAS) +#if defined(GGML_USE_HIP) #define GGML_COMMON_DECL_HIP #define GGML_COMMON_IMPL_HIP #else @@ -26,13 +26,13 @@ #include #include -#if defined(GGML_USE_HIPBLAS) +#if defined(GGML_USE_HIP) #include "vendors/hip.h" #elif defined(GGML_USE_MUSA) #include "vendors/musa.h" #else #include "vendors/cuda.h" -#endif // defined(GGML_USE_HIPBLAS) +#endif // defined(GGML_USE_HIP) #define STRINGIZE_IMPL(...) #__VA_ARGS__ #define STRINGIZE(...) STRINGIZE_IMPL(__VA_ARGS__) @@ -41,17 +41,28 @@ #define CUDART_HMAX 11070 // CUDA 11.7, min. ver. for which __hmax and __hmax2 are known to work (may be higher than needed) #define CUDART_HMASK 12000 // CUDA 12.0, min. ver. for half2 -> uint mask comparisons -#define CC_PASCAL 600 -#define MIN_CC_DP4A 610 // minimum compute capability for __dp4a, an intrinsic for byte-wise dot products -#define CC_VOLTA 700 -#define CC_TURING 750 -#define CC_AMPERE 800 -#define CC_OFFSET_AMD 1000000 -#define CC_RDNA1 (CC_OFFSET_AMD + 1010) -#define CC_RDNA2 (CC_OFFSET_AMD + 1030) -#define CC_RDNA3 (CC_OFFSET_AMD + 1100) -#define CC_QY1 210 -#define CC_QY2 220 +#define GGML_CUDA_CC_PASCAL 600 +#define GGML_CUDA_CC_DP4A 610 // minimum compute capability for __dp4a, an intrinsic for byte-wise dot products +#define GGML_CUDA_CC_VOLTA 700 +#define GGML_CUDA_CC_TURING 750 +#define GGML_CUDA_CC_AMPERE 800 +#define GGML_CUDA_CC_OFFSET_AMD 1000000 + +// GCN/CNDA, wave size is 64 +#define GGML_CUDA_CC_GCN4 (GGML_CUDA_CC_OFFSET_AMD + 803) // Tonga, Fiji, Polaris, minimum for fast fp16 +#define GGML_CUDA_CC_VEGA (GGML_CUDA_CC_OFFSET_AMD + 900) // Vega56/64, minimum for fp16 dual issue +#define GGML_CUDA_CC_VEGA20 (GGML_CUDA_CC_OFFSET_AMD + 906) // MI50/Radeon VII, minimum for dp4a +#define GGML_CUDA_CC_CDNA (GGML_CUDA_CC_OFFSET_AMD + 908) // MI100, minimum for MFMA, acc registers +#define GGML_CUDA_CC_CDNA2 (GGML_CUDA_CC_OFFSET_AMD + 910) // MI210, minimum acc register renameing +#define GGML_CUDA_CC_CDNA3 (GGML_CUDA_CC_OFFSET_AMD + 942) // MI300 + +// RNDA removes MFMA, dp4a, xnack, acc registers, wave size is 32 +#define GGML_CUDA_CC_RDNA1 (GGML_CUDA_CC_OFFSET_AMD + 1010) // RX 5000 +#define GGML_CUDA_CC_RDNA2 (GGML_CUDA_CC_OFFSET_AMD + 1030) // RX 6000, minimum for dp4a +#define GGML_CUDA_CC_RDNA3 (GGML_CUDA_CC_OFFSET_AMD + 1100) // RX 7000, minimum for WMMA + +#define GGML_CUDA_CC_QY1 210 +#define GGML_CUDA_CC_QY2 220 #define MATRIX_ROW_PADDING 512 // last row of quant. matrices is a multiple of this to avoid out-of-bounds memory accesses @@ -97,7 +108,7 @@ void ggml_cuda_error(const char * stmt, const char * func, const char * file, in #define CUBLAS_CHECK(err) CUDA_CHECK_GEN(err, CUBLAS_STATUS_SUCCESS, cublas_get_error_str) -#if !defined(GGML_USE_HIPBLAS) +#if !defined(GGML_USE_HIP) static const char * cu_get_error_str(CUresult err) { const char * err_str; cuGetErrorString(err, &err_str); @@ -120,50 +131,50 @@ typedef float dfloat; // dequantize float typedef float2 dfloat2; #endif // GGML_CUDA_F16 -#if (defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ >= CC_PASCAL +#if (defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL #define FP16_AVAILABLE -#endif // (defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ >= CC_PASCAL +#endif // (defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL #if defined(FP16_AVAILABLE) && __CUDA_ARCH__ != 610 #define FAST_FP16_AVAILABLE #endif // defined(FP16_AVAILABLE) && __CUDA_ARCH__ != 610 -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_VOLTA +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA #define FP16_MMA_AVAILABLE -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_VOLTA +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_TURING +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING #define INT8_MMA_AVAILABLE -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_TURING +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING -#if !(defined(GGML_USE_MUSA) && __MUSA_ARCH__ <= CC_QY1) +#if !(defined(GGML_USE_MUSA) && __MUSA_ARCH__ <= GGML_CUDA_CC_QY1) #define FLASH_ATTN_AVAILABLE -#endif // !(defined(GGML_USE_MUSA) && __MUSA_ARCH__ <= CC_QY1) +#endif // !(defined(GGML_USE_MUSA) && __MUSA_ARCH__ <= GGML_CUDA_CC_QY1) static constexpr bool fast_fp16_available(const int cc) { - return cc >= CC_PASCAL && cc != 610; + return cc >= GGML_CUDA_CC_PASCAL && cc != 610; } static constexpr bool fp16_mma_available(const int cc) { - return cc < CC_OFFSET_AMD && cc >= CC_VOLTA; + return cc < GGML_CUDA_CC_OFFSET_AMD && cc >= GGML_CUDA_CC_VOLTA; } static constexpr bool int8_mma_available(const int cc) { - return cc < CC_OFFSET_AMD && cc >= CC_TURING; + return cc < GGML_CUDA_CC_OFFSET_AMD && cc >= GGML_CUDA_CC_TURING; } [[noreturn]] static __device__ void no_device_code( const char * file_name, const int line, const char * function_name, const int arch, const char * arch_list) { -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) printf("%s:%d: ERROR: HIP kernel %s has no device code compatible with HIP arch %d.\n", file_name, line, function_name, arch); GGML_UNUSED(arch_list); #else printf("%s:%d: ERROR: CUDA kernel %s has no device code compatible with CUDA arch %d. ggml-cuda.cu was compiled for: %s\n", file_name, line, function_name, arch, arch_list); -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) __trap(); GGML_UNUSED(no_device_code); // suppress unused function warning @@ -176,30 +187,30 @@ static __device__ void no_device_code( #endif // __CUDA_ARCH__ static __device__ __forceinline__ int warp_reduce_sum(int x) { -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_AMPERE +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE return __reduce_add_sync(0xffffffff, x); #else #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - x += __shfl_xor_sync(0xffffffff, x, mask, 32); + for (int offset = 16; offset > 0; offset >>= 1) { + x += __shfl_xor_sync(0xffffffff, x, offset, 32); } return x; -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_AMPERE +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE } static __device__ __forceinline__ float warp_reduce_sum(float x) { #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - x += __shfl_xor_sync(0xffffffff, x, mask, 32); + for (int offset = 16; offset > 0; offset >>= 1) { + x += __shfl_xor_sync(0xffffffff, x, offset, 32); } return x; } static __device__ __forceinline__ float2 warp_reduce_sum(float2 a) { #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - a.x += __shfl_xor_sync(0xffffffff, a.x, mask, 32); - a.y += __shfl_xor_sync(0xffffffff, a.y, mask, 32); + for (int offset = 16; offset > 0; offset >>= 1) { + a.x += __shfl_xor_sync(0xffffffff, a.x, offset, 32); + a.y += __shfl_xor_sync(0xffffffff, a.y, offset, 32); } return a; } @@ -207,21 +218,21 @@ static __device__ __forceinline__ float2 warp_reduce_sum(float2 a) { static __device__ __forceinline__ half2 warp_reduce_sum(half2 a) { #ifdef FP16_AVAILABLE -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - const half2 a_other = __shfl_xor_sync(0xffffffff, a, mask, 32); + for (int offset = 16; offset > 0; offset >>= 1) { + const half2 a_other = __shfl_xor_sync(0xffffffff, a, offset, 32); reinterpret_cast(a.x) += __low2half(a_other); reinterpret_cast(a.y) += __high2half(a_other); } return a; #else #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - a = __hadd2(a, __shfl_xor_sync(0xffffffff, a, mask, 32)); + for (int offset = 16; offset > 0; offset >>= 1) { + a = __hadd2(a, __shfl_xor_sync(0xffffffff, a, offset, 32)); } return a; -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) #else NO_DEVICE_CODE; @@ -231,8 +242,8 @@ static __device__ __forceinline__ half2 warp_reduce_sum(half2 a) { static __device__ __forceinline__ float warp_reduce_max(float x) { #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - x = fmaxf(x, __shfl_xor_sync(0xffffffff, x, mask, 32)); + for (int offset = 16; offset > 0; offset >>= 1) { + x = fmaxf(x, __shfl_xor_sync(0xffffffff, x, offset, 32)); } return x; } @@ -240,11 +251,11 @@ static __device__ __forceinline__ float warp_reduce_max(float x) { static __device__ __forceinline__ half ggml_cuda_hmax(const half a, const half b) { #ifdef FP16_AVAILABLE -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && CUDART_VERSION < CUDART_HMAX +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && CUDART_VERSION < CUDART_HMAX return __float2half(fmaxf(__half2float(a), __half2float(b))); #else return __hmax(a, b); -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && CUDART_VERSION < CUDART_HMAX +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && CUDART_VERSION < CUDART_HMAX #else NO_DEVICE_CODE; @@ -254,7 +265,7 @@ static __device__ __forceinline__ half ggml_cuda_hmax(const half a, const half b } static __device__ __forceinline__ half2 ggml_cuda_hmax2(const half2 a, const half2 b) { -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) #if CUDART_VERSION >= CUDART_HMAX return __hmax2(a, b); @@ -269,20 +280,20 @@ static __device__ __forceinline__ half2 ggml_cuda_hmax2(const half2 a, const hal GGML_UNUSED(a); GGML_UNUSED(b); NO_DEVICE_CODE; -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) } static __device__ __forceinline__ half2 warp_reduce_max(half2 x) { -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL #pragma unroll - for (int mask = 16; mask > 0; mask >>= 1) { - x = ggml_cuda_hmax2(x, __shfl_xor_sync(0xffffffff, x, mask, 32)); + for (int offset = 16; offset > 0; offset >>= 1) { + x = ggml_cuda_hmax2(x, __shfl_xor_sync(0xffffffff, x, offset, 32)); } return x; #else GGML_UNUSED(x); NO_DEVICE_CODE; -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_PASCAL +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL } #if CUDART_VERSION < CUDART_HMASK @@ -294,7 +305,7 @@ static __device__ __forceinline__ uint32_t __hgt2_mask(const half2 a, const half #endif // CUDART_VERSION < CUDART_HMASK static __device__ __forceinline__ int ggml_cuda_dp4a(const int a, const int b, int c) { -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) #if defined(__gfx906__) || defined(__gfx908__) || defined(__gfx90a__) || defined(RDNA2) c = __builtin_amdgcn_sdot4(a, b, c, false); #elif defined(RDNA3) @@ -320,17 +331,17 @@ static __device__ __forceinline__ int ggml_cuda_dp4a(const int a, const int b, i #endif return c; -#else // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#else // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) -#if __CUDA_ARCH__ >= MIN_CC_DP4A +#if __CUDA_ARCH__ >= GGML_CUDA_CC_DP4A return __dp4a(a, b, c); -#else // __CUDA_ARCH__ >= MIN_CC_DP4A +#else // __CUDA_ARCH__ >= GGML_CUDA_CC_DP4A const int8_t * a8 = (const int8_t *) &a; const int8_t * b8 = (const int8_t *) &b; return c + a8[0]*b8[0] + a8[1]*b8[1] + a8[2]*b8[2] + a8[3]*b8[3]; -#endif // __CUDA_ARCH__ >= MIN_CC_DP4A +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_DP4A -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) } // TODO: move to ggml-common.h diff --git a/ggml/src/ggml-cuda/concat.cu b/ggml/src/ggml-cuda/concat.cu index dac10ec36..aafbaf803 100644 --- a/ggml/src/ggml-cuda/concat.cu +++ b/ggml/src/ggml-cuda/concat.cu @@ -94,7 +94,9 @@ static void concat_f32_cuda(const float * x, const float * y, float * dst, int n } // non-contiguous kernel (slow) -static __global__ void concat_f32_non_cont( +template +static __global__ void __launch_bounds__(CUDA_CONCAT_BLOCK_SIZE) + concat_f32_non_cont( const char * src0, const char * src1, char * dst, @@ -121,22 +123,28 @@ static __global__ void concat_f32_non_cont( uint64_t nb0, uint64_t nb1, uint64_t nb2, - uint64_t nb3, - int32_t dim) { + uint64_t nb3){ + static_assert(dim >= 0 && dim <= 3, "dim must be in [0, 3]"); + const int64_t i3 = blockIdx.z; const int64_t i2 = blockIdx.y; const int64_t i1 = blockIdx.x; - int64_t o[4] = {0, 0, 0, 0}; - o[dim] = dim == 0 ? ne00 : (dim == 1 ? ne01 : (dim == 2 ? ne02 : ne03)); - const float * x; - for (int i0 = threadIdx.x; i0 < ne0; i0 += blockDim.x) { + for (int64_t i0 = threadIdx.x; i0 < ne0; i0 += blockDim.x) { if (i0 < ne00 && i1 < ne01 && i2 < ne02 && i3 < ne03) { x = (const float *)(src0 + (i3 )*nb03 + (i2 )*nb02 + (i1 )*nb01 + (i0 )*nb00); } else { - x = (const float *)(src1 + (i3 - o[3])*nb13 + (i2 - o[2])*nb12 + (i1 - o[1])*nb11 + (i0 - o[0])*nb10); + if constexpr (dim == 0) { + x = (const float *) (src1 + i3 * nb13 + i2 * nb12 + i1 * nb11 + (i0 - ne00) * nb10); + } else if constexpr (dim == 1) { + x = (const float *) (src1 + i3 * nb13 + i2 * nb12 + (i1 - ne01) * nb11 + i0 * nb10); + } else if constexpr (dim == 2) { + x = (const float *) (src1 + i3 * nb13 + (i2 - ne02) * nb12 + i1 * nb11 + i0 * nb10); + } else if constexpr (dim == 3) { + x = (const float *) (src1 + (i3 - ne03) * nb13 + i2 * nb12 + i1 * nb11 + i0 * nb10); + } } float * y = (float *)(dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); @@ -182,15 +190,32 @@ void ggml_cuda_op_concat(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { } } else { dim3 grid_dim(dst->ne[1], dst->ne[2], dst->ne[3]); - concat_f32_non_cont<<>>( - (const char *)src0->data, - (const char *)src1->data, - ( char *)dst->data, + auto launch_kernel = [&](auto dim) { + concat_f32_non_cont<<>>( + (const char *) src0->data, (const char *) src1->data, (char *) dst->data, src0->ne[0], src0->ne[1], src0->ne[2], src0->ne[3], src0->nb[0], src0->nb[1], src0->nb[2], src0->nb[3], src1->ne[0], src1->ne[1], src1->ne[2], src1->ne[3], src1->nb[0], src1->nb[1], src1->nb[2], src1->nb[3], - dst->ne[0], dst->ne[1], dst->ne[2], dst->ne[3], - dst->nb[0], dst->nb[1], dst->nb[2], dst->nb[3], dim); + dst->ne[0], dst->ne[1], dst->ne[2], dst->ne[3], + dst->nb[0], dst->nb[1], dst->nb[2], dst->nb[3]); + }; + switch (dim) { + case 0: + launch_kernel(std::integral_constant{}); + break; + case 1: + launch_kernel(std::integral_constant{}); + break; + case 2: + launch_kernel(std::integral_constant{}); + break; + case 3: + launch_kernel(std::integral_constant{}); + break; + default: + GGML_ABORT("Invalid dim: %d", dim); + break; + } } } diff --git a/ggml/src/ggml-cuda/convert.cu b/ggml/src/ggml-cuda/convert.cu index c0a444707..5b0dfacef 100644 --- a/ggml/src/ggml-cuda/convert.cu +++ b/ggml/src/ggml-cuda/convert.cu @@ -26,7 +26,7 @@ static __global__ void dequantize_block(const void * __restrict__ vx, dst_t * __ template static __global__ void dequantize_block_q8_0_f16(const void * __restrict__ vx, half * __restrict__ y, const int64_t k) { -#if __CUDA_ARCH__ >= CC_PASCAL +#if __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL constexpr int nint = CUDA_Q8_0_NE_ALIGN/sizeof(int) + WARP_SIZE; const int64_t i0 = CUDA_Q8_0_NE_ALIGN*blockIdx.x; @@ -64,7 +64,7 @@ static __global__ void dequantize_block_q8_0_f16(const void * __restrict__ vx, h GGML_UNUSED(y); GGML_UNUSED(k); NO_DEVICE_CODE; -#endif // __CUDA_ARCH__ >= CC_PASCAL +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL } template @@ -599,7 +599,7 @@ to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type) { case GGML_TYPE_Q5_1: return dequantize_block_cuda; case GGML_TYPE_Q8_0: - if (ggml_cuda_info().devices[ggml_cuda_get_device()].cc >= CC_PASCAL) { + if (ggml_cuda_info().devices[ggml_cuda_get_device()].cc >= GGML_CUDA_CC_PASCAL) { return dequantize_block_q8_0_f16_cuda; } return dequantize_block_cuda; @@ -680,6 +680,8 @@ to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type) { return dequantize_row_iq3_s_cuda; case GGML_TYPE_F16: return convert_unary_cuda; + case GGML_TYPE_BF16: + return convert_unary_cuda; default: return nullptr; } diff --git a/ggml/src/ggml-cuda/count-equal.cu b/ggml/src/ggml-cuda/count-equal.cu index ffb053b10..08898115d 100644 --- a/ggml/src/ggml-cuda/count-equal.cu +++ b/ggml/src/ggml-cuda/count-equal.cu @@ -44,7 +44,7 @@ void ggml_cuda_count_equal(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { const int64_t ne = ggml_nelements(src0); GGML_ASSERT(ne < (1 << 30) && "atomicAdd implementation only supports int"); - const int64_t dne = GGML_PAD(ne / (4*nsm), CUDA_COUNT_EQUAL_CHUNK_SIZE); + const int64_t dne = GGML_PAD((ne + 4*nsm - 1) / (4*nsm), CUDA_COUNT_EQUAL_CHUNK_SIZE); CUDA_CHECK(cudaMemsetAsync(dst_d, 0, ggml_nbytes(dst), stream)); diff --git a/ggml/src/ggml-cuda/dmmv.cu b/ggml/src/ggml-cuda/dmmv.cu deleted file mode 100644 index 00e21b5d7..000000000 --- a/ggml/src/ggml-cuda/dmmv.cu +++ /dev/null @@ -1,699 +0,0 @@ -#include "dmmv.cuh" -#include "dequantize.cuh" -#include "convert.cuh" - -#ifndef K_QUANTS_PER_ITERATION -#define K_QUANTS_PER_ITERATION 2 -#else -static_assert(K_QUANTS_PER_ITERATION == 1 || K_QUANTS_PER_ITERATION == 2, "K_QUANTS_PER_ITERATION must be 1 or 2"); -#endif - -static __global__ void dequantize_mul_mat_vec_q2_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) { - - static_assert(16%K_QUANTS_PER_ITERATION == 0, "16 must be divisible by K_QUANTS_PER_ITERATION"); - - const int row = blockIdx.x*blockDim.y + threadIdx.y; - if (row > nrows) return; - - const int num_blocks_per_row = ncols / QK_K; - const int ib0 = row*num_blocks_per_row; - - const block_q2_K * x = (const block_q2_K *)vx + ib0; - - float tmp = 0; // partial sum for thread in warp - - const int tid = threadIdx.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...15 - const int ix = threadIdx.x%K_QUANTS_PER_ITERATION; // 0 or 0,1 - - const int step = 16/K_QUANTS_PER_ITERATION; - - const int im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const int in = tid - step*im; // 0...15 or 0...7 - - const int l0 = K_QUANTS_PER_ITERATION*in; // 0...15 or 0...14 in steps of 2 - const int q_offset = 32*im + l0; - const int s_offset = 8*im; - const int y_offset = 128*im + l0; - - uint32_t aux[4]; - const uint8_t * d = (const uint8_t *)aux; - const uint8_t * m = (const uint8_t *)(aux + 2); - - for (int i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - - const float * y = yy + i * QK_K + y_offset; - const uint8_t * q = x[i].qs + q_offset; - - const float dall = __low2half(x[i].dm); - const float dmin = __high2half(x[i].dm); - - const uint32_t * a = (const uint32_t *)(x[i].scales + s_offset); - aux[0] = a[0] & 0x0f0f0f0f; - aux[1] = a[1] & 0x0f0f0f0f; - aux[2] = (a[0] >> 4) & 0x0f0f0f0f; - aux[3] = (a[1] >> 4) & 0x0f0f0f0f; - - float sum1 = 0, sum2 = 0; - for (int l = 0; l < K_QUANTS_PER_ITERATION; ++l) { - sum1 += y[l+ 0] * d[0] * ((q[l+ 0] >> 0) & 3) - + y[l+32] * d[2] * ((q[l+ 0] >> 2) & 3) - + y[l+64] * d[4] * ((q[l+ 0] >> 4) & 3) - + y[l+96] * d[6] * ((q[l+ 0] >> 6) & 3) - + y[l+16] * d[1] * ((q[l+16] >> 0) & 3) - + y[l+48] * d[3] * ((q[l+16] >> 2) & 3) - + y[l+80] * d[5] * ((q[l+16] >> 4) & 3) - +y[l+112] * d[7] * ((q[l+16] >> 6) & 3); - sum2 += y[l+ 0] * m[0] + y[l+32] * m[2] + y[l+64] * m[4] + y[ l+96] * m[6] - + y[l+16] * m[1] + y[l+48] * m[3] + y[l+80] * m[5] + y[l+112] * m[7]; - - } - tmp += dall * sum1 - dmin * sum2; - - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (threadIdx.x == 0) { - dst[row] = tmp; - } -} - -static __global__ void dequantize_mul_mat_vec_q3_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) { - - const int row = blockIdx.x*blockDim.y + threadIdx.y; - if (row > nrows) return; - - const int num_blocks_per_row = ncols / QK_K; - const int ib0 = row*num_blocks_per_row; - - const block_q3_K * x = (const block_q3_K *)vx + ib0; - - float tmp = 0; // partial sum for thread in warp - - const uint16_t kmask1 = 0x0303; - const uint16_t kmask2 = 0x0f0f; - - const int tid = threadIdx.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const int ix = threadIdx.x%K_QUANTS_PER_ITERATION; // 0 or 0,1 - - const int n = K_QUANTS_PER_ITERATION; // iterations in the inner loop - const int step = 16/K_QUANTS_PER_ITERATION; - const int im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const int in = tid - step*im; // 0....15 or 0...7 - - const uint8_t m = 1 << (4*im); - - const int l0 = n*in; // 0...15 or 0...14 in steps of 2 - const int q_offset = 32*im + l0; - const int y_offset = 128*im + l0; - - uint16_t utmp[4]; - const int8_t * s = (const int8_t *)utmp; - - const uint16_t s_shift = 4*im; - - for (int i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - - const float * y = yy + i * QK_K + y_offset; - const uint8_t * q = x[i].qs + q_offset; - const uint8_t * h = x[i].hmask + l0; - - const uint16_t * a = (const uint16_t *)x[i].scales; - utmp[0] = ((a[0] >> s_shift) & kmask2) | (((a[4] >> (s_shift + 0)) & kmask1) << 4); - utmp[1] = ((a[1] >> s_shift) & kmask2) | (((a[5] >> (s_shift + 0)) & kmask1) << 4); - utmp[2] = ((a[2] >> s_shift) & kmask2) | (((a[4] >> (s_shift + 2)) & kmask1) << 4); - utmp[3] = ((a[3] >> s_shift) & kmask2) | (((a[5] >> (s_shift + 2)) & kmask1) << 4); - - const float d = x[i].d; - - float sum = 0; - for (int l = 0; l < n; ++l) { - sum += y[l+ 0] * (s[0] - 32) * (((q[l] >> 0) & 3) - (h[l] & (m << 0) ? 0 : 4)) - + y[l+32] * (s[2] - 32) * (((q[l] >> 2) & 3) - (h[l] & (m << 1) ? 0 : 4)) - + y[l+64] * (s[4] - 32) * (((q[l] >> 4) & 3) - (h[l] & (m << 2) ? 0 : 4)) - + y[l+96] * (s[6] - 32) * (((q[l] >> 6) & 3) - (h[l] & (m << 3) ? 0 : 4)); - sum += y[l+16] * (s[1] - 32) * (((q[l+16] >> 0) & 3) - (h[l+16] & (m << 0) ? 0 : 4)) - + y[l+48] * (s[3] - 32) * (((q[l+16] >> 2) & 3) - (h[l+16] & (m << 1) ? 0 : 4)) - + y[l+80] * (s[5] - 32) * (((q[l+16] >> 4) & 3) - (h[l+16] & (m << 2) ? 0 : 4)) - + y[l+112] * (s[7] - 32) * (((q[l+16] >> 6) & 3) - (h[l+16] & (m << 3) ? 0 : 4)); - } - tmp += d * sum; - - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (threadIdx.x == 0) { - dst[row] = tmp; - } -} - -static __global__ void dequantize_mul_mat_vec_q4_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) { - - const int row = blockIdx.x*blockDim.y + threadIdx.y; - if (row > nrows) return; - const int num_blocks_per_row = ncols / QK_K; - const int ib0 = row*num_blocks_per_row; - - const block_q4_K * x = (const block_q4_K *)vx + ib0; - - const uint16_t kmask1 = 0x3f3f; - const uint16_t kmask2 = 0x0f0f; - const uint16_t kmask3 = 0xc0c0; - - const int tid = threadIdx.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const int ix = threadIdx.x%K_QUANTS_PER_ITERATION; // 0 or 0,1 - - const int step = 8/K_QUANTS_PER_ITERATION; // 8 or 4 - - const int il = tid/step; // 0...3 - const int ir = tid - step*il; // 0...7 or 0...3 - const int n = 2 * K_QUANTS_PER_ITERATION; // 2 or 4 - - const int im = il/2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 - const int in = il%2; - - const int l0 = n*(2*ir + in); - const int q_offset = 32*im + l0; - const int y_offset = 64*im + l0; - - uint16_t aux[4]; - const uint8_t * sc = (const uint8_t *)aux; - -#if K_QUANTS_PER_ITERATION == 2 - uint32_t q32[4]; - const uint8_t * q4 = (const uint8_t *)q32; -#else - uint16_t q16[4]; - const uint8_t * q4 = (const uint8_t *)q16; -#endif - - float tmp = 0; // partial sum for thread in warp - - for (int i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - - const float * y1 = yy + i*QK_K + y_offset; - const float * y2 = y1 + 128; - - const float dall = __low2half(x[i].dm); - const float dmin = __high2half(x[i].dm); - - const uint16_t * a = (const uint16_t *)x[i].scales; - aux[0] = a[im+0] & kmask1; - aux[1] = a[im+2] & kmask1; - aux[2] = ((a[im+4] >> 0) & kmask2) | ((a[im+0] & kmask3) >> 2); - aux[3] = ((a[im+4] >> 4) & kmask2) | ((a[im+2] & kmask3) >> 2); - -#if K_QUANTS_PER_ITERATION == 2 - const uint32_t * q1 = (const uint32_t *)(x[i].qs + q_offset); - const uint32_t * q2 = q1 + 16; - - q32[0] = q1[0] & 0x0f0f0f0f; - q32[1] = q1[0] & 0xf0f0f0f0; - q32[2] = q2[0] & 0x0f0f0f0f; - q32[3] = q2[0] & 0xf0f0f0f0; - - float4 s = {0.f, 0.f, 0.f, 0.f}; - float smin = 0; - for (int l = 0; l < 4; ++l) { - s.x += y1[l] * q4[l+0]; s.y += y1[l+32] * q4[l+ 4]; - s.z += y2[l] * q4[l+8]; s.w += y2[l+32] * q4[l+12]; - smin += y1[l] * sc[2] + y1[l+32] * sc[3] + y2[l] * sc[6] + y2[l+32] * sc[7]; - } - tmp += dall * (s.x * sc[0] + s.y * sc[1] * 1.f/16.f + s.z * sc[4] + s.w * sc[5] * 1.f/16.f) - dmin * smin; -#else - const uint16_t * q1 = (const uint16_t *)(x[i].qs + q_offset); - const uint16_t * q2 = q1 + 32; - - q16[0] = q1[0] & 0x0f0f; - q16[1] = q1[0] & 0xf0f0; - q16[2] = q2[0] & 0x0f0f; - q16[3] = q2[0] & 0xf0f0; - - float4 s = {0.f, 0.f, 0.f, 0.f}; - float smin = 0; - for (int l = 0; l < 2; ++l) { - s.x += y1[l] * q4[l+0]; s.y += y1[l+32] * q4[l+2]; - s.z += y2[l] * q4[l+4]; s.w += y2[l+32] * q4[l+6]; - smin += y1[l] * sc[2] + y1[l+32] * sc[3] + y2[l] * sc[6] + y2[l+32] * sc[7]; - } - tmp += dall * (s.x * sc[0] + s.y * sc[1] * 1.f/16.f + s.z * sc[4] + s.w * sc[5] * 1.f/16.f) - dmin * smin; -#endif - - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (tid == 0) { - dst[row] = tmp; - } -} - -static __global__ void dequantize_mul_mat_vec_q5_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols) { - - const int row = blockIdx.x; - const int num_blocks_per_row = ncols / QK_K; - const int ib0 = row*num_blocks_per_row; - - const block_q5_K * x = (const block_q5_K *)vx + ib0; - - float tmp = 0; // partial sum for thread in warp - - const uint16_t kmask1 = 0x3f3f; - const uint16_t kmask2 = 0x0f0f; - const uint16_t kmask3 = 0xc0c0; - - const int tid = threadIdx.x/2; // 0...15 - const int ix = threadIdx.x%2; - - const int il = tid/4; // 0...3 - const int ir = tid - 4*il;// 0...3 - const int n = 2; - - const int im = il/2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 - const int in = il%2; - - const int l0 = n*(2*ir + in); - const int q_offset = 32*im + l0; - const int y_offset = 64*im + l0; - - const uint8_t hm1 = 1 << (2*im); - const uint8_t hm2 = hm1 << 4; - - uint16_t aux[4]; - const uint8_t * sc = (const uint8_t *)aux; - - uint16_t q16[8]; - const uint8_t * q4 = (const uint8_t *)q16; - - for (int i = ix; i < num_blocks_per_row; i += 2) { - - const uint8_t * ql1 = x[i].qs + q_offset; - const uint8_t * qh = x[i].qh + l0; - const float * y1 = yy + i*QK_K + y_offset; - const float * y2 = y1 + 128; - - const float dall = __low2half(x[i].dm); - const float dmin = __high2half(x[i].dm); - - const uint16_t * a = (const uint16_t *)x[i].scales; - aux[0] = a[im+0] & kmask1; - aux[1] = a[im+2] & kmask1; - aux[2] = ((a[im+4] >> 0) & kmask2) | ((a[im+0] & kmask3) >> 2); - aux[3] = ((a[im+4] >> 4) & kmask2) | ((a[im+2] & kmask3) >> 2); - - float4 sum = {0.f, 0.f, 0.f, 0.f}; - float smin = 0; - const uint16_t * q1 = (const uint16_t *)ql1; - const uint16_t * q2 = q1 + 32; - q16[0] = q1[0] & 0x0f0f; - q16[1] = q1[8] & 0x0f0f; - q16[2] = (q1[0] >> 4) & 0x0f0f; - q16[3] = (q1[8] >> 4) & 0x0f0f; - q16[4] = q2[0] & 0x0f0f; - q16[5] = q2[8] & 0x0f0f; - q16[6] = (q2[0] >> 4) & 0x0f0f; - q16[7] = (q2[8] >> 4) & 0x0f0f; - for (int l = 0; l < n; ++l) { - sum.x += y1[l+ 0] * (q4[l +0] + (qh[l+ 0] & (hm1 << 0) ? 16 : 0)) - + y1[l+16] * (q4[l +2] + (qh[l+16] & (hm1 << 0) ? 16 : 0)); - sum.y += y1[l+32] * (q4[l +4] + (qh[l+ 0] & (hm1 << 1) ? 16 : 0)) - + y1[l+48] * (q4[l +6] + (qh[l+16] & (hm1 << 1) ? 16 : 0)); - sum.z += y2[l+ 0] * (q4[l +8] + (qh[l+ 0] & (hm2 << 0) ? 16 : 0)) - + y2[l+16] * (q4[l+10] + (qh[l+16] & (hm2 << 0) ? 16 : 0)); - sum.w += y2[l+32] * (q4[l+12] + (qh[l+ 0] & (hm2 << 1) ? 16 : 0)) - + y2[l+48] * (q4[l+14] + (qh[l+16] & (hm2 << 1) ? 16 : 0)); - smin += (y1[l] + y1[l+16]) * sc[2] + (y1[l+32] + y1[l+48]) * sc[3] - + (y2[l] + y2[l+16]) * sc[6] + (y2[l+32] + y2[l+48]) * sc[7]; - } - tmp += dall * (sum.x * sc[0] + sum.y * sc[1] + sum.z * sc[4] + sum.w * sc[5]) - dmin * smin; - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (threadIdx.x == 0) { - dst[row] = tmp; - } -} - -static __global__ void dequantize_mul_mat_vec_q6_k(const void * __restrict__ vx, const float * __restrict__ yy, float * __restrict__ dst, const int ncols, int nrows) { - - static_assert(16%K_QUANTS_PER_ITERATION == 0, "16 must be divisible by K_QUANTS_PER_ITERATION"); - - const int row = blockIdx.x*blockDim.y + threadIdx.y; - if (row > nrows) return; - - const int num_blocks_per_row = ncols / QK_K; - const int ib0 = row*num_blocks_per_row; - - const block_q6_K * x = (const block_q6_K *)vx + ib0; - - const int tid = threadIdx.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const int ix = threadIdx.x%K_QUANTS_PER_ITERATION; // 0 or 0, 1 - - const int step = 16/K_QUANTS_PER_ITERATION; // 16 or 8 - - const int im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const int in = tid - step*im; // 0...15 or 0...7 - -#if K_QUANTS_PER_ITERATION == 1 - const int l0 = K_QUANTS_PER_ITERATION*in; // 0...15 - const int is = 0; -#else - const int l0 = 4 * in; // 0, 4, 8, ..., 28 - const int is = in / 4; -#endif - const int ql_offset = 64*im + l0; - const int qh_offset = 32*im + l0; - const int s_offset = 8*im + is; - const int y_offset = 128*im + l0; - - float tmp = 0; // partial sum for thread in warp - - for (int i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - - const float * y = yy + i * QK_K + y_offset; - const uint8_t * ql = x[i].ql + ql_offset; - const uint8_t * qh = x[i].qh + qh_offset; - const int8_t * s = x[i].scales + s_offset; - - const float d = x[i].d; - -#if K_QUANTS_PER_ITERATION == 1 - float sum = y[ 0] * s[0] * d * ((int8_t)((ql[ 0] & 0xF) | ((qh[ 0] & 0x03) << 4)) - 32) - + y[16] * s[1] * d * ((int8_t)((ql[16] & 0xF) | ((qh[16] & 0x03) << 4)) - 32) - + y[32] * s[2] * d * ((int8_t)((ql[32] & 0xF) | ((qh[ 0] & 0x0c) << 2)) - 32) - + y[48] * s[3] * d * ((int8_t)((ql[48] & 0xF) | ((qh[16] & 0x0c) << 2)) - 32) - + y[64] * s[4] * d * ((int8_t)((ql[ 0] >> 4) | ((qh[ 0] & 0x30) >> 0)) - 32) - + y[80] * s[5] * d * ((int8_t)((ql[16] >> 4) | ((qh[16] & 0x30) >> 0)) - 32) - + y[96] * s[6] * d * ((int8_t)((ql[32] >> 4) | ((qh[ 0] & 0xc0) >> 2)) - 32) - +y[112] * s[7] * d * ((int8_t)((ql[48] >> 4) | ((qh[16] & 0xc0) >> 2)) - 32); - tmp += sum; -#else - float sum = 0; - for (int l = 0; l < 4; ++l) { - sum += y[l+ 0] * s[0] * d * ((int8_t)((ql[l+ 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32) - + y[l+32] * s[2] * d * ((int8_t)((ql[l+32] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32) - + y[l+64] * s[4] * d * ((int8_t)((ql[l+ 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32) - + y[l+96] * s[6] * d * ((int8_t)((ql[l+32] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32); - } - tmp += sum; -#endif - - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (tid == 0) { - dst[row] = tmp; - } -} - -static __device__ void convert_f16(const void * vx, const int64_t ib, const int iqs, dfloat2 & v){ - const half * x = (const half *) vx; - // load 2 halfs into register in a single instruction - const half2 x_reg = *((half2 *) &(x[ib + iqs])); - // automatic half -> float type cast if dfloat == float - v.x = __low2float(x_reg); - v.y = __high2float(x_reg); -} - -static constexpr __device__ dequantize_kernel_t get_dequantize_kernel(ggml_type type) { - return type == GGML_TYPE_Q4_0 ? dequantize_q4_0 : - type == GGML_TYPE_Q4_1 ? dequantize_q4_1 : - type == GGML_TYPE_Q5_0 ? dequantize_q5_0 : - type == GGML_TYPE_Q5_1 ? dequantize_q5_1 : - type == GGML_TYPE_Q8_0 ? dequantize_q8_0 : - type == GGML_TYPE_F16 ? convert_f16 : - nullptr; -} - -template -static __global__ void dequantize_mul_mat_vec(const void * __restrict__ vx, const dfloat * __restrict__ y, float * __restrict__ dst, const int ncols, const int nrows) { - constexpr int qk = ggml_cuda_type_traits::qk; // quantized weights per x block - constexpr int qr = ggml_cuda_type_traits::qr; // number of quantized weights per data value in x block - constexpr dequantize_kernel_t dequantize_kernel = get_dequantize_kernel(type); - - const int64_t row = (int64_t)blockIdx.x*blockDim.y + threadIdx.y; - - if (row >= nrows) { - return; - } - - const int tid = threadIdx.x; - - const int iter_stride = 2*GGML_CUDA_DMMV_X; - const int vals_per_iter = iter_stride / WARP_SIZE; // num quantized vals per thread and i iter - const int y_offset = qr == 1 ? 1 : qk/2; - -// partial sum for each thread -#ifdef GGML_CUDA_F16 - half2 tmp = {0.0f, 0.0f}; // two sums for f16 to take advantage of half2 intrinsics -#else - float tmp = 0.0f; -#endif // GGML_CUDA_F16 - - for (int i = 0; i < ncols; i += iter_stride) { - const int col = i + vals_per_iter*tid; - const int64_t ib = ((int64_t)row*ncols + col)/qk; // x block index - const int iqs = (col%qk)/qr; // x quant index - const int iybs = col - col%qk; // y block start index - -// processing >2 values per i iter is faster for fast GPUs -#pragma unroll - for (int j = 0; j < vals_per_iter; j += 2) { - // process 2 vals per j iter - - // dequantize - // for qr = 2 the iqs needs to increase by 1 per j iter because 2 weights per data val - dfloat2 v; - dequantize_kernel(vx, ib, iqs + j/qr, v); - - // matrix multiplication - // for qr = 2 the y index needs to increase by 1 per j iter because of y_offset = qk/2 -#ifdef GGML_CUDA_F16 - if ( y_offset == 1 ) { - // load 2 dfloats into register in a single instruction - const dfloat2 y_reg = *((dfloat2 *) &(y[iybs + iqs + j/qr])); - tmp += __hmul2(v, y_reg); - } - else { - tmp += __hmul2(v, { - y[iybs + iqs + j/qr + 0], - y[iybs + iqs + j/qr + y_offset] - }); - } -#else - if ( y_offset == 1 ) { - // load 2 dfloats into register in a single instruction - const dfloat2 y_reg = *((dfloat2 *) &(y[iybs + iqs + j/qr])); - tmp += v.x * y_reg.x; - tmp += v.y * y_reg.y; - } - else { - tmp += v.x * y[iybs + iqs + j/qr + 0]; - tmp += v.y * y[iybs + iqs + j/qr + y_offset]; - } -#endif // GGML_CUDA_F16 - } - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (tid == 0) { -#ifdef GGML_CUDA_F16 - dst[row] = tmp.x + tmp.y; -#else - dst[row] = tmp; -#endif // GGML_CUDA_F16 - } -} - -static void dequantize_mul_mat_vec_q4_0_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - // the number of rows may exceed maximum grid size in the y or z dimensions, use the x dimension instead - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q4_1_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q5_0_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q5_1_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q8_0_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q2_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % QK_K == 0); - const int ny = 2; // very slightly faster than 1 even when K_QUANTS_PER_ITERATION = 2 - const int block_num_y = (nrows + ny - 1) / ny; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(32, ny, 1); - dequantize_mul_mat_vec_q2_k<<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q3_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % QK_K == 0); - const int ny = 2 / K_QUANTS_PER_ITERATION; - const int block_num_y = (nrows + ny - 1) / ny; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(32, ny, 1); - dequantize_mul_mat_vec_q3_k<<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q4_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % QK_K == 0); - const int ny = 2 / K_QUANTS_PER_ITERATION; - const int block_num_y = (nrows + ny - 1) / ny; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(32, ny, 1); - dequantize_mul_mat_vec_q4_k<<>>(vx, y, dst, ncols, nrows); -} - -static void dequantize_mul_mat_vec_q5_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % QK_K == 0); - const dim3 block_dims(32, 1, 1); - dequantize_mul_mat_vec_q5_k<<>>(vx, y, dst, ncols); -} - -static void dequantize_mul_mat_vec_q6_K_cuda(const void * vx, const float * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % QK_K == 0); - const int ny = 2 / K_QUANTS_PER_ITERATION; - const int block_num_y = (nrows + ny - 1) / ny; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(32, ny, 1); - dequantize_mul_mat_vec_q6_k<<>>(vx, y, dst, ncols, nrows); -} - -static void convert_mul_mat_vec_f16_cuda(const void * vx, const dfloat * y, float * dst, const int ncols, const int nrows, cudaStream_t stream) { - GGML_ASSERT(ncols % (GGML_CUDA_DMMV_X*2) == 0); - const int block_num_y = (nrows + GGML_CUDA_MMV_Y - 1) / GGML_CUDA_MMV_Y; - const dim3 block_nums(block_num_y, 1, 1); - const dim3 block_dims(WARP_SIZE, GGML_CUDA_MMV_Y, 1); - dequantize_mul_mat_vec - <<>>(vx, y, dst, ncols, nrows); -} - -void ggml_cuda_op_dequantize_mul_mat_vec( - ggml_backend_cuda_context & ctx, - const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const char * src0_dd_i, const float * src1_ddf_i, - const char * src1_ddq_i, float * dst_dd_i, const int64_t row_low, const int64_t row_high, const int64_t src1_ncols, - const int64_t src1_padded_row_size, cudaStream_t stream) { - GGML_UNUSED(ctx); - const int64_t ne00 = src0->ne[0]; - const int64_t row_diff = row_high - row_low; - - GGML_ASSERT(src1->type == GGML_TYPE_F32); - - // on some GPUs it is faster to convert src1 to half and to use half precision intrinsics -#ifdef GGML_CUDA_F16 - ggml_cuda_pool_alloc src1_dfloat_a(ctx.pool()); - half * src1_dfloat = nullptr; // dfloat == half - - bool src1_convert_f16 = - src0->type == GGML_TYPE_Q4_0 || src0->type == GGML_TYPE_Q4_1 || - src0->type == GGML_TYPE_Q5_0 || src0->type == GGML_TYPE_Q5_1 || - src0->type == GGML_TYPE_Q8_0 || src0->type == GGML_TYPE_F16; - - if (src1_convert_f16) { - src1_dfloat = src1_dfloat_a.alloc(ne00); - const to_fp16_cuda_t to_fp16_cuda = ggml_get_to_fp16_cuda(src1->type); - GGML_ASSERT(to_fp16_cuda != nullptr); - to_fp16_cuda(src1_ddf_i, src1_dfloat, ne00, stream); - } -#else - const dfloat * src1_dfloat = (const dfloat *) src1_ddf_i; // dfloat == float, no conversion -#endif // GGML_CUDA_F16 - - switch (src0->type) { - case GGML_TYPE_Q4_0: - dequantize_mul_mat_vec_q4_0_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q4_1: - dequantize_mul_mat_vec_q4_1_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q5_0: - dequantize_mul_mat_vec_q5_0_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q5_1: - dequantize_mul_mat_vec_q5_1_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q8_0: - dequantize_mul_mat_vec_q8_0_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q2_K: - dequantize_mul_mat_vec_q2_K_cuda(src0_dd_i, src1_ddf_i, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q3_K: - dequantize_mul_mat_vec_q3_K_cuda(src0_dd_i, src1_ddf_i, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q4_K: - dequantize_mul_mat_vec_q4_K_cuda(src0_dd_i, src1_ddf_i, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q5_K: - dequantize_mul_mat_vec_q5_K_cuda(src0_dd_i, src1_ddf_i, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_Q6_K: - dequantize_mul_mat_vec_q6_K_cuda(src0_dd_i, src1_ddf_i, dst_dd_i, ne00, row_diff, stream); - break; - case GGML_TYPE_F16: - convert_mul_mat_vec_f16_cuda(src0_dd_i, src1_dfloat, dst_dd_i, ne00, row_diff, stream); - break; - default: - GGML_ABORT("fatal error"); - break; - } - - GGML_UNUSED(src1); - GGML_UNUSED(dst); - GGML_UNUSED(src1_ddq_i); - GGML_UNUSED(src1_ncols); - GGML_UNUSED(src1_padded_row_size); -} - -bool ggml_cuda_dmmv_type_supported(ggml_type src0_type) { - return src0_type == GGML_TYPE_Q4_0 || src0_type == GGML_TYPE_Q4_1 || - src0_type == GGML_TYPE_Q5_0 || src0_type == GGML_TYPE_Q5_1 || - src0_type == GGML_TYPE_Q8_0 || src0_type == GGML_TYPE_Q2_K || - src0_type == GGML_TYPE_Q3_K || src0_type == GGML_TYPE_Q4_K || - src0_type == GGML_TYPE_Q5_K || src0_type == GGML_TYPE_Q6_K || - src0_type == GGML_TYPE_F16; -} diff --git a/ggml/src/ggml-cuda/fattn-common.cuh b/ggml/src/ggml-cuda/fattn-common.cuh index 1fb5c09c3..ee9752da6 100644 --- a/ggml/src/ggml-cuda/fattn-common.cuh +++ b/ggml/src/ggml-cuda/fattn-common.cuh @@ -517,9 +517,9 @@ constexpr __device__ dequantize_1_f32_t get_dequantize_1_f32(ggml_type type_V) { } template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_combine_results( const float * __restrict__ VKQ_parts, const float2 * __restrict__ VKQ_meta, diff --git a/ggml/src/ggml-cuda/fattn-tile-f16.cu b/ggml/src/ggml-cuda/fattn-tile-f16.cu index 5af02c7ec..4d314dacb 100644 --- a/ggml/src/ggml-cuda/fattn-tile-f16.cu +++ b/ggml/src/ggml-cuda/fattn-tile-f16.cu @@ -5,9 +5,9 @@ #define FATTN_KQ_STRIDE_TILE_F16 64 template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(nwarps*WARP_SIZE, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_tile_ext_f16( const char * __restrict__ Q, const char * __restrict__ K, diff --git a/ggml/src/ggml-cuda/fattn-tile-f32.cu b/ggml/src/ggml-cuda/fattn-tile-f32.cu index f402195ce..bb3360447 100644 --- a/ggml/src/ggml-cuda/fattn-tile-f32.cu +++ b/ggml/src/ggml-cuda/fattn-tile-f32.cu @@ -5,9 +5,9 @@ #define FATTN_KQ_STRIDE_TILE_F32 32 template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(nwarps*WARP_SIZE, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_tile_ext_f32( const char * __restrict__ Q, const char * __restrict__ K, diff --git a/ggml/src/ggml-cuda/fattn-vec-f16.cuh b/ggml/src/ggml-cuda/fattn-vec-f16.cuh index 2ed6509ac..34a2992c7 100644 --- a/ggml/src/ggml-cuda/fattn-vec-f16.cuh +++ b/ggml/src/ggml-cuda/fattn-vec-f16.cuh @@ -2,9 +2,9 @@ #include "fattn-common.cuh" template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_vec_ext_f16( const char * __restrict__ Q, const char * __restrict__ K, @@ -220,7 +220,6 @@ static __global__ void flash_attn_vec_ext_f16( for (int j = 0; j < ncols; ++j) { half kqmax_new_j = ncols == 1 ? kqmax_new : kqmax_new_arr[j]; - kqmax_new_j = warp_reduce_max(kqmax_new_j); if (threadIdx.x == 0) { kqmax_shared[j][threadIdx.y] = kqmax_new_j; } diff --git a/ggml/src/ggml-cuda/fattn-vec-f32.cuh b/ggml/src/ggml-cuda/fattn-vec-f32.cuh index bf5125902..a28fc8b7f 100644 --- a/ggml/src/ggml-cuda/fattn-vec-f32.cuh +++ b/ggml/src/ggml-cuda/fattn-vec-f32.cuh @@ -2,9 +2,9 @@ #include "fattn-common.cuh" template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_vec_ext_f32( const char * __restrict__ Q, const char * __restrict__ K, @@ -206,7 +206,6 @@ static __global__ void flash_attn_vec_ext_f32( for (int j = 0; j < ncols; ++j) { float kqmax_new_j = kqmax_new_arr[j]; - kqmax_new_j = warp_reduce_max(kqmax_new_j); if (threadIdx.x == 0) { kqmax_shared[j][threadIdx.y] = kqmax_new_j; } diff --git a/ggml/src/ggml-cuda/fattn-wmma-f16.cuh b/ggml/src/ggml-cuda/fattn-wmma-f16.cuh index b10d19d93..860d0e6dc 100644 --- a/ggml/src/ggml-cuda/fattn-wmma-f16.cuh +++ b/ggml/src/ggml-cuda/fattn-wmma-f16.cuh @@ -7,9 +7,9 @@ // D == head size, VKQ_stride == num VKQ rows calculated in parallel: template -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) __launch_bounds__(nwarps*WARP_SIZE, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void flash_attn_ext_f16( const char * __restrict__ Q, const char * __restrict__ K, diff --git a/ggml/src/ggml-cuda/fattn.cu b/ggml/src/ggml-cuda/fattn.cu index 83e5589a1..0b26b0f8e 100644 --- a/ggml/src/ggml-cuda/fattn.cu +++ b/ggml/src/ggml-cuda/fattn.cu @@ -13,9 +13,9 @@ static void ggml_cuda_flash_attn_ext_wmma_f16(ggml_backend_cuda_context & ctx, g const ggml_tensor * KQV = dst; const ggml_tensor * Q = dst->src[0]; - const int32_t precision = KQV->op_params[3]; + const enum ggml_prec prec = ggml_flash_attn_ext_get_prec(KQV); - if (precision != GGML_PREC_DEFAULT) { + if (prec != GGML_PREC_DEFAULT) { if (Q->ne[1] <= 32 || Q->ne[0] > 128) { constexpr int cols_per_block = 16; switch (Q->ne[0]) { @@ -301,11 +301,11 @@ void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst ggml_cuda_set_device(ctx.device); const int cc = ggml_cuda_info().devices[ggml_cuda_get_device()].cc; - const int32_t precision = KQV->op_params[3]; + const enum ggml_prec prec = ggml_flash_attn_ext_get_prec(KQV); // On AMD the tile kernels perform poorly, use the vec kernel instead: - if (cc >= CC_OFFSET_AMD) { - if (precision == GGML_PREC_DEFAULT && fast_fp16_available(cc)) { + if (cc >= GGML_CUDA_CC_OFFSET_AMD) { + if (prec == GGML_PREC_DEFAULT && fast_fp16_available(cc)) { ggml_cuda_flash_attn_ext_vec_f16(ctx, dst); } else { ggml_cuda_flash_attn_ext_vec_f32(ctx, dst); @@ -332,7 +332,7 @@ void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst } if (Q->ne[1] == 1 && Q->ne[0] % (2*WARP_SIZE) == 0) { - if (precision == GGML_PREC_DEFAULT) { + if (prec == GGML_PREC_DEFAULT) { ggml_cuda_flash_attn_ext_vec_f16(ctx, dst); return; } else if(Q->ne[0] <= 128) { diff --git a/ggml/src/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu similarity index 88% rename from ggml/src/ggml-cuda.cu rename to ggml/src/ggml-cuda/ggml-cuda.cu index e27c8e87d..9118edc72 100644 --- a/ggml/src/ggml-cuda.cu +++ b/ggml/src/ggml-cuda/ggml-cuda.cu @@ -16,11 +16,11 @@ #include "ggml-cuda/cpy.cuh" #include "ggml-cuda/cross-entropy-loss.cuh" #include "ggml-cuda/diagmask.cuh" -#include "ggml-cuda/dmmv.cuh" #include "ggml-cuda/fattn.cuh" #include "ggml-cuda/getrows.cuh" #include "ggml-cuda/im2col.cuh" #include "ggml-cuda/mmq.cuh" +#include "ggml-cuda/mmv.cuh" #include "ggml-cuda/mmvq.cuh" #include "ggml-cuda/norm.cuh" #include "ggml-cuda/opt-step-adamw.cuh" @@ -37,6 +37,7 @@ #include "ggml-cuda/unary.cuh" #include "ggml-cuda/upscale.cuh" #include "ggml-cuda/wkv6.cuh" +#include "ggml-cuda/gla.cuh" #include #include @@ -91,7 +92,7 @@ int ggml_cuda_get_device() { static cudaError_t ggml_cuda_device_malloc(void ** ptr, size_t size, int device) { ggml_cuda_set_device(device); -#if defined(GGML_USE_HIPBLAS) && defined(GGML_HIP_UMA) +#if defined(GGML_USE_HIP) && defined(GGML_HIP_UMA) auto res = hipMallocManaged(ptr, size); if (res == hipSuccess) { // if error we "need" to know why... @@ -100,7 +101,7 @@ static cudaError_t ggml_cuda_device_malloc(void ** ptr, size_t size, int device) return res; #else -#if !defined(GGML_USE_HIPBLAS) +#if !defined(GGML_USE_HIP) cudaError_t err; if (getenv("GGML_CUDA_ENABLE_UNIFIED_MEMORY") != nullptr) { @@ -113,7 +114,7 @@ static cudaError_t ggml_cuda_device_malloc(void ** ptr, size_t size, int device) return err; #else return cudaMalloc(ptr, size); -#endif // !defined(GGML_USE_HIPBLAS) +#endif // !defined(GGML_USE_HIP) #endif } @@ -151,7 +152,7 @@ static ggml_cuda_device_info ggml_cuda_init() { for (int id = 0; id < info.device_count; ++id) { int device_vmm = 0; -#if !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#if !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) CUdevice device; CU_CHECK(cuDeviceGet(&device, id)); CU_CHECK(cuDeviceGetAttribute(&device_vmm, CU_DEVICE_ATTRIBUTE_VIRTUAL_MEMORY_MANAGEMENT_SUPPORTED, device)); @@ -163,7 +164,7 @@ static ggml_cuda_device_info ggml_cuda_init() { alloc_prop.location.id = id; CU_CHECK(cuMemGetAllocationGranularity(&info.devices[id].vmm_granularity, &alloc_prop, CU_MEM_ALLOC_GRANULARITY_RECOMMENDED)); } -#endif // !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#endif // !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) info.devices[id].vmm = !!device_vmm; cudaDeviceProp prop; @@ -175,13 +176,13 @@ static ggml_cuda_device_info ggml_cuda_init() { info.devices[id].nsm = prop.multiProcessorCount; info.devices[id].smpb = prop.sharedMemPerBlock; -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) info.devices[id].smpbo = prop.sharedMemPerBlock; - info.devices[id].cc = 100*prop.major + 10*prop.minor + CC_OFFSET_AMD; + info.devices[id].cc = 100*prop.major + 10*prop.minor + GGML_CUDA_CC_OFFSET_AMD; #else info.devices[id].smpbo = prop.sharedMemPerBlockOptin; info.devices[id].cc = 100*prop.major + 10*prop.minor; -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) } for (int id = 0; id < info.device_count; ++id) { @@ -299,7 +300,7 @@ struct ggml_cuda_pool_leg : public ggml_cuda_pool { }; // pool with virtual memory -#if !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#if !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) struct ggml_cuda_pool_vmm : public ggml_cuda_pool { static const size_t CUDA_POOL_VMM_MAX_SIZE = 1ull << 35; // 32 GB @@ -393,14 +394,14 @@ struct ggml_cuda_pool_vmm : public ggml_cuda_pool { GGML_ASSERT(ptr == (void *) (pool_addr + pool_used)); } }; -#endif // !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#endif // !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) std::unique_ptr ggml_backend_cuda_context::new_pool_for_device(int device) { -#if !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#if !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) if (ggml_cuda_info().devices[device].vmm) { return std::unique_ptr(new ggml_cuda_pool_vmm(device)); } -#endif // !defined(GGML_USE_HIPBLAS) && !defined(GGML_CUDA_NO_VMM) +#endif // !defined(GGML_USE_HIP) && !defined(GGML_CUDA_NO_VMM) return std::unique_ptr(new ggml_cuda_pool_leg(device)); } @@ -1020,114 +1021,6 @@ typedef void (*ggml_cuda_op_mul_mat_t)( #define MUL_MAT_SRC1_COL_STRIDE 128 -static __global__ void mul_mat_p021_f16_f32( - const void * __restrict__ vx, const float * __restrict__ y, float * __restrict__ dst, - const int ncols_x, const int nrows_x, const int nchannels_x, const int nchannels_y) { - - const half * x = (const half *) vx; - - const int row_x = blockDim.y*blockIdx.y + threadIdx.y; - const int channel = blockDim.z*blockIdx.z + threadIdx.z; - const int channel_x = channel / (nchannels_y / nchannels_x); - - const int nrows_y = ncols_x; - const int nrows_dst = nrows_x; - const int row_dst = row_x; - - float tmp = 0.0f; - - for (int col_x0 = 0; col_x0 < ncols_x; col_x0 += blockDim.x) { - const int col_x = col_x0 + threadIdx.x; - - if (col_x >= ncols_x) { - break; - } - - // x is transposed and permuted - const int ix = row_x*nchannels_x*ncols_x + channel_x*ncols_x + col_x; - const float xi = __half2float(x[ix]); - - const int row_y = col_x; - - // y is not transposed but permuted - const int iy = channel*nrows_y + row_y; - - tmp += xi * y[iy]; - } - - // dst is not transposed and not permuted - const int idst = channel*nrows_dst + row_dst; - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (threadIdx.x == 0) { - dst[idst] = tmp; - } -} - -static __global__ void mul_mat_vec_nc_f16_f32( // nc == non-contiguous - const void * __restrict__ vx, const float * __restrict__ y, float * __restrict__ dst, const int ncols_x, const int nrows_x, - const int row_stride_x, const int channel_stride_x, const int channel_x_divisor) { - - const half * x = (const half *) vx; - - const int row_x = blockDim.y*blockIdx.y + threadIdx.y; - const int channel = blockDim.z*blockIdx.z + threadIdx.z; - const int channel_x = channel / channel_x_divisor; - - const int nrows_y = ncols_x; - const int nrows_dst = nrows_x; - const int row_dst = row_x; - - const int idst = channel*nrows_dst + row_dst; - - float tmp = 0.0f; - - for (int col_x0 = 0; col_x0 < ncols_x; col_x0 += blockDim.x) { - const int col_x = col_x0 + threadIdx.x; - - if (col_x >= ncols_x) { - break; - } - - const int row_y = col_x; - - const int ix = channel_x*channel_stride_x + row_x*row_stride_x + col_x; - const int iy = channel*nrows_y + row_y; - - const float xi = __half2float(x[ix]); - - tmp += xi * y[iy]; - } - - // sum up partial sums and write back result - tmp = warp_reduce_sum(tmp); - - if (threadIdx.x == 0) { - dst[idst] = tmp; - } -} - -static void ggml_mul_mat_p021_f16_f32_cuda( - const void * vx, const float * y, float * dst, const int ncols_x, const int nrows_x, - const int nchannels_x, const int nchannels_y, cudaStream_t stream) { - - const dim3 block_nums(1, nrows_x, nchannels_y); - const dim3 block_dims(WARP_SIZE, 1, 1); - mul_mat_p021_f16_f32<<>>(vx, y, dst, ncols_x, nrows_x, nchannels_x, nchannels_y); -} - -static void ggml_mul_mat_vec_nc_f16_f32_cuda( - const void * vx, const float * y, float * dst, const int ncols_x, const int nrows_x, const int row_stride_x, - const int nchannels_x, const int nchannels_y, const int channel_stride_x, cudaStream_t stream) { - - const dim3 block_nums(1, nrows_x, nchannels_y); - const dim3 block_dims(WARP_SIZE, 1, 1); - mul_mat_vec_nc_f16_f32<<>> - (vx, y, dst, ncols_x, nrows_x, row_stride_x, channel_stride_x, nchannels_y/nchannels_x); -} - static cudaError_t ggml_cuda_cpy_tensor_2d( void * dst, const struct ggml_tensor * src, int64_t i3, int64_t i2, int64_t i1_low, int64_t i1_high, cudaStream_t stream) { @@ -1189,7 +1082,7 @@ static void ggml_cuda_op_mul_mat_cublas( const int compute_capability = ggml_cuda_info().devices[id].cc; - if (compute_capability >= CC_VOLTA && (src0->type == GGML_TYPE_F16 || ggml_is_quantized(src0->type)) && ggml_is_contiguous(src0) && row_diff == src0->ne[1] && dst->op_params[0] == GGML_PREC_DEFAULT) { + if (compute_capability >= GGML_CUDA_CC_VOLTA && (src0->type == GGML_TYPE_F16 || ggml_is_quantized(src0->type)) && ggml_is_contiguous(src0) && row_diff == src0->ne[1] && dst->op_params[0] == GGML_PREC_DEFAULT) { // convert src0 and src1 to fp16, multiply as fp16, convert dst to fp32 ggml_cuda_pool_alloc src0_as_f16(ctx.pool(id)); if (src0->type != GGML_TYPE_F16) { @@ -1215,6 +1108,11 @@ static void ggml_cuda_op_mul_mat_cublas( const half alpha_f16 = 1.0f; const half beta_f16 = 0.0f; + cublasComputeType_t cu_compute_type = CUBLAS_COMPUTE_16F; + if (ggml_cuda_info().devices[ctx.device].cc == GGML_CUDA_CC_CDNA) { + cu_compute_type = CUBLAS_COMPUTE_32F; + } + CUBLAS_CHECK(cublasSetStream(ctx.cublas_handle(id), stream)); CUBLAS_CHECK( cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N, @@ -1222,7 +1120,7 @@ static void ggml_cuda_op_mul_mat_cublas( &alpha_f16, src0_ptr, CUDA_R_16F, ne00, src1_ptr, CUDA_R_16F, ne10, &beta_f16, dst_f16.get(), CUDA_R_16F, ldc, - CUBLAS_COMPUTE_16F, + cu_compute_type, CUBLAS_GEMM_DEFAULT_TENSOR_OP)); const to_fp32_cuda_t to_fp32_cuda = ggml_get_to_fp32_cuda(GGML_TYPE_F16); @@ -1325,7 +1223,7 @@ static void ggml_cuda_set_peer_access(const int n_tokens, int main_device) { static cudaError_t ggml_cuda_Memcpy2DPeerAsync( void * dst, int dstDevice, size_t dpitch, void * src, int srcDevice, size_t spitch, size_t width, size_t height, cudaStream_t stream) { -#if !defined(GGML_USE_HIPBLAS) && !defined(GGML_USE_MUSA) +#if !defined(GGML_USE_HIP) && !defined(GGML_USE_MUSA) // cudaMemcpy2DAsync may fail with copies between vmm pools of different devices cudaMemcpy3DPeerParms p = {}; p.dstDevice = dstDevice; @@ -1339,7 +1237,7 @@ static cudaError_t ggml_cuda_Memcpy2DPeerAsync( GGML_UNUSED(dstDevice); GGML_UNUSED(srcDevice); return cudaMemcpy2DAsync(dst, dpitch, src, spitch, width, height, cudaMemcpyDeviceToDevice, stream); -#endif // !defined(GGML_USE_HIPBLAS) && !defined(GGML_USE_MUSA) +#endif // !defined(GGML_USE_HIP) && !defined(GGML_USE_MUSA) } static void ggml_cuda_op_mul_mat( @@ -1654,58 +1552,6 @@ static void ggml_cuda_op_mul_mat( } } -static void ggml_cuda_mul_mat_vec_p021(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_permuted(src0) && ggml_is_permuted(src1)); - GGML_ASSERT(ggml_backend_buffer_is_cuda(src0->buffer)); - GGML_ASSERT(src0->nb[0] <= src0->nb[1] && src0->nb[2] <= src0->nb[3]); // 0213 permutation - GGML_ASSERT(src1->nb[0] <= src1->nb[1] && src1->nb[2] <= src1->nb[3]); // 0213 permutation - GGML_ASSERT(src0->type == GGML_TYPE_F16); - GGML_ASSERT(src1->type == GGML_TYPE_F32); - - const int64_t ne00 = src0->ne[0]; - const int64_t ne01 = src0->ne[1]; - const int64_t ne02 = src0->ne[2]; - - const int64_t ne12 = src1->ne[2]; - - cudaStream_t main_stream = ctx.stream(); - - void * src0_ddq = src0->data; - float * src1_ddf = (float *) src1->data; - float * dst_ddf = (float *) dst->data; - - ggml_mul_mat_p021_f16_f32_cuda(src0_ddq, src1_ddf, dst_ddf, ne00, ne01, ne02, ne12, main_stream); -} - -static void ggml_cuda_mul_mat_vec_nc(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(!ggml_is_transposed(src0)); - GGML_ASSERT(!ggml_is_transposed(src1)); - GGML_ASSERT(!ggml_is_permuted(src0)); - GGML_ASSERT(ggml_backend_buffer_is_cuda(src0->buffer)); - GGML_ASSERT(src0->type == GGML_TYPE_F16); - GGML_ASSERT(src1->type == GGML_TYPE_F32); - - const int64_t ne00 = src0->ne[0]; - const int64_t ne01 = src0->ne[1]; - const int64_t ne02 = src0->ne[2]; - - const int64_t nb01 = src0->nb[1]; - const int64_t nb02 = src0->nb[2]; - - const int64_t ne12 = src1->ne[2]; - - cudaStream_t main_stream = ctx.stream(); - - void * src0_ddq = src0->data; - float * src1_ddf = (float *) src1->data; - float * dst_ddf = (float *) dst->data; - - const int64_t row_stride_x = nb01 / sizeof(half); - const int64_t channel_stride_x = nb02 / sizeof(half); - - ggml_mul_mat_vec_nc_f16_f32_cuda(src0_ddq, src1_ddf, dst_ddf, ne00, ne01, row_stride_x, ne02, ne12, channel_stride_x, main_stream); -} - static __global__ void k_compute_batched_ptrs( const half * src0_as_f16, const half * src1_as_f16, char * dst, const void ** ptrs_src, void ** ptrs_dst, @@ -1767,6 +1613,10 @@ static void ggml_cuda_mul_mat_batched_cublas(ggml_backend_cuda_context & ctx, co cublasComputeType_t cu_compute_type = CUBLAS_COMPUTE_16F; cudaDataType_t cu_data_type = CUDA_R_16F; + if (ggml_cuda_info().devices[ctx.device].cc == GGML_CUDA_CC_CDNA) { + cu_compute_type = CUBLAS_COMPUTE_32F; + } + // dst strides size_t nbd2 = dst->nb[2]; size_t nbd3 = dst->nb[3]; @@ -1879,21 +1729,17 @@ static void ggml_cuda_mul_mat_batched_cublas(ggml_backend_cuda_context & ctx, co static void ggml_cuda_mul_mat(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { const bool split = ggml_backend_buft_is_cuda_split(src0->buffer->buft); - bool use_dequantize_mul_mat_vec = ggml_cuda_dmmv_type_supported(src0->type) + bool use_mul_mat_vec = (src0->type == GGML_TYPE_F16 || src0->type == GGML_TYPE_BF16) && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32 - && src0->ne[0] % (GGML_CUDA_DMMV_X*2) == 0 && src1->ne[1] == 1; - bool use_mul_mat_vec_q = ggml_is_quantized(src0->type) + && src0->ne[0] % 2 == 0 && src1->ne[1] == 1; + bool use_mul_mat_vec_q = ggml_is_quantized(src0->type) && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32 && src1->ne[1] <= MMVQ_MAX_BATCH_SIZE; - bool use_mul_mat_q = ggml_is_quantized(src0->type) + bool use_mul_mat_q = ggml_is_quantized(src0->type) && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32; - // if mmvq is available it's a better choice than dmmv: -#ifndef GGML_CUDA_FORCE_DMMV - use_dequantize_mul_mat_vec = use_dequantize_mul_mat_vec && !use_mul_mat_vec_q; -#endif // GGML_CUDA_FORCE_DMMV - - bool any_gpus_with_slow_fp16 = false; + bool any_gpus_with_slow_fp16 = false; + bool any_gpus_without_fp16_mma = false; if (split) { ggml_backend_cuda_split_buffer_type_context * buft_ctx = (ggml_backend_cuda_split_buffer_type_context *) src0->buffer->buft->context; @@ -1904,14 +1750,16 @@ static void ggml_cuda_mul_mat(ggml_backend_cuda_context & ctx, const ggml_tensor continue; } - const int cc = ggml_cuda_info().devices[id].cc; - use_mul_mat_q = use_mul_mat_q && ggml_cuda_should_use_mmq(src0->type, cc, src1->ne[1]); - any_gpus_with_slow_fp16 = any_gpus_with_slow_fp16 || !fast_fp16_available(cc); + const int cc = ggml_cuda_info().devices[id].cc; + use_mul_mat_q = use_mul_mat_q && ggml_cuda_should_use_mmq(src0->type, cc, src1->ne[1]); + any_gpus_with_slow_fp16 = any_gpus_with_slow_fp16 || !fast_fp16_available(cc); + any_gpus_without_fp16_mma = any_gpus_without_fp16_mma || !fp16_mma_available(cc); } } else { - const int cc = ggml_cuda_info().devices[ctx.device].cc; - use_mul_mat_q = use_mul_mat_q && ggml_cuda_should_use_mmq(src0->type, cc, src1->ne[1]); - any_gpus_with_slow_fp16 = any_gpus_with_slow_fp16 || !fast_fp16_available(cc); + const int cc = ggml_cuda_info().devices[ctx.device].cc; + use_mul_mat_q = use_mul_mat_q && ggml_cuda_should_use_mmq(src0->type, cc, src1->ne[1]); + any_gpus_with_slow_fp16 = any_gpus_with_slow_fp16 || !fast_fp16_available(cc); + any_gpus_without_fp16_mma = any_gpus_without_fp16_mma || !fp16_mma_available(cc); } // debug helpers @@ -1922,18 +1770,16 @@ static void ggml_cuda_mul_mat(ggml_backend_cuda_context & ctx, const ggml_tensor //printf("src0 is contiguous %d, transposed %d, type = %s, name = %s\n", ggml_is_contiguous(src0), ggml_is_transposed(src0), ggml_type_name(src0->type), src0->name); //printf("src1 is contiguous %d, transposed %d, type = %s, name = %s\n", ggml_is_contiguous(src1), ggml_is_transposed(src1), ggml_type_name(src1->type), src1->name); - if (!split && any_gpus_with_slow_fp16 && src0->type == GGML_TYPE_F16 && ggml_is_permuted(src0) && ggml_is_permuted(src1) && src1->ne[1] == 1) { - // FP32 precision KQ single-batch for batch size 1 without FlashAttention - ggml_cuda_mul_mat_vec_p021(ctx, src0, src1, dst); - } else if (!split && any_gpus_with_slow_fp16 && src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && src1->ne[1] == 1) { - // FP32 precision KQV single-batch for batch size 1 without FlashAttention - ggml_cuda_mul_mat_vec_nc(ctx, src0, src1, dst); + if (!split && use_mul_mat_vec && dst->ne[3] == 1 && (src0->ne[1] < MMV_MAX_ROWS || any_gpus_without_fp16_mma)) { + // the custom F16 vector kernel can be used over batched cuBLAS GEMM + // but this is only faster for GPUs without tensor cores or with a thin src0 matrix (particularly KQV in attention) + ggml_cuda_mul_mat_vec(ctx, src0, src1, dst); } else if (!split && src0->type == GGML_TYPE_F16 && (src1->type == GGML_TYPE_F16 || !any_gpus_with_slow_fp16) && !ggml_is_transposed(src0) && !ggml_is_transposed(src1) && src1->ne[2]*src1->ne[3] > 1) { - // KQ + KQV multi-batch without FlashAttention + // general KQ + KQV multi-batch without FlashAttention ggml_cuda_mul_mat_batched_cublas(ctx, src0, src1, dst); - } else if (use_dequantize_mul_mat_vec) { - ggml_cuda_op_mul_mat(ctx, src0, src1, dst, ggml_cuda_op_dequantize_mul_mat_vec, nullptr); + } else if (use_mul_mat_vec) { + ggml_cuda_op_mul_mat(ctx, src0, src1, dst, ggml_cuda_op_mul_mat_vec, nullptr); } else if (use_mul_mat_vec_q) { ggml_cuda_op_mul_mat(ctx, src0, src1, dst, ggml_cuda_op_mul_mat_vec_q, quantize_row_q8_1_cuda); } else if (use_mul_mat_q) { @@ -2295,6 +2141,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg case GGML_OP_ROPE: ggml_cuda_op_rope(ctx, dst); break; + case GGML_OP_ROPE_BACK: + ggml_cuda_op_rope_back(ctx, dst); + break; case GGML_OP_IM2COL: ggml_cuda_op_im2col(ctx, dst); break; @@ -2322,6 +2171,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg case GGML_OP_RWKV_WKV6: ggml_cuda_op_rwkv_wkv6(ctx, dst); break; + case GGML_OP_GATED_LINEAR_ATTN: + ggml_cuda_op_gated_linear_attn(ctx, dst); + break; case GGML_OP_CROSS_ENTROPY_LOSS_BACK: ggml_cuda_cross_entropy_loss_back(ctx, dst); break; @@ -2440,6 +2292,66 @@ static void ggml_backend_cuda_synchronize(ggml_backend_t backend) { } #ifdef USE_CUDA_GRAPH +static bool check_node_graph_compatibility_and_refresh_copy_ops(ggml_backend_cuda_context * cuda_ctx, ggml_cgraph * cgraph, + std::vector & ggml_cuda_cpy_fn_ptrs, bool use_cuda_graph) { + + // Loop over nodes in GGML graph to obtain info needed for CUDA graph + cuda_ctx->cuda_graph->updated_kernel_arg.clear(); + for (int i = 0; i < cgraph->n_nodes; i++) { + ggml_tensor * node = cgraph->nodes[i]; + + if (ggml_is_empty(node) || node->op == GGML_OP_RESHAPE || node->op == GGML_OP_TRANSPOSE || node->op == GGML_OP_VIEW || node->op == GGML_OP_PERMUTE || node->op == GGML_OP_NONE) { + continue; + } + + if (node->src[0] && node->src[0]->buffer && ggml_backend_buft_is_cuda_split(node->src[0]->buffer->buft)) { + use_cuda_graph = false; // Split buffers are not supported by CUDA graph capture +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to split buffer\n", __func__); +#endif + } + + if (node->op == GGML_OP_MUL_MAT_ID) { + use_cuda_graph = false; // This node type is not supported by CUDA graph capture +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to mul_mat_id\n", __func__); +#endif + } + + if (node->op == GGML_OP_ADD && node->src[1] && node->src[1]->ne[1] > 1) { + // disable CUDA graphs for batch size > 1 for now. + // Changes in batch size or context size can cause changes to the grid size of some kernels. + use_cuda_graph = false; +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to batch size > 1 [%s] [%ld %ld %ld %ld]\n", __func__, node->name, node->ne[0], node->ne[1], node->ne[2], node->ne[3]); +#endif + } + + if (node->op == GGML_OP_CPY) { + // store the copy op parameter which changes with each token. + cuda_ctx->cuda_graph->updated_kernel_arg.push_back((char **) &(node->src[1]->data)); + // store a pointer to each copy op CUDA kernel to identify it later + void * ptr = ggml_cuda_cpy_fn(node->src[0], node->src[1]); + if (!ptr) { + use_cuda_graph = false; +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to unsupported copy op\n", __func__); +#endif + } else { + if (std::find(ggml_cuda_cpy_fn_ptrs.begin(), ggml_cuda_cpy_fn_ptrs.end(), ptr) == ggml_cuda_cpy_fn_ptrs.end()) { + ggml_cuda_cpy_fn_ptrs.push_back(ptr); + } + } + } + + if (!use_cuda_graph) { + break; + } + } + + return use_cuda_graph; +} + static void set_ggml_graph_node_properties(ggml_tensor * node, ggml_graph_node_properties * graph_node_properties) { graph_node_properties->node_address = node->data; graph_node_properties->node_op = node->op; @@ -2490,149 +2402,105 @@ static bool ggml_graph_node_has_matching_properties(ggml_tensor * node, ggml_gra return true; } -#endif -static enum ggml_status ggml_backend_cuda_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) { - ggml_backend_cuda_context * cuda_ctx = (ggml_backend_cuda_context *)backend->context; +static void maintain_cuda_graph(ggml_backend_cuda_context * cuda_ctx, std::vector & ggml_cuda_cpy_fn_ptrs, bool cuda_graph_update_required) { - ggml_cuda_set_device(cuda_ctx->device); + if (cuda_graph_update_required) { + // Extract nodes from graph + // First call with null argument gets number of nodes in graph + CUDA_CHECK(cudaGraphGetNodes(cuda_ctx->cuda_graph->graph, nullptr, &cuda_ctx->cuda_graph->num_nodes)); + // Subsequent call with non-null argument gets nodes + cuda_ctx->cuda_graph->nodes.clear(); + cuda_ctx->cuda_graph->nodes.resize(cuda_ctx->cuda_graph->num_nodes); + cuda_ctx->cuda_graph->params.clear(); + cuda_ctx->cuda_graph->params.resize(cuda_ctx->cuda_graph->num_nodes); + if (cuda_ctx->cuda_graph->num_nodes > 0) { + CUDA_CHECK(cudaGraphGetNodes(cuda_ctx->cuda_graph->graph, cuda_ctx->cuda_graph->nodes.data(), &cuda_ctx->cuda_graph->num_nodes)); -#ifdef USE_CUDA_GRAPH - static const bool disable_cuda_graphs_due_to_env = (getenv("GGML_CUDA_DISABLE_GRAPHS") != nullptr); - - // Objects required for CUDA Graph - if (cuda_ctx->cuda_graph == nullptr) { - cuda_ctx->cuda_graph.reset(new ggml_cuda_graph()); - } - - bool use_cuda_graph = true; - bool cuda_graph_update_required = false; - // vector of pointers to CUDA cpy kernels, which are required to identify - // kernel parameters which need updated in the graph for each token - std::vector ggml_cuda_cpy_fn_ptrs; - - if (cuda_ctx->cuda_graph->graph == nullptr) { - if (ggml_cuda_info().devices[cuda_ctx->device].cc < CC_AMPERE) { - cuda_ctx->cuda_graph->disable_due_to_gpu_arch = true; -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to GPU architecture\n", __func__); -#endif - } - } - - // Disable CUDA graphs in presence of env var, old GPU, use-case which is changing too rapidly, - // or previous graph capture failure. - // Also disable for multi-gpu for now. TO DO investigate - if (disable_cuda_graphs_due_to_env - || cuda_ctx->cuda_graph->disable_due_to_gpu_arch - || cuda_ctx->cuda_graph->disable_due_to_too_many_updates - || cuda_ctx->cuda_graph->disable_due_to_failed_graph_capture) { - use_cuda_graph = false; - } - - if (use_cuda_graph) { - if (cuda_ctx->cuda_graph->instance == nullptr) { - cuda_graph_update_required = true; - } - - // Check if the graph size has changed - if (cuda_ctx->cuda_graph->ggml_graph_properties.size() != (size_t)cgraph->n_nodes) { - cuda_graph_update_required = true; - cuda_ctx->cuda_graph->ggml_graph_properties.resize(cgraph->n_nodes); - } - - // Loop over nodes in GGML graph to determine if CUDA graph update is required - // and store properties to allow this comparison for the next token - for (int i = 0; i < cgraph->n_nodes; i++) { - bool has_matching_properties = true; - if (!cuda_graph_update_required) { - has_matching_properties = ggml_graph_node_has_matching_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); - } - if (!has_matching_properties) { - cuda_graph_update_required = true; - } - set_ggml_graph_node_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); - } - - // Loop over nodes in GGML graph to obtain info needed for CUDA graph - cuda_ctx->cuda_graph->updated_kernel_arg.clear(); - for (int i = 0; i < cgraph->n_nodes; i++) { - ggml_tensor * node = cgraph->nodes[i]; - - if (ggml_is_empty(node) || node->op == GGML_OP_RESHAPE || node->op == GGML_OP_TRANSPOSE || node->op == GGML_OP_VIEW || node->op == GGML_OP_PERMUTE || node->op == GGML_OP_NONE) { - continue; - } - - if (node->src[0] && node->src[0]->buffer && ggml_backend_buft_is_cuda_split(node->src[0]->buffer->buft)) { - use_cuda_graph = false; // Split buffers are not supported by CUDA graph capture -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to split buffer\n", __func__); -#endif - } - - if (node->op == GGML_OP_MUL_MAT_ID) { - use_cuda_graph = false; // This node type is not supported by CUDA graph capture -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to mul_mat_id\n", __func__); -#endif - } - - if (node->op == GGML_OP_ADD && node->src[1] && node->src[1]->ne[1] > 1) { - // disable CUDA graphs for batch size > 1 for now. - // Changes in batch size or context size can cause changes to the grid size of some kernels. - use_cuda_graph = false; -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to batch size > 1 [%s] [%ld %ld %ld %ld]\n", __func__, node->name, node->ne[0], node->ne[1], node->ne[2], node->ne[3]); -#endif - } - - if (node->op == GGML_OP_CPY) { - // store the copy op parameter which changes with each token. - cuda_ctx->cuda_graph->updated_kernel_arg.push_back((char **) &(node->src[1]->data)); - // store a pointer to each copy op CUDA kernel to identify it later - void * ptr = ggml_cuda_cpy_fn(node->src[0], node->src[1]); - if (!ptr) { - use_cuda_graph = false; -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to unsupported copy op\n", __func__); -#endif - } else { - if (std::find(ggml_cuda_cpy_fn_ptrs.begin(), ggml_cuda_cpy_fn_ptrs.end(), ptr) == ggml_cuda_cpy_fn_ptrs.end()) { - ggml_cuda_cpy_fn_ptrs.push_back(ptr); + // Loop over nodes, and extract kernel parameters from each node + for (size_t i = 0; i < cuda_ctx->cuda_graph->num_nodes; i++) { + cudaGraphNodeType node_type; + CUDA_CHECK(cudaGraphNodeGetType(cuda_ctx->cuda_graph->nodes[i], &node_type)); + if (node_type == cudaGraphNodeTypeKernel) { + cudaError_t stat = cudaGraphKernelNodeGetParams(cuda_ctx->cuda_graph->nodes[i], &cuda_ctx->cuda_graph->params[i]); // Get params using runtime + if (stat == cudaErrorInvalidDeviceFunction) { + // Fails due to incorrect handling by CUDA runtime of CUDA BLAS node. + // We don't need to update blas nodes, so clear error and move on. + cudaGetLastError(); + } else { + GGML_ASSERT(stat == cudaSuccess); } } } - - if (!use_cuda_graph) { - break; + } + } else { + // One of the arguments to the copy kernel is updated for each token, hence we need to + // replace that argument with the updated value in the CUDA graph + // on update steps, the live parameters will already be captured + int k = 0; + for (size_t i = 0; i < cuda_ctx->cuda_graph->num_nodes; i++) { + if(count(ggml_cuda_cpy_fn_ptrs.begin(), ggml_cuda_cpy_fn_ptrs.end(), cuda_ctx->cuda_graph->params[i].func) > 0) { + char ** updated_kernel_arg_ptr = cuda_ctx->cuda_graph->updated_kernel_arg.at(k++); + cuda_ctx->cuda_graph->params[i].kernelParams[1] = updated_kernel_arg_ptr; + CUDA_CHECK(cudaGraphKernelNodeSetParams(cuda_ctx->cuda_graph->nodes[i], &cuda_ctx->cuda_graph->params[i])); } } - - // Disable CUDA graphs (from the next token) if the use-case is demanding too many consecutive graph updates. - if (use_cuda_graph && cuda_graph_update_required) { - cuda_ctx->cuda_graph->number_consecutive_updates++; - } else { - cuda_ctx->cuda_graph->number_consecutive_updates = 0; - } - - if (cuda_ctx->cuda_graph->number_consecutive_updates >= 4) { - cuda_ctx->cuda_graph->disable_due_to_too_many_updates = true; -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to too many consecutive updates\n", __func__); -#endif - } } +} - if (use_cuda_graph && cuda_graph_update_required) { // Start CUDA graph capture - CUDA_CHECK(cudaStreamBeginCapture(cuda_ctx->stream(), cudaStreamCaptureModeRelaxed)); - } +static bool is_cuda_graph_update_required(ggml_backend_cuda_context * cuda_ctx, ggml_cgraph * cgraph) { -#else - bool use_cuda_graph = false; bool cuda_graph_update_required = false; -#endif // USE_CUDA_GRAPH - bool graph_evaluated_or_captured = false; + if (cuda_ctx->cuda_graph->instance == nullptr) { + cuda_graph_update_required = true; + } + + // Check if the graph size has changed + if (cuda_ctx->cuda_graph->ggml_graph_properties.size() != (size_t)cgraph->n_nodes) { + cuda_graph_update_required = true; + cuda_ctx->cuda_graph->ggml_graph_properties.resize(cgraph->n_nodes); + } + + // Loop over nodes in GGML graph to determine if CUDA graph update is required + // and store properties to allow this comparison for the next token + for (int i = 0; i < cgraph->n_nodes; i++) { + bool has_matching_properties = true; + if (!cuda_graph_update_required) { + has_matching_properties = ggml_graph_node_has_matching_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); + } + if (!has_matching_properties) { + cuda_graph_update_required = true; + } + set_ggml_graph_node_properties(cgraph->nodes[i], &cuda_ctx->cuda_graph->ggml_graph_properties[i]); + } + + return cuda_graph_update_required; +} + +static void update_cuda_graph_executable(ggml_backend_cuda_context * cuda_ctx) { + + cudaGraphExecUpdateResultInfo result_info; + cudaError_t stat = cudaGraphExecUpdate(cuda_ctx->cuda_graph->instance, cuda_ctx->cuda_graph->graph, &result_info); + if (stat == cudaErrorGraphExecUpdateFailure) { +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: CUDA graph update failed\n", __func__); +#endif + // The pre-existing graph exec cannot be updated due to violated constraints + // so instead clear error and re-instantiate + cudaGetLastError(); + CUDA_CHECK(cudaGraphExecDestroy(cuda_ctx->cuda_graph->instance)); + cuda_ctx->cuda_graph->instance = nullptr; + CUDA_CHECK(cudaGraphInstantiate(&cuda_ctx->cuda_graph->instance, cuda_ctx->cuda_graph->graph, NULL, NULL, 0)); + } else { + GGML_ASSERT(stat == cudaSuccess); + } +} +#endif + +static void evaluate_and_capture_cuda_graph(ggml_backend_cuda_context * cuda_ctx, ggml_cgraph * cgraph, + [[maybe_unused]] std::vector & ggml_cuda_cpy_fn_ptrs, bool & graph_evaluated_or_captured, bool & use_cuda_graph, + bool & cuda_graph_update_required) { while (!graph_evaluated_or_captured) { // Only perform the graph execution if CUDA graphs are not enabled, or we are capturing the graph. @@ -2670,19 +2538,8 @@ static enum ggml_status ggml_backend_cuda_graph_compute(ggml_backend_t backend, CUDA_CHECK(cudaGraphDestroy(cuda_ctx->cuda_graph->graph)); cuda_ctx->cuda_graph->graph = nullptr; } - CUDA_CHECK(cudaStreamEndCapture(cuda_ctx->stream(), &cuda_ctx->cuda_graph->graph)); -#if 0 - if (disable_cuda_graphs_due_to_failed_capture) { - use_cuda_graph = false; - cuda_ctx->cuda_graph->disable_due_to_failed_graph_capture = true; -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: disabling CUDA graphs due to failed graph capture\n", __func__); -#endif - } else { - graph_evaluated_or_captured = true; // CUDA graph has been captured - } -#endif + CUDA_CHECK(cudaStreamEndCapture(cuda_ctx->stream(), &cuda_ctx->cuda_graph->graph)); graph_evaluated_or_captured = true; // CUDA graph has been captured } else { graph_evaluated_or_captured = true; // ggml graph has been directly evaluated @@ -2695,72 +2552,91 @@ static enum ggml_status ggml_backend_cuda_graph_compute(ggml_backend_t backend, } // Perform update to graph (if required for this token), and change copy parameter (required for every token) - - if (cuda_graph_update_required) { - // Extract nodes from graph - // First call with null argument gets number of nodes in graph - CUDA_CHECK(cudaGraphGetNodes(cuda_ctx->cuda_graph->graph, nullptr, &cuda_ctx->cuda_graph->num_nodes)); - // Subsequent call with non-null argument gets nodes - cuda_ctx->cuda_graph->nodes.clear(); - cuda_ctx->cuda_graph->nodes.resize(cuda_ctx->cuda_graph->num_nodes); - cuda_ctx->cuda_graph->params.clear(); - cuda_ctx->cuda_graph->params.resize(cuda_ctx->cuda_graph->num_nodes); - if (cuda_ctx->cuda_graph->num_nodes > 0) { - CUDA_CHECK(cudaGraphGetNodes(cuda_ctx->cuda_graph->graph, cuda_ctx->cuda_graph->nodes.data(), &cuda_ctx->cuda_graph->num_nodes)); - - // Loop over nodes, and extract kernel parameters from each node - for (size_t i = 0; i < cuda_ctx->cuda_graph->num_nodes; i++) { - cudaGraphNodeType node_type; - CUDA_CHECK(cudaGraphNodeGetType(cuda_ctx->cuda_graph->nodes[i], &node_type)); - if (node_type == cudaGraphNodeTypeKernel) { - cudaError_t stat = cudaGraphKernelNodeGetParams(cuda_ctx->cuda_graph->nodes[i], &cuda_ctx->cuda_graph->params[i]); // Get params using runtime - if (stat == cudaErrorInvalidDeviceFunction) { - // Fails due to incorrect handling by CUDA runtime of CUDA BLAS node. - // We don't need to update blas nodes, so clear error and move on. - cudaGetLastError(); - } else { - GGML_ASSERT(stat == cudaSuccess); - } - } - } - } - } - - // One of the arguments to the copy kernel is updated for each token, hence we need to - // replace that argument with the updated value in the CUDA graph - if (!cuda_graph_update_required) { // on update steps, the live parameters will already be captured - int k = 0; - for (size_t i = 0; i < cuda_ctx->cuda_graph->num_nodes; i++) { - if(count(ggml_cuda_cpy_fn_ptrs.begin(), ggml_cuda_cpy_fn_ptrs.end(), cuda_ctx->cuda_graph->params[i].func) > 0) { - char ** updated_kernel_arg_ptr = cuda_ctx->cuda_graph->updated_kernel_arg.at(k++); - cuda_ctx->cuda_graph->params[i].kernelParams[1] = updated_kernel_arg_ptr; - CUDA_CHECK(cudaGraphKernelNodeSetParams(cuda_ctx->cuda_graph->nodes[i], &cuda_ctx->cuda_graph->params[i])); - } - } - } + maintain_cuda_graph(cuda_ctx, ggml_cuda_cpy_fn_ptrs, cuda_graph_update_required); // Update graph executable - cudaGraphExecUpdateResultInfo result_info; - cudaError_t stat = cudaGraphExecUpdate(cuda_ctx->cuda_graph->instance, cuda_ctx->cuda_graph->graph, &result_info); - if (stat == cudaErrorGraphExecUpdateFailure) { -#ifndef NDEBUG - GGML_LOG_DEBUG("%s: CUDA graph update failed\n", __func__); -#endif - // The pre-existing graph exec cannot be updated due to violated constraints - // so instead clear error and re-instantiate - cudaGetLastError(); - CUDA_CHECK(cudaGraphExecDestroy(cuda_ctx->cuda_graph->instance)); - cuda_ctx->cuda_graph->instance = nullptr; - CUDA_CHECK(cudaGraphInstantiate(&cuda_ctx->cuda_graph->instance, cuda_ctx->cuda_graph->graph, NULL, NULL, 0)); - } else { - GGML_ASSERT(stat == cudaSuccess); - } + update_cuda_graph_executable(cuda_ctx); + // Launch graph CUDA_CHECK(cudaGraphLaunch(cuda_ctx->cuda_graph->instance, cuda_ctx->stream())); #else graph_evaluated_or_captured = true; -#endif // USE_CUDA_GRAPH +#endif // USE_CUDA_GRAPH } +} + +static enum ggml_status ggml_backend_cuda_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) { + ggml_backend_cuda_context * cuda_ctx = (ggml_backend_cuda_context *)backend->context; + + ggml_cuda_set_device(cuda_ctx->device); + + // vector of pointers to CUDA cpy kernels, which are required to identify + // kernel parameters which need updated in the graph for each token + std::vector ggml_cuda_cpy_fn_ptrs; + +#ifdef USE_CUDA_GRAPH + static const bool disable_cuda_graphs_due_to_env = (getenv("GGML_CUDA_DISABLE_GRAPHS") != nullptr); + + // Objects required for CUDA Graph + if (cuda_ctx->cuda_graph == nullptr) { + cuda_ctx->cuda_graph.reset(new ggml_cuda_graph()); + } + + bool use_cuda_graph = true; + bool cuda_graph_update_required = false; + + if (cuda_ctx->cuda_graph->graph == nullptr) { + if (ggml_cuda_info().devices[cuda_ctx->device].cc < GGML_CUDA_CC_AMPERE) { + cuda_ctx->cuda_graph->disable_due_to_gpu_arch = true; +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to GPU architecture\n", __func__); +#endif + } + } + + // Disable CUDA graphs in presence of env var, old GPU, use-case which is changing too rapidly, + // or previous graph capture failure. + // Also disable for multi-gpu for now. TO DO investigate + if (disable_cuda_graphs_due_to_env + || cuda_ctx->cuda_graph->disable_due_to_gpu_arch + || cuda_ctx->cuda_graph->disable_due_to_too_many_updates + || cuda_ctx->cuda_graph->disable_due_to_failed_graph_capture) { + use_cuda_graph = false; + } + + if (use_cuda_graph) { + cuda_graph_update_required = is_cuda_graph_update_required(cuda_ctx, cgraph); + + use_cuda_graph = check_node_graph_compatibility_and_refresh_copy_ops(cuda_ctx, cgraph, + ggml_cuda_cpy_fn_ptrs, use_cuda_graph); + + // Disable CUDA graphs (from the next token) if the use-case is demanding too many consecutive graph updates. + if (use_cuda_graph && cuda_graph_update_required) { + cuda_ctx->cuda_graph->number_consecutive_updates++; + } else { + cuda_ctx->cuda_graph->number_consecutive_updates = 0; + } + + if (cuda_ctx->cuda_graph->number_consecutive_updates >= 4) { + cuda_ctx->cuda_graph->disable_due_to_too_many_updates = true; +#ifndef NDEBUG + GGML_LOG_DEBUG("%s: disabling CUDA graphs due to too many consecutive updates\n", __func__); +#endif + } + } + + if (use_cuda_graph && cuda_graph_update_required) { // Start CUDA graph capture + CUDA_CHECK(cudaStreamBeginCapture(cuda_ctx->stream(), cudaStreamCaptureModeRelaxed)); + } + +#else + bool use_cuda_graph = false; + bool cuda_graph_update_required = false; +#endif // USE_CUDA_GRAPH + + bool graph_evaluated_or_captured = false; + + evaluate_and_capture_cuda_graph(cuda_ctx, cgraph, ggml_cuda_cpy_fn_ptrs, graph_evaluated_or_captured, use_cuda_graph, cuda_graph_update_required); return GGML_STATUS_SUCCESS; } @@ -2978,6 +2854,17 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g { struct ggml_tensor * a = op->src[0]; struct ggml_tensor * b = op->src[1]; + // for small weight matrices the active device can end up without any rows, don't use row split in those cases + // this avoids some edge cases (and the performance would not be good anyways) + if (a->buffer && ggml_backend_buft_is_cuda_split(a->buffer->buft)) { + ggml_backend_cuda_split_buffer_type_context * buft_ctx = (ggml_backend_cuda_split_buffer_type_context *) a->buffer->buft->context; + int64_t row_low; + int64_t row_high; + get_row_split(&row_low, &row_high, a, buft_ctx->tensor_split, dev_ctx->device); + if (row_low == row_high) { + return false; + } + } if (b->type == GGML_TYPE_F16 && a->type != GGML_TYPE_F16) { return false; } @@ -3013,6 +2900,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g case GGML_TYPE_IQ3_XXS: case GGML_TYPE_IQ4_NL: case GGML_TYPE_IQ4_XS: + case GGML_TYPE_BF16: #ifdef GGML_USE_MUSA if (a->type == GGML_TYPE_Q3_K) { return false; @@ -3140,7 +3028,11 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g case GGML_OP_SOFT_MAX: return true; case GGML_OP_ROPE: - return ggml_is_contiguous(op->src[0]); + case GGML_OP_ROPE_BACK: { + const size_t ts = ggml_type_size(op->src[0]->type); + const int64_t ne0_012 = op->src[0]->ne[0] * op->src[0]->ne[1] * op->src[0]->ne[2]; + return op->src[0]->nb[0] == ts && op->src[0]->nb[3] == ne0_012*ts; + } case GGML_OP_IM2COL: case GGML_OP_POOL_2D: case GGML_OP_SUM: @@ -3154,11 +3046,15 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_LEAKY_RELU: case GGML_OP_RWKV_WKV6: + case GGML_OP_GATED_LINEAR_ATTN: return true; case GGML_OP_FLASH_ATTN_EXT: { #ifndef FLASH_ATTN_AVAILABLE return false; #endif + if (op->src[1]->type == GGML_TYPE_BF16 || op->src[2]->type == GGML_TYPE_BF16) { + return false; + } if (op->src[0]->ne[0] == 64 && op->src[1]->type == GGML_TYPE_F16) { return true; } @@ -3169,7 +3065,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g return true; } const int cc = ggml_cuda_info().devices[dev_ctx->device].cc; - return cc >= CC_VOLTA && cc < CC_OFFSET_AMD && op->src[1]->type == GGML_TYPE_F16 && op->src[2]->type == GGML_TYPE_F16; + return cc >= GGML_CUDA_CC_VOLTA && cc < GGML_CUDA_CC_OFFSET_AMD && op->src[1]->type == GGML_TYPE_F16 && op->src[2]->type == GGML_TYPE_F16; } case GGML_OP_CROSS_ENTROPY_LOSS: case GGML_OP_CROSS_ENTROPY_LOSS_BACK: @@ -3192,6 +3088,7 @@ static int64_t get_op_batch_size(const ggml_tensor * op) { return op->ne[1]; case GGML_OP_MUL_MAT_ID: case GGML_OP_ROPE: + case GGML_OP_ROPE_BACK: return op->ne[2]; default: return ggml_nrows(op); @@ -3276,6 +3173,61 @@ static ggml_backend_dev_t ggml_backend_cuda_reg_get_device(ggml_backend_reg_t re return ctx->devices[index]; } +static ggml_backend_feature * ggml_backend_cuda_get_features(ggml_backend_reg_t reg) { + static std::vector features = []() { + std::vector features; + #define _STRINGIFY(...) #__VA_ARGS__ + #define STRINGIFY(...) _STRINGIFY(__VA_ARGS__) + + #ifdef __CUDA_ARCH_LIST__ + features.push_back({ "ARCHS", STRINGIFY(__CUDA_ARCH_LIST__) }); + #endif + + #ifdef GGML_CUDA_FORCE_MMQ + features.push_back({ "FORCE_MMQ", "1" }); + #endif + + #ifdef GGML_CUDA_FORCE_CUBLAS + features.push_back({ "FORCE_CUBLAS", "1" }); + #endif + + #ifdef GGML_CUDA_NO_VMM + features.push_back({ "NO_VMM", "1" }); + #endif + + #ifdef GGML_CUDA_NO_PEER_COPY + features.push_back({ "NO_PEER_COPY", "1" }); + #endif + + #ifdef GGML_CUDA_F16 + features.push_back({ "F16", "1" }); + #endif + + #ifdef GGML_CUDA_USE_GRAPHS + features.push_back({ "USE_GRAPHS", "1" }); + #endif + + #ifdef GGML_CUDA_PEER_MAX_BATCH_SIZE + features.push_back({ "PEER_MAX_BATCH_SIZE", STRINGIFY(GGML_CUDA_PEER_MAX_BATCH_SIZE) }); + #endif + + #ifdef GGML_CUDA_FA_ALL_QUANTS + features.push_back({ "FA_ALL_QUANTS", "1" }); + #endif + + #undef _STRINGIFY + #undef STRINGIFY + + features.push_back({ nullptr, nullptr }); + + return features; + }(); + + return features.data(); + + GGML_UNUSED(reg); +} + static void * ggml_backend_cuda_reg_get_proc_address(ggml_backend_reg_t reg, const char * name) { GGML_UNUSED(reg); if (strcmp(name, "ggml_backend_split_buffer_type") == 0) { @@ -3287,13 +3239,16 @@ static void * ggml_backend_cuda_reg_get_proc_address(ggml_backend_reg_t reg, con if (strcmp(name, "ggml_backend_unregister_host_buffer") == 0) { return (void *)ggml_backend_cuda_unregister_host_buffer; } + if (strcmp(name, "ggml_backend_get_features") == 0) { + return (void *)ggml_backend_cuda_get_features; + } return nullptr; } static const ggml_backend_reg_i ggml_backend_cuda_reg_interface = { /* .get_name = */ ggml_backend_cuda_reg_get_name, /* .get_device_count = */ ggml_backend_cuda_reg_get_device_count, - /* .get_device_get = */ ggml_backend_cuda_reg_get_device, + /* .get_device = */ ggml_backend_cuda_reg_get_device, /* .get_proc_address = */ ggml_backend_cuda_reg_get_proc_address, }; @@ -3319,16 +3274,17 @@ ggml_backend_reg_t ggml_backend_cuda_reg() { dev_ctx->description = prop.name; ggml_backend_dev_t dev = new ggml_backend_device { - /* .interface = */ ggml_backend_cuda_device_interface, - /* .reg = */ ®, - /* .context = */ dev_ctx + /* .iface = */ ggml_backend_cuda_device_interface, + /* .reg = */ ®, + /* .context = */ dev_ctx }; ctx->devices.push_back(dev); } reg = ggml_backend_reg { - /* .interface = */ ggml_backend_cuda_reg_interface, - /* .context = */ ctx + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_cuda_reg_interface, + /* .context = */ ctx }; } @@ -3359,3 +3315,5 @@ ggml_backend_t ggml_backend_cuda_init(int device) { return cuda_backend; } + +GGML_BACKEND_DL_IMPL(ggml_backend_cuda_reg) diff --git a/ggml/src/ggml-cuda/gla.cu b/ggml/src/ggml-cuda/gla.cu new file mode 100644 index 000000000..f7d615a82 --- /dev/null +++ b/ggml/src/ggml-cuda/gla.cu @@ -0,0 +1,93 @@ +#include "common.cuh" +#include "gla.cuh" + +template +static __global__ void gated_linear_attn_f32(const int B, const int T, const int C, const int H, const float scale, + const float * k, const float * v, const float * r, const float * td, const float * s, float * dst) { + const int tid = threadIdx.x; + const int bid = blockIdx.x; + + const int head_size = HEAD_SIZE; + const int batch_i = bid / H; + const int head_i = bid % H; + const int state_size = C * head_size; + const int n_seq_tokens = T / B; + + float state[head_size]; + __shared__ float _k[head_size], _r[head_size], _td[head_size]; + + #pragma unroll + for (int i = 0; i < head_size; i++) { + state[i] = s[batch_i * state_size + head_i * head_size * head_size + i * head_size + tid]; + } + + for (int t = batch_i * n_seq_tokens * C + head_i * head_size + tid; t < (batch_i + 1) * n_seq_tokens * C + head_i * head_size + tid; t += C) { + __syncthreads(); + _k[tid] = k[t]; + _r[tid] = r[t]; + _td[tid] = td[t]; + __syncthreads(); + + const float _v = v[t]; + float y = 0; + for (int j = 0; j < head_size; j += 4) { + const float4 & k = (float4 &)(_k[j]); + const float4 & r = (float4 &)(_r[j]); + const float4 & td = (float4 &)(_td[j]); + float4 & s = (float4 &)(state[j]); + float4 kv; + + kv.x = k.x * _v; + kv.y = k.y * _v; + kv.z = k.z * _v; + kv.w = k.w * _v; + + s.x = s.x * td.x + kv.x; + s.y = s.y * td.y + kv.y; + s.z = s.z * td.z + kv.z; + s.w = s.w * td.w + kv.w; + + y += r.x * s.x; + y += r.y * s.y; + y += r.z * s.z; + y += r.w * s.w; + } + dst[t] = y * scale; + } + + #pragma unroll + for (int i = 0; i < head_size; i++) { + dst[T * C + batch_i * state_size + head_i * head_size * head_size + i * head_size + tid] = state[i]; + } +} + +void ggml_cuda_op_gated_linear_attn(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + const float * k_d = (const float *)dst->src[0]->data; + const float * v_d = (const float *)dst->src[1]->data; + const float * r_d = (const float *)dst->src[2]->data; + const float * td_d = (const float *)dst->src[3]->data; + const float * s_d = (const float *)dst->src[4]->data; + + const int64_t B = dst->src[4]->ne[1]; + const int64_t T = dst->src[0]->ne[2]; + const int64_t C = dst->ne[0]; + const int64_t H = dst->src[0]->ne[1]; + + float scale; + memcpy(&scale, (float*)dst->op_params, sizeof(float)); + + float * dst_d = (float *)dst->data; + + cudaStream_t stream = ctx.stream(); + + GGML_ASSERT(dst->src[4]->type == GGML_TYPE_F32); + GGML_ASSERT(C % H == 0); + GGML_ASSERT(C / H == 64 || C / H == 128); + + + if (C / H == 64) { + gated_linear_attn_f32<64><<>>(B, T, C, H, scale, k_d, v_d, r_d, td_d, s_d, dst_d); + } else { + gated_linear_attn_f32<128><<>>(B, T, C, H, scale, k_d, v_d, r_d, td_d, s_d, dst_d); + } +} diff --git a/ggml/src/ggml-cuda/gla.cuh b/ggml/src/ggml-cuda/gla.cuh new file mode 100644 index 000000000..2c82ad7dd --- /dev/null +++ b/ggml/src/ggml-cuda/gla.cuh @@ -0,0 +1,3 @@ +#include "common.cuh" + +void ggml_cuda_op_gated_linear_attn(ggml_backend_cuda_context & ctx, ggml_tensor * dst); diff --git a/ggml/src/ggml-cuda/mma.cuh b/ggml/src/ggml-cuda/mma.cuh index a452a3cc3..7d11540af 100644 --- a/ggml/src/ggml-cuda/mma.cuh +++ b/ggml/src/ggml-cuda/mma.cuh @@ -171,7 +171,7 @@ struct mma_int_C_I16J8 { __device__ __forceinline__ void mma_K4(const mma_int_A_I16K4 & mma_A, const mma_int_B_J8K4 & mma_B) { #ifdef INT8_MMA_AVAILABLE -#if __CUDA_ARCH__ >= CC_AMPERE +#if __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE asm("mma.sync.aligned.m16n8k16.row.col.s32.s8.s8.s32 {%0, %1, %2, %3}, {%4, %5}, {%6}, {%0, %1, %2, %3};" : "+r"(x[0]), "+r"(x[1]), "+r"(x[2]), "+r"(x[3]) : "r"(mma_A.x[0]), "r"(mma_A.x[1]), "r"(mma_B.x[0])); @@ -183,7 +183,7 @@ struct mma_int_C_I16J8 { asm("mma.sync.aligned.m8n8k16.row.col.s32.s8.s8.s32 {%0, %1}, {%2}, {%3}, {%0, %1};" : "+r"(x[2]), "+r"(x[3]) : "r"(mma_A.x[1]), "r"(mma_B.x[0])); -#endif // __CUDA_ARCH__ >= CC_AMPERE +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE #else GGML_UNUSED(mma_A); GGML_UNUSED(mma_B); @@ -193,7 +193,7 @@ struct mma_int_C_I16J8 { __device__ __forceinline__ void mma_K8(const mma_int_A_I16K8 & mma_A, const mma_int_B_J8K8 & mma_B) { #ifdef INT8_MMA_AVAILABLE -#if __CUDA_ARCH__ >= CC_AMPERE +#if __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE asm("mma.sync.aligned.m16n8k32.row.col.s32.s8.s8.s32 {%0, %1, %2, %3}, {%4, %5, %6, %7}, {%8, %9}, {%0, %1, %2, %3};" : "+r"(x[0]), "+r"(x[1]), "+r"(x[2]), "+r"(x[3]) : "r"(mma_A.x[0]), "r"(mma_A.x[1]), "r"(mma_A.x[2]), "r"(mma_A.x[3]), "r"(mma_B.x[0]), "r"(mma_B.x[1])); @@ -211,7 +211,7 @@ struct mma_int_C_I16J8 { asm("mma.sync.aligned.m8n8k16.row.col.s32.s8.s8.s32 {%0, %1}, {%2}, {%3}, {%0, %1};" : "+r"(x[2]), "+r"(x[3]) : "r"(mma_A.x[3]), "r"(mma_B.x[1])); -#endif // __CUDA_ARCH__ >= CC_AMPERE +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE #else GGML_UNUSED(mma_A); GGML_UNUSED(mma_B); diff --git a/ggml/src/ggml-cuda/mmq.cu b/ggml/src/ggml-cuda/mmq.cu index ae5c68ab3..270251df4 100644 --- a/ggml/src/ggml-cuda/mmq.cu +++ b/ggml/src/ggml-cuda/mmq.cu @@ -27,7 +27,7 @@ void ggml_cuda_op_mul_mat_q( // The stream-k decomposition is only faster for recent NVIDIA GPUs. // Also its fixup needs to allocate a temporary buffer in the memory pool. // There are multiple parallel CUDA streams for src1_ncols != ne11 which would introduce a race condition for this buffer. - const bool use_stream_k = compute_capability >= CC_VOLTA && compute_capability < CC_OFFSET_AMD && src1_ncols == ne11; + const bool use_stream_k = compute_capability >= GGML_CUDA_CC_VOLTA && compute_capability < GGML_CUDA_CC_OFFSET_AMD && src1_ncols == ne11; const mmq_args args = {src0_dd_i, src1_ddq_i, dst_dd_i, ne00, row_diff, stride00, src1_padded_row_size, src1_ncols, ne11, nrows_dst, use_stream_k}; switch (src0->type) { @@ -136,7 +136,7 @@ bool ggml_cuda_should_use_mmq(enum ggml_type type, int cc, int64_t ne11) { return true; } - if (cc < MIN_CC_DP4A) { + if (cc < GGML_CUDA_CC_DP4A) { return false; } @@ -144,9 +144,9 @@ bool ggml_cuda_should_use_mmq(enum ggml_type type, int cc, int64_t ne11) { return true; #endif //GGML_CUDA_FORCE_MMQ - if (cc < CC_OFFSET_AMD) { - return cc < CC_VOLTA || ne11 < MMQ_DP4A_MAX_BATCH_SIZE; + if (cc < GGML_CUDA_CC_OFFSET_AMD) { + return cc < GGML_CUDA_CC_VOLTA || ne11 < MMQ_DP4A_MAX_BATCH_SIZE; } - return cc < CC_RDNA3 || ne11 < MMQ_DP4A_MAX_BATCH_SIZE; + return (cc < GGML_CUDA_CC_RDNA3 && cc != GGML_CUDA_CC_CDNA && cc != GGML_CUDA_CC_VEGA20) || ne11 < MMQ_DP4A_MAX_BATCH_SIZE; } diff --git a/ggml/src/ggml-cuda/mmq.cuh b/ggml/src/ggml-cuda/mmq.cuh index 021a25682..3cd508a1d 100644 --- a/ggml/src/ggml-cuda/mmq.cuh +++ b/ggml/src/ggml-cuda/mmq.cuh @@ -89,9 +89,9 @@ struct tile_x_sizes { static constexpr int get_mmq_x_max_host(const int cc) { return int8_mma_available(cc) ? 128 : #ifdef GGML_CUDA_FORCE_MMQ - cc >= CC_VOLTA && cc < CC_OFFSET_AMD ? 128 : 64; + cc >= GGML_CUDA_CC_VOLTA && cc < GGML_CUDA_CC_OFFSET_AMD ? 128 : 64; #else - cc >= CC_VOLTA && cc < CC_OFFSET_AMD ? MMQ_DP4A_MAX_BATCH_SIZE : 64; + cc >= GGML_CUDA_CC_VOLTA && cc < GGML_CUDA_CC_OFFSET_AMD ? MMQ_DP4A_MAX_BATCH_SIZE : 64; #endif // GGML_CUDA_FORCE_MMQ } @@ -100,43 +100,43 @@ static constexpr __device__ int get_mmq_x_max_device() { return 128; #else // INT8_MMA_AVAILABLE -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) return 128; -#else // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#else // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) -#if __CUDA_ARCH__ >= CC_VOLTA +#if __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA #ifdef GGML_CUDA_FORCE_MMQ return MMQ_DP4A_MAX_BATCH_SIZE; #else // GGML_CUDA_FORCE_MMQ return 128; #endif // GGML_CUDA_FORCE_MMQ -#else // __CUDA_ARCH__ >= CC_VOLTA +#else // __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA return 64; -#endif // __CUDA_ARCH__ >= CC_VOLTA +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) #endif // INT8_MMA_AVAILABLE } static constexpr int get_mmq_y_host(const int cc) { - return cc >= CC_OFFSET_AMD ? (cc == CC_RDNA1 ? 64 : 128) : (cc >= CC_VOLTA ? 128 : 64); + return cc >= GGML_CUDA_CC_OFFSET_AMD ? (cc == GGML_CUDA_CC_RDNA1 ? 64 : 128) : (cc >= GGML_CUDA_CC_VOLTA ? 128 : 64); } static constexpr __device__ int get_mmq_y_device() { -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) #if defined(RDNA1) return 64; #else return 128; #endif // defined RDNA1 #else -#if __CUDA_ARCH__ >= CC_VOLTA +#if __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA return 128; #else return 64; -#endif // __CUDA_ARCH__ >= CC_VOLTA -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) } #define MMQ_DP4A_TXS_Q4_0 tile_x_sizes{mmq_y*WARP_SIZE + mmq_y, mmq_y*WARP_SIZE/QI4_0 + mmq_y/QI4_0, 0} @@ -2569,17 +2569,17 @@ static __device__ void mul_mat_q_process_tile( // The mul_mat_q kernel implements "stream-k" work partitioning as described in https://arxiv.org/abs/2301.03598 template -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) -#if defined(RDNA3) || defined(RDNA2) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) +#if defined(RDNA3) || defined(RDNA2) || defined(CDNA) || defined(GCN) __launch_bounds__(WARP_SIZE*nwarps, 2) -#endif // defined(RDNA3) || defined(RDNA2) +#endif // defined(RDNA3) || defined(RDNA2) || defined(CDNA) || defined(GCN) #else -#if __CUDA_ARCH__ >= CC_VOLTA +#if __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA __launch_bounds__(WARP_SIZE*nwarps, 1) #else __launch_bounds__(WARP_SIZE*nwarps, 2) -#endif // __CUDA_ARCH__ >= CC_VOLTA -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) static __global__ void mul_mat_q( const char * __restrict__ x, const char * __restrict__ yc, float * __restrict__ dst, float * __restrict__ tmp_fixup, const int ne00, const int ne01, const int stride01, const int ne10, const int ne11, const int stride11, const int ne0) { @@ -2594,7 +2594,7 @@ static __global__ void mul_mat_q( constexpr int mmq_y = get_mmq_y_device(); // On AMD or old CUDA the performance with stream-k was worse, use conventional tiling instead: -#if (defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ < CC_VOLTA +#if (defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ < GGML_CUDA_CC_VOLTA { constexpr bool fixup = false; mul_mat_q_process_tile @@ -2602,7 +2602,7 @@ static __global__ void mul_mat_q( blockIdx.x, blockIdx.y, 0, ne00/qk); return; } -#endif // (defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ < CC_VOLTA +#endif // (defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ < GGML_CUDA_CC_VOLTA const int64_t blocks_per_ne00 = ne00 / qk; constexpr int blocks_per_iter = MMQ_ITER_K / qk; @@ -2765,14 +2765,14 @@ static void launch_mul_mat_q(ggml_backend_cuda_context & ctx, const mmq_args & a const int shmem = mmq_get_shmem(mmq_x, mmq_y, cc); -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static bool shmem_limit_raised[GGML_CUDA_MAX_DEVICES] = {false}; if (!shmem_limit_raised[id]) { CUDA_CHECK(cudaFuncSetAttribute(mul_mat_q, cudaFuncAttributeMaxDynamicSharedMemorySize, shmem)); CUDA_CHECK(cudaFuncSetAttribute(mul_mat_q, cudaFuncAttributeMaxDynamicSharedMemorySize, shmem)); shmem_limit_raised[id] = true; } -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) const int nty = (args.ne01 + mmq_y - 1) / mmq_y; const int ntx = (args.ne11 + mmq_x - 1) / mmq_x; @@ -2825,7 +2825,7 @@ void mul_mat_q_case(ggml_backend_cuda_context & ctx, const mmq_args & args, cuda const int mmq_x_max = get_mmq_x_max_host(cc); const int mmq_y = get_mmq_y_host(cc); const int block_num_y = (args.ne01 + mmq_y - 1) / mmq_y; - const bool use_stream_k = cc >= CC_VOLTA && cc < CC_OFFSET_AMD; + const bool use_stream_k = cc >= GGML_CUDA_CC_VOLTA && cc < GGML_CUDA_CC_OFFSET_AMD; int mmq_x_best = 0; int nparts_best = INT_MAX; diff --git a/ggml/src/ggml-cuda/mmv.cu b/ggml/src/ggml-cuda/mmv.cu new file mode 100644 index 000000000..ac45f2d17 --- /dev/null +++ b/ggml/src/ggml-cuda/mmv.cu @@ -0,0 +1,261 @@ +#include "common.cuh" +#include "mmv.cuh" + +template +static __global__ void mul_mat_vec( + const T * __restrict__ x, const float * __restrict__ y, float * __restrict__ dst, const int64_t ncols2, const int64_t stride_row, + const int64_t channel_ratio, const int64_t stride_channel_x, const int64_t stride_channel_y, const int64_t stride_channel_dst) { + const int64_t row = blockIdx.x; + const int64_t channel = blockIdx.z; + const int tid = threadIdx.x; + + x += (channel/channel_ratio)*stride_channel_x + row*stride_row; + y += channel *stride_channel_y; + dst += channel *stride_channel_dst; + + const float2 * y2 = (const float2 *) y; + + extern __shared__ char data_mmv[]; + float * buf_iw = (float *) data_mmv; + + if (block_size > WARP_SIZE) { + if (tid < WARP_SIZE) { + buf_iw[tid] = 0.0f; + } + __syncthreads(); + } + + float sumf; + + if constexpr (std::is_same::value) { + const half2 * x2 = (const half2 *) x; + + if (std::is_same::value) { + sumf = 0.0f; + + for (int64_t col2 = tid; col2 < ncols2; col2 += block_size) { + const float2 tmpx = __half22float2(x2[col2]); + const float2 tmpy = y2[col2]; + sumf += tmpx.x * tmpy.x; + sumf += tmpx.y * tmpy.y; + } + } else { +#ifdef FP16_AVAILABLE + half2 sumh2 = make_half2(0.0f, 0.0f); + + for (int64_t col2 = tid; col2 < ncols2; col2 += block_size) { + const float2 tmp = y2[col2]; + sumh2 += x2[col2] * make_half2(tmp.x, tmp.y); + } + + sumf = __low2float(sumh2) + __high2float(sumh2); +#else + NO_DEVICE_CODE; +#endif // FP16_AVAILABLE + } + } else if constexpr (std::is_same::value) { + const int * x2 = (const int *) x; + sumf = 0.0f; + + for (int64_t col2 = tid; col2 < ncols2; col2 += block_size) { + const int tmpx = x2[col2]; + const float2 tmpy = y2[col2]; + sumf += float(reinterpret_cast(&tmpx)[0]) * tmpy.x; + sumf += float(reinterpret_cast(&tmpx)[1]) * tmpy.y; + } + } else { + static_assert(std::is_same::value, "unsupported type"); + } + + sumf = warp_reduce_sum(sumf); + + if (block_size > WARP_SIZE) { + buf_iw[tid/WARP_SIZE] = sumf; + __syncthreads(); + if (tid >= WARP_SIZE) { + return; + } + sumf = buf_iw[tid]; + sumf = warp_reduce_sum(sumf); + } + + if (tid != 0) { + return; + } + + dst[row] = sumf; +} + +template +static void launch_mul_mat_vec_cuda( + const T * x, const float * y, float * dst, + const int64_t ncols, const int64_t nrows, const int64_t stride_row, const int64_t nchannels_x, const int64_t nchannels_y, + const int64_t stride_channel_x, const int64_t stride_channel_y, const int64_t stride_channel_dst, + cudaStream_t stream) { + GGML_ASSERT(ncols % 2 == 0); + GGML_ASSERT(stride_row % 2 == 0); + GGML_ASSERT(nchannels_y % nchannels_x == 0); + const int64_t channel_ratio = nchannels_y / nchannels_x; + + int64_t block_size_best = WARP_SIZE; + int64_t niter_best = (ncols + 2*WARP_SIZE - 1) / (2*WARP_SIZE); + for (int64_t block_size = 2*WARP_SIZE; block_size <= 256; block_size += WARP_SIZE) { + const int64_t niter = (ncols + 2*block_size - 1) / (2*block_size); + if (niter < niter_best) { + niter_best = niter; + block_size_best = block_size; + } + } + + const int smem = WARP_SIZE*sizeof(float); + const dim3 block_nums(nrows, 1, nchannels_y); + const dim3 block_dims(block_size_best, 1, 1); + switch (block_size_best) { + case 32: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 64: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 96: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 128: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 160: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 192: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 224: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + case 256: { + mul_mat_vec<<>> + (x, y, dst, ncols/2, stride_row, channel_ratio, stride_channel_x, stride_channel_y, stride_channel_dst); + } break; + default: { + GGML_ABORT("fatal error"); + } break; + } +} + +template +static void mul_mat_vec_cuda( + const T * x, const float * y, float * dst, + const int64_t ncols, const int64_t nrows, const int64_t stride_row, const int64_t nchannels_x, const int64_t nchannels_y, + const int64_t stride_channel_x, const int64_t stride_channel_y, const int64_t stride_channel_dst, + enum ggml_prec prec, cudaStream_t stream) { + switch (prec) { + case GGML_PREC_DEFAULT: { + launch_mul_mat_vec_cuda(x, y, dst, ncols, nrows, stride_row, nchannels_x, nchannels_y, + stride_channel_x, stride_channel_y, stride_channel_dst, stream); + } break; + case GGML_PREC_F32: { + launch_mul_mat_vec_cuda(x, y, dst, ncols, nrows, stride_row, nchannels_x, nchannels_y, + stride_channel_x, stride_channel_y, stride_channel_dst, stream); + } break; + } +} + +void ggml_cuda_mul_mat_vec(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src1->type == GGML_TYPE_F32); + GGML_ASSERT(dst->type == GGML_TYPE_F32); + + const int64_t ne00 = src0->ne[0]; + const int64_t ne01 = src0->ne[1]; + + GGML_ASSERT(src1->ne[1] == 1); + + const int cc = ggml_cuda_info().devices[ggml_cuda_get_device()].cc; + const enum ggml_prec prec = fast_fp16_available(cc) ? ggml_prec(dst->op_params[0]) : GGML_PREC_F32; + + const float * src1_d = (const float *) src1->data; + float * dst_d = (float *) dst->data; + + const int64_t ne02 = src0->ne[2]; + const int64_t ne12 = src1->ne[2]; + GGML_ASSERT(dst->ne[2] == ne12); + + GGML_ASSERT(src0->ne[3] == 1); + GGML_ASSERT(src1->ne[3] == 1); + GGML_ASSERT( dst->ne[3] == 1); + + const int64_t stride_row = src0->nb[1] / ggml_type_size(src0->type); + const int64_t channel_stride_x = src0->nb[2] / ggml_type_size(src0->type); + const int64_t channel_stride_y = src1->nb[2] / ggml_type_size(src1->type); + const int64_t channel_stride_dst = dst->nb[2] / ggml_type_size( dst->type); + + switch (src0->type) { + case GGML_TYPE_F16: { + const half * src0_d = (const half *) src0->data; + mul_mat_vec_cuda(src0_d, src1_d, dst_d, ne00, ne01, stride_row, ne02, ne12, + channel_stride_x, channel_stride_y, channel_stride_dst, prec, ctx.stream()); + } break; + case GGML_TYPE_BF16: { + const nv_bfloat16 * src0_d = (const nv_bfloat16 *) src0->data; + mul_mat_vec_cuda(src0_d, src1_d, dst_d, ne00, ne01, stride_row, ne02, ne12, + channel_stride_x, channel_stride_y, channel_stride_dst, prec, ctx.stream()); + } break; + default: + GGML_ABORT("unsupported type: %s", ggml_type_name(src0->type)); + } +} + +void ggml_cuda_op_mul_mat_vec( + ggml_backend_cuda_context & ctx, + const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const char * src0_dd_i, const float * src1_ddf_i, + const char * src1_ddq_i, float * dst_dd_i, const int64_t row_low, const int64_t row_high, const int64_t src1_ncols, + const int64_t src1_padded_row_size, cudaStream_t stream) { + + GGML_ASSERT(src1->type == GGML_TYPE_F32); + GGML_ASSERT(dst->type == GGML_TYPE_F32); + + const int64_t ne00 = src0->ne[0]; + const int64_t row_diff = row_high - row_low; + + GGML_ASSERT(src1_ncols == 1); + + const int cc = ggml_cuda_info().devices[ggml_cuda_get_device()].cc; + const enum ggml_prec prec = fast_fp16_available(cc) ? ggml_prec(dst->op_params[0]) : GGML_PREC_F32; + + + // ggml_cuda_op provides single, contiguous matrices + const int64_t stride_row = ne00; + const int64_t nchannels_x = 1; + const int64_t nchannels_y = 1; + const int64_t channel_stride_x = 0; + const int64_t channel_stride_y = 0; + const int64_t channel_stride_dst = 0; + + switch (src0->type) { + case GGML_TYPE_F16: { + const half * src0_d = (const half *) src0_dd_i; + mul_mat_vec_cuda(src0_d, src1_ddf_i, dst_dd_i, ne00, row_diff, stride_row, + nchannels_x, nchannels_y, channel_stride_x, channel_stride_y, channel_stride_dst, prec, stream); + } break; + case GGML_TYPE_BF16: { + const nv_bfloat16 * src0_d = (const nv_bfloat16 *) src0_dd_i; + mul_mat_vec_cuda(src0_d, src1_ddf_i, dst_dd_i, ne00, row_diff, stride_row, + nchannels_x, nchannels_y, channel_stride_x, channel_stride_y, channel_stride_dst, prec, stream); + } break; + default: + GGML_ABORT("unsupported type: %s", ggml_type_name(src0->type)); + } + + GGML_UNUSED(ctx); + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_ddq_i); + GGML_UNUSED(src1_ncols); + GGML_UNUSED(src1_padded_row_size); +} diff --git a/ggml/src/ggml-cuda/dmmv.cuh b/ggml/src/ggml-cuda/mmv.cuh similarity index 55% rename from ggml/src/ggml-cuda/dmmv.cuh rename to ggml/src/ggml-cuda/mmv.cuh index e727eb97f..78a1cd4a6 100644 --- a/ggml/src/ggml-cuda/dmmv.cuh +++ b/ggml/src/ggml-cuda/mmv.cuh @@ -1,20 +1,12 @@ #include "common.cuh" -// dmmv = dequantize_mul_mat_vec +// maximum number of src0 rows with which to use mul_mat_vec over cuBLAS if FP16 tensor cores are available +#define MMV_MAX_ROWS 512 -// TODO: remove this? -#ifndef GGML_CUDA_DMMV_X -#define GGML_CUDA_DMMV_X 32 -#endif +void ggml_cuda_mul_mat_vec(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); -#ifndef GGML_CUDA_MMV_Y -#define GGML_CUDA_MMV_Y 1 -#endif - -void ggml_cuda_op_dequantize_mul_mat_vec( +void ggml_cuda_op_mul_mat_vec( ggml_backend_cuda_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const char * src0_dd_i, const float * src1_ddf_i, const char * src1_ddq_i, float * dst_dd_i, const int64_t row_low, const int64_t row_high, const int64_t src1_ncols, const int64_t src1_padded_row_size, cudaStream_t stream); - -bool ggml_cuda_dmmv_type_supported(ggml_type src0_type); diff --git a/ggml/src/ggml-cuda/mmvq.cu b/ggml/src/ggml-cuda/mmvq.cu index 7dbbc9939..e3b912d87 100644 --- a/ggml/src/ggml-cuda/mmvq.cu +++ b/ggml/src/ggml-cuda/mmvq.cu @@ -48,10 +48,10 @@ static constexpr __device__ int get_vdr_mmvq(ggml_type type) { } template -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#if !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) // tell the compiler to use as many registers as it wants, see nwarps definition below __launch_bounds__((ncols_y <= 4 ? 4 : 2)*WARP_SIZE, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +#endif // !(defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)) static __global__ void mul_mat_vec_q( const void * __restrict__ vx, const void * __restrict__ vy, float * __restrict__ dst, const int ncols_x, const int nrows_x, const int nrows_y, const int nrows_dst) { @@ -62,13 +62,13 @@ static __global__ void mul_mat_vec_q( constexpr vec_dot_q_cuda_t vec_dot_q_cuda = get_vec_dot_q_cuda(type); -#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) && (defined(RDNA2) || defined(RDNA3)) +#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) && (defined(RDNA2) || defined(RDNA3)) constexpr int nwarps = 1; constexpr int rows_per_cuda_block = 1; #else constexpr int nwarps = ncols_y <= 4 ? 4 : 2; constexpr int rows_per_cuda_block = ncols_y == 1 ? 1 : 2; -#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) && !defined(RDNA2) && !defined(RDNA3) +#endif // defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__) && !defined(RDNA2) && !defined(RDNA3) const int tid = WARP_SIZE*threadIdx.y + threadIdx.x; const int row0 = rows_per_cuda_block*blockIdx.x; @@ -142,7 +142,7 @@ static void mul_mat_vec_q_cuda( int64_t nwarps = 1; int64_t rows_per_cuda_block = 1; - if (ggml_cuda_info().devices[id].cc < CC_RDNA2) { // NVIDIA and AMD older than RDNA2 + if (ggml_cuda_info().devices[id].cc < GGML_CUDA_CC_CDNA || ggml_cuda_info().devices[id].cc == GGML_CUDA_CC_RDNA1) { // NVIDIA and AMD older than RDNA2 but not CDNA switch(ncols_y) { case 1: nwarps = 4; diff --git a/ggml/src/ggml-cuda/opt-step-adamw.cu b/ggml/src/ggml-cuda/opt-step-adamw.cu index d6f13a9c6..35154f299 100644 --- a/ggml/src/ggml-cuda/opt-step-adamw.cu +++ b/ggml/src/ggml-cuda/opt-step-adamw.cu @@ -1,11 +1,11 @@ +#include "ggml-impl.h" #include "opt-step-adamw.cuh" #include static __global__ void opt_step_adamw_f32( - float * __restrict__ x, const float * __restrict__ g, float * __restrict__ g_m, float * __restrict__ g_v, const int64_t k, - const float alpha, const float beta1, const float beta2, const float eps, const float wd, - const float beta1h, const float beta2h) { + float * __restrict__ x, const float * __restrict__ g, float * __restrict__ g_m, float * __restrict__ g_v, + const float * __restrict__ pars, const int64_t k) { const int64_t i = (int64_t) blockIdx.x*blockDim.x + threadIdx.x; @@ -13,6 +13,14 @@ static __global__ void opt_step_adamw_f32( return; } + const float alpha = pars[0]; + const float beta1 = pars[1]; + const float beta2 = pars[2]; + const float eps = pars[3]; + const float wd = pars[4]; + const float beta1h = pars[5]; + const float beta2h = pars[6]; + const float gi = g[i]; const float gmi = g_m[i]*beta1 + gi*(1.0f - beta1); const float gvi = g_v[i]*beta2 + gi*gi*(1.0f - beta2); @@ -23,58 +31,48 @@ static __global__ void opt_step_adamw_f32( const float mh = gmi*beta1h; const float vh = sqrtf(gvi*beta2h) + eps; - x[i] = x[i]*(1.0f - alpha*wd) - mh/vh; + x[i] = x[i]*(1.0f - alpha*wd) - alpha*mh/vh; } static void opt_step_adamw_f32_cuda( - float * x, const float * g, float * g_m, float * g_v, const int64_t k, - const float alpha, const float beta1, const float beta2, const float eps, const float wd, - const float beta1h, const float beta2h, cudaStream_t stream) { + float * x, const float * g, float * g_m, float * g_v, const float * pars, const int64_t k, cudaStream_t stream) { const dim3 block_dims(CUDA_OPT_STEP_ADAMW_BLOCK_SIZE, 1, 1); const dim3 block_nums((k + CUDA_OPT_STEP_ADAMW_BLOCK_SIZE - 1) / CUDA_OPT_STEP_ADAMW_BLOCK_SIZE, 1, 1); - opt_step_adamw_f32<<>>(x, g, g_m, g_v, k, alpha, beta1, beta2, eps, wd, beta1h, beta2h); + opt_step_adamw_f32<<>>(x, g, g_m, g_v, pars, k); } void ggml_cuda_opt_step_adamw(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { - const ggml_tensor * src0 = dst->src[0]; - const ggml_tensor * src0_grad = dst->src[1]; - const ggml_tensor * src0_grad_m = dst->src[2]; - const ggml_tensor * src0_grad_v = dst->src[3]; + const ggml_tensor * src0 = dst->src[0]; + const ggml_tensor * src0_grad = dst->src[1]; + const ggml_tensor * src0_grad_m = dst->src[2]; + const ggml_tensor * src0_grad_v = dst->src[3]; + const ggml_tensor * adamw_params = dst->src[4]; - GGML_ASSERT(src0->type == GGML_TYPE_F32); - GGML_ASSERT(src0_grad->type == GGML_TYPE_F32); - GGML_ASSERT(src0_grad_m->type == GGML_TYPE_F32); - GGML_ASSERT(src0_grad_v->type == GGML_TYPE_F32); + GGML_ASSERT(src0->type == GGML_TYPE_F32); + GGML_ASSERT(src0_grad->type == GGML_TYPE_F32); + GGML_ASSERT(src0_grad_m->type == GGML_TYPE_F32); + GGML_ASSERT(src0_grad_v->type == GGML_TYPE_F32); + GGML_ASSERT(adamw_params->type == GGML_TYPE_F32); GGML_ASSERT(ggml_is_contiguous(src0)); GGML_ASSERT(ggml_is_contiguous(src0_grad)); GGML_ASSERT(ggml_is_contiguous(src0_grad_m)); GGML_ASSERT(ggml_is_contiguous(src0_grad_v)); + GGML_ASSERT(ggml_is_contiguous(adamw_params)); GGML_ASSERT(ggml_are_same_shape(src0, src0_grad)); GGML_ASSERT(ggml_are_same_shape(src0, src0_grad_m)); GGML_ASSERT(ggml_are_same_shape(src0, src0_grad_v)); + GGML_ASSERT(ggml_nelements(adamw_params) == 7); - float * src0_d = (float *) src0->data; - const float * src0_grad_d = (const float *) src0_grad->data; - float * src0_grad_m_d = (float *) src0_grad_m->data; - float * src0_grad_v_d = (float *) src0_grad_v->data; + float * src0_d = (float *) src0->data; + const float * src0_grad_d = (const float *) src0_grad->data; + float * src0_grad_m_d = (float *) src0_grad_m->data; + float * src0_grad_v_d = (float *) src0_grad_v->data; + const float * adamw_params_d = (const float *) adamw_params->data; cudaStream_t stream = ctx.stream(); const int64_t ne = ggml_nelements(src0); - int64_t iter; memcpy(&iter, &dst->op_params[0], sizeof(int64_t)); - float alpha; memcpy(&alpha, &dst->op_params[2], sizeof(float)); - float beta1; memcpy(&beta1, &dst->op_params[3], sizeof(float)); - float beta2; memcpy(&beta2, &dst->op_params[4], sizeof(float)); - float eps; memcpy(&eps, &dst->op_params[5], sizeof(float)); - float wd; memcpy(&wd, &dst->op_params[6], sizeof(float)); - - const float beta1h = alpha/(1.0f - powf(beta1, iter)); - const float beta2h = 1.0f/(1.0f - powf(beta2, iter)); - - opt_step_adamw_f32_cuda(src0_d, src0_grad_d, src0_grad_m_d, src0_grad_v_d, ne, alpha, beta1, beta2, eps, wd, beta1h, beta2h, stream); - - iter++; - memcpy(&dst->op_params[0], &iter, sizeof(int64_t)); + opt_step_adamw_f32_cuda(src0_d, src0_grad_d, src0_grad_m_d, src0_grad_v_d, adamw_params_d, ne, stream); } diff --git a/ggml/src/ggml-cuda/quantize.cu b/ggml/src/ggml-cuda/quantize.cu index 45408ce86..1702e4ce2 100644 --- a/ggml/src/ggml-cuda/quantize.cu +++ b/ggml/src/ggml-cuda/quantize.cu @@ -69,8 +69,8 @@ static __global__ void quantize_mmq_q8_1( // Exchange max. abs. value between vals_per_scale/4 threads. #pragma unroll - for (int mask = vals_per_scale/8; mask > 0; mask >>= 1) { - amax = fmaxf(amax, __shfl_xor_sync(0xFFFFFFFF, amax, mask, WARP_SIZE)); + for (int offset = vals_per_scale/8; offset > 0; offset >>= 1) { + amax = fmaxf(amax, __shfl_xor_sync(0xFFFFFFFF, amax, offset, WARP_SIZE)); } float sum; @@ -79,8 +79,8 @@ static __global__ void quantize_mmq_q8_1( // Exchange calculate sum across vals_per_sum/4 threads. #pragma unroll - for (int mask = vals_per_sum/8; mask > 0; mask >>= 1) { - sum += __shfl_xor_sync(0xFFFFFFFF, sum, mask, WARP_SIZE); + for (int offset = vals_per_sum/8; offset > 0; offset >>= 1) { + sum += __shfl_xor_sync(0xFFFFFFFF, sum, offset, WARP_SIZE); } } diff --git a/ggml/src/ggml-cuda/rope.cu b/ggml/src/ggml-cuda/rope.cu index 88f586d68..e1912fee1 100644 --- a/ggml/src/ggml-cuda/rope.cu +++ b/ggml/src/ggml-cuda/rope.cu @@ -4,6 +4,11 @@ struct rope_corr_dims { float v[2]; }; + +struct mrope_sections { + int v[4]; +}; + static __device__ float rope_yarn_ramp(const float low, const float high, const int i0) { const float y = (i0 / 2 - low) / max(0.001f, high - low); return 1.0f - min(1.0f, max(0.0f, y)); @@ -11,9 +16,10 @@ static __device__ float rope_yarn_ramp(const float low, const float high, const // YaRN algorithm based on LlamaYaRNScaledRotaryEmbedding.py from https://github.com/jquesnelle/yarn // MIT licensed. Copyright (c) 2023 Jeffrey Quesnelle and Bowen Peng. +template static __device__ void rope_yarn( - float theta_extrap, float freq_scale, rope_corr_dims corr_dims, int64_t i0, float ext_factor, float mscale, - float * cos_theta, float * sin_theta) { + const float theta_extrap, const float freq_scale, const rope_corr_dims corr_dims, const int64_t i0, const float ext_factor, + float mscale, float & cos_theta, float & sin_theta) { // Get n-d rotational scaling corrected for extrapolation float theta_interp = freq_scale * theta_extrap; float theta = theta_interp; @@ -24,24 +30,28 @@ static __device__ void rope_yarn( // Get n-d magnitude scaling corrected for interpolation mscale *= 1.0f + 0.1f * logf(1.0f / freq_scale); } - *cos_theta = cosf(theta) * mscale; - *sin_theta = sinf(theta) * mscale; + cos_theta = cosf(theta) * mscale; + sin_theta = sinf(theta) * mscale; + if (!forward) { + sin_theta *= -1.0f; + } } -template +template static __global__ void rope_norm( - const T * x, T * dst, int ne0, int n_dims, const int32_t * pos, float freq_scale, int p_delta_rows, - float ext_factor, float attn_factor, rope_corr_dims corr_dims, float theta_scale, const float * freq_factors) { + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int s1, const int s2, const int n_dims, + const int32_t * __restrict__ pos, const float freq_scale, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float theta_scale, const float * __restrict__ freq_factors) { const int i0 = 2*(blockDim.y*blockIdx.y + threadIdx.y); if (i0 >= ne0) { return; } - const int row = blockDim.x*blockIdx.x + threadIdx.x; + const int row_dst = blockDim.x*blockIdx.x + threadIdx.x; if (i0 >= n_dims) { - const int i = row*ne0 + i0; + const int i = row_dst*ne0 + i0; dst[i + 0] = x[i + 0]; dst[i + 1] = x[i + 1]; @@ -49,39 +59,43 @@ static __global__ void rope_norm( return; } - const int i = row*ne0 + i0; - const int i2 = row/p_delta_rows; + const int row_x = row_dst % ne1; + const int channel_x = row_dst / ne1; - const float theta_base = pos[i2]*powf(theta_scale, i0/2.0f); + const int idst = row_dst*ne0 + i0; + const int ix = channel_x*s2 + row_x*s1 + i0; + + const float theta_base = pos[channel_x]*powf(theta_scale, i0/2.0f); const float freq_factor = has_ff ? freq_factors[i0/2] : 1.0f; float cos_theta; float sin_theta; - rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta); + rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, cos_theta, sin_theta); - const float x0 = x[i + 0]; - const float x1 = x[i + 1]; + const float x0 = x[ix + 0]; + const float x1 = x[ix + 1]; - dst[i + 0] = x0*cos_theta - x1*sin_theta; - dst[i + 1] = x0*sin_theta + x1*cos_theta; + dst[idst + 0] = x0*cos_theta - x1*sin_theta; + dst[idst + 1] = x0*sin_theta + x1*cos_theta; } -template +template static __global__ void rope_neox( - const T * x, T * dst, int ne0, int n_dims, const int32_t * pos, float freq_scale, int p_delta_rows, - float ext_factor, float attn_factor, rope_corr_dims corr_dims, float theta_scale, const float * freq_factors) { + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int s1, const int s2, const int n_dims, + const int32_t * __restrict__ pos, const float freq_scale, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float theta_scale, const float * __restrict__ freq_factors) { const int i0 = 2*(blockDim.y*blockIdx.y + threadIdx.y); if (i0 >= ne0) { return; } - const int row = blockDim.x*blockIdx.x + threadIdx.x; + const int row_dst = blockDim.x*blockIdx.x + threadIdx.x; if (i0 >= n_dims) { - const int i = row*ne0 + i0; + const int i = row_dst*ne0 + i0; dst[i + 0] = x[i + 0]; dst[i + 1] = x[i + 1]; @@ -89,29 +103,140 @@ static __global__ void rope_neox( return; } - const int i = row*ne0 + i0/2; - const int i2 = row/p_delta_rows; + const int row_x = row_dst % ne1; + const int channel_x = row_dst / ne1; - const float theta_base = pos[i2]*powf(theta_scale, i0/2.0f); + const int idst = row_dst*ne0 + i0/2; + const int ix = channel_x*s2 + row_x*s1 + i0/2; + + const float theta_base = pos[channel_x]*powf(theta_scale, i0/2.0f); const float freq_factor = has_ff ? freq_factors[i0/2] : 1.0f; float cos_theta; float sin_theta; - rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta); + rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, cos_theta, sin_theta); - const float x0 = x[i + 0]; - const float x1 = x[i + n_dims/2]; + const float x0 = x[ix + 0]; + const float x1 = x[ix + n_dims/2]; - dst[i + 0] = x0*cos_theta - x1*sin_theta; - dst[i + n_dims/2] = x0*sin_theta + x1*cos_theta; + dst[idst + 0] = x0*cos_theta - x1*sin_theta; + dst[idst + n_dims/2] = x0*sin_theta + x1*cos_theta; } -template +template +static __global__ void rope_multi( + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int ne2, const int s1, const int s2, + const int n_dims, const int32_t * __restrict__ pos, const float freq_scale, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float theta_scale, const float * __restrict__ freq_factors, const mrope_sections sections) { + const int i0 = 2*(blockDim.y*blockIdx.y + threadIdx.y); + + if (i0 >= ne0) { + return; + } + + const int row_dst = blockDim.x*blockIdx.x + threadIdx.x; + + if (i0 >= n_dims) { + const int i = row_dst*ne0 + i0; + + dst[i + 0] = x[i + 0]; + dst[i + 1] = x[i + 1]; + + return; + } + + const int row_x = row_dst % ne1; + const int channel_x = row_dst / ne1; + + const int idst = row_dst*ne0 + i0/2; + const int ix = channel_x*s2 + row_x*s1 + i0/2; + + const int sect_dims = sections.v[0] + sections.v[1] + sections.v[2] + sections.v[3]; + const int sec_w = sections.v[1] + sections.v[0]; + const int sector = (i0 / 2) % sect_dims; + + float theta_base = 0.0; + if (sector < sections.v[0]) { + theta_base = pos[channel_x]*powf(theta_scale, i0/2.0f); + } + else if (sector >= sections.v[0] && sector < sec_w) { + theta_base = pos[channel_x + ne2 * 1]*powf(theta_scale, i0/2.0f); + } + else if (sector >= sec_w && sector < sec_w + sections.v[2]) { + theta_base = pos[channel_x + ne2 * 2]*powf(theta_scale, i0/2.0f); + } + else if (sector >= sec_w + sections.v[2]) { + theta_base = pos[channel_x + ne2 * 3]*powf(theta_scale, i0/2.0f); + } + + const float freq_factor = has_ff ? freq_factors[i0/2] : 1.0f; + + float cos_theta; + float sin_theta; + + rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, cos_theta, sin_theta); + + const float x0 = x[ix + 0]; + const float x1 = x[ix + n_dims/2]; + + dst[idst + 0] = x0*cos_theta - x1*sin_theta; + dst[idst + n_dims/2] = x0*sin_theta + x1*cos_theta; +} + +template +static __global__ void rope_vision( + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int ne2, const int s1, const int s2, const int n_dims, + const int32_t * __restrict__ pos, const float freq_scale, const float ext_factor, const float attn_factor, const rope_corr_dims corr_dims, + const float theta_scale, const float * __restrict__ freq_factors, const mrope_sections sections) { + const int i0 = 2*(blockDim.y*blockIdx.y + threadIdx.y); + + if (i0 >= ne0) { + return; + } + + const int row_dst = blockDim.x*blockIdx.x + threadIdx.x; + + const int row_x = row_dst % ne1; + const int channel_x = row_dst / ne1; + + const int idst = row_dst*ne0 + i0/2; + const int ix = channel_x*s2 + row_x*s1 + i0/2; + + const int sect_dims = sections.v[0] + sections.v[1]; + const int sec_w = sections.v[1] + sections.v[0]; + const int sector = (i0 / 2) % sect_dims; + + float theta_base = 0.0; + if (sector < sections.v[0]) { + const int p = sector; + theta_base = pos[channel_x]*powf(theta_scale, p); + } + else if (sector >= sections.v[0] && sector < sec_w) { + const int p = sector - sections.v[0]; + theta_base = pos[channel_x + ne2]*powf(theta_scale, p); + } + + const float freq_factor = has_ff ? freq_factors[i0/2] : 1.0f; + + float cos_theta; + float sin_theta; + + rope_yarn(theta_base/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, cos_theta, sin_theta); + + const float x0 = x[ix + 0]; + const float x1 = x[ix + n_dims]; + + dst[idst + 0] = x0*cos_theta - x1*sin_theta; + dst[idst + n_dims] = x0*sin_theta + x1*cos_theta; +} + +template static void rope_norm_cuda( - const T * x, T * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream) { + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int s1, const int s2, const int n_dims, const int nr, + const int32_t * __restrict__ pos, const float freq_scale, const float freq_base, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float * __restrict__ freq_factors, cudaStream_t stream) { GGML_ASSERT(ne0 % 2 == 0); const dim3 block_dims(1, CUDA_ROPE_BLOCK_SIZE, 1); const int n_blocks_x = (ne0 + 2*CUDA_ROPE_BLOCK_SIZE - 1) / (2*CUDA_ROPE_BLOCK_SIZE); @@ -120,22 +245,21 @@ static void rope_norm_cuda( const float theta_scale = powf(freq_base, -2.0f/n_dims); if (freq_factors == nullptr) { - rope_norm<<>>( - x, dst, ne0, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, - theta_scale, freq_factors - ); + rope_norm<<>>( + x, dst, ne0, ne1, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors); } else { - rope_norm<<>>( - x, dst, ne0, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, - theta_scale, freq_factors - ); + rope_norm<<>>( + x, dst, ne0, ne1, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors); } } -template +template static void rope_neox_cuda( - const T * x, T * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream) { + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int s1, const int s2, const int n_dims, const int nr, + const int32_t * __restrict__ pos, const float freq_scale, const float freq_base, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float * __restrict__ freq_factors, cudaStream_t stream) { GGML_ASSERT(ne0 % 2 == 0); const dim3 block_dims(1, CUDA_ROPE_BLOCK_SIZE, 1); const int n_blocks_x = (ne0 + 2*CUDA_ROPE_BLOCK_SIZE - 1) / (2*CUDA_ROPE_BLOCK_SIZE); @@ -144,48 +268,66 @@ static void rope_neox_cuda( const float theta_scale = powf(freq_base, -2.0f/n_dims); if (freq_factors == nullptr) { - rope_neox<<>>( - x, dst, ne0, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, - theta_scale, freq_factors - ); + rope_neox<<>>( + x, dst, ne0, ne1, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors); } else { - rope_neox<<>>( - x, dst, ne0, n_dims, pos, freq_scale, p_delta_rows, ext_factor, attn_factor, corr_dims, - theta_scale, freq_factors - ); + rope_neox<<>>( + x, dst, ne0, ne1, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors); } } -static void rope_norm_cuda_f16( - const half * x, half * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream) { +template +static void rope_multi_cuda( + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int ne2, const int s1, const int s2, const int n_dims, const int nr, + const int32_t * __restrict__ pos, const float freq_scale, const float freq_base, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float * __restrict__ freq_factors, const mrope_sections sections, cudaStream_t stream) { + GGML_ASSERT(ne0 % 2 == 0); + const dim3 block_dims(1, CUDA_ROPE_BLOCK_SIZE, 1); + const int n_blocks_x = (ne0 + 2*CUDA_ROPE_BLOCK_SIZE - 1) / (2*CUDA_ROPE_BLOCK_SIZE); + const dim3 block_nums(nr, n_blocks_x, 1); - rope_norm_cuda(x, dst, ne0, n_dims, nr, pos, freq_scale, p_delta_rows, freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); + const float theta_scale = powf(freq_base, -2.0f/n_dims); + + if (freq_factors == nullptr) { + rope_multi<<>>( + x, dst, ne0, ne1, ne2, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors, sections); + } else { + rope_multi<<>>( + x, dst, ne0, ne1, ne2, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors, sections); + } } -static void rope_norm_cuda_f32( - const float * x, float * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream) { +template +static void rope_vision_cuda( + const T * __restrict__ x, T * __restrict__ dst, const int ne0, const int ne1, const int ne2, const int s1, const int s2, const int n_dims, const int nr, + const int32_t * __restrict__ pos, const float freq_scale, const float freq_base, const float ext_factor, const float attn_factor, + const rope_corr_dims corr_dims, const float * __restrict__ freq_factors, const mrope_sections sections, cudaStream_t stream) { + GGML_ASSERT(ne0 % 2 == 0); + const dim3 block_dims(1, CUDA_ROPE_BLOCK_SIZE, 1); + const int n_blocks_x = (ne0 + 2*CUDA_ROPE_BLOCK_SIZE - 1) / (2*CUDA_ROPE_BLOCK_SIZE); + const dim3 block_nums(nr, n_blocks_x, 1); + // break down (head_dim, heads, seq) into (CUDA_ROPE_BLOCK_SIZE, x, heads * seq) + // where x ~= ceil(head_dim / CUDA_ROPE_BLOCK_SIZE); - rope_norm_cuda(x, dst, ne0, n_dims, nr, pos, freq_scale, p_delta_rows, freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); + const float theta_scale = powf(freq_base, -2.0f/n_dims); + + if (freq_factors == nullptr) { + rope_vision<<>>( + x, dst, ne0, ne1, ne2, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors, sections); + } else { + rope_vision<<>>( + x, dst, ne0, ne1, ne2, s1, s2, n_dims, pos, freq_scale, ext_factor, + attn_factor, corr_dims, theta_scale, freq_factors, sections); + } } -static void rope_neox_cuda_f16( - const half * x, half * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream) { - - rope_neox_cuda(x, dst, ne0, n_dims, nr, pos, freq_scale, p_delta_rows, freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); -} - -static void rope_neox_cuda_f32( - const float * x, float * dst, int ne0, int n_dims, int nr, const int32_t * pos, float freq_scale, int p_delta_rows, - float freq_base, float ext_factor, float attn_factor, rope_corr_dims corr_dims, const float * freq_factors, cudaStream_t stream -) { - - rope_neox_cuda(x, dst, ne0, n_dims, nr, pos, freq_scale, p_delta_rows, freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); -} - -void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { +template +void ggml_cuda_op_rope_impl(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { const ggml_tensor * src0 = dst->src[0]; const ggml_tensor * src1 = dst->src[1]; const ggml_tensor * src2 = dst->src[2]; @@ -196,20 +338,24 @@ void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { float * dst_d = (float *)dst->data; cudaStream_t stream = ctx.stream(); - GGML_ASSERT(ggml_is_contiguous(src0)); GGML_ASSERT(src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16); GGML_ASSERT( dst->type == GGML_TYPE_F32 || dst->type == GGML_TYPE_F16); GGML_ASSERT(src0->type == dst->type); - const int64_t ne00 = src0->ne[0]; - const int64_t ne01 = src0->ne[1]; + const int64_t ne00 = src0->ne[0]; // head dims + const int64_t ne01 = src0->ne[1]; // num heads + const int64_t ne02 = src0->ne[2]; // num heads const int64_t nr = ggml_nrows(src0); + const size_t s01 = src0->nb[1] / ggml_type_size(src0->type); + const size_t s02 = src0->nb[2] / ggml_type_size(src0->type); + //const int n_past = ((int32_t *) dst->op_params)[0]; const int n_dims = ((int32_t *) dst->op_params)[1]; const int mode = ((int32_t *) dst->op_params)[2]; //const int n_ctx = ((int32_t *) dst->op_params)[3]; const int n_ctx_orig = ((int32_t *) dst->op_params)[4]; + mrope_sections sections; // RoPE alteration for extended context float freq_base; @@ -225,8 +371,19 @@ void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { memcpy(&attn_factor, (int32_t *) dst->op_params + 8, sizeof(float)); memcpy(&beta_fast, (int32_t *) dst->op_params + 9, sizeof(float)); memcpy(&beta_slow, (int32_t *) dst->op_params + 10, sizeof(float)); + memcpy(§ions.v, (int32_t *) dst->op_params + 11, sizeof(int)*4); const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; + const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE; + const bool is_vision = mode == GGML_ROPE_TYPE_VISION; + + if (is_mrope) { + GGML_ASSERT(sections.v[0] > 0 || sections.v[1] > 0 || sections.v[2] > 0); + } + + if (is_vision) { + GGML_ASSERT(n_dims == ne00/2); + } const int32_t * pos = (const int32_t *) src1_d; @@ -241,31 +398,59 @@ void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { // compute if (is_neox) { if (src0->type == GGML_TYPE_F32) { - rope_neox_cuda_f32( - (const float *)src0_d, (float *)dst_d, ne00, n_dims, nr, pos, freq_scale, ne01, freq_base, ext_factor, - attn_factor, corr_dims, freq_factors, stream - ); + rope_neox_cuda( + (const float *) src0_d, (float *) dst_d, ne00, ne01, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); } else if (src0->type == GGML_TYPE_F16) { - rope_neox_cuda_f16( - (const half *)src0_d, (half *)dst_d, ne00, n_dims, nr, pos, freq_scale, ne01, freq_base, ext_factor, - attn_factor, corr_dims, freq_factors, stream - ); + rope_neox_cuda( + (const half *) src0_d, (half *) dst_d, ne00, ne01, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); + } else { + GGML_ABORT("fatal error"); + } + } else if (is_mrope && !is_vision) { + if (src0->type == GGML_TYPE_F32) { + rope_multi_cuda( + (const float *) src0_d, (float *) dst_d, ne00, ne01, ne02, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, sections, stream); + } else if (src0->type == GGML_TYPE_F16) { + rope_multi_cuda( + (const half *) src0_d, (half *) dst_d, ne00, ne01, ne02, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, sections, stream); + } else { + GGML_ABORT("fatal error"); + } + } else if (is_vision) { + if (src0->type == GGML_TYPE_F32) { + rope_vision_cuda( + (const float *) src0_d, (float *) dst_d, ne00, ne01, ne02, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, sections, stream); + } else if (src0->type == GGML_TYPE_F16) { + rope_vision_cuda( + (const half *) src0_d, (half *) dst_d, ne00, ne01, ne02, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, sections, stream); } else { GGML_ABORT("fatal error"); } } else { if (src0->type == GGML_TYPE_F32) { - rope_norm_cuda_f32( - (const float *)src0_d, (float *)dst_d, ne00, n_dims, nr, pos, freq_scale, ne01, freq_base, ext_factor, - attn_factor, corr_dims, freq_factors, stream - ); + rope_norm_cuda( + (const float *) src0_d, (float *) dst_d, ne00, ne01, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); } else if (src0->type == GGML_TYPE_F16) { - rope_norm_cuda_f16( - (const half *)src0_d, (half *)dst_d, ne00, n_dims, nr, pos, freq_scale, ne01, freq_base, ext_factor, - attn_factor, corr_dims, freq_factors, stream - ); + rope_norm_cuda( + (const half *) src0_d, (half *) dst_d, ne00, ne01, s01, s02, n_dims, nr, pos, freq_scale, + freq_base, ext_factor, attn_factor, corr_dims, freq_factors, stream); } else { GGML_ABORT("fatal error"); } } } + +void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + ggml_cuda_op_rope_impl(ctx, dst); +} + +void ggml_cuda_op_rope_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + ggml_cuda_op_rope_impl(ctx, dst); +} diff --git a/ggml/src/ggml-cuda/rope.cuh b/ggml/src/ggml-cuda/rope.cuh index 0f787a0b2..9139f3b22 100644 --- a/ggml/src/ggml-cuda/rope.cuh +++ b/ggml/src/ggml-cuda/rope.cuh @@ -3,3 +3,5 @@ #define CUDA_ROPE_BLOCK_SIZE 256 void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst); + +void ggml_cuda_op_rope_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst); diff --git a/ggml/src/ggml-cuda/sum.cu b/ggml/src/ggml-cuda/sum.cu index 0583e4fe0..e0dafc1d2 100644 --- a/ggml/src/ggml-cuda/sum.cu +++ b/ggml/src/ggml-cuda/sum.cu @@ -1,10 +1,8 @@ -#if !defined(GGML_USE_HIPBLAS) && !defined(GGML_USE_MUSA) && CUDART_VERSION >= 11700 +#if !defined(GGML_USE_HIP) && !defined(GGML_USE_MUSA) && CUDART_VERSION >= 11700 #define USE_CUB -#endif // !defined(GGML_USE_HIPBLAS) && !defined(GGML_USE_MUSA) && CUDART_VERSION >= 11700 +#endif // !defined(GGML_USE_HIP) && !defined(GGML_USE_MUSA) && CUDART_VERSION >= 11700 #ifdef USE_CUB -// On Windows CUB uses libraries with variables called CC_PASCAL which conflict with the define in common.cuh. -// For this reason CUB must be included BEFORE anything else. #include using namespace cub; #endif // USE_CUB diff --git a/ggml/src/ggml-cuda/vendors/cuda.h b/ggml/src/ggml-cuda/vendors/cuda.h index db9f6a165..1746b0732 100644 --- a/ggml/src/ggml-cuda/vendors/cuda.h +++ b/ggml/src/ggml-cuda/vendors/cuda.h @@ -3,6 +3,7 @@ #include #include #include +#include #include #if CUDART_VERSION < 11020 diff --git a/ggml/src/ggml-cuda/vendors/hip.h b/ggml/src/ggml-cuda/vendors/hip.h index 1f3c70c2e..c905b15d7 100644 --- a/ggml/src/ggml-cuda/vendors/hip.h +++ b/ggml/src/ggml-cuda/vendors/hip.h @@ -3,6 +3,7 @@ #include #include #include +#include #ifdef __HIP_PLATFORM_AMD__ // for rocblas_initialize() #include "rocblas/rocblas.h" @@ -95,6 +96,14 @@ #define __CUDA_ARCH__ 1300 +#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) +#define GCN +#endif + +#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__) +#define CDNA +#endif + #if defined(__gfx1100__) || defined(__gfx1101__) || defined(__gfx1102__) || defined(__gfx1103__) || \ defined(__gfx1150__) || defined(__gfx1151__) #define RDNA3 @@ -113,6 +122,8 @@ #define __has_builtin(x) 0 #endif +typedef hip_bfloat16 nv_bfloat16; + typedef int8_t int8x4_t __attribute__((ext_vector_type(4))); typedef uint8_t uint8x4_t __attribute__((ext_vector_type(4))); static __device__ __forceinline__ int __vsubss4(const int a, const int b) { diff --git a/ggml/src/ggml-cuda/vendors/musa.h b/ggml/src/ggml-cuda/vendors/musa.h index 1604b8229..6cc1b69ee 100644 --- a/ggml/src/ggml-cuda/vendors/musa.h +++ b/ggml/src/ggml-cuda/vendors/musa.h @@ -3,6 +3,7 @@ #include #include #include +#include #include #define CUBLAS_COMPUTE_16F CUDA_R_16F #define CUBLAS_COMPUTE_32F CUDA_R_32F @@ -132,3 +133,5 @@ #define cudaKernelNodeParams musaKernelNodeParams #define cudaStreamCaptureModeRelaxed musaStreamCaptureModeRelaxed #define cudaStreamEndCapture musaStreamEndCapture + +typedef mt_bfloat16 nv_bfloat16; diff --git a/ggml/src/ggml-cuda/wkv6.cu b/ggml/src/ggml-cuda/wkv6.cu index 42578341a..bbdafbee5 100644 --- a/ggml/src/ggml-cuda/wkv6.cu +++ b/ggml/src/ggml-cuda/wkv6.cu @@ -73,9 +73,9 @@ void ggml_cuda_op_rwkv_wkv6(ggml_backend_cuda_context & ctx, ggml_tensor * dst) const float * s_d = (const float *)dst->src[5]->data; const int64_t B = dst->src[5]->ne[1]; - const int64_t T = dst->src[0]->ne[3]; + const int64_t T = dst->src[0]->ne[2]; const int64_t C = dst->ne[0]; - const int64_t H = dst->src[0]->ne[2]; + const int64_t H = dst->src[0]->ne[1]; float * dst_d = (float *)dst->data; diff --git a/ggml/src/ggml-hip/CMakeLists.txt b/ggml/src/ggml-hip/CMakeLists.txt new file mode 100644 index 000000000..d090ba9bd --- /dev/null +++ b/ggml/src/ggml-hip/CMakeLists.txt @@ -0,0 +1,106 @@ +if (NOT EXISTS $ENV{ROCM_PATH}) + if (NOT EXISTS /opt/rocm) + set(ROCM_PATH /usr) + else() + set(ROCM_PATH /opt/rocm) + endif() +else() + set(ROCM_PATH $ENV{ROCM_PATH}) +endif() + +list(APPEND CMAKE_PREFIX_PATH ${ROCM_PATH}) +list(APPEND CMAKE_PREFIX_PATH "${ROCM_PATH}/lib64/cmake") + +# CMake on Windows doesn't support the HIP language yet +if (WIN32) + set(CXX_IS_HIPCC TRUE) +else() + string(REGEX MATCH "hipcc(\.bat)?$" CXX_IS_HIPCC "${CMAKE_CXX_COMPILER}") +endif() + +if (CXX_IS_HIPCC) + if (LINUX) + if (NOT ${CMAKE_CXX_COMPILER_ID} MATCHES "Clang") + message(WARNING "Only LLVM is supported for HIP, hint: CXX=/opt/rocm/llvm/bin/clang++") + endif() + + message(WARNING "Setting hipcc as the C++ compiler is legacy behavior." + " Prefer setting the HIP compiler directly. See README for details.") + endif() +else() + # Forward AMDGPU_TARGETS to CMAKE_HIP_ARCHITECTURES. + if (AMDGPU_TARGETS AND NOT CMAKE_HIP_ARCHITECTURES) + set(CMAKE_HIP_ARCHITECTURES ${AMDGPU_TARGETS}) + endif() + cmake_minimum_required(VERSION 3.21) + enable_language(HIP) +endif() + +find_package(hip REQUIRED) +find_package(hipblas REQUIRED) +find_package(rocblas REQUIRED) + +message(STATUS "HIP and hipBLAS found") + +file(GLOB GGML_HEADERS_ROCM "../ggml-cuda/*.cuh") +list(APPEND GGML_HEADERS_ROCM "../../include/ggml-cuda.h") + +file(GLOB GGML_SOURCES_ROCM "../ggml-cuda/*.cu") +file(GLOB SRCS "../ggml-cuda/template-instances/fattn-wmma*.cu") +list(APPEND GGML_SOURCES_ROCM ${SRCS}) +file(GLOB SRCS "../ggml-cuda/template-instances/mmq*.cu") +list(APPEND GGML_SOURCES_ROCM ${SRCS}) + +if (GGML_CUDA_FA_ALL_QUANTS) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*.cu") + list(APPEND GGML_SOURCES_ROCM ${SRCS}) + add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) +else() + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu") + list(APPEND GGML_SOURCES_ROCM ${SRCS}) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu") + list(APPEND GGML_SOURCES_ROCM ${SRCS}) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*f16-f16.cu") + list(APPEND GGML_SOURCES_ROCM ${SRCS}) +endif() + +ggml_add_backend_library(ggml-hip + ${GGML_HEADERS_ROCM} + ${GGML_SOURCES_ROCM} + ) + +# TODO: do not use CUDA definitions for HIP +if (NOT GGML_BACKEND_DL) + target_compile_definitions(ggml PUBLIC GGML_USE_CUDA) +endif() + +add_compile_definitions(GGML_USE_HIP) + +if (GGML_HIP_UMA) + add_compile_definitions(GGML_HIP_UMA) +endif() + +if (GGML_CUDA_FORCE_MMQ) + add_compile_definitions(GGML_CUDA_FORCE_MMQ) +endif() + +if (GGML_CUDA_FORCE_CUBLAS) + add_compile_definitions(GGML_CUDA_FORCE_CUBLAS) +endif() + +if (GGML_CUDA_NO_PEER_COPY) + add_compile_definitions(GGML_CUDA_NO_PEER_COPY) +endif() + +if (CXX_IS_HIPCC) + set_source_files_properties(${GGML_SOURCES_ROCM} PROPERTIES LANGUAGE CXX) + target_link_libraries(ggml-hip PRIVATE hip::device) +else() + set_source_files_properties(${GGML_SOURCES_ROCM} PROPERTIES LANGUAGE HIP) +endif() + +if (GGML_STATIC) + message(FATAL_ERROR "Static linking not supported for HIP/ROCm") +endif() + +target_link_libraries(ggml-hip PRIVATE ggml-base hip::host roc::rocblas roc::hipblas) diff --git a/ggml/src/ggml-impl.h b/ggml/src/ggml-impl.h index af29a26f0..eab017889 100644 --- a/ggml/src/ggml-impl.h +++ b/ggml/src/ggml-impl.h @@ -3,22 +3,42 @@ // GGML internal header #include "ggml.h" +#include "gguf.h" #include +#include #include // load `stdlib.h` before other headers to work around MinGW bug: https://sourceforge.net/p/mingw-w64/bugs/192/ #include #include #include +#ifdef __ARM_FEATURE_SVE +#include +#endif // __ARM_FEATURE_SVE + +#if defined(__ARM_NEON) && !defined(__CUDACC__) +// if YCM cannot find , make a symbolic link to it, for example: +// +// $ ln -sfn /Library/Developer/CommandLineTools/usr/lib/clang/13.1.6/include/arm_neon.h ./src/ +// +#include +#endif + +#if defined(__F16C__) +#include +#endif + #ifdef __cplusplus extern "C" { #endif -#undef MIN -#undef MAX +#ifndef MIN +# define MIN(a, b) ((a) < (b) ? (a) : (b)) +#endif -#define MIN(a, b) ((a) < (b) ? (a) : (b)) -#define MAX(a, b) ((a) > (b) ? (a) : (b)) +#ifndef MAX +# define MAX(a, b) ((a) > (b) ? (a) : (b)) +#endif // required for mmap as gguf only guarantees 32-byte alignment #define TENSOR_ALIGNMENT 32 @@ -28,13 +48,13 @@ extern "C" { // if C99 - static_assert is noop // ref: https://stackoverflow.com/a/53923785/4039976 #ifndef __cplusplus -#ifndef static_assert -#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201100L) -#define static_assert(cond, msg) _Static_assert(cond, msg) -#else -#define static_assert(cond, msg) struct global_scope_noop_trick -#endif -#endif + #ifndef static_assert + #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201100L) + #define static_assert(cond, msg) _Static_assert(cond, msg) + #else + #define static_assert(cond, msg) struct global_scope_noop_trick + #endif + #endif #endif static inline int ggml_up32(int n) { @@ -56,8 +76,8 @@ static inline int ggml_up(int n, int m) { // GGML_ATTRIBUTE_FORMAT(2, 3) -void ggml_log_internal (enum ggml_log_level level, const char * format, ...); -void ggml_log_callback_default(enum ggml_log_level level, const char * text, void * user_data); +GGML_API void ggml_log_internal (enum ggml_log_level level, const char * format, ...); +GGML_API void ggml_log_callback_default(enum ggml_log_level level, const char * text, void * user_data); #define GGML_LOG(...) ggml_log_internal(GGML_LOG_LEVEL_NONE , __VA_ARGS__) #define GGML_LOG_INFO(...) ggml_log_internal(GGML_LOG_LEVEL_INFO , __VA_ARGS__) @@ -120,14 +140,12 @@ struct ggml_map_custom1_op_params { void * userdata; }; - struct ggml_map_custom2_op_params { ggml_custom2_op_t fun; int n_tasks; void * userdata; }; - struct ggml_map_custom3_op_params { ggml_custom3_op_t fun; int n_tasks; @@ -182,7 +200,7 @@ void ggml_hash_set_reset(struct ggml_hash_set * hash_set); static bool ggml_hash_contains(const struct ggml_hash_set * hash_set, struct ggml_tensor * key); // returns GGML_HASHSET_FULL if table is full, otherwise the current index of the key or where it should be inserted -static size_t ggml_hash_find(const struct ggml_hash_set * hash_set, struct ggml_tensor * key); +static size_t ggml_hash_find(const struct ggml_hash_set * hash_set, const struct ggml_tensor * key); // returns GGML_HASHSET_ALREADY_EXISTS if key already exists, index otherwise, asserts if table is full static size_t ggml_hash_insert(struct ggml_hash_set * hash_set, struct ggml_tensor * key); @@ -196,7 +214,7 @@ static inline size_t ggml_hash(const struct ggml_tensor * p) { return (size_t)(uintptr_t)p >> 4; } -static size_t ggml_hash_find(const struct ggml_hash_set * hash_set, struct ggml_tensor * key) { +static size_t ggml_hash_find(const struct ggml_hash_set * hash_set, const struct ggml_tensor * key) { size_t h = ggml_hash(key) % hash_set->size; // linear probing @@ -267,30 +285,283 @@ enum ggml_cgraph_eval_order { }; struct ggml_cgraph { - int size; - int n_nodes; - int n_leafs; + int size; // maximum number of nodes/leafs/grads/grad_accs + int n_nodes; // number of nodes currently in use + int n_leafs; // number of leafs currently in use - struct ggml_tensor ** nodes; - struct ggml_tensor ** grads; - struct ggml_tensor ** leafs; + struct ggml_tensor ** nodes; // tensors with data that can change if the graph is evaluated + struct ggml_tensor ** grads; // the outputs of these tensors are the gradients of the nodes + struct ggml_tensor ** grad_accs; // accumulators for node gradients + struct ggml_tensor ** leafs; // tensors with constant data struct ggml_hash_set visited_hash_set; enum ggml_cgraph_eval_order order; }; +// returns a slice of cgraph with nodes [i0, i1) +// the slice does not have leafs or gradients +// if you need the gradients, get them from the original graph struct ggml_cgraph ggml_graph_view(struct ggml_cgraph * cgraph, int i0, int i1); // Memory allocation -void * ggml_aligned_malloc(size_t size); -void ggml_aligned_free(void * ptr, size_t size); +GGML_API void * ggml_aligned_malloc(size_t size); +GGML_API void ggml_aligned_free(void * ptr, size_t size); -// TODO: move to threading file -void ggml_critical_section_start(void); -void ggml_critical_section_end(void); +// FP16 to FP32 conversion + +#if defined(__ARM_NEON) + #if defined(_MSC_VER) || (defined(__CUDACC__) && __CUDACC_VER_MAJOR__ <= 11) + typedef uint16_t ggml_fp16_internal_t; + #else + typedef __fp16 ggml_fp16_internal_t; + #endif +#endif + +#if defined(__ARM_NEON) && !defined(_MSC_VER) && !(defined(__CUDACC__) && __CUDACC_VER_MAJOR__ <= 11) + #define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) + #define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) + + #define GGML_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) + + static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { + ggml_fp16_internal_t tmp; + memcpy(&tmp, &h, sizeof(ggml_fp16_t)); + return (float)tmp; + } + + static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { + ggml_fp16_t res; + ggml_fp16_internal_t tmp = f; + memcpy(&res, &tmp, sizeof(ggml_fp16_t)); + return res; + } + +#elif defined(__F16C__) + + #ifdef _MSC_VER + #define GGML_COMPUTE_FP16_TO_FP32(x) _mm_cvtss_f32(_mm_cvtph_ps(_mm_cvtsi32_si128(x))) + #define GGML_COMPUTE_FP32_TO_FP16(x) _mm_extract_epi16(_mm_cvtps_ph(_mm_set_ss(x), 0), 0) + #else + #define GGML_COMPUTE_FP16_TO_FP32(x) _cvtsh_ss(x) + #define GGML_COMPUTE_FP32_TO_FP16(x) _cvtss_sh(x, 0) + #endif + +#elif defined(__POWER9_VECTOR__) + + #define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) + #define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) + /* the inline asm below is about 12% faster than the lookup method */ + #define GGML_FP16_TO_FP32(x) GGML_COMPUTE_FP16_TO_FP32(x) + #define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) + + static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { + register float f; + register double d; + __asm__( + "mtfprd %0,%2\n" + "xscvhpdp %0,%0\n" + "frsp %1,%0\n" : + /* temp */ "=d"(d), + /* out */ "=f"(f): + /* in */ "r"(h)); + return f; + } + + static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { + register double d; + register ggml_fp16_t r; + __asm__( /* xscvdphp can work on double or single precision */ + "xscvdphp %0,%2\n" + "mffprd %1,%0\n" : + /* temp */ "=d"(d), + /* out */ "=r"(r): + /* in */ "f"(f)); + return r; + } + +#else + + // FP16 <-> FP32 + // ref: https://github.com/Maratyszcza/FP16 + + static inline float fp32_from_bits(uint32_t w) { + union { + uint32_t as_bits; + float as_value; + } fp32; + fp32.as_bits = w; + return fp32.as_value; + } + + static inline uint32_t fp32_to_bits(float f) { + union { + float as_value; + uint32_t as_bits; + } fp32; + fp32.as_value = f; + return fp32.as_bits; + } + + static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { + const uint32_t w = (uint32_t) h << 16; + const uint32_t sign = w & UINT32_C(0x80000000); + const uint32_t two_w = w + w; + + const uint32_t exp_offset = UINT32_C(0xE0) << 23; + #if (defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) || defined(__GNUC__) && !defined(__STRICT_ANSI__)) && (!defined(__cplusplus) || __cplusplus >= 201703L) + const float exp_scale = 0x1.0p-112f; + #else + const float exp_scale = fp32_from_bits(UINT32_C(0x7800000)); + #endif + const float normalized_value = fp32_from_bits((two_w >> 4) + exp_offset) * exp_scale; + + const uint32_t magic_mask = UINT32_C(126) << 23; + const float magic_bias = 0.5f; + const float denormalized_value = fp32_from_bits((two_w >> 17) | magic_mask) - magic_bias; + + const uint32_t denormalized_cutoff = UINT32_C(1) << 27; + const uint32_t result = sign | + (two_w < denormalized_cutoff ? fp32_to_bits(denormalized_value) : fp32_to_bits(normalized_value)); + return fp32_from_bits(result); + } + + static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { + #if (defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) || defined(__GNUC__) && !defined(__STRICT_ANSI__)) && (!defined(__cplusplus) || __cplusplus >= 201703L) + const float scale_to_inf = 0x1.0p+112f; + const float scale_to_zero = 0x1.0p-110f; + #else + const float scale_to_inf = fp32_from_bits(UINT32_C(0x77800000)); + const float scale_to_zero = fp32_from_bits(UINT32_C(0x08800000)); + #endif + float base = (fabsf(f) * scale_to_inf) * scale_to_zero; + + const uint32_t w = fp32_to_bits(f); + const uint32_t shl1_w = w + w; + const uint32_t sign = w & UINT32_C(0x80000000); + uint32_t bias = shl1_w & UINT32_C(0xFF000000); + if (bias < UINT32_C(0x71000000)) { + bias = UINT32_C(0x71000000); + } + + base = fp32_from_bits((bias >> 1) + UINT32_C(0x07800000)) + base; + const uint32_t bits = fp32_to_bits(base); + const uint32_t exp_bits = (bits >> 13) & UINT32_C(0x00007C00); + const uint32_t mantissa_bits = bits & UINT32_C(0x00000FFF); + const uint32_t nonsign = exp_bits + mantissa_bits; + return (sign >> 16) | (shl1_w > UINT32_C(0xFF000000) ? UINT16_C(0x7E00) : nonsign); + } + + #define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) + #define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) + +#endif // defined(__ARM_NEON) && (!defined(__MSC_VER) + +// precomputed f32 table for f16 (256 KB) +// defined in ggml.c, initialized in ggml_init() +GGML_API float ggml_table_f32_f16[1 << 16]; + +// On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32, +// so we define GGML_FP16_TO_FP32 and GGML_FP32_TO_FP16 elsewhere for NEON. +// This is also true for POWER9. +#if !defined(GGML_FP16_TO_FP32) +inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) { + uint16_t s; + memcpy(&s, &f, sizeof(uint16_t)); + return ggml_table_f32_f16[s]; +} + +#define GGML_FP16_TO_FP32(x) ggml_lookup_fp16_to_fp32(x) +#endif + +#if !defined(GGML_FP32_TO_FP16) +#define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) +#endif + +/** + * Converts brain16 to float32. + * + * The bfloat16 floating point format has the following structure: + * + * ┌sign + * │ + * │ ┌exponent + * │ │ + * │ │ ┌mantissa + * │ │ │ + * │┌──┴───┐┌─┴───┐ + * 0b0000000000000000 brain16 + * + * Since bf16 has the same number of exponent bits as a 32bit float, + * encoding and decoding numbers becomes relatively straightforward. + * + * ┌sign + * │ + * │ ┌exponent + * │ │ + * │ │ ┌mantissa + * │ │ │ + * │┌──┴───┐┌─┴───────────────────┐ + * 0b00000000000000000000000000000000 IEEE binary32 + * + * For comparison, the standard fp16 format has fewer exponent bits. + * + * ┌sign + * │ + * │ ┌exponent + * │ │ + * │ │ ┌mantissa + * │ │ │ + * │┌─┴─┐┌─┴──────┐ + * 0b0000000000000000 IEEE binary16 + * + * @see IEEE 754-2008 + */ +static inline float ggml_compute_bf16_to_fp32(ggml_bf16_t h) { + union { + float f; + uint32_t i; + } u; + u.i = (uint32_t)h.bits << 16; + return u.f; +} + +/** + * Converts float32 to brain16. + * + * This is binary identical with Google Brain float conversion. + * Floats shall round to nearest even, and NANs shall be quiet. + * Subnormals aren't flushed to zero, except perhaps when used. + * This code should vectorize nicely if using modern compilers. + */ +static inline ggml_bf16_t ggml_compute_fp32_to_bf16(float s) { + ggml_bf16_t h; + union { + float f; + uint32_t i; + } u; + u.f = s; + if ((u.i & 0x7fffffff) > 0x7f800000) { /* nan */ + h.bits = (u.i >> 16) | 64; /* force to quiet */ + return h; + } + h.bits = (u.i + (0x7fff + ((u.i >> 16) & 1))) >> 16; + return h; +} + +#define GGML_FP32_TO_BF16(x) ggml_compute_fp32_to_bf16(x) +#define GGML_BF16_TO_FP32(x) ggml_compute_bf16_to_fp32(x) #ifdef __cplusplus } #endif + +#ifdef __cplusplus +#include + +// expose GGUF internals for test code +GGML_API size_t gguf_type_size(enum gguf_type type); +GGML_API struct gguf_context * gguf_init_from_file_impl(FILE * file, struct gguf_init_params params); +GGML_API void gguf_write_to_buf(const struct gguf_context * ctx, std::vector & buf, bool only_meta); +#endif // __cplusplus diff --git a/ggml/src/ggml-kompute/CMakeLists.txt b/ggml/src/ggml-kompute/CMakeLists.txt new file mode 100644 index 000000000..c9109d5e8 --- /dev/null +++ b/ggml/src/ggml-kompute/CMakeLists.txt @@ -0,0 +1,166 @@ + +find_package(Vulkan COMPONENTS glslc REQUIRED) +find_program(glslc_executable NAMES glslc HINTS Vulkan::glslc) + +if (NOT glslc_executable) + message(FATAL_ERROR "glslc not found") +endif() + +ggml_add_backend_library(ggml-kompute + ggml-kompute.cpp + ../../include/ggml-kompute.h + ) + +target_link_libraries(ggml-kompute PRIVATE ggml-base kompute) +target_include_directories(ggml-kompute PRIVATE ${CMAKE_CURRENT_BINARY_DIR}) + +add_compile_definitions(VULKAN_HPP_DISPATCH_LOADER_DYNAMIC=1) + +function(compile_shader) + set(options) + set(oneValueArgs) + set(multiValueArgs SOURCES) + cmake_parse_arguments(compile_shader "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN}) + foreach(source ${compile_shader_SOURCES}) + get_filename_component(filename ${source} NAME) + set(spv_file ${filename}.spv) + add_custom_command( + OUTPUT ${spv_file} + DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/${source} + ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/common.comp + ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_getrows.comp + ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_mul_mv_q_n_pre.comp + ${CMAKE_CURRENT_SOURCE_DIR}/kompute-shaders/op_mul_mv_q_n.comp + COMMAND ${glslc_executable} --target-env=vulkan1.2 -o ${spv_file} ${CMAKE_CURRENT_SOURCE_DIR}/${source} + COMMENT "Compiling ${source} to ${spv_file}" + ) + + get_filename_component(RAW_FILE_NAME ${spv_file} NAME) + set(FILE_NAME "shader${RAW_FILE_NAME}") + string(REPLACE ".comp.spv" ".h" HEADER_FILE ${FILE_NAME}) + string(TOUPPER ${HEADER_FILE} HEADER_FILE_DEFINE) + string(REPLACE "." "_" HEADER_FILE_DEFINE "${HEADER_FILE_DEFINE}") + set(OUTPUT_HEADER_FILE "${HEADER_FILE}") + message(STATUS "${HEADER_FILE} generating ${HEADER_FILE_DEFINE}") + if(CMAKE_GENERATOR MATCHES "Visual Studio") + add_custom_command( + OUTPUT ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "/*THIS FILE HAS BEEN AUTOMATICALLY GENERATED - DO NOT EDIT*/" > ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#ifndef ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "namespace kp {" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "namespace shader_data {" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_BINARY_DIR}/bin/$/xxd -i ${RAW_FILE_NAME} >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "}}" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#endif // define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + DEPENDS ${spv_file} xxd + COMMENT "Converting to hpp: ${FILE_NAME} ${CMAKE_BINARY_DIR}/bin/$/xxd" + ) + else() + add_custom_command( + OUTPUT ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "/*THIS FILE HAS BEEN AUTOMATICALLY GENERATED - DO NOT EDIT*/" > ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#ifndef ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "namespace kp {" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "namespace shader_data {" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_BINARY_DIR}/bin/xxd -i ${RAW_FILE_NAME} >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo "}}" >> ${OUTPUT_HEADER_FILE} + COMMAND ${CMAKE_COMMAND} -E echo \"\#endif // define ${HEADER_FILE_DEFINE}\" >> ${OUTPUT_HEADER_FILE} + DEPENDS ${spv_file} xxd + COMMENT "Converting to hpp: ${FILE_NAME} ${CMAKE_BINARY_DIR}/bin/xxd" + ) + endif() + endforeach() +endfunction() + +if (EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/kompute/CMakeLists.txt") + message(STATUS "Kompute found") + set(KOMPUTE_OPT_LOG_LEVEL Error CACHE STRING "Kompute log level") + add_subdirectory(kompute) + + # Compile our shaders + compile_shader(SOURCES + kompute-shaders/op_scale.comp + kompute-shaders/op_scale_8.comp + kompute-shaders/op_add.comp + kompute-shaders/op_addrow.comp + kompute-shaders/op_mul.comp + kompute-shaders/op_silu.comp + kompute-shaders/op_relu.comp + kompute-shaders/op_gelu.comp + kompute-shaders/op_softmax.comp + kompute-shaders/op_norm.comp + kompute-shaders/op_rmsnorm.comp + kompute-shaders/op_diagmask.comp + kompute-shaders/op_mul_mat_mat_f32.comp + kompute-shaders/op_mul_mat_f16.comp + kompute-shaders/op_mul_mat_q8_0.comp + kompute-shaders/op_mul_mat_q4_0.comp + kompute-shaders/op_mul_mat_q4_1.comp + kompute-shaders/op_mul_mat_q4_k.comp + kompute-shaders/op_mul_mat_q6_k.comp + kompute-shaders/op_getrows_f32.comp + kompute-shaders/op_getrows_f16.comp + kompute-shaders/op_getrows_q4_0.comp + kompute-shaders/op_getrows_q4_1.comp + kompute-shaders/op_getrows_q6_k.comp + kompute-shaders/op_rope_norm_f16.comp + kompute-shaders/op_rope_norm_f32.comp + kompute-shaders/op_rope_neox_f16.comp + kompute-shaders/op_rope_neox_f32.comp + kompute-shaders/op_cpy_f16_f16.comp + kompute-shaders/op_cpy_f16_f32.comp + kompute-shaders/op_cpy_f32_f16.comp + kompute-shaders/op_cpy_f32_f32.comp + ) + + # Create a custom target for our generated shaders + add_custom_target(generated_shaders DEPENDS + shaderop_scale.h + shaderop_scale_8.h + shaderop_add.h + shaderop_addrow.h + shaderop_mul.h + shaderop_silu.h + shaderop_relu.h + shaderop_gelu.h + shaderop_softmax.h + shaderop_norm.h + shaderop_rmsnorm.h + shaderop_diagmask.h + shaderop_mul_mat_mat_f32.h + shaderop_mul_mat_f16.h + shaderop_mul_mat_q8_0.h + shaderop_mul_mat_q4_0.h + shaderop_mul_mat_q4_1.h + shaderop_mul_mat_q4_k.h + shaderop_mul_mat_q6_k.h + shaderop_getrows_f32.h + shaderop_getrows_f16.h + shaderop_getrows_q4_0.h + shaderop_getrows_q4_1.h + shaderop_getrows_q6_k.h + shaderop_rope_norm_f16.h + shaderop_rope_norm_f32.h + shaderop_rope_neox_f16.h + shaderop_rope_neox_f32.h + shaderop_cpy_f16_f16.h + shaderop_cpy_f16_f32.h + shaderop_cpy_f32_f16.h + shaderop_cpy_f32_f32.h + ) + + # Create a custom command that depends on the generated_shaders + add_custom_command( + OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp + COMMAND ${CMAKE_COMMAND} -E touch ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp + DEPENDS generated_shaders + COMMENT "Ensuring shaders are generated before compiling ggml-kompute.cpp" + ) + + # Add the stamp to the main sources to ensure dependency tracking + target_sources(ggml-kompute PRIVATE ${CMAKE_CURRENT_BINARY_DIR}/ggml-kompute.stamp) +else() + message(WARNING "Kompute not found") +endif() diff --git a/ggml/src/ggml-kompute.cpp b/ggml/src/ggml-kompute/ggml-kompute.cpp similarity index 92% rename from ggml/src/ggml-kompute.cpp rename to ggml/src/ggml-kompute/ggml-kompute.cpp index 2fea9e4cc..505792271 100644 --- a/ggml/src/ggml-kompute.cpp +++ b/ggml/src/ggml-kompute/ggml-kompute.cpp @@ -28,8 +28,10 @@ #include "shaderop_getrows_q4_0.h" #include "shaderop_getrows_q4_1.h" #include "shaderop_getrows_q6_k.h" -#include "shaderop_rope_f16.h" -#include "shaderop_rope_f32.h" +#include "shaderop_rope_norm_f16.h" +#include "shaderop_rope_norm_f32.h" +#include "shaderop_rope_neox_f16.h" +#include "shaderop_rope_neox_f32.h" #include "shaderop_cpy_f16_f16.h" #include "shaderop_cpy_f16_f32.h" #include "shaderop_cpy_f32_f16.h" @@ -345,7 +347,7 @@ void ggml_vk_allocate_descriptor_pool(struct ggml_kompute_context * ctx, size_t std::vector descriptorPoolSizes = { vk::DescriptorPoolSize( vk::DescriptorType::eStorageBuffer, - 3 * size // Descriptor count is number of possible tensors to pass into an algorithm + 4 * size // Descriptor count is number of possible tensors to pass into an algorithm ) }; @@ -788,7 +790,8 @@ static void ggml_vk_soft_max( const std::shared_ptr& out, uint32_t inAOff, uint32_t inBOff, uint32_t outOff, int32_t ne00, int32_t ne01, int32_t ne02, uint32_t ne03, - float scale + float scale, float max_bias, float m0, float m1, + uint32_t n_head_log2 ) { const static auto spirv = getSpirvShader(kp::shader_data::op_softmax_comp_spv, kp::shader_data::op_softmax_comp_spv_len); @@ -796,12 +799,14 @@ static void ggml_vk_soft_max( struct PushConstants { uint32_t inAOff, inBOff, outOff; int32_t ne00, ne01, ne02; - float scale; + float scale, max_bias, m0, m1; + uint32_t n_head_log2; int32_t mask; } pushConsts { safe_divide(inAOff, 4), safe_divide(inBOff, 4), safe_divide(outOff, 4), ne00, ne01, ne02, - scale, + scale, max_bias, m0, m1, + n_head_log2, bool(inB) }; @@ -911,9 +916,9 @@ static void ggml_vk_mul_mat_f16( const std::shared_ptr& out, uint32_t inAOff, uint32_t inBOff, uint32_t outOff, int32_t ne00, int32_t ne01, int32_t ne02, - uint32_t nb00, uint32_t nb01, uint32_t nb02, + uint32_t nb00, uint32_t nb01, uint32_t nb02, uint32_t nb03, int32_t ne10, int32_t ne11, int32_t ne12, int32_t ne13, - uint32_t nb10, uint32_t nb11, uint32_t nb12, + uint32_t nb10, uint32_t nb11, uint32_t nb12, uint32_t nb13, int32_t ne0, int32_t ne1, uint32_t r2, uint32_t r3 ) { @@ -923,17 +928,17 @@ static void ggml_vk_mul_mat_f16( struct PushConstants { uint32_t inAOff, inBOff, outOff; int32_t ne00, ne01, ne02; - uint32_t nb00, nb01, nb02; + uint32_t nb00, nb01, nb02, nb03; int32_t ne10, ne11, ne12; - uint32_t nb10, nb11, nb12; + uint32_t nb10, nb11, nb12, nb13; int32_t ne0, ne1; uint32_t r2, r3; } pushConsts { safe_divide(inAOff, 2), safe_divide(inBOff, 4), safe_divide(outOff, 4), ne00, ne01, ne02, - nb00, nb01, nb02, + nb00, nb01, nb02, nb03, ne10, ne11, ne12, - nb10, nb11, nb12, + nb10, nb11, nb12, nb13, ne0, ne1, r2, r3 }; @@ -1013,6 +1018,8 @@ static void ggml_vk_mul_mat_impl( int32_t ne00, int32_t ne01, int32_t ne02, int32_t ne10, int32_t ne11, int32_t ne12, int32_t ne13, int32_t ne0, int32_t ne1, + uint32_t nb01, uint32_t nb02, uint32_t nb03, + uint32_t nb11, uint32_t nb12, uint32_t nb13, uint32_t r2, uint32_t r3 ) { struct PushConstants { @@ -1020,19 +1027,23 @@ static void ggml_vk_mul_mat_impl( int32_t ne00, ne01, ne02; int32_t ne10, ne12; int32_t ne0, ne1; + uint32_t nb01, nb02, nb03; + uint32_t nb11, nb12, nb13; uint32_t r2, r3; } pushConsts { safe_divide(inAOff, block_size), safe_divide(inBOff, 4), safe_divide(outOff, 4), ne00, ne01, ne02, ne10, ne12, ne0, ne1, + nb01, nb02, nb03, + nb11, nb12, nb13, r2, r3 }; auto name = std::string(__func__) + "_" + suffix; std::shared_ptr s_algo = nullptr; if (!komputeManager()->hasAlgorithm(name)) { - const uint32_t local_x = ggml_vk_current_device().subgroupSize * 2; + const uint32_t local_x = (ggml_vk_current_device().subgroupSize * 2) / 8; s_algo = komputeManager()->algorithm(name, s_kompute_context->pool.get(), {inA, inB, out}, spirv, {unsigned((ne01 + 7)/8), unsigned(ne11), unsigned(ne12*ne13)}, {local_x}, {pushConsts}); } else { s_algo = komputeManager()->getAlgorithm(name); @@ -1074,19 +1085,26 @@ static void ggml_vk_mul_mat_q4_k( const std::shared_ptr& inB, const std::shared_ptr& out, uint32_t inAOff, uint32_t inBOff, uint32_t outOff, - int32_t ne00, int32_t ne01, int32_t ne02, int32_t ne10, - int32_t ne11, int32_t ne12, int32_t ne13, int32_t ne0, - int32_t ne1, int32_t r2, int32_t r3 + int32_t ne00, int32_t ne01, int32_t ne02, + int32_t ne10, int32_t ne11, int32_t ne12, int32_t ne13, + int32_t ne0, int32_t ne1, + uint32_t nb01, uint32_t nb02, uint32_t nb03, + uint32_t nb11, uint32_t nb12, uint32_t nb13, + uint32_t r2, uint32_t r3 ) { const static auto spirv = getSpirvShader(kp::shader_data::op_mul_mat_q4_k_comp_spv, kp::shader_data::op_mul_mat_q4_k_comp_spv_len); struct PushConstants { uint32_t inAOff, inBOff, outOff; - int32_t ne00, ne10, ne0, ne1, ne01, ne02, ne12, r2, r3; + int32_t ne00, ne10, ne0, ne1, ne01, ne02, ne12; + uint32_t nb01, nb02, nb03, nb11, nb12, nb13; + uint32_t r2, r3; } pushConsts { - 0, 0, 0, - ne00, ne10, ne0, ne1, ne01, ne02, ne12, r2, r3 + inAOff, safe_divide(inBOff, 4), safe_divide(outOff, 4), + ne00, ne10, ne0, ne1, ne01, ne02, ne12, + nb01, nb02, nb03, nb11, nb12, nb13, + r2, r3 }; std::shared_ptr s_algo = nullptr; @@ -1108,28 +1126,37 @@ static void ggml_vk_mul_mat_q6_k( const std::shared_ptr& inB, const std::shared_ptr& out, uint32_t inAOff, uint32_t inBOff, uint32_t outOff, - int32_t ne00, int32_t ne10, int32_t ne0, int32_t ne1, - int32_t ne01, int32_t ne11, int32_t ne12, int32_t ne02 + int32_t ne00, int32_t ne01, int32_t ne02, + int32_t ne10, int32_t ne11, int32_t ne12, int32_t ne13, + int32_t ne0, int32_t ne1, + uint32_t nb01, uint32_t nb02, uint32_t nb03, + uint32_t nb11, uint32_t nb12, uint32_t nb13, + uint32_t r2, uint32_t r3 ) { const static auto spirv = getSpirvShader(kp::shader_data::op_mul_mat_q6_k_comp_spv, kp::shader_data::op_mul_mat_q6_k_comp_spv_len); struct PushConstants { uint32_t inAOff, inBOff, outOff; - int32_t ne00, ne10, ne0, ne1, ne01, gqa; + int32_t ne00, ne10, ne0, ne1, ne01, ne02, ne12; + uint32_t nb01, nb02, nb03, nb11, nb12, nb13; + uint32_t r2, r3; } pushConsts { inAOff, safe_divide(inBOff, 4), safe_divide(outOff, 4), - ne00, ne10, ne0, ne1, ne01, ne12/ne02 + ne00, ne10, ne0, ne1, ne01, ne02, ne12, + nb01, nb02, nb03, nb11, nb12, nb13, + r2, r3 }; std::shared_ptr s_algo = nullptr; if (!komputeManager()->hasAlgorithm(__func__)) { - const uint32_t local_x = ggml_vk_current_device().subgroupSize * 2; - s_algo = komputeManager()->algorithm(__func__, s_kompute_context->pool.get(), {inA, inB, out}, spirv, {unsigned((ne01 + 1)/2), unsigned(ne11), unsigned(ne12)}, {local_x}, {pushConsts}); + const uint32_t local_x = 2; + const uint32_t local_y = ggml_vk_current_device().subgroupSize; + s_algo = komputeManager()->algorithm(__func__, s_kompute_context->pool.get(), {inA, inB, out}, spirv, {unsigned((ne01 + 1)/2), unsigned(ne11), unsigned(ne12)*unsigned(ne13)}, {local_x, local_y}, {pushConsts}); } else { s_algo = komputeManager()->getAlgorithm(__func__); s_algo->setTensors({inA, inB, out}); - s_algo->setWorkgroup({unsigned((ne01 + 1)/2), unsigned(ne11), unsigned(ne12)}); + s_algo->setWorkgroup({unsigned((ne01 + 1)/2), unsigned(ne11), unsigned(ne12)*unsigned(ne13)}); s_algo->setPushConstants({pushConsts}); s_algo->updateDescriptors(s_kompute_context->pool.get()); } @@ -1217,10 +1244,11 @@ static void ggml_vk_rope( kp::Sequence& seq, const std::shared_ptr& inA, const std::shared_ptr& inB, + const std::shared_ptr& inC, const std::shared_ptr& out, - uint32_t inAOff, uint32_t inBOff, uint32_t outOff, + uint32_t inAOff, uint32_t inBOff, uint32_t inCOff, uint32_t outOff, ggml_type src0t, int32_t n_dims, int32_t mode, int32_t n_ctx_orig, - float freq_base, float freq_scale, float ext_factor, float attn_factor, float beta_fast, float beta_slow, + float freq_base, float freq_scale, bool has_freq_factors, float ext_factor, float attn_factor, float beta_fast, float beta_slow, int32_t ne01, int32_t ne02, int32_t ne03, uint32_t nb00, uint32_t nb01, uint32_t nb02, uint32_t nb03, int32_t ne0, @@ -1228,11 +1256,17 @@ static void ggml_vk_rope( ) { GGML_ASSERT(src0t == GGML_TYPE_F16 || src0t == GGML_TYPE_F32); - static const auto spirv_f16 = getSpirvShader( - kp::shader_data::op_rope_f16_comp_spv, kp::shader_data::op_rope_f16_comp_spv_len + static const auto spirv_norm_f16 = getSpirvShader( + kp::shader_data::op_rope_norm_f16_comp_spv, kp::shader_data::op_rope_norm_f16_comp_spv_len ); - static const auto spirv_f32 = getSpirvShader( - kp::shader_data::op_rope_f32_comp_spv, kp::shader_data::op_rope_f32_comp_spv_len + static const auto spirv_norm_f32 = getSpirvShader( + kp::shader_data::op_rope_norm_f32_comp_spv, kp::shader_data::op_rope_norm_f32_comp_spv_len + ); + static const auto spirv_neox_f16 = getSpirvShader( + kp::shader_data::op_rope_neox_f16_comp_spv, kp::shader_data::op_rope_neox_f16_comp_spv_len + ); + static const auto spirv_neox_f32 = getSpirvShader( + kp::shader_data::op_rope_neox_f32_comp_spv, kp::shader_data::op_rope_neox_f32_comp_spv_len ); int type_size = src0t == GGML_TYPE_F16 ? 2 : 4; @@ -1247,32 +1281,40 @@ static void ggml_vk_rope( GGML_ASSERT(nb0 % type_size == 0); struct PushConstants { - uint32_t inAOff, inBOff, outOff; + uint32_t inAOff, inBOff, inCOff, outOff; int32_t n_dims, mode, n_ctx_orig; - float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + float freq_base, freq_scale; + bool has_freq_factors; + float ext_factor, attn_factor, beta_fast, beta_slow; uint32_t nb00, nb01, nb02, nb03; int32_t ne0; uint32_t nb0, nb1, nb2, nb3; } pushConsts { - safe_divide(inAOff, type_size), safe_divide(inBOff, 4), safe_divide(outOff, type_size), + safe_divide(inAOff, type_size), safe_divide(inBOff, 4), safe_divide(inCOff, type_size), safe_divide(outOff, type_size), n_dims, mode, n_ctx_orig, - freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow, + freq_base, freq_scale, + has_freq_factors, + ext_factor, attn_factor, beta_fast, beta_slow, nb00, nb01, nb02, nb03, ne0, nb0, nb1, nb2, nb3 }; - auto name = std::string(__func__) + (src0t == GGML_TYPE_F16 ? "_f16" : "_f32"); + auto & inC_ = inC ? inC : inA; + const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; + const bool is_f16 = src0t == GGML_TYPE_F16; + + auto name = std::string(__func__) + (is_neox ? "_neox" : "_norm") + (src0t == GGML_TYPE_F16 ? "_f16" : "_f32"); std::shared_ptr s_algo = nullptr; if (!komputeManager()->hasAlgorithm(name)) { + auto & spirv = is_neox ? is_f16 ? spirv_neox_f16 : spirv_neox_f32 : is_f16 ? spirv_norm_f16 : spirv_norm_f32; s_algo = komputeManager()->algorithm( - name, s_kompute_context->pool.get(), {inA, inB, out}, - src0t == GGML_TYPE_F16 ? spirv_f16 : spirv_f32, + name, s_kompute_context->pool.get(), {inA, inB, inC_, out}, spirv, {unsigned(ne01), unsigned(ne02), unsigned(ne03)}, {}, {pushConsts} ); } else { s_algo = komputeManager()->getAlgorithm(name); - s_algo->setTensors({inA, inB, out}); + s_algo->setTensors({inA, inB, inC_, out}); s_algo->setWorkgroup({unsigned(ne01), unsigned(ne02), unsigned(ne03)}); s_algo->setPushConstants({pushConsts}); s_algo->updateDescriptors(s_kompute_context->pool.get()); @@ -1351,11 +1393,15 @@ static void ggml_vk_cpy_f16_f32(Args&&... args) { } static bool ggml_backend_kompute_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { + int64_t n = ggml_nelements(op); switch (op->op) { case GGML_OP_UNARY: + if (n % 4 != 0) return false; switch (ggml_get_unary_op(op)) { - case GGML_UNARY_OP_RELU: case GGML_UNARY_OP_GELU: + if (n % 8 != 0) return false; + // fall through + case GGML_UNARY_OP_RELU: case GGML_UNARY_OP_SILU: return ggml_is_contiguous(op->src[0]); default: @@ -1373,8 +1419,18 @@ static bool ggml_backend_kompute_device_supports_op(ggml_backend_dev_t dev, cons case GGML_OP_SOFT_MAX: case GGML_OP_RMS_NORM: case GGML_OP_NORM: - case GGML_OP_ROPE: return true; + case GGML_OP_ROPE: + { + const int mode = ((const int32_t *) op->op_params)[2]; + if (mode & GGML_ROPE_TYPE_MROPE) { + return false; + } + if (mode & GGML_ROPE_TYPE_VISION) { + return false; + } + return true; + } case GGML_OP_DUP: case GGML_OP_CPY: case GGML_OP_CONT: @@ -1413,8 +1469,8 @@ static bool ggml_backend_kompute_device_supports_op(ggml_backend_dev_t dev, cons switch (op->src[0]->type) { case GGML_TYPE_F32: - case GGML_TYPE_Q6_K: return op->ne[3] == 1; + case GGML_TYPE_Q6_K: case GGML_TYPE_F16: case GGML_TYPE_Q8_0: case GGML_TYPE_Q4_0: @@ -1515,9 +1571,11 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml const static std::shared_ptr nullTensor = nullptr; uint32_t off_src0 = 0; uint32_t off_src1 = 0; + uint32_t off_src2 = 0; uint32_t off_dst = 0; const std::shared_ptr& id_src0 = src0 ? ggml_vk_get_tensor(src0, &off_src0) : nullTensor; const std::shared_ptr& id_src1 = src1 ? ggml_vk_get_tensor(src1, &off_src1) : nullTensor; + const std::shared_ptr& id_src2 = src2 ? ggml_vk_get_tensor(src2, &off_src2) : nullTensor; const std::shared_ptr& id_dst = dst ? ggml_vk_get_tensor(dst, &off_dst) : nullTensor; switch (dst->op) { @@ -1593,11 +1651,16 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml #pragma message("ref: https://github.com/ggerganov/llama.cpp/pull/5021") GGML_ASSERT(!src1 || src1t == GGML_TYPE_F32); -#pragma message("TODO: add ALiBi support") -#pragma message("ref: https://github.com/ggerganov/llama.cpp/pull/7192") - GGML_ASSERT(max_bias == 0.0f); + const int64_t nrows_x = ggml_nrows(src0); + const int64_t nrows_y = src0->ne[1]; - ggml_vk_soft_max(seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, ne00, ne01, ne02, ne03, scale); + const uint32_t n_head = nrows_x/nrows_y; + const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + ggml_vk_soft_max(seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, ne00, ne01, ne02, ne03, scale, max_bias, m0, m1, n_head_log2); } break; case GGML_OP_DIAG_MASK_INF: { @@ -1649,38 +1712,44 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml case GGML_TYPE_F16: ggml_vk_mul_mat_f16( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne01, ne02, nb00, nb01, nb02, ne10, ne11, ne12, ne13, nb10, nb11, nb12, + ne00, ne01, ne02, nb00, nb01, nb02, nb03, + ne10, ne11, ne12, ne13, nb10, nb11, nb12, nb13, ne0, ne1, r2, r3 ); break; case GGML_TYPE_Q8_0: ggml_vk_mul_mat_q8_0( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, r2, r3 + ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, + nb01, nb02, nb03, nb11, nb12, nb13, r2, r3 ); break; case GGML_TYPE_Q4_0: ggml_vk_mul_mat_q4_0( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, r2, r3 + ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, + nb01, nb02, nb03, nb11, nb12, nb13, r2, r3 ); break; case GGML_TYPE_Q4_1: ggml_vk_mul_mat_q4_1( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, r2, r3 + ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, + nb01, nb02, nb03, nb11, nb12, nb13, r2, r3 ); break; case GGML_TYPE_Q4_K: ggml_vk_mul_mat_q4_k( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, ne12/ne02, ne13/ne03 + ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, + nb01, nb02, nb03, nb11, nb12, nb13, r2, r3 ); break; case GGML_TYPE_Q6_K: ggml_vk_mul_mat_q6_k( seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, - ne00, ne10, ne0, ne1, ne01, ne11, ne12, ne02 + ne00, ne01, ne02, ne10, ne11, ne12, ne13, ne0, ne1, + nb01, nb02, nb03, nb11, nb12, nb13, r2, r3 ); break; default: { @@ -1709,13 +1778,6 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml } break; case GGML_OP_ROPE: { -#pragma message("TODO: implement phi3 frequency factors support") -#pragma message(" https://github.com/ggerganov/llama.cpp/pull/7225") - GGML_ASSERT(dst->src[2] == nullptr && "phi3 frequency factors not implemented yet"); - -#pragma message("TODO: update rope NORM mode to match NEOX mode") -#pragma message(" https://github.com/ggerganov/llama.cpp/pull/7634") - GGML_ASSERT(ne10 == ne02); GGML_ASSERT(src0t == dstt); // const int n_past = ((int32_t *) dst->op_params)[0]; @@ -1724,6 +1786,8 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml // skip 3, n_ctx used in GLM RoPE, unimplemented in Vulkan const int n_ctx_orig = ((int32_t *) dst->op_params)[4]; + const bool has_freq_factors = dst->src[2] != nullptr; + float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; memcpy(&freq_base, (int32_t *) dst->op_params + 5, sizeof(float)); memcpy(&freq_scale, (int32_t *) dst->op_params + 6, sizeof(float)); @@ -1732,8 +1796,8 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml memcpy(&beta_fast, (int32_t *) dst->op_params + 9, sizeof(float)); memcpy(&beta_slow, (int32_t *) dst->op_params + 10, sizeof(float)); ggml_vk_rope( - seq, id_src0, id_src1, id_dst, off_src0, off_src1, off_dst, src0t, n_dims, mode, n_ctx_orig, - freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow, + seq, id_src0, id_src1, id_src2, id_dst, off_src0, off_src1, off_src2, off_dst, src0t, n_dims, mode, n_ctx_orig, + freq_base, freq_scale, has_freq_factors, ext_factor, attn_factor, beta_fast, beta_slow, ne01, ne02, ne03, nb00, nb01, nb02, nb03, ne0, nb0, nb1, nb2, nb3 ); } break; @@ -2176,9 +2240,12 @@ static const struct ggml_backend_reg_i ggml_backend_kompute_reg_i = { ggml_backend_reg_t ggml_backend_kompute_reg() { static ggml_backend_reg reg = { - /* .iface = */ ggml_backend_kompute_reg_i, - /* .context = */ nullptr, + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_kompute_reg_i, + /* .context = */ nullptr, }; return ® } + +GGML_BACKEND_DL_IMPL(ggml_backend_kompute_reg) diff --git a/ggml/src/kompute b/ggml/src/ggml-kompute/kompute similarity index 100% rename from ggml/src/kompute rename to ggml/src/ggml-kompute/kompute diff --git a/ggml/src/kompute-shaders/common.comp b/ggml/src/ggml-kompute/kompute-shaders/common.comp similarity index 98% rename from ggml/src/kompute-shaders/common.comp rename to ggml/src/ggml-kompute/kompute-shaders/common.comp index 2aaddf704..dbe4cf804 100644 --- a/ggml/src/kompute-shaders/common.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/common.comp @@ -3,6 +3,7 @@ #extension GL_EXT_shader_explicit_arithmetic_types_float16: require #extension GL_EXT_shader_explicit_arithmetic_types_int8: require #extension GL_EXT_shader_explicit_arithmetic_types_int16: require +#extension GL_EXT_shader_explicit_arithmetic_types_int64: require #extension GL_EXT_control_flow_attributes: enable #extension GL_KHR_shader_subgroup_arithmetic : require #extension GL_EXT_debug_printf : enable diff --git a/ggml/src/kompute-shaders/op_add.comp b/ggml/src/ggml-kompute/kompute-shaders/op_add.comp similarity index 100% rename from ggml/src/kompute-shaders/op_add.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_add.comp diff --git a/ggml/src/kompute-shaders/op_addrow.comp b/ggml/src/ggml-kompute/kompute-shaders/op_addrow.comp similarity index 100% rename from ggml/src/kompute-shaders/op_addrow.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_addrow.comp diff --git a/ggml/src/kompute-shaders/op_cpy_f16_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f16.comp similarity index 100% rename from ggml/src/kompute-shaders/op_cpy_f16_f16.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f16.comp diff --git a/ggml/src/kompute-shaders/op_cpy_f16_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f32.comp similarity index 100% rename from ggml/src/kompute-shaders/op_cpy_f16_f32.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f32.comp diff --git a/ggml/src/kompute-shaders/op_cpy_f32_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f16.comp similarity index 100% rename from ggml/src/kompute-shaders/op_cpy_f32_f16.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f16.comp diff --git a/ggml/src/kompute-shaders/op_cpy_f32_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f32.comp similarity index 100% rename from ggml/src/kompute-shaders/op_cpy_f32_f32.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f32.comp diff --git a/ggml/src/kompute-shaders/op_diagmask.comp b/ggml/src/ggml-kompute/kompute-shaders/op_diagmask.comp similarity index 100% rename from ggml/src/kompute-shaders/op_diagmask.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_diagmask.comp diff --git a/ggml/src/kompute-shaders/op_gelu.comp b/ggml/src/ggml-kompute/kompute-shaders/op_gelu.comp similarity index 100% rename from ggml/src/kompute-shaders/op_gelu.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_gelu.comp diff --git a/ggml/src/kompute-shaders/op_getrows.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows.comp diff --git a/ggml/src/kompute-shaders/op_getrows_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f16.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows_f16.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows_f16.comp diff --git a/ggml/src/kompute-shaders/op_getrows_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f32.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows_f32.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows_f32.comp diff --git a/ggml/src/kompute-shaders/op_getrows_q4_0.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_0.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows_q4_0.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_0.comp diff --git a/ggml/src/kompute-shaders/op_getrows_q4_1.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_1.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows_q4_1.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_1.comp diff --git a/ggml/src/kompute-shaders/op_getrows_q6_k.comp b/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q6_k.comp similarity index 100% rename from ggml/src/kompute-shaders/op_getrows_q6_k.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_getrows_q6_k.comp diff --git a/ggml/src/kompute-shaders/op_mul.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul.comp similarity index 100% rename from ggml/src/kompute-shaders/op_mul.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul.comp diff --git a/ggml/src/kompute-shaders/op_mul_mat_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp similarity index 91% rename from ggml/src/kompute-shaders/op_mul_mat_f16.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp index 8f0a9031f..0ab1b2fc2 100644 --- a/ggml/src/kompute-shaders/op_mul_mat_f16.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp @@ -20,12 +20,14 @@ layout (push_constant) uniform parameter { uint nb00; uint nb01; uint nb02; + uint nb03; int ne10; int ne11; int ne12; uint nb10; uint nb11; uint nb12; + uint nb13; int ne0; int ne1; uint r2; @@ -42,7 +44,7 @@ void main() { const uint i12 = im%pcs.ne12; const uint i13 = im/pcs.ne12; - const uint offset0 = r0*pcs.nb01 + (i12/pcs.r2)*pcs.nb02 + (i13/pcs.r3)*pcs.nb02*pcs.ne02; + const uint offset0 = r0*pcs.nb01 + (i12/pcs.r2)*pcs.nb02 + (i13/pcs.r3)*pcs.nb03; const uint x = offset0 / 2 + pcs.inAOff; // Based from inA @@ -52,7 +54,7 @@ void main() { break; } - const uint y = (r1*pcs.nb11 + im*pcs.nb12) / 4 + pcs.inBOff; // Based from inB + const uint y = (r1*pcs.nb11 + i12*pcs.nb12 + i13*pcs.nb13) / 4 + pcs.inBOff; float sumf = 0; for (uint i = gl_SubgroupInvocationID.x; i < pcs.ne00; i += gl_SubgroupSize) { diff --git a/ggml/src/kompute-shaders/op_mul_mat_mat_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_mat_f32.comp similarity index 100% rename from ggml/src/kompute-shaders/op_mul_mat_mat_f32.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_mat_f32.comp diff --git a/ggml/src/kompute-shaders/op_mul_mat_q4_0.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_0.comp similarity index 100% rename from ggml/src/kompute-shaders/op_mul_mat_q4_0.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_0.comp diff --git a/ggml/src/kompute-shaders/op_mul_mat_q4_1.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_1.comp similarity index 100% rename from ggml/src/kompute-shaders/op_mul_mat_q4_1.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_1.comp diff --git a/ggml/src/kompute-shaders/op_mul_mat_q4_k.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp similarity index 89% rename from ggml/src/kompute-shaders/op_mul_mat_q4_k.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp index fc8e45aa9..a5752a3a0 100644 --- a/ggml/src/kompute-shaders/op_mul_mat_q4_k.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp @@ -24,8 +24,14 @@ layout (push_constant) uniform parameter { int ne01; int ne02; int ne12; - int r2; - int r3; + uint nb01; + uint nb02; + uint nb03; + uint nb11; + uint nb12; + uint nb13; + uint r2; + uint r3; } pcs; void main() { @@ -50,10 +56,11 @@ void main() { const uint i12 = im%pcs.ne12; const uint i13 = im/pcs.ne12; - const uint offset0 = (i12/pcs.r2)*(nb*pcs.ne01) + (i13/pcs.r3)*(nb*pcs.ne01*pcs.ne02); + const uint offset0 = first_row*(pcs.nb01/SIZE_OF_BLOCK) + (i12/pcs.r2)*(pcs.nb02/SIZE_OF_BLOCK) + (i13/pcs.r3)*(pcs.nb03/SIZE_OF_BLOCK); + const uint offset1 = r1*pcs.nb11 + (i12 )*pcs.nb12 + (i13 )*pcs.nb13; - const uint xblk = ib_row + offset0 + pcs.inAOff; - const uint y = r1*pcs.ne10 + im*pcs.ne00*pcs.ne1 + pcs.inBOff; + const uint xblk = offset0 + pcs.inAOff; + const uint y = (offset1 / 4) + pcs.inBOff; float yl[16]; float yh[16]; @@ -74,7 +81,7 @@ void main() { } for (int row = 0; row < N_DST; row++) { - uint row_idx = row * nb; + uint row_idx = row * (pcs.nb01 / SIZE_OF_BLOCK); uint16_t sc_0 = u8BufToU16(inA[blk_idx + row_idx].scales, iq * 2 + 0); uint16_t sc_1 = u8BufToU16(inA[blk_idx + row_idx].scales, iq * 2 + 2); diff --git a/ggml/src/kompute-shaders/op_mul_mat_q6_k.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp similarity index 86% rename from ggml/src/kompute-shaders/op_mul_mat_q6_k.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp index c9baebdf4..d331d1a70 100644 --- a/ggml/src/kompute-shaders/op_mul_mat_q6_k.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp @@ -21,7 +21,16 @@ layout (push_constant) uniform parameter { int ne0; int ne1; int ne01; - int gqa; + int ne02; + int ne12; + uint nb01; + uint nb02; + uint nb03; + uint nb11; + uint nb12; + uint nb13; + uint r2; + uint r3; } pcs; void main() { @@ -34,12 +43,15 @@ void main() { const uint r0 = gl_WorkGroupID.x; const uint r1 = gl_WorkGroupID.y; - const uint r2 = gl_WorkGroupID.z; + const uint im = gl_WorkGroupID.z; const uint row = (r0 * gl_NumSubgroups + gl_SubgroupID); - const uint offset0 = r2/pcs.gqa*(nb*pcs.ne0); - const uint x = row * nb + offset0; // Based from inA without base offset - const uint yy = r1*pcs.ne10 + r2*pcs.ne00*pcs.ne1+pcs.inBOff; // Based from inB + + const uint i12 = im%pcs.ne12; + const uint i13 = im/pcs.ne12; + + const uint x = row*(pcs.nb01/SIZE_OF_BLOCK) + (i12/pcs.r2)*(pcs.nb02/SIZE_OF_BLOCK) + (i13/pcs.r3)*(pcs.nb03/SIZE_OF_BLOCK); + const uint yy = (r1*pcs.nb11 + i12*pcs.nb12 + i13*pcs.nb13) / 4 + pcs.inBOff; float sumf = 0; @@ -89,6 +101,6 @@ void main() { const float tot = subgroupAdd(sumf); if (subgroupElect()) { - out_[r1*pcs.ne0 + r2*pcs.ne0*pcs.ne1 + row + pcs.outOff] = tot; + out_[r1*pcs.ne0 + im*pcs.ne0*pcs.ne1 + row + pcs.outOff] = tot; } } diff --git a/ggml/src/kompute-shaders/op_mul_mat_q8_0.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q8_0.comp similarity index 100% rename from ggml/src/kompute-shaders/op_mul_mat_q8_0.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q8_0.comp diff --git a/ggml/src/kompute-shaders/op_mul_mv_q_n.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp similarity index 76% rename from ggml/src/kompute-shaders/op_mul_mv_q_n.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp index 440b5ab2c..a6517cc1f 100644 --- a/ggml/src/kompute-shaders/op_mul_mv_q_n.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp @@ -14,10 +14,15 @@ void main() { const uint i12 = im%pcs.ne12; const uint i13 = im/pcs.ne12; - const uint offset0 = first_row * nb + (i12/pcs.r2)*(nb*pcs.ne01) + (i13/pcs.r3)*(nb*pcs.ne01*pcs.ne02); + // pointers to src0 rows + uint ax[N_ROWS]; + for (int row = 0; row < N_ROWS; ++row) { + const uint offset0 = (first_row + row)*(pcs.nb01/SIZE_OF_BLOCK) + (i12/pcs.r2)*(pcs.nb02/SIZE_OF_BLOCK) + (i13/pcs.r3)*(pcs.nb03/SIZE_OF_BLOCK); - const uint x = offset0; // Based from inA without base offset - const uint y = r1*uint(pcs.ne10)+im*pcs.ne00*pcs.ne1+pcs.inBOff; // Based from inB + ax[row] = offset0 + pcs.inAOff; + } + + const uint y = (r1*pcs.nb11 + i12*pcs.nb12 + i13*pcs.nb13) / 4 + pcs.inBOff; float sumf[N_ROWS] = {0.0f, 0.0f, 0.0f, 0.0f}; @@ -32,8 +37,7 @@ void main() { for (uint ib = ix; ib < nb; ib += 16) { for (int row = 0; row < N_ROWS; row++) { - const uint block_index = x + ib + row * nb; - sumf[row] += block_q_n_dot_y(block_index, yb, il); + sumf[row] += block_q_n_dot_y(ax[row] + ib, yb, il); } yb += BLOCKS_IN_QUANT * 16; diff --git a/ggml/src/kompute-shaders/op_mul_mv_q_n_pre.comp b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp similarity index 80% rename from ggml/src/kompute-shaders/op_mul_mv_q_n_pre.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp index 7912b09ac..a9a2f2218 100644 --- a/ggml/src/kompute-shaders/op_mul_mv_q_n_pre.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp @@ -1,5 +1,5 @@ layout(local_size_x_id = 0) in; -layout(local_size_y = 1) in; +layout(local_size_y = 8) in; layout(local_size_z = 1) in; layout (binding = 0) readonly buffer tensorInA { uint8_t inA[]; }; @@ -17,6 +17,12 @@ layout (push_constant) uniform parameter { int ne12; int ne0; int ne1; + uint nb01; + uint nb02; + uint nb03; + uint nb11; + uint nb12; + uint nb13; uint r2; uint r3; } pcs; diff --git a/ggml/src/kompute-shaders/op_norm.comp b/ggml/src/ggml-kompute/kompute-shaders/op_norm.comp similarity index 100% rename from ggml/src/kompute-shaders/op_norm.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_norm.comp diff --git a/ggml/src/kompute-shaders/op_relu.comp b/ggml/src/ggml-kompute/kompute-shaders/op_relu.comp similarity index 100% rename from ggml/src/kompute-shaders/op_relu.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_relu.comp diff --git a/ggml/src/kompute-shaders/op_rmsnorm.comp b/ggml/src/ggml-kompute/kompute-shaders/op_rmsnorm.comp similarity index 100% rename from ggml/src/kompute-shaders/op_rmsnorm.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_rmsnorm.comp diff --git a/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp new file mode 100644 index 000000000..63659cbfe --- /dev/null +++ b/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp @@ -0,0 +1,52 @@ +#version 450 + +#include "rope_common.comp" + +layout(binding = 0) buffer restrict readonly tensorInA { float16_t inA[]; }; +layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; +layout(binding = 2) buffer restrict readonly tensorInC { float inC[]; }; +layout(binding = 3) buffer restrict writeonly tensorOut { float16_t out_[]; }; + +void main() { + const uint i3 = gl_WorkGroupID.z; + const uint i2 = gl_WorkGroupID.y; + const uint i1 = gl_WorkGroupID.x; + + float corr_dims[2]; + rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); + + const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); + + float theta_base = float(inB[pcs.inBOff + i2]); + float inv_ndims = -1.f/pcs.n_dims; + + float cos_theta; + float sin_theta; + + for (uint i0 = 2*gl_LocalInvocationIndex; i0 < pcs.ne0; i0 += 2*gl_WorkGroupSize.x) { + if (i0 < pcs.n_dims) { + uint ic = i0/2; + + float theta = theta_base * pow(pcs.freq_base, inv_ndims*i0); + + const float freq_factor = pcs.has_freq_factors ? inC[pcs.inCOff + ic] : 1.0f; + + rope_yarn(theta/freq_factor, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); + + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + ic*pcs.nb00) / 2) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + ic*pcs.nb0) / 2) + pcs.outOff; // Based from out_ + + const float x0 = float(inA[src]); + const float x1 = float(inA[src+pcs.n_dims/2]); + + out_[dst_data] = float16_t(x0*cos_theta - x1*sin_theta); + out_[dst_data+pcs.n_dims/2] = float16_t(x0*sin_theta + x1*cos_theta); + } else { + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ + + out_[dst_data] = inA[src]; + out_[dst_data+1] = inA[src+1]; + } + } +} diff --git a/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp new file mode 100644 index 000000000..4df56204d --- /dev/null +++ b/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp @@ -0,0 +1,52 @@ +#version 450 + +#include "rope_common.comp" + +layout(binding = 0) buffer restrict readonly tensorInA { float inA[]; }; +layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; +layout(binding = 2) buffer restrict readonly tensorInC { float inC[]; }; +layout(binding = 3) buffer restrict writeonly tensorOut { float out_[]; }; + +void main() { + const uint i3 = gl_WorkGroupID.z; + const uint i2 = gl_WorkGroupID.y; + const uint i1 = gl_WorkGroupID.x; + + float corr_dims[2]; + rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); + + const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); + + float theta_base = float(inB[pcs.inBOff + i2]); + float inv_ndims = -1.f/pcs.n_dims; + + float cos_theta; + float sin_theta; + + for (uint i0 = 2*gl_LocalInvocationIndex; i0 < pcs.ne0; i0 += 2*gl_WorkGroupSize.x) { + if (i0 < pcs.n_dims) { + uint ic = i0/2; + + float theta = theta_base * pow(pcs.freq_base, inv_ndims*i0); + + const float freq_factor = pcs.has_freq_factors ? inC[pcs.inCOff + ic] : 1.0f; + + rope_yarn(theta/freq_factor, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); + + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + ic*pcs.nb00) / 4) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + ic*pcs.nb0) / 4) + pcs.outOff; // Based from out_ + + const float x0 = inA[src]; + const float x1 = inA[src+pcs.n_dims/2]; + + out_[dst_data] = x0*cos_theta - x1*sin_theta; + out_[dst_data+pcs.n_dims/2] = x0*sin_theta + x1*cos_theta; + } else { + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ + + out_[dst_data] = inA[src]; + out_[dst_data+1] = inA[src+1]; + } + } +} diff --git a/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp b/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp new file mode 100644 index 000000000..a3c0eda8b --- /dev/null +++ b/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp @@ -0,0 +1,52 @@ +#version 450 + +#include "rope_common.comp" + +layout(binding = 0) buffer restrict readonly tensorInA { float16_t inA[]; }; +layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; +layout(binding = 2) buffer restrict readonly tensorInC { float inC[]; }; +layout(binding = 3) buffer restrict writeonly tensorOut { float16_t out_[]; }; + +void main() { + const uint i3 = gl_WorkGroupID.z; + const uint i2 = gl_WorkGroupID.y; + const uint i1 = gl_WorkGroupID.x; + + float corr_dims[2]; + rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); + + const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); + + float theta_base = float(inB[pcs.inBOff + i2]); + float inv_ndims = -1.f/pcs.n_dims; + + float cos_theta; + float sin_theta; + + for (uint i0 = 2*gl_LocalInvocationIndex; i0 < pcs.ne0; i0 += 2*gl_WorkGroupSize.x) { + if (i0 < pcs.n_dims) { + uint ic = i0/2; + + float theta = theta_base * pow(pcs.freq_base, inv_ndims*i0); + + const float freq_factor = pcs.has_freq_factors ? inC[pcs.inCOff + ic] : 1.0f; + + rope_yarn(theta/freq_factor, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); + + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ + + const float x0 = float(inA[src]); + const float x1 = float(inA[src+1]); + + out_[dst_data] = float16_t(x0*cos_theta - x1*sin_theta); + out_[dst_data+1] = float16_t(x0*sin_theta + x1*cos_theta); + } else { + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ + + out_[dst_data] = inA[src]; + out_[dst_data+1] = inA[src+1]; + } + } +} diff --git a/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp b/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp new file mode 100644 index 000000000..b7963ae72 --- /dev/null +++ b/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp @@ -0,0 +1,52 @@ +#version 450 + +#include "rope_common.comp" + +layout(binding = 0) buffer restrict readonly tensorInA { float inA[]; }; +layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; +layout(binding = 2) buffer restrict readonly tensorInC { float inC[]; }; +layout(binding = 3) buffer restrict writeonly tensorOut { float out_[]; }; + +void main() { + const uint i3 = gl_WorkGroupID.z; + const uint i2 = gl_WorkGroupID.y; + const uint i1 = gl_WorkGroupID.x; + + float corr_dims[2]; + rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); + + const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); + + float theta_base = float(inB[pcs.inBOff + i2]); + float inv_ndims = -1.f/pcs.n_dims; + + float cos_theta; + float sin_theta; + + for (uint i0 = 2*gl_LocalInvocationIndex; i0 < pcs.ne0; i0 += 2*gl_WorkGroupSize.x) { + if (i0 < pcs.n_dims) { + uint ic = i0/2; + + float theta = theta_base * pow(pcs.freq_base, inv_ndims*i0); + + const float freq_factor = pcs.has_freq_factors ? inC[pcs.inCOff + ic] : 1.0f; + + rope_yarn(theta/freq_factor, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); + + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ + + const float x0 = inA[src]; + const float x1 = inA[src+1]; + + out_[dst_data] = x0*cos_theta - x1*sin_theta; + out_[dst_data+1] = x0*sin_theta + x1*cos_theta; + } else { + const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in + const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ + + out_[dst_data] = inA[src]; + out_[dst_data+1] = inA[src+1]; + } + } +} diff --git a/ggml/src/kompute-shaders/op_scale.comp b/ggml/src/ggml-kompute/kompute-shaders/op_scale.comp similarity index 100% rename from ggml/src/kompute-shaders/op_scale.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_scale.comp diff --git a/ggml/src/kompute-shaders/op_scale_8.comp b/ggml/src/ggml-kompute/kompute-shaders/op_scale_8.comp similarity index 100% rename from ggml/src/kompute-shaders/op_scale_8.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_scale_8.comp diff --git a/ggml/src/kompute-shaders/op_silu.comp b/ggml/src/ggml-kompute/kompute-shaders/op_silu.comp similarity index 100% rename from ggml/src/kompute-shaders/op_silu.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_silu.comp diff --git a/ggml/src/kompute-shaders/op_softmax.comp b/ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp similarity index 78% rename from ggml/src/kompute-shaders/op_softmax.comp rename to ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp index 7bc9176ca..4165295bf 100644 --- a/ggml/src/kompute-shaders/op_softmax.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp @@ -18,6 +18,10 @@ layout(push_constant) uniform PushConstants { int ne01; int ne02; float scale; + float max_bias; + float m0; + float m1; + uint n_head_log2; int mask; } pcs; @@ -34,17 +38,29 @@ void main() { const uint pmask = i01*pcs.ne00 + pcs.inBOff; // Based from inB const uint pdst = extra_off + pcs.outOff; // Based from out_ + float slope = 1.0f; + + // ALiBi + if (pcs.max_bias > 0.0f) { + int64_t h = i02; + + float base = h < pcs.n_head_log2 ? pcs.m0 : pcs.m1; + int64_t exp = h < pcs.n_head_log2 ? h + 1 : 2*(h - pcs.n_head_log2) + 1; + + slope = pow(base, float(exp)); + } + // parallel max float localMax = uintBitsToFloat(0xFF800000); for (uint i00 = gl_SubgroupInvocationID.x; i00 < pcs.ne00; i00 += 32) { - localMax = max(localMax, inA[psrc0 + i00]*pcs.scale + (pcs.mask!=0 ? inB[pmask + i00] : 0.0f)); + localMax = max(localMax, inA[psrc0 + i00]*pcs.scale + (pcs.mask!=0 ? slope*inB[pmask + i00] : 0.0f)); } float max_ = subgroupMax(localMax); // parallel sum float localSum = 0.0f; for (uint i00 = gl_SubgroupInvocationID.x; i00 < pcs.ne00; i00 += 32) { - const float exp_psrc0 = exp(inA[psrc0 + i00]*pcs.scale + (pcs.mask!=0 ? inB[pmask + i00] : 0.0f) - max_); + const float exp_psrc0 = exp(inA[psrc0 + i00]*pcs.scale + (pcs.mask!=0 ? slope*inB[pmask + i00] : 0.0f) - max_); localSum += exp_psrc0; out_[pdst + i00] = exp_psrc0; } diff --git a/ggml/src/kompute-shaders/rope_common.comp b/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp similarity index 98% rename from ggml/src/kompute-shaders/rope_common.comp rename to ggml/src/ggml-kompute/kompute-shaders/rope_common.comp index df4702896..0fca640dc 100644 --- a/ggml/src/kompute-shaders/rope_common.comp +++ b/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp @@ -8,12 +8,14 @@ layout(local_size_x = 1) in; layout (push_constant) uniform parameter { uint inAOff; uint inBOff; + uint inCOff; uint outOff; int n_dims; int mode; int n_ctx_orig; float freq_base; float freq_scale; + bool has_freq_factors; float ext_factor; float attn_factor; float beta_fast; diff --git a/ggml/src/ggml-metal/CMakeLists.txt b/ggml/src/ggml-metal/CMakeLists.txt new file mode 100644 index 000000000..89fcde2fa --- /dev/null +++ b/ggml/src/ggml-metal/CMakeLists.txt @@ -0,0 +1,121 @@ +find_library(FOUNDATION_LIBRARY Foundation REQUIRED) +find_library(METAL_FRAMEWORK Metal REQUIRED) +find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) + +message(STATUS "Metal framework found") + +ggml_add_backend_library(ggml-metal + ggml-metal.m + ) + +target_link_libraries(ggml-metal PRIVATE + ${FOUNDATION_LIBRARY} + ${METAL_FRAMEWORK} + ${METALKIT_FRAMEWORK} + ) + +if (GGML_METAL_NDEBUG) + add_compile_definitions(GGML_METAL_NDEBUG) +endif() + +if (GGML_METAL_USE_BF16) + add_compile_definitions(GGML_METAL_USE_BF16) +endif() + +# copy metal files to bin directory +configure_file(../ggml-common.h ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h COPYONLY) +configure_file(ggml-metal.metal ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal COPYONLY) +configure_file(ggml-metal-impl.h ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal-impl.h COPYONLY) + +if (GGML_METAL_EMBED_LIBRARY) + enable_language(ASM) + + add_compile_definitions(GGML_METAL_EMBED_LIBRARY) + + set(METALLIB_COMMON "${CMAKE_CURRENT_SOURCE_DIR}/../ggml-common.h") + set(METALLIB_SOURCE "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal.metal") + set(METALLIB_IMPL "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal-impl.h") + + file(MAKE_DIRECTORY "${CMAKE_BINARY_DIR}/autogenerated") + + # merge ggml-common.h and ggml-metal.metal into a single file + set(METALLIB_EMBED_ASM "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.s") + set(METALLIB_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.metal") + set(METALLIB_SOURCE_EMBED_TMP "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.metal.tmp") + + add_custom_command( + OUTPUT ${METALLIB_EMBED_ASM} + COMMAND echo "Embedding Metal library" + COMMAND sed -e '/__embed_ggml-common.h__/r ${METALLIB_COMMON}' -e '/__embed_ggml-common.h__/d' < ${METALLIB_SOURCE} > ${METALLIB_SOURCE_EMBED_TMP} + COMMAND sed -e '/\#include \"ggml-metal-impl.h\"/r ${METALLIB_IMPL}' -e '/\#include \"ggml-metal-impl.h\"/d' < ${METALLIB_SOURCE_EMBED_TMP} > ${METALLIB_SOURCE_EMBED} + COMMAND echo ".section __DATA,__ggml_metallib" > ${METALLIB_EMBED_ASM} + COMMAND echo ".globl _ggml_metallib_start" >> ${METALLIB_EMBED_ASM} + COMMAND echo "_ggml_metallib_start:" >> ${METALLIB_EMBED_ASM} + COMMAND echo ".incbin \\\"${METALLIB_SOURCE_EMBED}\\\"" >> ${METALLIB_EMBED_ASM} + COMMAND echo ".globl _ggml_metallib_end" >> ${METALLIB_EMBED_ASM} + COMMAND echo "_ggml_metallib_end:" >> ${METALLIB_EMBED_ASM} + DEPENDS ../ggml-common.h ggml-metal.metal ggml-metal-impl.h + COMMENT "Generate assembly for embedded Metal library" + ) + + target_sources(ggml-metal PRIVATE ${METALLIB_EMBED_ASM}) +else() + if (GGML_METAL_SHADER_DEBUG) + # custom command to do the following: + # xcrun -sdk macosx metal -fno-fast-math -c ggml-metal.metal -o ggml-metal.air + # xcrun -sdk macosx metallib ggml-metal.air -o default.metallib + # + # note: this is the only way I found to disable fast-math in Metal. it's ugly, but at least it works + # disabling fast math is needed in order to pass tests/test-backend-ops + # note: adding -fno-inline fixes the tests when using MTL_SHADER_VALIDATION=1 + # note: unfortunately, we have to call it default.metallib instead of ggml.metallib + # ref: https://github.com/ggerganov/whisper.cpp/issues/1720 + set(XC_FLAGS -fno-fast-math -fno-inline -g) + else() + set(XC_FLAGS -O3) + endif() + + # Append macOS metal versioning flags + if (GGML_METAL_MACOSX_VERSION_MIN) + message(STATUS "Adding -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN} flag to metal compilation") + list (APPEND XC_FLAGS -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN}) + endif() + + if (GGML_METAL_STD) + message(STATUS "Adding -std=${GGML_METAL_STD} flag to metal compilation") + list (APPEND XC_FLAGS -std=${GGML_METAL_STD}) + endif() + + add_custom_command( + OUTPUT ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib + COMMAND xcrun -sdk macosx metal ${XC_FLAGS} -c ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal -o ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air + COMMAND xcrun -sdk macosx metallib ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air -o ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib + COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.air + COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h + COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal + DEPENDS ggml-metal.metal ggml-common.h + COMMENT "Compiling Metal kernels" + ) + + # FIXME: only add to the ggml-metal target? + add_custom_target( + ggml-metal-lib ALL + DEPENDS ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib + ) +endif() # GGML_METAL_EMBED_LIBRARY + +if (NOT GGML_METAL_EMBED_LIBRARY) + install( + FILES src/ggml-metal/ggml-metal.metal + PERMISSIONS + OWNER_READ + OWNER_WRITE + GROUP_READ + WORLD_READ + DESTINATION ${CMAKE_INSTALL_BINDIR}) + + install( + FILES ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib + DESTINATION ${CMAKE_INSTALL_BINDIR} + ) +endif() diff --git a/ggml/src/ggml-metal/ggml-metal-impl.h b/ggml/src/ggml-metal/ggml-metal-impl.h new file mode 100644 index 000000000..e3dc25f16 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-impl.h @@ -0,0 +1,288 @@ +#ifndef GGML_METAL_IMPL +#define GGML_METAL_IMPL + +// kernel argument structs +// +// - element counters (e.g. ne00) typically use int32_t to reduce register usage +// however, be careful from int overflows when using those in the kernel implementation +// +// - strides (e.g. nb00) use uint64_t + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + int32_t ne03; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne10; + int32_t ne11; + int32_t ne12; + int32_t ne13; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + int32_t ne0; + int32_t ne1; + int32_t ne2; + int32_t ne3; + uint64_t nb0; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; + int32_t dim; +} ggml_metal_kargs_concat; + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + int32_t ne03; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne10; + int32_t ne11; + int32_t ne12; + int32_t ne13; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + int32_t ne0; + int32_t ne1; + int32_t ne2; + int32_t ne3; + uint64_t nb0; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; + uint64_t offs; +} ggml_metal_kargs_bin; + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + int32_t ne03; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne0; + int32_t ne1; + int32_t ne2; + int32_t ne3; + uint64_t nb0; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; +} ggml_metal_kargs_repeat; + +typedef struct { + int64_t ne00; + int64_t ne01; + int64_t ne02; + int64_t ne03; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int64_t ne0; + int64_t ne1; + int64_t ne2; + int64_t ne3; + uint64_t nb0; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; +} ggml_metal_kargs_cpy; + +typedef struct { + int64_t ne10; + int64_t ne11; + int64_t ne12; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; + uint64_t offs; + bool inplace; +} ggml_metal_kargs_set; + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + int32_t ne03; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne0; + int32_t ne1; + int32_t ne2; + int32_t ne3; + uint64_t nb0; + uint64_t nb1; + uint64_t nb2; + uint64_t nb3; + int32_t n_past; + int32_t n_dims; + int32_t n_ctx_orig; + float freq_base; + float freq_scale; + float ext_factor; + float attn_factor; + float beta_fast; + float beta_slow; +} ggml_metal_kargs_rope; + +typedef struct { + int32_t ne01; + int32_t ne02; + int32_t ne03; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne11; + int32_t ne_12_2; // assume K and V are same shape + int32_t ne_12_3; + uint64_t nb_12_1; + uint64_t nb_12_2; + uint64_t nb_12_3; + uint64_t nb31; + int32_t ne1; + int32_t ne2; + float scale; + float max_bias; + float m0; + float m1; + uint16_t n_head_log2; + float logit_softcap; +} ggml_metal_kargs_flash_attn_ext; + +typedef struct { + int32_t ne00; + int32_t ne02; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne12; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + int32_t ne0; + int32_t ne1; + int16_t r2; + int16_t r3; +} ggml_metal_kargs_mul_mm; + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne10; + int32_t ne11; + int32_t ne12; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + int32_t ne0; + int32_t ne1; + int16_t r2; + int16_t r3; +} ggml_metal_kargs_mul_mv; + +typedef struct { + int32_t ne00; + int32_t ne01; + int32_t ne02; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + uint64_t nb03; + int32_t ne10; + int32_t ne11; + int32_t ne12; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + uint64_t nb13; + int32_t ne0; + int32_t ne1; + int16_t r2; + int16_t r3; + int16_t nsg; + int16_t nxpsg; + int16_t r1ptg; +} ggml_metal_kargs_mul_mv_ext; + +typedef struct { + int32_t nei0; + int32_t nei1; + uint64_t nbi1; + int32_t ne00; + int32_t ne02; + uint64_t nb01; + uint64_t nb02; + int32_t ne11; + int32_t ne12; + int32_t ne13; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + int32_t ne0; + int32_t ne1; +} ggml_metal_kargs_mul_mm_id; + +typedef struct { + int32_t nei0; + int32_t nei1; + uint64_t nbi1; + int32_t ne00; + int32_t ne01; + int32_t ne02; + uint64_t nb00; + uint64_t nb01; + uint64_t nb02; + int32_t ne10; + int32_t ne11; + int32_t ne12; + int32_t ne13; + uint64_t nb10; + uint64_t nb11; + uint64_t nb12; + int32_t ne0; + int32_t ne1; + uint64_t nb1; +} ggml_metal_kargs_mul_mv_id; + +typedef struct { + int32_t ne00; + int32_t ne00_4; + uint64_t nb01; + float eps; +} ggml_metal_kargs_norm; + +typedef struct { + int32_t ne00; + int32_t ne00_4; + uint64_t nb01; + float eps; +} ggml_metal_kargs_rms_norm; + +#endif // GGML_METAL_IMPL diff --git a/ggml/src/ggml-metal.m b/ggml/src/ggml-metal/ggml-metal.m similarity index 73% rename from ggml/src/ggml-metal.m rename to ggml/src/ggml-metal/ggml-metal.m index f13adee38..a85502ee0 100644 --- a/ggml/src/ggml-metal.m +++ b/ggml/src/ggml-metal/ggml-metal.m @@ -2,6 +2,7 @@ #import "ggml-impl.h" #import "ggml-backend-impl.h" +#import "ggml-metal-impl.h" #import @@ -39,6 +40,7 @@ static struct ggml_backend_metal_device_context { bool has_simdgroup_reduction; bool has_simdgroup_mm; bool has_bfloat; + bool use_bfloat; char name[128]; } g_ggml_ctx_dev_main = { @@ -47,6 +49,7 @@ static struct ggml_backend_metal_device_context { /*.has_simdgroup_reduction =*/ false, /*.has_simdgroup_mm =*/ false, /*.has_bfloat =*/ false, + /*.use_bfloat =*/ false, /*.name =*/ "", }; @@ -65,6 +68,12 @@ static id ggml_backend_metal_device_acq(struct ggml_backend_metal_dev ctx->has_bfloat = [ctx->mtl_device supportsFamily:MTLGPUFamilyMetal3_GGML]; ctx->has_bfloat |= [ctx->mtl_device supportsFamily:MTLGPUFamilyApple6]; +#if defined(GGML_METAL_USE_BF16) + ctx->use_bfloat = ctx->has_bfloat; +#else + ctx->use_bfloat = false; +#endif + strncpy(ctx->name, [[ctx->mtl_device name] UTF8String], sizeof(ctx->name) - 1); } @@ -117,6 +126,7 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_GELU_QUICK_4, GGML_METAL_KERNEL_TYPE_SILU, GGML_METAL_KERNEL_TYPE_SILU_4, + GGML_METAL_KERNEL_TYPE_ELU, GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16, GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16_4, GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32, @@ -165,6 +175,46 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32, GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32, GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4, + GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5, GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32, GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32, GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32, @@ -256,8 +306,11 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_IM2COL_F32, GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F16, GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F32, + GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32, + GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32, GGML_METAL_KERNEL_TYPE_UPSCALE_F32, GGML_METAL_KERNEL_TYPE_PAD_F32, + GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32, GGML_METAL_KERNEL_TYPE_ARANGE_F32, GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32, GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_ASC, @@ -269,6 +322,12 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H112, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H256, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H64, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H80, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H96, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H112, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H128, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H64, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H80, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H96, @@ -300,17 +359,21 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q8_0_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q8_0_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H128, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_1_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q8_0_H128, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H256, + GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_1_H256, GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q8_0_H256, + GGML_METAL_KERNEL_TYPE_SET_I32, + GGML_METAL_KERNEL_TYPE_SET_F32, GGML_METAL_KERNEL_TYPE_CPY_F32_F32, GGML_METAL_KERNEL_TYPE_CPY_F32_F16, GGML_METAL_KERNEL_TYPE_CPY_F32_BF16, @@ -332,6 +395,7 @@ enum ggml_metal_kernel_type { GGML_METAL_KERNEL_TYPE_SUM_ROWS, GGML_METAL_KERNEL_TYPE_POOL_2D_AVG_F32, GGML_METAL_KERNEL_TYPE_POOL_2D_MAX_F32, + GGML_METAL_KERNEL_TYPE_ARGMAX, GGML_METAL_KERNEL_TYPE_COUNT }; @@ -446,6 +510,35 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de #endif NSString * path_lib = [bundle pathForResource:@"default" ofType:@"metallib"]; + if (path_lib == nil) { + // Try to find the resource in the directory where the current binary located. + NSString * current_binary = [[NSProcessInfo processInfo] arguments][0]; + NSString * bin_dir = [current_binary stringByDeletingLastPathComponent]; + NSString * default_metallib_path = [NSString pathWithComponents:@[bin_dir, @"default.metallib"]]; + if ([[NSFileManager defaultManager] isReadableFileAtPath:default_metallib_path]) { + GGML_LOG_INFO("%s: found '%s'\n", __func__, [default_metallib_path UTF8String]); + NSDictionary * atts = [[NSFileManager defaultManager] attributesOfItemAtPath:default_metallib_path error:&error]; + if (atts && atts[NSFileType] == NSFileTypeSymbolicLink) { + // Optionally, if this is a symlink, try to resolve it. + default_metallib_path = [[NSFileManager defaultManager] destinationOfSymbolicLinkAtPath:default_metallib_path error:&error]; + if (default_metallib_path && [default_metallib_path length] > 0 && ![[default_metallib_path substringToIndex:1] isEqualToString:@"/"]) { + // It is a relative path, adding the binary directory as directory prefix. + default_metallib_path = [NSString pathWithComponents:@[bin_dir, default_metallib_path]]; + } + if (!default_metallib_path || ![[NSFileManager defaultManager] isReadableFileAtPath:default_metallib_path]) { + // Link to the resource could not be resolved. + default_metallib_path = nil; + } else { + GGML_LOG_INFO("%s: symlink resolved '%s'\n", __func__, [default_metallib_path UTF8String]); + } + } + } else { + // The resource couldn't be found in the binary's directory. + default_metallib_path = nil; + } + path_lib = default_metallib_path; + } + if (try_metallib && path_lib != nil) { // pre-compiled library found NSURL * libURL = [NSURL fileURLWithPath:path_lib]; @@ -496,6 +589,14 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de // dictionary of preprocessor macros NSMutableDictionary * prep = [NSMutableDictionary dictionary]; + if (ctx_dev->use_bfloat) { + [prep setObject:@"1" forKey:@"GGML_METAL_USE_BF16"]; + } + +#if GGML_METAL_EMBED_LIBRARY + [prep setObject:@"1" forKey:@"GGML_METAL_EMBED_LIBRARY"]; +#endif + MTLCompileOptions * options = [MTLCompileOptions new]; options.preprocessorMacros = prep; @@ -548,7 +649,8 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_LOG_INFO("%s: simdgroup reduction = %s\n", __func__, ctx_dev->has_simdgroup_reduction ? "true" : "false"); GGML_LOG_INFO("%s: simdgroup matrix mul. = %s\n", __func__, ctx_dev->has_simdgroup_mm ? "true" : "false"); - GGML_LOG_INFO("%s: bfloat = %s\n", __func__, ctx_dev->has_bfloat ? "true" : "false"); + GGML_LOG_INFO("%s: has bfloat = %s\n", __func__, ctx_dev->has_bfloat ? "true" : "false"); + GGML_LOG_INFO("%s: use bfloat = %s\n", __func__, ctx_dev->use_bfloat ? "true" : "false"); GGML_LOG_INFO("%s: hasUnifiedMemory = %s\n", __func__, ctx_dev->mtl_device.hasUnifiedMemory ? "true" : "false"); ctx->capture_next_compute = false; @@ -575,16 +677,14 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de ctx->kernels[i].pipeline = nil; } - /* - GGML_LOG_INFO("%s: loaded %-40s %16p | th_max = %4d | th_width = %4d\n", __func__, "kernel_"#name, (void *) kernel->pipeline, \ - (int) kernel->pipeline.maxTotalThreadsPerThreadgroup, \ - (int) kernel->pipeline.threadExecutionWidth); \ - */ #define GGML_METAL_ADD_KERNEL(e, name, supported) \ if (supported) { \ struct ggml_metal_kernel * kernel = &ctx->kernels[e]; \ id metal_function = [metal_library newFunctionWithName:@"kernel_"#name]; \ kernel->pipeline = [device newComputePipelineStateWithFunction:metal_function error:&error]; \ + GGML_LOG_DEBUG("%s: loaded %-40s %16p | th_max = %4d | th_width = %4d\n", __func__, "kernel_"#name, (void *) kernel->pipeline, \ + (int) kernel->pipeline.maxTotalThreadsPerThreadgroup, \ + (int) kernel->pipeline.threadExecutionWidth); \ [metal_function release]; \ if (error) { \ GGML_LOG_ERROR("%s: error: load pipeline error: %s\n", __func__, [[error description] UTF8String]); \ @@ -597,7 +697,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de const bool has_simdgroup_mm = ctx_dev->has_simdgroup_mm; const bool has_simdgroup_reduction = ctx_dev->has_simdgroup_reduction; - const bool has_bfloat = ctx_dev->has_bfloat; + const bool use_bfloat = ctx_dev->use_bfloat; // simd_sum and simd_max requires MTLGPUFamilyApple7 @@ -625,6 +725,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_QUICK_4, gelu_quick_4, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SILU, silu, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SILU_4, silu_4, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ELU, elu, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16, soft_max_f16, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16_4, soft_max_f16_4, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32, soft_max_f32, has_simdgroup_reduction); @@ -633,7 +734,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF_8, diag_mask_inf_8, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_F32, get_rows_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_F16, get_rows_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_BF16, get_rows_bf16, has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_BF16, get_rows_bf16, use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_0, get_rows_q4_0, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_1, get_rows_q4_1, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_0, get_rows_q5_0, true); @@ -660,10 +761,10 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SSM_CONV_F32, ssm_conv_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32, ssm_scan_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32, mul_mv_f32_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32, mul_mv_bf16_f32, has_simdgroup_reduction && has_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW, mul_mv_bf16_f32_1row, has_simdgroup_reduction && has_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4, mul_mv_bf16_f32_l4, has_simdgroup_reduction && has_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16, mul_mv_bf16_bf16, has_simdgroup_reduction && has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32, mul_mv_bf16_f32, has_simdgroup_reduction && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW, mul_mv_bf16_f32_1row, has_simdgroup_reduction && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4, mul_mv_bf16_f32_l4, has_simdgroup_reduction && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16, mul_mv_bf16_bf16, has_simdgroup_reduction && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32, mul_mv_f16_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW, mul_mv_f16_f32_1row, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4, mul_mv_f16_f32_l4, has_simdgroup_reduction); @@ -673,6 +774,46 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32, mul_mv_q5_0_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32, mul_mv_q5_1_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32, mul_mv_q8_0_f32, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2, mul_mv_ext_f16_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3, mul_mv_ext_f16_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4, mul_mv_ext_f16_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5, mul_mv_ext_f16_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2, mul_mv_ext_q4_0_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3, mul_mv_ext_q4_0_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4, mul_mv_ext_q4_0_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5, mul_mv_ext_q4_0_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2, mul_mv_ext_q4_1_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3, mul_mv_ext_q4_1_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4, mul_mv_ext_q4_1_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5, mul_mv_ext_q4_1_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2, mul_mv_ext_q5_0_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3, mul_mv_ext_q5_0_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4, mul_mv_ext_q5_0_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5, mul_mv_ext_q5_0_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2, mul_mv_ext_q5_1_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3, mul_mv_ext_q5_1_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4, mul_mv_ext_q5_1_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5, mul_mv_ext_q5_1_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2, mul_mv_ext_q8_0_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3, mul_mv_ext_q8_0_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4, mul_mv_ext_q8_0_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5, mul_mv_ext_q8_0_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2, mul_mv_ext_q4_K_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3, mul_mv_ext_q4_K_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4, mul_mv_ext_q4_K_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5, mul_mv_ext_q4_K_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2, mul_mv_ext_q5_K_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3, mul_mv_ext_q5_K_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4, mul_mv_ext_q5_K_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5, mul_mv_ext_q5_K_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2, mul_mv_ext_q6_K_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3, mul_mv_ext_q6_K_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4, mul_mv_ext_q6_K_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5, mul_mv_ext_q6_K_f32_r1_5, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2, mul_mv_ext_iq4_nl_f32_r1_2, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3, mul_mv_ext_iq4_nl_f32_r1_3, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4, mul_mv_ext_iq4_nl_f32_r1_4, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5, mul_mv_ext_iq4_nl_f32_r1_5, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32, mul_mv_q2_K_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32, mul_mv_q3_K_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32, mul_mv_q4_K_f32, has_simdgroup_reduction); @@ -692,7 +833,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_1ROW, mul_mv_id_f16_f32_1row, has_simdgroup_reduction); //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_L4, mul_mv_id_f16_f32_l4, has_simdgroup_reduction); //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F16, mul_mv_id_f16_f16, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_BF16_F32, mul_mv_id_bf16_f32, has_simdgroup_reduction && has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_BF16_F32, mul_mv_id_bf16_f32, has_simdgroup_reduction && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_0_F32, mul_mv_id_q4_0_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_1_F32, mul_mv_id_q4_1_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_0_F32, mul_mv_id_q5_0_f32, has_simdgroup_reduction); @@ -714,7 +855,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_XS_F32, mul_mv_id_iq4_xs_f32, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32, mul_mm_f32_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32, mul_mm_f16_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32, mul_mm_bf16_f32, has_simdgroup_mm && has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32, mul_mm_bf16_f32, has_simdgroup_mm && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32, mul_mm_q4_0_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32, mul_mm_q4_1_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32, mul_mm_q5_0_f32, has_simdgroup_mm); @@ -736,7 +877,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32, mul_mm_iq4_xs_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F32, mul_mm_id_f32_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F32, mul_mm_id_f16_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_BF16_F32, mul_mm_id_bf16_f32, has_simdgroup_mm && has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_BF16_F32, mul_mm_id_bf16_f32, has_simdgroup_mm && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_0_F32, mul_mm_id_q4_0_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_1_F32, mul_mm_id_q4_1_f32, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_0_F32, mul_mm_id_q5_0_f32, has_simdgroup_mm); @@ -764,8 +905,11 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_F32, im2col_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F16, im2col_ext_f16, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F32, im2col_ext_f32, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32, conv_transpose_1d_f32_f32, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32, conv_transpose_1d_f16_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_UPSCALE_F32, upscale_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_PAD_F32, pad_f32, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32, pad_reflect_1d_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32, timestep_embedding_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARANGE_F32, arange_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_ASC, argsort_f32_i32_asc, true); @@ -777,6 +921,12 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H112, flash_attn_ext_f16_h112, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H128, flash_attn_ext_f16_h128, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_F16_H256, flash_attn_ext_f16_h256, has_simdgroup_mm); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H64, flash_attn_ext_bf16_h64, has_simdgroup_mm && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H80, flash_attn_ext_bf16_h80, has_simdgroup_mm && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H96, flash_attn_ext_bf16_h96, has_simdgroup_mm && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H112, flash_attn_ext_bf16_h112, has_simdgroup_mm && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H128, flash_attn_ext_bf16_h128, has_simdgroup_mm && use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H256, flash_attn_ext_bf16_h256, has_simdgroup_mm && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H64, flash_attn_ext_q4_0_h64, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H80, flash_attn_ext_q4_0_h80, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q4_0_H96, flash_attn_ext_q4_0_h96, has_simdgroup_mm); @@ -808,24 +958,28 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q8_0_H128, flash_attn_ext_q8_0_h128, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_Q8_0_H256, flash_attn_ext_q8_0_h256, has_simdgroup_mm); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H128, flash_attn_ext_vec_f16_h128, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H128, flash_attn_ext_vec_bf16_h128, has_simdgroup_reduction && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H128, flash_attn_ext_vec_q4_0_h128, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H128, flash_attn_ext_vec_q4_1_h128, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H128, flash_attn_ext_vec_q5_0_h128, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_1_H128, flash_attn_ext_vec_q5_1_h128, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q8_0_H128, flash_attn_ext_vec_q8_0_h128, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H256, flash_attn_ext_vec_f16_h256, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H256, flash_attn_ext_vec_bf16_h256, has_simdgroup_reduction && use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H256, flash_attn_ext_vec_q4_0_h256, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H256, flash_attn_ext_vec_q4_1_h256, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H256, flash_attn_ext_vec_q5_0_h256, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_1_H256, flash_attn_ext_vec_q5_1_h256, has_simdgroup_reduction); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q8_0_H256, flash_attn_ext_vec_q8_0_h256, has_simdgroup_reduction); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_F32, set_f32, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_I32, set_i32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_F32, cpy_f32_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_F16, cpy_f32_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_BF16, cpy_f32_bf16, has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_BF16, cpy_f32_bf16, use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F32, cpy_f16_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F16, cpy_f16_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_F32, cpy_bf16_f32, has_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_BF16, cpy_bf16_bf16, has_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_F32, cpy_bf16_f32, use_bfloat); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_BF16, cpy_bf16_bf16, use_bfloat); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0, cpy_f32_q8_0, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0, cpy_f32_q4_0, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1, cpy_f32_q4_1, true); @@ -838,6 +992,7 @@ static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t de GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SIN, sin, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_COS, cos, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SUM_ROWS, sum_rows, true); + GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARGMAX, argmax, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_POOL_2D_AVG_F32, pool_2d_avg_f32, true); GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_POOL_2D_MAX_F32, pool_2d_max_f32, true); } @@ -917,9 +1072,9 @@ static id ggml_metal_get_buffer(struct ggml_tensor * t, size_t * offs static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_context * ctx_dev, const struct ggml_tensor * op) { const bool has_simdgroup_mm = ctx_dev->has_simdgroup_mm; const bool has_simdgroup_reduction = ctx_dev->has_simdgroup_reduction; - const bool has_bfloat = ctx_dev->has_bfloat; + const bool use_bfloat = ctx_dev->use_bfloat; - if (!has_bfloat) { + if (!use_bfloat) { for (size_t i = 0, n = 3; i < n; ++i) { if (op->src[i] != NULL && op->src[i]->type == GGML_TYPE_BF16) { return false; @@ -936,6 +1091,7 @@ static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_contex case GGML_UNARY_OP_GELU: case GGML_UNARY_OP_GELU_QUICK: case GGML_UNARY_OP_SILU: + case GGML_UNARY_OP_ELU: return ggml_is_contiguous(op->src[0]); default: return false; @@ -954,6 +1110,7 @@ static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_contex case GGML_OP_REPEAT: case GGML_OP_SCALE: case GGML_OP_CLAMP: + case GGML_OP_CONV_TRANSPOSE_1D: return true; case GGML_OP_SQR: case GGML_OP_SQRT: @@ -962,12 +1119,24 @@ static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_contex return ggml_is_contiguous(op->src[0]); case GGML_OP_SUM_ROWS: case GGML_OP_SOFT_MAX: - case GGML_OP_RMS_NORM: case GGML_OP_GROUP_NORM: return has_simdgroup_reduction; + case GGML_OP_RMS_NORM: + return has_simdgroup_reduction && (op->ne[0] % 4 == 0); + case GGML_OP_ARGMAX: case GGML_OP_NORM: - case GGML_OP_ROPE: return true; + case GGML_OP_ROPE: + { + const int mode = ((const int32_t *) op->op_params)[2]; + if (mode & GGML_ROPE_TYPE_MROPE) { + return false; + } + if (mode & GGML_ROPE_TYPE_VISION) { + return false; + } + return true; + } case GGML_OP_IM2COL: return op->src[0]->type == GGML_TYPE_F16; case GGML_OP_POOL_1D: @@ -975,6 +1144,7 @@ static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_contex case GGML_OP_POOL_2D: case GGML_OP_UPSCALE: case GGML_OP_PAD: + case GGML_OP_PAD_REFLECT_1D: case GGML_OP_ARANGE: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_ARGSORT: @@ -1032,6 +1202,16 @@ static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_contex return false; }; } + case GGML_OP_SET: + { + switch (op->src[0]->type) { + case GGML_TYPE_F32: + case GGML_TYPE_I32: + return true; + default: + return false; + }; + } case GGML_OP_DIAG_MASK_INF: case GGML_OP_GET_ROWS: { @@ -1111,7 +1291,7 @@ static void ggml_metal_encode_node( const uint64_t nb20 = src2 ? src2->nb[0] : 0; GGML_UNUSED(nb20); const uint64_t nb21 = src2 ? src2->nb[1] : 0; const uint64_t nb22 = src2 ? src2->nb[2] : 0; - const uint64_t nb23 = src2 ? src2->nb[3] : 0; + const uint64_t nb23 = src2 ? src2->nb[3] : 0; GGML_UNUSED(nb23); const int64_t ne0 = dst ? dst->ne[0] : 0; const int64_t ne1 = dst ? dst->ne[1] : 0; @@ -1162,35 +1342,39 @@ static void ggml_metal_encode_node( const int32_t dim = ((const int32_t *) dst->op_params)[0]; + ggml_metal_kargs_concat args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.dim =*/ dim, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:3]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:4]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:5]; - [encoder setBytes:&ne03 length:sizeof(ne03) atIndex:6]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:7]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:8]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:9]; - [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:10]; - [encoder setBytes:&ne10 length:sizeof(ne10) atIndex:11]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:12]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:13]; - [encoder setBytes:&ne13 length:sizeof(ne13) atIndex:14]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:15]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:16]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:17]; - [encoder setBytes:&nb13 length:sizeof(nb13) atIndex:18]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:19]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:20]; - [encoder setBytes:&ne2 length:sizeof(ne2) atIndex:21]; - [encoder setBytes:&ne3 length:sizeof(ne3) atIndex:22]; - [encoder setBytes:&nb0 length:sizeof(nb0) atIndex:23]; - [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:24]; - [encoder setBytes:&nb2 length:sizeof(nb2) atIndex:25]; - [encoder setBytes:&nb3 length:sizeof(nb3) atIndex:26]; - [encoder setBytes:&dim length:sizeof(dim) atIndex:27]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; const int nth = MIN(1024, ne0); @@ -1208,8 +1392,6 @@ static void ggml_metal_encode_node( bool bcast_row = false; - int64_t nb = ne00; // used by the "row" kernels - id pipeline = nil; if (ggml_nelements(src1) == ne10 && ggml_is_contiguous(src1) && ne00 % 4 == 0 && ne10 % 4 == 0) { @@ -1218,7 +1400,6 @@ static void ggml_metal_encode_node( // src1 is a row GGML_ASSERT(ne11 == 1); - nb = ne00 / 4; switch (dst->op) { case GGML_OP_ADD: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ADD_ROW].pipeline; break; case GGML_OP_SUB: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SUB_ROW].pipeline; break; @@ -1238,36 +1419,39 @@ static void ggml_metal_encode_node( } } + ggml_metal_kargs_bin args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.offs =*/ offs, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:3]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:4]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:5]; - [encoder setBytes:&ne03 length:sizeof(ne03) atIndex:6]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:7]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:8]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:9]; - [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:10]; - [encoder setBytes:&ne10 length:sizeof(ne10) atIndex:11]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:12]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:13]; - [encoder setBytes:&ne13 length:sizeof(ne13) atIndex:14]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:15]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:16]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:17]; - [encoder setBytes:&nb13 length:sizeof(nb13) atIndex:18]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:19]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:20]; - [encoder setBytes:&ne2 length:sizeof(ne2) atIndex:21]; - [encoder setBytes:&ne3 length:sizeof(ne3) atIndex:22]; - [encoder setBytes:&nb0 length:sizeof(nb0) atIndex:23]; - [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:24]; - [encoder setBytes:&nb2 length:sizeof(nb2) atIndex:25]; - [encoder setBytes:&nb3 length:sizeof(nb3) atIndex:26]; - [encoder setBytes:&offs length:sizeof(offs) atIndex:27]; - [encoder setBytes:&nb length:sizeof(nb) atIndex:28]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; if (bcast_row) { const int64_t n = ggml_nelements(dst)/4; @@ -1291,25 +1475,29 @@ static void ggml_metal_encode_node( default: GGML_ABORT("fatal error"); } + ggml_metal_kargs_repeat args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:2]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:3]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:4]; - [encoder setBytes:&ne03 length:sizeof(ne03) atIndex:5]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:6]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:7]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:8]; - [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:9]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:10]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:11]; - [encoder setBytes:&ne2 length:sizeof(ne2) atIndex:12]; - [encoder setBytes:&ne3 length:sizeof(ne3) atIndex:13]; - [encoder setBytes:&nb0 length:sizeof(nb0) atIndex:14]; - [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:15]; - [encoder setBytes:&nb2 length:sizeof(nb2) atIndex:16]; - [encoder setBytes:&nb3 length:sizeof(nb3) atIndex:17]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne0); @@ -1338,25 +1526,29 @@ static void ggml_metal_encode_node( const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F32].pipeline; + ggml_metal_kargs_cpy args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; - [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:3]; - [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:4]; - [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:5]; - [encoder setBytes:&nb00 length:sizeof(uint64_t) atIndex:6]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:7]; - [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:8]; - [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:9]; - [encoder setBytes:&ne0 length:sizeof( int64_t) atIndex:10]; - [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:11]; - [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:12]; - [encoder setBytes:&ne3 length:sizeof( int64_t) atIndex:13]; - [encoder setBytes:&nb0 length:sizeof(uint64_t) atIndex:14]; - [encoder setBytes:&nb1 length:sizeof(uint64_t) atIndex:15]; - [encoder setBytes:&nb2 length:sizeof(uint64_t) atIndex:16]; - [encoder setBytes:&nb3 length:sizeof(uint64_t) atIndex:17]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00); @@ -1365,35 +1557,39 @@ static void ggml_metal_encode_node( const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ADD].pipeline; + ggml_metal_kargs_bin args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ pnb1, + /*.nb02 =*/ pnb2, + /*.nb03 =*/ pnb3, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ pnb1, + /*.nb2 =*/ pnb2, + /*.nb3 =*/ pnb3, + /*.offs =*/ offs, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:3]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:4]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:5]; - [encoder setBytes:&ne03 length:sizeof(ne03) atIndex:6]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:7]; - [encoder setBytes:&pnb1 length:sizeof(pnb1) atIndex:8]; - [encoder setBytes:&pnb2 length:sizeof(pnb2) atIndex:9]; - [encoder setBytes:&pnb3 length:sizeof(pnb3) atIndex:10]; - [encoder setBytes:&ne10 length:sizeof(ne10) atIndex:11]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:12]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:13]; - [encoder setBytes:&ne13 length:sizeof(ne13) atIndex:14]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:15]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:16]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:17]; - [encoder setBytes:&nb13 length:sizeof(nb13) atIndex:18]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:19]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:20]; - [encoder setBytes:&ne2 length:sizeof(ne2) atIndex:21]; - [encoder setBytes:&ne3 length:sizeof(ne3) atIndex:22]; - [encoder setBytes:&nb0 length:sizeof(nb0) atIndex:23]; - [encoder setBytes:&pnb1 length:sizeof(pnb1) atIndex:24]; - [encoder setBytes:&pnb2 length:sizeof(pnb2) atIndex:25]; - [encoder setBytes:&pnb3 length:sizeof(pnb3) atIndex:26]; - [encoder setBytes:&offs length:sizeof(offs) atIndex:27]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00); @@ -1434,10 +1630,10 @@ static void ggml_metal_encode_node( memcpy(&max, ((const int32_t *) dst->op_params) + 1, sizeof(float)); [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&min length:sizeof(min) atIndex:2]; - [encoder setBytes:&max length:sizeof(max) atIndex:3]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; + [encoder setBytes:&min length:sizeof(min) atIndex:2]; + [encoder setBytes:&max length:sizeof(max) atIndex:3]; const int64_t n = ggml_nelements(dst); @@ -1541,6 +1737,18 @@ static void ggml_metal_encode_node( [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; } break; + case GGML_UNARY_OP_ELU: + { + id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ELU].pipeline; + + [encoder setComputePipelineState:pipeline]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; + + const int64_t n = ggml_nelements(dst); + + [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; + } break; default: { GGML_LOG_WARN("%s: node %3d, op = %8s not implemented\n", __func__, idx, ggml_op_name(dst->op)); @@ -1609,6 +1817,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SUM_ROWS].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -1684,6 +1893,8 @@ static void ggml_metal_encode_node( const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + // TODO: add ggml_metal_kargs struct + // TODO: optimize (see https://github.com/ggerganov/llama.cpp/pull/10238/commits/7941b6b9ec29a2866fec6fa6c51612515ca509f6) [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; if (id_src1) { @@ -1700,6 +1911,7 @@ static void ggml_metal_encode_node( [encoder setBytes:&m0 length:sizeof(m0) atIndex:8]; [encoder setBytes:&m1 length:sizeof(m1) atIndex:9]; [encoder setBytes:&n_head_log2 length:sizeof(n_head_log2) atIndex:10]; + [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; [encoder dispatchThreadgroups:MTLSizeMake(ne01*ne02*ne03, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; @@ -1716,6 +1928,7 @@ static void ggml_metal_encode_node( pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF].pipeline; } + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -1740,6 +1953,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SSM_CONV_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; @@ -1810,6 +2024,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; @@ -1852,336 +2067,495 @@ static void ggml_metal_encode_node( GGML_ASSERT(ne12 % ne02 == 0); GGML_ASSERT(ne13 % ne03 == 0); - const uint r2 = ne12/ne02; - const uint r3 = ne13/ne03; + const uint32_t r2 = ne12/ne02; + const uint32_t r3 = ne13/ne03; // find the break-even point where the matrix-matrix kernel becomes more efficient compared // to the matrix-vector kernel - int ne11_mm_min = 1; + const int ne11_mm_min = 4; -#if 0 - // the numbers below are measured on M2 Ultra for 7B and 13B models - // these numbers do not translate to other devices or model sizes - // TODO: need to find a better approach - if ([device.name isEqualToString:@"Apple M2 Ultra"]) { - switch (src0t) { - case GGML_TYPE_F16: ne11_mm_min = 2; break; - case GGML_TYPE_Q8_0: ne11_mm_min = 7; break; - case GGML_TYPE_Q2_K: ne11_mm_min = 15; break; - case GGML_TYPE_Q3_K: ne11_mm_min = 7; break; - case GGML_TYPE_Q4_0: - case GGML_TYPE_Q4_1: ne11_mm_min = 15; break; - case GGML_TYPE_Q4_K: ne11_mm_min = 11; break; - case GGML_TYPE_Q5_0: // not tested yet - case GGML_TYPE_Q5_1: ne11_mm_min = 13; break; // not tested yet - case GGML_TYPE_Q5_K: ne11_mm_min = 7; break; - case GGML_TYPE_Q6_K: ne11_mm_min = 7; break; - default: ne11_mm_min = 1; break; - } - } -#endif + // first try to use small-batch mat-mv kernels + // these should be efficient for BS [2, ~8] + if (src1t == GGML_TYPE_F32 && (ne00%256 == 0) && + ( + ( + ( + src0t == GGML_TYPE_F16 || // TODO: helper function + src0t == GGML_TYPE_Q4_0 || + src0t == GGML_TYPE_Q4_1 || + src0t == GGML_TYPE_Q5_0 || + src0t == GGML_TYPE_Q5_1 || + src0t == GGML_TYPE_Q8_0 || + src0t == GGML_TYPE_IQ4_NL || + false) && (ne11 >= 2 && ne11 <= 8) + ) || + ( + ( + src0t == GGML_TYPE_Q4_K || + src0t == GGML_TYPE_Q5_K || + src0t == GGML_TYPE_Q6_K || + false) && (ne11 >= 4 && ne11 <= 8) + ) + ) + ) { + // TODO: determine the optimal parameters based on grid utilization + // I still don't know why we should not always use the maximum available threads: + // + // nsg = pipeline.maxTotalThreadsPerThreadgroup / 32 + // + // my current hypothesis is that the work grid is not evenly divisible for different nsg + // values and there can be some tail effects when nsg is high. need to confirm this + // + const int nsg = 2; // num simdgroups per threadgroup + const int nxpsg = ne11 < 3 ? 16 : 8; // num threads along row per simdgroup + const int nypsg = 32/nxpsg; // num threads along col per simdgroup (i.e. a simdgroup processes that many src0 rows at a time) + const int r0ptg = nypsg*nsg; // num src0 rows per threadgroup + int r1ptg = 4; // num src1 rows per threadgroup - // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs - // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel - if ([device supportsFamily:MTLGPUFamilyApple7] && - !ggml_is_transposed(src0) && - !ggml_is_transposed(src1) && - src1t == GGML_TYPE_F32 && - ne00 % 32 == 0 && ne00 >= 64 && - (ne11 > ne11_mm_min || (ggml_is_quantized(src0t) && ne12 > 1))) { - //printf("matrix: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); + // note: not sure how optimal are those across all different hardware. there might be someting cleverer + switch (ne11) { + case 2: + r1ptg = 2; break; + case 3: + case 6: + r1ptg = 3; break; + case 4: + case 7: + case 8: + r1ptg = 4; break; + case 5: + r1ptg = 5; break; + }; - // some Metal matrix data types require aligned pointers - // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) - switch (src0->type) { - case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; - case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; - case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; - default: break; - } + id pipeline = nil; - id pipeline = nil; + switch (src0->type) { + case GGML_TYPE_F16: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q4_0: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q4_1: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q5_0: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q5_1: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q8_0: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q4_K: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q5_K: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_Q6_K: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + case GGML_TYPE_IQ4_NL: + switch (r1ptg) { + case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2].pipeline; break; + case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3].pipeline; break; + case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4].pipeline; break; + case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5].pipeline; break; + default: GGML_ABORT("not implemented"); + } break; + default: GGML_ABORT("not implemented"); + } - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32 ].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32 ].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32 ].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32 ].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32 ].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32 ].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_1_F32 ].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q8_0_F32 ].pipeline; break; - case GGML_TYPE_Q2_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q2_K_F32 ].pipeline; break; - case GGML_TYPE_Q3_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q3_K_F32 ].pipeline; break; - case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32 ].pipeline; break; - case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32 ].pipeline; break; - case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32 ].pipeline; break; - case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32].pipeline; break; - case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32 ].pipeline; break; - case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32].pipeline; break; - case GGML_TYPE_IQ3_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_S_F32 ].pipeline; break; - case GGML_TYPE_IQ2_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_S_F32 ].pipeline; break; - case GGML_TYPE_IQ1_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_S_F32 ].pipeline; break; - case GGML_TYPE_IQ1_M: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_M_F32 ].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_NL_F32 ].pipeline; break; - case GGML_TYPE_IQ4_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32 ].pipeline; break; - default: GGML_ABORT("MUL MAT-MAT not implemented"); - } + ggml_metal_kargs_mul_mv_ext args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + /*.nsg =*/ nsg, + /*.nxpsg =*/ nxpsg, + /*.r1ptg =*/ r1ptg, + }; - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:3]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:4]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:5]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:6]; - [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:7]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:8]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:9]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:10]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:11]; - [encoder setBytes:&nb13 length:sizeof(nb13) atIndex:12]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:13]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:14]; - [encoder setBytes:&r2 length:sizeof(r2) atIndex:15]; - [encoder setBytes:&r3 length:sizeof(r3) atIndex:16]; - [encoder setThreadgroupMemoryLength:8192 atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake( (ne11 + 31)/32, (ne01 + 63)/64, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(128, 1, 1)]; - } else { - int nth0 = 32; - int nth1 = 1; - int nrows = 1; - //printf("vector: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); + [encoder setComputePipelineState:pipeline]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - id pipeline = nil; + //printf("ne01 = %lld nr0ptg = %d\n", ne01, nr0ptg); + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + r0ptg - 1)/r0ptg, (ne11 + r1ptg - 1)/r1ptg, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; + } else + // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs + // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel + if ([device supportsFamily:MTLGPUFamilyApple7] && + !ggml_is_transposed(src0) && + !ggml_is_transposed(src1) && + src1t == GGML_TYPE_F32 && + ne00 % 32 == 0 && ne00 >= 64 && + (ne11 > ne11_mm_min || (ggml_is_quantized(src0t) && ne12 > 1))) { + //printf("matrix: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); - // use custom matrix x vector kernel - switch (src0t) { - case GGML_TYPE_F32: - { - GGML_ASSERT(src1t == GGML_TYPE_F32); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32].pipeline; + // some Metal matrix data types require aligned pointers + // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) + switch (src0->type) { + case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; + case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; + case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; + default: break; + } + + id pipeline = nil; + + switch (src0->type) { + case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32 ].pipeline; break; + case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32 ].pipeline; break; + case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32 ].pipeline; break; + case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32 ].pipeline; break; + case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32 ].pipeline; break; + case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32 ].pipeline; break; + case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_1_F32 ].pipeline; break; + case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q8_0_F32 ].pipeline; break; + case GGML_TYPE_Q2_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q2_K_F32 ].pipeline; break; + case GGML_TYPE_Q3_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q3_K_F32 ].pipeline; break; + case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32 ].pipeline; break; + case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32 ].pipeline; break; + case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32 ].pipeline; break; + case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32].pipeline; break; + case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32 ].pipeline; break; + case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32].pipeline; break; + case GGML_TYPE_IQ3_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_S_F32 ].pipeline; break; + case GGML_TYPE_IQ2_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_S_F32 ].pipeline; break; + case GGML_TYPE_IQ1_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_S_F32 ].pipeline; break; + case GGML_TYPE_IQ1_M: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_M_F32 ].pipeline; break; + case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_NL_F32 ].pipeline; break; + case GGML_TYPE_IQ4_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32 ].pipeline; break; + default: GGML_ABORT("MUL MAT-MAT not implemented"); + } + + ggml_metal_kargs_mul_mm args = { + /*.ne00 =*/ ne00, + /*.ne02 =*/ ne02, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + }; + + [encoder setComputePipelineState:pipeline]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; + + [encoder setThreadgroupMemoryLength:8192 atIndex:0]; + [encoder dispatchThreadgroups:MTLSizeMake( (ne11 + 31)/32, (ne01 + 63)/64, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(128, 1, 1)]; + } else { + int nth0 = 32; + int nth1 = 1; + int nrows = 1; + //printf("vector: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); + + id pipeline = nil; + + // use custom matrix x vector kernel + switch (src0t) { + case GGML_TYPE_F32: + { + GGML_ASSERT(src1t == GGML_TYPE_F32); + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32].pipeline; + nrows = 4; + } break; + case GGML_TYPE_F16: + { + nth0 = 32; + nth1 = 1; + if (src1t == GGML_TYPE_F32) { + if (ne11 * ne12 < 4) { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW].pipeline; + } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4].pipeline; + nrows = ne11; + } else { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32].pipeline; nrows = 4; - } break; - case GGML_TYPE_F16: - { - nth0 = 32; - nth1 = 1; - if (src1t == GGML_TYPE_F32) { - if (ne11 * ne12 < 4) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW].pipeline; - } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4].pipeline; - nrows = ne11; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32].pipeline; - nrows = 4; - } - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F16].pipeline; - nrows = 4; - } - } break; - case GGML_TYPE_BF16: - { - nth0 = 32; - nth1 = 1; - if (src1t == GGML_TYPE_F32) { - if (ne11 * ne12 < 4) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW].pipeline; - } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4].pipeline; - nrows = ne11; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32].pipeline; - nrows = 4; - } - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16].pipeline; - nrows = 4; - } - } break; - case GGML_TYPE_Q4_0: - { - nth0 = 8; - nth1 = 8; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_0_F32].pipeline; - } break; - case GGML_TYPE_Q4_1: - { - nth0 = 8; - nth1 = 8; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_1_F32].pipeline; - } break; - case GGML_TYPE_Q5_0: - { - nth0 = 8; - nth1 = 8; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32].pipeline; - } break; - case GGML_TYPE_Q5_1: - { - nth0 = 8; - nth1 = 8; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32].pipeline; - } break; - case GGML_TYPE_Q8_0: - { - nth0 = 8; - nth1 = 8; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32].pipeline; - } break; - case GGML_TYPE_Q2_K: - { - nth0 = 2; - nth1 = 32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32].pipeline; - } break; - case GGML_TYPE_Q3_K: - { - nth0 = 2; - nth1 = 32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32].pipeline; - } break; - case GGML_TYPE_Q4_K: - { - nth0 = 4; //1; - nth1 = 8; //32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32].pipeline; - } break; - case GGML_TYPE_Q5_K: - { - nth0 = 2; - nth1 = 32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32].pipeline; - } break; - case GGML_TYPE_Q6_K: - { - nth0 = 2; - nth1 = 32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XXS: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XS: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_XXS: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_S: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_S_F32].pipeline; - } break; - case GGML_TYPE_IQ2_S: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_S: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_M: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_M_F32].pipeline; - } break; - case GGML_TYPE_IQ4_NL: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_NL_F32].pipeline; - } break; - case GGML_TYPE_IQ4_XS: - { - nth0 = 4; - nth1 = 16; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_XS_F32].pipeline; - } break; - default: - { - GGML_LOG_ERROR("Asserting on type %d\n", (int)src0t); - GGML_ABORT("not implemented"); } - }; + } else { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F16].pipeline; + nrows = 4; + } + } break; + case GGML_TYPE_BF16: + { + nth0 = 32; + nth1 = 1; + if (src1t == GGML_TYPE_F32) { + if (ne11 * ne12 < 4) { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW].pipeline; + } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4].pipeline; + nrows = ne11; + } else { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32].pipeline; + nrows = 4; + } + } else { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16].pipeline; + nrows = 4; + } + } break; + case GGML_TYPE_Q4_0: + { + nth0 = 8; + nth1 = 8; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_0_F32].pipeline; + } break; + case GGML_TYPE_Q4_1: + { + nth0 = 8; + nth1 = 8; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_1_F32].pipeline; + } break; + case GGML_TYPE_Q5_0: + { + nth0 = 8; + nth1 = 8; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32].pipeline; + } break; + case GGML_TYPE_Q5_1: + { + nth0 = 8; + nth1 = 8; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32].pipeline; + } break; + case GGML_TYPE_Q8_0: + { + nth0 = 8; + nth1 = 8; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32].pipeline; + } break; + case GGML_TYPE_Q2_K: + { + nth0 = 2; + nth1 = 32; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32].pipeline; + } break; + case GGML_TYPE_Q3_K: + { + nth0 = 2; + nth1 = 32; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32].pipeline; + } break; + case GGML_TYPE_Q4_K: + { + nth0 = 4; //1; + nth1 = 8; //32; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32].pipeline; + } break; + case GGML_TYPE_Q5_K: + { + nth0 = 2; + nth1 = 32; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32].pipeline; + } break; + case GGML_TYPE_Q6_K: + { + nth0 = 2; + nth1 = 32; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32].pipeline; + } break; + case GGML_TYPE_IQ2_XXS: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32].pipeline; + } break; + case GGML_TYPE_IQ2_XS: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32].pipeline; + } break; + case GGML_TYPE_IQ3_XXS: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32].pipeline; + } break; + case GGML_TYPE_IQ3_S: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_S_F32].pipeline; + } break; + case GGML_TYPE_IQ2_S: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_S_F32].pipeline; + } break; + case GGML_TYPE_IQ1_S: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_S_F32].pipeline; + } break; + case GGML_TYPE_IQ1_M: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_M_F32].pipeline; + } break; + case GGML_TYPE_IQ4_NL: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_NL_F32].pipeline; + } break; + case GGML_TYPE_IQ4_XS: + { + nth0 = 4; + nth1 = 16; + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_XS_F32].pipeline; + } break; + default: + { + GGML_LOG_ERROR("Asserting on type %d\n", (int)src0t); + GGML_ABORT("not implemented"); + } + }; - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:3]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:4]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:5]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:6]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:7]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:8]; - [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:9]; - [encoder setBytes:&ne10 length:sizeof(ne10) atIndex:10]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:11]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:12]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:13]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:14]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:15]; - [encoder setBytes:&nb13 length:sizeof(nb13) atIndex:16]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:17]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:18]; - [encoder setBytes:&r2 length:sizeof(r2) atIndex:19]; - [encoder setBytes:&r3 length:sizeof(r3) atIndex:20]; + ggml_metal_kargs_mul_mv args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + }; - if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 || src0t == GGML_TYPE_Q5_0 || - src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 || src0t == GGML_TYPE_Q2_K || - src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_IQ2_XXS || src0t == GGML_TYPE_IQ2_XS) { - const int mem_size = src0t == GGML_TYPE_IQ2_XXS ? 256*8+128 : 512*8+128; - [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_IQ3_XXS || src0t == GGML_TYPE_IQ3_S) { - const int mem_size = src0t == GGML_TYPE_IQ3_XXS ? 256*4+128 : 512*4; - [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_IQ4_NL || src0t == GGML_TYPE_IQ4_XS) { - const int mem_size = 32*sizeof(float); - [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_Q4_K) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_Q3_K) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_Q5_K) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - else if (src0t == GGML_TYPE_Q6_K) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 1)/2, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } else { - const int64_t ny = (ne11 + nrows - 1)/nrows; - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ny, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; - } - } + [encoder setComputePipelineState:pipeline]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; + + if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 || src0t == GGML_TYPE_Q5_0 || + src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 || src0t == GGML_TYPE_Q2_K || + src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S) { + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_IQ2_XXS || src0t == GGML_TYPE_IQ2_XS) { + const int mem_size = src0t == GGML_TYPE_IQ2_XXS ? 256*8+128 : 512*8+128; + [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_IQ3_XXS || src0t == GGML_TYPE_IQ3_S) { + const int mem_size = src0t == GGML_TYPE_IQ3_XXS ? 256*4+128 : 512*4; + [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_IQ4_NL || src0t == GGML_TYPE_IQ4_XS) { + const int mem_size = 32*sizeof(float); + [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_Q4_K) { + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_Q3_K) { + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_Q5_K) { + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 3)/4, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + else if (src0t == GGML_TYPE_Q6_K) { + [encoder dispatchThreadgroups:MTLSizeMake((ne01 + 1)/2, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } else { + const int64_t ny = (ne11 + nrows - 1)/nrows; + [encoder dispatchThreadgroups:MTLSizeMake(ne01, ny, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)]; + } + } } break; case GGML_OP_MUL_MAT_ID: { @@ -2257,27 +2631,30 @@ static void ggml_metal_encode_node( default: GGML_ABORT("MUL_MAT_ID not implemented"); } + ggml_metal_kargs_mul_mm_id args = { + /*.nei0 =*/ ne20, + /*.nei1 =*/ ne21, + /*.nbi1 =*/ nb21, + /*.ne00 =*/ ne00, + /*.ne02 =*/ ne02, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - [encoder setBytes:&ne20 length:sizeof(ne20) atIndex:4]; - [encoder setBytes:&ne21 length:sizeof(ne21) atIndex:5]; - [encoder setBytes:&nb21 length:sizeof(nb21) atIndex:6]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:7]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:8]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:9]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:10]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:11]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:12]; - [encoder setBytes:&ne13 length:sizeof(ne13) atIndex:13]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:14]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:15]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:16]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:17]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:18]; - [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:19]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; + [encoder setBuffer:id_src2 offset:offs_src2 atIndex:4]; [encoder setThreadgroupMemoryLength:GGML_PAD(8192 + dst_rows*4/*sizeof(ushort2)*/, 16) atIndex:0]; @@ -2436,30 +2813,34 @@ static void ggml_metal_encode_node( GGML_ASSERT(ne00 >= nth0*nth1); } + ggml_metal_kargs_mul_mv_id args = { + /*.nei0 =*/ ne20, + /*.nei1 =*/ ne21, + /*.nbi1 =*/ nb21, + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.nb1 =*/ nb1, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - [encoder setBytes:&ne20 length:sizeof(ne20) atIndex:4]; - [encoder setBytes:&ne21 length:sizeof(ne21) atIndex:5]; - [encoder setBytes:&nb21 length:sizeof(nb21) atIndex:6]; - [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:7]; - [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:8]; - [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:9]; - [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:10]; - [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:11]; - [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:12]; - [encoder setBytes:&ne10 length:sizeof(ne10) atIndex:13]; - [encoder setBytes:&ne11 length:sizeof(ne11) atIndex:14]; - [encoder setBytes:&ne12 length:sizeof(ne12) atIndex:15]; - [encoder setBytes:&ne13 length:sizeof(ne13) atIndex:16]; - [encoder setBytes:&nb10 length:sizeof(nb10) atIndex:17]; - [encoder setBytes:&nb11 length:sizeof(nb11) atIndex:18]; - [encoder setBytes:&nb12 length:sizeof(nb12) atIndex:19]; - [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:20]; - [encoder setBytes:&ne1 length:sizeof(ne1) atIndex:21]; - [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:22]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; + [encoder setBuffer:id_src2 offset:offs_src2 atIndex:4]; const int64_t _ne1 = 1; const int tgz = dst_rows; @@ -2532,6 +2913,7 @@ static void ggml_metal_encode_node( default: GGML_ABORT("not implemented"); } + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; @@ -2555,20 +2937,28 @@ static void ggml_metal_encode_node( float eps; memcpy(&eps, dst->op_params, sizeof(float)); + id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_RMS_NORM].pipeline; + int nth = 32; // SIMD width - while (nth < ne00/4 && nth < 1024) { + while (nth < ne00/4 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { nth *= 2; } - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_RMS_NORM].pipeline; + nth = MIN(nth, ne00/4); + + ggml_metal_kargs_rms_norm args = { + /*.ne00 =*/ ne00, + /*.ne00_4 =*/ ne00/4, + /*.nb01 =*/ nb01, + /*.eps =*/ eps, + }; [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:3]; - [encoder setBytes:&eps length:sizeof( float) atIndex:4]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; + [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; const int64_t nrows = ggml_nrows(src0); @@ -2577,7 +2967,6 @@ static void ggml_metal_encode_node( } break; case GGML_OP_GROUP_NORM: { - GGML_ASSERT(ne00 % 4 == 0); GGML_ASSERT(ggml_is_contiguous(src0)); float eps; @@ -2593,6 +2982,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GROUP_NORM].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2610,22 +3000,35 @@ static void ggml_metal_encode_node( } break; case GGML_OP_NORM: { + GGML_ASSERT(ne00 % 4 == 0); GGML_ASSERT(ggml_is_contiguous_1(src0)); float eps; memcpy(&eps, dst->op_params, sizeof(float)); - const int nth = MIN(256, ne00); - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_NORM].pipeline; + int nth = 32; // SIMD width + + while (nth < ne00/4 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { + nth *= 2; + } + + nth = MIN(nth, ne00/4); + + ggml_metal_kargs_norm args = { + /*.ne00 =*/ ne00, + /*.ne00_4 =*/ ne00/4, + /*.nb01 =*/ nb01, + /*.eps =*/ eps, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:3]; - [encoder setBytes:&eps length:sizeof( float) atIndex:4]; - [encoder setThreadgroupMemoryLength:GGML_PAD(nth*sizeof(float), 16) atIndex:0]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; + + [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; const int64_t nrows = ggml_nrows(src0); @@ -2633,7 +3036,9 @@ static void ggml_metal_encode_node( } break; case GGML_OP_ROPE: { - GGML_ASSERT(ne10 == ne02); + // make sure we have one or more position id(ne10) per token(ne02) + GGML_ASSERT(ne10 % ne02 == 0); + GGML_ASSERT(ne10 >= ne02); const int nth = MIN(1024, ne00); @@ -2675,40 +3080,44 @@ static void ggml_metal_encode_node( }; } + ggml_metal_kargs_rope args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.n_past =*/ n_past, + /*.n_dims =*/ n_dims, + /*.n_ctx_orig =*/ n_ctx_orig, + /*.freq_base =*/ freq_base, + /*.freq_scale =*/ freq_scale, + /*.ext_factor =*/ ext_factor, + /*.attn_factor =*/ attn_factor, + /*.beta_fast =*/ beta_fast, + /*.beta_slow =*/ beta_slow, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; if (id_src2 != nil) { - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; + [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:2]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:3]; } - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:4]; - [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:5]; - [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:6]; - [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:7]; - [encoder setBytes:&nb00 length:sizeof(uint64_t) atIndex:8]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:9]; - [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:10]; - [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:11]; - [encoder setBytes:&ne0 length:sizeof( int64_t) atIndex:12]; - [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:13]; - [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:14]; - [encoder setBytes:&ne3 length:sizeof( int64_t) atIndex:15]; - [encoder setBytes:&nb0 length:sizeof(uint64_t) atIndex:16]; - [encoder setBytes:&nb1 length:sizeof(uint64_t) atIndex:17]; - [encoder setBytes:&nb2 length:sizeof(uint64_t) atIndex:18]; - [encoder setBytes:&nb3 length:sizeof(uint64_t) atIndex:19]; - [encoder setBytes:&n_past length:sizeof( int) atIndex:20]; - [encoder setBytes:&n_dims length:sizeof( int) atIndex:21]; - [encoder setBytes:&n_ctx_orig length:sizeof( int) atIndex:22]; - [encoder setBytes:&freq_base length:sizeof( float) atIndex:23]; - [encoder setBytes:&freq_scale length:sizeof( float) atIndex:24]; - [encoder setBytes:&ext_factor length:sizeof( float) atIndex:25]; - [encoder setBytes:&attn_factor length:sizeof( float) atIndex:26]; - [encoder setBytes:&beta_fast length:sizeof( float) atIndex:27]; - [encoder setBytes:&beta_slow length:sizeof( float) atIndex:28]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; } break; @@ -2765,6 +3174,7 @@ static void ggml_metal_encode_node( default: GGML_ABORT("fatal error"); }; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src1 offset:offs_src1 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2794,6 +3204,49 @@ static void ggml_metal_encode_node( [encoder dispatchThreadgroups:MTLSizeMake(IC, OH, OW) threadsPerThreadgroup:MTLSizeMake(N, KH, KW)]; } } break; + case GGML_OP_CONV_TRANSPOSE_1D: + { + GGML_ASSERT(ggml_is_contiguous(src0)); + GGML_ASSERT(ggml_is_contiguous(src1)); + GGML_ASSERT(src0->type == GGML_TYPE_F16 || src0->type == GGML_TYPE_F32); + GGML_ASSERT(src1->type == GGML_TYPE_F32); + GGML_ASSERT( dst->type == GGML_TYPE_F32); + + const int32_t s0 = ((const int32_t *)(dst->op_params))[0]; + + const int32_t IC = src1->ne[1]; + const int32_t IL = src1->ne[0]; + + const int32_t K = src0->ne[0]; + + const int32_t OL = dst->ne[0]; + const int32_t OC = dst->ne[1]; + + id pipeline; + + switch (src0->type) { + case GGML_TYPE_F32: { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32].pipeline; + } break; + case GGML_TYPE_F16: { + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32].pipeline; + } break; + default: GGML_ABORT("fatal error"); + }; + + [encoder setComputePipelineState:pipeline]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; + [encoder setBytes:&IC length:sizeof( int32_t) atIndex:3]; + [encoder setBytes:&IL length:sizeof( int32_t) atIndex:4]; + [encoder setBytes:&K length:sizeof( int32_t) atIndex:5]; + [encoder setBytes:&s0 length:sizeof( int32_t) atIndex:6]; + [encoder setBytes:&nb0 length:sizeof(uint64_t) atIndex:7]; + [encoder setBytes:&nb1 length:sizeof(uint64_t) atIndex:8]; + + [encoder dispatchThreadgroups:MTLSizeMake(OL, OC, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; + } break; case GGML_OP_UPSCALE: { GGML_ASSERT(src0->type == GGML_TYPE_F32); @@ -2805,6 +3258,7 @@ static void ggml_metal_encode_node( const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_UPSCALE_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2839,6 +3293,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_PAD_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2861,6 +3316,38 @@ static void ggml_metal_encode_node( const int nth = MIN(1024, ne0); + [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + } break; + case GGML_OP_PAD_REFLECT_1D: + { + GGML_ASSERT(src0->type == GGML_TYPE_F32); + + const int32_t p0 = ((const int32_t *)(dst->op_params))[0]; + const int32_t p1 = ((const int32_t *)(dst->op_params))[1]; + + id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32].pipeline; + + [encoder setComputePipelineState:pipeline]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; + [encoder setBytes:&ne00 length:sizeof(ne00) atIndex:2]; + [encoder setBytes:&ne01 length:sizeof(ne01) atIndex:3]; + [encoder setBytes:&ne02 length:sizeof(ne02) atIndex:4]; + [encoder setBytes:&ne03 length:sizeof(ne03) atIndex:5]; + [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:6]; + [encoder setBytes:&nb00 length:sizeof(nb00) atIndex:7]; + [encoder setBytes:&nb01 length:sizeof(nb01) atIndex:8]; + [encoder setBytes:&nb02 length:sizeof(nb02) atIndex:9]; + [encoder setBytes:&nb03 length:sizeof(nb03) atIndex:10]; + [encoder setBytes:&nb0 length:sizeof(nb0) atIndex:11]; + [encoder setBytes:&nb1 length:sizeof(nb1) atIndex:12]; + [encoder setBytes:&nb2 length:sizeof(nb2) atIndex:13]; + [encoder setBytes:&nb3 length:sizeof(nb3) atIndex:14]; + [encoder setBytes:&p0 length:sizeof(p0) atIndex:15]; + [encoder setBytes:&p1 length:sizeof(p1) atIndex:16]; + + const int nth = MIN(1024, ne0); + [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; } break; case GGML_OP_ARANGE: @@ -2875,6 +3362,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARANGE_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_dst offset:offs_dst atIndex:0]; [encoder setBytes:&ne0 length:sizeof(ne0) atIndex:1]; @@ -2896,6 +3384,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2934,6 +3423,7 @@ static void ggml_metal_encode_node( default: GGML_ABORT("fatal error"); }; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -2952,6 +3442,7 @@ static void ggml_metal_encode_node( id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_LEAKY_RELU_F32].pipeline; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -3014,6 +3505,8 @@ static void ggml_metal_encode_node( bool use_vec_kernel = false; + // TODO: add vec kernels for (ne00%64 == 0) and maybe also for (ne00%32 == 0) + // for now avoiding mainly to keep the number of templates/kernels a bit lower if (ne01 >= 4 || (ne00%128 != 0)) { switch (src1->type) { case GGML_TYPE_F16: @@ -3033,6 +3526,23 @@ static void ggml_metal_encode_node( } } } break; + case GGML_TYPE_BF16: + { + switch (ne00) { + case 64: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H64 ].pipeline; break; + case 80: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H80 ].pipeline; break; + case 96: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H96 ].pipeline; break; + case 112: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H112].pipeline; break; + case 128: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H128].pipeline; break; + case 256: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_BF16_H256].pipeline; break; + default: + { + GGML_LOG_ERROR("unsupported size: %lld\n", ne00); + GGML_LOG_ERROR("add template specialization for this size\n"); + GGML_ABORT("add template specialization for this size"); + } + } + } break; case GGML_TYPE_Q4_0: { switch (ne00) { @@ -3133,6 +3643,7 @@ static void ggml_metal_encode_node( { switch (src1->type) { case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H128].pipeline; break; + case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H128].pipeline; break; case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H128].pipeline; break; case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H128].pipeline; break; case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H128].pipeline; break; @@ -3150,6 +3661,7 @@ static void ggml_metal_encode_node( { switch (src1->type) { case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_F16_H256].pipeline; break; + case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_BF16_H256].pipeline; break; case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_0_H256].pipeline; break; case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q4_1_H256].pipeline; break; case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_FLASH_ATTN_EXT_VEC_Q5_0_H256].pipeline; break; @@ -3172,40 +3684,41 @@ static void ggml_metal_encode_node( } } + ggml_metal_kargs_flash_attn_ext args = { + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, + /*.ne_12_2 =*/ ne12, + /*.ne_12_3 =*/ ne13, + /*.nb_12_1 =*/ nb11, + /*.nb_12_2 =*/ nb12, + /*.nb_12_3 =*/ nb13, + /*.nb31 =*/ nb31, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.scale =*/ scale, + /*.max_bias =*/ max_bias, + /*.m0 =*/ m0, + /*.m1 =*/ m1, + /*.n_head_log2 =*/ n_head_log2, + /*.logit_softcap =*/ logit_softcap, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; if (id_src3) { - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; + [encoder setBuffer:id_src3 offset:offs_src3 atIndex:4]; } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:3]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:4]; } - [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; - [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:5]; - [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:6]; - [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:7]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:8]; - [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:9]; - [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:10]; - [encoder setBytes:&ne11 length:sizeof( int64_t) atIndex:11]; - [encoder setBytes:&ne12 length:sizeof( int64_t) atIndex:12]; - [encoder setBytes:&ne13 length:sizeof( int64_t) atIndex:13]; - [encoder setBytes:&nb11 length:sizeof(uint64_t) atIndex:14]; - [encoder setBytes:&nb12 length:sizeof(uint64_t) atIndex:15]; - [encoder setBytes:&nb13 length:sizeof(uint64_t) atIndex:16]; - [encoder setBytes:&nb21 length:sizeof(uint64_t) atIndex:17]; - [encoder setBytes:&nb22 length:sizeof(uint64_t) atIndex:18]; - [encoder setBytes:&nb23 length:sizeof(uint64_t) atIndex:19]; - [encoder setBytes:&nb31 length:sizeof(uint64_t) atIndex:20]; - [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:21]; - [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:22]; - [encoder setBytes:&scale length:sizeof( float) atIndex:23]; - [encoder setBytes:&max_bias length:sizeof( float) atIndex:24]; - [encoder setBytes:&m0 length:sizeof(m0) atIndex:25]; - [encoder setBytes:&m1 length:sizeof(m1) atIndex:26]; - [encoder setBytes:&n_head_log2 length:sizeof(n_head_log2) atIndex:27]; - [encoder setBytes:&logit_softcap length:sizeof(logit_softcap) atIndex:28]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:5]; if (!use_vec_kernel) { // half8x8 kernel @@ -3216,11 +3729,14 @@ static void ggml_metal_encode_node( GGML_ASSERT(nqptg % 8 == 0); GGML_ASSERT(ncpsg % 32 == 0); + // 2*(2*ncpsg + nqptg)*(nsg) + // ncpsg soft_max values + ncpsg mask values + a diagonal scaling matrix (in float) + // // 16*32*(nsg) // the shared memory needed for the simdgroups to load the KV cache // each thread loads (dequantizes) 16 head elements, there are 32 threads in th SG // -#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*(ncpsg + nqptg)*(nsg)) + 16*32*(nsg))*(sizeof(float)/2), 16)) +#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*(2*ncpsg + nqptg)*(nsg)) + 16*32*(nsg))*(sizeof(float)/2), 16)) int64_t nsgmax = 2; @@ -3254,12 +3770,12 @@ static void ggml_metal_encode_node( // ne00 + 2*ncpsg*(nsg) // for each query, we load it as f16 in shared memory (ne00) - // and store the attention scores (nqptg x ncpsg) as f32 + // and store the soft_max values and the mask // - // 2*ne00*(nsg) - // each simdgroup has a full f32 head vector in shared mem to accumulate results + // ne00*(nsg) + // each simdgroup has a full f16 head vector in shared mem to accumulate results // -#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*ncpsg*(nsg)) + 2*ne00*(nsg))*(sizeof(float)/2), 16)) +#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*ncpsg*(nsg)) + ne00*(nsg))*(sizeof(float)/2), 16)) int64_t nsgmax = 2; @@ -3337,28 +3853,94 @@ static void ggml_metal_encode_node( default: GGML_ABORT("not implemented"); } + ggml_metal_kargs_cpy args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; - [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:3]; - [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:4]; - [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:5]; - [encoder setBytes:&nb00 length:sizeof(uint64_t) atIndex:6]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:7]; - [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:8]; - [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:9]; - [encoder setBytes:&ne0 length:sizeof( int64_t) atIndex:10]; - [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:11]; - [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:12]; - [encoder setBytes:&ne3 length:sizeof( int64_t) atIndex:13]; - [encoder setBytes:&nb0 length:sizeof(uint64_t) atIndex:14]; - [encoder setBytes:&nb1 length:sizeof(uint64_t) atIndex:15]; - [encoder setBytes:&nb2 length:sizeof(uint64_t) atIndex:16]; - [encoder setBytes:&nb3 length:sizeof(uint64_t) atIndex:17]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; } break; + case GGML_OP_SET: + { + GGML_ASSERT(ggml_are_same_shape(src0, dst)); + GGML_ASSERT(ggml_is_contiguous(dst) && ggml_is_contiguous(src0)); + + // src0 and dst as viewed during set + const size_t dst_nb0 = ggml_element_size(src0); + + const size_t dst_nb1 = ((int32_t *) dst->op_params)[0]; + const size_t dst_nb2 = ((int32_t *) dst->op_params)[1]; + const size_t dst_nb3 = ((int32_t *) dst->op_params)[2]; + const size_t offset = ((int32_t *) dst->op_params)[3]; + const bool inplace = (bool) ((int32_t *) dst->op_params)[4]; + + if (!inplace) { + memcpy(((char *) dst->data), ((char *) src0->data), ggml_nbytes(dst)); + } + + const int im0 = (ne10 == 0 ? 0 : ne10-1); + const int im1 = (ne11 == 0 ? 0 : ne11-1); + const int im2 = (ne12 == 0 ? 0 : ne12-1); + const int im3 = (ne13 == 0 ? 0 : ne13-1); + + GGML_ASSERT(offset + im0*dst_nb0 + im1*dst_nb1 + im2*dst_nb2 + im3*dst_nb3 <= ggml_nbytes(dst)); + + id pipeline = nil; + + switch (src0t) { + case GGML_TYPE_F32: + GGML_ASSERT(nb10 == sizeof(float)); + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_F32].pipeline; break; + case GGML_TYPE_I32: + GGML_ASSERT(nb10 == sizeof(int32_t)); + pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_I32].pipeline; break; + default: GGML_ABORT("fatal error"); + } + + ggml_metal_kargs_set args = { + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.nb1 =*/ dst_nb1, + /*.nb2 =*/ dst_nb2, + /*.nb3 =*/ dst_nb3, + /*.offs =*/ offset, + /*.inplace =*/ inplace, + }; + + const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne10); + + [encoder setComputePipelineState:pipeline]; + [encoder setBytes:&args length:sizeof(args) atIndex:0]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; + + [encoder dispatchThreadgroups:MTLSizeMake(ne11, ne12, ne13) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + } break; case GGML_OP_POOL_2D: { GGML_ASSERT(ggml_is_contiguous(src0)); @@ -3400,6 +3982,7 @@ static void ggml_metal_encode_node( const int64_t n_threads = MIN((int64_t)[pipeline maxTotalThreadsPerThreadgroup], parallel_elements); const int64_t n_tg = (parallel_elements + n_threads - 1) / n_threads; + // TODO: add ggml_metal_kargs struct [encoder setComputePipelineState:pipeline]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; @@ -3417,6 +4000,31 @@ static void ggml_metal_encode_node( [encoder dispatchThreadgroups:MTLSizeMake(n_tg, 1, 1) threadsPerThreadgroup:MTLSizeMake(n_threads, 1, 1)]; } break; + case GGML_OP_ARGMAX: + { + GGML_ASSERT(src0->type == GGML_TYPE_F32); + GGML_ASSERT(ggml_is_contiguous_1(src0)); + GGML_ASSERT(nb00 == ggml_type_size(src0->type)); + + const int64_t nrows = ggml_nrows(src0); + + int nth = 32; // SIMD width + while (nth < ne00 && nth*ne01*ne02*ne03 < 256) { + nth *= 2; + } + + id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARGMAX].pipeline; + + [encoder setComputePipelineState:pipeline]; + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; + [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; + [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:3]; + [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; + [encoder setThreadgroupMemoryLength:32*sizeof(int32_t) atIndex:1]; + + [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + } break; default: { GGML_LOG_ERROR("%s: error: node %3d, op = %8s not implemented\n", __func__, idx, ggml_op_name(dst->op)); @@ -3587,6 +4195,12 @@ static void * ggml_backend_metal_buffer_get_base(ggml_backend_buffer_t buffer) { return ctx->all_data; } +static void ggml_backend_metal_buffer_memset_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { + memset((char *)tensor->data + offset, value, size); + + UNUSED(buffer); +} + static void ggml_backend_metal_buffer_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { memcpy((char *)tensor->data + offset, data, size); @@ -3619,7 +4233,7 @@ static struct ggml_backend_buffer_i ggml_backend_metal_buffer_i = { /* .free_buffer = */ ggml_backend_metal_buffer_free_buffer, /* .get_base = */ ggml_backend_metal_buffer_get_base, /* .init_tensor = */ NULL, - /* .memset_tensor = */ NULL, + /* .memset_tensor = */ ggml_backend_metal_buffer_memset_tensor, /* .set_tensor = */ ggml_backend_metal_buffer_set_tensor, /* .get_tensor = */ ggml_backend_metal_buffer_get_tensor, /* .cpy_tensor = */ ggml_backend_metal_buffer_cpy_tensor, @@ -4216,19 +4830,45 @@ static ggml_backend_dev_t ggml_backend_metal_reg_device_get(ggml_backend_reg_t r GGML_UNUSED(index); } +static struct ggml_backend_feature g_ggml_backend_metal_features[] = { +#if defined(GGML_METAL_EMBED_LIBRARY) + { "EMBED_LIBRARY", "1" }, +#endif +#if defined(GGML_METAL_USE_BF16) + { "BF16", "1" }, +#endif + { nil, nil }, +}; + +static struct ggml_backend_feature * ggml_backend_metal_get_features(ggml_backend_reg_t reg) { + return g_ggml_backend_metal_features; + + GGML_UNUSED(reg); +} + +static void * ggml_backend_metal_get_proc_address(ggml_backend_reg_t reg, const char * name) { + if (strcmp(name, "ggml_backend_get_features") == 0) { + return (void *)ggml_backend_metal_get_features; + } + + return NULL; + + GGML_UNUSED(reg); +} static struct ggml_backend_reg_i ggml_backend_metal_reg_i = { /* .get_name = */ ggml_backend_metal_reg_get_name, /* .device_count = */ ggml_backend_metal_reg_device_count, /* .device_get = */ ggml_backend_metal_reg_device_get, - /* .get_proc_address = */ NULL, + /* .get_proc_address = */ ggml_backend_metal_get_proc_address, }; ggml_backend_reg_t ggml_backend_metal_reg(void) { // TODO: make this thread-safe somehow? { g_ggml_backend_metal_reg = (struct ggml_backend_reg) { - /* .iface = */ ggml_backend_metal_reg_i, - /* .context = */ NULL, + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_metal_reg_i, + /* .context = */ NULL, }; g_ggml_backend_metal_device = (struct ggml_backend_device) { @@ -4240,3 +4880,5 @@ ggml_backend_reg_t ggml_backend_metal_reg(void) { return &g_ggml_backend_metal_reg; } + +GGML_BACKEND_DL_IMPL(ggml_backend_metal_reg) diff --git a/ggml/src/ggml-metal.metal b/ggml/src/ggml-metal/ggml-metal.metal similarity index 56% rename from ggml/src/ggml-metal.metal rename to ggml/src/ggml-metal/ggml-metal.metal index 16b5da3ff..8ba43904d 100644 --- a/ggml/src/ggml-metal.metal +++ b/ggml/src/ggml-metal/ggml-metal.metal @@ -1,6 +1,12 @@ #define GGML_COMMON_DECL_METAL #define GGML_COMMON_IMPL_METAL -#include "ggml-common.h" +#if defined(GGML_METAL_EMBED_LIBRARY) +__embed_ggml-common.h__ +#else +// TODO: this should not be a relative path, but can't figure out how to set Metal include paths in Package.swift +#include "../ggml-common.h" +#endif +#include "ggml-metal-impl.h" #include @@ -15,14 +21,14 @@ using namespace metal; // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf // // cmd: -// .../usr/bin/metal -dM -E -c ggml/src/ggml-metal.metal -// .../usr/bin/metal -dM -E -c -target air64-apple-ios14.0 ggml/src/ggml-metal.metal +// .../usr/bin/metal -dM -E -c ggml/src/ggml-metal/ggml-metal.metal +// .../usr/bin/metal -dM -E -c -target air64-apple-ios14.0 ggml/src/ggml-metal/ggml-metal.metal // -#if __METAL_VERSION__ < 310 -#define GGML_METAL_NO_BFLOAT +#if __METAL_VERSION__ < 310 && defined(GGML_METAL_USE_BF16) +#undef GGML_METAL_USE_BF16 #endif -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) typedef matrix bfloat4x4; #endif @@ -41,7 +47,12 @@ void dequantize_f16(device const half4x4 * src, short il, thread type4x4 & reg) reg = (type4x4)(*src); } -#if !defined(GGML_METAL_NO_BFLOAT) +template +void dequantize_f16_t4(device const half4 * src, short il, thread type4 & reg) { + reg = (type4)(*(src + il)); +} + +#if defined(GGML_METAL_USE_BF16) template void dequantize_bf16(device const bfloat4x4 * src, short il, thread type4x4 & reg) { reg = (type4x4)(*src); @@ -49,7 +60,7 @@ void dequantize_bf16(device const bfloat4x4 * src, short il, thread type4x4 & re #endif template -void dequantize_q4_0(device const block_q4_0 *xb, short il, thread type4x4 & reg) { +void dequantize_q4_0(device const block_q4_0 * xb, short il, thread type4x4 & reg) { device const uint16_t * qs = ((device const uint16_t *)xb + 1); const float d1 = il ? (xb->d / 16.h) : xb->d; const float d2 = d1 / 256.f; @@ -57,14 +68,33 @@ void dequantize_q4_0(device const block_q4_0 *xb, short il, thread type4x4 & reg const ushort mask0 = il ? 0x00F0 : 0x000F; const ushort mask1 = mask0 << 8; - for (int i=0;i<8;i++) { - reg[i/2][2*(i%2)+0] = d1 * (qs[i] & mask0) + md; - reg[i/2][2*(i%2)+1] = d2 * (qs[i] & mask1) + md; + float4x4 reg_f; + + for (int i = 0; i < 8; i++) { + reg_f[i/2][2*(i%2) + 0] = d1 * (qs[i] & mask0) + md; + reg_f[i/2][2*(i%2) + 1] = d2 * (qs[i] & mask1) + md; + } + + reg = (type4x4) reg_f; +} + +template +void dequantize_q4_0_t4(device const block_q4_0 * xb, short il, thread type4 & reg) { + device const uint16_t * qs = ((device const uint16_t *)xb + 1); + const float d1 = (il/4) ? (xb->d / 16.h) : xb->d; + const float d2 = d1 / 256.f; + const float md = -8.h * xb->d; + const ushort mask0 = (il/4) ? 0x00F0 : 0x000F; + const ushort mask1 = mask0 << 8; + + for (int i = 0; i < 2; i++) { + reg[2*i + 0] = d1 * (qs[2*(il%4) + i] & mask0) + md; + reg[2*i + 1] = d2 * (qs[2*(il%4) + i] & mask1) + md; } } template -void dequantize_q4_1(device const block_q4_1 *xb, short il, thread type4x4 & reg) { +void dequantize_q4_1(device const block_q4_1 * xb, short il, thread type4x4 & reg) { device const uint16_t * qs = ((device const uint16_t *)xb + 2); const float d1 = il ? (xb->d / 16.h) : xb->d; const float d2 = d1 / 256.f; @@ -72,14 +102,33 @@ void dequantize_q4_1(device const block_q4_1 *xb, short il, thread type4x4 & reg const ushort mask0 = il ? 0x00F0 : 0x000F; const ushort mask1 = mask0 << 8; - for (int i=0;i<8;i++) { - reg[i/2][2*(i%2)+0] = ((qs[i] & mask0) * d1) + m; - reg[i/2][2*(i%2)+1] = ((qs[i] & mask1) * d2) + m; + float4x4 reg_f; + + for (int i = 0; i < 8; i++) { + reg_f[i/2][2*(i%2) + 0] = ((qs[i] & mask0) * d1) + m; + reg_f[i/2][2*(i%2) + 1] = ((qs[i] & mask1) * d2) + m; + } + + reg = (type4x4) reg_f; +} + +template +void dequantize_q4_1_t4(device const block_q4_1 * xb, short il, thread type4 & reg) { + device const uint16_t * qs = ((device const uint16_t *)xb + 2); + const float d1 = (il/4) ? (xb->d / 16.h) : xb->d; + const float d2 = d1 / 256.f; + const float m = xb->m; + const ushort mask0 = (il/4) ? 0x00F0 : 0x000F; + const ushort mask1 = mask0 << 8; + + for (int i = 0; i < 2; i++) { + reg[2*i + 0] = d1 * (qs[2*(il%4) + i] & mask0) + m; + reg[2*i + 1] = d2 * (qs[2*(il%4) + i] & mask1) + m; } } template -void dequantize_q5_0(device const block_q5_0 *xb, short il, thread type4x4 & reg) { +void dequantize_q5_0(device const block_q5_0 * xb, short il, thread type4x4 & reg) { device const uint16_t * qs = ((device const uint16_t *)xb + 3); const float d = xb->d; const float md = -16.h * xb->d; @@ -92,6 +141,8 @@ void dequantize_q5_0(device const block_q5_0 *xb, short il, thread type4x4 & reg const int gh_mv = il ? 12 : 0; const int gh_bk = il ? 0 : 4; + float4x4 reg_f; + for (int i = 0; i < 8; i++) { // extract the 5-th bits for x0 and x1 const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10; @@ -101,13 +152,45 @@ void dequantize_q5_0(device const block_q5_0 *xb, short il, thread type4x4 & reg const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0); const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1); - reg[i/2][2*(i%2)+0] = d * x0 + md; - reg[i/2][2*(i%2)+1] = d * x1 + md; + reg_f[i/2][2*(i%2) + 0] = d * x0 + md; + reg_f[i/2][2*(i%2) + 1] = d * x1 + md; + } + + reg = (type4x4) reg_f; +} + +template +void dequantize_q5_0_t4(device const block_q5_0 * xb, short il, thread type4 & reg) { + device const uint16_t * qs = ((device const uint16_t *)xb + 3); + const float d = xb->d; + const float md = -16.h * xb->d; + const ushort mask = (il/4) ? 0x00F0 : 0x000F; + + const uint32_t qh = *((device const uint32_t *)xb->qh); + + const int x_mv = (il/4) ? 4 : 0; + + const int gh_mv = (il/4) ? 12 : 0; + const int gh_bk = (il/4) ? 0 : 4; + + for (int ii = 0; ii < 2; ii++) { + int i = 2*(il%4) + ii; + + // extract the 5-th bits for x0 and x1 + const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10; + const uint8_t xh_1 = ((qh >> (gh_mv + 2*i+1)) << gh_bk) & 0x10; + + // combine the 4-bits from qs with the 5th bit + const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0); + const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1); + + reg[2*ii + 0] = d * x0 + md; + reg[2*ii + 1] = d * x1 + md; } } template -void dequantize_q5_1(device const block_q5_1 *xb, short il, thread type4x4 & reg) { +void dequantize_q5_1(device const block_q5_1 * xb, short il, thread type4x4 & reg) { device const uint16_t * qs = ((device const uint16_t *)xb + 4); const float d = xb->d; const float m = xb->m; @@ -120,6 +203,8 @@ void dequantize_q5_1(device const block_q5_1 *xb, short il, thread type4x4 & reg const int gh_mv = il ? 12 : 0; const int gh_bk = il ? 0 : 4; + float4x4 reg_f; + for (int i = 0; i < 8; i++) { // extract the 5-th bits for x0 and x1 const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10; @@ -129,18 +214,64 @@ void dequantize_q5_1(device const block_q5_1 *xb, short il, thread type4x4 & reg const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0); const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1); - reg[i/2][2*(i%2)+0] = d * x0 + m; - reg[i/2][2*(i%2)+1] = d * x1 + m; + reg_f[i/2][2*(i%2) + 0] = d * x0 + m; + reg_f[i/2][2*(i%2) + 1] = d * x1 + m; + } + + reg = (type4x4) reg_f; +} + +template +void dequantize_q5_1_t4(device const block_q5_1 * xb, short il, thread type4 & reg) { + device const uint16_t * qs = ((device const uint16_t *)xb + 4); + const float d = xb->d; + const float m = xb->m; + const ushort mask = (il/4) ? 0x00F0 : 0x000F; + + const uint32_t qh = *((device const uint32_t *)xb->qh); + + const int x_mv = (il/4) ? 4 : 0; + + const int gh_mv = (il/4) ? 12 : 0; + const int gh_bk = (il/4) ? 0 : 4; + + for (int ii = 0; ii < 2; ii++) { + int i = 2*(il%4) + ii; + + // extract the 5-th bits for x0 and x1 + const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10; + const uint8_t xh_1 = ((qh >> (gh_mv + 2*i+1)) << gh_bk) & 0x10; + + // combine the 4-bits from qs with the 5th bit + const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0); + const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1); + + reg[2*ii + 0] = d * x0 + m; + reg[2*ii + 1] = d * x1 + m; } } template void dequantize_q8_0(device const block_q8_0 *xb, short il, thread type4x4 & reg) { device const int8_t * qs = ((device const int8_t *)xb->qs); - const half d = xb->d; + const float d = xb->d; + + float4x4 reg_f; for (int i = 0; i < 16; i++) { - reg[i/4][i%4] = (qs[i + 16*il] * d); + reg_f[i/4][i%4] = (qs[i + 16*il] * d); + } + + reg = (type4x4) reg_f; +} + +template +void dequantize_q8_0_t4(device const block_q8_0 *xb, short il, thread type4 & reg) { + device const int8_t * qs = ((device const int8_t *)xb->qs); + const float d = xb->d; + + for (int i = 0; i < 4; i++) { + reg[i] = (qs[4*(il%4) + i + 16*(il/4)] * d); } } @@ -198,7 +329,7 @@ static inline uchar2 get_scale_min_k4_just2(int j, int k, device const uchar * q } template -void dequantize_q4_K(device const block_q4_K *xb, short il, thread type4x4 & reg) { +void dequantize_q4_K(device const block_q4_K * xb, short il, thread type4x4 & reg) { device const uchar * q = xb->qs; short is = (il/4) * 2; @@ -210,7 +341,7 @@ void dequantize_q4_K(device const block_q4_K *xb, short il, thread type4x4 & reg const float dl = d * sc[0]; const float ml = min * sc[1]; - const ushort mask = il<2 ? 0x0F : 0xF0; + const ushort mask = il < 2 ? 0x0F : 0xF0; for (int i = 0; i < 16; ++i) { reg[i/4][i%4] = dl * (q[i] & mask) - ml; } @@ -443,6 +574,19 @@ void dequantize_iq4_nl(device const block_iq4_nl * xb, short il, thread type4x4 } } +template +void dequantize_iq4_nl_t4(device const block_iq4_nl * xb, short il, thread type4 & reg) { + device const uint16_t * q4 = (device const uint16_t *)xb->qs; + const float d = xb->d; + uint32_t aux32; + thread const uint8_t * q8 = (thread const uint8_t *)&aux32; + aux32 = ((q4[2*(il%4)] | (q4[2*(il%4)+1] << 16)) >> 4*(il/4)) & 0x0f0f0f0f; + reg[0] = d * kvalues_iq4nl_f[q8[0]]; + reg[1] = d * kvalues_iq4nl_f[q8[1]]; + reg[2] = d * kvalues_iq4nl_f[q8[2]]; + reg[3] = d * kvalues_iq4nl_f[q8[3]]; +} + template void dequantize_iq4_xs(device const block_iq4_xs * xb, short il, thread type4x4 & reg) { // il is 0...15 for QK_K = 256 => index of block of 32 is il/2 @@ -472,240 +616,131 @@ enum ggml_sort_order { // pros: works for non-contiguous tensors, supports broadcast across all dims // cons: not very efficient kernel void kernel_add( + constant ggml_metal_kargs_bin & args, device const char * src0, device const char * src1, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - constant int64_t & offs, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig.z; - const int64_t i02 = tgpig.y; - const int64_t i01 = tgpig.x; + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig.z; + const int i02 = tgpig.y; + const int i01 = tgpig.x; - const int64_t i13 = i03 % ne13; - const int64_t i12 = i02 % ne12; - const int64_t i11 = i01 % ne11; + const int i13 = i03%args.ne13; + const int i12 = i02%args.ne12; + const int i11 = i01%args.ne11; - device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01 + offs; - device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; - device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1 + offs; + device const char * src0_ptr = src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + args.offs; + device const char * src1_ptr = src1 + i13*args.nb13 + i12*args.nb12 + i11*args.nb11; + device char * dst_ptr = dst + i03*args.nb3 + i02*args.nb2 + i01*args.nb1 + args.offs; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - const int i10 = i0 % ne10; - *((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) + *((device float *)(src1_ptr + i10*nb10)); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + const int i10 = i0%args.ne10; + *((device float *)(dst_ptr + i0*args.nb0)) = *((device float *)(src0_ptr + i0*args.nb00)) + *((device float *)(src1_ptr + i10*args.nb10)); } } kernel void kernel_sub( + constant ggml_metal_kargs_bin & args, device const char * src0, device const char * src1, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - constant int64_t & offs, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig.z; - const int64_t i02 = tgpig.y; - const int64_t i01 = tgpig.x; + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig.z; + const int i02 = tgpig.y; + const int i01 = tgpig.x; - const int64_t i13 = i03 % ne13; - const int64_t i12 = i02 % ne12; - const int64_t i11 = i01 % ne11; + const int i13 = i03%args.ne13; + const int i12 = i02%args.ne12; + const int i11 = i01%args.ne11; - device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01 + offs; - device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; - device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1 + offs; + device const char * src0_ptr = src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + args.offs; + device const char * src1_ptr = src1 + i13*args.nb13 + i12*args.nb12 + i11*args.nb11; + device char * dst_ptr = dst + i03*args.nb3 + i02*args.nb2 + i01*args.nb1 + args.offs; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - const int i10 = i0 % ne10; - *((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) - *((device float *)(src1_ptr + i10*nb10)); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + const int i10 = i0%args.ne10; + *((device float *)(dst_ptr + i0*args.nb0)) = *((device float *)(src0_ptr + i0*args.nb00)) - *((device float *)(src1_ptr + i10*args.nb10)); } } kernel void kernel_mul( + constant ggml_metal_kargs_bin & args, device const char * src0, device const char * src1, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig.z; - const int64_t i02 = tgpig.y; - const int64_t i01 = tgpig.x; + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig.z; + const int i02 = tgpig.y; + const int i01 = tgpig.x; - const int64_t i13 = i03 % ne13; - const int64_t i12 = i02 % ne12; - const int64_t i11 = i01 % ne11; + const int i13 = i03%args.ne13; + const int i12 = i02%args.ne12; + const int i11 = i01%args.ne11; - device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01; - device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; - device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1; + device const char * src0_ptr = src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01; + device const char * src1_ptr = src1 + i13*args.nb13 + i12*args.nb12 + i11*args.nb11; + device char * dst_ptr = dst + i03*args.nb3 + i02*args.nb2 + i01*args.nb1; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - const int i10 = i0 % ne10; - *((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) * *((device float *)(src1_ptr + i10*nb10)); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + const int i10 = i0%args.ne10; + *((device float *)(dst_ptr + i0*args.nb0)) = *((device float *)(src0_ptr + i0*args.nb00)) * *((device float *)(src1_ptr + i10*args.nb10)); } } kernel void kernel_div( + constant ggml_metal_kargs_bin & args, device const char * src0, device const char * src1, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig.z; - const int64_t i02 = tgpig.y; - const int64_t i01 = tgpig.x; + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig.z; + const int i02 = tgpig.y; + const int i01 = tgpig.x; - const int64_t i13 = i03 % ne13; - const int64_t i12 = i02 % ne12; - const int64_t i11 = i01 % ne11; + const int i13 = i03%args.ne13; + const int i12 = i02%args.ne12; + const int i11 = i01%args.ne11; - device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01; - device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; - device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1; + device const char * src0_ptr = src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01; + device const char * src1_ptr = src1 + i13*args.nb13 + i12*args.nb12 + i11*args.nb11; + device char * dst_ptr = dst + i03*args.nb3 + i02*args.nb2 + i01*args.nb1; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - const int i10 = i0 % ne10; - *((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) / *((device float *)(src1_ptr + i10*nb10)); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + const int i10 = i0%args.ne10; + *((device float *)(dst_ptr + i0*args.nb0)) = *((device float *)(src0_ptr + i0*args.nb00)) / *((device float *)(src1_ptr + i10*args.nb10)); } } template kernel void kernel_repeat( + constant ggml_metal_kargs_repeat & args, device const char * src0, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i3 = tgpig.z; - const int64_t i2 = tgpig.y; - const int64_t i1 = tgpig.x; + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i3 = tgpig.z; + const int i2 = tgpig.y; + const int i1 = tgpig.x; - const int64_t i03 = i3 % ne03; - const int64_t i02 = i2 % ne02; - const int64_t i01 = i1 % ne01; + const int i03 = i3%args.ne03; + const int i02 = i2%args.ne02; + const int i01 = i1%args.ne01; - device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01; - device char * dst_ptr = dst + i3*nb3 + i2*nb2 + i1*nb1 ; + device const char * src0_ptr = src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01; + device char * dst_ptr = dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - const int i00 = i0 % ne00; - *((device T *)(dst_ptr + i0*nb0)) = *((device T *)(src0_ptr + i00*nb00)); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + const int i00 = i0%args.ne00; + *((device T *)(dst_ptr + i0*args.nb0)) = *((device T *)(src0_ptr + i00*args.nb00)); } } @@ -719,38 +754,42 @@ template [[host_name("kernel_repeat_i16")]] kernel kernel_repeat_t kernel_repeat // assumption: src1 is a row // broadcast src1 into src0 kernel void kernel_add_row( + constant ggml_metal_kargs_bin & args, device const float4 * src0, device const float4 * src1, device float4 * dst, - constant uint64_t & nb [[buffer(28)]], uint tpig[[thread_position_in_grid]]) { + const uint nb = args.ne00/4; dst[tpig] = src0[tpig] + src1[tpig % nb]; } kernel void kernel_sub_row( + constant ggml_metal_kargs_bin & args, device const float4 * src0, device const float4 * src1, device float4 * dst, - constant uint64_t & nb [[buffer(28)]], uint tpig[[thread_position_in_grid]]) { + const uint nb = args.ne00/4; dst[tpig] = src0[tpig] - src1[tpig % nb]; } kernel void kernel_mul_row( + constant ggml_metal_kargs_bin & args, device const float4 * src0, device const float4 * src1, device float4 * dst, - constant uint64_t & nb [[buffer(28)]], uint tpig[[thread_position_in_grid]]) { + const uint nb = args.ne00/4; dst[tpig] = src0[tpig] * src1[tpig % nb]; } kernel void kernel_div_row( + constant ggml_metal_kargs_bin & args, device const float4 * src0, device const float4 * src1, device float4 * dst, - constant uint64_t & nb [[buffer(28)]], uint tpig[[thread_position_in_grid]]) { + const uint nb = args.ne00/4; dst[tpig] = src0[tpig] / src1[tpig % nb]; } @@ -861,6 +900,14 @@ kernel void kernel_silu_4( dst[tpig] = x / (1.0f + exp(-x)); } +kernel void kernel_elu( + device const float * src0, + device float * dst, + uint tpig[[thread_position_in_grid]]) { + device const float & x = src0[tpig]; + dst[tpig] = (x > 0.0f) ? x : (exp(x) - 1.0f); +} + kernel void kernel_sqr( device const float * src0, device float * dst, @@ -1319,103 +1366,170 @@ kernel void kernel_ssm_scan_f32( } } -kernel void kernel_norm( - device const void * src0, - device float * dst, - constant int64_t & ne00, - constant uint64_t & nb01, - constant float & eps, - threadgroup float * sum [[threadgroup(0)]], - uint tgpig[[threadgroup_position_in_grid]], - uint tpitg[[thread_position_in_threadgroup]], - uint ntg[[threads_per_threadgroup]]) { - device const float * x = (device const float *) ((device const char *) src0 + tgpig*nb01); - // MEAN - // parallel sum - sum[tpitg] = 0.0f; - for (int i00 = tpitg; i00 < ne00; i00 += ntg) { - sum[tpitg] += x[i00]; - } - // reduce - threadgroup_barrier(mem_flags::mem_threadgroup); - for (uint i = ntg/2; i > 0; i /= 2) { - if (tpitg < i) { - sum[tpitg] += sum[tpitg + i]; +kernel void kernel_argmax( + device const void * x, + device int32_t * dst, + constant int64_t & ncols, + constant uint64_t & nb01, + threadgroup float * shared_maxval [[threadgroup(0)]], + threadgroup int32_t * shared_argmax [[threadgroup(1)]], + uint tgpig[[threadgroup_position_in_grid]], + uint tpitg[[thread_position_in_threadgroup]], + uint sgitg[[simdgroup_index_in_threadgroup]], + uint tiisg[[thread_index_in_simdgroup]], + uint ntg[[threads_per_threadgroup]]) { + device const float * x_row = (device const float *) ((device const char *) x + tgpig * nb01); + + float lmax = -INFINITY; + int32_t larg = -1; + + for (int i00 = tpitg; i00 < ncols; i00 += ntg) { + if (x_row[i00] > lmax) { + lmax = x_row[i00]; + larg = i00; } - threadgroup_barrier(mem_flags::mem_threadgroup); - } - const float mean = sum[0] / ne00; - - // recenter and VARIANCE - threadgroup_barrier(mem_flags::mem_threadgroup); - device float * y = dst + tgpig*ne00; - sum[tpitg] = 0.0f; - for (int i00 = tpitg; i00 < ne00; i00 += ntg) { - y[i00] = x[i00] - mean; - sum[tpitg] += y[i00] * y[i00]; } - // reduce - threadgroup_barrier(mem_flags::mem_threadgroup); - for (uint i = ntg/2; i > 0; i /= 2) { - if (tpitg < i) { - sum[tpitg] += sum[tpitg + i]; - } - threadgroup_barrier(mem_flags::mem_threadgroup); - } - const float variance = sum[0] / ne00; + // find the argmax value in the block + float max_val = simd_max(lmax); + int32_t arg_val = simd_max(select(-1, larg, lmax == max_val)); - const float scale = 1.0f/sqrt(variance + eps); - for (int i00 = tpitg; i00 < ne00; i00 += ntg) { - y[i00] = y[i00] * scale; - } -} - -kernel void kernel_rms_norm( - device const void * src0, - device float * dst, - constant int64_t & ne00, - constant uint64_t & nb01, - constant float & eps, - threadgroup float * buf [[threadgroup(0)]], - uint tgpig[[threadgroup_position_in_grid]], - uint tpitg[[thread_position_in_threadgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]], - uint tiisg[[thread_index_in_simdgroup]], - uint ntg[[threads_per_threadgroup]]) { - device const float4 * x = (device const float4 *) ((device const char *) src0 + tgpig*nb01); - - float4 sumf = 0; - float all_sum = 0; - - // parallel sum - for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) { - sumf += x[i00] * x[i00]; - } - all_sum = sumf[0] + sumf[1] + sumf[2] + sumf[3]; - all_sum = simd_sum(all_sum); if (ntg > N_SIMDWIDTH) { if (sgitg == 0) { - buf[tiisg] = 0.0f; + shared_maxval[tiisg] = -INFINITY; + shared_argmax[tiisg] = -1; } threadgroup_barrier(mem_flags::mem_threadgroup); if (tiisg == 0) { - buf[sgitg] = all_sum; + shared_maxval[sgitg] = max_val; + shared_argmax[sgitg] = arg_val; } threadgroup_barrier(mem_flags::mem_threadgroup); - all_sum = buf[tiisg]; - all_sum = simd_sum(all_sum); + max_val = shared_maxval[tiisg]; + arg_val = shared_argmax[tiisg]; + + float max_val_reduced = simd_max(max_val); + int32_t arg_val_reduced = simd_max(select(-1, arg_val, max_val == max_val_reduced)); + + dst[tgpig] = arg_val_reduced; + + return; } - const float mean = all_sum/ne00; - const float scale = 1.0f/sqrt(mean + eps); + dst[tgpig] = arg_val; +} - device float4 * y = (device float4 *) (dst + tgpig*ne00); - for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) { +kernel void kernel_norm( + constant ggml_metal_kargs_norm & args, + device const char * src0, + device char * dst, + threadgroup float * shmem_f32 [[threadgroup(0)]], + uint tgpig[[threadgroup_position_in_grid]], + ushort tpitg[[thread_position_in_threadgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort ntg[[threads_per_threadgroup]]) { + if (sgitg == 0) { + shmem_f32[tiisg] = 0.0f; + } + + device const float4 * x = (device const float4 *) (src0 + tgpig*args.nb01); + + float4 sumf4(0.0f); + + float sumf = 0.0f; + + for (int i00 = tpitg; i00 < args.ne00_4; i00 += ntg) { + sumf4 += x[i00]; + } + sumf = sumf4[0] + sumf4[1] + sumf4[2] + sumf4[3]; + sumf = simd_sum(sumf); + + threadgroup_barrier(mem_flags::mem_threadgroup); + + if (tiisg == 0) { + shmem_f32[sgitg] = sumf; + } + + threadgroup_barrier(mem_flags::mem_threadgroup); + + sumf = shmem_f32[tiisg]; + sumf = simd_sum(sumf); + + const float mean = sumf/args.ne00; + + device float4 * y = (device float4 *) dst + tgpig*args.ne00_4; + + sumf = 0.0f; + for (int i00 = tpitg; i00 < args.ne00_4; i00 += ntg) { + y[i00] = x[i00] - mean; + sumf += dot(y[i00], y[i00]); + } + sumf = simd_sum(sumf); + + threadgroup_barrier(mem_flags::mem_threadgroup); + + if (tiisg == 0) { + shmem_f32[sgitg] = sumf; + } + + threadgroup_barrier(mem_flags::mem_threadgroup); + + sumf = shmem_f32[tiisg]; + sumf = simd_sum(sumf); + + const float variance = sumf/args.ne00; + + const float scale = 1.0f/sqrt(variance + args.eps); + for (int i00 = tpitg; i00 < args.ne00_4; i00 += ntg) { + y[i00] = y[i00] * scale; + } +} + +kernel void kernel_rms_norm( + constant ggml_metal_kargs_rms_norm & args, + device const char * src0, + device char * dst, + threadgroup float * shmem_f32 [[threadgroup(0)]], + uint tgpig[[threadgroup_position_in_grid]], + ushort tpitg[[thread_position_in_threadgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort ntg[[threads_per_threadgroup]]) { + if (sgitg == 0) { + shmem_f32[tiisg] = 0.0f; + } + + device const float4 * x = (device const float4 *) (src0 + tgpig*args.nb01); + + float sumf = 0.0f; + + // parallel sum + for (int i00 = tpitg; i00 < args.ne00_4; i00 += ntg) { + sumf += dot(x[i00], x[i00]); + } + sumf = simd_sum(sumf); + + threadgroup_barrier(mem_flags::mem_threadgroup); + + if (tiisg == 0) { + shmem_f32[sgitg] = sumf; + } + + threadgroup_barrier(mem_flags::mem_threadgroup); + + sumf = shmem_f32[tiisg]; + sumf = simd_sum(sumf); + + const float mean = sumf/args.ne00; + const float scale = 1.0f/sqrt(mean + args.eps); + + device float4 * y = (device float4 *) dst + tgpig*args.ne00_4; + for (int i00 = tpitg; i00 < args.ne00_4; i00 += ntg) { y[i00] = x[i00] * scale; } } @@ -1603,31 +1717,17 @@ inline float block_q_n_dot_y(device const block_q5_1 * qb_curr, float sumy, thre // quantizations where the block size is 32. It also does not // guard against the number of rows not being divisible by // N_DST, so this is another explicit assumption of the implementation. -template +template void mul_vec_q_n_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { - const int nb = ne00/QK4_0; + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { + const int nb = args.ne00/QK4_0; const int r0 = tgpig.x; const int r1 = tgpig.y; @@ -1635,19 +1735,19 @@ void mul_vec_q_n_f32_impl( const int first_row = (r0 * nsg + sgitg) * nr; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - //const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + //const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - //device const block_q_type * x = (device const block_q_type *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + //device const block_q_type * x = (device const block_q_type *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); // pointers to src0 rows device const block_q_type * ax[nr]; for (int row = 0; row < nr; ++row) { - const uint offset0 = (first_row + row)*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + const uint64_t offset0 = (first_row + row)*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; ax[row] = (device const block_q_type *) ((device char *) src0 + offset0); } @@ -1655,10 +1755,10 @@ void mul_vec_q_n_f32_impl( float yl[16]; // src1 vector cache float sumf[nr] = {0.f}; - const int ix = (tiisg/2); - const int il = (tiisg%2)*8; + const short ix = (tiisg/2); + const short il = (tiisg%2)*8; - device const float * yb = y + ix * QK4_0 + il; + device const float * yb = y + ix*QK4_0 + il; // each thread in a SIMD group deals with half a block. for (int ib = ix; ib < nb; ib += nw/2) { @@ -1683,324 +1783,511 @@ void mul_vec_q_n_f32_impl( yb += QK4_0 * 16; } + device float * dst_f32 = (device float *) dst + im*args.ne0*args.ne1 + r1*args.ne0; + for (int row = 0; row < nr; ++row) { const float tot = simd_sum(sumf[row]); - if (tiisg == 0 && first_row + row < ne01) { - dst[im*ne0*ne1 + r1*ne0 + first_row + row] = tot; + + if (tiisg == 0 && first_row + row < args.ne01) { + dst_f32[first_row + row] = tot; } } } kernel void kernel_mul_mv_q4_0_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(src0,src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + mul_vec_q_n_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q4_1_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(src0,src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + mul_vec_q_n_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q5_0_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(src0,src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + mul_vec_q_n_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q5_1_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(src0,src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + mul_vec_q_n_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } - #define NB_Q8_0 8 +template void kernel_mul_mv_q8_0_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { const int nr = N_DST; const int nsg = N_SIMDGROUP; const int nw = N_SIMDWIDTH; - const int nb = ne00/QK8_0; + const int nb = args.ne00/QK8_0; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr; + const int first_row = (r0*nsg + sgitg)*nr; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - //const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + //const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - //device const block_q8_0 * x = (device const block_q8_0 *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + //device const block_q8_0 * x = (device const block_q8_0 *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); // pointers to src0 rows device const block_q8_0 * ax[nr]; for (int row = 0; row < nr; ++row) { - const uint offset0 = (first_row + row)*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + const uint64_t offset0 = (first_row + row)*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; ax[row] = (device const block_q8_0 *) ((device char *) src0 + offset0); } float yl[NB_Q8_0]; - float sumf[nr]={0.f}; + float sumf[nr] = { 0.f }; - const int ix = tiisg/4; - const int il = tiisg%4; + const short ix = tiisg/4; + const short il = tiisg%4; - device const float * yb = y + ix * QK8_0 + NB_Q8_0*il; + device const float * yb = y + ix*QK8_0 + il*NB_Q8_0; // each thread in a SIMD group deals with NB_Q8_0 quants at a time for (int ib = ix; ib < nb; ib += nw/4) { - for (int i = 0; i < NB_Q8_0; ++i) { + for (short i = 0; i < NB_Q8_0; ++i) { yl[i] = yb[i]; } for (int row = 0; row < nr; row++) { - device const int8_t * qs = ax[row][ib].qs + NB_Q8_0*il; + device const int8_t * qs = ax[row][ib].qs + il*NB_Q8_0; float sumq = 0.f; - for (int iq = 0; iq < NB_Q8_0; ++iq) { + for (short iq = 0; iq < NB_Q8_0; ++iq) { sumq += qs[iq] * yl[iq]; } sumf[row] += sumq*ax[row][ib].d; } - yb += NB_Q8_0 * nw; + yb += nw*NB_Q8_0; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < nr; ++row) { const float tot = simd_sum(sumf[row]); - if (tiisg == 0 && first_row + row < ne01) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot; + + if (tiisg == 0 && first_row + row < args.ne01) { + dst_f32[first_row + row] = tot; } } } [[host_name("kernel_mul_mv_q8_0_f32")]] kernel void kernel_mul_mv_q8_0_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q8_0_f32_impl(src0,src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + kernel_mul_mv_q8_0_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } +// mat-vec kernel processing in chunks of float4 +// chpb - chunks per quantization block +template +void kernel_mul_mv_ext_q4_f32_impl( + constant ggml_metal_kargs_mul_mv_ext & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + const short chpt = 4; // chunks per thread + + //const short nxpsg = (32); + const short nypsg = (32/nxpsg); + + const short tx = tiisg%nxpsg; + const short ty = tiisg/nxpsg; + + const int i01 = tgpig.x*(nypsg*args.nsg) + nypsg*sgitg + ty; + const int i11 = tgpig.y*r1ptg; + const int i1m = tgpig.z; + + const int i12 = i1m%args.ne12; + const int i13 = i1m/args.ne12; + + const uint64_t offset0 = i01*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = i11*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; + + device const q_t * xq = (i01 < args.ne01) ? (device const q_t *) (src0 + offset0) + tx/chpb : (device const q_t *) src0; + + device const float4 * y4[r1ptg]; + + for (int ir1 = 0; ir1 < r1ptg; ++ir1) { + y4[ir1] = (i11 + ir1 < args.ne11) ? (device const float4 *) (src1 + offset1 + ir1*args.nb11) + tx : (device const float4 *) src1; + } + + float sumf[r1ptg] = { [ 0 ... r1ptg - 1 ] = 0.0f }; + + short cch = tx%chpb; // current chunk index + + for (int ich = tx; 4*ich < args.ne00; ich += chpt*nxpsg) { + float4 lx[chpt]; + +#pragma unroll(chpt) + for (short ch = 0; ch < chpt; ++ch) { + deq_t4(xq, cch, lx[ch]); + + cch += nxpsg; + if (cch >= chpb) { + xq += cch/chpb; + cch %= chpb; + } + } + +#pragma unroll(chpt) + for (short ch = 0; ch < chpt; ++ch) { +#pragma unroll(r1ptg) + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + sumf[ir1] += dot(lx[ch], y4[ir1][ch*nxpsg]); + + } + } + +#pragma unroll(r1ptg) + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + y4[ir1] += chpt*nxpsg; + } + } + + // reduce only the threads in each row + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + if (nxpsg >= 32) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 16); + } + if (nxpsg >= 16) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 8); + } + if (nxpsg >= 8) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 4); + } + if (nxpsg >= 4) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 2); + } + if (nxpsg >= 2) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 1); + } + + //sumf[ir1] = simd_sum(sumf[ir1]); + } + + if (tx == 0) { + for (short ir1 = 0; ir1 < r1ptg && i11 + ir1 < args.ne11; ++ir1) { + device float * dst_f32 = (device float *) dst + (uint64_t)i1m*args.ne0*args.ne1 + (uint64_t)(i11 + ir1)*args.ne0; + + if (i01 < args.ne01) { + dst_f32[i01] = sumf[ir1]; + } + } + } +} + +// mat-vec kernel processing in chunks of float4x4 +template +void kernel_mul_mv_ext_q4x4_f32_impl( + constant ggml_metal_kargs_mul_mv_ext & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + const short chpt = 1; + + //const short nxpsg = (32); + const short nypsg = (32/nxpsg); + + const short tx = tiisg%nxpsg; + const short ty = tiisg/nxpsg; + + const int i01 = tgpig.x*(nypsg*args.nsg) + nypsg*sgitg + ty; + const int i11 = tgpig.y*r1ptg; + const int i1m = tgpig.z; + + const int i12 = i1m%args.ne12; + const int i13 = i1m/args.ne12; + + const uint64_t offset0 = i01*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = i11*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; + + device const q_t * xq = (i01 < args.ne01) ? (device const q_t *) (src0 + offset0) + tx/chpb : (device const q_t *) src0; + + device const float4x4 * y4x4[r1ptg]; + + for (int ir1 = 0; ir1 < r1ptg; ++ir1) { + y4x4[ir1] = (i11 + ir1 < args.ne11) ? (device const float4x4 *) (src1 + offset1 + ir1*args.nb11) + tx : (device const float4x4 *) src1; + } + + float sumf[r1ptg] = { [ 0 ... r1ptg - 1 ] = 0.0f }; + + short cch = tx%chpb; + + for (int ich = tx; 16*ich < args.ne00; ich += chpt*nxpsg) { + float4x4 lx[chpt]; + +#pragma unroll(chpt) + for (short ch = 0; ch < chpt; ++ch) { + deq_t4x4(xq, cch, lx[ch]); + + cch += nxpsg; + if (cch >= chpb) { + xq += cch/chpb; + cch %= chpb; + } + } + +#pragma unroll(chpt) + for (short ch = 0; ch < chpt; ++ch) { +#pragma unroll(r1ptg) + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + sumf[ir1] += + dot(lx[ch][0], y4x4[ir1][ch*nxpsg][0]) + + dot(lx[ch][1], y4x4[ir1][ch*nxpsg][1]) + + dot(lx[ch][2], y4x4[ir1][ch*nxpsg][2]) + + dot(lx[ch][3], y4x4[ir1][ch*nxpsg][3]); + + } + } + +#pragma unroll(r1ptg) + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + y4x4[ir1] += chpt*nxpsg; + } + } + + for (short ir1 = 0; ir1 < r1ptg; ++ir1) { + if (nxpsg >= 32) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 16); + } + if (nxpsg >= 16) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 8); + } + if (nxpsg >= 8) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 4); + } + if (nxpsg >= 4) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 2); + } + if (nxpsg >= 2) { + sumf[ir1] += simd_shuffle_down(sumf[ir1], 1); + } + + //sumf[ir1] = simd_sum(sumf[ir1]); + } + + if (tx == 0) { + for (short ir1 = 0; ir1 < r1ptg && i11 + ir1 < args.ne11; ++ir1) { + device float * dst_f32 = (device float *) dst + (uint64_t)i1m*args.ne0*args.ne1 + (uint64_t)(i11 + ir1)*args.ne0; + + if (i01 < args.ne01) { + dst_f32[i01] = sumf[ir1]; + } + } + } +} + +// dispatchers needed for compile-time nxpsg +// epb - elements per quantization block +template +kernel void kernel_mul_mv_ext_q4_f32_disp( + constant ggml_metal_kargs_mul_mv_ext & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + switch (args.nxpsg) { + case 4: kernel_mul_mv_ext_q4_f32_impl<4, r1ptg, q_t, epb/4, deq_t4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 8: kernel_mul_mv_ext_q4_f32_impl<8, r1ptg, q_t, epb/4, deq_t4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 16: kernel_mul_mv_ext_q4_f32_impl<16, r1ptg, q_t, epb/4, deq_t4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 32: kernel_mul_mv_ext_q4_f32_impl<32, r1ptg, q_t, epb/4, deq_t4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + } +} + +template +kernel void kernel_mul_mv_ext_q4x4_f32_disp( + constant ggml_metal_kargs_mul_mv_ext & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + switch (args.nxpsg) { + case 4: kernel_mul_mv_ext_q4x4_f32_impl<4, r1ptg, q_t, epb/16, deq_t4x4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 8: kernel_mul_mv_ext_q4x4_f32_impl<8, r1ptg, q_t, epb/16, deq_t4x4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 16: kernel_mul_mv_ext_q4x4_f32_impl<16, r1ptg, q_t, epb/16, deq_t4x4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + case 32: kernel_mul_mv_ext_q4x4_f32_impl<32, r1ptg, q_t, epb/16, deq_t4x4>(args, src0, src1, dst, tgpig, tiisg, sgitg); break; + } +} + +typedef decltype(kernel_mul_mv_ext_q4_f32_disp <2, block_q8_0, 32, dequantize_q8_0_t4>) mul_mv_ext_q4_f32_t; +typedef decltype(kernel_mul_mv_ext_q4x4_f32_disp<2, block_q4_K, 256, dequantize_q4_K>) mul_mv_ext_q4x4_f32_t; + +template [[host_name("kernel_mul_mv_ext_f16_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, half4, 4, dequantize_f16_t4>; +template [[host_name("kernel_mul_mv_ext_f16_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, half4, 4, dequantize_f16_t4>; +template [[host_name("kernel_mul_mv_ext_f16_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, half4, 4, dequantize_f16_t4>; +template [[host_name("kernel_mul_mv_ext_f16_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, half4, 4, dequantize_f16_t4>; + +template [[host_name("kernel_mul_mv_ext_q4_0_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_q4_0, 32, dequantize_q4_0_t4>; +template [[host_name("kernel_mul_mv_ext_q4_0_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_q4_0, 32, dequantize_q4_0_t4>; +template [[host_name("kernel_mul_mv_ext_q4_0_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_q4_0, 32, dequantize_q4_0_t4>; +template [[host_name("kernel_mul_mv_ext_q4_0_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_q4_0, 32, dequantize_q4_0_t4>; + +template [[host_name("kernel_mul_mv_ext_q4_1_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_q4_1, 32, dequantize_q4_1_t4>; +template [[host_name("kernel_mul_mv_ext_q4_1_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_q4_1, 32, dequantize_q4_1_t4>; +template [[host_name("kernel_mul_mv_ext_q4_1_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_q4_1, 32, dequantize_q4_1_t4>; +template [[host_name("kernel_mul_mv_ext_q4_1_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_q4_1, 32, dequantize_q4_1_t4>; + +template [[host_name("kernel_mul_mv_ext_q5_0_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_q5_0, 32, dequantize_q5_0_t4>; +template [[host_name("kernel_mul_mv_ext_q5_0_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_q5_0, 32, dequantize_q5_0_t4>; +template [[host_name("kernel_mul_mv_ext_q5_0_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_q5_0, 32, dequantize_q5_0_t4>; +template [[host_name("kernel_mul_mv_ext_q5_0_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_q5_0, 32, dequantize_q5_0_t4>; + +template [[host_name("kernel_mul_mv_ext_q5_1_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_q5_1, 32, dequantize_q5_1_t4>; +template [[host_name("kernel_mul_mv_ext_q5_1_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_q5_1, 32, dequantize_q5_1_t4>; +template [[host_name("kernel_mul_mv_ext_q5_1_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_q5_1, 32, dequantize_q5_1_t4>; +template [[host_name("kernel_mul_mv_ext_q5_1_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_q5_1, 32, dequantize_q5_1_t4>; + +template [[host_name("kernel_mul_mv_ext_q8_0_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_q8_0, 32, dequantize_q8_0_t4>; +template [[host_name("kernel_mul_mv_ext_q8_0_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_q8_0, 32, dequantize_q8_0_t4>; +template [[host_name("kernel_mul_mv_ext_q8_0_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_q8_0, 32, dequantize_q8_0_t4>; +template [[host_name("kernel_mul_mv_ext_q8_0_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_q8_0, 32, dequantize_q8_0_t4>; + +template [[host_name("kernel_mul_mv_ext_iq4_nl_f32_r1_2")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<2, block_iq4_nl, 32, dequantize_iq4_nl_t4>; +template [[host_name("kernel_mul_mv_ext_iq4_nl_f32_r1_3")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<3, block_iq4_nl, 32, dequantize_iq4_nl_t4>; +template [[host_name("kernel_mul_mv_ext_iq4_nl_f32_r1_4")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<4, block_iq4_nl, 32, dequantize_iq4_nl_t4>; +template [[host_name("kernel_mul_mv_ext_iq4_nl_f32_r1_5")]] kernel mul_mv_ext_q4_f32_t kernel_mul_mv_ext_q4_f32_disp<5, block_iq4_nl, 32, dequantize_iq4_nl_t4>; + +template [[host_name("kernel_mul_mv_ext_q4_K_f32_r1_2")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<2, block_q4_K, 256, dequantize_q4_K>; +template [[host_name("kernel_mul_mv_ext_q4_K_f32_r1_3")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<3, block_q4_K, 256, dequantize_q4_K>; +template [[host_name("kernel_mul_mv_ext_q4_K_f32_r1_4")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<4, block_q4_K, 256, dequantize_q4_K>; +template [[host_name("kernel_mul_mv_ext_q4_K_f32_r1_5")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<5, block_q4_K, 256, dequantize_q4_K>; + +template [[host_name("kernel_mul_mv_ext_q5_K_f32_r1_2")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<2, block_q5_K, 256, dequantize_q5_K>; +template [[host_name("kernel_mul_mv_ext_q5_K_f32_r1_3")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<3, block_q5_K, 256, dequantize_q5_K>; +template [[host_name("kernel_mul_mv_ext_q5_K_f32_r1_4")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<4, block_q5_K, 256, dequantize_q5_K>; +template [[host_name("kernel_mul_mv_ext_q5_K_f32_r1_5")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<5, block_q5_K, 256, dequantize_q5_K>; + +template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_2")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<2, block_q6_K, 256, dequantize_q6_K>; +template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_3")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<3, block_q6_K, 256, dequantize_q6_K>; +template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_4")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<4, block_q6_K, 256, dequantize_q6_K>; +template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_5")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<5, block_q6_K, 256, dequantize_q6_K>; + #define N_MV_T_T 4 -template +template void kernel_mul_mv_impl( - device const char * src0, - device const char * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb00, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne11, - int64_t ne12, - uint64_t nb10, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - uint3 tgpig, - uint tiisg) { - const int64_t r0 = tgpig.x; - const int64_t rb = tgpig.y*N_MV_T_T; - const int64_t im = tgpig.z; + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig, + ushort tiisg) { + const int r0 = tgpig.x; + const int rb = tgpig.y*N_MV_T_T; + const int im = tgpig.z; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; device const T0 * x = (device const T0 *) (src0 + offset0); - if (ne00 < 128) { + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1; + + if (args.ne00 < 128) { for (int row = 0; row < N_MV_T_T; ++row) { int r1 = rb + row; - if (r1 >= ne11) { + if (r1 >= args.ne11) { break; } - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; device const T1 * y = (device const T1 *) (src1 + offset1); float sumf = 0; - for (int i = tiisg; i < ne00; i += 32) { + for (int i = tiisg; i < args.ne00; i += 32) { sumf += (T0) x[i] * (T1) y[i]; } float all_sum = simd_sum(sumf); if (tiisg == 0) { - dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + dst_f32[(uint64_t)r1*args.ne0 + r0] = all_sum; } } } else { device const T04 * x4 = (device const T04 *) x; for (int row = 0; row < N_MV_T_T; ++row) { int r1 = rb + row; - if (r1 >= ne11) { + if (r1 >= args.ne11) { break; } - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; device const T1 * y = (device const T1 *) (src1 + offset1); device const T14 * y4 = (device const T14 *) y; float sumf = 0; - for (int i = tiisg; i < ne00/4; i += 32) { - for (int k = 0; k < 4; ++k) sumf += (float) (x4[i][k] * y4[i][k]); + for (int i = tiisg; i < args.ne00/4; i += 32) { + sumf += dot((float4) x4[i], (float4) y4[i]); } float all_sum = simd_sum(sumf); if (tiisg == 0) { - for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (float) (x[i] * y[i]); - dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + for (int i = 4*(args.ne00/4); i < args.ne00; ++i) all_sum += (float) (x[i] * y[i]); + dst_f32[(uint64_t)r1*args.ne0 + r0] = all_sum; } } } @@ -2008,51 +2295,17 @@ void kernel_mul_mv_impl( template kernel void kernel_mul_mv( - device const char * src0, - device const char * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]]) { - kernel_mul_mv_impl( + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]]) { + kernel_mul_mv_impl( + args, src0, src1, dst, - ne00, - ne01, - ne02, - nb00, - nb01, - nb02, - nb03, - ne10, - ne11, - ne12, - nb10, - nb11, - nb12, - nb13, - ne0, - ne1, - r2, - r3, tgpig, tiisg); } @@ -2062,72 +2315,57 @@ typedef decltype(kernel_mul_mv) mul_mv_t; template [[host_name("kernel_mul_mv_f32_f32")]] kernel mul_mv_t kernel_mul_mv; template [[host_name("kernel_mul_mv_f16_f32")]] kernel mul_mv_t kernel_mul_mv; template [[host_name("kernel_mul_mv_f16_f16")]] kernel mul_mv_t kernel_mul_mv; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mv_bf16_f32")]] kernel mul_mv_t kernel_mul_mv; template [[host_name("kernel_mul_mv_bf16_bf16")]] kernel mul_mv_t kernel_mul_mv; #endif template kernel void kernel_mul_mv_1row( - device const char * src0, - device const char * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]]) { - const int64_t r0 = tgpig.x; - const int64_t r1 = tgpig.y; - const int64_t im = tgpig.z; + const int r0 = tgpig.x; + const int r1 = tgpig.y; + const int im = tgpig.z; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; device const T * x = (device const T *) (src0 + offset0); device const float * y = (device const float *) (src1 + offset1); + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + float sumf = 0; - if (ne00 < 128) { - for (int i = tiisg; i < ne00; i += 32) { + if (args.ne00 < 128) { + for (int i = tiisg; i < args.ne00; i += 32) { sumf += (float) x[i] * (float) y[i]; } float all_sum = simd_sum(sumf); if (tiisg == 0) { - dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + dst_f32[r0] = all_sum; } } else { device const T4 * x4 = (device const T4 *) x; device const float4 * y4 = (device const float4 *) y; - for (int i = tiisg; i < ne00/4; i += 32) { - for (int k = 0; k < 4; ++k) sumf += (float) (x4[i][k] * y4[i][k]); + for (int i = tiisg; i < args.ne00/4; i += 32) { + sumf += dot((float4) x4[i], y4[i]); } float all_sum = simd_sum(sumf); if (tiisg == 0) { - for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (float) (x[i] * y[i]); - dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + for (int i = 4*(args.ne00/4); i < args.ne00; ++i) all_sum += (float) (x[i] * y[i]); + dst_f32[r0] = all_sum; } } } @@ -2135,61 +2373,46 @@ kernel void kernel_mul_mv_1row( typedef decltype(kernel_mul_mv_1row) mul_mv_1row_t; template [[host_name("kernel_mul_mv_f16_f32_1row")]] kernel mul_mv_1row_t kernel_mul_mv_1row; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mv_bf16_f32_1row")]] kernel mul_mv_1row_t kernel_mul_mv_1row; #endif // Assumes row size (ne00) is a multiple of 4 template kernel void kernel_mul_mv_l4( - device const char * src0, - device const char * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]]) { - const int nrows = ne11; - const int64_t r0 = tgpig.x; - const int64_t im = tgpig.z; + const int nrows = args.ne11; + const int r0 = tgpig.x; + const int im = tgpig.z; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; device const T4 * x4 = (device const T4 *) (src0 + offset0); + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1; + for (int r1 = 0; r1 < nrows; ++r1) { - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; device const float4 * y4 = (device const float4 *) (src1 + offset1); float sumf = 0; - for (int i = tiisg; i < ne00/4; i += 32) { - for (int k = 0; k < 4; ++k) sumf += (float) (x4[i][k] * y4[i][k]); + for (int i = tiisg; i < args.ne00/4; i += 32) { + sumf += dot((float4) x4[i], y4[i]); } float all_sum = simd_sum(sumf); if (tiisg == 0) { - dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + dst_f32[(uint64_t)r1*args.ne0 + r0] = all_sum; } } } @@ -2197,7 +2420,7 @@ kernel void kernel_mul_mv_l4( typedef decltype(kernel_mul_mv_l4) mul_mv_l4_t; template [[host_name("kernel_mul_mv_f16_f32_l4")]] kernel mul_mv_l4_t kernel_mul_mv_l4; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mv_bf16_f32_l4")]] kernel mul_mv_l4_t kernel_mul_mv_l4; #endif @@ -2209,7 +2432,7 @@ static float rope_yarn_ramp(const float low, const float high, const int i0) { // YaRN algorithm based on LlamaYaRNScaledRotaryEmbedding.py from https://github.com/jquesnelle/yarn // MIT licensed. Copyright (c) 2023 Jeffrey Quesnelle and Bowen Peng. static void rope_yarn( - float theta_extrap, float freq_scale, float corr_dims[2], int64_t i0, float ext_factor, float mscale, + float theta_extrap, float freq_scale, float corr_dims[2], int i0, float ext_factor, float mscale, thread float * cos_theta, thread float * sin_theta) { // Get n-d rotational scaling corrected for extrapolation float theta_interp = freq_scale * theta_extrap; @@ -2241,65 +2464,41 @@ static void rope_yarn_corr_dims( template kernel void kernel_rope_norm( - device const void * src0, - device const int32_t * src1, - device const float * src2, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - constant int & n_past, - constant int & n_dims, - constant int & n_ctx_orig, - constant float & freq_base, - constant float & freq_scale, - constant float & ext_factor, - constant float & attn_factor, - constant float & beta_fast, - constant float & beta_slow, - uint tiitg[[thread_index_in_threadgroup]], - uint3 tptg[[threads_per_threadgroup]], - uint3 tgpig[[threadgroup_position_in_grid]]) { - const int64_t i3 = tgpig[2]; - const int64_t i2 = tgpig[1]; - const int64_t i1 = tgpig[0]; + constant ggml_metal_kargs_rope & args, + device const char * src0, + device const char * src1, + device const char * src2, + device char * dst, + ushort tiitg[[thread_index_in_threadgroup]], + ushort3 tptg [[threads_per_threadgroup]], + uint3 tgpig[[threadgroup_position_in_grid]]) { + const int i3 = tgpig[2]; + const int i2 = tgpig[1]; + const int i1 = tgpig[0]; float corr_dims[2]; - rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow, corr_dims); + rope_yarn_corr_dims(args.n_dims, args.n_ctx_orig, args.freq_base, args.beta_fast, args.beta_slow, corr_dims); - device const int32_t * pos = src1; + device const int32_t * pos = (device const int32_t *) src1; const float theta_base = (float) pos[i2]; - const float inv_ndims = -1.f/n_dims; + const float inv_ndims = -1.f/args.n_dims; float cos_theta; float sin_theta; - for (int64_t i0 = 2*tiitg; i0 < ne0; i0 += 2*tptg.x) { - if (i0 < n_dims) { - const int64_t ic = i0/2; + for (int i0 = 2*tiitg; i0 < args.ne0; i0 += 2*tptg.x) { + if (i0 < args.n_dims) { + const int ic = i0/2; - const float theta = theta_base * pow(freq_base, inv_ndims*i0); + const float theta = theta_base * pow(args.freq_base, inv_ndims*i0); - const float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + const float freq_factor = src2 != src0 ? ((device const float *) src2)[ic] : 1.0f; - rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta); + rope_yarn(theta/freq_factor, args.freq_scale, corr_dims, i0, args.ext_factor, args.attn_factor, &cos_theta, &sin_theta); - device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); - device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device const T * const src = (device T *)(src0 + i3*args.nb03 + i2*args.nb02 + i1*args.nb01 + i0*args.nb00); + device T * dst_data = (device T *)( dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); const float x0 = src[0]; const float x1 = src[1]; @@ -2307,8 +2506,8 @@ kernel void kernel_rope_norm( dst_data[0] = x0*cos_theta - x1*sin_theta; dst_data[1] = x0*sin_theta + x1*cos_theta; } else { - device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); - device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device const T * const src = (device T *)(src0 + i3*args.nb03 + i2*args.nb02 + i1*args.nb01 + i0*args.nb00); + device T * dst_data = (device T *)( dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); dst_data[0] = src[0]; dst_data[1] = src[1]; @@ -2318,74 +2517,50 @@ kernel void kernel_rope_norm( template kernel void kernel_rope_neox( - device const void * src0, - device const int32_t * src1, - device const float * src2, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - constant int & n_past, - constant int & n_dims, - constant int & n_ctx_orig, - constant float & freq_base, - constant float & freq_scale, - constant float & ext_factor, - constant float & attn_factor, - constant float & beta_fast, - constant float & beta_slow, - uint tiitg[[thread_index_in_threadgroup]], - uint3 tptg[[threads_per_threadgroup]], - uint3 tgpig[[threadgroup_position_in_grid]]) { - const int64_t i3 = tgpig[2]; - const int64_t i2 = tgpig[1]; - const int64_t i1 = tgpig[0]; + constant ggml_metal_kargs_rope & args, + device const char * src0, + device const char * src1, + device const char * src2, + device char * dst, + ushort tiitg[[thread_index_in_threadgroup]], + ushort3 tptg [[threads_per_threadgroup]], + uint3 tgpig[[threadgroup_position_in_grid]]) { + const int i3 = tgpig[2]; + const int i2 = tgpig[1]; + const int i1 = tgpig[0]; float corr_dims[2]; - rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow, corr_dims); + rope_yarn_corr_dims(args.n_dims, args.n_ctx_orig, args.freq_base, args.beta_fast, args.beta_slow, corr_dims); - device const int32_t * pos = src1; + device const int32_t * pos = (device const int32_t *) src1; const float theta_base = (float) pos[i2]; - const float inv_ndims = -1.f/n_dims; + const float inv_ndims = -1.f/args.n_dims; float cos_theta; float sin_theta; - for (int64_t i0 = 2*tiitg; i0 < ne0; i0 += 2*tptg.x) { - if (i0 < n_dims) { - const int64_t ic = i0/2; + for (int i0 = 2*tiitg; i0 < args.ne0; i0 += 2*tptg.x) { + if (i0 < args.n_dims) { + const int ic = i0/2; - const float theta = theta_base * pow(freq_base, inv_ndims*i0); + const float theta = theta_base * pow(args.freq_base, inv_ndims*i0); - const float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + const float freq_factor = src2 != src0 ? ((device const float *) src2)[ic] : 1.0f; - rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta); + rope_yarn(theta/freq_factor, args.freq_scale, corr_dims, i0, args.ext_factor, args.attn_factor, &cos_theta, &sin_theta); - device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); - device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + device const T * const src = (device T *)(src0 + i3*args.nb03 + i2*args.nb02 + i1*args.nb01 + ic*args.nb00); + device T * dst_data = (device T *)( dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + ic*args.nb0); const float x0 = src[0]; - const float x1 = src[n_dims/2]; + const float x1 = src[args.n_dims/2]; - dst_data[0] = x0*cos_theta - x1*sin_theta; - dst_data[n_dims/2] = x0*sin_theta + x1*cos_theta; + dst_data[0] = x0*cos_theta - x1*sin_theta; + dst_data[args.n_dims/2] = x0*sin_theta + x1*cos_theta; } else { - device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); - device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device const T * const src = (device T *)(src0 + i3*args.nb03 + i2*args.nb02 + i1*args.nb01 + i0*args.nb00); + device T * dst_data = (device T *)( dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); dst_data[0] = src[0]; dst_data[1] = src[1]; @@ -2440,20 +2615,34 @@ kernel void kernel_im2col( uint3 tgpg[[threadgroups_per_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { - const int32_t iiw = tgpig[2] * s0 + tpitg[2] * d0 - p0; - const int32_t iih = tgpig[1] * s1 + tpitg[1] * d1 - p1; +// const int64_t IC = tgpg[0]; + const int64_t OH = tgpg[1]; + const int64_t OW = tgpg[2]; - const int32_t offset_dst = - (tpitg[0] * tgpg[1] * tgpg[2] + tgpig[1] * tgpg[2] + tgpig[2]) * CHW + - (tgpig[0] * (ntg[1] * ntg[2]) + tpitg[1] * ntg[2] + tpitg[2]); +// const int64_t N = ntg[0]; + const int64_t KH = ntg[1]; + const int64_t KW = ntg[2]; + + const int64_t in = tpitg[0]; + const int64_t ikh = tpitg[1]; + const int64_t ikw = tpitg[2]; + + const int64_t iic = tgpig[0]; + const int64_t ioh = tgpig[1]; + const int64_t iow = tgpig[2]; + + const int64_t iiw = iow*s0 + ikw*d0 - p0; + const int64_t iih = ioh*s1 + ikh*d1 - p1; + + const int64_t offset_dst = (in*OH*OW + ioh*OW + iow)*CHW + (iic*(KH*KW) + ikh*KW + ikw); device T * pdst = (device T *) (dst); if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) { pdst[offset_dst] = 0.0f; } else { - const int32_t offset_src = tpitg[0] * ofs0 + tgpig[0] * ofs1; - pdst[offset_dst] = x[offset_src + iih * IW + iiw]; + const int64_t offset_src = in*ofs0 + iic*ofs1 + iih*IW + iiw; + pdst[offset_dst] = x[offset_src]; } } @@ -2504,25 +2693,25 @@ kernel void kernel_im2col_ext( uint3 tgpg[[threadgroups_per_grid]], // tgpg[0] = D x IC x KH x KW, CHW = IC x KH x KW uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { // [M, 1, 1] - const int32_t KHW = KH * KW; // KHW == ntg[1] * ntg[2], KW == ntg[2] + const int64_t KHW = KH * KW; // KHW == ntg[1] * ntg[2], KW == ntg[2] - const int32_t d = tgpig[0] / CHW; - const int32_t chw = tgpig[0] % CHW; - const int32_t tgpig_0 = chw / KHW; // 0 ~ (IC - 1) - const int32_t HW = tgpig[0] % KHW; + const int64_t d = tgpig[0] / CHW; + const int64_t chw = tgpig[0] % CHW; + const int64_t tgpig_0 = chw / KHW; // 0 ~ (IC - 1) + const int64_t HW = tgpig[0] % KHW; - const int32_t tpitg_0 = (d * ntg[0]) + tpitg[0]; + const int64_t tpitg_0 = (d * ntg[0]) + tpitg[0]; if (tpitg_0 >= N) { return; } - const int32_t tpitg_1 = HW / KW; - const int32_t tpitg_2 = HW % KW; + const int64_t tpitg_1 = HW / KW; + const int64_t tpitg_2 = HW % KW; - const int32_t iiw = tgpig[2] * s0 + tpitg_2 * d0 - p0; - const int32_t iih = tgpig[1] * s1 + tpitg_1 * d1 - p1; + const int64_t iiw = tgpig[2] * s0 + tpitg_2 * d0 - p0; + const int64_t iih = tgpig[1] * s1 + tpitg_1 * d1 - p1; - const int32_t offset_dst = + const int64_t offset_dst = (tpitg_0 * tgpg[1] * tgpg[2] + tgpig[1] * tgpg[2] + tgpig[2]) * CHW + (tgpig_0 * KHW + tpitg_1 * KW + tpitg_2); @@ -2531,7 +2720,7 @@ kernel void kernel_im2col_ext( if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) { pdst[offset_dst] = 0.0f; } else { - const int32_t offset_src = tpitg_0 * ofs0 + tgpig_0 * ofs1; + const int64_t offset_src = tpitg_0 * ofs0 + tgpig_0 * ofs1; pdst[offset_dst] = x[offset_src + iih * IW + iiw]; } } @@ -2539,6 +2728,79 @@ kernel void kernel_im2col_ext( template [[host_name("kernel_im2col_ext_f32")]] kernel im2col_ext_t kernel_im2col_ext; template [[host_name("kernel_im2col_ext_f16")]] kernel im2col_ext_t kernel_im2col_ext; +typedef void (conv_transpose_1d_t)( + device const float * src0, + device const float * src1, + device char * dst, + constant int32_t & IC, + constant int32_t & IL, + constant int32_t & K, + constant int32_t & s0, + constant uint64_t & nb0, + constant uint64_t & nb1, + uint3 tgpig[[threadgroup_position_in_grid]], + uint3 tgpg[[threadgroups_per_grid]]); + +template +kernel void kernel_conv_transpose_1d( + device const T * src0, + device const float * src1, + device char * dst, + constant int32_t & IC, + constant int32_t & IL, + constant int32_t & K, + constant int32_t & s0, + constant uint64_t & nb0, + constant uint64_t & nb1, + uint3 tgpig[[threadgroup_position_in_grid]], + uint3 tgpg[[threadgroups_per_grid]]) { + + float v = 0.0f; + + for (int64_t c = 0; c < IC; c++) { + const int32_t kernel_offset = c * tgpg[1] * K + K * tgpig[1]; + const int32_t input_offset = c * IL; + + for (int64_t i = 0; i < IL; i++) { + if (tgpig[0] >= i * s0 && tgpig[0] < i * s0 + K) { + v += src0[kernel_offset + tgpig[0] - i * s0] * src1[input_offset + i]; + } + } + } + + device float * dst_ptr = (device float *) (dst + tgpig[0] * nb0 + tgpig[1] * nb1); + + dst_ptr[0] = v; +} + +template [[host_name("kernel_conv_transpose_1d_f32_f32")]] +kernel void kernel_conv_transpose_1d( + device const float * src0, + device const float * src1, + device char * dst, + constant int32_t & IC, + constant int32_t & IL, + constant int32_t & K, + constant int32_t & s0, + constant uint64_t & nb0, + constant uint64_t & nb1, + uint3 tgpig[[threadgroup_position_in_grid]], + uint3 tgpg[[threadgroups_per_grid]]); + +template [[host_name("kernel_conv_transpose_1d_f16_f32")]] +kernel void kernel_conv_transpose_1d( + device const half * src0, + device const float * src1, + device char * dst, + constant int32_t & IC, + constant int32_t & IL, + constant int32_t & K, + constant int32_t & s0, + constant uint64_t & nb0, + constant uint64_t & nb1, + uint3 tgpig[[threadgroup_position_in_grid]], + uint3 tgpg[[threadgroups_per_grid]]); + kernel void kernel_upscale_f32( device const char * src0, device char * dst, @@ -2635,6 +2897,53 @@ kernel void kernel_pad_f32( } } +kernel void kernel_pad_reflect_1d_f32( + device const char * src0, + device char * dst, + constant int64_t & ne00, + constant int64_t & ne01, + constant int64_t & ne02, + constant int64_t & ne03, + constant int64_t & ne0, + constant uint64_t & nb00, + constant uint64_t & nb01, + constant uint64_t & nb02, + constant uint64_t & nb03, + constant uint64_t & nb0, + constant uint64_t & nb1, + constant uint64_t & nb2, + constant uint64_t & nb3, + constant int32_t & p0, + constant int32_t & p1, + uint3 tgpig[[threadgroup_position_in_grid]], + uint3 tgpg[[threadgroups_per_grid]], + uint3 tpitg[[thread_position_in_threadgroup]], + uint3 ntg[[threads_per_threadgroup]]) { + + const int64_t i3 = tgpig.z; + const int64_t i2 = tgpig.y; + const int64_t i1 = tgpig.x; + + const int64_t i03 = i3; + const int64_t i02 = i2; + const int64_t i01 = i1; + + device const float * src0_ptr = (device const float *) (src0 + i03*nb03 + i02*nb02 + i01*nb01); + device float * dst_ptr = (device float *) (dst + i3*nb3 + i2*nb2 + i1*nb1); + + if (i1 < ne01 && i2 < ne02 && i3 < ne03) { + for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { + if (i0 < p0) { + dst_ptr[i0] = src0_ptr[p0 - i0]; + } else if (i0 < ne0 - p1) { + dst_ptr[i0] = src0_ptr[i0 - p0]; + } else { + dst_ptr[i0] = src0_ptr[(ne0 - p1 - p0) - (p1 + 1 - (ne0 - i0)) - 1]; + } + } + } +} + kernel void kernel_arange_f32( device char * dst, constant int64_t & ne0, @@ -2755,44 +3064,45 @@ kernel void kernel_leaky_relu_f32( } // ref: https://arxiv.org/pdf/2307.08691.pdf -// D - head size, Q - queries per threadgroup, KV - key/value processed per each simdgroup, C - cache items per threadgroup -template +template< + typename q_t, // query types in shared memory + typename q4_t, + typename q8x8_t, + typename k_t, // key types in shared memory + typename k4x4_t, + typename k8x8_t, + typename v_t, // value types in shared memory + typename v4x4_t, + typename v8x8_t, + typename qk_t, // Q*K types + typename qk8x8_t, + typename s_t, // soft-max types + typename s8x8_t, + typename o_t, // attention accumulation types + typename o4_t, + typename o8x8_t, + typename kd4x4_t, // key type in device memory + short nl_k, + void (*deq_k)(device const kd4x4_t *, short, thread k4x4_t &), + typename vd4x4_t, // key type in device memory + short nl_v, + void (*deq_v)(device const vd4x4_t *, short, thread v4x4_t &), + short D, // head size + short Q = 8, // queries per threadgroup + short KV = 8, // key/value processed per each simdgroup + short C = 32> // cache items per threadgroup kernel void kernel_flash_attn_ext( - device const char * q, - device const char * k, - device const char * v, - device const char * mask, - device float * dst, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant uint64_t & nb21, - constant uint64_t & nb22, - constant uint64_t & nb23, - constant uint64_t & nb31, - constant int64_t & ne1, - constant int64_t & ne2, - constant float & scale, - constant float & max_bias, - constant float & m0, - constant float & m1, - constant uint32_t & n_head_log2, - constant float & logit_softcap, - threadgroup half * shared [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]], - ushort tiisg[[thread_index_in_simdgroup]], - ushort sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_flash_attn_ext & args, + device const char * q, + device const char * k, + device const char * v, + device const char * mask, + device char * dst, + threadgroup half * shmem_f16 [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 ntg[[threads_per_threadgroup]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { const short nsg = ntg.y; // number of simdgroups const int iq3 = tgpig[2]; @@ -2803,219 +3113,231 @@ kernel void kernel_flash_attn_ext( const short D8 = D/8; const short D16 = D/16; const short NW = N_SIMDWIDTH; - const short SH = (C + Q); // shared memory per simdgroup in (half) + const short SH = (2*C + Q); // shared memory per simdgroup (s_t == float) - const short T = D + 2*nsg*SH; // shared memory size per query in (half) - const short TF = T/2; // shared memory size per query in (float) - const short T4 = T/4; // shared memory size per query in (half4) + const short TS = nsg*SH; // shared memory size per query in (s_t == float) + const short T = D + 2*TS; // shared memory size per query in (half) - threadgroup half * sq = (threadgroup half *) (shared + 0*D); // holds the query data - threadgroup half4 * sq4 = (threadgroup half4 *) (shared + 0*D); // same as above but in half4 - threadgroup float * ss = (threadgroup float *) (shared + 2*sgitg*SH + 1*D); // scratch buffer for attention and diagonal matrix + threadgroup q_t * sq = (threadgroup q_t *) (shmem_f16 + 0*D); // holds the query data + threadgroup q4_t * sq4 = (threadgroup q4_t *) (shmem_f16 + 0*D); // same as above but in q4_t + threadgroup o_t * so = (threadgroup o_t *) (shmem_f16 + 0*D); // reuse query data for accumulation + threadgroup o4_t * so4 = (threadgroup o4_t *) (shmem_f16 + 0*D); // same as above but in o4_t + threadgroup s_t * ss = (threadgroup s_t *) (shmem_f16 + 2*sgitg*SH + Q*D); // scratch buffer for attention, mask and diagonal matrix - threadgroup half * skv = (threadgroup half *) (shared + sgitg*(4*16*KV) + Q*T); // scratch buffer to load K and V in shared memory - threadgroup half4x4 * skv4 = (threadgroup half4x4 *) (shared + sgitg*(4*16*KV) + Q*T); // same as above but in half4x4 + threadgroup k_t * sk = (threadgroup k_t *) (shmem_f16 + sgitg*(4*16*KV) + Q*T); // scratch buffer to load K in shared memory + threadgroup k4x4_t * sk4x4 = (threadgroup k4x4_t *) (shmem_f16 + sgitg*(4*16*KV) + Q*T); // same as above but in k4x4_t + + threadgroup v_t * sv = (threadgroup v_t *) (shmem_f16 + sgitg*(4*16*KV) + Q*T); // scratch buffer to load V in shared memory + threadgroup v4x4_t * sv4x4 = (threadgroup v4x4_t *) (shmem_f16 + sgitg*(4*16*KV) + Q*T); // same as above but in v4x4_t // store the result for all queries in local memory in 8x8 matrices (the O matrix from the paper) - simdgroup_half8x8 lo[D8]; + o8x8_t lo[D8]; // load heads from Q to shared memory for (short j = sgitg; j < Q; j += nsg) { - device const float4 * q4 = (device const float4 *) ((device const char *) q + ((iq1 + j)*nb01 + iq2*nb02 + iq3*nb03)); + device const float4 * q4 = (device const float4 *) ((device const char *) q + ((iq1 + j)*args.nb01 + iq2*args.nb02 + iq3*args.nb03)); for (short i = tiisg; i < D4; i += NW) { - if (iq1 + j < ne01) { - sq4[j*T4 + i] = (half4) q4[i]; + if (iq1 + j < args.ne01) { + sq4[j*D4 + i] = (q4_t) q4[i]; } else { - sq4[j*T4 + i] = 0.0h; + sq4[j*D4 + i] = (q4_t) 0.0f; } } } // zero out lo for (short i = 0; i < D8; ++i) { - lo[i] = make_filled_simdgroup_matrix(0.0h); + lo[i] = make_filled_simdgroup_matrix((o_t) 0.0f); } // zero out shared memory SH for (short j = 0; j < Q; ++j) { for (short i = tiisg; i < SH; i += NW) { - ss[j*TF + i] = 0.0f; + ss[j*TS + i] = 0.0f; } } threadgroup_barrier(mem_flags::mem_threadgroup); { - float S[Q] = { [0 ... Q-1] = 0.0f }; - float M[Q] = { [0 ... Q-1] = -FLT_MAX/2 }; + half S[Q] = { [0 ... Q-1] = 0.0f }; + half M[Q] = { [0 ... Q-1] = -__FLT16_MAX__/2 }; // thread indices inside the simdgroup + // TODO: see if we can utilize quad-group functions for better performance + // https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (6.9.3) const short tx = tiisg%4; const short ty = tiisg/4; - // assume K and V are same shape - const short ne22 = ne12; - const short ne23 = ne13; + // broadcast kv + //const short rk2 = args.ne02/args.ne12; + //const short rk3 = args.ne03/args.ne13; - // broadcast k - const short rk2 = ne02/ne12; - const short rk3 = ne03/ne13; - - const short ik2 = iq2/rk2; - const short ik3 = iq3/rk3; - - // broadcast v - const short rv2 = ne02/ne22; - const short rv3 = ne03/ne23; - - const short iv2 = iq2/rv2; - const short iv3 = iq3/rv3; + const short ikv2 = iq2/(args.ne02/args.ne_12_2); + const short ikv3 = iq3/(args.ne03/args.ne_12_3); // load the queries from shared memory into local memory - simdgroup_half8x8 mq[D8]; + q8x8_t mq[D8]; for (short i = 0; i < D8; ++i) { - simdgroup_load(mq[i], sq + i*8, T); + simdgroup_load(mq[i], sq + i*8, D); } - // pointer to the mask - device const half * mp = (device const half *) (mask + iq1*nb31); + const bool has_mask = mask != q; - float slope = 1.0f; + half slope = 1.0f; // ALiBi - if (max_bias > 0.0f) { - const uint32_t h = iq2; + if (args.max_bias > 0.0f) { + const short h = iq2; - const float base = h < n_head_log2 ? m0 : m1; - const int exph = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + const half base = h < args.n_head_log2 ? args.m0 : args.m1; + const short exph = h < args.n_head_log2 ? h + 1 : 2*(h - args.n_head_log2) + 1; slope = pow(base, exph); } // loop over the KV cache // each simdgroup handles blocks of Q rows and C columns - for (int ic0 = 0; ic0 < ne11; ic0 += C*nsg) { + for (int ic0 = 0; ic0 < args.ne11; ic0 += C*nsg) { const int ic = ic0 + C*sgitg; - if (ic >= ne11) { + if (ic >= args.ne11) { break; } + if (has_mask) { + // used to detect blocks full of -INF + half smax = -INFINITY; + + // load the mask in shared memory + #pragma unroll(Q) + for (short j = 0; j < Q; ++j) { + device const half * pm = (device const half *) ((device const char *) mask + (iq1 + j)*args.nb31); + + const half m = pm[ic + tiisg]; + + ss[j*TS + C + tiisg] = m; + smax = max(smax, m); + } + + smax = simd_max(smax); + + if (smax == -INFINITY) { + continue; + } + } + // Q*K^T { for (short cc = 0; cc < C/8; ++cc) { - simdgroup_float8x8 mqk = make_filled_simdgroup_matrix(0.h); + qk8x8_t mqk = make_filled_simdgroup_matrix((qk_t) 0.0f); // this is compile-time check, so it does not have runtime overhead - if (is_same::value) { + if (is_same::value) { // we can read directly from global memory - device const half * pk = (device const half *) ((device const char *) k + ((ic + 8*cc)*nb11 + ik2*nb12 + ik3*nb13)); + device const k_t * pk = (device const k_t *) ((device const char *) k + ((ic + 8*cc)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); + #pragma unroll(D8) for (short i = 0; i < D8; ++i) { - simdgroup_half8x8 mk; - simdgroup_load(mk, pk + i*8, nb11/sizeof(half), 0, true); // transpose + k8x8_t mk; + simdgroup_load(mk, pk + i*8, args.nb_12_1/sizeof(k_t), 0, true); // transpose // TODO: use ne10 simdgroup_multiply_accumulate(mqk, mq[i], mk, mqk); } } else { for (short ii = 0; ii < D16; ii += 4) { - device const block_q * pk4 = (device const block_q *) ((device const char *) k + ((ic + 8*cc + ty)*nb11 + ik2*nb12 + ik3*nb13)); + device const kd4x4_t * pk4x4 = (device const kd4x4_t *) ((device const char *) k + ((ic + 8*cc + ty)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); if (D16%4 == 0) { // the head is evenly divisible by 4*16 = 64, so no need for bound checks - half4x4 tmp; - dequantize_func(pk4 + (ii + tx)/nl, (ii + tx)%nl, tmp); - skv4[4*ty + tx] = tmp; + { + k4x4_t tmp; + deq_k(pk4x4 + (ii + tx)/nl_k, (ii + tx)%nl_k, tmp); + sk4x4[4*ty + tx] = tmp; + } simdgroup_barrier(mem_flags::mem_threadgroup); -#pragma unroll + #pragma unroll(4) for (short k = 0; k < 4; ++k) { - simdgroup_half8x8 mk; + k8x8_t mk; - simdgroup_load(mk, skv + 16*k + 0*8, 4*16, 0, true); // transpose + simdgroup_load(mk, sk + 16*k + 0*8, 4*16, 0, true); // transpose simdgroup_multiply_accumulate(mqk, mq[2*(ii + k) + 0], mk, mqk); - simdgroup_load(mk, skv + 16*k + 1*8, 4*16, 0, true); // transpose + simdgroup_load(mk, sk + 16*k + 1*8, 4*16, 0, true); // transpose simdgroup_multiply_accumulate(mqk, mq[2*(ii + k) + 1], mk, mqk); } } else { if (ii + tx < D16) { - half4x4 tmp; - dequantize_func(pk4 + (ii + tx)/nl, (ii + tx)%nl, tmp); - skv4[4*ty + tx] = tmp; + k4x4_t tmp; + deq_k(pk4x4 + (ii + tx)/nl_k, (ii + tx)%nl_k, tmp); + sk4x4[4*ty + tx] = tmp; } simdgroup_barrier(mem_flags::mem_threadgroup); for (short k = 0; k < 4 && ii + k < D16; ++k) { - simdgroup_half8x8 mk; + k8x8_t mk; - simdgroup_load(mk, skv + 16*k + 0*8, 4*16, 0, true); // transpose + simdgroup_load(mk, sk + 16*k + 0*8, 4*16, 0, true); // transpose simdgroup_multiply_accumulate(mqk, mq[2*(ii + k) + 0], mk, mqk); - simdgroup_load(mk, skv + 16*k + 1*8, 4*16, 0, true); // transpose + simdgroup_load(mk, sk + 16*k + 1*8, 4*16, 0, true); // transpose simdgroup_multiply_accumulate(mqk, mq[2*(ii + k) + 1], mk, mqk); } } } } - simdgroup_store(mqk, ss + 8*cc, TF, 0, false); + // cast qk_t -> s_t + //s8x8_t mqks(1.0f); + //simdgroup_multiply(mqks, mqk, mqks); + //simdgroup_store(mqks, ss + 8*cc, TS, 0, false); + + simdgroup_store(mqk, ss + 8*cc, TS, 0, false); } } - // used to detect blocks full of -INF - float smax = -INFINITY; - // online softmax { - float ms[Q]; - - for (short j = 0; j < Q; ++j) { - const float m = M[j]; + for (ushort j = 0; j < Q; ++j) { + const half m = M[j]; // scale and apply the logitcap / mask - float s = ss[j*TF + tiisg]*scale; + half s = ss[j*TS + tiisg]*args.scale; - if (logit_softcap != 0.0f) { - s = logit_softcap*precise::tanh(s); + if (args.logit_softcap != 0.0f) { + s = args.logit_softcap*precise::tanh(s); } - if (mask != q) { - // mqk = mqk + mask*slope - s += slope*mp[ic + j*nb31/sizeof(half) + tiisg]; - } + // mqk = mqk + mask*slope + s += slope*ss[j*TS + C + tiisg]; - smax = simd_max(max(smax, s)); M[j] = simd_max(max(M[j], s)); - ms[j] = exp(m - M[j]); - const float vs = exp(s - M[j]); + const half ms = exp(m - M[j]); + const half vs = exp(s - M[j]); - S[j] = S[j]*ms[j] + simd_sum(vs); + S[j] = S[j]*ms + simd_sum(vs); // the P matrix from the paper (Q rows, C columns) - ss[j*TF + tiisg] = vs; - } + ss[j*TS + tiisg] = vs; - // create a QxQ diagonal matrix for rescaling the output - if (tiisg < Q) { - ss[tiisg*TF + C + tiisg] = ms[tiisg]; + // create a QxQ diagonal matrix for rescaling the output + if (tiisg == j) { + ss[j*TS + 2*C + j] = ms; + } } } - // skip -INF blocks - if (smax == -INFINITY) { - continue; - } - // O = diag(ms)*O { - simdgroup_float8x8 mm; - simdgroup_load(mm, ss + C, TF, 0, false); + s8x8_t mm; + simdgroup_load(mm, ss + 2*C, TS, 0, false); + #pragma unroll(D8) for (short i = 0; i < D8; ++i) { simdgroup_multiply(lo[i], mm, lo[i]); } @@ -3024,57 +3346,60 @@ kernel void kernel_flash_attn_ext( // O = O + (Q*K^T)*V { for (short cc = 0; cc < C/8; ++cc) { - simdgroup_float8x8 ms; - simdgroup_load(ms, ss + 8*cc, TF, 0, false); + s8x8_t ms; + simdgroup_load(ms, ss + 8*cc, TS, 0, false); - if (is_same::value) { + if (is_same::value) { // we can read directly from global memory - device const half * pv = (device const half *) ((device const char *) v + ((ic + 8*cc)*nb21 + iv2*nb22 + iv3*nb23)); -#pragma unroll + device const v_t * pv = (device const v_t *) ((device const char *) v + ((ic + 8*cc)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); + + #pragma unroll(D8) for (short i = 0; i < D8; ++i) { - simdgroup_half8x8 mv; - simdgroup_load(mv, pv + i*8, nb21/sizeof(half), 0, false); + v8x8_t mv; + simdgroup_load(mv, pv + i*8, args.nb_12_1/sizeof(v_t), 0, false); // TODO: use ne20 simdgroup_multiply_accumulate(lo[i], ms, mv, lo[i]); } } else { for (short ii = 0; ii < D16; ii += 4) { - device const block_q * pv4 = (device const block_q *) ((device const char *) v + ((ic + 8*cc + ty)*nb21 + iv2*nb22 + iv3*nb23)); + device const vd4x4_t * pv4x4 = (device const vd4x4_t *) ((device const char *) v + ((ic + 8*cc + ty)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); if (D16%4 == 0) { // no need for bound checks - half4x4 tmp; - dequantize_func(pv4 + (ii + tx)/nl, (ii + tx)%nl, tmp); - skv4[4*ty + tx] = tmp; + { + v4x4_t tmp; + deq_v(pv4x4 + (ii + tx)/nl_v, (ii + tx)%nl_v, tmp); + sv4x4[4*ty + tx] = tmp; + } simdgroup_barrier(mem_flags::mem_threadgroup); -#pragma unroll + #pragma unroll(4) for (short k = 0; k < 4; ++k) { - simdgroup_half8x8 mv; + v8x8_t mv; - simdgroup_load(mv, skv + 16*k + 0*8, 4*16, 0, false); + simdgroup_load(mv, sv + 16*k + 0*8, 4*16, 0, false); simdgroup_multiply_accumulate(lo[2*(ii + k) + 0], ms, mv, lo[2*(ii + k) + 0]); - simdgroup_load(mv, skv + 16*k + 1*8, 4*16, 0, false); + simdgroup_load(mv, sv + 16*k + 1*8, 4*16, 0, false); simdgroup_multiply_accumulate(lo[2*(ii + k) + 1], ms, mv, lo[2*(ii + k) + 1]); } } else { if (ii + tx < D16) { - half4x4 tmp; - dequantize_func(pv4 + (ii + tx)/nl, (ii + tx)%nl, tmp); - skv4[4*ty + tx] = tmp; + v4x4_t tmp; + deq_v(pv4x4 + (ii + tx)/nl_v, (ii + tx)%nl_v, tmp); + sv4x4[4*ty + tx] = tmp; } simdgroup_barrier(mem_flags::mem_threadgroup); for (short k = 0; k < 4 && ii + k < D16; ++k) { - simdgroup_half8x8 mv; + v8x8_t mv; - simdgroup_load(mv, skv + 16*k + 0*8, 4*16, 0, false); + simdgroup_load(mv, sv + 16*k + 0*8, 4*16, 0, false); simdgroup_multiply_accumulate(lo[2*(ii + k) + 0], ms, mv, lo[2*(ii + k) + 0]); - simdgroup_load(mv, skv + 16*k + 1*8, 4*16, 0, false); + simdgroup_load(mv, sv + 16*k + 1*8, 4*16, 0, false); simdgroup_multiply_accumulate(lo[2*(ii + k) + 1], ms, mv, lo[2*(ii + k) + 1]); } } @@ -3087,23 +3412,23 @@ kernel void kernel_flash_attn_ext( // these are needed for reducing the results from the simdgroups (reuse the ss buffer) for (short j = 0; j < Q; ++j) { if (tiisg == 0) { - ss[j*TF + 0] = S[j]; - ss[j*TF + 1] = M[j]; + ss[j*TS + 0] = S[j]; + ss[j*TS + 1] = M[j]; } } } // reduce the warps sequentially - for (short sg = 1; sg < nsg; ++sg) { - float S = { 0.0f }; - float M = { -FLT_MAX/2 }; + for (ushort sg = 1; sg < nsg; ++sg) { + half S = { 0.0f }; + half M = { -__FLT16_MAX__/2 }; threadgroup_barrier(mem_flags::mem_threadgroup); // each simdgroup stores its output to shared memory, reusing sq if (sgitg == sg) { for (short i = 0; i < D8; ++i) { - simdgroup_store(lo[i], sq + i*8, T, 0, false); + simdgroup_store(lo[i], so + i*8, D, 0, false); } } @@ -3112,39 +3437,41 @@ kernel void kernel_flash_attn_ext( // the first simdgroup accumulates the results from the other simdgroups if (sgitg == 0) { for (short j = 0; j < Q; ++j) { - const float S0 = ss[j*TF + 0]; - const float S1 = ss[j*TF + sg*SH + 0]; + const half S0 = ss[j*TS + 0]; + const half S1 = ss[j*TS + sg*SH + 0]; - const float M0 = ss[j*TF + 1]; - const float M1 = ss[j*TF + sg*SH + 1]; + const half M0 = ss[j*TS + 1]; + const half M1 = ss[j*TS + sg*SH + 1]; M = max(M0, M1); - const float ms0 = exp(M0 - M); - const float ms1 = exp(M1 - M); + const half ms0 = exp(M0 - M); + const half ms1 = exp(M1 - M); S = S0*ms0 + S1*ms1; if (tiisg == 0) { - ss[j*TF + 0] = S; - ss[j*TF + 1] = M; + ss[j*TS + 0] = S; + ss[j*TS + 1] = M; - ss[j*TF + C + j ] = ms0; - ss[j*TF + C + j + sg*SH] = ms1; + ss[j*TS + 2*C + j ] = ms0; + ss[j*TS + 2*C + j + sg*SH] = ms1; } } // O_0 = diag(ms0)*O_0 + diag(ms1)*O_1 { - simdgroup_half8x8 t; - simdgroup_float8x8 ms0; - simdgroup_float8x8 ms1; + s8x8_t ms0; + s8x8_t ms1; - simdgroup_load(ms0, ss + C, TF, 0, false); - simdgroup_load(ms1, ss + C + sg*SH, TF, 0, false); + simdgroup_load(ms0, ss + 2*C, TS, 0, false); + simdgroup_load(ms1, ss + 2*C + sg*SH, TS, 0, false); + #pragma unroll(D8) for (short i = 0; i < D8; ++i) { - simdgroup_load (t, sq + i*8, T, 0, false); + o8x8_t t; + + simdgroup_load (t, so + i*8, D, 0, false); simdgroup_multiply(t, ms1, t); simdgroup_multiply_accumulate(lo[i], ms0, lo[i], t); @@ -3156,7 +3483,7 @@ kernel void kernel_flash_attn_ext( // store result to shared memory (reuse sq) if (sgitg == 0) { for (short i = 0; i < D8; ++i) { - simdgroup_store(lo[i], sq + i*8, T, 0, false); + simdgroup_store(lo[i], so + i*8, D, 0, false); } } @@ -3164,99 +3491,113 @@ kernel void kernel_flash_attn_ext( // final rescale with 1/S and store to global memory if (sgitg == 0) { - for (short j = 0; j < Q && iq1 + j < ne01; ++j) { - const float S = ss[j*TF + 0]; + for (short j = 0; j < Q && iq1 + j < args.ne01; ++j) { + const float S = ss[j*TS + 0]; for (short i = tiisg; i < D4; i += NW) { - dst4[(iq3*ne2*ne1 + iq2 + (iq1 + j)*ne1)*D4 + i] = (float4) sq4[j*T4 + i]/S; + dst4[((uint64_t)iq3*args.ne2*args.ne1 + iq2 + (uint64_t)(iq1 + j)*args.ne1)*D4 + i] = (float4) so4[j*D4 + i]/S; } } } } -typedef decltype(kernel_flash_attn_ext) flash_attn_ext_t; +// TODO: this is quite ugly. in the future these types will be hardcoded in the kernel, but for now keep them as +// template to be able to explore different combinations +// +#define FA_TYPES \ + half, half4, simdgroup_half8x8, \ + half, half4x4, simdgroup_half8x8, \ + half, half4x4, simdgroup_half8x8, \ + float, simdgroup_float8x8, \ + float, simdgroup_float8x8, \ + half, half4, simdgroup_half8x8 -template [[host_name("kernel_flash_attn_ext_f16_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_f16_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_f16_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_f16_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_f16_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_f16_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +typedef decltype(kernel_flash_attn_ext) flash_attn_ext_t; -template [[host_name("kernel_flash_attn_ext_q4_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_f16_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q4_1_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +#if defined(GGML_METAL_USE_BF16) +template [[host_name("kernel_flash_attn_ext_bf16_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_bf16_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_bf16_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_bf16_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_bf16_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_bf16_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +#endif -template [[host_name("kernel_flash_attn_ext_q5_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q5_1_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q4_1_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -template [[host_name("kernel_flash_attn_ext_q8_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -// NOTE: can use half instead of float precision for some extra perf -// D - head size, Q - queries per threadgroup, C - cache items per threadgroup -template +template [[host_name("kernel_flash_attn_ext_q5_1_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_1_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_1_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_1_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_1_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q5_1_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; + +template [[host_name("kernel_flash_attn_ext_q8_0_h64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q8_0_h80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q8_0_h96" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q8_0_h112")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q8_0_h128")]] kernel flash_attn_ext_t kernel_flash_attn_ext; +template [[host_name("kernel_flash_attn_ext_q8_0_h256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; + +#undef FA_TYPES + +template< + typename q4_t, // query types in shared memory + typename q4x4_t, + typename k4x4_t, // key types in shared memory + typename v4x4_t, // value types in shared memory + typename qk_t, // Q*K types + typename s_t, // soft-max types + typename s4_t, + typename s4x4_t, + typename o4x4_t, // attention accumulation types + typename kd4x4_t, // key type in device memory + short nl_k, + void (*deq_k)(device const kd4x4_t *, short, thread k4x4_t &), + typename vd4x4_t, // key type in device memory + short nl_v, + void (*deq_v)(device const vd4x4_t *, short, thread v4x4_t &), + short D, // head size + short Q = 1, // queries per threadgroup + short C = 32> // cache items per threadgroup kernel void kernel_flash_attn_ext_vec( - device const char * q, - device const char * k, - device const char * v, - device const char * mask, - device float * dst, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant uint64_t & nb21, - constant uint64_t & nb22, - constant uint64_t & nb23, - constant uint64_t & nb31, - constant int64_t & ne1, - constant int64_t & ne2, - constant float & scale, - constant float & max_bias, - constant float & m0, - constant float & m1, - constant uint32_t & n_head_log2, - constant float & logit_softcap, - threadgroup half * shared [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]], - ushort tiisg[[thread_index_in_simdgroup]], - ushort sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_flash_attn_ext & args, + device const char * q, + device const char * k, + device const char * v, + device const char * mask, + device char * dst, + threadgroup half * shmem_f16 [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 ntg[[threads_per_threadgroup]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { const short nsg = ntg.y; // number of simdgroups const int iq3 = tgpig[2]; @@ -3266,122 +3607,126 @@ kernel void kernel_flash_attn_ext_vec( const short D4 = D/4; const short D16 = D/16; const short NW = N_SIMDWIDTH; - const short NW4 = NW/4; - const short SH = C; // shared memory per simdgroup in (half) + const short NL = NW/4; // note: this can be adjusted to support D%64 == 0 and D%32 == 0 + const short SH = 2*C; // shared memory per simdgroup - const short T = D + 2*nsg*SH; // shared memory size per query in (half) + const short T = D + nsg*SH; // shared memory size per query in (half) - //threadgroup half * sq = (threadgroup half *) (shared + 0*D); // holds the query data - threadgroup half4 * sq4 = (threadgroup half4 *) (shared + 0*D); // same as above but in half4 - threadgroup half4x4 * sq44 = (threadgroup half4x4 *) (shared + 0*D); // same as above but in half4x4 - threadgroup float * ss = (threadgroup float *) (shared + 2*sgitg*SH + 1*D); // scratch buffer for attention - threadgroup float4 * ss4 = (threadgroup float4 *) (shared + 2*sgitg*SH + 1*D); // same as above but in half4 - threadgroup float4x4 * sr44 = (threadgroup float4x4 *) (shared + 2*sgitg*D + Q*T); // scratch buffer for the results + //threadgroup q_t * sq = (threadgroup q_t *) (shmem_f16 + 0*D); // holds the query data + threadgroup q4_t * sq4 = (threadgroup q4_t *) (shmem_f16 + 0*D); // same as above but in q4_t + threadgroup q4x4_t * sq4x4 = (threadgroup q4x4_t *) (shmem_f16 + 0*D); // same as above but in q4x4_t + threadgroup s_t * ss = (threadgroup s_t *) (shmem_f16 + sgitg*SH + Q*D); // scratch buffer for attention + threadgroup s4_t * ss4 = (threadgroup s4_t *) (shmem_f16 + sgitg*SH + Q*D); // same as above but in s4_t + threadgroup half * sm = (threadgroup half *) (shmem_f16 + sgitg*SH + C + Q*D); // scratch buffer for mask + threadgroup o4x4_t * sr4x4 = (threadgroup o4x4_t *) (shmem_f16 + sgitg*D + Q*T); // scratch buffer for the results // store the result for all queries in local memory in 8x8 matrices (the O matrix from the paper) - float4x4 lo[D16/NW4]; + o4x4_t lo[D16/NL]; // load heads from Q to shared memory - device const float4 * q4 = (device const float4 *) ((device const char *) q + (iq1*nb01 + iq2*nb02 + iq3*nb03)); + device const float4 * q4 = (device const float4 *) ((device const char *) q + (iq1*args.nb01 + iq2*args.nb02 + iq3*args.nb03)); for (short i = tiisg; i < D4; i += NW) { - if (iq1 < ne01) { - sq4[i] = (half4) q4[i]; + if (iq1 < args.ne01) { + sq4[i] = (q4_t) q4[i]; } else { - sq4[i] = 0.0h; + sq4[i] = (q4_t) 0.0f; } } // zero out lo - for (short i = 0; i < D16/NW4; i += NW4) { - lo[i] = float4x4(0.0f); + for (short i = 0; i < D16/NL; ++i) { + lo[i] = (o4x4_t) 0.0f; } // zero out shared memory SH for (short i = tiisg; i < SH/4; i += NW) { - ss4[i] = 0.0h; + ss4[i] = (s4_t) 0.0f; } threadgroup_barrier(mem_flags::mem_threadgroup); { - float S = 0.0f; - float M = -FLT_MAX/2; + half S = 0.0f; + half M = -__FLT16_MAX__/2; // thread indices inside the simdgroup - const short tx = tiisg%8; - const short ty = tiisg/8; + const short tx = tiisg%NL; + const short ty = tiisg/NL; - // assume K and V are same shape - const short ne22 = ne12; - const short ne23 = ne13; + // broadcast kv + //const short rk2 = args.ne02/args.ne12; + //const short rk3 = args.ne03/args.ne13; - // broadcast k - const short rk2 = ne02/ne12; - const short rk3 = ne03/ne13; - - const short ik2 = iq2/rk2; - const short ik3 = iq3/rk3; - - // broadcast v - const short rv2 = ne02/ne22; - const short rv3 = ne03/ne23; - - const short iv2 = iq2/rv2; - const short iv3 = iq3/rv3; + const short ikv2 = iq2/(args.ne02/args.ne_12_2); + const short ikv3 = iq3/(args.ne03/args.ne_12_3); // load the queries from shared memory into local memory - float4x4 mq[D16/NW4]; + q4x4_t mq[D16/NL]; - for (short ii = 0; ii < D16; ii += NW4) { - mq[ii/NW4] = (float4x4) sq44[ii + tx]; + #pragma unroll(D16/NL) + for (short ii = 0; ii < D16; ii += NL) { + mq[ii/NL] = sq4x4[ii + tx]; } - // pointer to the mask - device const half * mp = (device const half *) (mask + iq1*nb31); + const bool has_mask = mask != q; - float slope = 1.0f; + // pointer to the mask + device const half * pm = (device const half *) (mask + iq1*args.nb31); + + half slope = 1.0f; // ALiBi - if (max_bias > 0.0f) { - const uint32_t h = iq2; + if (args.max_bias > 0.0f) { + const short h = iq2; - const float base = h < n_head_log2 ? m0 : m1; - const int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + const half base = h < args.n_head_log2 ? args.m0 : args.m1; + const short exph = h < args.n_head_log2 ? h + 1 : 2*(h - args.n_head_log2) + 1; - slope = pow(base, exp); + slope = pow(base, exph); } // loop over the KV cache // each simdgroup handles blocks of Q rows and C columns - for (int ic0 = 0; ic0 < ne11; ic0 += C*nsg) { + for (int ic0 = 0; ic0 < args.ne11; ic0 += C*nsg) { const int ic = ic0 + C*sgitg; - if (ic >= ne11) { + if (ic >= args.ne11) { break; } + if (has_mask) { + sm[tiisg] = pm[ic + tiisg]; + } + // Q*K^T { - // each simdgroup processes 1 query and 4 keys + // each simdgroup processes 1 query and 4 (NW/NL) keys for (short cc = 0; cc < C/4; ++cc) { - float mqk = 0.0; + qk_t mqka[4] = { 0.0, 0.0, 0.0, 0.0 }; - device const block_q * pk = (device const block_q *) ((device const char *) k + ((ic + 4*cc + ty)*nb11 + ik2*nb12 + ik3*nb13)); + device const kd4x4_t * pk = (device const kd4x4_t *) ((device const char *) k + ((ic + 4*cc + ty)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); -#pragma unroll - for (short ii = 0; ii < D16; ii += NW4) { + #pragma unroll(D16/NL) + for (short ii = 0; ii < D16; ii += NL) { const short i = ii + tx; - float4x4 mk; - dequantize_func(pk + i/nl, i%nl, mk); + k4x4_t mk; + deq_k(pk + i/nl_k, i%nl_k, mk); - mqk += - dot(mq[ii/NW4][0], mk[0]) + - dot(mq[ii/NW4][1], mk[1]) + - dot(mq[ii/NW4][2], mk[2]) + - dot(mq[ii/NW4][3], mk[3]); + // note: this is less precise than the version below + //mqka[0] += dot(mq[ii/NL][0], mk[0]); + //mqka[1] += dot(mq[ii/NL][1], mk[1]); + //mqka[2] += dot(mq[ii/NL][2], mk[2]); + //mqka[3] += dot(mq[ii/NL][3], mk[3]); + + mqka[0] += dot((float4) mq[ii/NL][0], (float4) mk[0]); + mqka[1] += dot((float4) mq[ii/NL][1], (float4) mk[1]); + mqka[2] += dot((float4) mq[ii/NL][2], (float4) mk[2]); + mqka[3] += dot((float4) mq[ii/NL][3], (float4) mk[3]); } + qk_t mqk = mqka[0] + mqka[1] + mqka[2] + mqka[3]; + // simdgroup reduce // [ 0 .. 7] -> [ 0] // [ 8 .. 15] -> [ 8] @@ -3395,13 +3740,13 @@ kernel void kernel_flash_attn_ext_vec( // mqk = mqk*scale + mask*slope if (tx == 0) { - mqk *= scale; + mqk *= args.scale; - if (logit_softcap != 0.0f) { - mqk = logit_softcap*precise::tanh(mqk); + if (args.logit_softcap != 0.0f) { + mqk = args.logit_softcap*precise::tanh(mqk); } - mqk += (mask != q) ? ((float) mp[ic + 4*cc + ty])*slope : (float) 0.0f; + mqk += sm[4*cc + ty]*slope; ss[4*cc + ty] = mqk; } @@ -3412,25 +3757,23 @@ kernel void kernel_flash_attn_ext_vec( // online softmax { - const short p = tiisg; - - const float m = M; - const float s = ss[p]; + const half m = M; + const half s = ss[tiisg]; M = simd_max(max(M, s)); - const float ms = exp(m - M); - const float vs = exp(s - M); + const half ms = exp(m - M); + const half vs = exp(s - M); S = S*ms + simd_sum(vs); // the P matrix from the paper (Q rows, C columns) - ss[p] = vs; + ss[tiisg] = vs; // O = diag(ms)*O -#pragma unroll - for (short ii = 0; ii < D16; ii += NW4) { - lo[ii/NW4] *= ms; + #pragma unroll(D16/NL) + for (short ii = 0; ii < D16; ii += NL) { + lo[ii/NL] *= ms; } } @@ -3438,20 +3781,19 @@ kernel void kernel_flash_attn_ext_vec( // O = O + (Q*K^T)*V { -#pragma unroll for (short cc = 0; cc < C/4; ++cc) { - device const block_q * pv4 = (device const block_q *) ((device const char *) v + ((ic + 4*cc + ty)*nb21 + iv2*nb22 + iv3*nb23)); + device const vd4x4_t * pv4 = (device const vd4x4_t *) ((device const char *) v + ((ic + 4*cc + ty)*args.nb_12_1 + ikv2*args.nb_12_2 + ikv3*args.nb_12_3)); - const float4x4 lss(ss[4*cc + ty]); + const s4x4_t ms(ss[4*cc + ty]); -#pragma unroll - for (short ii = 0; ii < D16; ii += NW4) { + #pragma unroll(D16/NL) + for (short ii = 0; ii < D16; ii += NL) { const short i = ii + tx; - float4x4 mv; - dequantize_func(pv4 + i/nl, i%nl, mv); + v4x4_t mv; + deq_v(pv4 + i/nl_v, i%nl_v, mv); - lo[ii/NW4] += mv*lss; + lo[ii/NL] += mv*ms; } } } @@ -3459,8 +3801,8 @@ kernel void kernel_flash_attn_ext_vec( // these are needed for reducing the results from the simdgroups (reuse the ss buffer) if (tiisg == 0) { - ss[0] = S; - ss[1] = M; + ss[0] = (s_t) S; + ss[1] = (s_t) M; } } @@ -3473,23 +3815,37 @@ kernel void kernel_flash_attn_ext_vec( // [ 5, 13, 21, 29] -> [ 5] // [ 6, 14, 22, 30] -> [ 6] // [ 7, 15, 23, 31] -> [ 7] - for (short ii = 0; ii < D16; ii += NW4) { - lo[ii/NW4][0] += simd_shuffle_down(lo[ii/NW4][0], 16); - lo[ii/NW4][0] += simd_shuffle_down(lo[ii/NW4][0], 8); + for (short ii = 0; ii < D16; ii += NL) { + lo[ii/NL][0] += simd_shuffle_down(lo[ii/NL][0], 16); + lo[ii/NL][0] += simd_shuffle_down(lo[ii/NL][0], 8); + //lo[ii/NL][0] += simd_shuffle_down(lo[ii/NL][0], 4); + //lo[ii/NL][0] += simd_shuffle_down(lo[ii/NL][0], 2); + //lo[ii/NL][0] += simd_shuffle_down(lo[ii/NL][0], 1); - lo[ii/NW4][1] += simd_shuffle_down(lo[ii/NW4][1], 16); - lo[ii/NW4][1] += simd_shuffle_down(lo[ii/NW4][1], 8); + lo[ii/NL][1] += simd_shuffle_down(lo[ii/NL][1], 16); + lo[ii/NL][1] += simd_shuffle_down(lo[ii/NL][1], 8); + //lo[ii/NL][1] += simd_shuffle_down(lo[ii/NL][1], 4); + //lo[ii/NL][1] += simd_shuffle_down(lo[ii/NL][1], 2); + //lo[ii/NL][1] += simd_shuffle_down(lo[ii/NL][1], 1); - lo[ii/NW4][2] += simd_shuffle_down(lo[ii/NW4][2], 16); - lo[ii/NW4][2] += simd_shuffle_down(lo[ii/NW4][2], 8); + lo[ii/NL][2] += simd_shuffle_down(lo[ii/NL][2], 16); + lo[ii/NL][2] += simd_shuffle_down(lo[ii/NL][2], 8); + //lo[ii/NL][2] += simd_shuffle_down(lo[ii/NL][2], 4); + //lo[ii/NL][2] += simd_shuffle_down(lo[ii/NL][2], 2); + //lo[ii/NL][2] += simd_shuffle_down(lo[ii/NL][2], 1); - lo[ii/NW4][3] += simd_shuffle_down(lo[ii/NW4][3], 16); - lo[ii/NW4][3] += simd_shuffle_down(lo[ii/NW4][3], 8); + lo[ii/NL][3] += simd_shuffle_down(lo[ii/NL][3], 16); + lo[ii/NL][3] += simd_shuffle_down(lo[ii/NL][3], 8); + //lo[ii/NL][3] += simd_shuffle_down(lo[ii/NL][3], 4); + //lo[ii/NL][3] += simd_shuffle_down(lo[ii/NL][3], 2); + //lo[ii/NL][3] += simd_shuffle_down(lo[ii/NL][3], 1); } + threadgroup_barrier(mem_flags::mem_threadgroup); + // store results to shared memory - for (short i = tiisg; i < D16; i += NW4) { - sr44[i] = lo[i/NW4]; + for (short i = tiisg; i < D16; i += NL) { + sr4x4[i] = lo[i/NL]; } threadgroup_barrier(mem_flags::mem_threadgroup); @@ -3497,18 +3853,18 @@ kernel void kernel_flash_attn_ext_vec( // parallel reduce for (short r = nsg/2; r > 0; r >>= 1) { if (sgitg < r) { - const float S0 = ss[ 0]; - const float S1 = ss[r*SH + 0]; + const half S0 = ss[ 0]; + const half S1 = ss[r*SH + 0]; - const float M0 = ss[ 1]; - const float M1 = ss[r*SH + 1]; + const half M0 = ss[ 1]; + const half M1 = ss[r*SH + 1]; - const float M = max(M0, M1); + const half M = max(M0, M1); - const float ms0 = exp(M0 - M); - const float ms1 = exp(M1 - M); + const half ms0 = exp(M0 - M); + const half ms1 = exp(M1 - M); - const float S = S0*ms0 + S1*ms1; + const half S = S0*ms0 + S1*ms1; if (tiisg == 0) { ss[0] = S; @@ -3517,7 +3873,7 @@ kernel void kernel_flash_attn_ext_vec( // O_0 = diag(ms0)*O_0 + diag(ms1)*O_1 for (short i = tiisg; i < D16; i += NW) { - sr44[i] = sr44[i]*ms0 + sr44[i + r*D16]*ms1; + sr4x4[i] = sr4x4[i]*ms0 + sr4x4[i + r*D16]*ms1; } } @@ -3531,65 +3887,101 @@ kernel void kernel_flash_attn_ext_vec( const float S = ss[0]; for (short i = tiisg; i < D16; i += NW) { - dst44[(iq3*ne2*ne1 + iq2 + (iq1)*ne1)*D16 + i] = sr44[i]/S; + dst44[((uint64_t)iq3*args.ne2*args.ne1 + iq2 + (uint64_t)iq1*args.ne1)*D16 + i] = (float4x4) sr4x4[i]/S; } } } -typedef decltype(kernel_flash_attn_ext_vec) flash_attn_ext_vec_t; +// note: I think the s_t can be half instead of float, because the Q*K scaling is done before storing to shared mem +// in the other (non-vec) kernel, we need s_t to also be float because we scale during the soft_max +// +#define FA_TYPES \ + half4, half4x4, \ + half4x4, \ + half4x4, \ + float, \ + half, half4, half4x4, \ + half4x4 -template [[host_name("kernel_flash_attn_ext_vec_f16_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q4_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q4_1_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q5_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q5_1_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q8_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +typedef decltype(kernel_flash_attn_ext_vec) flash_attn_ext_vec_t; -template [[host_name("kernel_flash_attn_ext_vec_f16_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q4_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q4_1_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q5_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q5_1_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -template [[host_name("kernel_flash_attn_ext_vec_q8_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_f16_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +#if defined(GGML_METAL_USE_BF16) +template [[host_name("kernel_flash_attn_ext_vec_bf16_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +#endif +template [[host_name("kernel_flash_attn_ext_vec_q4_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q4_1_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q5_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q5_1_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q8_0_h128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; + +template [[host_name("kernel_flash_attn_ext_vec_f16_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +#if defined(GGML_METAL_USE_BF16) +template [[host_name("kernel_flash_attn_ext_vec_bf16_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +#endif +template [[host_name("kernel_flash_attn_ext_vec_q4_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q4_1_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q5_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q5_1_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; +template [[host_name("kernel_flash_attn_ext_vec_q8_0_h256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; + +#undef FA_TYPES + +template +kernel void kernel_set( + constant ggml_metal_kargs_set & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i13 = tgpig[2]; + const int i12 = tgpig[1]; + const int i11 = tgpig[0]; + + const int64_t n = i13*args.ne12*args.ne11*args.ne10 + i12*args.ne11*args.ne10 + i11*args.ne10; + + const int64_t i3 = n / (args.ne12*args.ne11*args.ne10); + const int64_t i2 = (n - i3*args.ne12*args.ne11*args.ne10) / (args.ne11*args.ne10); + const int64_t i1 = (n - i3*args.ne12*args.ne11*args.ne10 - i2*args.ne11*args.ne10) / args.ne10; + + device T * dst_data = (device T *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + args.offs); + + for (int64_t i10 = tpitg.x; i10 < args.ne10; i10 += ntg.x) { + device const T * src = (device T *) (src1 + i13*args.nb13 + i12*args.nb12 + i11*args.nb11 + i10*args.nb10); + dst_data[i10] = (T) src[0]; + } +} + +typedef decltype(kernel_set) kernel_set_t; + +template [[host_name("kernel_set_f32")]] kernel kernel_set_t kernel_set; +template [[host_name("kernel_set_i32")]] kernel kernel_set_t kernel_set; template kernel void kernel_cpy( - device const void * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0); + const int64_t i3 = n/(args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0)/(args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0)/args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0); - device T1 * dst_data = (device T1 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device T1 * dst_data = (device T1 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) { - device const T0 * src = (device T0 *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x; i00 < args.ne00; i00 += ntg.x) { + device const T0 * src = (device T0 *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); dst_data[i00] = (T1) src[0]; } } @@ -3598,53 +3990,38 @@ typedef decltype(kernel_cpy) kernel_cpy_t; template [[host_name("kernel_cpy_f32_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_f32_f16")]] kernel kernel_cpy_t kernel_cpy; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_cpy_f32_bf16")]] kernel kernel_cpy_t kernel_cpy; #endif template [[host_name("kernel_cpy_f16_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_f16_f16")]] kernel kernel_cpy_t kernel_cpy; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_cpy_bf16_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_bf16_bf16")]] kernel kernel_cpy_t kernel_cpy; #endif kernel void kernel_cpy_f32_q8_0( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK8_0; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK8_0; - device block_q8_0 * dst_data = (device block_q8_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_q8_0 * dst_data = (device block_q8_0 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK8_0; i00 < ne00; i00 += ntg.x*QK8_0) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK8_0; i00 < args.ne00; i00 += ntg.x*QK8_0) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float amax = 0.0f; // absolute max @@ -3667,42 +4044,27 @@ kernel void kernel_cpy_f32_q8_0( } kernel void kernel_cpy_f32_q4_0( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_0; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK4_0; - device block_q4_0 * dst_data = (device block_q4_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_q4_0 * dst_data = (device block_q4_0 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK4_0; i00 < ne00; i00 += ntg.x*QK4_0) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK4_0; i00 < args.ne00; i00 += ntg.x*QK4_0) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float amax = 0.0f; // absolute max float max = 0.0f; @@ -3734,42 +4096,27 @@ kernel void kernel_cpy_f32_q4_0( } kernel void kernel_cpy_f32_q4_1( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_1; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK4_1; - device block_q4_1 * dst_data = (device block_q4_1 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_q4_1 * dst_data = (device block_q4_1 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK4_1; i00 < ne00; i00 += ntg.x*QK4_1) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK4_1; i00 < args.ne00; i00 += ntg.x*QK4_1) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float min = FLT_MAX; float max = -FLT_MAX; @@ -3800,42 +4147,27 @@ kernel void kernel_cpy_f32_q4_1( } kernel void kernel_cpy_f32_q5_0( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK5_0; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK5_0; - device block_q5_0 * dst_data = (device block_q5_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_q5_0 * dst_data = (device block_q5_0 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK5_0; i00 < ne00; i00 += ntg.x*QK5_0) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK5_0; i00 < args.ne00; i00 += ntg.x*QK5_0) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float amax = 0.0f; // absolute max float max = 0.0f; @@ -3873,42 +4205,27 @@ kernel void kernel_cpy_f32_q5_0( } kernel void kernel_cpy_f32_q5_1( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK5_1; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK5_1; - device block_q5_1 * dst_data = (device block_q5_1 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_q5_1 * dst_data = (device block_q5_1 *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK5_1; i00 < ne00; i00 += ntg.x*QK5_1) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK5_1; i00 < args.ne00; i00 += ntg.x*QK5_1) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float max = src[0]; float min = src[0]; @@ -3956,42 +4273,27 @@ static inline int best_index_int8(int n, constant float * val, float x) { } kernel void kernel_cpy_f32_iq4_nl( - device const float * src0, - device void * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { - const int64_t i03 = tgpig[2]; - const int64_t i02 = tgpig[1]; - const int64_t i01 = tgpig[0]; + constant ggml_metal_kargs_cpy & args, + device const char * src0, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { + const int i03 = tgpig[2]; + const int i02 = tgpig[1]; + const int i01 = tgpig[0]; - const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + const int64_t n = i03*args.ne02*args.ne01*args.ne00 + i02*args.ne01*args.ne00 + i01*args.ne00; - const int64_t i3 = n / (ne2*ne1*ne0); - const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); - const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; - const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_NL; + const int64_t i3 = n / (args.ne2*args.ne1*args.ne0); + const int64_t i2 = (n - i3*args.ne2*args.ne1*args.ne0) / (args.ne1*args.ne0); + const int64_t i1 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0) / args.ne0; + const int64_t i0 = (n - i3*args.ne2*args.ne1*args.ne0 - i2*args.ne1*args.ne0 - i1*args.ne0)/QK4_NL; - device block_iq4_nl * dst_data = (device block_iq4_nl *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device block_iq4_nl * dst_data = (device block_iq4_nl *) (dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); - for (int64_t i00 = tpitg.x*QK4_NL; i00 < ne00; i00 += ntg.x*QK4_NL) { - device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + for (int64_t i00 = tpitg.x*QK4_NL; i00 < args.ne00; i00 += ntg.x*QK4_NL) { + device const float * src = (device float *)(src0 + i03*args.nb03 + i02*args.nb02 + i01*args.nb01 + i00*args.nb00); float amax = 0.0f; // absolute max float max = 0.0f; @@ -4027,104 +4329,66 @@ kernel void kernel_cpy_f32_iq4_nl( } dst_data[i00/QK4_NL].d = sumq2 > 0 ? sumqx/sumq2 : d; - } } kernel void kernel_concat( + constant ggml_metal_kargs_concat & args, device const char * src0, device const char * src1, device char * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant int64_t & ne03, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant int64_t & ne2, - constant int64_t & ne3, - constant uint64_t & nb0, - constant uint64_t & nb1, - constant uint64_t & nb2, - constant uint64_t & nb3, - constant int32_t & dim, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { + uint3 tgpig[[threadgroup_position_in_grid]], + ushort3 tpitg[[thread_position_in_threadgroup]], + ushort3 ntg[[threads_per_threadgroup]]) { - const int64_t i3 = tgpig.z; - const int64_t i2 = tgpig.y; - const int64_t i1 = tgpig.x; + const int i3 = tgpig.z; + const int i2 = tgpig.y; + const int i1 = tgpig.x; - int64_t o[4] = {0, 0, 0, 0}; - o[dim] = dim == 0 ? ne00 : (dim == 1 ? ne01 : (dim == 2 ? ne02 : ne03)); + int o[4] = {0, 0, 0, 0}; + o[args.dim] = args.dim == 0 ? args.ne00 : (args.dim == 1 ? args.ne01 : (args.dim == 2 ? args.ne02 : args.ne03)); device const float * x; - for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) { - if (i0 < ne00 && i1 < ne01 && i2 < ne02 && i3 < ne03) { - x = (device const float *)(src0 + (i3 )*nb03 + (i2 )*nb02 + (i1 )*nb01 + (i0 )*nb00); + for (int i0 = tpitg.x; i0 < args.ne0; i0 += ntg.x) { + if (i0 < args.ne00 && i1 < args.ne01 && i2 < args.ne02 && i3 < args.ne03) { + x = (device const float *)(src0 + (i3 )*args.nb03 + (i2 )*args.nb02 + (i1 )*args.nb01 + (i0 )*args.nb00); } else { - x = (device const float *)(src1 + (i3 - o[3])*nb13 + (i2 - o[2])*nb12 + (i1 - o[1])*nb11 + (i0 - o[0])*nb10); + x = (device const float *)(src1 + (i3 - o[3])*args.nb13 + (i2 - o[2])*args.nb12 + (i1 - o[1])*args.nb11 + (i0 - o[0])*args.nb10); } - device float * y = (device float *)(dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + device float * y = (device float *)(dst + i3*args.nb3 + i2*args.nb2 + i1*args.nb1 + i0*args.nb0); *y = *x; } } +template void kernel_mul_mv_q2_K_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_q2_K * x = (device const block_q2_K *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_q2_K * x = (device const block_q2_K *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; @@ -4173,92 +4437,64 @@ void kernel_mul_mv_q2_K_f32_impl( (acc1[3] + 1.f/256.f * acc2[3]) * (sc[6] & 0xF) * 1.f/64.f) - dmin * (sumy[0] * (sc[0] & 0xF0) + sumy[1] * (sc[2] & 0xF0) + sumy[2] * (sc[4] & 0xF0) + sumy[3] * (sc[6] & 0xF0)); - qs += nb01/2; - sc += nb01; - dh += nb01/2; + qs += args.nb01/2; + sc += args.nb01; + dh += args.nb01/2; } y4 += 4 * QK_K; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } [[host_name("kernel_mul_mv_q2_K_f32")]] kernel void kernel_mul_mv_q2_K_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q2_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q2_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_q3_K_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; - const int64_t r0 = tgpig.x; - const int64_t r1 = tgpig.y; - const int64_t im = tgpig.z; + const int r0 = tgpig.x; + const int r1 = tgpig.y; + const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * 2; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_q3_K * x = (device const block_q3_K *) ((device char *) src0 + offset0); - device const float * yy = (device const float *) ((device char *) src1 + offset1); + device const block_q3_K * x = (device const block_q3_K *) (src0 + offset0); + device const float * yy = (device const float *) (src1 + offset1); float yl[32]; @@ -4288,9 +4524,10 @@ void kernel_mul_mv_q3_K_f32_impl( const ushort4 hm = mm[2*ip + il/2]; - const int shift = 2*il; - const float v1 = il == 0 ? 4.f : 64.f; - const float v2 = 4.f * v1; + const short shift = 2*il; + + const float v1 = il == 0 ? 4.f : 64.f; + const float v2 = 4.f * v1; const uint16_t s_shift1 = 4*ip; const uint16_t s_shift2 = s_shift1 + il; @@ -4359,10 +4596,10 @@ void kernel_mul_mv_q3_K_f32_impl( sumf1[row] += d1 * (scales[1] - 32); sumf2[row] += d2 * (scales[3] - 32); - q += nb01/2; - h += nb01/2; - a += nb01/2; - dh += nb01/2; + q += args.nb01/2; + h += args.nb01/2; + a += args.nb01/2; + dh += args.nb01/2; } y1 += 4 * QK_K; @@ -4372,66 +4609,39 @@ void kernel_mul_mv_q3_K_f32_impl( const float sumf = (sumf1[row] + 0.25f * sumf2[row]) / (1 << shift); sumf1[row] = simd_sum(sumf); } + + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + if (tiisg == 0) { for (int row = 0; row < 2; ++row) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = sumf1[row]; + dst_f32[first_row + row] = sumf1[row]; } } } [[host_name("kernel_mul_mv_q3_K_f32")]] kernel void kernel_mul_mv_q3_K_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q3_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q3_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_q4_K_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { const uint16_t kmask1 = 0x3f3f; const uint16_t kmask2 = 0x0f0f; @@ -4442,21 +4652,21 @@ void kernel_mul_mv_q4_K_f32_impl( const int iq = it/4; // 0 or 1 const int ir = it%4; // 0...3 - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; //const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; const int first_row = r0 * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_q4_K * x = (device const block_q4_K *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_q4_K * x = (device const block_q4_K *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[16]; float yh[16]; @@ -4509,92 +4719,64 @@ void kernel_mul_mv_q4_K_f32_impl( (acc2[2] + 1.f/256.f * acc2[3]) * sc8[5] * 1.f/16.f) - dmin * (sumy[0] * sc8[2] + sumy[1] * sc8[3] + sumy[2] * sc8[6] + sumy[3] * sc8[7]); - q1 += nb01/2; - sc += nb01/2; - dh += nb01/2; + q1 += args.nb01/2; + sc += args.nb01/2; + dh += args.nb01/2; } y4 += 4 * QK_K; } + device float * dst_f32 = (device float *) dst + (int64_t)im*args.ne0*args.ne1 + (int64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } [[host_name("kernel_mul_mv_q4_K_f32")]] kernel void kernel_mul_mv_q4_K_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q4_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q4_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_q5_K_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; - const int64_t r0 = tgpig.x; - const int64_t r1 = tgpig.y; + const int r0 = tgpig.x; + const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * 2; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_q5_K * x = (device const block_q5_K *) ((device char *) src0 + offset0); - device const float * yy = (device const float *) ((device char *) src1 + offset1); + device const block_q5_K * x = (device const block_q5_K *) (src0 + offset0); + device const float * yy = (device const float *) (src1 + offset1); float sumf[2]={0.f}; @@ -4668,98 +4850,70 @@ void kernel_mul_mv_q5_K_f32_impl( sc8[5] * (acc1[3]/16.f + 16.f*acc2[3])) - dmin * (sumy[0] * sc8[2] + sumy[1] * sc8[3] + sumy[2] * sc8[6] + sumy[3] * sc8[7]); - q1 += nb01; - qh += nb01; - dh += nb01/2; - a += nb01/2; + q1 += args.nb01; + qh += args.nb01; + dh += args.nb01/2; + a += args.nb01/2; } y1 += 4 * QK_K; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < 2; ++row) { const float tot = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot; + dst_f32[first_row + row] = tot; } } } [[host_name("kernel_mul_mv_q5_K_f32")]] kernel void kernel_mul_mv_q5_K_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q5_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q5_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_q6_K_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { const uint8_t kmask1 = 0x03; const uint8_t kmask2 = 0x0C; const uint8_t kmask3 = 0x30; const uint8_t kmask4 = 0xC0; - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; - const int64_t r0 = tgpig.x; - const int64_t r1 = tgpig.y; - const int im = tgpig.z; + const int r0 = tgpig.x; + const int r1 = tgpig.y; + const int im = tgpig.z; - const int row = 2 * r0 + sgitg; + const int row = 2*r0 + sgitg; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_q6_K * x = (device const block_q6_K *) ((device char *) src0 + offset0); - device const float * yy = (device const float *) ((device char *) src1 + offset1); + device const block_q6_K * x = (device const block_q6_K *) (src0 + offset0); + device const float * yy = (device const float *) (src1 + offset1); float sumf = 0; @@ -4776,7 +4930,6 @@ void kernel_mul_mv_q6_K_f32_impl( const int q_offset_h = 32*ip + l0; for (int i = ix; i < nb; i += 2) { - device const uint8_t * q1 = x[i].ql + q_offset_l; device const uint8_t * q2 = q1 + 32; device const uint8_t * qh = x[i].qh + q_offset_h; @@ -4798,98 +4951,70 @@ void kernel_mul_mv_q6_K_f32_impl( } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + const float tot = simd_sum(sumf); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + row] = tot; + dst_f32[row] = tot; } } [[host_name("kernel_mul_mv_q6_K_f32")]] kernel void kernel_mul_mv_q6_K_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q6_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q6_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } // ======================= "True" 2-bit +template void kernel_mul_mv_iq2_xxs_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq2_xxs * x = (device const block_iq2_xxs *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq2_xxs * x = (device const block_iq2_xxs *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; const int nb32 = nb * (QK_K / 32); - threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values; - threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 256); + threadgroup uint64_t * svalues = (threadgroup uint64_t *)(shmem); + threadgroup uint8_t * ssigns = (threadgroup uint8_t *)(svalues + 256); { int nval = 4; int pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) values[pos + i] = iq2xxs_grid[pos + i]; + for (int i = 0; i < nval; ++i) svalues[pos + i] = iq2xxs_grid[pos + i]; nval = 2; pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i]; + for (int i = 0; i < nval; ++i) ssigns[pos+i] = ksigns_iq2xs[pos+i]; threadgroup_barrier(mem_flags::mem_threadgroup); } @@ -4919,114 +5044,85 @@ void kernel_mul_mv_iq2_xxs_f32_impl( float sum = 0; for (int l = 0; l < 4; ++l) { - const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + aux8[l]); - const uint8_t signs = shared_signs[(aux32 >> 7*l) & 127]; + const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(svalues + aux8[l]); + const uint8_t signs = ssigns[(aux32 >> 7*l) & 127]; for (int j = 0; j < 8; ++j) { sum += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f); } } sumf[row] += d * sum; - dh += nb01/2; - q2 += nb01/2; + dh += args.nb01/2; + q2 += args.nb01/2; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f; + dst_f32[first_row + row] = all_sum * 0.25f; } } } [[host_name("kernel_mul_mv_iq2_xxs_f32")]] kernel void kernel_mul_mv_iq2_xxs_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - - kernel_mul_mv_iq2_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + kernel_mul_mv_iq2_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_iq2_xs_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq2_xs * x = (device const block_iq2_xs *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq2_xs * x = (device const block_iq2_xs *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; const int nb32 = nb * (QK_K / 32); - threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values; - threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 512); + threadgroup uint64_t * svalues = (threadgroup uint64_t *)(shmem); + threadgroup uint8_t * ssigns = (threadgroup uint8_t *)(svalues + 512); { int nval = 8; int pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) values[pos + i] = iq2xs_grid[pos + i]; + for (int i = 0; i < nval; ++i) svalues[pos + i] = iq2xs_grid[pos + i]; nval = 2; pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i]; + for (int i = 0; i < nval; ++i) ssigns[pos+i] = ksigns_iq2xs[pos+i]; threadgroup_barrier(mem_flags::mem_threadgroup); } @@ -5058,122 +5154,94 @@ void kernel_mul_mv_iq2_xs_f32_impl( float sum1 = 0, sum2 = 0; for (int l = 0; l < 2; ++l) { - const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511)); - const uint8_t signs = shared_signs[(q2[l] >> 9)]; + const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(svalues + (q2[l] & 511)); + const uint8_t signs = ssigns[(q2[l] >> 9)]; for (int j = 0; j < 8; ++j) { sum1 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f); } } for (int l = 2; l < 4; ++l) { - const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511)); - const uint8_t signs = shared_signs[(q2[l] >> 9)]; + const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(svalues + (q2[l] & 511)); + const uint8_t signs = ssigns[(q2[l] >> 9)]; for (int j = 0; j < 8; ++j) { sum2 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f); } } sumf[row] += d1 * sum1 + d2 * sum2; - dh += nb01/2; - q2 += nb01/2; - sc += nb01; + dh += args.nb01/2; + q2 += args.nb01/2; + sc += args.nb01; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f; + dst_f32[first_row + row] = all_sum * 0.25f; } } } [[host_name("kernel_mul_mv_iq2_xs_f32")]] kernel void kernel_mul_mv_iq2_xs_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq2_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq2_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_iq3_xxs_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq3_xxs * x = (device const block_iq3_xxs *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq3_xxs * x = (device const block_iq3_xxs *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; const int nb32 = nb * (QK_K / 32); - threadgroup uint32_t * values = (threadgroup uint32_t *)shared_values; - threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 256); + threadgroup uint32_t * svalues = (threadgroup uint32_t *)(shmem); + threadgroup uint8_t * ssigns = (threadgroup uint8_t *)(svalues + 256); { int nval = 4; int pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) values[pos + i] = iq3xxs_grid[pos + i]; + for (int i = 0; i < nval; ++i) svalues[pos + i] = iq3xxs_grid[pos + i]; nval = 2; pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i]; + for (int i = 0; i < nval; ++i) ssigns[pos+i] = ksigns_iq2xs[pos+i]; threadgroup_barrier(mem_flags::mem_threadgroup); } @@ -5182,7 +5250,6 @@ void kernel_mul_mv_iq3_xxs_f32_impl( device const float * y4 = y + 32 * ix; for (int ib32 = ix; ib32 < nb32; ib32 += 32) { - for (int i = 0; i < 32; ++i) { yl[i] = y4[i]; } @@ -5196,16 +5263,15 @@ void kernel_mul_mv_iq3_xxs_f32_impl( device const half * dh = &xr->d; for (int row = 0; row < N_DST; row++) { - const float db = dh[0]; const uint32_t aux32 = gas[0] | (gas[1] << 16); const float d = db * (0.5f + (aux32 >> 28)); float2 sum = {0}; for (int l = 0; l < 4; ++l) { - const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(values + q3[2*l+0]); - const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(values + q3[2*l+1]); - const uint8_t signs = shared_signs[(aux32 >> 7*l) & 127]; + const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(svalues + q3[2*l+0]); + const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(svalues + q3[2*l+1]); + const uint8_t signs = ssigns[(aux32 >> 7*l) & 127]; for (int j = 0; j < 4; ++j) { sum[0] += yl[8*l + j + 0] * grid1[j] * (signs & kmask_iq2xs[j+0] ? -1.f : 1.f); sum[1] += yl[8*l + j + 4] * grid2[j] * (signs & kmask_iq2xs[j+4] ? -1.f : 1.f); @@ -5213,103 +5279,75 @@ void kernel_mul_mv_iq3_xxs_f32_impl( } sumf[row] += d * (sum[0] + sum[1]); - dh += nb01/2; - q3 += nb01; - gas += nb01/2; + dh += args.nb01/2; + q3 += args.nb01; + gas += args.nb01/2; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.5f; + dst_f32[first_row + row] = all_sum * 0.5f; } } } [[host_name("kernel_mul_mv_iq3_xxs_f32")]] kernel void kernel_mul_mv_iq3_xxs_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq3_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq3_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_iq3_s_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq3_s * x = (device const block_iq3_s *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq3_s * x = (device const block_iq3_s *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; const int nb32 = nb * (QK_K / 32); - threadgroup uint32_t * values = (threadgroup uint32_t *)shared_values; + threadgroup uint32_t * svalues = (threadgroup uint32_t *) shmem; { int nval = 8; int pos = (32*sgitg + tiisg)*nval; - for (int i = 0; i < nval; ++i) values[pos + i] = iq3s_grid[pos + i]; + for (int i = 0; i < nval; ++i) svalues[pos + i] = iq3s_grid[pos + i]; threadgroup_barrier(mem_flags::mem_threadgroup); } @@ -5340,8 +5378,8 @@ void kernel_mul_mv_iq3_s_f32_impl( float2 sum = {0}; for (int l = 0; l < 4; ++l) { - const threadgroup uint32_t * table1 = qh[0] & kmask_iq2xs[2*l+0] ? values + 256 : values; - const threadgroup uint32_t * table2 = qh[0] & kmask_iq2xs[2*l+1] ? values + 256 : values; + const threadgroup uint32_t * table1 = qh[0] & kmask_iq2xs[2*l+0] ? svalues + 256 : svalues; + const threadgroup uint32_t * table2 = qh[0] & kmask_iq2xs[2*l+1] ? svalues + 256 : svalues; const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(table1 + qs[2*l+0]); const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(table2 + qs[2*l+1]); for (int j = 0; j < 4; ++j) { @@ -5351,105 +5389,77 @@ void kernel_mul_mv_iq3_s_f32_impl( } sumf[row] += d * (sum[0] + sum[1]); - dh += nb01/2; - qs += nb01; - qh += nb01; - sc += nb01; - signs += nb01; + dh += args.nb01/2; + qs += args.nb01; + qh += args.nb01; + sc += args.nb01; + signs += args.nb01; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } [[host_name("kernel_mul_mv_iq3_s_f32")]] kernel void kernel_mul_mv_iq3_s_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq3_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq3_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_iq2_s_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq2_s * x = (device const block_iq2_s *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq2_s * x = (device const block_iq2_s *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; const int nb32 = nb * (QK_K / 32); - //threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values; + //threadgroup uint64_t * svalues = (threadgroup uint64_t *) shmem; //{ // int nval = 32; // int pos = (32*sgitg + tiisg)*nval; - // for (int i = 0; i < nval; ++i) values[pos + i] = iq2s_grid[pos + i]; + // for (int i = 0; i < nval; ++i) svalues[pos + i] = iq2s_grid[pos + i]; // threadgroup_barrier(mem_flags::mem_threadgroup); //} @@ -5481,8 +5491,8 @@ void kernel_mul_mv_iq2_s_f32_impl( float2 sum = {0}; for (int l = 0; l < 2; ++l) { - //const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(values + (qs[l+0] | ((qh[0] << (8-2*l)) & 0x300))); - //const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(values + (qs[l+2] | ((qh[0] << (4-2*l)) & 0x300))); + //const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(svalues + (qs[l+0] | ((qh[0] << (8-2*l)) & 0x300))); + //const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(svalues + (qs[l+2] | ((qh[0] << (4-2*l)) & 0x300))); constant uint8_t * grid1 = (constant uint8_t *)(iq2s_grid + (qs[l+0] | ((qh[0] << (8-2*l)) & 0x300))); constant uint8_t * grid2 = (constant uint8_t *)(iq2s_grid + (qs[l+2] | ((qh[0] << (4-2*l)) & 0x300))); for (int j = 0; j < 8; ++j) { @@ -5492,94 +5502,66 @@ void kernel_mul_mv_iq2_s_f32_impl( } sumf[row] += d1 * sum[0] + d2 * sum[1]; - dh += nb01/2; - qs += nb01; - qh += nb01; - sc += nb01; - signs += nb01; + dh += args.nb01/2; + qs += args.nb01; + qh += args.nb01; + sc += args.nb01; + signs += args.nb01; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f; + dst_f32[first_row + row] = all_sum * 0.25f; } } } [[host_name("kernel_mul_mv_iq2_s_f32")]] kernel void kernel_mul_mv_iq2_s_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq2_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq2_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } +template void kernel_mul_mv_iq1_s_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_value, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq1_s * x = (device const block_iq1_s *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq1_s * x = (device const block_iq1_s *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; @@ -5622,61 +5604,50 @@ void kernel_mul_mv_iq1_s_f32_impl( } sumf[row] += (float)dh[0] * (sum + sumy * (qh[0] & 0x8000 ? -1 - IQ1S_DELTA : -1 + IQ1S_DELTA)) * (2*((qh[0] >> 12) & 7) + 1); - dh += nb01/2; - qs += nb01; - qh += nb01/2; + dh += args.nb01/2; + qs += args.nb01; + qh += args.nb01/2; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } +template void kernel_mul_mv_iq1_m_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_value, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - const int nb = ne00/QK_K; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq1_m * x = (device const block_iq1_m *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq1_m * x = (device const block_iq1_m *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); float yl[32]; float sumf[N_DST]={0.f}, all_sum; @@ -5728,66 +5699,55 @@ void kernel_mul_mv_iq1_m_f32_impl( sumf[row] += (float)scale.f16 * ((sum[0] + delta1) * (2*((sc[ib/2] >> (6*(ib%2)+0)) & 7) + 1) + (sum[1] + delta2) * (2*((sc[ib/2] >> (6*(ib%2)+3)) & 7) + 1)); - sc += nb01/2; - qs += nb01; - qh += nb01; + sc += args.nb01/2; + qs += args.nb01; + qh += args.nb01; } y4 += 32 * 32; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < N_DST; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } +template void kernel_mul_mv_iq4_nl_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values_i8, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - threadgroup float * shared_values = (threadgroup float *)shared_values_i8; - const int nb = ne00/QK4_NL; + threadgroup float * shmem_f32 = (threadgroup float *) shmem; + const int nb = args.ne00/QK4_NL; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * 2 + sgitg) * 2; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq4_nl * x = (device const block_iq4_nl *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq4_nl * x = (device const block_iq4_nl *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); const int ix = tiisg/2; // 0...15 const int it = tiisg%2; // 0 or 1 - shared_values[tiisg] = kvalues_iq4nl_f[tiisg%16]; + shmem_f32[tiisg] = kvalues_iq4nl_f[tiisg%16]; threadgroup_barrier(mem_flags::mem_threadgroup); float4 yl[4]; @@ -5805,7 +5765,7 @@ void kernel_mul_mv_iq4_nl_f32_impl( device const float4 * y4 = (device const float4 *)yb; yl[0] = y4[0]; yl[1] = y4[4]; yl[2] = y4[1]; yl[3] = y4[5]; - for (int row = 0; row < 2 && first_row + row < ne01; ++row) { + for (int row = 0; row < 2 && first_row + row < args.ne01; ++row) { device const block_iq4_nl & xb = x[row*nb + ib]; device const uint16_t * q4 = (device const uint16_t *)(xb.qs + 8*it); @@ -5815,16 +5775,16 @@ void kernel_mul_mv_iq4_nl_f32_impl( aux32[0] = q4[0] | (q4[1] << 16); aux32[1] = (aux32[0] >> 4) & 0x0f0f0f0f; aux32[0] &= 0x0f0f0f0f; - qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]}; - qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]}; + qf1 = {shmem_f32[q8[0]], shmem_f32[q8[1]], shmem_f32[q8[2]], shmem_f32[q8[3]]}; + qf2 = {shmem_f32[q8[4]], shmem_f32[q8[5]], shmem_f32[q8[6]], shmem_f32[q8[7]]}; acc1 += yl[0] * qf1; acc2 += yl[1] * qf2; aux32[0] = q4[2] | (q4[3] << 16); aux32[1] = (aux32[0] >> 4) & 0x0f0f0f0f; aux32[0] &= 0x0f0f0f0f; - qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]}; - qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]}; + qf1 = {shmem_f32[q8[0]], shmem_f32[q8[1]], shmem_f32[q8[2]], shmem_f32[q8[3]]}; + qf2 = {shmem_f32[q8[4]], shmem_f32[q8[5]], shmem_f32[q8[6]], shmem_f32[q8[7]]}; acc1 += yl[2] * qf1; acc2 += yl[3] * qf2; @@ -5837,60 +5797,49 @@ void kernel_mul_mv_iq4_nl_f32_impl( yb += 16 * QK4_NL; } - for (int row = 0; row < 2 && first_row + row < ne01; ++row) { + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + + for (int row = 0; row < 2 && first_row + row < args.ne01; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } +template void kernel_mul_mv_iq4_xs_f32_impl( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values_i8, - uint3 tgpig, - uint tiisg, - uint sgitg) { + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { - threadgroup float * shared_values = (threadgroup float *)shared_values_i8; - const int nb = ne00/QK_K; + threadgroup float * shmem_f32 = (threadgroup float *) shmem; + const int nb = args.ne00/QK_K; const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; const int first_row = (r0 * 2 + sgitg) * 2; - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; - const uint offset0 = first_row*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; - const uint offset1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + const uint64_t offset0 = first_row*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const block_iq4_xs * x = (device const block_iq4_xs *) ((device char *) src0 + offset0); - device const float * y = (device const float *) ((device char *) src1 + offset1); + device const block_iq4_xs * x = (device const block_iq4_xs *) (src0 + offset0); + device const float * y = (device const float *) (src1 + offset1); const int ix = tiisg/16; // 0 or 1 const int it = tiisg%16; // 0...15 const int ib = it/2; const int il = it%2; - shared_values[tiisg] = kvalues_iq4nl_f[tiisg%16]; + shmem_f32[tiisg] = kvalues_iq4nl_f[tiisg%16]; threadgroup_barrier(mem_flags::mem_threadgroup); float4 yl[4]; @@ -5904,28 +5853,26 @@ void kernel_mul_mv_iq4_xs_f32_impl( float4 qf1, qf2; for (int ibl = ix; ibl < nb; ibl += 2) { - device const float4 * y4 = (device const float4 *)yb; yl[0] = y4[0]; yl[1] = y4[4]; yl[2] = y4[1]; yl[3] = y4[5]; for (int row = 0; row < 2; ++row) { - device const block_iq4_xs & xb = x[row*nb + ibl]; device const uint32_t * q4 = (device const uint32_t *)(xb.qs + 16*ib + 8*il); float4 acc1 = {0.f}, acc2 = {0.f}; - aux32[0] = q4[0] & 0x0f0f0f0f; + aux32[0] = (q4[0] ) & 0x0f0f0f0f; aux32[1] = (q4[0] >> 4) & 0x0f0f0f0f; - qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]}; - qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]}; + qf1 = {shmem_f32[q8[0]], shmem_f32[q8[1]], shmem_f32[q8[2]], shmem_f32[q8[3]]}; + qf2 = {shmem_f32[q8[4]], shmem_f32[q8[5]], shmem_f32[q8[6]], shmem_f32[q8[7]]}; acc1 += yl[0] * qf1; acc2 += yl[1] * qf2; - aux32[0] = q4[1] & 0x0f0f0f0f; + aux32[0] = (q4[1] ) & 0x0f0f0f0f; aux32[1] = (q4[1] >> 4) & 0x0f0f0f0f; - qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]}; - qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]}; + qf1 = {shmem_f32[q8[0]], shmem_f32[q8[1]], shmem_f32[q8[2]], shmem_f32[q8[3]]}; + qf2 = {shmem_f32[q8[4]], shmem_f32[q8[5]], shmem_f32[q8[6]], shmem_f32[q8[7]]}; acc1 += yl[2] * qf1; acc2 += yl[3] * qf2; @@ -5939,134 +5886,68 @@ void kernel_mul_mv_iq4_xs_f32_impl( yb += 2 * QK_K; } + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + for (int row = 0; row < 2; ++row) { all_sum = simd_sum(sumf[row]); if (tiisg == 0) { - dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum; + dst_f32[first_row + row] = all_sum; } } } [[host_name("kernel_mul_mv_iq1_s_f32")]] kernel void kernel_mul_mv_iq1_s_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq1_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_iq1_s_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } [[host_name("kernel_mul_mv_iq1_m_f32")]] kernel void kernel_mul_mv_iq1_m_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq1_m_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_iq1_m_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } [[host_name("kernel_mul_mv_iq4_nl_f32")]] kernel void kernel_mul_mv_iq4_nl_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq4_nl_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq4_nl_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } [[host_name("kernel_mul_mv_iq4_xs_f32")]] kernel void kernel_mul_mv_iq4_xs_f32( - device const void * src0, - device const float * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq4_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb01, nb02, nb03, ne10, ne12, nb11, nb12, nb13, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg); + kernel_mul_mv_iq4_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } template @@ -6170,134 +6051,141 @@ kernel void kernel_get_rows_i32( // each block_q contains 16*nl weights template -kernel void kernel_mul_mm(device const uchar * src0, - device const uchar * src1, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne02, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant uint64_t & nb03, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant uint64_t & nb13, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint & r2, - constant uint & r3, - threadgroup uchar * shared_memory [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiitg[[thread_index_in_threadgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { +kernel void kernel_mul_mm( + constant ggml_metal_kargs_mul_mm & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiitg[[thread_index_in_threadgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - threadgroup T * sa = (threadgroup T *)(shared_memory); - threadgroup float * sb = (threadgroup float *)(shared_memory + 4096); + threadgroup T * sa = (threadgroup T *)(shmem); + threadgroup float * sb = (threadgroup float *)(shmem + 4096); - const uint r0 = tgpig.y; - const uint r1 = tgpig.x; - const uint im = tgpig.z; + const int r0 = tgpig.y; + const int r1 = tgpig.x; + const int im = tgpig.z; // if this block is of 64x32 shape or smaller - short n_rows = (ne0 - r0 * BLOCK_SIZE_M < BLOCK_SIZE_M) ? (ne0 - r0 * BLOCK_SIZE_M) : BLOCK_SIZE_M; - short n_cols = (ne1 - r1 * BLOCK_SIZE_N < BLOCK_SIZE_N) ? (ne1 - r1 * BLOCK_SIZE_N) : BLOCK_SIZE_N; + const short n_rows = (args.ne0 - r0*BLOCK_SIZE_M < BLOCK_SIZE_M) ? (args.ne0 - r0*BLOCK_SIZE_M) : BLOCK_SIZE_M; + const short n_cols = (args.ne1 - r1*BLOCK_SIZE_N < BLOCK_SIZE_N) ? (args.ne1 - r1*BLOCK_SIZE_N) : BLOCK_SIZE_N; // a thread shouldn't load data outside of the matrix - short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1; - short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1; + const short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1; + const short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1; simdgroup_T8x8 ma[4]; simdgroup_float8x8 mb[2]; - simdgroup_float8x8 c_res[8]; - for (int i = 0; i < 8; i++){ - c_res[i] = make_filled_simdgroup_matrix(0.f); + simdgroup_float8x8 mc[8]; + + for (short i = 0; i < 8; i++){ + mc[i] = make_filled_simdgroup_matrix(0.f); } short il = (tiitg % THREAD_PER_ROW); - const uint i12 = im%ne12; - const uint i13 = im/ne12; + const int i12 = im%args.ne12; + const int i13 = im/args.ne12; - uint offset0 = (i12/r2)*nb02 + (i13/r3)*nb03; - ushort offset1 = il/nl; + const uint64_t offset0 = (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const short offset1 = il/nl; + + device const block_q * x = (device const block_q *)(src0 + + args.nb01*(r0*BLOCK_SIZE_M + thread_row) + offset0) + offset1; - device const block_q * x = (device const block_q *)(src0 + (r0 * BLOCK_SIZE_M + thread_row) * nb01 + offset0) + offset1; device const float * y = (device const float *)(src1 - + nb13 * i13 - + nb12 * i12 - + nb11 * (r1 * BLOCK_SIZE_N + thread_col) - + nb10 * (BLOCK_SIZE_K / THREAD_PER_COL * (tiitg % THREAD_PER_COL))); + + args.nb13*i13 + + args.nb12*i12 + + args.nb11*(r1*BLOCK_SIZE_N + thread_col) + + args.nb10*(BLOCK_SIZE_K / THREAD_PER_COL * (tiitg % THREAD_PER_COL))); - for (int loop_k = 0; loop_k < ne00; loop_k += BLOCK_SIZE_K) { + for (int loop_k = 0; loop_k < args.ne00; loop_k += BLOCK_SIZE_K) { // load data and store to threadgroup memory T4x4 temp_a; dequantize_func(x, il, temp_a); + threadgroup_barrier(mem_flags::mem_threadgroup); #pragma unroll(16) - for (int i = 0; i < 16; i++) { - *(sa + SG_MAT_SIZE * ((tiitg / THREAD_PER_ROW / 8) \ - + (tiitg % THREAD_PER_ROW) * 16 + (i / 8) * 8) \ - + (tiitg / THREAD_PER_ROW) % 8 + (i & 7) * 8) = temp_a[i/4][i%4]; + for (short i = 0; i < 16; i++) { + *(sa + SG_MAT_SIZE * ((tiitg/THREAD_PER_ROW/8) \ + + (tiitg%THREAD_PER_ROW)*16 + (i/8)*8) \ + + (tiitg/THREAD_PER_ROW)%8 + (i&7)*8) = temp_a[i/4][i%4]; } - *(threadgroup float2x4 *)(sb + (tiitg % THREAD_PER_COL) * 8 * 32 + 8 * (tiitg / THREAD_PER_COL)) = *((device float2x4 *)y); + *(threadgroup float2x4 *)(sb + 32*8*(tiitg%THREAD_PER_COL) + 8*(tiitg/THREAD_PER_COL)) = *((device float2x4 *) y); il = (il + 2 < nl) ? il + 2 : il % 2; - x = (il < 2) ? x + (2+nl-1)/nl : x; + x = (il < 2) ? x + (2 + nl - 1)/nl : x; y += BLOCK_SIZE_K; threadgroup_barrier(mem_flags::mem_threadgroup); // load matrices from threadgroup memory and conduct outer products - threadgroup T * lsma = (sa + THREAD_MAT_M * SG_MAT_SIZE * (sgitg % 2)); - threadgroup float * lsmb = (sb + THREAD_MAT_N * SG_MAT_SIZE * (sgitg / 2)); + threadgroup const T * lsma = (sa + THREAD_MAT_M*SG_MAT_SIZE*(sgitg%2)); + threadgroup const float * lsmb = (sb + THREAD_MAT_N*SG_MAT_SIZE*(sgitg/2)); #pragma unroll(4) - for (int ik = 0; ik < BLOCK_SIZE_K / 8; ik++) { + for (short ik = 0; ik < BLOCK_SIZE_K/8; ik++) { #pragma unroll(4) - for (int i = 0; i < 4; i++) { - simdgroup_load(ma[i],lsma + SG_MAT_SIZE * i); - } - simdgroup_barrier(mem_flags::mem_none); - #pragma unroll(2) - for (int i = 0; i < 2; i++) { - simdgroup_load(mb[i],lsmb + SG_MAT_SIZE * i); + for (short i = 0; i < 4; i++) { + simdgroup_load(ma[i], lsma + SG_MAT_SIZE * i); } - lsma += BLOCK_SIZE_M / SG_MAT_ROW * SG_MAT_SIZE; - lsmb += BLOCK_SIZE_N / SG_MAT_ROW * SG_MAT_SIZE; + simdgroup_barrier(mem_flags::mem_none); + + #pragma unroll(2) + for (short i = 0; i < 2; i++) { + simdgroup_load(mb[i], lsmb + SG_MAT_SIZE * i); + } #pragma unroll(8) - for (int i = 0; i < 8; i++){ - simdgroup_multiply_accumulate(c_res[i], mb[i/4], ma[i%4], c_res[i]); + for (short i = 0; i < 8; i++){ + simdgroup_multiply_accumulate(mc[i], mb[i/4], ma[i%4], mc[i]); } + + lsma += (BLOCK_SIZE_M/SG_MAT_ROW)*SG_MAT_SIZE; + lsmb += (BLOCK_SIZE_N/SG_MAT_ROW)*SG_MAT_SIZE; } } - if ((r0 + 1) * BLOCK_SIZE_M <= ne0 && (r1 + 1) * BLOCK_SIZE_N <= ne1) { - device float * C = dst + (BLOCK_SIZE_M * r0 + 32 * (sgitg & 1)) \ - + (BLOCK_SIZE_N * r1 + 16 * (sgitg >> 1)) * ne0 + im*ne1*ne0; - for (int i = 0; i < 8; i++) { - simdgroup_store(c_res[i], C + 8 * (i%4) + 8 * ne0 * (i/4), ne0); + if ((r0 + 1) * BLOCK_SIZE_M <= args.ne0 && (r1 + 1) * BLOCK_SIZE_N <= args.ne1) { + device float * C = (device float *) dst + + (BLOCK_SIZE_M * r0 + 32*(sgitg & 1)) + \ + (BLOCK_SIZE_N * r1 + 16*(sgitg >> 1)) * args.ne0 + im*args.ne1*args.ne0; + + for (short i = 0; i < 8; i++) { + simdgroup_store(mc[i], C + 8 * (i%4) + 8 * args.ne0 * (i/4), args.ne0); } } else { // block is smaller than 64x32, we should avoid writing data outside of the matrix threadgroup_barrier(mem_flags::mem_threadgroup); - threadgroup float * temp_str = ((threadgroup float *)shared_memory) \ - + 32 * (sgitg&1) + (16 * (sgitg>>1)) * BLOCK_SIZE_M; - for (int i = 0; i < 8; i++) { - simdgroup_store(c_res[i], temp_str + 8 * (i%4) + 8 * BLOCK_SIZE_M * (i/4), BLOCK_SIZE_M); + threadgroup float * temp_str = ((threadgroup float *) shmem) \ + + 32*(sgitg&1) + (16*(sgitg >> 1))*BLOCK_SIZE_M; + for (short i = 0; i < 8; i++) { + simdgroup_store(mc[i], temp_str + 8*(i%4) + 8*BLOCK_SIZE_M*(i/4), BLOCK_SIZE_M); } threadgroup_barrier(mem_flags::mem_threadgroup); - device float * C = dst + (BLOCK_SIZE_M * r0) + (BLOCK_SIZE_N * r1) * ne0 + im*ne1*ne0; if (sgitg == 0) { - for (int i = 0; i < n_rows; i++) { - for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) { - *(C + i + j * ne0) = *(temp_str + i + j * BLOCK_SIZE_M); + for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) { + device float * D = (device float *) dst + (r0*BLOCK_SIZE_M) + (r1*BLOCK_SIZE_N + j)*args.ne0 + im*args.ne1*args.ne0; + device float4 * D4 = (device float4 *) D; + + threadgroup float * C = temp_str + (j*BLOCK_SIZE_M); + threadgroup float4 * C4 = (threadgroup float4 *) C; + + int i = 0; + for (; i < n_rows/4; i++) { + *(D4 + i) = *(C4 + i); + } + + i *= 4; + for (; i < n_rows; i++) { + *(D + i) = *(C + i); } } } @@ -6305,36 +6193,37 @@ kernel void kernel_mul_mm(device const uchar * src0, } // same as kernel_mul_mm_impl, but src1 and dst are accessed via indices stored in rowids +// TODO: this kernel needs to be reimplemented from scratch for better performance template void kernel_mul_mm_id_impl( - device const uchar * src0, - device const uchar * src1, + int32_t ne00, + int32_t ne02, + uint64_t nb01, + uint64_t nb02, + int32_t ne11, + int32_t ne12, + uint64_t nb10, + uint64_t nb11, + uint64_t nb12, + int32_t ne0, + int32_t ne1, + int64_t ne0ne1, + device const char * src0, + device const char * src1, threadgroup ushort2 * rowids, - device float * dst, - constant int64_t & ne00, - constant int64_t & ne02, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant int64_t & ne11, - constant int64_t & ne12, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant int64_t & ne0, - int64_t ne1, - int64_t ne0ne1, - threadgroup uchar * shared_memory, - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiitg[[thread_index_in_threadgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + device char * dst, + threadgroup char * shmem, + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiitg[[thread_index_in_threadgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { - threadgroup half * sa = (threadgroup half *)(shared_memory); - threadgroup float * sb = (threadgroup float *)(shared_memory + 4096); + threadgroup half * sa = (threadgroup half *)(shmem); + threadgroup float * sb = (threadgroup float *)(shmem + 4096); - const uint r0 = tgpig.y; - const uint r1 = tgpig.x; + const int r0 = tgpig.y; + const int r1 = tgpig.x; - if (r1 * BLOCK_SIZE_N >= ne1) return; + if (r1*BLOCK_SIZE_N >= ne1) return; // if this block is of 64x32 shape or smaller short n_rows = (ne0 - r0 * BLOCK_SIZE_M < BLOCK_SIZE_M) ? (ne0 - r0 * BLOCK_SIZE_M) : BLOCK_SIZE_M; @@ -6346,9 +6235,9 @@ void kernel_mul_mm_id_impl( simdgroup_half8x8 ma[4]; simdgroup_float8x8 mb[2]; - simdgroup_float8x8 c_res[8]; + simdgroup_float8x8 mc[8]; for (int i = 0; i < 8; i++){ - c_res[i] = make_filled_simdgroup_matrix(0.f); + mc[i] = make_filled_simdgroup_matrix(0.f); } short il = (tiitg % THREAD_PER_ROW); @@ -6386,11 +6275,14 @@ void kernel_mul_mm_id_impl( threadgroup half * lsma = (sa + THREAD_MAT_M * SG_MAT_SIZE * (sgitg % 2)); threadgroup float * lsmb = (sb + THREAD_MAT_N * SG_MAT_SIZE * (sgitg / 2)); + #pragma unroll(BLOCK_SIZE_K/8) for (int ik = 0; ik < BLOCK_SIZE_K / 8; ik++) { + #pragma unroll(4) for (int i = 0; i < 4; i++) { simdgroup_load(ma[i], lsma + SG_MAT_SIZE * i); } simdgroup_barrier(mem_flags::mem_none); + #pragma unroll(2) for (int i = 0; i < 2; i++) { simdgroup_load(mb[i], lsmb + SG_MAT_SIZE * i); } @@ -6398,29 +6290,42 @@ void kernel_mul_mm_id_impl( lsma += BLOCK_SIZE_M / SG_MAT_ROW * SG_MAT_SIZE; lsmb += BLOCK_SIZE_N / SG_MAT_ROW * SG_MAT_SIZE; + #pragma unroll(8) for (int i = 0; i < 8; i++){ - simdgroup_multiply_accumulate(c_res[i], mb[i/4], ma[i%4], c_res[i]); + simdgroup_multiply_accumulate(mc[i], mb[i/4], ma[i%4], mc[i]); } } } { threadgroup_barrier(mem_flags::mem_threadgroup); - threadgroup float * temp_str = ((threadgroup float *)shared_memory) \ + threadgroup float * temp_str = ((threadgroup float *) shmem) \ + 32 * (sgitg&1) + (16 * (sgitg>>1)) * BLOCK_SIZE_M; for (int i = 0; i < 8; i++) { - simdgroup_store(c_res[i], temp_str + 8 * (i%4) + 8 * BLOCK_SIZE_M * (i/4), BLOCK_SIZE_M); + simdgroup_store(mc[i], temp_str + 8 * (i%4) + 8 * BLOCK_SIZE_M * (i/4), BLOCK_SIZE_M); } threadgroup_barrier(mem_flags::mem_threadgroup); - device float * C = dst + (BLOCK_SIZE_M * r0); if (sgitg == 0) { for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) { threadgroup const auto & jid = rowids[r1 * BLOCK_SIZE_N + j]; - int joff = jid[0] * ne0 + jid[1] * ne0ne1; - for (int i = 0; i < n_rows; i++) { - *(C + i + joff) = *(temp_str + i + j * BLOCK_SIZE_M); + int64_t joff = jid[0]*ne0 + jid[1]*ne0ne1; + + device float * D = (device float *) dst + (r0*BLOCK_SIZE_M) + joff; + device float4 * D4 = (device float4 *) D; + + threadgroup float * C = temp_str + (j*BLOCK_SIZE_M); + threadgroup float4 * C4 = (threadgroup float4 *) C; + + int i = 0; + for (; i < n_rows/4; i++) { + *(D4 + i) = *(C4 + i); + } + + i *= 4; + for (; i < n_rows; i++) { + *(D + i) = *(C + i); } } } @@ -6429,48 +6334,34 @@ void kernel_mul_mm_id_impl( template kernel void kernel_mul_mm_id( - device const uchar * src0s, - device const uchar * src1, - device float * dst, - device const uchar * ids, - constant int64_t & nei0, - constant int64_t & nei1, - constant uint64_t & nbi1, - constant int64_t & ne00, - constant int64_t & ne02, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint64_t & nb1, - threadgroup uchar * shared_memory [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiitg[[thread_index_in_threadgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { + constant ggml_metal_kargs_mul_mm_id & args, + device const char * src0s, + device const char * src1, + device char * dst, + device const char * ids, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiitg[[thread_index_in_threadgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { const int32_t i02 = tgpig.z; + tgpig.z = 0; - device const uchar * src0 = src0s + i02*nb02; + device const char * src0 = src0s + i02*args.nb02; // row indices - threadgroup ushort2 * rowids = (threadgroup ushort2 *)(shared_memory + 8192); + threadgroup ushort2 * rowids = (threadgroup ushort2 *)(shmem + 8192); // TODO: parallelize this loop - int64_t _ne1 = 0; - for (ushort ii1 = 0; ii1 < nei1; ii1++) { - for (ushort ii0 = 0; ii0 < nei0; ii0++) { - int32_t id = ((device int32_t *) (ids + ii1*nbi1))[ii0]; + int32_t _ne1 = 0; + for (ushort ii1 = 0; ii1 < args.nei1; ii1++) { + for (ushort ii0 = 0; ii0 < args.nei0; ii0++) { + int32_t id = ((device int32_t *) (ids + ii1*args.nbi1))[ii0]; if (id == i02) { - //if (tiitg == 0) { + if (tiitg == 0) { rowids[_ne1] = ushort2(ii0, ii1); - //} + } _ne1++; } } @@ -6479,23 +6370,23 @@ kernel void kernel_mul_mm_id( threadgroup_barrier(mem_flags::mem_threadgroup); kernel_mul_mm_id_impl( + args.ne00, + args.ne02, + args.nb01, + args.nb02, + args.ne11, + args.ne12, + args.nb10, + args.nb11, + args.nb12, + args.ne0, + _ne1, + (int64_t)args.ne0*args.ne1, src0, src1, rowids, dst, - ne00, - ne02, - nb01, - nb02, - ne11, - ne12, - nb10, - nb11, - nb12, - ne0, - _ne1, - ne0*ne1, - shared_memory, + shmem, tgpig, tiitg, sgitg); @@ -6511,7 +6402,7 @@ typedef decltype(kernel_get_rows_f) get_rows_f_t; template [[host_name("kernel_get_rows_f32")]] kernel get_rows_f_t kernel_get_rows_f; template [[host_name("kernel_get_rows_f16")]] kernel get_rows_f_t kernel_get_rows_f; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_get_rows_bf16")]] kernel get_rows_f_t kernel_get_rows_f; #endif @@ -6545,7 +6436,7 @@ typedef decltype(kernel_mul_mm; template [[host_name("kernel_mul_mm_f16_f32")]] kernel mat_mm_t kernel_mul_mm; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mm_bf16_f32")]] kernel mat_mm_t kernel_mul_mm; #endif template [[host_name("kernel_mul_mm_q4_0_f32")]] kernel mat_mm_t kernel_mul_mm; @@ -6576,7 +6467,7 @@ typedef decltype(kernel_mul_mm_id) mat_mm_id_t; template [[host_name("kernel_mul_mm_id_f32_f32")]] kernel mat_mm_id_t kernel_mul_mm_id; template [[host_name("kernel_mul_mm_id_f16_f32")]] kernel mat_mm_id_t kernel_mul_mm_id; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mm_id_bf16_f32")]] kernel mat_mm_id_t kernel_mul_mm_id; #endif template [[host_name("kernel_mul_mm_id_q4_0_f32")]] kernel mat_mm_id_t kernel_mul_mm_id; @@ -6604,194 +6495,110 @@ template [[host_name("kernel_mul_mm_id_iq4_xs_f32")]] kernel mat_mm_id_t kernel // typedef void (kernel_mul_mv_impl_t)( - device const char * src0, - device const char * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb00, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne11, - int64_t ne12, - uint64_t nb10, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - uint3 tgpig, - uint tiisg); + ggml_metal_kargs_mul_mv args, + device const char * src0, + device const char * src1, + device char * dst, + uint3 tgpig, + ushort tiisg); typedef void (kernel_mul_mv2_impl_t)( - device const void * src0, - device const float * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne12, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiisg, - uint sgitg); + ggml_metal_kargs_mul_mv args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg); template void mmv_fn( - device const char * src0, - device const char * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb00, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne11, - int64_t ne12, - int64_t ne13, - uint64_t nb10, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint64_t nb1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiitg, - uint tiisg, - uint sgitg) { - impl_fn(src0,src1,dst,ne00,ne01,ne02,nb00,nb01,nb02,nb03,ne10,ne11,ne12,nb10,nb11,nb12,nb13,ne0,ne1,r2,r3,tgpig,tiisg); + ggml_metal_kargs_mul_mv args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiitg, + ushort tiisg, + ushort sgitg) { + impl_fn(args, src0, src1, dst, tgpig, tiisg); } template void mmv_fn( - device const char * src0, - device const char * src1, - device float * dst, - int64_t ne00, - int64_t ne01, - int64_t ne02, - uint64_t nb00, - uint64_t nb01, - uint64_t nb02, - uint64_t nb03, - int64_t ne10, - int64_t ne11, - int64_t ne12, - int64_t ne13, - uint64_t nb10, - uint64_t nb11, - uint64_t nb12, - uint64_t nb13, - int64_t ne0, - int64_t ne1, - uint64_t nb1, - uint r2, - uint r3, - threadgroup int8_t * shared_values, - uint3 tgpig, - uint tiitg, - uint tiisg, - uint sgitg) { - impl_fn(src0,(const device float *)src1,dst,ne00,ne01,ne02,nb01,nb02,nb03,ne10,ne12,nb11,nb12,nb13,ne0,ne1,r2,r3,shared_values,tgpig,tiisg,sgitg); + ggml_metal_kargs_mul_mv args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiitg, + ushort tiisg, + ushort sgitg) { + impl_fn(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -typedef decltype(mmv_fn>) mul_mv_impl_fn_t; +typedef decltype(mmv_fn>) mul_mv_impl_fn_t; template kernel void kernel_mul_mv_id( - device const char * src0s, - device const char * src1, - device float * dst, - device const char * ids, - constant int64_t & nei0, - constant int64_t & nei1, - constant uint64_t & nbi1, - constant int64_t & ne00, - constant int64_t & ne01, - constant int64_t & ne02, - constant uint64_t & nb00, - constant uint64_t & nb01, - constant uint64_t & nb02, - constant int64_t & ne10, - constant int64_t & ne11, - constant int64_t & ne12, - constant int64_t & ne13, - constant uint64_t & nb10, - constant uint64_t & nb11, - constant uint64_t & nb12, - constant int64_t & ne0, - constant int64_t & ne1, - constant uint64_t & nb1, - threadgroup int8_t * shared_values [[threadgroup(0)]], - uint3 tgpig[[threadgroup_position_in_grid]], - uint tiitg[[thread_index_in_threadgroup]], - uint tiisg[[thread_index_in_simdgroup]], - uint sgitg[[simdgroup_index_in_threadgroup]]) { - const int iid1 = tgpig.z/nei0; - const int idx = tgpig.z%nei0; + constant ggml_metal_kargs_mul_mv_id & args, + device const char * src0s, + device const char * src1, + device char * dst, + device const char * ids, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiitg[[thread_index_in_threadgroup]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + const int iid1 = tgpig.z/args.nei0; + const int idx = tgpig.z%args.nei0; tgpig.z = 0; - const int32_t i02 = ((device const int32_t *) (ids + iid1*nbi1))[idx]; + const int32_t i02 = ((device const int32_t *) (ids + iid1*args.nbi1))[idx]; - const int64_t i11 = idx % ne11; + const int64_t i11 = idx % args.ne11; const int64_t i12 = iid1; const int64_t i1 = idx; const int64_t i2 = i12; - device const char * src0_cur = src0s + i02*nb02; - device const char * src1_cur = src1 + i11*nb11 + i12*nb12; - device float * dst_cur = dst + i1*ne0 + i2*ne1*ne0; + device const char * src0_cur = src0s + i02*args.nb02; + device const char * src1_cur = src1 + i11*args.nb11 + i12*args.nb12; + + device char * dst_cur = dst + (i1*args.ne0 + i2*args.ne1*args.ne0)*sizeof(float); + + ggml_metal_kargs_mul_mv args0 = { + /*.ne00 =*/ args.ne00, + /*.ne01 =*/ args.ne01, + /*.ne02 =*/ 1, // args.ne02, + /*.nb00 =*/ args.nb00, + /*.nb01 =*/ args.nb01, + /*.nb02 =*/ args.nb02, + /*.nb03 =*/ args.nb02, // args.ne02 == 1 + /*.ne10 =*/ args.ne10, + /*.ne11 =*/ 1, // args.ne11, + /*.ne12 =*/ 1, // args.ne12, + /*.nb10 =*/ args.nb10, + /*.nb11 =*/ args.nb11, + /*.nb12 =*/ args.nb12, + /*.nb13 =*/ args.nb12, // ne12 == 1 + /*.ne0 =*/ args.ne0, + /*.ne1 =*/ 1, // args.ne1, + /*.r2 =*/ 1, + /*.r3 =*/ 1, + }; impl_fn( + args0, /* src0 */ src0_cur, /* src1 */ src1_cur, /* dst */ dst_cur, - /* ne00 */ ne00, - /* ne01 */ ne01, - /* ne02 */ 1, // ne02, - /* nb00 */ nb00, - /* nb01 */ nb01, - /* nb02 */ nb02, - /* nb03 */ nb02, // ne02 == 1 - /* ne10 */ ne10, - /* ne11 */ 1, // ne11, - /* ne12 */ 1, // ne12, - /* ne13 */ 1, // ne13, - /* nb10 */ nb10, - /* nb11 */ nb11, - /* nb12 */ nb12, - /* ne13 */ nb12, // ne12 == 1 - /* ne0 */ ne0, - /* ne1 */ 1, // ne1, - /* nb1 */ nb1, - /* r2 */ 1, - /* r3 */ 1, - shared_values, + shmem, tgpig, tiitg, tiisg, @@ -6802,7 +6609,7 @@ typedef decltype(kernel_mul_mv_id>>; template [[host_name("kernel_mul_mv_id_f16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -#if !defined(GGML_METAL_NO_BFLOAT) +#if defined(GGML_METAL_USE_BF16) template [[host_name("kernel_mul_mv_id_bf16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; #endif template [[host_name("kernel_mul_mv_id_q8_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>; diff --git a/ggml/src/ggml-musa/CMakeLists.txt b/ggml/src/ggml-musa/CMakeLists.txt new file mode 100644 index 000000000..415b2b2e0 --- /dev/null +++ b/ggml/src/ggml-musa/CMakeLists.txt @@ -0,0 +1,107 @@ +if (NOT EXISTS $ENV{MUSA_PATH}) + if (NOT EXISTS /opt/musa) + set(MUSA_PATH /usr/local/musa) + else() + set(MUSA_PATH /opt/musa) + endif() +else() + set(MUSA_PATH $ENV{MUSA_PATH}) +endif() + +set(CMAKE_C_COMPILER "${MUSA_PATH}/bin/clang") +set(CMAKE_C_EXTENSIONS OFF) +set(CMAKE_CXX_COMPILER "${MUSA_PATH}/bin/clang++") +set(CMAKE_CXX_EXTENSIONS OFF) + +list(APPEND CMAKE_MODULE_PATH "${MUSA_PATH}/cmake") + +find_package(MUSAToolkit) + +if (MUSAToolkit_FOUND) + message(STATUS "MUSA Toolkit found") + + if (NOT DEFINED MUSA_ARCHITECTURES) + set(MUSA_ARCHITECTURES "21;22") + endif() + message(STATUS "Using MUSA architectures: ${MUSA_ARCHITECTURES}") + + file(GLOB GGML_HEADERS_MUSA "../ggml-cuda/*.cuh") + list(APPEND GGML_HEADERS_MUSA "../../include/ggml-cuda.h") + + file(GLOB GGML_SOURCES_MUSA "../ggml-cuda/*.cu") + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-wmma*.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + file(GLOB SRCS "../ggml-cuda/template-instances/mmq*.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + + if (GGML_CUDA_FA_ALL_QUANTS) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) + else() + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + file(GLOB SRCS "../ggml-cuda/template-instances/fattn-vec*f16-f16.cu") + list(APPEND GGML_SOURCES_MUSA ${SRCS}) + endif() + + set_source_files_properties(${GGML_SOURCES_MUSA} PROPERTIES LANGUAGE CXX) + foreach(SOURCE ${GGML_SOURCES_MUSA}) + set(COMPILE_FLAGS "-x musa -mtgpu") + foreach(ARCH ${MUSA_ARCHITECTURES}) + set(COMPILE_FLAGS "${COMPILE_FLAGS} --cuda-gpu-arch=mp_${ARCH}") + endforeach() + set_property(SOURCE ${SOURCE} PROPERTY COMPILE_FLAGS ${COMPILE_FLAGS}) + endforeach() + + ggml_add_backend_library(ggml-musa + ${GGML_HEADERS_MUSA} + ${GGML_SOURCES_MUSA} + ) + + # TODO: do not use CUDA definitions for MUSA + target_compile_definitions(ggml PUBLIC GGML_USE_CUDA) + + add_compile_definitions(GGML_USE_MUSA) + add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${GGML_CUDA_PEER_MAX_BATCH_SIZE}) + + if (GGML_CUDA_GRAPHS) + add_compile_definitions(GGML_CUDA_USE_GRAPHS) + endif() + + if (GGML_CUDA_FORCE_MMQ) + add_compile_definitions(GGML_CUDA_FORCE_MMQ) + endif() + + if (GGML_CUDA_FORCE_CUBLAS) + add_compile_definitions(GGML_CUDA_FORCE_CUBLAS) + endif() + + if (GGML_CUDA_NO_VMM) + add_compile_definitions(GGML_CUDA_NO_VMM) + endif() + + if (GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16) + add_compile_definitions(GGML_CUDA_F16) + endif() + + if (GGML_CUDA_NO_PEER_COPY) + add_compile_definitions(GGML_CUDA_NO_PEER_COPY) + endif() + + if (GGML_STATIC) + target_link_libraries(ggml-musa PRIVATE MUSA::musart_static MUSA::mublas_static) + else() + target_link_libraries(ggml-musa PRIVATE MUSA::musart MUSA::mublas) + endif() + + if (GGML_CUDA_NO_VMM) + # No VMM requested, no need to link directly with the musa driver lib (libmusa.so) + else() + target_link_libraries(ggml-musa PRIVATE MUSA::musa_driver) + endif() +else() + message(FATAL_ERROR "MUSA Toolkit not found") +endif() diff --git a/ggml/src/ggml-opencl/CMakeLists.txt b/ggml/src/ggml-opencl/CMakeLists.txt new file mode 100644 index 000000000..45328a657 --- /dev/null +++ b/ggml/src/ggml-opencl/CMakeLists.txt @@ -0,0 +1,147 @@ +find_package(OpenCL REQUIRED) +find_package(Python3 REQUIRED) + +set(TARGET_NAME ggml-opencl) + +ggml_add_backend_library(${TARGET_NAME} + ggml-opencl.cpp + ../../include/ggml-opencl.h) +target_link_libraries(${TARGET_NAME} PRIVATE ${OpenCL_LIBRARIES}) +target_include_directories(${TARGET_NAME} PRIVATE ${OpenCL_INCLUDE_DIRS}) + +if (GGML_OPENCL_PROFILING) + message(STATUS "OpenCL profiling enabled (increases CPU overhead)") + add_compile_definitions(GGML_OPENCL_PROFILING) +endif () + +add_compile_definitions(GGML_OPENCL_SOA_Q) + +if (GGML_OPENCL_USE_ADRENO_KERNELS) + message(STATUS "OpenCL will use matmul kernels optimized for Adreno") + add_compile_definitions(GGML_OPENCL_USE_ADRENO_KERNELS) +endif () + +if (GGML_OPENCL_EMBED_KERNELS) + add_compile_definitions(GGML_OPENCL_EMBED_KERNELS) + + set(OPENCL_CL_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl.cl.h") + set(OPENCL_MM_CL_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_mm.cl.h") + set(OPENCL_CVT_CL_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_cvt.cl.h") + + set(OPENCL_GEMV_NOSHUFFLE_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_gemv_noshuffle.cl.h") + set(OPENCL_GEMV_NOSHUFFLE_GENERAL_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_gemv_noshuffle_general.cl.h") + set(OPENCL_MUL_MAT_Ab_Bi_8x4_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_mul_mat_Ab_Bi_8x4.cl.h") + set(OPENCL_TRANSPOSE_16_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_transpose_16.cl.h") + set(OPENCL_TRANSPOSE_32_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_transpose_32.cl.h") + set(OPENCL_TRANSPOSE_32_16_SOURCE_EMBED "${CMAKE_BINARY_DIR}/autogenerated/ggml-opencl_transpose_32_16.cl.h") + + set(EMBED_KERNEL_SCRIPT "${CMAKE_CURRENT_SOURCE_DIR}/kernels/embed_kernel.py") + file(MAKE_DIRECTORY "${CMAKE_BINARY_DIR}/autogenerated") + + include_directories("${CMAKE_BINARY_DIR}/autogenerated") + + # Python must be accessible from command line + add_custom_command( + OUTPUT ${OPENCL_CL_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl.cl + ${OPENCL_CL_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_MM_CL_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_mm.cl + ${OPENCL_MM_CL_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_mm.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_mm.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_CVT_CL_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_cvt.cl + ${OPENCL_CVT_CL_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_cvt.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_cvt.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_GEMV_NOSHUFFLE_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_gemv_noshuffle.cl + ${OPENCL_GEMV_NOSHUFFLE_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_gemv_noshuffle.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_gemv_noshuffle.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_GEMV_NOSHUFFLE_GENERAL_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_gemv_noshuffle_general.cl + ${OPENCL_GEMV_NOSHUFFLE_GENERAL_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_gemv_noshuffle_general.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_gemv_noshuffle_general.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_MUL_MAT_Ab_Bi_8x4_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl + ${OPENCL_MUL_MAT_Ab_Bi_8x4_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_mul_mat_Ab_Bi_8x4.cl.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_TRANSPOSE_16_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_transpose_16.cl + ${OPENCL_TRANSPOSE_16_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_transpose_16.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_transpose_16.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_TRANSPOSE_32_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_transpose_32.cl + ${OPENCL_TRANSPOSE_32_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_transpose_32.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_transpose_32.cl.h" + ) + + add_custom_command( + OUTPUT ${OPENCL_TRANSPOSE_32_16_SOURCE_EMBED} + COMMAND ${Python3_EXECUTABLE} ${EMBED_KERNEL_SCRIPT} + ${CMAKE_CURRENT_SOURCE_DIR}/kernels/ggml-opencl_transpose_32_16.cl + ${OPENCL_TRANSPOSE_32_16_SOURCE_EMBED} + DEPENDS kernels/ggml-opencl_transpose_32_16.cl ${EMBED_KERNEL_SCRIPT} + COMMENT "Generate ggml-opencl_transpose_32_16.cl.h" + ) + + target_sources(${TARGET_NAME} PRIVATE + ${OPENCL_CL_SOURCE_EMBED} + ${OPENCL_MM_CL_SOURCE_EMBED} + ${OPENCL_CVT_CL_SOURCE_EMBED} + ${OPENCL_GEMV_NOSHUFFLE_SOURCE_EMBED} + ${OPENCL_GEMV_NOSHUFFLE_GENERAL_SOURCE_EMBED} + ${OPENCL_MUL_MAT_Ab_Bi_8x4_SOURCE_EMBED} + ${OPENCL_TRANSPOSE_16_SOURCE_EMBED} + ${OPENCL_TRANSPOSE_32_SOURCE_EMBED} + ${OPENCL_TRANSPOSE_32_16_SOURCE_EMBED}) +else () + # copy ggml-opencl.cl to bin directory + configure_file(kernels/ggml-opencl.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl.cl COPYONLY) + configure_file(kernels/ggml-opencl_mm.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_mm.cl COPYONLY) + configure_file(kernels/ggml-opencl_cvt.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_cvt.cl COPYONLY) + + configure_file(kernels/ggml-opencl_gemv_noshuffle.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_gemv_noshuffle.cl COPYONLY) + configure_file(kernels/ggml-opencl_gemv_noshuffle_general.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_gemv_noshuffle_general.cl COPYONLY) + configure_file(kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_mul_mat_Ab_Bi_8x4.cl COPYONLY) + configure_file(kernels/ggml-opencl_transpose_16.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_transpose_16.cl COPYONLY) + configure_file(kernels/ggml-opencl_transpose_32.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_transpose_32.cl COPYONLY) + configure_file(kernels/ggml-opencl_transpose_32_16.cl ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-opencl_transpose_32_16.cl COPYONLY) +endif () diff --git a/ggml/src/ggml-opencl/ggml-opencl.cpp b/ggml/src/ggml-opencl/ggml-opencl.cpp new file mode 100644 index 000000000..ed90e471a --- /dev/null +++ b/ggml/src/ggml-opencl/ggml-opencl.cpp @@ -0,0 +1,4004 @@ +#define CL_TARGET_OPENCL_VERSION 220 +#define CL_USE_DEPRECATED_OPENCL_1_2_APIS + +// suppress warnings in CL headers for GCC and Clang +#pragma GCC diagnostic ignored "-Woverlength-strings" +#ifdef __clang__ +#pragma GCC diagnostic ignored "-Wgnu-anonymous-struct" +#endif + +#include "ggml-opencl.h" +#include "ggml-backend.h" +#include "ggml-impl.h" +#include "ggml-backend-impl.h" +#include "ggml.h" + +#include + +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#undef MIN +#undef MAX +#define MIN(a, b) ((a) < (b) ? (a) : (b)) +#define MAX(a, b) ((a) > (b) ? (a) : (b)) + +#define UNUSED(x) (void)(x) + +#define CL_CHECK(err) \ + do { \ + cl_int err_ = (err); \ + if (err_ != CL_SUCCESS) { \ + GGML_LOG_ERROR("ggml_opencl: %s error %d at %s:%d\n", \ + #err, err_, __FILE__, __LINE__); \ + GGML_ASSERT(0); \ + } \ + } while (0) + +//------------------------------------------------------------------------------ +// OpenCL +//------------------------------------------------------------------------------ + +bool ggml_cl_compute_forward(ggml_backend_t backend, struct ggml_tensor * tensor); + +enum GPU_FAMILY { + ADRENO, + INTEL, + UNKNOWN, +}; + +enum ADRENO_GPU_GEN { + ADRENO_UNKNOWN, + A7X, + A8X, + X1E, +}; + +static ADRENO_GPU_GEN get_adreno_gpu_gen(const char *device_name) { + if (strstr(device_name, "730") || + strstr(device_name, "740") || + strstr(device_name, "750")) { + return ADRENO_GPU_GEN::A7X; + } + + if (strstr(device_name, "830")) { + return ADRENO_GPU_GEN::A8X; + } + + if (strstr(device_name, "X1")) { + return ADRENO_GPU_GEN::X1E; + } + + return ADRENO_GPU_GEN::ADRENO_UNKNOWN; +} + +static int get_adreno_cl_compiler_version(const char *driver_version) { + std::string driver_ver_str(driver_version); + size_t compiler_ver_pos = driver_ver_str.find("E031"); + size_t compiler_ver_len = 13; + size_t compiler_ver_offset = 5; + + if (compiler_ver_pos == std::string::npos) { + compiler_ver_pos = driver_ver_str.find("DX"); + if (compiler_ver_pos == std::string::npos) { + return -1; + } + compiler_ver_len = 11; + compiler_ver_offset = 3; + } + + std::string compiler_ver_str = driver_ver_str.substr(compiler_ver_pos, compiler_ver_len); + std::string major_ver_str = compiler_ver_str.substr(compiler_ver_offset, 2); + return std::atoi(major_ver_str.c_str()); +} + +// backend device context +struct ggml_backend_opencl_device_context { + cl_platform_id platform; + std::string platform_name; + + cl_device_id device; + std::string device_name; +}; + +// backend context +struct ggml_backend_opencl_context { + cl_device_id device; + std::string device_name; + + std::string driver_version; + + GPU_FAMILY gpu_family; + ADRENO_GPU_GEN adreno_gen; + + cl_int alignment; + size_t max_alloc_size; + bool fp16_support; + + int adreno_wave_size; + + cl_context context; + cl_command_queue queue; + + cl_program program; + cl_program program_1; + cl_program program_2; + + cl_kernel kernel_add, kernel_add_row; + cl_kernel kernel_mul, kernel_mul_row; + cl_kernel kernel_scale; + cl_kernel kernel_silu, kernel_silu_4; + cl_kernel kernel_gelu, kernel_gelu_4; + cl_kernel kernel_relu; + cl_kernel kernel_clamp; + cl_kernel kernel_norm; + cl_kernel kernel_rms_norm; + cl_kernel kernel_diag_mask_inf, kernel_diag_mask_inf_8; + cl_kernel kernel_soft_max, kernel_soft_max_4; + cl_kernel kernel_get_rows_f32, kernel_get_rows_f16, kernel_get_rows_q4_0; + cl_kernel kernel_rope_norm_f32, kernel_rope_norm_f16, kernel_rope_neox_f32, kernel_rope_neox_f16; + cl_kernel kernel_cpy_f16_f16, kernel_cpy_f16_f32, kernel_cpy_f32_f16, kernel_cpy_f32_f32; + cl_kernel kernel_mul_mat_f32_f32; + cl_kernel kernel_mul_mat_f16_f16; + cl_kernel kernel_mul_mat_f16_f32_1row; + cl_kernel kernel_mul_mat_f16_f32; + cl_kernel kernel_mul_mat_f16_f32_l4; + cl_kernel kernel_mul_mat_q4_0_f32, kernel_mul_mat_q4_0_f32_v; + cl_kernel kernel_convert_block_q4_0, kernel_restore_block_q4_0, kernel_mul_mat_q4_0_f32_flat; + cl_kernel kernel_mul_mat_q4_0_f32_8x_flat; + cl_kernel kernel_convert_block_q4_0_noshuffle, kernel_mul_mat_q4_0_f32_flat_v0, + kernel_mul_mat_q4_0_f32_flat_img_v0; + cl_kernel kernel_mul_mat_q4_0_f32_1d_8x_flat, kernel_mul_mat_q4_0_f32_1d_16x_flat; + cl_kernel kernel_mul_mv_q6_K_f32; + +#ifdef GGML_OPENCL_USE_ADRENO_KERNELS + // Transpose kernels + cl_program program_transpose_32; + cl_program program_transpose_32_16; + cl_program program_transpose_16; + cl_kernel kernel_transpose_32; + cl_kernel kernel_transpose_32_16; + cl_kernel kernel_transpose_16; + + cl_mem A_s_d_max; // max scale buffer size for transpose + cl_mem A_q_d_max; // max weight buffer size for transpose + cl_mem B_d_max; // max activation buffer size for transpose + + // Gemm and Gemv related programs, kernels, etc + cl_program program_CL_gemm; + cl_program program_CL_gemv_general; + cl_program program_CL_gemv_4096_1_11008; + cl_program program_CL_gemv_4096_1_4096; + cl_program program_CL_gemv_11008_1_4096; + cl_program program_CL_gemv_32000_1_4096; + cl_kernel CL_mul_mat_Ab_Bi_8x4; + cl_kernel CL_mul_mat_vec_q4_0_f32_1d_4x_flat_general; + cl_kernel CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_11008; + cl_kernel CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_4096; + cl_kernel CL_mul_mat_vec_q4_0_f32_1d_4x_flat_11008_1_4096; + cl_kernel CL_mul_mat_vec_q4_0_f32_1d_4x_flat_32000_1_4096; +#endif // GGML_OPENCL_USE_ADRENO_KERNELS +}; + +static ggml_backend_device g_ggml_backend_opencl_device; +static ggml_backend_opencl_device_context g_ggml_ctx_dev_main { + /*.platform =*/ nullptr, + /*.platform_nane =*/ "", + /*.device =*/ nullptr, + /*.device_name =*/ "", +}; + +static int ggml_backend_opencl_n_devices = 0; + +// Profiling +#ifdef GGML_OPENCL_PROFILING +struct ProfilingInfo { + std::string op_name; + std::string kernel_name; + // Kernel execution time in nanoseconds. + cl_ulong duration_ns; + // Global and local work sizes. + size_t global_size[3]; + size_t local_size[3]; + // Op output size. + size_t output_size[4]; +}; + +std::vector g_profiling_info; +#endif + +inline std::string read_file(const std::string &path) { + std::ifstream ifs(path); + if (!ifs) { + return ""; + } + std::string text; + ifs.seekg(0, std::ios::end); + text.resize(ifs.tellg()); + ifs.seekg(0, std::ios::beg); + ifs.read(&text[0], text.size()); + return text; +} + +static cl_program build_program_from_source(cl_context ctx, cl_device_id dev, const char* program_buffer, const std::string &compile_opts) { + cl_program p; + char *program_log; + size_t program_size; + size_t log_size; + int err; + + program_size = strlen(program_buffer); + + p = clCreateProgramWithSource(ctx, 1, (const char**)&program_buffer, &program_size, &err); + if(err < 0) { + GGML_LOG_ERROR("OpenCL error creating program"); + exit(1); + } + + err = clBuildProgram(p, 0, NULL, compile_opts.c_str(), NULL, NULL); + if(err < 0) { + clGetProgramBuildInfo(p, dev, CL_PROGRAM_BUILD_LOG, 0, NULL, &log_size); + program_log = (char*) malloc(log_size + 1); + program_log[log_size] = '\0'; + clGetProgramBuildInfo(p, dev, CL_PROGRAM_BUILD_LOG, log_size + 1, program_log, NULL); + GGML_LOG_ERROR("ggml_opencl: kernel compile error:\n\n%s\n", program_log); + free(program_log); + exit(1); + } + + return p; +} + +static ggml_backend_opencl_context * ggml_cl2_init(ggml_backend_dev_t dev) { + static bool initialized = false; + static ggml_backend_opencl_context *backend_ctx = nullptr; + + if (initialized) { + return backend_ctx; + } + + ggml_backend_opencl_device_context *dev_ctx = (ggml_backend_opencl_device_context *)dev->context; + GGML_ASSERT(dev_ctx); + GGML_ASSERT(dev_ctx->platform == nullptr); + GGML_ASSERT(dev_ctx->device == nullptr); + GGML_ASSERT(backend_ctx == nullptr); + + initialized = true; + backend_ctx = new ggml_backend_opencl_context(); + backend_ctx->gpu_family = GPU_FAMILY::UNKNOWN; + + cl_int err; + +#ifdef GGML_PROFILE_OPENCL + GGML_LOG_INFO("ggml_opencl: OpenCL profiling enabled\n"); +#endif + + struct cl_device; + struct cl_platform { + cl_platform_id id; + unsigned number; + char name[128]; + char vendor[128]; + struct cl_device * devices; + unsigned n_devices; + struct cl_device * default_device; + }; + + struct cl_device { + struct cl_platform * platform; + cl_device_id id; + unsigned number; + cl_device_type type; + char name[128]; + }; + + enum { NPLAT = 16, NDEV = 16 }; + + struct cl_platform platforms[NPLAT]; + unsigned n_platforms = 0; + struct cl_device devices[NDEV]; + unsigned n_devices = 0; + struct cl_device * default_device = NULL; + + cl_platform_id platform_ids[NPLAT]; + if (clGetPlatformIDs(NPLAT, platform_ids, &n_platforms) != CL_SUCCESS) { + GGML_LOG_ERROR("ggml_opencl: plaform IDs not available.\n"); + return backend_ctx; + } + + for (unsigned i = 0; i < n_platforms; i++) { + struct cl_platform * p = &platforms[i]; + p->number = i; + p->id = platform_ids[i]; + CL_CHECK(clGetPlatformInfo(p->id, CL_PLATFORM_NAME, sizeof(p->name), &p->name, NULL)); + CL_CHECK(clGetPlatformInfo(p->id, CL_PLATFORM_VENDOR, sizeof(p->vendor), &p->vendor, NULL)); + + cl_device_id device_ids[NDEV]; + cl_int clGetDeviceIDsError = clGetDeviceIDs(p->id, CL_DEVICE_TYPE_ALL, NDEV, device_ids, &p->n_devices); + if (clGetDeviceIDsError == CL_DEVICE_NOT_FOUND) { + p->n_devices = 0; + } else { + CL_CHECK(clGetDeviceIDsError); + } + p->devices = p->n_devices > 0 ? &devices[n_devices] : NULL; + p->default_device = NULL; + + for (unsigned j = 0; j < p->n_devices; j++) { + struct cl_device * d = &devices[n_devices]; + d->number = n_devices++; + d->id = device_ids[j]; + d->platform = p; + CL_CHECK(clGetDeviceInfo(d->id, CL_DEVICE_NAME, sizeof(d->name), &d->name, NULL)); + CL_CHECK(clGetDeviceInfo(d->id, CL_DEVICE_TYPE, sizeof(d->type), &d->type, NULL)); + + if (p->default_device == NULL && d->type == CL_DEVICE_TYPE_GPU) { + p->default_device = d; + } + } + + if (default_device == NULL && p->default_device != NULL) { + default_device = p->default_device; + } + } + + if (n_devices == 0) { + GGML_LOG_ERROR("ggml_opencl: could find any OpenCL devices.\n"); + return backend_ctx; + } + + char * user_platform_string = getenv("GGML_OPENCL_PLATFORM"); + char * user_device_string = getenv("GGML_OPENCL_DEVICE"); + int user_platform_number = -1; + int user_device_number = -1; + + unsigned n; + if (user_platform_string != NULL && sscanf(user_platform_string, " %u", &n) == 1 && n < n_platforms) { + user_platform_number = (int)n; + } + if (user_device_string != NULL && sscanf(user_device_string, " %u", &n) == 1 && n < n_devices) { + user_device_number = (int)n; + } + if (user_platform_number != -1 && user_device_number != -1) { + cl_platform* platform = &platforms[user_platform_number]; + if ((unsigned)user_device_number >= platform->n_devices) { + GGML_LOG_ERROR("ggml_opencl: invalid device number %d\n", user_device_number); + exit(1); + } + default_device = &platform->devices[user_device_number]; + } else { + + struct cl_device * selected_devices = devices; + unsigned n_selected_devices = n_devices; + + if (user_platform_number == -1 && user_platform_string != NULL && user_platform_string[0] != 0) { + for (unsigned i = 0; i < n_platforms; i++) { + struct cl_platform * p = &platforms[i]; + if (strstr(p->name, user_platform_string) != NULL || + strstr(p->vendor, user_platform_string) != NULL) { + user_platform_number = (int)i; + break; + } + } + if (user_platform_number == -1) { + GGML_LOG_ERROR("ggml_opencl: no platform matching '%s' was found.\n", user_platform_string); + exit(1); + } + } + if (user_platform_number != -1) { + struct cl_platform * p = &platforms[user_platform_number]; + selected_devices = p->devices; + n_selected_devices = p->n_devices; + default_device = p->default_device; + if (n_selected_devices == 0) { + GGML_LOG_ERROR("ggml_opencl: selected platform '%s' does not have any devices.\n", p->name); + exit(1); + } + } + + if (user_device_number == -1 && user_device_string != NULL && user_device_string[0] != 0) { + for (unsigned i = 0; i < n_selected_devices; i++) { + struct cl_device * d = &selected_devices[i]; + if (strstr(d->name, user_device_string) != NULL) { + user_device_number = d->number; + break; + } + } + if (user_device_number == -1) { + GGML_LOG_ERROR("ggml_opencl: no device matching '%s' was found.\n", user_device_string); + exit(1); + } + } + if (user_device_number != -1) { + selected_devices = &devices[user_device_number]; + n_selected_devices = 1; + default_device = &selected_devices[0]; + } + + GGML_ASSERT(n_selected_devices > 0); + + if (default_device == NULL) { + default_device = &selected_devices[0]; + } + } + + GGML_LOG_INFO("ggml_opencl: selecting platform: '%s'\n", default_device->platform->name); + GGML_LOG_INFO("ggml_opencl: selecting device: '%s'\n", default_device->name); + if (default_device->type != CL_DEVICE_TYPE_GPU) { + GGML_LOG_WARN("ggml_opencl: warning, not a GPU: '%s'.\n", default_device->name); + } + + dev_ctx->platform = default_device->platform->id; + dev_ctx->device = default_device->id; + backend_ctx->device = default_device->id; + + if (strstr(default_device->name, "Adreno")) { + backend_ctx->gpu_family = GPU_FAMILY::ADRENO; + backend_ctx->adreno_gen = get_adreno_gpu_gen(default_device->name); + + // Default wave size is 128, A8x uses 64. + if (backend_ctx->adreno_gen == ADRENO_GPU_GEN::A8X) { + backend_ctx->adreno_wave_size = 64; + } else if (backend_ctx->adreno_gen == ADRENO_GPU_GEN::A7X || + backend_ctx->adreno_gen == ADRENO_GPU_GEN::X1E) { + backend_ctx->adreno_wave_size = 128; + } else { + backend_ctx->adreno_wave_size = 128; + GGML_LOG_WARN("ggml_opencl: Unsupported Adreno GPU: %s, " + "using wave size %d, " + "may not work as expected\n", + backend_ctx->device_name.c_str(), backend_ctx->adreno_wave_size); + } + } else if (strstr(default_device->name, "Intel")) { + backend_ctx->gpu_family = GPU_FAMILY::INTEL; + } else { + GGML_LOG_ERROR("Unsupported GPU: %s\n", default_device->name); + backend_ctx->gpu_family = GPU_FAMILY::UNKNOWN; + return backend_ctx; + } + +#ifdef GGML_OPENCL_USE_ADRENO_KERNELS + if (backend_ctx->gpu_family != GPU_FAMILY::ADRENO) { + GGML_LOG_ERROR("ggml_opencl: Adreno-specific kernels should not be enabled for non-Adreno GPUs; " + "run on an Adreno GPU or recompile with CMake option `-DGGML_OPENCL_USE_ADRENO_KERNELS=OFF`\n"); + return backend_ctx; + } +#endif + + // Populate backend device name + dev_ctx->platform_name = default_device->platform->name; + dev_ctx->device_name = default_device->name; + backend_ctx->device_name = default_device->name; + + // A local ref of cl_device_id for convenience + cl_device_id device = backend_ctx->device; + + // Check device OpenCL version, OpenCL 2.0 or above is required + size_t device_ver_str_size; + clGetDeviceInfo(device, CL_DEVICE_VERSION, 0, NULL, &device_ver_str_size); + char *device_ver_buffer = (char *)alloca(device_ver_str_size + 1); + clGetDeviceInfo(device, CL_DEVICE_VERSION, device_ver_str_size, device_ver_buffer, NULL); + device_ver_buffer[device_ver_str_size] = '\0'; + GGML_LOG_INFO("ggml_opencl: device OpenCL version: %s\n", device_ver_buffer); + + if (strstr(device_ver_buffer, "OpenCL 2") == NULL && + strstr(device_ver_buffer, "OpenCL 3") == NULL) { + GGML_LOG_ERROR("ggml_opencl: OpenCL 2.0 or above is required\n"); + return backend_ctx; + } + + // Check driver version + size_t driver_version_str_size; + clGetDeviceInfo(device, CL_DRIVER_VERSION, 0, NULL, &driver_version_str_size); + char *driver_version = (char *)alloca(driver_version_str_size + 1); + clGetDeviceInfo(device, CL_DRIVER_VERSION, driver_version_str_size, driver_version, NULL); + driver_version[driver_version_str_size] = '\0'; + GGML_LOG_INFO("ggml_opencl: OpenCL driver: %s\n", driver_version); + backend_ctx->driver_version = driver_version; + + int adreno_cl_compiler_version = get_adreno_cl_compiler_version(driver_version); + bool has_vector_subgroup_broadcast = + adreno_cl_compiler_version >= 47 || adreno_cl_compiler_version == 17; + GGML_LOG_INFO("ggml_opencl: vector subgroup broadcast support: %s\n", + has_vector_subgroup_broadcast ? "true" : "false"); + + size_t ext_str_size; + clGetDeviceInfo(device, CL_DEVICE_EXTENSIONS, 0, NULL, &ext_str_size); + char *ext_buffer = (char *)alloca(ext_str_size + 1); + clGetDeviceInfo(device, CL_DEVICE_EXTENSIONS, ext_str_size, ext_buffer, NULL); + ext_buffer[ext_str_size] = '\0'; // ensure it is null terminated + // Check if ext_buffer contains cl_khr_fp16 + backend_ctx->fp16_support = strstr(ext_buffer, "cl_khr_fp16") != NULL; + GGML_LOG_INFO("ggml_opencl: device FP16 support: %s\n", backend_ctx->fp16_support ? "true" : "false"); + + // fp16 is required + if (!backend_ctx->fp16_support) { + GGML_LOG_ERROR("ggml_opencl: device does not support FP16\n"); + return backend_ctx; + } + + // If OpenCL 3.0 is supported, then check for cl_khr_subgroups, which becomes + // optional in OpenCL 3.0 (cl_khr_subgroup is mandatory in OpenCL 2.x) + if (strstr(device_ver_buffer, "OpenCL 3") && + strstr(ext_buffer, "cl_khr_subgroups") == NULL && + strstr(ext_buffer, "cl_intel_subgroups") == NULL) { + GGML_LOG_ERROR("ggml_opencl: device does not support subgroups (cl_khr_subgroups or cl_intel_subgroups) " + "(note that subgroups is an optional feature in OpenCL 3.0)\n"); + return backend_ctx; + } + + CL_CHECK(clGetDeviceInfo(device, CL_DEVICE_MEM_BASE_ADDR_ALIGN, sizeof(cl_uint), &backend_ctx->alignment, NULL)); + GGML_LOG_INFO("ggml_opencl: mem base addr align: %u\n", backend_ctx->alignment); + + clGetDeviceInfo(device, CL_DEVICE_MAX_MEM_ALLOC_SIZE, sizeof(size_t), &backend_ctx->max_alloc_size, NULL); + GGML_LOG_INFO("ggml_opencl: max mem alloc size: %zu MB\n", backend_ctx->max_alloc_size/1024/1024); + + // Check SVM. + cl_device_svm_capabilities svm_caps; + CL_CHECK(clGetDeviceInfo(device, CL_DEVICE_SVM_CAPABILITIES, sizeof(cl_device_svm_capabilities), &svm_caps, 0)); + GGML_LOG_INFO("ggml_opencl: SVM coarse grain buffer support: %s\n", + svm_caps & CL_DEVICE_SVM_COARSE_GRAIN_BUFFER ? "true" : "false"); + GGML_LOG_INFO("ggml_opencl: SVM fine grain buffer support: %s\n", + svm_caps & CL_DEVICE_SVM_FINE_GRAIN_BUFFER ? "true" : "false"); + GGML_LOG_INFO("ggml_opencl: SVM fine grain system support: %s\n", + svm_caps & CL_DEVICE_SVM_FINE_GRAIN_SYSTEM ? "true" : "false"); + GGML_LOG_INFO("ggml_opencl: SVM atomics support: %s\n", + svm_caps & CL_DEVICE_SVM_ATOMICS ? "true" : "false"); + + // Print out configurations +#ifdef GGML_OPENCL_SOA_Q + GGML_LOG_INFO("ggml_opencl: flattening quantized weights representation as struct of arrays (GGML_OPENCL_SOA_Q)\n"); +#endif // GGML_OPENCL_SOA_Q + +#ifdef GGML_OPENCL_USE_ADRENO_KERNELS + GGML_LOG_INFO("ggml_opencl: using kernels optimized for Adreno (GGML_OPENCL_USE_ADRENO_KERNELS)\n"); +#endif // GGML_OPENCL_USE_ADRENO_KERNELS + + cl_context_properties properties[] = { + (intptr_t)CL_CONTEXT_PLATFORM, (intptr_t)dev_ctx->platform, 0 + }; + + CL_CHECK((backend_ctx->context = clCreateContext(properties, 1, &device, NULL, NULL, &err), err)); + + // A local ref of cl_context for convenience + cl_context context = backend_ctx->context; + + //CL_CHECK((queue = clCreateCommandQueue(context, device, CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE, &err), + // (err != CL_INVALID_QUEUE_PROPERTIES && err != CL_INVALID_VALUE ? err : + // (queue = clCreateCommandQueue(context, device, 0, &err), err) + //))); + cl_command_queue_properties command_queue_props = 0; +#ifdef GGML_OPENCL_PROFILING + command_queue_props |= CL_QUEUE_PROFILING_ENABLE; +#endif + CL_CHECK((backend_ctx->queue = clCreateCommandQueue(context, device, command_queue_props, &err), err)); + +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src { + #include "ggml-opencl.cl.h" + }; +#else + const std::string kernel_src = read_file("ggml-opencl.cl"); +#endif + + std::string compile_opts = + "-cl-std=CL2.0 -cl-mad-enable -cl-unsafe-math-optimizations " + "-cl-finite-math-only -cl-fast-relaxed-math "; + backend_ctx->program = build_program_from_source(context, device, kernel_src.c_str(), compile_opts); + + // Non matmul kernels. + CL_CHECK((backend_ctx->kernel_get_rows_f32 = clCreateKernel(backend_ctx->program, "kernel_get_rows_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_get_rows_f16 = clCreateKernel(backend_ctx->program, "kernel_get_rows_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_get_rows_q4_0 = clCreateKernel(backend_ctx->program, "kernel_get_rows_q4_0", &err), err)); + CL_CHECK((backend_ctx->kernel_add = clCreateKernel(backend_ctx->program, "kernel_add", &err), err)); + CL_CHECK((backend_ctx->kernel_add_row = clCreateKernel(backend_ctx->program, "kernel_add_row", &err), err)); + CL_CHECK((backend_ctx->kernel_mul = clCreateKernel(backend_ctx->program, "kernel_mul", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_row = clCreateKernel(backend_ctx->program, "kernel_mul_row", &err), err)); + CL_CHECK((backend_ctx->kernel_scale = clCreateKernel(backend_ctx->program, "kernel_scale", &err), err)); + CL_CHECK((backend_ctx->kernel_silu = clCreateKernel(backend_ctx->program, "kernel_silu", &err), err)); + CL_CHECK((backend_ctx->kernel_silu_4 = clCreateKernel(backend_ctx->program, "kernel_silu_4", &err), err)); + CL_CHECK((backend_ctx->kernel_gelu = clCreateKernel(backend_ctx->program, "kernel_gelu", &err), err)); + CL_CHECK((backend_ctx->kernel_gelu_4 = clCreateKernel(backend_ctx->program, "kernel_gelu_4", &err), err)); + CL_CHECK((backend_ctx->kernel_relu = clCreateKernel(backend_ctx->program, "kernel_relu", &err), err)); + CL_CHECK((backend_ctx->kernel_clamp = clCreateKernel(backend_ctx->program, "kernel_clamp", &err), err)); + CL_CHECK((backend_ctx->kernel_norm = clCreateKernel(backend_ctx->program, "kernel_norm", &err), err)); + CL_CHECK((backend_ctx->kernel_rms_norm = clCreateKernel(backend_ctx->program, "kernel_rms_norm", &err), err)); + CL_CHECK((backend_ctx->kernel_diag_mask_inf = clCreateKernel(backend_ctx->program, "kernel_diag_mask_inf", &err), err)); + CL_CHECK((backend_ctx->kernel_diag_mask_inf_8 = clCreateKernel(backend_ctx->program, "kernel_diag_mask_inf_8", &err), err)); + CL_CHECK((backend_ctx->kernel_soft_max = clCreateKernel(backend_ctx->program, "kernel_soft_max", &err), err)); + CL_CHECK((backend_ctx->kernel_soft_max_4 = clCreateKernel(backend_ctx->program, "kernel_soft_max_4", &err), err)); + CL_CHECK((backend_ctx->kernel_rope_norm_f32 = clCreateKernel(backend_ctx->program, "kernel_rope_norm_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_rope_norm_f16 = clCreateKernel(backend_ctx->program, "kernel_rope_norm_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_rope_neox_f32 = clCreateKernel(backend_ctx->program, "kernel_rope_neox_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_rope_neox_f16 = clCreateKernel(backend_ctx->program, "kernel_rope_neox_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_cpy_f16_f16 = clCreateKernel(backend_ctx->program, "kernel_cpy_f16_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_cpy_f16_f32 = clCreateKernel(backend_ctx->program, "kernel_cpy_f16_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_cpy_f32_f16 = clCreateKernel(backend_ctx->program, "kernel_cpy_f32_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_cpy_f32_f32 = clCreateKernel(backend_ctx->program, "kernel_cpy_f32_f32", &err), err)); + + // Matmul kernels. + CL_CHECK((backend_ctx->kernel_mul_mat_f32_f32 = clCreateKernel(backend_ctx->program, "kernel_mul_mat_f32_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_f16_f16 = clCreateKernel(backend_ctx->program, "kernel_mul_mat_f16_f16", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_f16_f32_1row = clCreateKernel(backend_ctx->program, "kernel_mul_mat_f16_f32_1row", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_f16_f32 = clCreateKernel(backend_ctx->program, "kernel_mul_mat_f16_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_f16_f32_l4 = clCreateKernel(backend_ctx->program, "kernel_mul_mat_f16_f32_l4", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32 = clCreateKernel(backend_ctx->program, "kernel_mul_mat_q4_0_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_v = clCreateKernel(backend_ctx->program, "kernel_mul_mat_q4_0_f32_v", &err), err)); + + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_flat = clCreateKernel(backend_ctx->program, "kernel_mul_mat_q4_0_f32_flat", &err), err)); + CL_CHECK((backend_ctx->kernel_convert_block_q4_0 = clCreateKernel(backend_ctx->program, "kernel_convert_block_q4_0", &err), err)); + CL_CHECK((backend_ctx->kernel_restore_block_q4_0 = clCreateKernel(backend_ctx->program, "kernel_restore_block_q4_0", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_8x_flat = clCreateKernel(backend_ctx->program, "kernel_mul_mat_q4_0_f32_8x_flat", &err), err)); + + // Load additional mulmat kernels. +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src_1 { + #include "ggml-opencl_mm.cl.h" + }; +#else + const std::string kernel_src_1 = read_file("ggml-opencl_mm.cl"); +#endif + backend_ctx->program_1 = build_program_from_source(context, device, kernel_src_1.c_str(), compile_opts); + + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_1d_8x_flat = clCreateKernel(backend_ctx->program_1, "kernel_mul_mat_q4_0_f32_1d_8x_flat", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_1d_16x_flat = clCreateKernel(backend_ctx->program_1, "kernel_mul_mat_q4_0_f32_1d_16x_flat", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mv_q6_K_f32 = clCreateKernel(backend_ctx->program_1, "kernel_mul_mv_q6_K_f32", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_flat_v0 = clCreateKernel(backend_ctx->program_1, "kernel_mul_mat_q4_0_f32_flat_v0", &err), err)); + CL_CHECK((backend_ctx->kernel_mul_mat_q4_0_f32_flat_img_v0 = clCreateKernel(backend_ctx->program_1, "kernel_mul_mat_q4_0_f32_flat_img_v0", &err), err)); + + // Load additional data conversion kernels. +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src_2 { + #include "ggml-opencl_cvt.cl.h" + }; +#else + const std::string kernel_src_2 = read_file("ggml-opencl_cvt.cl"); +#endif + backend_ctx->program_2 = build_program_from_source(context, device, kernel_src_2.c_str(), compile_opts); + + CL_CHECK((backend_ctx->kernel_convert_block_q4_0_noshuffle = clCreateKernel(backend_ctx->program_2, "kernel_convert_block_q4_0_noshuffle", &err), err)); + + // Kernels for Adreno +#ifdef GGML_OPENCL_USE_ADRENO_KERNELS +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string transpose_32_src { + #include "ggml-opencl_transpose_32.cl.h" + }; +#else + const std::string transpose_32_src = read_file("ggml-opencl_transpose_32.cl"); +#endif + backend_ctx->program_transpose_32 = build_program_from_source(context, device, transpose_32_src.c_str(), compile_opts); + CL_CHECK((backend_ctx->kernel_transpose_32 = clCreateKernel(backend_ctx->program_transpose_32, "kernel_transpose_32", &err), err)); + +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string transpose_32_16_src { + #include "ggml-opencl_transpose_32_16.cl.h" + }; +#else + const std::string transpose_32_16_src = read_file("ggml-opencl_transpose_32_16.cl"); +#endif + backend_ctx->program_transpose_32_16 = build_program_from_source(context, device, transpose_32_16_src.c_str(), compile_opts); + CL_CHECK((backend_ctx->kernel_transpose_32_16 = clCreateKernel(backend_ctx->program_transpose_32_16, "kernel_transpose_32_16", &err), err)); + +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string transpose_16_src { + #include "ggml-opencl_transpose_16.cl.h" + }; +#else + const std::string transpose_16_src = read_file("ggml-opencl_transpose_16.cl"); +#endif + backend_ctx->program_transpose_16 = build_program_from_source(context, device, transpose_16_src.c_str(), compile_opts); + CL_CHECK((backend_ctx->kernel_transpose_16 = clCreateKernel(backend_ctx->program_transpose_16, "kernel_transpose_16", &err), err)); + + // Gemv general + std::string CL_gemv_compile_opts = + " -cl-std=CL2.0 " + " -cl-mad-enable " + " -DSIMDGROUP_WIDTH=" + std::to_string(backend_ctx->adreno_wave_size); + if (has_vector_subgroup_broadcast) { + CL_gemv_compile_opts += " -DVECTOR_SUB_GROUP_BROADCAT "; + } +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src_CL_gemv_general { + #include "ggml-opencl_gemv_noshuffle_general.cl.h" + }; +#else + const std::string kernel_src_CL_gemv_general = read_file("ggml-opencl_gemv_noshuffle_general.cl"); +#endif + + backend_ctx->program_CL_gemv_general = build_program_from_source( + context, device, kernel_src_CL_gemv_general.c_str(), CL_gemv_compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_general = clCreateKernel(backend_ctx->program_CL_gemv_general, "kernel_gemv_noshuffle", &err), err)); + + // Gemv 2048, 16384 + CL_gemv_compile_opts = + " -cl-std=CL2.0 " + " -cl-mad-enable " + " -DLINE_STRIDE_A=2048 " + " -DBLOCK_STRIDE_A=16384 " + " -DSIMDGROUP_WIDTH=" + std::to_string(backend_ctx->adreno_wave_size); + if (has_vector_subgroup_broadcast) { + CL_gemv_compile_opts += " -DVECTOR_SUB_GROUP_BROADCAT "; + } +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src_CL_gemv { + #include "ggml-opencl_gemv_noshuffle.cl.h" + }; +#else + const std::string kernel_src_CL_gemv = read_file("ggml-opencl_gemv_noshuffle.cl"); +#endif + + backend_ctx->program_CL_gemv_4096_1_4096 = build_program_from_source( + context, device, kernel_src_CL_gemv.c_str(), CL_gemv_compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_4096 = clCreateKernel(backend_ctx->program_CL_gemv_4096_1_4096, "kernel_gemv_noshuffle", &err), err)); + + // Gemv 2048, 16384 + CL_gemv_compile_opts = + " -cl-std=CL2.0 " + " -cl-mad-enable " + " -DLINE_STRIDE_A=2048 " + " -DBLOCK_STRIDE_A=16384 " + " -DSIMDGROUP_WIDTH=" + std::to_string(backend_ctx->adreno_wave_size); + if (has_vector_subgroup_broadcast) { + CL_gemv_compile_opts += " -DVECTOR_SUB_GROUP_BROADCAT "; + } + + backend_ctx->program_CL_gemv_4096_1_11008 = build_program_from_source( + context, device, kernel_src_CL_gemv.c_str(), CL_gemv_compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_11008 = clCreateKernel(backend_ctx->program_CL_gemv_4096_1_11008, "kernel_gemv_noshuffle", &err), err)); + + // Gemv 5504, 44032 + CL_gemv_compile_opts = + " -cl-std=CL2.0 " + " -cl-mad-enable " + " -DLINE_STRIDE_A=5504 " + " -DBLOCK_STRIDE_A=44032 " + " -DSIMDGROUP_WIDTH=" + std::to_string(backend_ctx->adreno_wave_size); + if (has_vector_subgroup_broadcast) { + CL_gemv_compile_opts += " -DVECTOR_SUB_GROUP_BROADCAT "; + } + + backend_ctx->program_CL_gemv_11008_1_4096 = build_program_from_source( + context, device, kernel_src_CL_gemv.c_str(), CL_gemv_compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_11008_1_4096 = clCreateKernel(backend_ctx->program_CL_gemv_11008_1_4096, "kernel_gemv_noshuffle", &err), err)); + + // Gemv 16000, 128000 + CL_gemv_compile_opts = + " -cl-std=CL2.0 " + " -cl-mad-enable " + " -DLINE_STRIDE_A=16000 " + " -DBLOCK_STRIDE_A=128000 " + " -DSIMDGROUP_WIDTH=" + std::to_string(backend_ctx->adreno_wave_size); + if (has_vector_subgroup_broadcast) { + CL_gemv_compile_opts += " -DVECTOR_SUB_GROUP_BROADCAT "; + } + + backend_ctx->program_CL_gemv_32000_1_4096 = build_program_from_source(context, device, kernel_src_CL_gemv.c_str(), CL_gemv_compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_32000_1_4096 = clCreateKernel(backend_ctx->program_CL_gemv_32000_1_4096, "kernel_gemv_noshuffle", &err), err)); + + // Gemm +#ifdef GGML_OPENCL_EMBED_KERNELS + const std::string kernel_src_CL_gemm { + #include "ggml-opencl_mul_mat_Ab_Bi_8x4.cl.h" + }; +#else + const std::string kernel_src_CL_gemm = read_file("ggml-opencl_mul_mat_Ab_Bi_8x4.cl"); +#endif + backend_ctx->program_CL_gemm = build_program_from_source(context, device, kernel_src_CL_gemm.c_str(), compile_opts); + CL_CHECK((backend_ctx->CL_mul_mat_Ab_Bi_8x4 = clCreateKernel(backend_ctx->program_CL_gemm, "kernel_mul_mat_Ab_Bi_8x4", &err), err)); + + // Allocate intermediate buffers and images + size_t max_A_q_d_bytes = 311164928; + size_t max_A_s_d_bytes = 38895616; + size_t max_B_d_bytes = 45088768; + + CL_CHECK((backend_ctx->A_q_d_max = clCreateBuffer(context, 0, max_A_q_d_bytes, NULL, &err), err)); + CL_CHECK((backend_ctx->A_s_d_max = clCreateBuffer(context, 0, max_A_s_d_bytes, NULL, &err), err)); + CL_CHECK((backend_ctx->B_d_max = clCreateBuffer(context, 0, max_B_d_bytes, NULL, &err), err)); +#endif // GGML_OPENCL_USE_ADRENO_KERNELS + + // For now we support a single devices + ggml_backend_opencl_n_devices = 1; + + return backend_ctx; +} + +static void ggml_cl2_free(void) { +#ifdef GGML_OPENCL_PROFILING + FILE * fperf = fopen("cl_profiling.csv", "w"); + if (!fperf) { + GGML_LOG_ERROR("Failed to open cl_profiling.csv\n"); + return; + } + + float total_kernel_time = 0; + fprintf(fperf, "op name, kernel name, duration (ms), global size, local size, output size\n"); + for (const ProfilingInfo & info : g_profiling_info) { + total_kernel_time += info.duration_ns/1.e6f; + fprintf(fperf, "%s,%s,%f,%zux%zux%zu,%zux%zux%zu,%zux%zux%zux%zu\n", + info.op_name.c_str(), info.kernel_name.c_str(), info.duration_ns/1.e6f, + info.global_size[0], info.global_size[1], info.global_size[2], + info.local_size[0], info.local_size[2], info.local_size[2], + info.output_size[0], info.output_size[1], info.output_size[2], info.output_size[3]); + } + fclose(fperf); + + GGML_LOG_INFO("ggml_opencl: total kernel time: %f\n", total_kernel_time); +#endif +} + +//------------------------------------------------------------------------------ +// Tensor extra management +//------------------------------------------------------------------------------ +struct ggml_tensor_extra_cl { + // The buffer object that holds the data. + cl_mem data_device; + // The offset into the buffer object. This is primarily for scratch buffer + // and view operation. + // NB: this offset no longer includes view offset (view_offs). Whenever this + // offset is used, view_offs should be considered. + cl_ulong offset; + // The actual size of the cl_mem object. This is needed when returning the + // block to the pool. + size_t actual_size; + + void reset() { + data_device = nullptr; + offset = 0; + actual_size = 0; + } +}; + +// Additional tensor extra structs for quantized tensors. +// These tensors are loaded from files and should not be allocated in scratch -- +// they should always be allocated from the pool. Hence, they do not have an +// `offset`, which indicate their locations in the scratch buffer. +struct ggml_tensor_extra_cl_q4_0 { + // Quantized values. + cl_mem q = nullptr; + // Quantized values in image1d_buffer_t. + cl_mem q_img = nullptr; + // Scales. + cl_mem d = nullptr; + // Scales in image1d_buffer_t. + cl_mem d_img = nullptr; + // Size of quantized values. + size_t size_q = 0; + // Size of scales. + size_t size_d = 0; + + ~ggml_tensor_extra_cl_q4_0() { + reset(); + } + + void reset() { + // q and d are subbuffers into the bigger buffer allocated in ggml_backend_buffer. + // They must be properly released so that the original buffer can be + // properly released to avoid memory leak. + if (q != nullptr) { + CL_CHECK(clReleaseMemObject(q)); + q = nullptr; + } + if (d != nullptr) { + CL_CHECK(clReleaseMemObject(d)); + d = nullptr; + } + // Currently, q_img and d_img are only initialized when SMALL_ALLOC is + // enabled. They point to the images in ggml_backend_opencl_buffer_context. + // So, there is no need to release them here. + // TODO: initialize them for non SMALL_PATH path, or remove them. + q_img = nullptr; + d_img = nullptr; + size_q = 0; + size_d = 0; + } +}; + +//------------------------------------------------------------------------------ +// Backend API +//------------------------------------------------------------------------------ + +// +// backend +// +static const char * ggml_backend_opencl_name(ggml_backend_t backend) { + return "OpenCL"; + + UNUSED(backend); +} + +static void ggml_backend_opencl_free(ggml_backend_t backend) { + ggml_cl2_free(); + + GGML_UNUSED(backend); +} + +static void ggml_backend_opencl_set_tensor_async(ggml_backend_t backend, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + GGML_UNUSED(backend); + GGML_UNUSED(tensor); + GGML_UNUSED(data); + GGML_UNUSED(offset); + GGML_UNUSED(size); +} + +static void ggml_backend_opencl_get_tensor_async(ggml_backend_t backend, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + GGML_UNUSED(backend); + GGML_UNUSED(tensor); + GGML_UNUSED(data); + GGML_UNUSED(offset); + GGML_UNUSED(size); +} + +static bool ggml_backend_opencl_cpy_tensor_async(ggml_backend_t backend, const ggml_tensor * src, ggml_tensor * dst) { + GGML_UNUSED(backend); + GGML_UNUSED(src); + GGML_UNUSED(dst); + return false; +} + +static void ggml_backend_opencl_synchronize(ggml_backend_t backend) { + GGML_UNUSED(backend); +} + +static ggml_status ggml_backend_opencl_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) { + for (int i = 0; i < cgraph->n_nodes; i++) { + ggml_tensor * node = cgraph->nodes[i]; + + if (node->op == GGML_OP_RESHAPE || node->op == GGML_OP_TRANSPOSE || node->op == GGML_OP_VIEW || node->op == GGML_OP_PERMUTE || node->op == GGML_OP_NONE) { + continue; + } + + bool ok = ggml_cl_compute_forward(backend, node); + if (!ok) { + GGML_LOG_ERROR("%s: error: op not supported %s (%s)\n", __func__, node->name, ggml_op_name(node->op)); + } + GGML_ASSERT(ok); + } + + return GGML_STATUS_SUCCESS; +} + +static bool ggml_opencl_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { + GGML_UNUSED(dev); + + switch (op->op) { + case GGML_OP_NONE: + return true; + case GGML_OP_GET_ROWS: + switch (op->src[0]->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + return true; + case GGML_TYPE_Q4_0: +#ifdef GGML_OPENCL_SOA_Q + // We do not support flattened Q4_0 (and possibly other Q's) + return false; +#else // GGML_OPENCL_SOA_Q + return true; +#endif // GGML_OPENCL_SOA_Q + default: + return false; + } + case GGML_OP_CPY: + case GGML_OP_DUP: + case GGML_OP_CONT: + switch (op->src[0]->type) { + case GGML_TYPE_F32: + switch (op->type) { + case GGML_TYPE_F16: + case GGML_TYPE_F32: + return true; + default: + return false; + } + case GGML_TYPE_F16: + switch (op->type) { + case GGML_TYPE_F16: + case GGML_TYPE_F32: + return true; + default: + return false; + } + default: + return false; + } + case GGML_OP_ADD: + case GGML_OP_SCALE: + case GGML_OP_MUL: + return true; + case GGML_OP_UNARY: + switch (ggml_get_unary_op(op)) { + case GGML_UNARY_OP_GELU: + case GGML_UNARY_OP_SILU: + case GGML_UNARY_OP_RELU: + return ggml_is_contiguous(op->src[0]); + default: + return false; + } + case GGML_OP_CLAMP: + case GGML_OP_SOFT_MAX: + case GGML_OP_NORM: + case GGML_OP_RMS_NORM: + return true; + case GGML_OP_MUL_MAT: + if (op->src[0]->type == GGML_TYPE_F16) { + return true; + } else if (op->src[0]->type == GGML_TYPE_F32) { + return op->src[1]->type == GGML_TYPE_F32 && ggml_is_contiguous(op->src[0]) && ggml_is_contiguous(op->src[1]); + } else if (op->src[0]->type == GGML_TYPE_Q4_0 || + op->src[0]->type == GGML_TYPE_Q6_K) { + return op->src[1]->type == GGML_TYPE_F32 && ggml_is_contiguous(op->src[0]) && ggml_is_contiguous(op->src[1]); + } + return false; + case GGML_OP_RESHAPE: + case GGML_OP_VIEW: + case GGML_OP_PERMUTE: + case GGML_OP_TRANSPOSE: + return true; + case GGML_OP_DIAG_MASK_INF: + return op->ne[3] == 1; + case GGML_OP_ROPE: + return true; + default: + return false; + } +} + +// Forward declaration - implementation appears later in the file. +static const char * ggml_backend_opencl_buffer_type_get_name(ggml_backend_buffer_type_t buffer_type); + +static ggml_guid_t ggml_backend_opencl_guid() { + static ggml_guid guid = { 0xde, 0xe0, 0x70, 0xa2, 0x73, 0x4e, 0x4d, 0xbc, 0xb0, 0xc7, 0x4f, 0xd4, 0x6d, 0x4e, 0x90, 0xfe }; + return &guid; +} + +static ggml_backend_i ggml_backend_opencl_i = { + /* .get_name = */ ggml_backend_opencl_name, + /* .free = */ ggml_backend_opencl_free, + /* .set_tensor_async = */ NULL, /* ggml_backend_opencl_set_tensor_async */ + /* .get_tensor_async = */ NULL, /* ggml_backend_opencl_get_tensor_async */ + /* .cpy_tensor_async = */ NULL, /* ggml_backend_opencl_cpy_tensor_async */ + /* .synchronize = */ NULL, /* ggml_backend_opencl_synchronize */ + /* .graph_plan_create = */ NULL, + /* .graph_plan_free = */ NULL, + /* .graph_plan_update = */ NULL, + /* .graph_plan_compute = */ NULL, + /* .graph_compute = */ ggml_backend_opencl_graph_compute, + /* .event_record = */ NULL, + /* .event_wait = */ NULL, +}; + +ggml_backend_t ggml_backend_opencl_init(void) { + ggml_backend_dev_t dev = ggml_backend_reg_dev_get(ggml_backend_opencl_reg(), 0); + ggml_backend_opencl_context *backend_ctx = ggml_cl2_init(dev); + + ggml_backend_t backend = new ggml_backend { + /* .guid = */ ggml_backend_opencl_guid(), + /* .interface = */ ggml_backend_opencl_i, + /* .device = */ dev, + /* .context = */ backend_ctx + }; + + return backend; +} + +bool ggml_backend_is_opencl(ggml_backend_t backend) { + return backend && backend->iface.get_name == ggml_backend_opencl_name; +} + +// +// buffer +// +struct ggml_backend_opencl_buffer_context { + // A buffer context can hold multiple cl_mem objects. This is for flattening + // quantized weights and should be used with GGML_OPENCL_SMALL_ALLOC where + // each tensor is allocated a separate buffer. When flattening is enabled + // with small allocation, each tensor is backed by two cl_mem objects (for + // quants and scales) packed into a backend_opencl_buffer. + ggml_backend_opencl_buffer_context(cl_mem buf) + : name("OpenCL") { + buffer.push_back(buf); + } + + ~ggml_backend_opencl_buffer_context() { + for (cl_mem buf : buffer) { + CL_CHECK(clReleaseMemObject(buf)); + } + for (cl_mem im : img) { + CL_CHECK(clReleaseMemObject(im)); + } + + // Delete all extras to trigger their destructors + for (ggml_tensor_extra_cl * e : temp_tensor_extras) { + delete e; + } + for (ggml_tensor_extra_cl * e : temp_tensor_extras_in_use) { + delete e; + } + for (ggml_tensor_extra_cl_q4_0 * e : temp_tensor_extras_q4_0) { + delete e; + } + for (ggml_tensor_extra_cl_q4_0 * e : temp_tensor_extras_q4_0_in_use) { + delete e; + } + } + + ggml_tensor_extra_cl * ggml_opencl_alloc_temp_tensor_extra() { + ggml_tensor_extra_cl * extra; + if (temp_tensor_extras.empty()) { + extra = new ggml_tensor_extra_cl(); + } else { + extra = temp_tensor_extras.back(); + temp_tensor_extras.pop_back(); + } + + temp_tensor_extras_in_use.push_back(extra); + + extra->reset(); + return extra; + } + + ggml_tensor_extra_cl_q4_0 * ggml_opencl_alloc_temp_tensor_extra_q4_0() { + ggml_tensor_extra_cl_q4_0 * extra; + if (temp_tensor_extras_q4_0.empty()) { + extra = new ggml_tensor_extra_cl_q4_0(); + } else { + extra = temp_tensor_extras_q4_0.back(); + temp_tensor_extras_q4_0.pop_back(); + } + + temp_tensor_extras_q4_0_in_use.push_back(extra); + + extra->reset(); + return extra; + } + + void reset() { + for (ggml_tensor_extra_cl * e : temp_tensor_extras_in_use) { + temp_tensor_extras.push_back(e); + } + temp_tensor_extras_in_use.clear(); + + for (ggml_tensor_extra_cl_q4_0 * e : temp_tensor_extras_q4_0_in_use) { + temp_tensor_extras_q4_0.push_back(e); + } + temp_tensor_extras_q4_0_in_use.clear(); + } + + // Pools for extras. Available extras are in `temp_tensor_extras`. Extras + // being used are in `temp_tensor_extras_in_use`. At the first run, new + // extras get created and put in `in_use`. When the buffer is reset via + // the `reset` callback, all extras in `in_use` get moved to available extras + // for reuse. + std::vector temp_tensor_extras; + std::vector temp_tensor_extras_in_use; + std::vector temp_tensor_extras_q4_0; + std::vector temp_tensor_extras_q4_0_in_use; + + // The buffer_context is initially created by ggml_backend_buft_alloc_buffer + // before any tensor is initialized (at the beginning of alloc_tensor_range). + // Hence, there is alway a buffer object in this vector. When each tensor is + // being initialized, this original buffer object will be released if both + // flattening and small allocation are enabled, and additional buffer + // objects will be created in init_tensor to represent flattened quantized + // weights. + std::vector buffer; + // These are image1d_buffer_t objects that wrap around the quants and scales. + // For Q4_0 quantization, there should be two of them - one for quants and + // one for scales. They should be populated only when flattening and small + // allocation are enabled. + std::vector img; + std::string name; +}; + +static void * const cl_ptr_base = (void *)(uintptr_t) 0x1000; + +static void ggml_backend_opencl_buffer_free_buffer(ggml_backend_buffer_t buffer) { + ggml_backend_opencl_buffer_context * ctx = (ggml_backend_opencl_buffer_context *) buffer->context; + delete ctx; +} + +static void * ggml_backend_opencl_buffer_get_base(ggml_backend_buffer_t buffer) { + return cl_ptr_base; + + GGML_UNUSED(buffer); +} + +static void ggml_backend_opencl_buffer_init_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor) { + ggml_backend_opencl_buffer_context * ctx = (ggml_backend_opencl_buffer_context *) buffer->context; + + ggml_cl2_init(buffer->buft->device); + + if (tensor->view_src != nullptr) { + GGML_ASSERT(tensor->view_src->buffer->buft == buffer->buft); + + ggml_tensor_extra_cl * view_extra = (ggml_tensor_extra_cl *) tensor->view_src->extra; + GGML_ASSERT(view_extra && "view_extra is nullptr?"); + + // Reuse extra of the parent tensor. The offset of this view tensor + // becomes `extra->offset + view_offs` and needs to be calculated when + // it is used. This changes is needed because of the change to + // ggml_alloc.c in https://github.com/ggerganov/llama.cpp/pull/7640. + // `buffer` passed in here will always be `tensor->buffer`. It is OK + // to allocate extras from the same buffer context for ordinary + // intermediate tensors. But for views into kv cache tensors, doing so + // would mess up the extras used by kv cache. + // Before #7640, `buffer` is for intermediate tensors, which is always + // different from that of kv cache tensors. + // + // NB: now extra->offset no longer accounts for view_offs. + // NB: this should not apply to weight tensors (for end-to-end runs, but + // may apply for test-backend-ops). + // FIXME: if any unexpected results are seen, double check the offset - + // there could be other places that need fix. + tensor->extra = view_extra; + } else { + { + size_t offset = (char *)tensor->data - (char *)cl_ptr_base; + + ggml_tensor_extra_cl * extra = ctx->ggml_opencl_alloc_temp_tensor_extra(); + extra->offset = offset; + extra->data_device = ctx->buffer[0]; + extra->actual_size = ggml_nbytes(tensor); + + tensor->extra = extra; + } + } +} + +// The optimized gemm and gemv kernels are used for large matrices without batch. +// tensor is the quantized weights matrix. +inline bool use_adreno_kernels(const ggml_tensor *tensor) { + return tensor->ne[0] >= 512 && tensor->ne[1] >= 512 && + tensor->ne[2] == 1 && tensor->ne[3] == 1; +} + +static void ggml_backend_opencl_buffer_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_backend_opencl_context *backend_ctx = ggml_cl2_init(buffer->buft->device); + + cl_context context = backend_ctx->context; + cl_command_queue queue = backend_ctx->queue; + +#ifdef GGML_OPENCL_SOA_Q + // We separate the quantized bits and scale from block_q4_0 by using an + // additional kernel, where each thread handles a block. We first read the + // original weights into a temporary buffer, then create two separate + // buffers for quantized bits and scales, which are then populated by the + // conversion kernel. + if (tensor->type == GGML_TYPE_Q4_0) { + // Tensors should have been preallocated, therefore they should + // already have ggml_tensor_extra_cl as extra. + ggml_tensor_extra_cl * extra_orig = (ggml_tensor_extra_cl *)tensor->extra; + GGML_ASSERT(extra_orig && "Tesnors in OpenCL backend should have been allocated and initialized"); + + // Allocate the new extra and create aliases from the original. + ggml_backend_opencl_buffer_context * ctx = (ggml_backend_opencl_buffer_context *) buffer->context; + ggml_tensor_extra_cl_q4_0 * extra = ctx->ggml_opencl_alloc_temp_tensor_extra_q4_0(); + + size_t size_d = ggml_nelements(tensor)/ggml_blck_size(tensor->type)*sizeof(ggml_fp16_t); + size_t size_q = ggml_nelements(tensor)/ggml_blck_size(tensor->type)*ggml_blck_size(tensor->type)/2; + GGML_ASSERT(size_d + size_q == ggml_nbytes(tensor) && "Incorrect tensor size"); + + cl_int err; + cl_mem data_device = clCreateBuffer(context, CL_MEM_READ_WRITE, + ggml_nbytes(tensor), NULL, &err); + CL_CHECK(err); + CL_CHECK(clEnqueueWriteBuffer( + queue, data_device, CL_TRUE, 0, + ggml_nbytes(tensor), data, 0, NULL, NULL)); + + // We consider the specified offset arg as always, although For weights + // the offset arg should be 0 (we do not assert this). + //GGML_ASSERT(offset == 0); + + // We create subbuffers from the original tensor buffer for scales and + // quants - i.e., scales and quants are aliases into the buffer obejct + // that backs the original tensor. This is a cleaner way to adapt to the + // new memory management. + // In the old code, we allocate new buffers for scales and quants + // respectively, which could still be done but would result in double + // allocation; properly deallocating the preallocated buffer that backs + // the tensors is tricky and would leak the backend specific information + // into the general backend code. + // Does this create misaligned subbuffers (alignment is 1024) in certain + // cases ? + cl_buffer_region region; + + // The original tensor memory is divided into scales and quants, i.e., + // we first store scales, then quants. + // Create subbuffer for scales. + region.origin = extra_orig->offset + tensor->view_offs + offset; + region.size = size_d; + extra->d = clCreateSubBuffer( + extra_orig->data_device, CL_MEM_READ_WRITE, + CL_BUFFER_CREATE_TYPE_REGION, ®ion, &err); + CL_CHECK(err); + + // Create subbuffer for quants. + region.origin = extra_orig->offset + tensor->view_offs + offset + size_d; + region.size = size_q; + extra->q = clCreateSubBuffer( + extra_orig->data_device, CL_MEM_READ_WRITE, + CL_BUFFER_CREATE_TYPE_REGION, ®ion, &err); + CL_CHECK(err); + + //cl_kernel kernel = backend_ctx->kernel_convert_block_q4_0; + #ifdef GGML_OPENCL_USE_ADRENO_KERNELS + cl_kernel kernel = backend_ctx->kernel_convert_block_q4_0; + + // The optimized kernels need weights in natural order, so unshuffle. + if (use_adreno_kernels(tensor)) { + kernel = backend_ctx->kernel_convert_block_q4_0_noshuffle; + } + #else + cl_kernel kernel = backend_ctx->kernel_convert_block_q4_0; + #endif // GGML_OPENCL_USE_ADRENO_KERNELS + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra->q)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra->d)); + + size_t global_work_size[] = {(size_t)ggml_nelements(tensor)/ggml_blck_size(tensor->type), 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + CL_CHECK(clReleaseMemObject(data_device)); + + tensor->extra = extra; + + // transpose the weights and scales + #ifdef GGML_OPENCL_USE_ADRENO_KERNELS + // Only do transpose for large, non batched matrix + // TODO: use preallocated images instead of sub-buffer then image + if (use_adreno_kernels(tensor)) { + // <----------------------------------------------------------------------------------> // + // start transpose + // <----------------------------------------------------------------------------------> // + int M = tensor->ne[1]; // ne01 + int K = tensor->ne[0]; // ne00 + + // transpose is out of place, so we need to allocate transposed buffers + // <----------------------------------------------------------------------------------> // + // use sub_buffer of max buffer size instead + + size_t q_size_bytes = K * M / 8 * sizeof(float); + cl_buffer_region region; + region.origin = 0; + region.size = q_size_bytes; + cl_mem qT_d = clCreateSubBuffer( + backend_ctx->A_q_d_max, + 0, + CL_BUFFER_CREATE_TYPE_REGION, + ®ion, + &err); + // cl_mem qT_d = clCreateBuffer(context, CL_MEM_READ_WRITE, q_size_bytes, NULL, &err); + CL_CHECK(err); + + // size_t d_size_bytes = M * (K / 32) / 2 * sizeof(float); + size_t d_size_bytes = M * (K / 32) * 2; + region.origin = 0; + region.size = d_size_bytes; + cl_mem dT_d = clCreateSubBuffer( + backend_ctx->A_s_d_max, + 0, + CL_BUFFER_CREATE_TYPE_REGION, + ®ion, + &err); + // cl_mem dT_d = clCreateBuffer(context, CL_MEM_READ_WRITE, d_size_bytes, NULL, &err); + CL_CHECK(err); + + // <----------------------------------------------------------------------------------> // + + + // create images from the buffers + // <----------------------------------------------------------------------------------> // + cl_mem q_d_image1D; + cl_mem d_d_image1D; + cl_mem qT_d_image1D; + cl_mem dT_d_image1D; + + cl_image_format img_fmt_1d = { CL_RGBA, CL_FLOAT }; + cl_image_desc img_desc_1d; + + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.image_width = M * K / 8 / 4; + img_desc_1d.buffer = extra->q; + q_d_image1D = clCreateImage(context, 0, &img_fmt_1d, &img_desc_1d, NULL, &err); + CL_CHECK(err); + + img_fmt_1d = { CL_RGBA, CL_FLOAT }; + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.image_width = M * K / 8 / 4; + img_desc_1d.buffer = qT_d; + qT_d_image1D = clCreateImage(context, 0, &img_fmt_1d, &img_desc_1d, NULL, &err); + CL_CHECK(err); + + img_fmt_1d = { CL_RGBA, CL_FLOAT }; + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.image_width = M * K / 32 / 4 / 2; + img_desc_1d.buffer = extra->d; + d_d_image1D = clCreateImage(context, 0, &img_fmt_1d, &img_desc_1d, NULL, &err); + CL_CHECK(err); + + img_fmt_1d = { CL_RGBA, CL_FLOAT }; + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.image_width = M * K / 32 / 4 / 2; + img_desc_1d.buffer = dT_d; + dT_d_image1D = clCreateImage(context, 0, &img_fmt_1d, &img_desc_1d, NULL, &err); + CL_CHECK(err); + // <----------------------------------------------------------------------------------> // + + // set up and call the transpose kernels + // <----------------------------------------------------------------------------------> // + // weights + int height_q = M / 8; + int width_q = K / 8 / 4; + kernel = backend_ctx->kernel_transpose_16; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &q_d_image1D)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &qT_d_image1D)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(int), &height_q)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(int), &width_q)); + + size_t local_size_q[3] = {4, 16, 1}; + size_t global_size_q[3] = {static_cast(width_q), static_cast(height_q), 1}; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_size_q, local_size_q, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + + // scales + int height_s = M / 8; + int width_s = K / 32 / 8; + + kernel = backend_ctx->kernel_transpose_16; + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &d_d_image1D)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &dT_d_image1D)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(int), &height_s)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(int), &width_s)); + + size_t local_size_s[3] = {4, 16, 1}; + size_t global_size_s[3] = {static_cast(width_s), static_cast(height_s), 1}; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_size_s, local_size_s, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + // <----------------------------------------------------------------------------------> // + + // copy transposed buffer contents to original buffers + // <----------------------------------------------------------------------------------> // + // weights + CL_CHECK(clEnqueueCopyBuffer(queue, qT_d, extra->q, 0, 0, q_size_bytes, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + + // scales + CL_CHECK(clEnqueueCopyBuffer(queue, dT_d, extra->d, 0, 0, d_size_bytes, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + // <----------------------------------------------------------------------------------> // + + // deallocate transpose buffers + // <----------------------------------------------------------------------------------> // + CL_CHECK(clReleaseMemObject(qT_d)); + CL_CHECK(clReleaseMemObject(dT_d)); + + // deallocate temporary images + CL_CHECK(clReleaseMemObject(q_d_image1D)); + CL_CHECK(clReleaseMemObject(d_d_image1D)); + CL_CHECK(clReleaseMemObject(qT_d_image1D)); + CL_CHECK(clReleaseMemObject(dT_d_image1D)); + // <----------------------------------------------------------------------------------> // + // end transpose + // <----------------------------------------------------------------------------------> // + } + #endif // GGML_OPENCL_USE_ADRENO_KERNELS + + return; + } +#endif // GGML_OPENCL_SOA_Q + + ggml_tensor_extra_cl * extra = (ggml_tensor_extra_cl *) tensor->extra; + GGML_ASSERT(extra); + + CL_CHECK(clEnqueueWriteBuffer( + queue, extra->data_device, CL_TRUE, extra->offset + offset, + size, data, 0, NULL, NULL)); + + GGML_UNUSED(buffer); +} + +static void ggml_backend_opencl_buffer_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + GGML_ASSERT(tensor->extra); + + ggml_backend_opencl_context *backend_ctx = ggml_cl2_init(buffer->buft->device); + + cl_context context = backend_ctx->context; + cl_command_queue queue = backend_ctx->queue; + + // Make sure all previously submitted commands are finished. + CL_CHECK(clFinish(queue)); + +#ifdef GGML_OPENCL_SOA_Q + // In end-to-end runs, get_tensor is usually used to get back the logits, + // where we can simply do clEnqueueReadBuffer since they are f32. + // However, in test-backend-ops, the GPU graph is copied to the CPU backend, + // which requires reading back quantized weight tensors. + // To properly support this, we need to restore block_q4_0 struct arrays + // from the flattened buffers. + if (tensor->type == GGML_TYPE_Q4_0) { + ggml_tensor_extra_cl_q4_0 * extra = (ggml_tensor_extra_cl_q4_0 *)tensor->extra; + + cl_int err; + cl_mem data_device = clCreateBuffer(context, CL_MEM_READ_WRITE, + ggml_nbytes(tensor), NULL, &err); + CL_CHECK(err); + + cl_kernel kernel = backend_ctx->kernel_restore_block_q4_0; + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra->q)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra->d)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &data_device)); + + size_t global_work_size[] = {(size_t)ggml_nelements(tensor)/ggml_blck_size(tensor->type), 1, 1}; + size_t local_work_size[] = {1, 1, 1}; + + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, + global_work_size, local_work_size, 0, NULL, &evt)); + CL_CHECK(clWaitForEvents(1, &evt)); + CL_CHECK(clEnqueueReadBuffer( + queue, data_device, CL_TRUE, offset, + size, data, 0, NULL, NULL)); + CL_CHECK(clReleaseMemObject(data_device)); + return; + } +#endif // GGML_OPENCL_SOA_Q + + ggml_tensor_extra_cl * extra = (ggml_tensor_extra_cl *) tensor->extra; + + CL_CHECK(clEnqueueReadBuffer( + queue, extra->data_device, CL_TRUE, extra->offset + tensor->view_offs + offset, + size, data, 0, NULL, NULL)); + + GGML_UNUSED(buffer); +} + +static void ggml_backend_opencl_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + ggml_backend_dev_t dev = buffer->buft->device; + ggml_backend_opencl_context *backend_ctx = ggml_cl2_init(dev); + cl_command_queue queue = backend_ctx->queue; + + ggml_backend_opencl_buffer_context * ctx = (ggml_backend_opencl_buffer_context *) buffer->context; + for (cl_mem buf : ctx->buffer) { + CL_CHECK(clEnqueueFillBuffer(queue, buf, &value, sizeof(value), 0, buffer->size, 0, NULL, NULL)); + } + CL_CHECK(clFinish(queue)); +} + +static void ggml_backend_opencl_buffer_reset(ggml_backend_buffer_t buffer) { + ggml_backend_opencl_buffer_context * ctx = (ggml_backend_opencl_buffer_context *) buffer->context; + ctx->reset(); +} + +static ggml_backend_buffer_i ggml_backend_opencl_buffer_interface = { + /* .free_buffer = */ ggml_backend_opencl_buffer_free_buffer, + /* .get_base = */ ggml_backend_opencl_buffer_get_base, + /* .init_tensor = */ ggml_backend_opencl_buffer_init_tensor, + /* .memset_tensor = */ NULL, + /* .set_tensor = */ ggml_backend_opencl_buffer_set_tensor, + /* .get_tensor = */ ggml_backend_opencl_buffer_get_tensor, + /* .cpy_tensor = */ NULL, + /* .clear = */ ggml_backend_opencl_buffer_clear, + /* .reset = */ ggml_backend_opencl_buffer_reset, +}; + +// +// buffer type +// + +static const char * ggml_backend_opencl_buffer_type_get_name(ggml_backend_buffer_type_t buffer_type) { + return "OpenCL"; + + GGML_UNUSED(buffer_type); +} + +static ggml_backend_buffer_t ggml_backend_opencl_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buffer_type, size_t size) { + ggml_backend_opencl_context *backend_ctx = ggml_cl2_init(buffer_type->device); + + // clCreateBuffer returns -61 for size 0 + size = std::max(size, (size_t)1); + + cl_int err; + cl_mem mem = clCreateBuffer(backend_ctx->context, CL_MEM_READ_WRITE, size, NULL, &err); + if (err != CL_SUCCESS) { + GGML_LOG_INFO("%s: failed to allocate %.2f MiB\n", __func__, size / 1024.0 / 1024.0); + return nullptr; + } + + ggml_backend_opencl_buffer_context * ctx = new ggml_backend_opencl_buffer_context(mem); + + return ggml_backend_buffer_init(buffer_type, ggml_backend_opencl_buffer_interface, ctx, size); +} + +static size_t ggml_backend_opencl_buffer_type_get_alignment(ggml_backend_buffer_type_t buffer_type) { + // FIXME: not thread safe, device may not be initialized yet + static cl_uint alignment = -1; + if (alignment == (cl_uint)-1) { + ggml_backend_opencl_context * backend_ctx = ggml_cl2_init(buffer_type->device); + alignment = backend_ctx->alignment; + } + return alignment; +} + +static size_t ggml_backend_opencl_buffer_type_get_max_size(ggml_backend_buffer_type_t buffer_type) { + static size_t max_size = -1; + if (max_size == (size_t)-1) { + ggml_backend_opencl_context * backend_ctx = ggml_cl2_init(buffer_type->device); + max_size = backend_ctx->max_alloc_size; + } + return max_size; +} + +static bool ggml_backend_opencl_buffer_type_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend) { + return ggml_backend_is_opencl(backend); + + UNUSED(buft); +} + +static ggml_backend_buffer_type_i ggml_backend_opencl_buffer_type_interface = { + /* .get_name = */ ggml_backend_opencl_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_opencl_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_opencl_buffer_type_get_alignment, + /* .get_max_size = */ ggml_backend_opencl_buffer_type_get_max_size, + /* .get_alloc_size = */ NULL, + /* .is_host = */ NULL, +}; + +ggml_backend_buffer_type_t ggml_backend_opencl_buffer_type() { + static ggml_backend_buffer_type buffer_type = { + /* .iface = */ ggml_backend_opencl_buffer_type_interface, + /* .device = */ &g_ggml_backend_opencl_device, + /* .context = */ nullptr, + }; + + return &buffer_type; +} + +// +// backend device +// + +static const char * ggml_backend_opencl_device_get_name(ggml_backend_dev_t dev) { + return "GPUOpenCL"; + + GGML_UNUSED(dev); +} + +static const char * ggml_backend_opencl_device_get_description(ggml_backend_dev_t dev) { + ggml_backend_opencl_device_context *dev_ctx = (ggml_backend_opencl_device_context *) dev->context; + return dev_ctx->device_name.c_str(); +} + +static void ggml_backend_opencl_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { + *free = 1; + *total = 1; + + GGML_UNUSED(dev); +} + +static enum ggml_backend_dev_type ggml_backend_opencl_device_get_type(ggml_backend_dev_t dev) { + return GGML_BACKEND_DEVICE_TYPE_GPU; + + GGML_UNUSED(dev); +} + +static void ggml_backend_opencl_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { + props->name = ggml_backend_opencl_device_get_name(dev); + props->description = ggml_backend_opencl_device_get_description(dev); + props->type = ggml_backend_opencl_device_get_type(dev); + ggml_backend_opencl_device_get_memory(dev, &props->memory_free, &props->memory_total); + props->caps = ggml_backend_dev_caps { + /* .async = */ false, + /* .host_buffer = */ false, + /* .buffer_from_host_ptr = */ false, + /* .events = */ false, + }; +} + +static ggml_backend_t ggml_backend_opencl_device_init(ggml_backend_dev_t dev, const char * params) { + ggml_backend_opencl_context * backend_ctx = ggml_cl2_init(dev); + + ggml_backend_t backend = new ggml_backend { + /* .guid = */ ggml_backend_opencl_guid(), + /* .interface = */ ggml_backend_opencl_i, + /* .device = */ dev, + /* .context = */ backend_ctx, + }; + + return backend; + + GGML_UNUSED(params); +} + +static ggml_backend_buffer_type_t ggml_backend_opencl_device_get_buffer_type(ggml_backend_dev_t dev) { + return ggml_backend_opencl_buffer_type(); + + GGML_UNUSED(dev); +} + +static ggml_backend_buffer_t ggml_backend_opencl_device_buffer_from_ptr(ggml_backend_dev_t dev, void * ptr, size_t size, size_t max_tensor_size) { + GGML_UNUSED(dev); + GGML_UNUSED(ptr); + GGML_UNUSED(size); + GGML_UNUSED(max_tensor_size); + return nullptr; +} + +static bool ggml_backend_opencl_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { + return ggml_opencl_supports_op(dev, op); +} + +static bool ggml_backend_opencl_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { + return buft->iface.get_name == ggml_backend_opencl_buffer_type_get_name; + + GGML_UNUSED(dev); +} + +static struct ggml_backend_device_i ggml_backend_opencl_device_i = { + /* .get_name = */ ggml_backend_opencl_device_get_name, + /* .get_description = */ ggml_backend_opencl_device_get_description, + /* .get_memory = */ ggml_backend_opencl_device_get_memory, + /* .get_type = */ ggml_backend_opencl_device_get_type, + /* .get_props = */ ggml_backend_opencl_device_get_props, + /* .init_backend = */ ggml_backend_opencl_device_init, + /* .get_buffer_type = */ ggml_backend_opencl_device_get_buffer_type, + /* .get_host_buffer_type = */ NULL, + /* .buffer_from_host_ptr = */ ggml_backend_opencl_device_buffer_from_ptr, + /* .supports_op = */ ggml_backend_opencl_device_supports_op, + /* .supports_buft = */ ggml_backend_opencl_device_supports_buft, + /* .offload_op = */ NULL, + /* .event_new = */ NULL, + /* .event_free = */ NULL, + /* .event_synchronize = */ NULL, +}; + +// Backend registry + +static const char * ggml_backend_opencl_reg_get_name(ggml_backend_reg_t reg) { + return "OpenCL"; + + GGML_UNUSED(reg); +} + +static size_t ggml_backend_opencl_reg_device_count(ggml_backend_reg_t reg) { + return ggml_backend_opencl_n_devices; + + GGML_UNUSED(reg); +} + +static ggml_backend_dev_t ggml_backend_opencl_reg_device_get(ggml_backend_reg_t reg, size_t index) { + GGML_ASSERT(index == 0); + + return &g_ggml_backend_opencl_device; + + GGML_UNUSED(reg); + GGML_UNUSED(index); +} + +static struct ggml_backend_reg_i ggml_backend_opencl_reg_i = { + /* .get_name = */ ggml_backend_opencl_reg_get_name, + /* .device_count = */ ggml_backend_opencl_reg_device_count, + /* .device_get = */ ggml_backend_opencl_reg_device_get, + /* .get_proc_address = */ NULL, +}; + +ggml_backend_reg_t ggml_backend_opencl_reg(void) { + // TODO: make this thread-safe somehow? + static ggml_backend_reg reg; + static bool initialized = false; + + if (!initialized) { + reg = ggml_backend_reg { + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_opencl_reg_i, + /* .context = */ NULL, + }; + + g_ggml_backend_opencl_device = ggml_backend_device { + /* .iface = */ ggml_backend_opencl_device_i, + /* .reg = */ ®, + /* .context = */ &g_ggml_ctx_dev_main, + }; + + ggml_cl2_init(&g_ggml_backend_opencl_device); + + initialized = true; + } + + return ® +} + +GGML_BACKEND_DL_IMPL(ggml_backend_opencl_reg) + +//------------------------------------------------------------------------------ +// Debugging utils +//------------------------------------------------------------------------------ +#if 0 +#define QK4_0 32 +typedef struct { + ggml_fp16_t d; // delta + uint8_t qs[QK4_0 / 2]; // nibbles / quants +} block_q4_0; +static_assert(sizeof(block_q4_0) == sizeof(ggml_fp16_t) + QK4_0 / 2, + "wrong q4_0 block size/padding"); + +#include +#ifdef __cplusplus +#include "half.hpp" +#endif + +static void dump_tensor(ggml_backend_t backend, const struct ggml_tensor * tensor) { + void * buf = malloc(ggml_nbytes(tensor)); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; +#ifdef GGML_OPENCL_SOA_Q + void * buf_q; + void * buf_d; +#endif + +#ifdef GGML_USE_OPENCL + // Make sure everything is done. + CL_CHECK(clFinish(queue)); + +#ifdef GGML_OPENCL_SOA_Q + if (tensor->type == GGML_TYPE_Q4_0) { + ggml_tensor_extra_cl_q4_0 * extra = (ggml_tensor_extra_cl_q4_0 *) tensor->extra; + GGML_ASSERT(extra); + + size_t size_q = ggml_nelements(tensor)/QK4_0 * QK4_0/2; + size_t size_d = ggml_nelements(tensor)/QK4_0 * sizeof(ggml_fp16_t); + GGML_ASSERT(size_q + size_d == ggml_nbytes(tensor)); + buf_q = malloc(size_q); + buf_d = malloc(size_d); + + CL_CHECK(clEnqueueReadBuffer(queue, extra->q, CL_TRUE, 0, size_q, buf_q, 0, NULL, NULL)); + CL_CHECK(clEnqueueReadBuffer(queue, extra->d, CL_TRUE, 0, size_d, buf_d, 0, NULL, NULL)); + CL_CHECK(clFinish(queue)); + } else { + // Read out the tensor from GPU memory. + ggml_tensor_extra_cl * extra = (ggml_tensor_extra_cl *) tensor->extra; + GGML_ASSERT(extra); + + CL_CHECK(clEnqueueReadBuffer(queue, extra->data_device, CL_TRUE, + extra->offset, ggml_nbytes(tensor), buf, 0, NULL, NULL)); + CL_CHECK(clFinish(queue)); + } +#else + // Read out the tensor from GPU memory. + ggml_tensor_extra_cl * extra = (ggml_tensor_extra_cl *) tensor->extra; + GGML_ASSERT(extra); + + CL_CHECK(clEnqueueReadBuffer(queue, extra->data_device, CL_TRUE, + extra->offset, ggml_nbytes(tensor), buf, 0, NULL, NULL)); + CL_CHECK(clFinish(queue)); +#endif // GGML_OPENCL_SOA_Q +#endif // GGML_USE_OPENCL + + // Open file and dump. + char fname[512]; + sprintf(fname, "./tensor-dumps/%s.txt", tensor->name); + FILE * f = fopen(fname, "w"); + if (!f) { + printf("Failed to open %s\n", fname); + return; + } + + if (tensor->type == GGML_TYPE_F32) { + float * data = (float *) buf; + for (int i = 0; i < ggml_nelements(tensor); ++i) { + if (isnan(data[i])) { + printf("NaN found: %s\n", tensor->name); + break; + } + fprintf(f, "%f\n", data[i]); + } + } else if (tensor->type == GGML_TYPE_I32) { + int * data = (int *) buf; + for (int i = 0; i < ggml_nelements(tensor); ++i) { + if (isnan(data[i])) { + printf("NaN found: %s\n", tensor->name); + break; + } + fprintf(f, "%d\n", data[i]); + } + } else if (tensor->type == GGML_TYPE_F16) { +#ifdef __cplusplus + half_float::half * data = (half_float::half *) buf; + for (int i = 0; i < ggml_nelements(tensor); ++i) { + if (std::isnan(data[i])) { + printf("NaN found: %s\n", tensor->name); + break; + } + fprintf(f, "%f\n", float(data[i])); + } +#endif + } else if (tensor->type == GGML_TYPE_Q4_0) { +#ifdef GGML_OPENCL_SOA_Q + ggml_fp16_t * data_d = (ggml_fp16_t *)buf_d; + unsigned char * data_q = (unsigned char *)buf_q; + + for (int i = 0; i < ggml_nelements(tensor)/QK4_0; ++i) { + fprintf(f, "%04x, ", data_d[i]); + for (int k = 0; k < QK4_0/2; ++k) { + fprintf(f, "%02x, ", data_q[k]); + } + fprintf(f, "\n"); + data_q += QK4_0/2; + } + free(buf_d); + free(buf_q); +#else + block_q4_0 * data = (block_q4_0 *) buf; + for (int i = 0; i < ggml_nelements(tensor)/QK4_0; ++i) { + fprintf(f, "%04x, ", data[i].d); + for (int k = 0; k < QK4_0/2; ++k) { + fprintf(f, "%02x, ", data[i].qs[k]); + } + fprintf(f, "\n"); + } +#endif // GGML_OPENCL_SOA_Q + } + free(buf); + fflush(f); + fclose(f); +} +#else +#define dump_tensor(tensor) +#endif + +//------------------------------------------------------------------------------ +// Profiling utility +//------------------------------------------------------------------------------ +#ifdef GGML_OPENCL_PROFILING +void populateProfilingInfo( + ProfilingInfo& info, cl_event evt, cl_kernel kernel, + size_t global_size[3], size_t local_size[3], + const ggml_tensor * tensor) { + cl_ulong start; + cl_ulong end; + CL_CHECK(clWaitForEvents(1, &evt)); + CL_CHECK(clGetEventProfilingInfo( + evt, CL_PROFILING_COMMAND_START, sizeof(cl_ulong), &start, NULL)); + CL_CHECK(clGetEventProfilingInfo( + evt, CL_PROFILING_COMMAND_END, sizeof(cl_ulong), &end, NULL)); + + char kernel_name[512]; + CL_CHECK(clGetKernelInfo(kernel, CL_KERNEL_FUNCTION_NAME, + sizeof(kernel_name), kernel_name, NULL)); + + info.duration_ns = end - start; + info.op_name = tensor->name; + info.kernel_name = kernel_name; + info.local_size[0] = local_size[0]; + info.local_size[1] = local_size[1]; + info.local_size[2] = local_size[2]; + info.global_size[0] = global_size[0]; + info.global_size[1] = global_size[1]; + info.global_size[2] = global_size[2]; + info.output_size[0] = tensor->ne[0]; + info.output_size[1] = tensor->ne[1]; + info.output_size[2] = tensor->ne[2]; + info.output_size[3] = tensor->ne[3]; +} +#endif + +//------------------------------------------------------------------------------ +// Ops +//------------------------------------------------------------------------------ + +static bool ggml_cl_can_mul_mat(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst) { + const int64_t ne10 = src1->ne[0]; + + const int64_t ne0 = dst->ne[0]; + const int64_t ne1 = dst->ne[1]; + + // TODO: find the optimal values for these + return (src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16 || ggml_is_quantized(src0->type)) && + src1->type == GGML_TYPE_F32 && + dst->type == GGML_TYPE_F32 && + (ne0 >= 32 && ne1 >= 32 && ne10 >= 32); +} + +static void ggml_cl_nop(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + UNUSED(backend); + UNUSED(src0); + UNUSED(src1); + UNUSED(dst); +} + +static void ggml_cl_get_rows(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + const int ne00 = src0 ? src0->ne[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + const cl_ulong nb02 = src0 ? src0->nb[2] : 0; + const int ne10 = src1 ? src1->ne[0] : 0; + const cl_ulong nb10 = src1 ? src1->nb[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; + const cl_ulong nb11 = src1 ? src1->nb[1] : 0; + const cl_ulong nb1 = dst ? dst->nb[1] : 0; + const cl_ulong nb2 = dst ? dst->nb[2] : 0; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel; + + switch (src0->type) { + case GGML_TYPE_F32: + kernel = backend_ctx->kernel_get_rows_f32; + break; + case GGML_TYPE_F16: + kernel = backend_ctx->kernel_get_rows_f16; + break; + case GGML_TYPE_Q4_0: + kernel = backend_ctx->kernel_get_rows_q4_0; + break; + default: + GGML_ASSERT(false && "not implemented"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(cl_ulong), &nb2)); + + size_t global_work_size[] = {(size_t)ne10, (size_t)ne11, 1}; + size_t local_work_size[] = {1, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_add(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + const cl_ulong nb00 = src0 ? src0->nb[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + const cl_ulong nb02 = src0 ? src0->nb[2] : 0; + const cl_ulong nb03 = src0 ? src0->nb[3] : 0; + + const int ne10 = src1 ? src1->ne[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; + const int ne12 = src1 ? src1->ne[2] : 0; + const int ne13 = src1 ? src1->ne[3] : 0; UNUSED(ne13); + + const cl_ulong nb10 = src1 ? src1->nb[0] : 0; + const cl_ulong nb11 = src1 ? src1->nb[1] : 0; + const cl_ulong nb12 = src1 ? src1->nb[2] : 0; + const cl_ulong nb13 = src1 ? src1->nb[3] : 0; UNUSED(nb13); + + const int ne0 = dst ? dst->ne[0] : 0; + const int ne1 = dst ? dst->ne[1] : 0; + const int ne2 = dst ? dst->ne[2] : 0; + const int ne3 = dst ? dst->ne[3] : 0; + + const cl_ulong nb0 = dst ? dst->nb[0] : 0; + const cl_ulong nb1 = dst ? dst->nb[1] : 0; + const cl_ulong nb2 = dst ? dst->nb[2] : 0; + const cl_ulong nb3 = dst ? dst->nb[3] : 0; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + bool bcast_row = false; + cl_kernel kernel; + + if (ggml_nelements(src1) == ne10 && ggml_is_contiguous(src1) && ne00 % 4 == 0 && ne10 % 4 == 0) { + GGML_ASSERT(ggml_is_contiguous(src0)); + + // src1 is a row + GGML_ASSERT(ne11 == 1); + + bcast_row = true; + int ne = ne00 / 4; + kernel = backend_ctx->kernel_add_row; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne)); + } else { + kernel = backend_ctx->kernel_add; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne03)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &ne11)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(int), &ne13)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 20, sizeof(cl_ulong), &nb12)); + CL_CHECK(clSetKernelArg(kernel, 21, sizeof(cl_ulong), &nb13)); + CL_CHECK(clSetKernelArg(kernel, 22, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 23, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 24, sizeof(int), &ne2)); + CL_CHECK(clSetKernelArg(kernel, 25, sizeof(int), &ne3)); + CL_CHECK(clSetKernelArg(kernel, 26, sizeof(cl_ulong), &nb0)); + CL_CHECK(clSetKernelArg(kernel, 27, sizeof(cl_ulong), &nb1)); + CL_CHECK(clSetKernelArg(kernel, 28, sizeof(cl_ulong), &nb2)); + CL_CHECK(clSetKernelArg(kernel, 29, sizeof(cl_ulong), &nb3)); + } + + if (bcast_row) { + int n = ggml_nelements(dst)/4; + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } else { + unsigned int nth = MIN(64, ne0); + size_t global_work_size[] = {ne01*nth, (size_t)ne02, (size_t)ne03}; + size_t local_work_size[] = {nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } +} + +static void ggml_cl_mul(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + const cl_ulong nb00 = src0 ? src0->nb[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + const cl_ulong nb02 = src0 ? src0->nb[2] : 0; + const cl_ulong nb03 = src0 ? src0->nb[3] : 0; + + const int ne10 = src1 ? src1->ne[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; + const int ne12 = src1 ? src1->ne[2] : 0; + const int ne13 = src1 ? src1->ne[3] : 0; UNUSED(ne13); + + const cl_ulong nb10 = src1 ? src1->nb[0] : 0; + const cl_ulong nb11 = src1 ? src1->nb[1] : 0; + const cl_ulong nb12 = src1 ? src1->nb[2] : 0; + const cl_ulong nb13 = src1 ? src1->nb[3] : 0; UNUSED(nb13); + + const int ne0 = dst ? dst->ne[0] : 0; + const int ne1 = dst ? dst->ne[1] : 0; + const int ne2 = dst ? dst->ne[2] : 0; + const int ne3 = dst ? dst->ne[3] : 0; + + const cl_ulong nb0 = dst ? dst->nb[0] : 0; + const cl_ulong nb1 = dst ? dst->nb[1] : 0; + const cl_ulong nb2 = dst ? dst->nb[2] : 0; + const cl_ulong nb3 = dst ? dst->nb[3] : 0; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + bool bcast_row = false; + cl_kernel kernel; + + if (ggml_nelements(src1) == ne10 && ggml_is_contiguous(src1) && ne00 % 4 == 0 && ne10 % 4 == 0) { + GGML_ASSERT(ggml_is_contiguous(src0)); + + // src1 is a row + GGML_ASSERT(ne11 == 1); + + bcast_row = true; + int ne = ne00 / 4; + kernel = backend_ctx->kernel_mul_row; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne)); + } else { + kernel = backend_ctx->kernel_mul; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne03)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &ne11)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(int), &ne13)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 20, sizeof(cl_ulong), &nb12)); + CL_CHECK(clSetKernelArg(kernel, 21, sizeof(cl_ulong), &nb13)); + CL_CHECK(clSetKernelArg(kernel, 22, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 23, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 24, sizeof(int), &ne2)); + CL_CHECK(clSetKernelArg(kernel, 25, sizeof(int), &ne3)); + CL_CHECK(clSetKernelArg(kernel, 26, sizeof(cl_ulong), &nb0)); + CL_CHECK(clSetKernelArg(kernel, 27, sizeof(cl_ulong), &nb1)); + CL_CHECK(clSetKernelArg(kernel, 28, sizeof(cl_ulong), &nb2)); + CL_CHECK(clSetKernelArg(kernel, 29, sizeof(cl_ulong), &nb3)); + } + + if (bcast_row) { + int n = ggml_nelements(dst)/4; + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } else { + unsigned int nth = MIN(64, ne0); + size_t global_work_size[] = {ne01*nth, (size_t)ne02, (size_t)ne03}; + size_t local_work_size[] = {nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } +} + +static void ggml_cl_gelu(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel; + + int n = ggml_nelements(dst); + + if (n % 4 == 0) { + kernel = backend_ctx->kernel_gelu_4; + n /= 4; + } else { + kernel = backend_ctx->kernel_gelu; + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL); +#endif +} + +static void ggml_cl_silu(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel; + + int n = ggml_nelements(dst); + + if (n % 4 == 0) { + kernel = backend_ctx->kernel_silu_4; + n /= 4; + } else { + kernel = backend_ctx->kernel_silu; + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_relu(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel = backend_ctx->kernel_relu; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + + const int64_t n = ggml_nelements(dst); + + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_clamp(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + float min; + float max; + memcpy(&min, ((int32_t *) dst->op_params) + 0, sizeof(float)); + memcpy(&max, ((int32_t *) dst->op_params) + 1, sizeof(float)); + + cl_kernel kernel = backend_ctx->kernel_clamp; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(float), &min)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(float), &max)); + + const int64_t n = ggml_nelements(dst); + + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_norm(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + float eps; + memcpy(&eps, dst->op_params, sizeof(float)); + + const int ne00 = src0 ? src0->ne[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + + GGML_ASSERT(ggml_is_contiguous_1(src0)); + + const int nth = MIN(64, ne00); + + cl_kernel kernel = backend_ctx->kernel_norm; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(float), &eps)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(float)*nth, NULL)); + + const int64_t nrows = ggml_nrows(src0); + + size_t global_work_size[] = {(size_t)nrows*nth, 1, 1}; + size_t local_work_size[] = {(size_t)nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_rms_norm(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_backend_opencl_device_context * dev_ctx = + (ggml_backend_opencl_device_context *)backend->device->context; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + float eps; + memcpy(&eps, dst->op_params, sizeof(float)); + + const int ne00 = src0 ? src0->ne[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + + GGML_ASSERT(ne00 % 4 == 0); + GGML_ASSERT(ggml_is_contiguous_1(src0)); + + const int nth = MIN(64, ne00); + + const int64_t nrows = ggml_nrows(src0); + + size_t global_work_size[] = {(size_t)nrows*nth, 1, 1}; + size_t local_work_size[] = {(size_t)nth, 1, 1}; + + cl_kernel kernel = backend_ctx->kernel_rms_norm; + + // Note, this kernel declares local memory in kernel args and the size + // depends on subgroup size. + // Retrieve subgroup size. + // Note, this requires OpenCL 2.1 and above + size_t sgs; + CL_CHECK(clGetKernelSubGroupInfo(kernel, dev_ctx->device, + CL_KERNEL_MAX_SUB_GROUP_SIZE_FOR_NDRANGE, + sizeof(local_work_size), local_work_size, + sizeof(size_t), &sgs, NULL)); + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(float), &eps)); + // This is local memory - the size depends on subgroup size. + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(float)*nth/sgs, NULL)); + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_mul_mat(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + const enum ggml_type src0t = src0 ? src0->type : GGML_TYPE_COUNT; + const enum ggml_type src1t = src1 ? src1->type : GGML_TYPE_COUNT; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + +#ifdef GGML_OPENCL_SOA_Q + ggml_tensor_extra_cl_q4_0 * extra0_q4_0 = (ggml_tensor_extra_cl_q4_0 *)src0->extra; +#endif + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + const cl_ulong nb00 = src0 ? src0->nb[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + const cl_ulong nb02 = src0 ? src0->nb[2] : 0; + const cl_ulong nb03 = src0 ? src0->nb[3] : 0; + + const int ne10 = src1 ? src1->ne[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; + const int ne12 = src1 ? src1->ne[2] : 0; + const int ne13 = src1 ? src1->ne[3] : 0; + + const cl_ulong nb10 = src1 ? src1->nb[0] : 0; + const cl_ulong nb11 = src1 ? src1->nb[1] : 0; + const cl_ulong nb12 = src1 ? src1->nb[2] : 0; + const cl_ulong nb13 = src1 ? src1->nb[3] : 0; + + const int ne0 = dst ? dst->ne[0] : 0; + const int ne1 = dst ? dst->ne[1] : 0; + + int r2 = ne12/ne02; + int r3 = ne13/ne03; + + GGML_ASSERT(ne00 == ne10); + + int nth0 = 32; + int nth1 = 1; + int nrows = 1; + // The number of values produced by each subgroup + int ndst = 4; + + cl_kernel kernel; + +#ifdef GGML_OPENCL_USE_ADRENO_KERNELS + cl_context context = backend_ctx->context; + + if (ne01 && ne1 && use_adreno_kernels(src0)) { + + // init CL objects + // <--------------------------------------------> // + cl_int status; + cl_image_format img_fmt_1d; + cl_image_desc img_desc_1d; + cl_buffer_region region; + cl_mem A_image1d = nullptr; + cl_mem B_image1d = nullptr; + cl_mem B_sub_buffer = nullptr; + cl_mem C_d = nullptr; + // for B transpose + cl_mem B_d = nullptr; + cl_mem B_d_input_image = nullptr; + // <--------------------------------------------> // + + // define matrix dimensions + // <--------------------------------------------> // + int M = ne01; + int N = ne1; + int K = ne00; + int padding; + // <--------------------------------------------> // + + // q4_0 x fp32 + if(src0t == GGML_TYPE_Q4_0 && src1t == GGML_TYPE_F32) { + // TODO: remove duplicate definitions of image description + format -- move to top + + // create an image for A + // <--------------------------------------------> // + if (N == 1) { + img_fmt_1d = { CL_R, CL_UNSIGNED_INT32}; + } else { + img_fmt_1d = { CL_R, CL_FLOAT}; + } + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.image_width = M * K / 2 / 4; // Divide by 4 for char -> float + img_desc_1d.buffer = extra0_q4_0->q; + A_image1d = clCreateImage( + context, + CL_MEM_READ_ONLY, + &img_fmt_1d, + &img_desc_1d, + NULL, + &status); + CL_CHECK(status); + // <--------------------------------------------> // + + + // create a sub_buffer for B + // <--------------------------------------------> // + region.origin = (extra1->offset); + region.size = K * N * sizeof(float); + B_sub_buffer = clCreateSubBuffer( + extra1->data_device, + 0, + CL_BUFFER_CREATE_TYPE_REGION, + ®ion, + &status); + CL_CHECK(status); + // <--------------------------------------------> // + + // transpose activation for Skyler's gemm + if (N != 1) { + //how many extra elements beyond multiple of 8 + int extra_elements = N % 8; + + //how much padding to add + padding = 0; + if (extra_elements > 0){ + padding = 8 - extra_elements; + } + + // Specify the starting offset (in bytes) + region.origin = 0; + // Specify the size of the sub-buffer (divide by 2 for FP16) + region.size = K * (N + padding) * sizeof(float)/2; + B_d = clCreateSubBuffer( + backend_ctx->B_d_max, + 0, + CL_BUFFER_CREATE_TYPE_REGION, + ®ion, + &status); + CL_CHECK(status); + + cl_image_format image_format_B_d_input = { CL_RGBA, CL_FLOAT }; + cl_image_desc image_desc_B_d_input = { + CL_MEM_OBJECT_IMAGE1D_BUFFER, + static_cast(K * N / 4), + 0, 0, 0, 0, 0, 0, 0, { B_sub_buffer } + }; + B_d_input_image = clCreateImage( + context, + 0, + &image_format_B_d_input, + &image_desc_B_d_input, + NULL, + &status); + CL_CHECK(status); + + cl_image_format image_format_B_d_output = { CL_RGBA, CL_HALF_FLOAT }; //(CL_HALF_FLOAT for FP16) + cl_image_desc image_desc_B_d_output = { + CL_MEM_OBJECT_IMAGE1D_BUFFER, + static_cast(K * (N + padding)/4), + 0, 0, 0, 0, 0, 0, 0, { B_d } + }; + B_image1d = clCreateImage( + context, + 0, + &image_format_B_d_output, + &image_desc_B_d_output, + NULL, + &status); + CL_CHECK(status); + + int height_B = N/4; + int width_B = K/4; + int padded_height_B = (N + padding)/4; + + kernel = backend_ctx->kernel_transpose_32_16; + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &B_d_input_image)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &B_image1d)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(int), &height_B)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(int), &width_B)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &padded_height_B)); + + size_t local_size_t[2] = { 1, 16 }; + //WGS tuning + if (ne0 == 4096 && ne1 == 128 && ne10 == 4096) { + local_size_t[0]=4; + local_size_t[1]=8; + } else if (ne0 == 11008 && ne1 == 128 && ne10 == 4096) { + local_size_t[0]=2; + local_size_t[1]=8; + } else if(ne0 == 4096 && ne1 == 128 && ne10 == 11008) { + local_size_t[0]=1; + local_size_t[1]=8; + } else if(ne0 == 32000 && ne1 == 128 && ne10 == 4096) { + local_size_t[0]=2; + local_size_t[1]=8; + } + + size_t global_size_t[2] = { + static_cast(width_B), + static_cast(padded_height_B) + }; + + #ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 2, NULL, global_size_t, local_size_t, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_size_t, local_size_t, dst); + #else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 2, NULL, global_size_t, local_size_t, 0, NULL, NULL)); + #endif + } else { + // no need to transpose B in other cases + // create an image for B from sub_buffer + // <--------------------------------------------> // + img_fmt_1d = {CL_RGBA, CL_FLOAT}; + + memset(&img_desc_1d, 0, sizeof(img_desc_1d)); + img_desc_1d.image_width = K * N / 4; + img_desc_1d.image_type = CL_MEM_OBJECT_IMAGE1D_BUFFER; + img_desc_1d.buffer = B_sub_buffer; + B_image1d = clCreateImage( + context, + CL_MEM_READ_ONLY, + &img_fmt_1d, + &img_desc_1d, + NULL, + &status); + CL_CHECK(status); + // <--------------------------------------------> // + } + + // choose gemm or gemv kernel + // <--------------------------------------------> // + if (N == 1) { + kernel = backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_general; + if (M == 4096 && K == 4096) { + kernel = backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_4096; + } else if (M == 4096 && K == 11008) { + kernel = backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_4096_1_11008; + } else if (M == 11008 && K == 4096) { + kernel = backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_11008_1_4096; + } else if (M == 32000 && K == 4096) { + kernel = backend_ctx->CL_mul_mat_vec_q4_0_f32_1d_4x_flat_32000_1_4096; + } + } else { + kernel = backend_ctx->CL_mul_mat_Ab_Bi_8x4; + } + // <--------------------------------------------> // + + // set kernel args + // <--------------------------------------------> // + cl_uint k_arg = 0; + + if (N == 1) { + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_mem), &A_image1d)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_mem), &extra0_q4_0->d)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_mem), &B_image1d)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_ulong), &extra1->offset)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(cl_ulong), &extrad->offset)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, k_arg++, sizeof(int), &r3)); + } else { + region.origin = extrad->offset; // Specify the starting offset (in bytes) + region.size = M * N * sizeof(float); // Specify the size of the sub-buffer + C_d = clCreateSubBuffer(extrad->data_device, CL_MEM_WRITE_ONLY, CL_BUFFER_CREATE_TYPE_REGION, ®ion, &status); + CL_CHECK(status); + + int padded_N = ne1 + padding; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0_q4_0->q)); //A_q_dextra0_q4_0->q + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra0_q4_0->d)); //A_s_d + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &B_image1d)); //B_d + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_mem), &C_d)); //C_d + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne01)); //M + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(int), &padded_N)); //N with padding + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); //K + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne1)); //N without padding + } + // <--------------------------------------------> // + + // choose workgroup size + // <--------------------------------------------> // + size_t global_work_size[3] = { + 64, static_cast((M+63)/64), static_cast((N+31)/32)}; + size_t local_work_size[3] = {64, 2, 4}; + + global_work_size[0] = (size_t)(ceil((float)ne1/8)); + global_work_size[1] = (size_t)(ne01/4); + global_work_size[2] = (size_t)(1); + + local_work_size[0] = (size_t)(1); //4x32 for FP32 + local_work_size[1] = (size_t)(128); + local_work_size[2] = (size_t)(1); + + //WGS tuning + if (ne0 == 4096 && ne1 == 128 && ne10 == 4096) { + local_work_size[0] = 1; + local_work_size[1] = 128; + } else if (ne0 == 11008 && ne1 == 128 && ne10 == 4096) { + local_work_size[0] = 2; + local_work_size[1] = 64; + } else if (ne0 == 4096 && ne1 == 128 && ne10 == 11008) { + local_work_size[0] = 2; + local_work_size[1] = 64; + } else if (ne0 == 32000 && ne1 == 128 && ne10 == 4096) { + local_work_size[0] = 2; + local_work_size[1] = 64; + } + + if (N == 1) { + local_work_size[0] = backend_ctx->adreno_wave_size; // localsize + local_work_size[1] = 4; // reduce factor + local_work_size[2] = 1; + + global_work_size[0] = M / 2; + global_work_size[1] = 4; // reduce factor + global_work_size[2] = 1; + } + // <--------------------------------------------> // + + // enqueue kernel with profiling + // <--------------------------------------------> // + #ifdef GGML_OPENCL_PROFILING + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); + // enqueue kernel without profiling + #else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); + #endif + // <--------------------------------------------> // + + // deallocate sub buffers and images + // <--------------------------------------------> // + CL_CHECK(clReleaseMemObject(A_image1d)); + CL_CHECK(clReleaseMemObject(B_sub_buffer)); + CL_CHECK(clReleaseMemObject(B_image1d)); + + if (N != 1) { + CL_CHECK(clReleaseMemObject(B_d)); + CL_CHECK(clReleaseMemObject(B_d_input_image)); + CL_CHECK(clReleaseMemObject(C_d)); + } + // <--------------------------------------------> // + + return; + } + } // if (ne01 && ne1) +#endif // GGML_OPENCL_USE_ADRENO_KERNELS + + if (!ggml_is_transposed(src0) && + !ggml_is_transposed(src1) && + src1t == GGML_TYPE_F32 && + ne00%32 == 0 && + ne11 > 2) { +#ifdef GGML_OPENCL_SOA_Q + // Set up kernel. + switch(src0t) { + case GGML_TYPE_Q4_0: + // This should have been satisfied. + GGML_ASSERT(ne11 == ne1); + GGML_ASSERT(ne01 == ne0); + + if (backend_ctx->gpu_family == INTEL) { + nth0 = 16; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32_1d_16x_flat; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 64; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32_1d_8x_flat; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0_q4_0->q)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra0_q4_0->d)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &r3)); + break; + default: + break; + } + + // Launch kernel. + if (src0t == GGML_TYPE_Q4_0) { + size_t global_work_size[] = {(size_t)(ne01 + 7)/8*nth0, (size_t)ne11*nth1, (size_t)ne12*ne13}; + size_t local_work_size[] = {(size_t)nth0, (size_t)nth1, 1}; + + if (backend_ctx->gpu_family == INTEL) { + // Set global size for Intel. It uses 16x output values. + global_work_size[0] = (size_t)(ne01 + 15)/16*nth0; + global_work_size[1] = (size_t)ne11*nth1; + global_work_size[2] = (size_t)ne12*ne13; + } + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + return; + } +#else // GGML_OPENCL_SOA_Q + // TODO: add block_q4_0 variant. +#endif // GGML_OPENCL_SOA_Q + } + + // use custom matrix x vector kernel + switch (src0t) { + case GGML_TYPE_F32: + //GGML_ASSERT(ne02 == ne12); + GGML_ASSERT(src1t == GGML_TYPE_F32); + kernel = backend_ctx->kernel_mul_mat_f32_f32; + nrows = 4; + + if (backend_ctx->gpu_family == INTEL) { + nth0 = 32; + nth1 = 1; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 64; + nth1 = 1; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &ne11)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(cl_ulong), &nb12)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(cl_ulong), &nb13)); + CL_CHECK(clSetKernelArg(kernel, 20, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 21, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 22, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 23, sizeof(int), &r3)); + break; + case GGML_TYPE_F16: + //GGML_ASSERT(ne02 == ne12); + if (backend_ctx->gpu_family == INTEL) { + nth0 = 32; + nth1 = 1; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 64; + nth1 = 1; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + if (src1t == GGML_TYPE_F32) { + if (ne11 * ne12 < 4) { + kernel = backend_ctx->kernel_mul_mat_f16_f32_1row; + } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { + kernel = backend_ctx->kernel_mul_mat_f16_f32_l4; + nrows = ne11; + } else { + kernel = backend_ctx->kernel_mul_mat_f16_f32; + nrows = 4; + } + } else { + kernel = backend_ctx->kernel_mul_mat_f16_f16; + nrows = 4; + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &ne11)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(cl_ulong), &nb12)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(cl_ulong), &nb13)); + CL_CHECK(clSetKernelArg(kernel, 20, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 21, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 22, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 23, sizeof(int), &r3)); + break; + case GGML_TYPE_Q4_0: + // This should have been satisfied. + GGML_ASSERT(ne11 == ne1); + GGML_ASSERT(ne01 == ne0); + +#ifdef GGML_OPENCL_SOA_Q + if (backend_ctx->gpu_family == INTEL) { + nth0 = 16; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32_8x_flat; + ndst = 8; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 64; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32_8x_flat; + ndst =8; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0_q4_0->q)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), &extra0_q4_0->d)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &r3)); +#else // GGML_OPENCL_SOA_Q + if (backend_ctx->gpu_family == INTEL) { + // Use 1D local size. Each workgroup is a SIMD group. Each SIMD + // group produces N_DST (4 for Q4_0 kernel) values in the result. + // The number of workgroups on dim 0 (the leading dimension) is + // the nearest multiple of 4 that covers ne0 (equals ne01). + nth0 = 16; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32; + ndst = 4; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 64; + nth1 = 1; + + kernel = backend_ctx->kernel_mul_mat_q4_0_f32_v; + ndst = 4; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &r3)); +#endif // GGML_OPENCL_SOA_Q + break; + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q8_0: + case GGML_TYPE_Q2_K: + case GGML_TYPE_Q3_K: + case GGML_TYPE_Q4_K: + case GGML_TYPE_Q5_K: + case GGML_TYPE_Q6_K: + kernel = backend_ctx->kernel_mul_mv_q6_K_f32; + + if (backend_ctx->gpu_family == INTEL) { + nth0 = 2; + nth1 = 16; + } else if (backend_ctx->gpu_family == ADRENO) { + nth0 = 2; + nth1 = 64; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &r2)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &r3)); + break; + default: + GGML_ASSERT(false && "not implemented"); + } + + if (src0t == GGML_TYPE_Q4_0 || + src0t == GGML_TYPE_Q4_1 || + src0t == GGML_TYPE_Q8_0 || + src0t == GGML_TYPE_Q2_K) { + // Each SIMD group produces N_DST values in the result. Assuming each + // workgroup has N_SIMDGROUP SIMD groups, then each workgroup will + // produce N_DST*N_SIMDGROUP values in the result. Hence, the grid size + // (number of workgroups) will be a nearest multiple of + // N_DST*N_SIMDGROUP to cover the size of the dimension. Below, 4 is + // N_DST*N_SIMDGROUP (see the kernel for Q4_0 matmul). + size_t global_work_size[] = {(size_t)(ne01 + ndst-1)/ndst*nth0, (size_t)ne11*nth1, (size_t)ne12*ne13}; + size_t local_work_size[] = {(size_t)nth0, (size_t)nth1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } else if (src0t == GGML_TYPE_Q4_K) { + GGML_ASSERT(false && "not implemented"); + } else if (src0t == GGML_TYPE_Q3_K) { + GGML_ASSERT(false && "not implemented"); + } else if (src0t == GGML_TYPE_Q5_K) { + GGML_ASSERT(false && "not implemented"); + } else if (src0t == GGML_TYPE_Q6_K) { + size_t global_work_size[] = {(size_t)(ne01+1)/2*nth0, (size_t)ne11*nth1, (size_t)ne12*ne13}; + size_t local_work_size[] = {(size_t)nth0, (size_t)nth1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } else { + int64_t ny = (ne11 + nrows - 1)/nrows; + + size_t global_work_size[] = {(size_t)ne01*nth0, (size_t)ny*nth1, (size_t)ne12*ne13}; + size_t local_work_size[] = {(size_t)nth0, (size_t)nth1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } +} + +static void ggml_cl_scale(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + GGML_UNUSED(src1); + + GGML_ASSERT(ggml_is_contiguous(src0)); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + float scale; + memcpy(&scale, dst->op_params, sizeof(scale)); + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel = backend_ctx->kernel_scale; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(float), &scale)); + + int n = ggml_nelements(dst)/4; + + size_t global_work_size[] = {(size_t)n, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_cpy(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + + // GGML_OP_CPY happens between src0 and src1. + // GGML_OP_DUP and GGML_OP_CONT happen between src0 and dst. + UNUSED(dst); + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + const cl_ulong nb00 = src0 ? src0->nb[0] : 0; + const cl_ulong nb01 = src0 ? src0->nb[1] : 0; + const cl_ulong nb02 = src0 ? src0->nb[2] : 0; + const cl_ulong nb03 = src0 ? src0->nb[3] : 0; + + const int ne10 = src1 ? src1->ne[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; + const int ne12 = src1 ? src1->ne[2] : 0; + const int ne13 = src1 ? src1->ne[3] : 0; + + const cl_ulong nb10 = src1 ? src1->nb[0] : 0; + const cl_ulong nb11 = src1 ? src1->nb[1] : 0; + const cl_ulong nb12 = src1 ? src1->nb[2] : 0; + const cl_ulong nb13 = src1 ? src1->nb[3] : 0; + + const enum ggml_type src0t = src0 ? src0->type : GGML_TYPE_COUNT; + const enum ggml_type src1t = src1 ? src1->type : GGML_TYPE_COUNT; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + + cl_kernel kernel; + + switch (src0t) { + case GGML_TYPE_F32: + switch (src1t) { + case GGML_TYPE_F16: + kernel = backend_ctx->kernel_cpy_f32_f16; + break; + case GGML_TYPE_F32: + kernel = backend_ctx->kernel_cpy_f32_f32; + break; + default: + GGML_ASSERT(false && "not implemented"); + } + break; + case GGML_TYPE_F16: + switch (src1t) { + case GGML_TYPE_F16: + kernel = backend_ctx->kernel_cpy_f16_f16; + break; + case GGML_TYPE_F32: + kernel = backend_ctx->kernel_cpy_f16_f32; + break; + default: + GGML_ASSERT(false && "not implemented"); + } + break; + default: + GGML_ASSERT(false && "not implemented"); + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne03)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(int), &ne10)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &ne11)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(int), &ne12)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(int), &ne13)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(cl_ulong), &nb10)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(cl_ulong), &nb11)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(cl_ulong), &nb12)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(cl_ulong), &nb13)); + + const int nth = MIN(64, ne00); + + size_t global_work_size[] = {(size_t)ne01*nth, (size_t)ne02, (size_t)ne03}; + size_t local_work_size[] = {(size_t)nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, src1); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_dup(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + ggml_cl_cpy(backend, src0, dst, nullptr); + UNUSED(src1); +} + +static void ggml_cl_diag_mask_inf(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + UNUSED(src1); + + int n_past = ((int32_t *)(dst->op_params))[0]; + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_kernel kernel; + + if (ne00%8 == 0) { + kernel = backend_ctx->kernel_diag_mask_inf_8; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &n_past)); + + size_t global_work_size[] = {(size_t)ne00*ne01*ne02/8, 1, 1}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } else { + kernel = backend_ctx->kernel_diag_mask_inf; + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &n_past)); + + size_t global_work_size[] = {(size_t)ne00, (size_t)ne01, (size_t)ne02}; + size_t local_work_size[] = {64, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif + } +} + +static void ggml_cl_soft_max(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + // Softmax can now fuse KQ mask and KQ scale, which used to be two additional + // ops before softmax. It now also fuses alibi if `max_bias > 0`. For llama, + // alibi is not used; however, for some other models, it is used. + // KQ_mask + if (src1) { + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + } + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + ggml_tensor_extra_cl * extra1 = src1 ? (ggml_tensor_extra_cl *)src1->extra : nullptr; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + cl_ulong offset1 = extra1 ? extra1->offset + src1->view_offs : offset0; + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + float scale, max_bias; + memcpy(&scale, dst->op_params + 0, sizeof(float)); + memcpy(&max_bias, dst->op_params + 1, sizeof(float)); + + const int nrows_x = ggml_nrows(src0); + const int nrows_y = src0->ne[1]; + + const int n_head = nrows_x/nrows_y; + const int n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + // Local size must be wave size. Each workgroup is a wave, working on a row, + // where a row corresponds to leading dimension. + int nth = MIN(32, ne00); + + if (backend_ctx->gpu_family == INTEL) { + // This is the same as the initial value. + nth = MIN(32, ne00); + } + else if (backend_ctx->gpu_family == ADRENO) { + nth = 64; + } else { + GGML_ASSERT(false && "TODO: Unknown GPU"); + } + + cl_kernel kernel; + + if (ne00%4 == 0) { + kernel = backend_ctx->kernel_soft_max_4; + } else { + kernel = backend_ctx->kernel_soft_max; + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), extra1 ? &extra1->data_device : &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(float), &scale)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(float), &max_bias)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(float), &m0)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(float), &m1)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(int), &n_head_log2)); + + size_t global_work_size[] = {(size_t)ne01*nth, (size_t)ne02, (size_t)ne03}; + size_t local_work_size[] = {(size_t)nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +static void ggml_cl_rope(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { + GGML_ASSERT(src0); + GGML_ASSERT(src0->extra); + GGML_ASSERT(src1); + GGML_ASSERT(src1->extra); + GGML_ASSERT(dst); + GGML_ASSERT(dst->extra); + + ggml_backend_opencl_context *backend_ctx = (ggml_backend_opencl_context *)backend->context; + cl_command_queue queue = backend_ctx->queue; + + ggml_tensor_extra_cl * extra0 = (ggml_tensor_extra_cl *)src0->extra; + ggml_tensor_extra_cl * extra1 = (ggml_tensor_extra_cl *)src1->extra; + ggml_tensor_extra_cl * extrad = (ggml_tensor_extra_cl *)dst->extra; + + cl_ulong offset0 = extra0->offset + src0->view_offs; + cl_ulong offset1 = extra1->offset + src1->view_offs; + cl_ulong offsetd = extrad->offset + dst->view_offs; + + ggml_tensor * src2 = dst->src[2]; + ggml_tensor_extra_cl * extra2 = src2 ? (ggml_tensor_extra_cl *)src2->extra : nullptr; + + cl_ulong offset2 = extra2 ? extra2->offset + src2->view_offs : offset0; + + const int ne00 = src0 ? src0->ne[0] : 0; + const int ne01 = src0 ? src0->ne[1] : 0; + const int ne02 = src0 ? src0->ne[2] : 0; + const int ne03 = src0 ? src0->ne[3] : 0; + + const int nb00 = src0 ? src0->nb[0] : 0; + const int nb01 = src0 ? src0->nb[1] : 0; + const int nb02 = src0 ? src0->nb[2] : 0; + const int nb03 = src0 ? src0->nb[3] : 0; + + const int ne10 = src1 ? src1->ne[0] : 0; + const int ne11 = src1 ? src1->ne[1] : 0; UNUSED(ne11); + const int ne12 = src1 ? src1->ne[2] : 0; UNUSED(ne12); + const int ne13 = src1 ? src1->ne[3] : 0; UNUSED(ne13); + + const int ne0 = dst ? dst->ne[0] : 0; + const int ne1 = dst ? dst->ne[1] : 0; + const int ne2 = dst ? dst->ne[2] : 0; + const int ne3 = dst ? dst->ne[3] : 0; + + const int nb0 = dst ? dst->nb[0] : 0; + const int nb1 = dst ? dst->nb[1] : 0; + const int nb2 = dst ? dst->nb[2] : 0; + const int nb3 = dst ? dst->nb[3] : 0; + + GGML_ASSERT(ne10 == ne02); + + int nth = MIN(64, ne00); + + const int n_past = ((int *) dst->op_params)[0]; + const int n_dims = ((int *) dst->op_params)[1]; + const int mode = ((int *) dst->op_params)[2]; + const int n_ctx_orig = ((int32_t *) dst->op_params)[4]; + + float freq_base; + float freq_scale; + float ext_factor; + float attn_factor; + float beta_fast; + float beta_slow; + + memcpy(&freq_base, (int32_t *) dst->op_params + 5, sizeof(float)); + memcpy(&freq_scale, (int32_t *) dst->op_params + 6, sizeof(float)); + memcpy(&ext_factor, (int32_t *) dst->op_params + 7, sizeof(float)); + memcpy(&attn_factor, (int32_t *) dst->op_params + 8, sizeof(float)); + memcpy(&beta_fast, (int32_t *) dst->op_params + 9, sizeof(float)); + memcpy(&beta_slow, (int32_t *) dst->op_params + 10, sizeof(float)); + + const bool is_neox = mode & 2; + + cl_kernel kernel; + + if (!is_neox) { + switch (src0->type) { + case GGML_TYPE_F32: + kernel = backend_ctx->kernel_rope_norm_f32; + break; + case GGML_TYPE_F16: + kernel = backend_ctx->kernel_rope_norm_f16; + break; + default: + GGML_ASSERT(false); + }; + } else { + switch (src0->type) { + case GGML_TYPE_F32: + kernel = backend_ctx->kernel_rope_neox_f32; + break; + case GGML_TYPE_F16: + kernel = backend_ctx->kernel_rope_neox_f16; + break; + default: + GGML_ASSERT(false); + }; + } + + CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_ulong), &offset0)); + CL_CHECK(clSetKernelArg(kernel, 2, sizeof(cl_mem), &extra1->data_device)); + CL_CHECK(clSetKernelArg(kernel, 3, sizeof(cl_ulong), &offset1)); + CL_CHECK(clSetKernelArg(kernel, 4, sizeof(cl_mem), extra2 ? &extra2->data_device : &extra0->data_device)); + CL_CHECK(clSetKernelArg(kernel, 5, sizeof(cl_ulong), &offset2)); + CL_CHECK(clSetKernelArg(kernel, 6, sizeof(cl_mem), &extrad->data_device)); + CL_CHECK(clSetKernelArg(kernel, 7, sizeof(cl_ulong), &offsetd)); + CL_CHECK(clSetKernelArg(kernel, 8, sizeof(int), &ne00)); + CL_CHECK(clSetKernelArg(kernel, 9, sizeof(int), &ne01)); + CL_CHECK(clSetKernelArg(kernel, 10, sizeof(int), &ne02)); + CL_CHECK(clSetKernelArg(kernel, 11, sizeof(int), &ne03)); + CL_CHECK(clSetKernelArg(kernel, 12, sizeof(cl_ulong), &nb00)); + CL_CHECK(clSetKernelArg(kernel, 13, sizeof(cl_ulong), &nb01)); + CL_CHECK(clSetKernelArg(kernel, 14, sizeof(cl_ulong), &nb02)); + CL_CHECK(clSetKernelArg(kernel, 15, sizeof(cl_ulong), &nb03)); + CL_CHECK(clSetKernelArg(kernel, 16, sizeof(int), &ne0)); + CL_CHECK(clSetKernelArg(kernel, 17, sizeof(int), &ne1)); + CL_CHECK(clSetKernelArg(kernel, 18, sizeof(int), &ne2)); + CL_CHECK(clSetKernelArg(kernel, 19, sizeof(int), &ne3)); + CL_CHECK(clSetKernelArg(kernel, 20, sizeof(cl_ulong), &nb0)); + CL_CHECK(clSetKernelArg(kernel, 21, sizeof(cl_ulong), &nb1)); + CL_CHECK(clSetKernelArg(kernel, 22, sizeof(cl_ulong), &nb2)); + CL_CHECK(clSetKernelArg(kernel, 23, sizeof(cl_ulong), &nb3)); + CL_CHECK(clSetKernelArg(kernel, 24, sizeof(int), &n_past)); + CL_CHECK(clSetKernelArg(kernel, 25, sizeof(int), &n_dims)); + CL_CHECK(clSetKernelArg(kernel, 26, sizeof(int), &n_ctx_orig)); + CL_CHECK(clSetKernelArg(kernel, 27, sizeof(float), &freq_base)); + CL_CHECK(clSetKernelArg(kernel, 28, sizeof(float), &freq_scale)); + CL_CHECK(clSetKernelArg(kernel, 29, sizeof(float), &ext_factor)); + CL_CHECK(clSetKernelArg(kernel, 30, sizeof(float), &attn_factor)); + CL_CHECK(clSetKernelArg(kernel, 31, sizeof(float), &beta_fast)); + CL_CHECK(clSetKernelArg(kernel, 32, sizeof(float), &beta_slow)); + + size_t global_work_size[] = {(size_t)ne01*nth, (size_t)ne02, (size_t)ne03}; + size_t local_work_size[] = {(size_t)nth, 1, 1}; + +#ifdef GGML_OPENCL_PROFILING + cl_event evt; + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, &evt)); + + g_profiling_info.emplace_back(); + populateProfilingInfo(g_profiling_info.back(), evt, kernel, global_work_size, local_work_size, dst); +#else + CL_CHECK(clEnqueueNDRangeKernel(queue, kernel, 3, NULL, global_work_size, local_work_size, 0, NULL, NULL)); +#endif +} + +//------------------------------------------------------------------------------ +// Op offloading +//------------------------------------------------------------------------------ + +typedef void (*ggml_cl_func_t)(ggml_backend_t backend, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); + +bool ggml_cl_compute_forward(ggml_backend_t backend, struct ggml_tensor * tensor) { + ggml_cl_func_t func = nullptr; + + ggml_tensor * src0 = tensor->src[0]; + ggml_tensor * src1 = tensor->src[1]; + + const bool any_on_device = tensor->extra + || (src0 != nullptr && src0->extra) + || (src1 != nullptr && src1->extra); + + switch (tensor->op) { + case GGML_OP_GET_ROWS: + if (!any_on_device) { + return false; + } + func = ggml_cl_get_rows; + break; + case GGML_OP_CPY: + if (!any_on_device) { + return false; + } + func = ggml_cl_cpy; + break; + case GGML_OP_DUP: + case GGML_OP_CONT: + if (!any_on_device) { + return false; + } + func = ggml_cl_dup; + break; + case GGML_OP_ADD: + if (!any_on_device) { + return false; + } + GGML_ASSERT(ggml_is_contiguous(src0)); + GGML_ASSERT(ggml_is_contiguous(src1)); + func = ggml_cl_add; + break; + case GGML_OP_MUL: + if (!any_on_device) { + return false; + } + func = ggml_cl_mul; + break; + case GGML_OP_UNARY: + switch (ggml_get_unary_op(tensor)) { + case GGML_UNARY_OP_GELU: + if (!any_on_device) { + return false; + } + func = ggml_cl_gelu; + break; + case GGML_UNARY_OP_SILU: + if (!any_on_device) { + return false; + } + func = ggml_cl_silu; + break; + case GGML_UNARY_OP_RELU: + if (!any_on_device) { + return false; + } + func = ggml_cl_relu; + break; + default: + return false; + } break; + case GGML_OP_CLAMP: + if (!any_on_device) { + return false; + } + func = ggml_cl_clamp; + break; + case GGML_OP_NORM: + if (!any_on_device) { + return false; + } + func = ggml_cl_norm; + break; + case GGML_OP_RMS_NORM: + if (!any_on_device) { + return false; + } + func = ggml_cl_rms_norm; + break; + case GGML_OP_MUL_MAT: + if (!any_on_device && !ggml_cl_can_mul_mat(tensor->src[0], tensor->src[1], tensor)) { + return false; + } + func = ggml_cl_mul_mat; + break; + case GGML_OP_SCALE: + if (!any_on_device) { + return false; + } + func = ggml_cl_scale; + break; + case GGML_OP_RESHAPE: + case GGML_OP_VIEW: + case GGML_OP_PERMUTE: + case GGML_OP_TRANSPOSE: + if (!any_on_device) { + return false; + } + func = ggml_cl_nop; + break; + case GGML_OP_DIAG_MASK_INF: + if (!any_on_device) { + return false; + } + func = ggml_cl_diag_mask_inf; + break; + case GGML_OP_SOFT_MAX: + if (!any_on_device) { + return false; + } + func = ggml_cl_soft_max; + break; + case GGML_OP_ROPE: + if (!any_on_device) { + return false; + } + func = ggml_cl_rope; + break; + default: + return false; + } + + func(backend, tensor->src[0], tensor->src[1], tensor); + return true; +} diff --git a/ggml/src/ggml-opencl/kernels/embed_kernel.py b/ggml/src/ggml-opencl/kernels/embed_kernel.py new file mode 100644 index 000000000..b5d1d7242 --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/embed_kernel.py @@ -0,0 +1,26 @@ +# + +import sys +import logging +logger = logging.getLogger("opencl-embed-kernel") + + +def main(): + logging.basicConfig(level=logging.INFO) + + if len(sys.argv) != 3: + logger.info("Usage: python embed_kernel.py ") + sys.exit(1) + + ifile = open(sys.argv[1], "r") + ofile = open(sys.argv[2], "w") + + for i in ifile: + ofile.write('R"({})"\n'.format(i)) + + ifile.close() + ofile.close() + + +if __name__ == "__main__": + main() diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl.cl new file mode 100644 index 000000000..d1cdf709b --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl.cl @@ -0,0 +1,2683 @@ +#ifdef cl_khr_fp16 +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#elif defined(cl_amd_fp16) +#pragma OPENCL EXTENSION cl_amd_fp16 : enable +#else +#error "Half precision floating point not supportedby OpenCL implementation on your device." +#endif + +#ifdef cl_khr_subgroups +#pragma OPENCL EXTENSION cl_khr_subgroups : enable +#elif defined(cl_intel_subgroups) +#pragma OPENCL EXTENSION cl_intel_subgroups : enable +#else +#error "Subgroup not supported on your device." +#endif + +#ifdef cl_intel_required_subgroup_size +// Always use subgroup size of 32 on Intel. +#pragma OPENCL EXTENSION cl_intel_required_subgroup_size : enable +#define INTEL_GPU 1 +#define REQD_SUBGROUP_SIZE_16 __attribute__((intel_reqd_sub_group_size(16))) +#define REQD_SUBGROUP_SIZE_32 __attribute__((intel_reqd_sub_group_size(32))) +#elif defined(cl_qcom_reqd_sub_group_size) +// Always use subgroups size of 64 on Adreno. +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable +#define ADRENO_GPU 1 +#define REQD_SUBGROUP_SIZE_64 __attribute__((qcom_reqd_sub_group_size("half"))) +#define REQD_SUBGROUP_SIZE_128 __attribute__((qcom_reqd_sub_group_size("full"))) +#else +// TODO: do not know how to choose subgroup size on other GPUs. +#error "Selecting subgroup size is not supported on your device." +#endif + +#define QK4_0 32 +#define QR4_0 2 +#define QK4_1 32 +#define QR4_1 2 +#define QK5_0 32 +#define QR5_0 2 +#define QK5_1 32 +#define QR5_1 2 +#define QK8_0 32 +#define QR8_0 1 +#define QK_K 256 +#define K_QUANTS_PER_ITERATION 2 + +typedef char int8_t; +typedef uchar uint8_t; +typedef short int16_t; +typedef ushort uint16_t; +typedef int int32_t; +typedef uint uint32_t; + +//------------------------------------------------------------------------------ +// block_q4_0 +//------------------------------------------------------------------------------ +struct block_q4_0 +{ + half d; + uint8_t qs[QK4_0 / 2]; +}; + +//------------------------------------------------------------------------------ +// block_q4_1 +//------------------------------------------------------------------------------ +struct block_q4_1 +{ + half d; + half m; + uint8_t qs[QK4_1 / 2]; +}; + +//------------------------------------------------------------------------------ +// block_q5_0 +//------------------------------------------------------------------------------ +struct block_q5_0 +{ + half d; + uint32_t qh; + uint8_t qs[QK5_0 / 2]; +}; + +//------------------------------------------------------------------------------ +// block_q5_1 +//------------------------------------------------------------------------------ +struct block_q5_1 +{ + half d; + half m; + uint32_t qh; + uint8_t qs[QK5_1 / 2]; +}; + +//------------------------------------------------------------------------------ +// block_q8_0 +//------------------------------------------------------------------------------ +struct block_q8_0 +{ + half d; + int8_t qs[QK8_0]; +}; + +//------------------------------------------------------------------------------ +// block_q2_K +//------------------------------------------------------------------------------ +struct block_q2_K +{ + uint8_t scales[16]; + uint8_t qs[64]; + half d; + half dmin; +}; + +//------------------------------------------------------------------------------ +// block_q3_K +//------------------------------------------------------------------------------ +struct block_q3_K +{ + uint8_t hmask[32]; + uint8_t qs[64]; + uint8_t scales[12]; + half d; +}; + +//------------------------------------------------------------------------------ +// block_q4_K +//------------------------------------------------------------------------------ +struct block_q4_K +{ + half d; + half dmin; + uint8_t scales[12]; + uint8_t qs[128]; +}; + +//------------------------------------------------------------------------------ +// block_q5_K +//------------------------------------------------------------------------------ +struct block_q5_K +{ + half d; + half dmin; + uint8_t scales[12]; + uint8_t qh[32]; + uint8_t qs[128]; +}; + +//------------------------------------------------------------------------------ +// block_q6_K +//------------------------------------------------------------------------------ +struct block_q6_K +{ + uint8_t ql[128]; + uint8_t qh[64]; + int8_t scales[16]; + half d; +}; + +//------------------------------------------------------------------------------ +// dequantize_q4_0_f32, dequantize_q4_0_f16 +//------------------------------------------------------------------------------ +void dequantize_q4_0_f32(global struct block_q4_0 * xb, short il, float16 * reg) { + global ushort * qs = ((global ushort *)xb + 1); + float d1 = il ? (xb->d / 16.h) : xb->d; + float d2 = d1 / 256.f; + float md = -8.h * xb->d; + ushort mask0 = il ? 0x00F0 : 0x000F; + ushort mask1 = mask0 << 8; + + reg->s0 = d1 * (qs[0] & mask0) + md; + reg->s1 = d2 * (qs[0] & mask1) + md; + + reg->s2 = d1 * (qs[1] & mask0) + md; + reg->s3 = d2 * (qs[1] & mask1) + md; + + reg->s4 = d1 * (qs[2] & mask0) + md; + reg->s5 = d2 * (qs[2] & mask1) + md; + + reg->s6 = d1 * (qs[3] & mask0) + md; + reg->s7 = d2 * (qs[3] & mask1) + md; + + reg->s8 = d1 * (qs[4] & mask0) + md; + reg->s9 = d2 * (qs[4] & mask1) + md; + + reg->sa = d1 * (qs[5] & mask0) + md; + reg->sb = d2 * (qs[5] & mask1) + md; + + reg->sc = d1 * (qs[6] & mask0) + md; + reg->sd = d2 * (qs[6] & mask1) + md; + + reg->se = d1 * (qs[7] & mask0) + md; + reg->sf = d2 * (qs[7] & mask1) + md; +} + +void dequantize_q4_0_f16(global struct block_q4_0 * xb, short il, half16 * reg) { + global ushort * qs = ((global ushort *)xb + 1); + half d1 = il ? (xb->d / 16.h) : xb->d; + half d2 = d1 / 256.h; + half md = -8.h * xb->d; + ushort mask0 = il ? 0x00F0 : 0x000F; + ushort mask1 = mask0 << 8; + + reg->s0 = d1 * (qs[0] & mask0) + md; + reg->s1 = d2 * (qs[0] & mask1) + md; + + reg->s2 = d1 * (qs[1] & mask0) + md; + reg->s3 = d2 * (qs[1] & mask1) + md; + + reg->s4 = d1 * (qs[2] & mask0) + md; + reg->s5 = d2 * (qs[2] & mask1) + md; + + reg->s6 = d1 * (qs[3] & mask0) + md; + reg->s7 = d2 * (qs[3] & mask1) + md; + + reg->s8 = d1 * (qs[4] & mask0) + md; + reg->s9 = d2 * (qs[4] & mask1) + md; + + reg->sa = d1 * (qs[5] & mask0) + md; + reg->sb = d2 * (qs[5] & mask1) + md; + + reg->sc = d1 * (qs[6] & mask0) + md; + reg->sd = d2 * (qs[6] & mask1) + md; + + reg->se = d1 * (qs[7] & mask0) + md; + reg->sf = d2 * (qs[7] & mask1) + md; +} + +//------------------------------------------------------------------------------ +// add +//------------------------------------------------------------------------------ + +// general-purpose kernel for addition of two tensors +// pros: works for non-contiguous tensors, supports broadcast across dims 1, 2 and 3 +// cons: not very efficient +kernel void kernel_add( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global char * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + int ne13, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + src0 = src0 + offset0; + src1 = src1 + offset1; + dst = dst + offsetd; + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int i13 = i03 % ne13; + int i12 = i02 % ne12; + int i11 = i01 % ne11; + + global char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01; + global char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; + global char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1; + + for (int i0 = get_local_id(0); i0 < ne0; i0 += get_local_size(0)) { + const int i10 = i0 % ne10; + *((global float *)(dst_ptr + i0*nb0)) = *((global float *)(src0_ptr + i0*nb00)) + *((global float *)(src1_ptr + i10*nb10)); + } +} + +// assumption: src1 is a row +// broadcast src1 into src0 +kernel void kernel_add_row( + global float4 * src0, + ulong offset0, + global float4 * src1, + ulong offset1, + global float4 * dst, + ulong offsetd, + int ne +) { + src0 = (global float4*)((global char*)src0 + offset0); + src1 = (global float4*)((global char*)src1 + offset1); + dst = (global float4*)((global char*)dst + offsetd); + + // This performs better than using %. + uint gid = get_global_id(0); + uint idx1 = gid - (gid/ne)*ne; // get_global_id(0) % ne + dst[gid] = src0[gid] + src1[idx1]; +} + +//------------------------------------------------------------------------------ +// mul +//------------------------------------------------------------------------------ +kernel void kernel_mul( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global char * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + int ne13, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + src0 = src0 + offset0; + src1 = src1 + offset1; + dst = dst + offsetd; + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int i13 = i03 % ne13; + int i12 = i02 % ne12; + int i11 = i01 % ne11; + + global char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01; + global char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11; + global char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1; + + for (int i0 = get_local_id(0); i0 < ne0; i0 += get_local_size(0)) { + const int i10 = i0 % ne10; + *((global float *)(dst_ptr + i0*nb0)) = *((global float *)(src0_ptr + i0*nb00)) * *((global float *)(src1_ptr + i10*nb10)); + } +} + +// assumption: src1 is a row +// broadcast src1 into src0 +kernel void kernel_mul_row( + global float4 * src0, + ulong offset0, + global float4 * src1, + ulong offset1, + global float4 * dst, + ulong offsetd, + int ne +) { + src0 = (global float4*)((global char*)src0 + offset0); + src1 = (global float4*)((global char*)src1 + offset1); + dst = (global float4*)((global char*)dst + offsetd); + + // This performs better than using %. + uint gid = get_global_id(0); + uint idx1 = gid - (gid/ne)*ne; // get_global_id(0) % ne + dst[gid] = src0[gid] * src1[idx1]; +} + +//------------------------------------------------------------------------------ +// scale +//------------------------------------------------------------------------------ +kernel void kernel_scale( + global float4 * src0, + ulong offset0, + global float4 * dst, + ulong offsetd, + float scale +) { + src0 = (global float4*)((global char*)src0 + offset0); + dst = (global float4*)((global char*)dst + offsetd); + dst[get_global_id(0)] = src0[get_global_id(0)] * scale; +} + +//------------------------------------------------------------------------------ +// gelu +//------------------------------------------------------------------------------ +#define GELU_COEF_A 0.044715f +#define SQRT_2_OVER_PI 0.79788456080286535587989211986876f + +kernel void kernel_gelu( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + float x = src0[get_global_id(0)]; + + dst[get_global_id(0)] = 0.5f*x*(1.0f + tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x))); +} + +kernel void kernel_gelu_4( + global float4 * src0, + ulong offset0, + global float4 * dst, + ulong offsetd +) { + src0 = (global float4*)((global char*)src0 + offset0); + dst = (global float4*)((global char*)dst + offsetd); + + float4 x = src0[get_global_id(0)]; + + dst[get_global_id(0)] = 0.5f*x*(1.0f + tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x))); +} + +//------------------------------------------------------------------------------ +// silu +//------------------------------------------------------------------------------ +kernel void kernel_silu( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + float x = src0[get_global_id(0)]; + dst[get_global_id(0)] = x / (1.0f + exp(-x)); +} + +kernel void kernel_silu_4( + global float4 * src0, + ulong offset0, + global float4 * dst, + ulong offsetd +) { + src0 = (global float4*)((global char*)src0 + offset0); + dst = (global float4*)((global char*)dst + offsetd); + + float4 x = src0[get_global_id(0)]; + dst[get_global_id(0)] = x / (1.0f + exp(-x)); +} + +//------------------------------------------------------------------------------ +// relu +//------------------------------------------------------------------------------ +kernel void kernel_relu( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + dst[get_global_id(0)] = fmax(0.0f, src0[get_global_id(0)]); +} + +//------------------------------------------------------------------------------ +// clamp +//------------------------------------------------------------------------------ +kernel void kernel_clamp( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd, + float min, + float max +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + dst[get_global_id(0)] = src0[get_global_id(0)] < min ? + min : + (src0[get_global_id(0)] > max ? max : src0[get_global_id(0)]); +} + +//------------------------------------------------------------------------------ +// norm +//------------------------------------------------------------------------------ +kernel void kernel_norm( + global void * src0, + ulong offset0, + global float * dst, + ulong offsetd, + int ne00, + ulong nb01, + float eps, + local float * sum +) { + src0 = (global void*)((global char*)src0 + offset0); + dst = (global void*)((global char*)dst + offsetd); + + global float * x = (global float *) ((global char *) src0 + get_group_id(0)*nb01); + + // MEAN + // parallel sum + sum[get_local_id(0)] = 0.0f; + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + sum[get_local_id(0)] += x[i00]; + } + // reduce + barrier(CLK_LOCAL_MEM_FENCE); + for (uint i = get_local_size(0)/2; i > 0; i /= 2) { + if (get_local_id(0) < i) { + sum[get_local_id(0)] += sum[get_local_id(0) + i]; + } + barrier(CLK_LOCAL_MEM_FENCE); + } + float mean = sum[0] / ne00; + + // recenter and VARIANCE + barrier(CLK_LOCAL_MEM_FENCE); + global float * y = dst + get_group_id(0)*ne00; + sum[get_local_id(0)] = 0.0f; + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + y[i00] = x[i00] - mean; + sum[get_local_id(0)] += y[i00] * y[i00]; + } + + // reduce + barrier(CLK_LOCAL_MEM_FENCE); + for (uint i = get_local_size(0)/2; i > 0; i /= 2) { + if (get_local_id(0) < i) { + sum[get_local_id(0)] += sum[get_local_id(0) + i]; + } + barrier(CLK_LOCAL_MEM_FENCE); + } + float variance = sum[0] / ne00; + + float scale = 1.0f/sqrt(variance + eps); + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + y[i00] = y[i00] * scale; + } +} + +//------------------------------------------------------------------------------ +// rms_norm +//------------------------------------------------------------------------------ +// This kernel depends on subgroup size. +kernel void kernel_rms_norm( + global void * src0, + ulong offset0, + global float * dst, + ulong offsetd, + int ne00, + ulong nb01, + float eps, + local float * sum // Note, the size depends on number of subgroups +) { + src0 = (global void*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + global float4 * x = (global float4 *) ((global char *) src0 + get_group_id(0)*nb01); + global float * x_scalar = (global float *) x; + float4 sumf = 0; + float all_sum = 0; + + // parallel sum + for (int i00 = get_local_id(0); i00 < ne00/4; i00 += get_local_size(0)) { + sumf += x[i00] * x[i00]; + } + all_sum = sumf.s0 + sumf.s1 + sumf.s2 + sumf.s3; + all_sum = sub_group_reduce_add(all_sum); + if (get_sub_group_local_id() == 0) { + sum[get_sub_group_id()] = all_sum; + } + + barrier(CLK_LOCAL_MEM_FENCE); + // broadcast + for (uint i = get_local_size(0) / get_max_sub_group_size() / 2; i > 0; i /= 2) { + if (get_local_id(0) < i) { + sum[get_local_id(0)] += sum[get_local_id(0) + i]; + } + } + if (get_local_id(0) == 0) { + for (int i = 4 * (ne00 / 4); i < ne00; i++) { + sum[0] += x_scalar[i]; + } + sum[0] /= ne00; + } + + barrier(CLK_LOCAL_MEM_FENCE); + + const float mean = sum[0]; + const float scale = 1.0f/sqrt(mean + eps); + + global float4 * y = (global float4 *) (dst + get_group_id(0)*ne00); + global float * y_scalar = (global float *) y; + for (int i00 = get_local_id(0); i00 < ne00/4; i00 += get_local_size(0)) { + y[i00] = x[i00] * scale; + } + if (get_local_id(0) == 0) { + for (int i00 = 4 * (ne00 / 4); i00 < ne00; i00++) { + y_scalar[i00] = x_scalar[i00] * scale; + } + } +} + +//------------------------------------------------------------------------------ +// diag_mask_inf kernels +//------------------------------------------------------------------------------ +kernel void kernel_diag_mask_inf( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int n_past +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + int i02 = get_global_id(2); + int i01 = get_global_id(1); + int i00 = get_global_id(0); + + if (i00 > n_past + i01) { + dst[i02*ne01*ne00 + i01*ne00 + i00] = -INFINITY; + } else { + dst[i02*ne01*ne00 + i01*ne00 + i00] = src0[i02*ne01*ne00 + i01*ne00 + i00]; + } +} + +kernel void kernel_diag_mask_inf_8( + global float4 * src0, + ulong offset0, + global float4 * dst, + ulong offsetd, + int ne00, + int ne01, + int n_past +) { + src0 = (global float4*)((global char*)src0 + offset0); + dst = (global float4*)((global char*)dst + offsetd); + + int i = 2*get_global_id(0); + + dst[i+0] = src0[i+0]; + dst[i+1] = src0[i+1]; + int i4 = 4*i; + int i02 = i4/(ne00*ne01); i4 -= i02*ne00*ne01; + int i01 = i4/(ne00); i4 -= i01*ne00; + int i00 = i4; + for (int k = 3; k >= 0; --k) { + if (i00 + 4 + k <= n_past + i01) { + break; + } + (&dst[i+1])[k] = -INFINITY; + if (i00 + k > n_past + i01) { + (&dst[i])[k] = -INFINITY; + } + } +} + +//------------------------------------------------------------------------------ +// softmax +//------------------------------------------------------------------------------ +kernel void kernel_soft_max( + global float * src0, + ulong offset0, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + float scale, + float max_bias, + float m0, + float m1, + int n_head_log2 +) { + src0 = (global float*)((global char*)src0 + offset0); + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + global float * psrc0 = src0 + i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + global float * pmask = src1 != src0 ? src1 + i01*ne00 : 0; + global float * pdst = dst + i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + + float slope = 1.0f; + + // ALiBi + if (max_bias > 0.0f) { + int h = i02; + + float base = h < n_head_log2 ? m0 : m1; + int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + + slope = pow(base, exp); + } + + // parallel max + float lmax = -INFINITY; + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + lmax = fmax(lmax, psrc0[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f)); + } + float max = sub_group_reduce_max(lmax); + + // parallel sum + float lsum = 0.0f; + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + float exp_psrc0 = exp((psrc0[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f)) - max); + lsum += exp_psrc0; + // Remember the result of exp here. exp is expensive, so we really do not + // wish to compute it twice. + pdst[i00] = exp_psrc0; + } + + const float sum = sub_group_reduce_add(lsum); + + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + pdst[i00] /= sum; + } +} + +#ifdef ADRENO_GPU +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_soft_max_4( + global float * src0, + ulong offset0, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + float scale, + float max_bias, + float m0, + float m1, + int n_head_log2 +) { + src0 = (global float*)((global char*)src0 + offset0); + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + global float4 * psrc4 = (global float4 *)(src0 + i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00); + global float4 * pmask = src1 != src0 ? (global float4 *)(src1 + i01*ne00) : 0; + global float4 * pdst4 = (global float4 *)(dst + i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00); + + float slope = 1.0f; + + // ALiBi + if (max_bias > 0.0f) { + int h = i02; + + float base = h < n_head_log2 ? m0 : m1; + int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + + slope = pow(base, exp); + } + + // parallel max + float4 lmax4 = -INFINITY; + for (int i00 = get_local_id(0); i00 < ne00/4; i00 += get_local_size(0)) { + lmax4 = fmax(lmax4, psrc4[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f)); + } + float lmax = fmax(fmax(lmax4.s0, lmax4.s1), fmax(lmax4.s2, lmax4.s3)); + + const float max = sub_group_reduce_max(lmax); + + // parallel sum + float4 lsum4 = 0.0f; + for (int i00 = get_local_id(0); i00 < ne00/4; i00 += get_local_size(0)) { + const float4 exp_psrc4 = exp((psrc4[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f)) - max); + lsum4 += exp_psrc4; + pdst4[i00] = exp_psrc4; + } + float lsum = lsum4.s0 + lsum4.s1 + lsum4.s2 + lsum4.s3; + + const float sum = sub_group_reduce_add(lsum); + + for (int i00 = get_local_id(0); i00 < ne00/4; i00 += get_local_size(0)) { + pdst4[i00] /= sum; + } +} + +//------------------------------------------------------------------------------ +// kernel_rope +//------------------------------------------------------------------------------ +float rope_yarn_ramp(float low, float high, int i0) { + const float y = (i0 / 2 - low) / max(0.001f, high - low); + return 1.0f - min(1.0f, max(0.0f, y)); +} + +// YaRN algorithm based on LlamaYaRNScaledRotaryEmbedding.py from https://github.com/jquesnelle/yarn +// MIT licensed. Copyright (c) 2023 Jeffrey Quesnelle and Bowen Peng. +float2 rope_yarn( + float theta_extrap, float freq_scale, float2 corr_dims, int i0, float ext_factor, float mscale +) { + // Get n-d rotational scaling corrected for extrapolation + float theta_interp = freq_scale * theta_extrap; + float theta = theta_interp; + if (ext_factor != 0.0f) { + float ramp_mix = rope_yarn_ramp(corr_dims.s0, corr_dims.s1, i0) * ext_factor; + theta = theta_interp * (1 - ramp_mix) + theta_extrap * ramp_mix; + + // Get n-d magnitude scaling corrected for interpolation + mscale *= 1.0f + 0.1f * log(1.0f / freq_scale); + } + return (float2)(cos(theta) * mscale, sin(theta) * mscale); +} + +// Apparently solving `n_rot = 2pi * x * base^((2 * max_pos_emb) / n_dims)` for x, we get +// `corr_fac(n_rot) = n_dims * log(max_pos_emb / (n_rot * 2pi)) / (2 * log(base))` +float rope_yarn_corr_factor(int n_dims, int n_ctx_orig, float n_rot, float base) { + return n_dims * log(n_ctx_orig / (n_rot * 2 * M_PI_F)) / (2 * log(base)); +} + +float2 rope_yarn_corr_dims( + int n_dims, int n_ctx_orig, float freq_base, float beta_fast, float beta_slow +) { + // start and end correction dims + return (float2)( + max(0.0f, floor(rope_yarn_corr_factor(n_dims, n_ctx_orig, beta_fast, freq_base))), + min(n_dims - 1.0f, ceil(rope_yarn_corr_factor(n_dims, n_ctx_orig, beta_slow, freq_base))) + ); +} + +kernel void kernel_rope_norm_f32( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * src2, + ulong offset2, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3, + int n_past, + int n_dims, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + src2 = (global float*)((global char*)src2 + offset2); + dst = (global float*)((global char*)dst + offsetd); + + int i3 = get_group_id(2); + int i2 = get_group_id(1); + int i1 = get_group_id(0); + + float2 corr_dims = rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow); + + global int * pos = src1; + + float theta_base = (float) pos[i2]; + float inv_ndims = -1.f/n_dims; + + for (int i0 = 2*get_local_id(0); i0 < ne0; i0 += 2*get_local_size(0)) { + if (i0 < n_dims) { + int ic = i0/2; + + float theta = theta_base * pow(freq_base, inv_ndims*i0); + + float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + + float2 cos_sin_theta = rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor); + + global float * src = (global float *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global float * dst_data = (global float *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + float x0 = src[0]; + float x1 = src[1]; + + dst_data[0] = x0*cos_sin_theta.s0 - x1*cos_sin_theta.s1; + dst_data[1] = x0*cos_sin_theta.s1 + x1*cos_sin_theta.s0; + } else { + global float * src = (global float *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global float * dst_data = (global float *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } + } +} + +kernel void kernel_rope_norm_f16( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * src2, + ulong offset2, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3, + int n_past, + int n_dims, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + src2 = (global float*)((global char*)src2 + offset2); + dst = (global float*)((global char*)dst + offsetd); + + int i3 = get_group_id(2); + int i2 = get_group_id(1); + int i1 = get_group_id(0); + + float2 corr_dims = rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow); + + global int * pos = src1; + + float theta_base = (float) pos[i2]; + float inv_ndims = -1.f/n_dims; + + for (int i0 = 2*get_local_id(0); i0 < ne0; i0 += 2*get_local_size(0)) { + if (i0 < n_dims) { + int ic = i0/2; + + float theta = theta_base * pow(freq_base, inv_ndims*i0); + + float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + + float2 cos_sin_theta = rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor); + + global half * src = (global half *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global half * dst_data = (global half *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + float x0 = src[0]; + float x1 = src[1]; + + dst_data[0] = x0*cos_sin_theta.s0 - x1*cos_sin_theta.s1; + dst_data[1] = x0*cos_sin_theta.s1 + x1*cos_sin_theta.s0; + } else { + global half * src = (global half *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global half * dst_data = (global half *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } + } +} + +kernel void kernel_rope_neox_f32( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * src2, + ulong offset2, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3, + int n_past, + int n_dims, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + src2 = (global float*)((global char*)src2 + offset2); + dst = (global float*)((global char*)dst + offsetd); + + int i3 = get_group_id(2); + int i2 = get_group_id(1); + int i1 = get_group_id(0); + + float2 corr_dims = rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow); + + global int * pos = src1; + + float theta_base = (float) pos[i2]; + float inv_ndims = -1.f/n_dims; + + for (int i0 = 2*get_local_id(0); i0 < ne0; i0 += 2*get_local_size(0)) { + if (i0 < n_dims) { + int ic = i0/2; + + const float theta = theta_base * pow(freq_base, inv_ndims*i0); + + const float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + + float2 cos_sin_theta = rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor); + + global float * src = (global float *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + global float * dst_data = (global float *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = src[0]; + const float x1 = src[n_dims/2]; + + dst_data[0] = x0*cos_sin_theta.s0 - x1*cos_sin_theta.s1; + dst_data[n_dims/2] = x0*cos_sin_theta.s1 + x1*cos_sin_theta.s0; + } else { + global float * const src = (global float *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global float * dst_data = (global float *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } + } +} + +kernel void kernel_rope_neox_f16( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * src2, + ulong offset2, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3, + int n_past, + int n_dims, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + src2 = (global float*)((global char*)src2 + offset2); + dst = (global float*)((global char*)dst + offsetd); + + int i3 = get_group_id(2); + int i2 = get_group_id(1); + int i1 = get_group_id(0); + + float2 corr_dims = rope_yarn_corr_dims(n_dims, n_ctx_orig, freq_base, beta_fast, beta_slow); + + global int * pos = src1; + + float theta_base = (float) pos[i2]; + float inv_ndims = -1.f/n_dims; + + for (int i0 = 2*get_local_id(0); i0 < ne0; i0 += 2*get_local_size(0)) { + if (i0 < n_dims) { + int ic = i0/2; + + const float theta = theta_base * pow(freq_base, inv_ndims*i0); + + const float freq_factor = src2 != src0 ? src2[ic] : 1.0f; + + float2 cos_sin_theta = rope_yarn(theta/freq_factor, freq_scale, corr_dims, i0, ext_factor, attn_factor); + + global half * src = (global half *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + ic*nb00); + global half * dst_data = (global half *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + ic*nb0); + + const float x0 = src[0]; + const float x1 = src[n_dims/2]; + + dst_data[0] = x0*cos_sin_theta.s0 - x1*cos_sin_theta.s1; + dst_data[n_dims/2] = x0*cos_sin_theta.s1 + x1*cos_sin_theta.s0; + } else { + global half * const src = (global half *)((global char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00); + global half * dst_data = (global half *)((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + dst_data[0] = src[0]; + dst_data[1] = src[1]; + } + } +} + +//------------------------------------------------------------------------------ +// cpy +//------------------------------------------------------------------------------ + +kernel void kernel_cpy_f16_f16( + global half * src0, + ulong offset0, + global half * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + src0 = (global half*)((global char*)src0 + offset0); + dst = (global half*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + + int i3 = n / (ne2*ne1*ne0); + int i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); + int i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; + int i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0); + + global half * dst_data = (global half *) ((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + global const half * src = (global half *)((global char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + dst_data[i00] = src[0]; + } +} + +kernel void kernel_cpy_f16_f32( + global half * src0, + ulong offset0, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + + src0 = (global half*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + + int i3 = n / (ne2*ne1*ne0); + int i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); + int i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; + int i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0); + + global float * dst_data = (global float *) ((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + global half * src = (global half *)((global char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + dst_data[i00] = src[0]; + } +} + +kernel void kernel_cpy_f32_f16( + global float * src0, + ulong offset0, + global half * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global half*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + + int i3 = n / (ne2*ne1*ne0); + int i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); + int i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; + int i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0); + + global half * dst_data = (global half *) ((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + global const float * src = (global float *)((global char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + + dst_data[i00] = src[0]; + } +} + +kernel void kernel_cpy_f32_f32( + global float * src0, + ulong offset0, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne03, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne0, + int ne1, + int ne2, + int ne3, + ulong nb0, + ulong nb1, + ulong nb2, + ulong nb3 +) { + src0 = (global float*)((global char*)src0 + offset0); + dst = (global float*)((global char*)dst + offsetd); + + int i03 = get_group_id(2); + int i02 = get_group_id(1); + int i01 = get_group_id(0); + + int n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00; + + int i3 = n / (ne2*ne1*ne0); + int i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0); + int i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0; + int i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0); + + global float * dst_data = (global float *) ((global char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0); + + for (int i00 = get_local_id(0); i00 < ne00; i00 += get_local_size(0)) { + global const float * src = (global float *)((global char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00); + + dst_data[i00] = src[0]; + } +} + +//------------------------------------------------------------------------------ +// get_rows +//------------------------------------------------------------------------------ +kernel void kernel_get_rows_f32( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + ulong nb01, + ulong nb02, + int ne10, + ulong nb10, + ulong nb11, + ulong nb1, + ulong nb2 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int i10 = get_group_id(0); + int i11 = get_group_id(1); + + int r = ((global int *) ((global char *) src1 + i11*nb11 + i10*nb10))[0]; + + int i02 = i11; + + for (int ind = get_local_id(0); ind < ne00; ind += get_local_size(0)) { + ((global float *) ((global char *) dst + i11*nb2 + i10*nb1))[ind] = + ((global float *) ((global char *) src0 + r*nb01 + i02*nb02))[ind]; + } +} + +kernel void kernel_get_rows_f16( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + ulong nb01, + ulong nb02, + int ne10, + ulong nb10, + ulong nb11, + ulong nb1, + ulong nb2 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int i10 = get_group_id(0); + int i11 = get_group_id(1); + + int r = ((global int32_t *) ((global char *) src1 + i11*nb11 + i10*nb10))[0]; + + int i02 = i11; + + for (int ind = get_local_id(0); ind < ne00; ind += get_local_size(0)) { + ((global float *) ((global char *) dst + i11*nb2 + i10*nb1))[ind] = + ((global half *) ((global char *) src0 + r*nb01 + i02*nb02))[ind]; + } +} + +kernel void kernel_get_rows_q4_0( + global void * src0, + ulong offset0, + global int * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + ulong nb01, + ulong nb02, + int ne10, + ulong nb10, + ulong nb11, + ulong nb1, + ulong nb2 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global int*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + const int NL = 2; + + int i10 = get_group_id(0); + int i11 = get_group_id(1); + + int r = ((global int32_t *) ((global char *) src1 + i11*nb11 + i10*nb10))[0]; + + int i02 = i11; + + for (int ind = get_local_id(0); ind < ne00/16; ind += get_local_size(0)) { + float16 temp; + dequantize_q4_0_f32( + ((global struct block_q4_0 *) ((global char *) src0 + r*nb01 + i02*nb02)) + ind/NL, ind%NL, &temp); + *(((global float16 *) ((global char *) dst + i11*nb2 + i10*nb1)) + ind) = temp; + } +} + +//------------------------------------------------------------------------------ +// mul_mat_f32_f32 +//------------------------------------------------------------------------------ +#define N_F32_F32 4 + +kernel void kernel_mul_mat_f32_f32( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global char*)((global char*)src0 + offset0); + src1 = (global char*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int r0 = get_group_id(0); + int rb = get_group_id(1)*N_F32_F32; + int im = get_group_id(2); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + + global float * x = (global float *) (src0 + offset_src0); + + if (ne00 < 128) { + for (int row = 0; row < N_F32_F32; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global float * y = (global float *) (src1 + offset_src1); + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00; i += get_max_sub_group_size()) { + sumf += (float) x[i] * (float) y[i]; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } else { + global float4 * x4 = (global float4 *)x; + for (int row = 0; row < N_F32_F32; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global float * y = (global float *) (src1 + offset_src1); + global float4 * y4 = (global float4 *) y; + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00/4; i += get_max_sub_group_size()) { + sumf += (float) x4[i].s0 * y4[i].s0; + sumf += (float) x4[i].s1 * y4[i].s1; + sumf += (float) x4[i].s2 * y4[i].s2; + sumf += (float) x4[i].s3 * y4[i].s3; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + for (int i = 4*(ne00/4); i < ne00; ++i) { + all_sum += (float) x[i] * y[i]; + } + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } +} + +//------------------------------------------------------------------------------ +// mul_mat_f16_f16 +//------------------------------------------------------------------------------ +#define N_F16_F16 4 + +kernel void kernel_mul_mat_f16_f16( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int r2, + int r3) +{ + src0 = (global char*)((global char*)src0 + offset0); + src1 = (global char*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int r0 = get_group_id(0); + int rb = get_group_id(1)*N_F16_F16; + int im = get_group_id(2); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + + global half * x = (global half *) (src0 + offset_src0); + + if (ne00 < 128) { + for (int row = 0; row < N_F16_F16; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global half * y = (global half *) (src1 + offset_src1); + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00; i += get_max_sub_group_size()) { + sumf += (half) x[i] * (half) y[i]; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } else { + global half4 * x4 = (global half4 *)x; + for (int row = 0; row < N_F16_F16; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global half * y = (global half *) (src1 + offset_src1); + global half4 * y4 = (global half4 *) y; + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00/4; i += get_max_sub_group_size()) { + sumf += (half) x4[i].s0 * y4[i].s0; + sumf += (half) x4[i].s1 * y4[i].s1; + sumf += (half) x4[i].s2 * y4[i].s2; + sumf += (half) x4[i].s3 * y4[i].s3; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + for (int i = 4*(ne00/4); i < ne00; ++i) { + all_sum += (half) x[i] * y[i]; + } + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } +} + +//------------------------------------------------------------------------------ +// mul_mat_f16_f32_1row +//------------------------------------------------------------------------------ +kernel void kernel_mul_mat_f16_f32_1row( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global char*)((global char*)src0 + offset0); + src1 = (global char*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global half * x = (global half *) (src0 + offset_src0); + global float * y = (global float *) (src1 + offset_src1); + + float sumf = 0; + if (ne00 < 128) { + for (int i = get_sub_group_local_id(); i < ne00; i += get_max_sub_group_size()) { + sumf += (float) x[i] * (float) y[i]; + } + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } else { + global half4 * x4 = (global half4 *) x; + global float4 * y4 = (global float4 *) y; + for (int i = get_sub_group_local_id(); i < ne00/4; i += get_max_sub_group_size()) { + sumf += (float) x4[i].s0 * y4[i].s0; + sumf += (float) x4[i].s1 * y4[i].s1; + sumf += (float) x4[i].s2 * y4[i].s2; + sumf += (float) x4[i].s3 * y4[i].s3; + } + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + for (int i = 4*(ne00/4); i < ne00; ++i) { + all_sum += (float) x[i] * y[i]; + } + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + +} + +//------------------------------------------------------------------------------ +// mul_mat_f16_f32 +//------------------------------------------------------------------------------ +#define N_F16_F32 4 + +#ifdef ADRENO_GPU +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_f16_f32( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global char*)((global char*)src0 + offset0); + src1 = (global char*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int r0 = get_group_id(0); + int rb = get_group_id(1)*N_F16_F32; + int im = get_group_id(2); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + + global half * x = (global half *) (src0 + offset_src0); + + if (ne00 < 128) { + for (int row = 0; row < N_F16_F32; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global float * y = (global float *) (src1 + offset_src1); + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00; i += get_max_sub_group_size()) { + sumf += convert_float(x[i]) * y[i]; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } else { + global half4 * x4 = (global half4 *)x; + for (int row = 0; row < N_F16_F32; ++row) { + int r1 = rb + row; + if (r1 >= ne11) { + break; + } + + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global float * y = (global float *) (src1 + offset_src1); + global float4 * y4 = (global float4 *) y; + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00/4; i += get_max_sub_group_size()) { + sumf += convert_float(x4[i].s0) * y4[i].s0; + sumf += convert_float(x4[i].s1) * y4[i].s1; + sumf += convert_float(x4[i].s2) * y4[i].s2; + sumf += convert_float(x4[i].s3) * y4[i].s3; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + for (int i = 4*(ne00/4); i < ne00; ++i) { + all_sum += (float) x[i] * y[i]; + } + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } + } +} + +//------------------------------------------------------------------------------ +// mul_mat_f16_f32_l4 +//------------------------------------------------------------------------------ +// Assumes row size (ne00) is a multiple of 4 +#ifdef ADRENO_GPU +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_f16_f32_l4( + global char * src0, + ulong offset0, + global char * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + ulong nb00, + ulong nb01, + ulong nb02, + ulong nb03, + int ne10, + int ne11, + int ne12, + ulong nb10, + ulong nb11, + ulong nb12, + ulong nb13, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global char*)((global char*)src0 + offset0); + src1 = (global char*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + int nrows = ne11; + int r0 = get_group_id(0); + int im = get_group_id(2); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb03; + + global half4 * x4 = (global half4 *) (src0 + offset_src0); + + for (int r1 = 0; r1 < nrows; ++r1) { + ulong offset_src1 = r1*nb11 + (i12 )*nb12 + (i13 )*nb13; + + global float4 * y4 = (global float4 *) (src1 + offset_src1); + + float sumf = 0; + for (int i = get_sub_group_local_id(); i < ne00/4; i += get_max_sub_group_size()) { + sumf += convert_float(x4[i].s0) * y4[i].s0; + sumf += convert_float(x4[i].s1) * y4[i].s1; + sumf += convert_float(x4[i].s2) * y4[i].s2; + sumf += convert_float(x4[i].s3) * y4[i].s3; + } + + float all_sum = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum; + } + } +} + +//------------------------------------------------------------------------------ +// mul_vec_q_n_f32 +//------------------------------------------------------------------------------ +// function for calculate inner product between half a q4_0 block and 16 floats (yl), sumy is SUM(yl[i]) +// il indicates where the q4 quants begin (0 or QK4_0/4) +// we assume that the yl's have been multiplied with the appropriate scale factor +// that corresponds to the missing bit shifts (1, 1/16, 1/256, 1/4096) +inline float block_q_4_0_dot_y( + global struct block_q4_0 * qb_curr, + float sumy, + private float * yl, + int il +) { + float d = qb_curr->d; + float2 acc = 0.f; + global ushort * qs = ((global ushort *)qb_curr + 1 + il/2); + for (int i = 0; i < 8; i+=2) { + acc.s0 += yl[i + 0] * (qs[i / 2] & 0x000F) + + yl[i + 1] * (qs[i / 2] & 0x0F00); + acc.s1 += yl[i + 8] * (qs[i / 2] & 0x00F0) + + yl[i + 9] * (qs[i / 2] & 0xF000); + } + return d * (sumy * -8.f + acc.s0 + acc.s1); +} + +#ifdef INTEL_GPU +#define N_DST 4 // each SIMD group works on 4 rows +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 16 +#elif defined (ADRENO_GPU) +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +inline void mul_vec_q_n_f32( + global void * src0, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + + const ulong nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is essenatially the linear global + // id of a SIMD group in the grid. + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset0 = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + + global struct block_q4_0 * x = (global struct block_q4_0 *) src0 + offset0; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float yl[16]; // src1 vector cache + float sumf[N_DST]={0.f}; + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix * QK4_0 + il; + + // each thread in a SIMD group deals with half a block. + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0; + for (int i = 0; i < 8; i += 2) { + sumy += yb[i] + yb[i+1]; + yl[i+0] = yb[i+ 0]; + yl[i+1] = yb[i+ 1]/256.f; + sumy += yb[i+16] + yb[i+17]; + yl[i+8] = yb[i+16]/16.f; + yl[i+9] = yb[i+17]/4096.f; + } + + for (int row = 0; row < N_DST; row++) { + sumf[row] += block_q_4_0_dot_y(x+ib+row*nb, sumy, yl, il); + } + + // One thread in a SIMD group (i.e., subgroup) handles a half block, + // hence then entire SIMD group handles SIMDWIDTH/2 blocks. + // y points to the activation matrix (of type float). Therefore for + // one thread, the # of blocks y should advance is SIMDWIDTH/2 (because + // SIMDWIDTH/2 blocks are processed by a SIMD group) - in terms of + // floats, it is QK4_0 * (SIMDWIDTH/2), where QK4_0 is the block size. + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + // The above does not work for Adreno - it produces incorrect results for + // row = 1, 2, 3 and only row = 0 gives the correct result. + // If N_DST is changed, the below array must be initialized accordingly. + // This also seems to perform better on Intel. + float tot[N_DST] = { + sub_group_reduce_add(sumf[0]), sub_group_reduce_add(sumf[1]), + sub_group_reduce_add(sumf[2]), sub_group_reduce_add(sumf[3])}; + for (int row = 0; row < N_DST; ++row) { + if (get_sub_group_local_id() == 0 && first_row + row < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot[row]; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32( + global void * src0, + ulong offset0, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_vec_q_n_f32(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} + +// +// This variant unrolls the loops and uses vector types instead of pointers. +// It improves performance on Adreno but not so much on Intel. +// +inline float block_q_4_0_dot_y_v( + global struct block_q4_0 * qb_curr, + float sumy, + float16 yl, + int il +) { + float d = qb_curr->d; + float acc = 0.f; + global ushort * qs = ((global ushort *)qb_curr + 1 + il/2); + + acc += yl.s0 * (qs[0] & 0x000F); + acc += yl.s1 * (qs[0] & 0x0F00); + acc += yl.s8 * (qs[0] & 0x00F0); + acc += yl.s9 * (qs[0] & 0xF000); + + acc += yl.s2 * (qs[1] & 0x000F); + acc += yl.s3 * (qs[1] & 0x0F00); + acc += yl.sa * (qs[1] & 0x00F0); + acc += yl.sb * (qs[1] & 0xF000); + + acc += yl.s4 * (qs[2] & 0x000F); + acc += yl.s5 * (qs[2] & 0x0F00); + acc += yl.sc * (qs[2] & 0x00F0); + acc += yl.sd * (qs[2] & 0xF000); + + acc += yl.s6 * (qs[3] & 0x000F); + acc += yl.s7 * (qs[3] & 0x0F00); + acc += yl.se * (qs[3] & 0x00F0); + acc += yl.sf * (qs[3] & 0xF000); + + return d * (sumy * -8.f + acc); +} + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 4 // each SIMD group works on 4 rows +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 16 +#elif defined (ADRENO_GPU) +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +inline void mul_vec_q_n_f32_v( + global void * src0, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + const ulong nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is essenatially the linear global + // id of a SIMD group in the grid. + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset0 = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + + global struct block_q4_0 * x = (global struct block_q4_0 *) src0 + offset0; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float16 yl; // src1 vector cache + float4 sumf = (float4)(0.f, 0.f, 0.f, 0.f); + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix * QK4_0 + il; + + // each thread in a SIMD group deals with half a block. + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0; + + sumy += yb[0]; + sumy += yb[1]; + sumy += yb[2]; + sumy += yb[3]; + sumy += yb[4]; + sumy += yb[5]; + sumy += yb[6]; + sumy += yb[7]; + + sumy += yb[16]; + sumy += yb[17]; + sumy += yb[18]; + sumy += yb[19]; + sumy += yb[20]; + sumy += yb[21]; + sumy += yb[22]; + sumy += yb[23]; + + + yl.s0 = yb[0]; + yl.s1 = yb[1]/256.f; + + yl.s2 = yb[2]; + yl.s3 = yb[3]/256.f; + + yl.s4 = yb[4]; + yl.s5 = yb[5]/256.f; + + yl.s6 = yb[6]; + yl.s7 = yb[7]/256.f; + + yl.s8 = yb[16]/16.f; + yl.s9 = yb[17]/4096.f; + + yl.sa = yb[18]/16.f; + yl.sb = yb[19]/4096.f; + + yl.sc = yb[20]/16.f; + yl.sd = yb[21]/4096.f; + + yl.se = yb[22]/16.f; + yl.sf = yb[23]/4096.f; + + sumf.s0 += block_q_4_0_dot_y_v(x+ib+0*nb, sumy, yl, il); + sumf.s1 += block_q_4_0_dot_y_v(x+ib+1*nb, sumy, yl, il); + sumf.s2 += block_q_4_0_dot_y_v(x+ib+2*nb, sumy, yl, il); + sumf.s3 += block_q_4_0_dot_y_v(x+ib+3*nb, sumy, yl, il); + + // One thread in a SIMD group (i.e., subgroup) handles a half block, + // hence then entire SIMD group handles SIMDWIDTH/2 blocks. + // y points to the activation matrix (of type float). Therefore for + // one thread, the # of blocks y should advance is SIMDWIDTH/2 (because + // SIMDWIDTH/2 blocks are processed by a SIMD group) - in terms of + // floats, it is QK4_0 * (SIMDWIDTH/2), where QK4_0 is the block size. + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + // The above does not work for Adreno - it produces incorrect results for + // row = 1, 2, 3 and only row = 0 gives the correct result. + // If N_DST is changed, the below array must be initialized accordingly. + // This also seems to perform better on Intel. + float4 tot = (float4)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1), + sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3) + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_v( + global void * src0, + ulong offset0, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_vec_q_n_f32_v(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} + +//------------------------------------------------------------------------------ +// kernel_convert_block_q4_0 +// Convert the block_q4_0 format to 2 separate arrays (AOS -> SOA). +// This kernel does not deshuffle the bits. +//------------------------------------------------------------------------------ +kernel void kernel_convert_block_q4_0( + global struct block_q4_0 * src0, + global uchar * dst_q, + global half * dst_d +) { + global struct block_q4_0 * b = (global struct block_q4_0 *) src0 + get_global_id(0); + global uchar * q = (global uchar *) dst_q + QK4_0/2*get_global_id(0); + global half * d = (global half *) dst_d + get_global_id(0); + + *d = b->d; + + for (int i = 0; i < QK4_0/2; ++i) { + q[i] = b->qs[i]; + } +} + +kernel void kernel_restore_block_q4_0( + global uchar * src_q, + global half * src_d, + global struct block_q4_0 * dst +) { + global struct block_q4_0 * b = (global struct block_q4_0 *) dst + get_global_id(0); + global uchar * q = (global uchar *) src_q + QK4_0/2*get_global_id(0); + global half * d = (global half *) src_d + get_global_id(0); + + b->d = *d; + for (int i = 0; i < QK4_0/2; ++i) { + b->qs[i] = q[i]; + } +} + +//------------------------------------------------------------------------------ +// mul_vec_q_n_f32_flat +// +// This variation uses flat arrays (struct of arrays, SOA) representation for +// quant tensors. +//------------------------------------------------------------------------------ + +// This function requires the original shuffled weights. +// As a reminder, the original weights are shuffled so that (q[0], q[16]) are +// packed together in a byte, so are (q[1], q[17]) and so on. +inline float block_q_4_0_dot_y_flat( + global uchar * x, + global half * dh, + float sumy, + float16 yl, + int il +) { + float d = *dh; + global ushort * qs = ((global ushort *)x + il/2); + float acc = 0.f; + + acc += yl.s0 * (qs[0] & 0x000F); + acc += yl.s1 * (qs[0] & 0x0F00); + acc += yl.s8 * (qs[0] & 0x00F0); + acc += yl.s9 * (qs[0] & 0xF000); + + acc += yl.s2 * (qs[1] & 0x000F); + acc += yl.s3 * (qs[1] & 0x0F00); + acc += yl.sa * (qs[1] & 0x00F0); + acc += yl.sb * (qs[1] & 0xF000); + + acc += yl.s4 * (qs[2] & 0x000F); + acc += yl.s5 * (qs[2] & 0x0F00); + acc += yl.sc * (qs[2] & 0x00F0); + acc += yl.sd * (qs[2] & 0xF000); + + acc += yl.s6 * (qs[3] & 0x000F); + acc += yl.s7 * (qs[3] & 0x0F00); + acc += yl.se * (qs[3] & 0x00F0); + acc += yl.sf * (qs[3] & 0xF000); + + return d * (sumy * -8.f + acc); +} + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 4 // each SIMD group works on 4 rows +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 32 +#elif defined (ADRENO_GPU) +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +inline void mul_vec_q_n_f32_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + const ulong nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is the linear global id of + // a SIMD group in the grid. Each SIMD group produces N_DST values in the + // result, hence uses nb blocks, i.e., the offset becomes first_row*nb. + // Currently with llama2 7B, im is always 0. + // TODO: how to handle im/gqa*(nb*ne0)? + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = (first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02)) * QK4_0/2; + + global uchar * x = (global uchar *) src0_q + offset0_q; + global half * d = (global half *) src0_d + offset0_d; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float16 yl; + float4 sumf = (float4)(0.f, 0.f, 0.f, 0.f); + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix*QK4_0 + il; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0.f; + + sumy += yb[0]; + sumy += yb[1]; + sumy += yb[2]; + sumy += yb[3]; + sumy += yb[4]; + sumy += yb[5]; + sumy += yb[6]; + sumy += yb[7]; + + sumy += yb[16]; + sumy += yb[17]; + sumy += yb[18]; + sumy += yb[19]; + sumy += yb[20]; + sumy += yb[21]; + sumy += yb[22]; + sumy += yb[23]; + + yl.s0 = yb[0]; + yl.s1 = yb[1]/256.f; + + yl.s2 = yb[2]; + yl.s3 = yb[3]/256.f; + + yl.s4 = yb[4]; + yl.s5 = yb[5]/256.f; + + yl.s6 = yb[6]; + yl.s7 = yb[7]/256.f; + + yl.s8 = yb[16]/16.f; + yl.s9 = yb[17]/4096.f; + + yl.sa = yb[18]/16.f; + yl.sb = yb[19]/4096.f; + + yl.sc = yb[20]/16.f; + yl.sd = yb[21]/4096.f; + + yl.se = yb[22]/16.f; + yl.sf = yb[23]/4096.f; + + sumf.s0 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 0*nb*QK4_0/2, d + ib + 0*nb, sumy, yl, il); + sumf.s1 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 1*nb*QK4_0/2, d + ib + 1*nb, sumy, yl, il); + sumf.s2 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 2*nb*QK4_0/2, d + ib + 2*nb, sumy, yl, il); + sumf.s3 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 3*nb*QK4_0/2, d + ib + 3*nb, sumy, yl, il); + + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + float4 tot = (float4)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1), + sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3) + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_vec_q_n_f32_flat(src0_q, src0_d, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} + +// +// This variant outputs 8 values. +// +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 8 // each SIMD group works on 8 rows +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 32 +#elif defined (ADRENO_GPU) +#define N_DST 8 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +inline void mul_vec_q_n_f32_8x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + const ulong nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is the linear global id of + // a SIMD group in the grid. Each SIMD group produces N_DST values in the + // result, hence uses nb blocks, i.e., the offset becomes first_row*nb. + // Currently with llama2 7B, im is always 0. + // TODO: how to handle im/gqa*(nb*ne0)? + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = (first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02)) * QK4_0/2; + + global uchar * x = (global uchar *) src0_q + offset0_q; + global half * d = (global half *) src0_d + offset0_d; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float16 yl; + float8 sumf = 0.f; + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix*QK4_0 + il; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0.f; + + sumy += yb[0]; + sumy += yb[1]; + sumy += yb[2]; + sumy += yb[3]; + sumy += yb[4]; + sumy += yb[5]; + sumy += yb[6]; + sumy += yb[7]; + + sumy += yb[16]; + sumy += yb[17]; + sumy += yb[18]; + sumy += yb[19]; + sumy += yb[20]; + sumy += yb[21]; + sumy += yb[22]; + sumy += yb[23]; + + yl.s0 = yb[0]; + yl.s1 = yb[1]/256.f; + + yl.s2 = yb[2]; + yl.s3 = yb[3]/256.f; + + yl.s4 = yb[4]; + yl.s5 = yb[5]/256.f; + + yl.s6 = yb[6]; + yl.s7 = yb[7]/256.f; + + yl.s8 = yb[16]/16.f; + yl.s9 = yb[17]/4096.f; + + yl.sa = yb[18]/16.f; + yl.sb = yb[19]/4096.f; + + yl.sc = yb[20]/16.f; + yl.sd = yb[21]/4096.f; + + yl.se = yb[22]/16.f; + yl.sf = yb[23]/4096.f; + + sumf.s0 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 0*nb*QK4_0/2, d + ib + 0*nb, sumy, yl, il); + sumf.s1 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 1*nb*QK4_0/2, d + ib + 1*nb, sumy, yl, il); + sumf.s2 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 2*nb*QK4_0/2, d + ib + 2*nb, sumy, yl, il); + sumf.s3 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 3*nb*QK4_0/2, d + ib + 3*nb, sumy, yl, il); + + sumf.s4 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 4*nb*QK4_0/2, d + ib + 4*nb, sumy, yl, il); + sumf.s5 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 5*nb*QK4_0/2, d + ib + 5*nb, sumy, yl, il); + sumf.s6 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 6*nb*QK4_0/2, d + ib + 6*nb, sumy, yl, il); + sumf.s7 += block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 7*nb*QK4_0/2, d + ib + 7*nb, sumy, yl, il); + + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + float8 tot = (float8)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1), + sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3), + sub_group_reduce_add(sumf.s4), sub_group_reduce_add(sumf.s5), + sub_group_reduce_add(sumf.s6), sub_group_reduce_add(sumf.s7) + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } + + if (first_row + 4 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 4] = tot.s4; + } + if (first_row + 5 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 5] = tot.s5; + } + if (first_row + 6 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 6] = tot.s6; + } + if (first_row + 7 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 7] = tot.s7; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_8x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_vec_q_n_f32_8x_flat(src0_q, src0_d, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_cvt.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_cvt.cl new file mode 100644 index 000000000..e2024332f --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_cvt.cl @@ -0,0 +1,106 @@ +//------------------------------------------------------------------------------ +// This file is contains additional kernels for data conversion. +// These kernels are used when loading the model, so its performance is less +// important. +//------------------------------------------------------------------------------ +#ifdef cl_khr_fp16 +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#elif defined(cl_amd_fp16) +#pragma OPENCL EXTENSION cl_amd_fp16 : enable +#else +#error "Half precision floating point not supportedby OpenCL implementation on your device." +#endif + +#ifdef cl_khr_subgroups +#pragma OPENCL EXTENSION cl_khr_subgroups : enable +#elif defined(cl_intel_subgroups) +#pragma OPENCL EXTENSION cl_intel_subgroups : enable +#else +#error "Subgroup not supported on your device." +#endif + +#ifdef cl_intel_required_subgroup_size +// Always use subgroup size of 32 on Intel. +#pragma OPENCL EXTENSION cl_intel_required_subgroup_size : enable +#define INTEL_GPU 1 +#define REQD_SUBGROUP_SIZE_16 __attribute__((intel_reqd_sub_group_size(16))) +#define REQD_SUBGROUP_SIZE_32 __attribute__((intel_reqd_sub_group_size(32))) +#elif defined(cl_qcom_reqd_sub_group_size) +// Always use subgroups size of 64 on Adreno. +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable +#define ADRENO_GPU 1 +#define REQD_SUBGROUP_SIZE_64 __attribute__((qcom_reqd_sub_group_size("half"))) +#define REQD_SUBGROUP_SIZE_128 __attribute__((qcom_reqd_sub_group_size("full"))) +#else +// TODO: do not know how to choose subgroup size on other GPUs. +#error "Selecting subgroup size is not supported on your device." +#endif + +#define QK4_0 32 +#define QR4_0 2 +#define QK4_1 32 +#define QR4_1 2 +#define QK5_0 32 +#define QR5_0 2 +#define QK5_1 32 +#define QR5_1 2 +#define QK8_0 32 +#define QR8_0 1 +#define QK_K 256 +#define K_QUANTS_PER_ITERATION 2 + +typedef char int8_t; +typedef uchar uint8_t; +typedef short int16_t; +typedef ushort uint16_t; +typedef int int32_t; +typedef uint uint32_t; + +//------------------------------------------------------------------------------ +// block_q4_0 +//------------------------------------------------------------------------------ +struct block_q4_0 +{ + half d; + uint8_t qs[QK4_0 / 2]; +}; + +//------------------------------------------------------------------------------ +// mul_vec_q_n_f32_flat_noshuffle +// +// This variation uses flat arrays (struct of arrays, SOA) representation for +// quant tensors. It also uses non shuffled bit order for weights. +// +// The shuffled version is kept in the original file because moving it here +// seems to result in worse performance for adreno. +//------------------------------------------------------------------------------ + +kernel void kernel_convert_block_q4_0_noshuffle( + global struct block_q4_0 * src0, + global uchar * dst_q, + global half * dst_d +) { + global struct block_q4_0 * b = (global struct block_q4_0 *) src0 + get_global_id(0); + global uchar * q = (global uchar *) dst_q + QK4_0/2*get_global_id(0); + global half * d = (global half *) dst_d + get_global_id(0); + + *d = b->d; + for (int i = 0; i < QK4_0/4; ++i) { + uchar x0 = b->qs[2*i + 0]; + uchar x1 = b->qs[2*i + 1]; + + q[i + 0 ] = convert_uchar(x0 & 0x0F) | convert_uchar((x1 & 0x0F) << 4); + q[i + QK4_0/4] = convert_uchar((x0 & 0xF0) >> 4) | convert_uchar(x1 & 0xF0); + +#ifdef ADRENO_GPU + // Workaround for adreno - must have the following printf statement for + // the kernel to work properly. Otherwise it produces incorrect result. + // convert_uchar above also seems necessary. + // Compare against a large number so that it does not print anything. + // get_sub_group_local_id() also works. + if (get_global_id(0) == 65536*4096) { + printf("%04x - %02x\n", *(global ushort*)d, ((x0 & 0xF0) >> 4) | (x1 & 0xF0)); + } +#endif + } +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle.cl new file mode 100644 index 000000000..5e195411d --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle.cl @@ -0,0 +1,265 @@ +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#pragma OPENCL EXTENSION cl_khr_subgroups : enable +#pragma OPENCL EXTENSION cl_qcom_subgroup_uniform_load: enable +#pragma OPENCL EXTENSION cl_qcom_subgroup_constant_load: enable +#pragma OPENCL EXTENSION cl_qcom_extra_vector_types : enable +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable + +// assume +#define QK4_0 32 +#define N_SIMDGROUP 4 + +#define dequantizeBlockAccum_ns_sgbroadcast_1_hi(total_sums, bits4, scale, y) \ + float shared_y; \ + shared_y = sub_group_broadcast(y.s0, 0); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 0); \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 0); \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 0); \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 0); \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 0); \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 0); \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 0); \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s0, 1); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 1); \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 1); \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 1); \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 1); \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 1); \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 1); \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 1); \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_1_lo(total_sums, bits4, scale, y) \ + shared_y = sub_group_broadcast(y.s0, 2); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 2); \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 2); \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 2); \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 2); \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 2); \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 2); \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 2); \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s0, 3); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 3); \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 3); \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 3); \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 3); \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 3); \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 3); \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 3); \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_8_hi(total_sums, bits4, scale, y) \ + float8 shared_y; \ + shared_y = sub_group_broadcast(y, 0); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + shared_y = sub_group_broadcast(y, 1); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_8_lo(total_sums, bits4, scale, y) \ + shared_y = sub_group_broadcast(y, 2); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + shared_y = sub_group_broadcast(y, 3); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + + +__attribute__((qcom_reqd_sub_group_size("full"))) +__kernel void kernel_gemv_noshuffle( + __read_only image1d_buffer_t src0_q, // quantized A + global half2 * src0_d, // A scales + __read_only image1d_buffer_t src1, // B + ulong offset1, // offset to B (0) + global float * dst, // C + ulong offsetd, // offset to C (0) + uint K, // K + int ne01, // M + int ne02, // 1 + int ne10, // K + int ne12, // 1 + int ne0, // M + int ne1, // N + int r2, // 1 + int r3) +{ + uint groupId = get_local_id(1); + uint gid = get_global_id(0); + ushort slid = get_sub_group_local_id(); + + __private uint4 regA; + __private half2 regS; + __private float8 regB; + + __private float2 totalSum = (float2)(0.0f); + + // loop along K in block granularity, skip 4 blocks every iter + for (uint k = groupId; k < (K / QK4_0); k += N_SIMDGROUP) { + regS = src0_d[gid + k * LINE_STRIDE_A]; // each fiber loads scale of two rows + // first 4 fibers in each wave load 8 B values to its private scope + if (slid < 4) { + regB.s0123 = read_imagef(src1, (slid * 2 + k * 8)); + regB.s4567 = read_imagef(src1, (1 + slid * 2 + k * 8)); + } + + // load half weights for two blocks in consecutive rows + regA.s0 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 0)).x; + regA.s1 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 1)).x; + regA.s2 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 2)).x; + regA.s3 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 3)).x; +#ifdef VECTOR_SUB_GROUP_BROADCAT + dequantizeBlockAccum_ns_sgbroadcast_8_hi(totalSum, as_ushort8(regA), regS, regB); +#else + dequantizeBlockAccum_ns_sgbroadcast_1_hi(totalSum, as_ushort8(regA), regS, regB); +#endif // VECTOR_SUB_GROUP_BROADCAT + + regA.s0 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 4)).x; + regA.s1 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 5)).x; + regA.s2 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 6)).x; + regA.s3 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 7)).x; +#ifdef VECTOR_SUB_GROUP_BROADCAT + dequantizeBlockAccum_ns_sgbroadcast_8_lo(totalSum, as_ushort8(regA), regS, regB); +#else + dequantizeBlockAccum_ns_sgbroadcast_1_lo(totalSum, as_ushort8(regA), regS, regB); +#endif // VECTOR_SUB_GROUP_BROADCAT + } + + // reduction in local memory, assumes #wave=4 + __local float2 reduceLM[SIMDGROUP_WIDTH * 3]; + if (groupId == 1) reduceLM[SIMDGROUP_WIDTH * 0 + slid] = totalSum; + if (groupId == 2) reduceLM[SIMDGROUP_WIDTH * 1 + slid] = totalSum; + if (groupId == 3) reduceLM[SIMDGROUP_WIDTH * 2 + slid] = totalSum; + barrier(CLK_LOCAL_MEM_FENCE); + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 0 + slid]; + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 1 + slid]; + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 2 + slid]; + + // 2 outputs per fiber in wave 0 + if (groupId == 0) { + dst = (global float*)((global char*)dst + offsetd); + vstore2(totalSum, 0, &(dst[gid * 2])); + } + +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle_general.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle_general.cl new file mode 100644 index 000000000..5bdd4d067 --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_gemv_noshuffle_general.cl @@ -0,0 +1,271 @@ +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#pragma OPENCL EXTENSION cl_khr_subgroups : enable +#pragma OPENCL EXTENSION cl_qcom_subgroup_uniform_load: enable +#pragma OPENCL EXTENSION cl_qcom_subgroup_constant_load: enable +#pragma OPENCL EXTENSION cl_qcom_extra_vector_types : enable +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable + +// assume +#define QK4_0 32 +#define N_SIMDGROUP 4 + +#define dequantizeBlockAccum_ns_sgbroadcast_1_hi(total_sums, bits4, scale, y) \ + float shared_y; \ + shared_y = sub_group_broadcast(y.s0, 0); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 0); \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 0); \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 0); \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 0); \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 0); \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 0); \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 0); \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s0, 1); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 1); \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 1); \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 1); \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 1); \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 1); \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 1); \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 1); \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_1_lo(total_sums, bits4, scale, y) \ + shared_y = sub_group_broadcast(y.s0, 2); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 2); \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 2); \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 2); \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 2); \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 2); \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 2); \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 2); \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s0, 3); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s1, 3); \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s2, 3); \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s3, 3); \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s4, 3); \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s5, 3); \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s6, 3); \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y; \ + shared_y = sub_group_broadcast(y.s7, 3); \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_8_hi(total_sums, bits4, scale, y) \ + float8 shared_y; \ + shared_y = sub_group_broadcast(y, 0); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + shared_y = sub_group_broadcast(y, 1); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + + +#define dequantizeBlockAccum_ns_sgbroadcast_8_lo(total_sums, bits4, scale, y) \ + shared_y = sub_group_broadcast(y, 2); \ + total_sums.s0 += ((bits4.s0 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s0 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s0 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s0 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s2 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s2 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s2 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s2 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s1 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s1 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s1 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s1 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s3 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s3 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s3 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s3 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + shared_y = sub_group_broadcast(y, 3); \ + total_sums.s0 += ((bits4.s4 & 0x000F) - 8) * scale.s0 * shared_y.s0; \ + total_sums.s0 += (((bits4.s4 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s1; \ + total_sums.s0 += (((bits4.s4 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s2; \ + total_sums.s0 += (((bits4.s4 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s3; \ + total_sums.s0 += ((bits4.s6 & 0x000F) - 8) * scale.s0 * shared_y.s4; \ + total_sums.s0 += (((bits4.s6 & 0x00F0) >> 4) - 8) * scale.s0 * shared_y.s5; \ + total_sums.s0 += (((bits4.s6 & 0x0F00) >> 8) - 8) * scale.s0 * shared_y.s6; \ + total_sums.s0 += (((bits4.s6 & 0xF000) >> 12) - 8) * scale.s0 * shared_y.s7; \ + total_sums.s1 += ((bits4.s5 & 0x000F) - 8) * scale.s1 * shared_y.s0; \ + total_sums.s1 += (((bits4.s5 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s1; \ + total_sums.s1 += (((bits4.s5 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s2; \ + total_sums.s1 += (((bits4.s5 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s3; \ + total_sums.s1 += ((bits4.s7 & 0x000F) - 8) * scale.s1 * shared_y.s4; \ + total_sums.s1 += (((bits4.s7 & 0x00F0) >> 4) - 8) * scale.s1 * shared_y.s5; \ + total_sums.s1 += (((bits4.s7 & 0x0F00) >> 8) - 8) * scale.s1 * shared_y.s6; \ + total_sums.s1 += (((bits4.s7 & 0xF000) >> 12) - 8) * scale.s1 * shared_y.s7; \ + + +__attribute__((qcom_reqd_sub_group_size("full"))) +__kernel void kernel_gemv_noshuffle( + __read_only image1d_buffer_t src0_q, // quantized A + global half2 * src0_d, // A scales + __read_only image1d_buffer_t src1, // B + ulong offset1, // offset to B (0) + global float * dst, // C + ulong offsetd, // offset to C (0) + int ne00, // K + int ne01, // M + int ne02, // 1 + int ne10, // K + int ne12, // 1 + int ne0, // M + int ne1, // N + int r2, // 1 + int r3) +{ + uint groupId = get_local_id(1); + uint gid = get_global_id(0); + ushort slid = get_sub_group_local_id(); + + uint K = ne00; + uint M = ne01; + + uint LINE_STRIDE_A = M / 2; + uint BLOCK_STRIDE_A = N_SIMDGROUP * M; + + __private uint4 regA; + __private half2 regS; + __private float8 regB; + + __private float2 totalSum = (float2)(0.0f); + + // loop along K in block granularity, skip 4 blocks every iter + for (uint k = groupId; k < (K / QK4_0); k += N_SIMDGROUP) { + regS = src0_d[gid + k * LINE_STRIDE_A]; // each fiber loads scale of two rows + // first 4 fibers in each wave load 8 B values to its private scope + if (slid < 4) { + regB.s0123 = read_imagef(src1, (slid * 2 + k * 8)); + regB.s4567 = read_imagef(src1, (1 + slid * 2 + k * 8)); + } + + // load half weights for two blocks in consecutive rows + regA.s0 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 0)).x; + regA.s1 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 1)).x; + regA.s2 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 2)).x; + regA.s3 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 3)).x; +#ifdef VECTOR_SUB_GROUP_BROADCAT + dequantizeBlockAccum_ns_sgbroadcast_8_hi(totalSum, as_ushort8(regA), regS, regB); +#else + dequantizeBlockAccum_ns_sgbroadcast_1_hi(totalSum, as_ushort8(regA), regS, regB); +#endif // VECTOR_SUB_GROUP_BROADCAT + + regA.s0 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 4)).x; + regA.s1 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 5)).x; + regA.s2 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 6)).x; + regA.s3 = read_imageui(src0_q, (gid + k * BLOCK_STRIDE_A + LINE_STRIDE_A * 7)).x; +#ifdef VECTOR_SUB_GROUP_BROADCAT + dequantizeBlockAccum_ns_sgbroadcast_8_lo(totalSum, as_ushort8(regA), regS, regB); +#else + dequantizeBlockAccum_ns_sgbroadcast_1_lo(totalSum, as_ushort8(regA), regS, regB); +#endif // VECTOR_SUB_GROUP_BROADCAT + } + + // reduction in local memory, assumes #wave=4 + __local float2 reduceLM[SIMDGROUP_WIDTH * 3]; + if (groupId == 1) reduceLM[SIMDGROUP_WIDTH * 0 + slid] = totalSum; + if (groupId == 2) reduceLM[SIMDGROUP_WIDTH * 1 + slid] = totalSum; + if (groupId == 3) reduceLM[SIMDGROUP_WIDTH * 2 + slid] = totalSum; + barrier(CLK_LOCAL_MEM_FENCE); + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 0 + slid]; + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 1 + slid]; + if (groupId == 0) totalSum += reduceLM[SIMDGROUP_WIDTH * 2 + slid]; + + // 2 outputs per fiber in wave 0 + if (groupId == 0) { + dst = (global float*)((global char*)dst + offsetd); + vstore2(totalSum, 0, &(dst[gid * 2])); + } + +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_mm.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_mm.cl new file mode 100644 index 000000000..e19e9a2f4 --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_mm.cl @@ -0,0 +1,1225 @@ +//------------------------------------------------------------------------------ +// This file is contains additional mulmat kernels +// (and potentially other kernels). +//------------------------------------------------------------------------------ +#ifdef cl_khr_fp16 +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#elif defined(cl_amd_fp16) +#pragma OPENCL EXTENSION cl_amd_fp16 : enable +#else +#error "Half precision floating point not supportedby OpenCL implementation on your device." +#endif + +#ifdef cl_khr_subgroups +#pragma OPENCL EXTENSION cl_khr_subgroups : enable +#elif defined(cl_intel_subgroups) +#pragma OPENCL EXTENSION cl_intel_subgroups : enable +#else +#error "Subgroup not supported on your device." +#endif + +#ifdef cl_intel_required_subgroup_size +// Always use subgroup size of 32 on Intel. +#pragma OPENCL EXTENSION cl_intel_required_subgroup_size : enable +#define INTEL_GPU 1 +#define REQD_SUBGROUP_SIZE_16 __attribute__((intel_reqd_sub_group_size(16))) +#define REQD_SUBGROUP_SIZE_32 __attribute__((intel_reqd_sub_group_size(32))) +#elif defined(cl_qcom_reqd_sub_group_size) +// Always use subgroups size of 64 on Adreno. +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable +#define ADRENO_GPU 1 +#define REQD_SUBGROUP_SIZE_64 __attribute__((qcom_reqd_sub_group_size("half"))) +#define REQD_SUBGROUP_SIZE_128 __attribute__((qcom_reqd_sub_group_size("full"))) +#else +// TODO: do not know how to choose subgroup size on other GPUs. +#error "Selecting subgroup size is not supported on your device." +#endif + +#define QK4_0 32 +#define QR4_0 2 +#define QK4_1 32 +#define QR4_1 2 +#define QK5_0 32 +#define QR5_0 2 +#define QK5_1 32 +#define QR5_1 2 +#define QK8_0 32 +#define QR8_0 1 +#define QK_K 256 +#define K_QUANTS_PER_ITERATION 2 + +typedef char int8_t; +typedef uchar uint8_t; +typedef short int16_t; +typedef ushort uint16_t; +typedef int int32_t; +typedef uint uint32_t; + +//------------------------------------------------------------------------------ +// block_q4_0 +//------------------------------------------------------------------------------ +struct block_q4_0 +{ + half d; + uint8_t qs[QK4_0 / 2]; +}; + +//------------------------------------------------------------------------------ +// block_q6_K +//------------------------------------------------------------------------------ +// 6-bit quantization +// weight is represented as x = a * q +// 16 blocks of 16 elements each +// Effectively 6.5625 bits per weight +typedef struct { + uint8_t ql[QK_K/2]; // quants, lower 4 bits + uint8_t qh[QK_K/4]; // quants, upper 2 bits + int8_t scales[QK_K/16]; // scales, quantized with 8 bits + half d; // super-block scale +} block_q6_K; + +//------------------------------------------------------------------------------ +// These are the variant for matmatmul, based on the matvecmul kernel with +// flattened block_q4_0. +//------------------------------------------------------------------------------ + +// Common dot prod. +inline float mm_block_q_4_0_dot_y_flat( + global uchar * x, + global half * dh, + float sumy, + float16 yl, + int il +) { + float d = *dh; + global ushort * qs = ((global ushort *)x + il/2); + float acc = 0.f; + + acc += yl.s0 * (qs[0] & 0x000F); + acc += yl.s1 * (qs[0] & 0x0F00); + acc += yl.s8 * (qs[0] & 0x00F0); + acc += yl.s9 * (qs[0] & 0xF000); + + acc += yl.s2 * (qs[1] & 0x000F); + acc += yl.s3 * (qs[1] & 0x0F00); + acc += yl.sa * (qs[1] & 0x00F0); + acc += yl.sb * (qs[1] & 0xF000); + + acc += yl.s4 * (qs[2] & 0x000F); + acc += yl.s5 * (qs[2] & 0x0F00); + acc += yl.sc * (qs[2] & 0x00F0); + acc += yl.sd * (qs[2] & 0xF000); + + acc += yl.s6 * (qs[3] & 0x000F); + acc += yl.s7 * (qs[3] & 0x0F00); + acc += yl.se * (qs[3] & 0x00F0); + acc += yl.sf * (qs[3] & 0xF000); + + return d * (sumy * -8.f + acc); +} + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 8 // each SIMD group works on 8 rows (in weights matrix) +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 16 +#elif defined (ADRENO_GPU) +#define N_DST 8 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif +// +// This variant performs 1d blocking with 8x output. +// Eeach simdgroup outputs 8 values on `n0` dim (row in the output matrix). +// +inline void mul_mat_q_n_f32_1d_8x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + const int nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is the linear global id of + // a SIMD group in the grid. Each SIMD group produces N_DST values in the + // result, hence uses nb blocks, i.e., the offset becomes first_row*nb. + // Currently with llama2 7B, im is always 0. + // TODO: how to handle im/gqa*(nb*ne0)? + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = (first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02)) * QK4_0/2; + + global uchar * x = (global uchar *) src0_q + offset0_q; + global half * d = (global half *) src0_d + offset0_d; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float16 yl; + float8 sumf = (float8)(0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f); + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix*QK4_0 + il; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0.f; + + sumy += yb[0]; + sumy += yb[1]; + sumy += yb[2]; + sumy += yb[3]; + sumy += yb[4]; + sumy += yb[5]; + sumy += yb[6]; + sumy += yb[7]; + + sumy += yb[16]; + sumy += yb[17]; + sumy += yb[18]; + sumy += yb[19]; + sumy += yb[20]; + sumy += yb[21]; + sumy += yb[22]; + sumy += yb[23]; + + yl.s0 = yb[0]; + yl.s1 = yb[1]/256.f; + + yl.s2 = yb[2]; + yl.s3 = yb[3]/256.f; + + yl.s4 = yb[4]; + yl.s5 = yb[5]/256.f; + + yl.s6 = yb[6]; + yl.s7 = yb[7]/256.f; + + yl.s8 = yb[16]/16.f; + yl.s9 = yb[17]/4096.f; + + yl.sa = yb[18]/16.f; + yl.sb = yb[19]/4096.f; + + yl.sc = yb[20]/16.f; + yl.sd = yb[21]/4096.f; + + yl.se = yb[22]/16.f; + yl.sf = yb[23]/4096.f; + + sumf.s0 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 0*nb*QK4_0/2, d + ib + 0*nb, sumy, yl, il); + sumf.s1 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 1*nb*QK4_0/2, d + ib + 1*nb, sumy, yl, il); + sumf.s2 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 2*nb*QK4_0/2, d + ib + 2*nb, sumy, yl, il); + sumf.s3 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 3*nb*QK4_0/2, d + ib + 3*nb, sumy, yl, il); + + sumf.s4 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 4*nb*QK4_0/2, d + ib + 4*nb, sumy, yl, il); + sumf.s5 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 5*nb*QK4_0/2, d + ib + 5*nb, sumy, yl, il); + sumf.s6 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 6*nb*QK4_0/2, d + ib + 6*nb, sumy, yl, il); + sumf.s7 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 7*nb*QK4_0/2, d + ib + 7*nb, sumy, yl, il); + + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + float8 tot = (float8)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1), + sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3), + sub_group_reduce_add(sumf.s4), sub_group_reduce_add(sumf.s5), + sub_group_reduce_add(sumf.s6), sub_group_reduce_add(sumf.s7) + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } + + if (first_row + 4 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 4] = tot.s4; + } + if (first_row + 5 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 5] = tot.s5; + } + if (first_row + 6 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 6] = tot.s6; + } + if (first_row + 7 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 7] = tot.s7; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_1d_8x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_mat_q_n_f32_1d_8x_flat(src0_q, src0_d, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 16 // each SIMD group works on 8 rows (in weights matrix) +#define N_SIMDGROUP 1 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // assuming SIMD group size is 16 +#elif defined (ADRENO_GPU) +#define N_DST 16 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif +// +// This variant performs 1d blocking with 16x output. +// Eeach simdgroup outputs 16 values on `n0` dim (row in the output matrix). +// +inline void mul_mat_q_n_f32_1d_16x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + global float * dst, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + const int nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + // (r0 * N_SIMDGROUP + get_sub_group_id()) is the linear global id of + // a SIMD group in the grid. Each SIMD group produces N_DST values in the + // result, hence uses nb blocks, i.e., the offset becomes first_row*nb. + // Currently with llama2 7B, im is always 0. + // TODO: how to handle im/gqa*(nb*ne0)? + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = (first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02)) * QK4_0/2; + + global uchar * x = (global uchar *) src0_q + offset0_q; + global half * d = (global half *) src0_d + offset0_d; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float16 yl; + float16 sumf = (float16)(0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, + 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f); + + int ix = get_sub_group_local_id()/2; + int il = 8*(get_sub_group_local_id()%2); + + global float * yb = y + ix*QK4_0 + il; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/2) { + float sumy = 0.f; + + sumy += yb[0]; + sumy += yb[1]; + sumy += yb[2]; + sumy += yb[3]; + sumy += yb[4]; + sumy += yb[5]; + sumy += yb[6]; + sumy += yb[7]; + + sumy += yb[16]; + sumy += yb[17]; + sumy += yb[18]; + sumy += yb[19]; + sumy += yb[20]; + sumy += yb[21]; + sumy += yb[22]; + sumy += yb[23]; + + yl.s0 = yb[0]; + yl.s1 = yb[1]/256.f; + + yl.s2 = yb[2]; + yl.s3 = yb[3]/256.f; + + yl.s4 = yb[4]; + yl.s5 = yb[5]/256.f; + + yl.s6 = yb[6]; + yl.s7 = yb[7]/256.f; + + yl.s8 = yb[16]/16.f; + yl.s9 = yb[17]/4096.f; + + yl.sa = yb[18]/16.f; + yl.sb = yb[19]/4096.f; + + yl.sc = yb[20]/16.f; + yl.sd = yb[21]/4096.f; + + yl.se = yb[22]/16.f; + yl.sf = yb[23]/4096.f; + + sumf.s0 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 0*nb*QK4_0/2, d + ib + 0*nb, sumy, yl, il); + sumf.s1 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 1*nb*QK4_0/2, d + ib + 1*nb, sumy, yl, il); + sumf.s2 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 2*nb*QK4_0/2, d + ib + 2*nb, sumy, yl, il); + sumf.s3 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 3*nb*QK4_0/2, d + ib + 3*nb, sumy, yl, il); + + sumf.s4 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 4*nb*QK4_0/2, d + ib + 4*nb, sumy, yl, il); + sumf.s5 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 5*nb*QK4_0/2, d + ib + 5*nb, sumy, yl, il); + sumf.s6 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 6*nb*QK4_0/2, d + ib + 6*nb, sumy, yl, il); + sumf.s7 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 7*nb*QK4_0/2, d + ib + 7*nb, sumy, yl, il); + + sumf.s8 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 8*nb*QK4_0/2, d + ib + 8*nb, sumy, yl, il); + sumf.s9 += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 9*nb*QK4_0/2, d + ib + 9*nb, sumy, yl, il); + sumf.sa += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 10*nb*QK4_0/2, d + ib + 10*nb, sumy, yl, il); + sumf.sb += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 11*nb*QK4_0/2, d + ib + 11*nb, sumy, yl, il); + + sumf.sc += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 12*nb*QK4_0/2, d + ib + 12*nb, sumy, yl, il); + sumf.sd += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 13*nb*QK4_0/2, d + ib + 13*nb, sumy, yl, il); + sumf.se += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 14*nb*QK4_0/2, d + ib + 14*nb, sumy, yl, il); + sumf.sf += mm_block_q_4_0_dot_y_flat(x + ib*QK4_0/2 + 15*nb*QK4_0/2, d + ib + 15*nb, sumy, yl, il); + + yb += QK4_0 * (N_SIMDWIDTH/2); + } + + float16 tot = (float16)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1), + sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3), + sub_group_reduce_add(sumf.s4), sub_group_reduce_add(sumf.s5), + sub_group_reduce_add(sumf.s6), sub_group_reduce_add(sumf.s7), + + sub_group_reduce_add(sumf.s8), sub_group_reduce_add(sumf.s9), + sub_group_reduce_add(sumf.sa), sub_group_reduce_add(sumf.sb), + sub_group_reduce_add(sumf.sc), sub_group_reduce_add(sumf.sd), + sub_group_reduce_add(sumf.se), sub_group_reduce_add(sumf.sf) + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } + + if (first_row + 4 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 4] = tot.s4; + } + if (first_row + 5 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 5] = tot.s5; + } + if (first_row + 6 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 6] = tot.s6; + } + if (first_row + 7 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 7] = tot.s7; + } + + if (first_row + 8 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 8] = tot.s8; + } + if (first_row + 9 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 9] = tot.s9; + } + if (first_row + 10 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 10] = tot.sa; + } + if (first_row + 11 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 11] = tot.sb; + } + + if (first_row + 12 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 12] = tot.sc; + } + if (first_row + 13 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 13] = tot.sd; + } + if (first_row + 14 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 14] = tot.se; + } + if (first_row + 15 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 15] = tot.sf; + } + } +} + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_1d_16x_flat( + global uchar * src0_q, + global half * src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + mul_mat_q_n_f32_1d_16x_flat(src0_q, src0_d, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3); +} + +//------------------------------------------------------------------------------ +// kernel_mul_mat_q4_0_f32_flat_v0 +//------------------------------------------------------------------------------ +inline float block_q_4_0_dot_y_flat_v2( + half x, + half d, + float sumy, + float4 yl +) { + uchar2 q = as_uchar2(x); + float acc = 0.0f; + + acc += (q.s0 & 0x0F) * yl.s0; + acc += (q.s1 & 0x0F) * yl.s1; + + acc += (q.s0 & 0xF0) * yl.s2; + acc += (q.s1 & 0xF0) * yl.s3; + + return d * (sumy * -8.f + acc);; +} + +inline float block_q_4_0_dot_y_flat_v4( + float x, + half d, + float sumy, + float8 yl +) { + uchar4 q = as_uchar4(x); + float acc = 0.0f; + + acc += (q.s0 & 0x0F) * yl.s0; + acc += (q.s1 & 0x0F) * yl.s1; + acc += (q.s2 & 0x0F) * yl.s2; + acc += (q.s3 & 0x0F) * yl.s3; + + acc += (q.s0 & 0xF0) * yl.s4; + acc += (q.s1 & 0xF0) * yl.s5; + acc += (q.s2 & 0xF0) * yl.s6; + acc += (q.s3 & 0xF0) * yl.s7; + + return d * (sumy * -8.f + acc);; +} + +inline float block_q_4_0_dot_y_flat_v8( + float2 x, + half d, + float sumy, + float16 yl +) { + uchar8 q = as_uchar8(x); + float acc = 0.0f; + + acc += (q.s0 & 0x0F) * yl.s0; + acc += (q.s1 & 0x0F) * yl.s1; + acc += (q.s2 & 0x0F) * yl.s2; + acc += (q.s3 & 0x0F) * yl.s3; + acc += (q.s4 & 0x0F) * yl.s4; + acc += (q.s5 & 0x0F) * yl.s5; + acc += (q.s6 & 0x0F) * yl.s6; + acc += (q.s7 & 0x0F) * yl.s7; + + acc += (q.s0 & 0xF0) * yl.s8; + acc += (q.s1 & 0xF0) * yl.s9; + acc += (q.s2 & 0xF0) * yl.sa; + acc += (q.s3 & 0xF0) * yl.sb; + acc += (q.s4 & 0xF0) * yl.sc; + acc += (q.s5 & 0xF0) * yl.sd; + acc += (q.s6 & 0xF0) * yl.se; + acc += (q.s7 & 0xF0) * yl.sf; + + return d * (sumy * -8.f + acc);; +} + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define THREADS_PER_BLK 4 // Number of threads per block, or each thread process 1/THREADS_PER_BLK of a block +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 16 +#elif defined (ADRENO_GPU) +#define THREADS_PER_BLK 4 +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +#if THREADS_PER_BLK == 2 // Each thread processes 1/2 block +# define ACT_TY float16 +# define Q_BLK_LD_TY float2 +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v8 +#elif THREADS_PER_BLK == 4 // Each thread processes 1/4 block +# define ACT_TY float8 +# define Q_BLK_LD_TY float +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v4 +#elif THREADS_PER_BLK == 8 // Each thread processes 1/8 block +# define ACT_TY float4 +# define Q_BLK_LD_TY half +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v2 +#endif + +#define BTYES_PER_THREAD_IN_BLK (QK4_0/2/THREADS_PER_BLK) + +#if N_DST == 2 +# define SUM_TY float2 +#elif N_DST == 4 +# define SUM_TY float4 +#elif N_DST == 8 +# define SUM_TY float8 +#elif N_DST == 16 +# define SUM_TY float16 +#endif + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_flat_v0( + global uchar * src0_q, + global half * src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + const int nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = (first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02)) * QK4_0/2; + + global uchar * x = (global uchar *) src0_q + offset0_q; + global half * d = (global half *) src0_d + offset0_d; + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + int ix = get_sub_group_local_id()/THREADS_PER_BLK; + int il = get_sub_group_local_id()%THREADS_PER_BLK; + + global float * yb = y + ix*QK4_0 + BTYES_PER_THREAD_IN_BLK*il; + + // Registers for caching activation + ACT_TY yl = 0.f; + + // Registers for caching quants + Q_BLK_LD_TY q_blk_0 = 0, q_blk_1 = 0; +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + Q_BLK_LD_TY q_blk_2 = 0, q_blk_3 = 0; +#endif +#if N_DST == 8 || N_DST == 16 + Q_BLK_LD_TY q_blk_4 = 0, q_blk_5 = 0, q_blk_6 = 0, q_blk_7 = 0; +#endif + + // Partial sum + SUM_TY sumf = 0.f; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/THREADS_PER_BLK) { + float sumy = 0.f; + + q_blk_0 = *(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 0*nb*QK4_0/2); + q_blk_1 = *(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 1*nb*QK4_0/2); +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + q_blk_2 = *(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 2*nb*QK4_0/2); + q_blk_3 = *(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 3*nb*QK4_0/2); +#endif +#if N_DST == 8 || N_DST == 16 + q_blk_4 = (*(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 4*nb*QK4_0/2)); + q_blk_5 = (*(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 5*nb*QK4_0/2)); + q_blk_6 = (*(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 6*nb*QK4_0/2)); + q_blk_7 = (*(global Q_BLK_LD_TY*)(x + ib*QK4_0/2 + BTYES_PER_THREAD_IN_BLK*il + 7*nb*QK4_0/2)); +#endif + + // Load activation +#if THREADS_PER_BLK == 2 // Each thread processes 1/2 block + yl.s01234567 = *(global float8 *)(yb); + yl.s89abcdef = *(global float8 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; + sumy += yl.s3; + sumy += yl.s4; + sumy += yl.s5; + sumy += yl.s6; + sumy += yl.s7; + sumy += yl.s8; yl.s8 /= 16.f; + sumy += yl.s9; yl.s9 /= 16.f; + sumy += yl.sa; yl.sa /= 16.f; + sumy += yl.sb; yl.sb /= 16.f; + sumy += yl.sc; yl.sc /= 16.f; + sumy += yl.sd; yl.sd /= 16.f; + sumy += yl.se; yl.se /= 16.f; + sumy += yl.sf; yl.sf /= 16.f; +#elif THREADS_PER_BLK == 4 // Each thread processes 1/4 block + yl.s0123 = *(global float4 *)(yb); + yl.s4567 = *(global float4 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; + sumy += yl.s3; + sumy += yl.s4; yl.s4 /= 16.f; + sumy += yl.s5; yl.s5 /= 16.f; + sumy += yl.s6; yl.s6 /= 16.f; + sumy += yl.s7; yl.s7 /= 16.f; +#elif THREADS_PER_BLK == 8 // Each thread processes 1/8 block + yl.s01 = *(global float2 *)(yb); + yl.s23 = *(global float2 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; yl.s2 /= 16.f; + sumy += yl.s3; yl.s3 /= 16.f; +#endif + + sumf.s0 += block_q_4_0_dot_y_flat(q_blk_0, *(d + ib + 0*nb), sumy, yl); + sumf.s1 += block_q_4_0_dot_y_flat(q_blk_1, *(d + ib + 1*nb), sumy, yl); +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + sumf.s2 += block_q_4_0_dot_y_flat(q_blk_2, *(d + ib + 2*nb), sumy, yl); + sumf.s3 += block_q_4_0_dot_y_flat(q_blk_3, *(d + ib + 3*nb), sumy, yl); +#endif +#if N_DST == 8 || N_DST == 16 + sumf.s4 += block_q_4_0_dot_y_flat(q_blk_4, *(d + ib + 4*nb), sumy, yl); + sumf.s5 += block_q_4_0_dot_y_flat(q_blk_5, *(d + ib + 5*nb), sumy, yl); + sumf.s6 += block_q_4_0_dot_y_flat(q_blk_6, *(d + ib + 6*nb), sumy, yl); + sumf.s7 += block_q_4_0_dot_y_flat(q_blk_7, *(d + ib + 7*nb), sumy, yl); +#endif + + yb += QK4_0 * (N_SIMDWIDTH/THREADS_PER_BLK); + } + + SUM_TY tot = (SUM_TY)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1) +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + , sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3) +#endif +#if N_DST == 8 || N_DST == 16 + , sub_group_reduce_add(sumf.s4), sub_group_reduce_add(sumf.s5) + , sub_group_reduce_add(sumf.s6), sub_group_reduce_add(sumf.s7) +#endif + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } +#endif +#if N_DST == 8 || N_DST == 16 + if (first_row + 4 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 4] = tot.s4; + } + if (first_row + 5 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 5] = tot.s5; + } + if (first_row + 6 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 6] = tot.s6; + } + if (first_row + 7 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 7] = tot.s7; + } +#endif + } +} + +//------------------------------------------------------------------------------ +// Using image1d_buffer_t + +#if defined(cl_qcom_subgroup_shuffle) +#pragma OPENCL EXTENSION cl_qcom_subgroup_shuffle : enable +float qcom_sub_group_reduce_add(float sum) { + sum += qcom_sub_group_shuffle_down(sum, 32, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + sum += qcom_sub_group_shuffle_down(sum, 16, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + sum += qcom_sub_group_shuffle_down(sum, 8, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + sum += qcom_sub_group_shuffle_down(sum, 4, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + sum += qcom_sub_group_shuffle_down(sum, 2, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + sum += qcom_sub_group_shuffle_down(sum, 1, CLK_SUB_GROUP_SHUFFLE_WIDTH_WAVE_SIZE_QCOM, 0.f); + return sum; +} +#define sub_group_reduce_add qcom_sub_group_reduce_add +#else +#define sub_group_reduce_add sub_group_reduce_add +#endif + +#undef THREADS_PER_BLK +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define THREADS_PER_BLK 4 // Number of threads per block, or each thread process 1/THREADS_PER_BLK of a block +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 16 +#elif defined (ADRENO_GPU) +#define THREADS_PER_BLK 4 +#define N_DST 4 +#define N_SIMDGROUP 1 +#define N_SIMDWIDTH 64 +#endif + +#if THREADS_PER_BLK == 2 // Each thread processes 1/2 block +# define ACT_TY float16 +# define Q_BLK_LD_TY float2 +# define EXTRACT_BLK_DATA(tmp, part) *((float2*)&tmp + part) +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v8 +#elif THREADS_PER_BLK == 4 // Each thread processes 1/4 block +# define ACT_TY float8 +# define Q_BLK_LD_TY float +# define EXTRACT_BLK_DATA(tmp, part) *((float*)&tmp + part) +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v4 +#elif THREADS_PER_BLK == 8 // Each thread processes 1/8 block +# define ACT_TY float4 +# define Q_BLK_LD_TY half +# define EXTRACT_BLK_DATA(tmp, part) *((half*)&tmp + part) +# define block_q_4_0_dot_y_flat block_q_4_0_dot_y_flat_v2 +#endif + +#define BTYES_PER_THREAD_IN_BLK (QK4_0/2/THREADS_PER_BLK) + +#if N_DST == 2 +# define SUM_TY float2 +#elif N_DST == 4 +# define SUM_TY float4 +#elif N_DST == 8 +# define SUM_TY float8 +#elif N_DST == 16 +# define SUM_TY float16 +#endif + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mat_q4_0_f32_flat_img_v0( + read_only image1d_buffer_t src0_q, + read_only image1d_buffer_t src0_d, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + const int nb = ne00/QK4_0; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + int first_row = (r0 * N_SIMDGROUP + get_sub_group_id()) * N_DST; + + int i12 = im%ne12; + int i13 = im/ne12; + + // The number of scales is the same as the number of blocks. + ulong offset0_d = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + // Each block contains QK4_0/2 uchars, hence offset for qs is as follows. + ulong offset0_q = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + + global float * y = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + int ix = get_sub_group_local_id()/THREADS_PER_BLK; + int il = get_sub_group_local_id()%THREADS_PER_BLK; + + global float * yb = y + ix*QK4_0 + BTYES_PER_THREAD_IN_BLK*il; + + // Registers for caching activation + ACT_TY yl = 0.f; + + // Registers for caching quants + Q_BLK_LD_TY q_blk_0 = 0, q_blk_1 = 0; +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + Q_BLK_LD_TY q_blk_2 = 0, q_blk_3 = 0; +#endif +#if N_DST == 8 || N_DST == 16 + Q_BLK_LD_TY q_blk_4 = 0, q_blk_5 = 0, q_blk_6 = 0, q_blk_7 = 0; +#endif + + // Partial sum + SUM_TY sumf = 0.f; + + for (int ib = ix; ib < nb; ib += N_SIMDWIDTH/THREADS_PER_BLK) { + float sumy = 0.f;; + + float4 tmp; + tmp = read_imagef(src0_q, offset0_q + ib + 0*nb); + q_blk_0 = EXTRACT_BLK_DATA(tmp, il); + tmp = read_imagef(src0_q, offset0_q + ib + 1*nb); + q_blk_1 = EXTRACT_BLK_DATA(tmp, il); +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + tmp = read_imagef(src0_q, offset0_q + ib + 2*nb); + q_blk_2 = EXTRACT_BLK_DATA(tmp, il); + tmp = read_imagef(src0_q, offset0_q + ib + 3*nb); + q_blk_3 = EXTRACT_BLK_DATA(tmp, il); +#endif +#if N_DST == 8 || N_DST == 16 + tmp = read_imagef(src0_q, offset0_q + ib + 4*nb); + q_blk_4 = EXTRACT_BLK_DATA(tmp, il); + tmp = read_imagef(src0_q, offset0_q + ib + 5*nb); + q_blk_5 = EXTRACT_BLK_DATA(tmp, il); + tmp = read_imagef(src0_q, offset0_q + ib + 6*nb); + q_blk_6 = EXTRACT_BLK_DATA(tmp, il); + tmp = read_imagef(src0_q, offset0_q + ib + 7*nb); + q_blk_7 = EXTRACT_BLK_DATA(tmp, il); +#endif + + // Load activation +#if THREADS_PER_BLK == 2 // Each thread processes 1/2 block + yl.s01234567 = *(global float8 *)(yb); + yl.s89abcdef = *(global float8 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; + sumy += yl.s3; + sumy += yl.s4; + sumy += yl.s5; + sumy += yl.s6; + sumy += yl.s7; + sumy += yl.s8; yl.s8 /= 16.f; + sumy += yl.s9; yl.s9 /= 16.f; + sumy += yl.sa; yl.sa /= 16.f; + sumy += yl.sb; yl.sb /= 16.f; + sumy += yl.sc; yl.sc /= 16.f; + sumy += yl.sd; yl.sd /= 16.f; + sumy += yl.se; yl.se /= 16.f; + sumy += yl.sf; yl.sf /= 16.f; +#elif THREADS_PER_BLK == 4 // Each thread processes 1/4 block + yl.s0123 = *(global float4 *)(yb); + yl.s4567 = *(global float4 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; + sumy += yl.s3; + sumy += yl.s4; yl.s4 /= 16.f; + sumy += yl.s5; yl.s5 /= 16.f; + sumy += yl.s6; yl.s6 /= 16.f; + sumy += yl.s7; yl.s7 /= 16.f; +#elif THREADS_PER_BLK == 8 // Each thread processes 1/8 block + yl.s01 = *(global float2 *)(yb); + yl.s23 = *(global float2 *)(yb + 16); + + sumy += yl.s0; + sumy += yl.s1; + sumy += yl.s2; yl.s2 /= 16.f; + sumy += yl.s3; yl.s3 /= 16.f; +#endif + + sumf.s0 += block_q_4_0_dot_y_flat(q_blk_0, read_imageh(src0_d, offset0_d + ib + 0*nb).s0, sumy, yl); + sumf.s1 += block_q_4_0_dot_y_flat(q_blk_1, read_imageh(src0_d, offset0_d + ib + 1*nb).s0, sumy, yl); +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + sumf.s2 += block_q_4_0_dot_y_flat(q_blk_2, read_imageh(src0_d, offset0_d + ib + 2*nb).s0, sumy, yl); + sumf.s3 += block_q_4_0_dot_y_flat(q_blk_3, read_imageh(src0_d, offset0_d + ib + 3*nb).s0, sumy, yl); +#endif +#if N_DST == 8 || N_DST == 16 + sumf.s4 += block_q_4_0_dot_y_flat(q_blk_4, read_imageh(src0_d, offset0_d + ib + 4*nb).s0, sumy, yl); + sumf.s5 += block_q_4_0_dot_y_flat(q_blk_5, read_imageh(src0_d, offset0_d + ib + 5*nb).s0, sumy, yl); + sumf.s6 += block_q_4_0_dot_y_flat(q_blk_6, read_imageh(src0_d, offset0_d + ib + 6*nb).s0, sumy, yl); + sumf.s7 += block_q_4_0_dot_y_flat(q_blk_7, read_imageh(src0_d, offset0_d + ib + 7*nb).s0, sumy, yl); +#endif + + yb += QK4_0 * (N_SIMDWIDTH/THREADS_PER_BLK); + } + + SUM_TY tot = (SUM_TY)( + sub_group_reduce_add(sumf.s0), sub_group_reduce_add(sumf.s1) +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + , sub_group_reduce_add(sumf.s2), sub_group_reduce_add(sumf.s3) +#endif +#if N_DST == 8 || N_DST == 16 + , sub_group_reduce_add(sumf.s4), sub_group_reduce_add(sumf.s5) + , sub_group_reduce_add(sumf.s6), sub_group_reduce_add(sumf.s7) +#endif + ); + + if (get_sub_group_local_id() == 0) { + if (first_row + 0 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 0] = tot.s0; + } + if (first_row + 1 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 1] = tot.s1; + } +#if N_DST == 4 || N_DST == 8 || N_DST == 16 + if (first_row + 2 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 2] = tot.s2; + } + if (first_row + 3 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 3] = tot.s3; + } +#endif +#if N_DST == 8 || N_DST == 16 + if (first_row + 4 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 4] = tot.s4; + } + if (first_row + 5 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 5] = tot.s5; + } + if (first_row + 6 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 6] = tot.s6; + } + if (first_row + 7 < ne01) { + dst[r1*ne0 + im*ne0*ne1 + first_row + 7] = tot.s7; + } +#endif + } +} + +//------------------------------------------------------------------------------ +// kernel_mul_mv_q6_K_f32 +//------------------------------------------------------------------------------ + +#undef N_DST +#undef N_SIMDGROUP +#undef N_SIMDWIDTH + +#ifdef INTEL_GPU +#define N_DST 1 // number of rows each SIMD group works on +#define N_SIMDGROUP 2 // number of SIMD groups in a thread group +#define N_SIMDWIDTH 16 // SIMD group size +#elif defined (ADRENO_GPU) +#define N_DST 1 +#define N_SIMDGROUP 2 +#define N_SIMDWIDTH 64 +#endif + +#define BLOCK_STRIDE (N_SIMDWIDTH/16) // number of blocks each subgroup processes + +#ifdef INTEL_GPU +REQD_SUBGROUP_SIZE_16 +#elif defined (ADRENO_GPU) +REQD_SUBGROUP_SIZE_64 +#endif +kernel void kernel_mul_mv_q6_K_f32( + global void * src0, + ulong offset0, + global float * src1, + ulong offset1, + global float * dst, + ulong offsetd, + int ne00, + int ne01, + int ne02, + int ne10, + int ne12, + int ne0, + int ne1, + int r2, + int r3 +) { + src0 = (global void*)((global char*)src0 + offset0); + src1 = (global float*)((global char*)src1 + offset1); + dst = (global float*)((global char*)dst + offsetd); + + uchar kmask1 = 0x03; + uchar kmask2 = 0x0C; + uchar kmask3 = 0x30; + uchar kmask4 = 0xC0; + + int nb = ne00/QK_K; + + int r0 = get_group_id(0); + int r1 = get_group_id(1); + int im = get_group_id(2); + + int row = N_SIMDGROUP * r0 + get_sub_group_id(); + + int i12 = im%ne12; + int i13 = im/ne12; + + ulong offset_src0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02); + + global block_q6_K * x = (global block_q6_K *) src0 + row*nb + offset_src0; + global float * yy = (global float *) src1 + r1*ne10 + im*ne00*ne1; + + float sumf = 0; + + // For Q6_K quantization, 16 values forms a subblock, 16 subblock forms a + // block. Values in a subblock shares a scale that is quantized with 8 bits; + // the entire block shares a single floating point scale. + // For work distribution, each thread processes a subblock (16 weights), hence + // 16 threads process a (super) block -- a subgroup thus handles SIMDWIDTH/16 + // (super) blocks -- this is the block stride. + // The 16 threads that process a (super) block are split into 2 portions, each has + // 8 threads; each portion works on 8 subblocks. + // For subgroup of 16 threads, the entire subgroup works on a single (super) block + // before moving to the next (super) block. Thread0 - thread7 work on the + // first 8 subblocks; thread8 - thread15 works on the last 8 subblocks. + // Thread0 - thread3 work on subblocks 0, 2, 4, 6; thread4 - thread7 work on + // subblocks 1, 3, 5, 7. Each thread does not work on an entire subblock, but + // works on a total of 16 weight values. + int tid = get_sub_group_local_id()/BLOCK_STRIDE; // first block_stride groups have tid=0 + int ix = get_sub_group_local_id()%BLOCK_STRIDE; // first block is 0..block_stride-1 + int ip = tid/8; // first or second half of (super) block (0 or 1) + int il = tid%8; // each half has 8 parts, one per scale + int n = 4; // 4 scales at a time (and 4 sums) + int l0 = n*il; // offset into half-block, 0..28 + int is = 8*ip + l0/16; // 0, 1, 8, 9 + + int y_offset = 128*ip + l0; + int q_offset_l = 64*ip + l0; + int q_offset_h = 32*ip + l0; + + for (int i = ix; i < nb; i += BLOCK_STRIDE) { + + global uint8_t * q1 = x[i].ql + q_offset_l; + global uint8_t * q2 = q1 + QK_K/8; + global uint8_t * qh = x[i].qh + q_offset_h; + global int8_t * sc = x[i].scales + is; + + global float * y = yy + i * QK_K + y_offset; + + float dall = x[i].d; + + float4 sums = {0.f, 0.f, 0.f, 0.f}; + + sums.s0 += y[0+ 0] * ((float)((q1[0] & 0xF) | ((qh[0] & kmask1) << 4)) - 32.f); + sums.s1 += y[0+32] * ((float)((q2[0] & 0xF) | ((qh[0] & kmask2) << 2)) - 32.f); + sums.s2 += y[0+64] * ((float)((q1[0] >> 4) | ((qh[0] & kmask3) << 0)) - 32.f); + sums.s3 += y[0+96] * ((float)((q2[0] >> 4) | ((qh[0] & kmask4) >> 2)) - 32.f); + + sums.s0 += y[1+ 0] * ((float)((q1[1] & 0xF) | ((qh[1] & kmask1) << 4)) - 32.f); + sums.s1 += y[1+32] * ((float)((q2[1] & 0xF) | ((qh[1] & kmask2) << 2)) - 32.f); + sums.s2 += y[1+64] * ((float)((q1[1] >> 4) | ((qh[1] & kmask3) << 0)) - 32.f); + sums.s3 += y[1+96] * ((float)((q2[1] >> 4) | ((qh[1] & kmask4) >> 2)) - 32.f); + + sums.s0 += y[2+ 0] * ((float)((q1[2] & 0xF) | ((qh[2] & kmask1) << 4)) - 32.f); + sums.s1 += y[2+32] * ((float)((q2[2] & 0xF) | ((qh[2] & kmask2) << 2)) - 32.f); + sums.s2 += y[2+64] * ((float)((q1[2] >> 4) | ((qh[2] & kmask3) << 0)) - 32.f); + sums.s3 += y[2+96] * ((float)((q2[2] >> 4) | ((qh[2] & kmask4) >> 2)) - 32.f); + + sums.s0 += y[3+ 0] * ((float)((q1[3] & 0xF) | ((qh[3] & kmask1) << 4)) - 32.f); + sums.s1 += y[3+32] * ((float)((q2[3] & 0xF) | ((qh[3] & kmask2) << 2)) - 32.f); + sums.s2 += y[3+64] * ((float)((q1[3] >> 4) | ((qh[3] & kmask3) << 0)) - 32.f); + sums.s3 += y[3+96] * ((float)((q2[3] >> 4) | ((qh[3] & kmask4) >> 2)) - 32.f); + + sumf += dall * (sums.s0 * sc[0] + sums.s1 * sc[2] + sums.s2 * sc[4] + sums.s3 * sc[6]); + } + + float tot = sub_group_reduce_add(sumf); + if (get_sub_group_local_id() == 0) { + dst[r1*ne0 + im*ne0*ne1 + row] = tot; + } +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl new file mode 100644 index 000000000..57768c803 --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_mul_mat_Ab_Bi_8x4.cl @@ -0,0 +1,130 @@ +// src0_q, src0_d, src1 are transposed as a preprocessing step +// 4-bit weights are transposed in groups of 4 (unsigned short int) +// consider weights originally "next to each other", now "on top of each other" +// each fiber computes a 8x4 tile of output elements +// using unshuffled weights + +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#pragma OPENCL EXTENSION cl_qcom_reqd_sub_group_size : enable + +__attribute__((qcom_reqd_sub_group_size("full"))) +kernel void kernel_mul_mat_Ab_Bi_8x4( + global const ushort * src0_q, // quantized A + global const half * src0_d, // A scales + __read_only image1d_buffer_t src1, // B (1d image) + global float * dst, // C + int m, // M + int n, // N with padding + int k, // K + int n_no_padding // N without padding +) { + + int m_4 = m >> 2; + int n_4 = n >> 2; + + int gy = get_global_id(0); + int gx = get_global_id(1); + int gx_2 = gx << 2; + + half8 c0 = 0, c1 = 0, c2 = 0, c3 = 0; // 8x4 output elements + half8 B; // registers for activations + half4 dequantized_weights; // registers for dequantized weights + __global const ushort* weight_ptr = src0_q + gx_2; // pointer for weights + __global const half* scale_ptr = src0_d + gx_2; // pointer for scales + + for(int i=0; i> 4) - 8) * scale.s0; // dequantize a row of the 16 weights + dequantized_weights.s1 = (((bits4.s1 & (0x00F0)) >> 4) - 8) * scale.s1; + dequantized_weights.s2 = (((bits4.s2 & (0x00F0)) >> 4) - 8) * scale.s2; + dequantized_weights.s3 = (((bits4.s3 & (0x00F0)) >> 4) - 8) * scale.s3; + c0 += B * dequantized_weights.s0; //vector-scalar multiplication to accumulate + c1 += B * dequantized_weights.s1; + c2 += B * dequantized_weights.s2; + c3 += B * dequantized_weights.s3; + + // j=2 + B.s0123 = read_imageh(src1, gy*2 + (i+2)*(n_4)); + B.s4567 = read_imageh(src1, gy*2 + (i+2)*(n_4)+1); + dequantized_weights.s0 = (((bits4.s0 & (0x0F00)) >> 8) - 8) * scale.s0; // dequantize a row of the 16 weights + dequantized_weights.s1 = (((bits4.s1 & (0x0F00)) >> 8) - 8) * scale.s1; + dequantized_weights.s2 = (((bits4.s2 & (0x0F00)) >> 8) - 8) * scale.s2; + dequantized_weights.s3 = (((bits4.s3 & (0x0F00)) >> 8) - 8) * scale.s3; + c0 += B * dequantized_weights.s0; // vector-scalar multiplication to accumulate + c1 += B * dequantized_weights.s1; + c2 += B * dequantized_weights.s2; + c3 += B * dequantized_weights.s3; + + // j=3 + B.s0123 = read_imageh(src1, gy*2 + (i+3)*(n_4)); + B.s4567 = read_imageh(src1, gy*2 + (i+3)*(n_4)+1); + dequantized_weights.s0 = (((bits4.s0 & (0xF000)) >> 12) - 8) * scale.s0; // dequantize a row of the 16 weights + dequantized_weights.s1 = (((bits4.s1 & (0xF000)) >> 12) - 8) * scale.s1; + dequantized_weights.s2 = (((bits4.s2 & (0xF000)) >> 12) - 8) * scale.s2; + dequantized_weights.s3 = (((bits4.s3 & (0xF000)) >> 12) - 8) * scale.s3; + c0 += B * dequantized_weights.s0; // vector-scalar multiplication to accumulate + c1 += B * dequantized_weights.s1; + c2 += B * dequantized_weights.s2; + c3 += B * dequantized_weights.s3; + } + + int idx = (gy<<3)*m + (gx<<2); // vectorized store 16 elements + + // conditional check if store is to a valid location. Required when N is not a multiple of 8 + // if statements allow registers to be reused for each store + // provides a performance boost due to reduced register footprint, which increases number of concurrent waves + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s0, c1.s0, c2.s0, c3.s0), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s1, c1.s1, c2.s1, c3.s1), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s2, c1.s2, c2.s2, c3.s2), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s3, c1.s3, c2.s3, c3.s3), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s4, c1.s4, c2.s4, c3.s4), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s5, c1.s5, c2.s5, c3.s5), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s6, c1.s6, c2.s6, c3.s6), 0, dst + idx); + idx += m; + } + if(idx+3 < m*n_no_padding){ + vstore4((float4)(c0.s7, c1.s7, c2.s7, c3.s7), 0, dst + idx); + } +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_16.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_16.cl new file mode 100644 index 000000000..d59a0c05d --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_16.cl @@ -0,0 +1,32 @@ +// 16-bit transpose, loading/storing an 8x8 tile of elements + +kernel void kernel_transpose_16( + __read_only image1d_buffer_t input, + __write_only image1d_buffer_t output, + const uint rows, + const uint cols +) { + + const int i = get_global_id(0); + const int j = get_global_id(1); + const int i_3 = i<<3; + const int j_3 = j<<3; + + ushort8 temp0 = as_ushort8(read_imagef(input, (j_3+0)*cols+i)); + ushort8 temp1 = as_ushort8(read_imagef(input, (j_3+1)*cols+i)); + ushort8 temp2 = as_ushort8(read_imagef(input, (j_3+2)*cols+i)); + ushort8 temp3 = as_ushort8(read_imagef(input, (j_3+3)*cols+i)); + ushort8 temp4 = as_ushort8(read_imagef(input, (j_3+4)*cols+i)); + ushort8 temp5 = as_ushort8(read_imagef(input, (j_3+5)*cols+i)); + ushort8 temp6 = as_ushort8(read_imagef(input, (j_3+6)*cols+i)); + ushort8 temp7 = as_ushort8(read_imagef(input, (j_3+7)*cols+i)); + + write_imagef(output, (i_3+0)*rows+j, as_float4((ushort8)(temp0.s0, temp1.s0, temp2.s0, temp3.s0, temp4.s0, temp5.s0, temp6.s0, temp7.s0))); + write_imagef(output, (i_3+1)*rows+j, as_float4((ushort8)(temp0.s1, temp1.s1, temp2.s1, temp3.s1, temp4.s1, temp5.s1, temp6.s1, temp7.s1))); + write_imagef(output, (i_3+2)*rows+j, as_float4((ushort8)(temp0.s2, temp1.s2, temp2.s2, temp3.s2, temp4.s2, temp5.s2, temp6.s2, temp7.s2))); + write_imagef(output, (i_3+3)*rows+j, as_float4((ushort8)(temp0.s3, temp1.s3, temp2.s3, temp3.s3, temp4.s3, temp5.s3, temp6.s3, temp7.s3))); + write_imagef(output, (i_3+4)*rows+j, as_float4((ushort8)(temp0.s4, temp1.s4, temp2.s4, temp3.s4, temp4.s4, temp5.s4, temp6.s4, temp7.s4))); + write_imagef(output, (i_3+5)*rows+j, as_float4((ushort8)(temp0.s5, temp1.s5, temp2.s5, temp3.s5, temp4.s5, temp5.s5, temp6.s5, temp7.s5))); + write_imagef(output, (i_3+6)*rows+j, as_float4((ushort8)(temp0.s6, temp1.s6, temp2.s6, temp3.s6, temp4.s6, temp5.s6, temp6.s6, temp7.s6))); + write_imagef(output, (i_3+7)*rows+j, as_float4((ushort8)(temp0.s7, temp1.s7, temp2.s7, temp3.s7, temp4.s7, temp5.s7, temp6.s7, temp7.s7))); +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32.cl new file mode 100644 index 000000000..914ec0193 --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32.cl @@ -0,0 +1,25 @@ +// 32-bit transpose, loading/storing a 4x4 tile of elements + +kernel void kernel_transpose_32( + __read_only image1d_buffer_t input, + __write_only image1d_buffer_t output, + const uint rows, + const uint cols +) { + + const int i = get_global_id(0); + const int j = get_global_id(1); + const int i_2 = i<<2; + const int j_2 = j<<2; + + float4 temp0 = read_imagef(input, (j_2+0)*cols+i); + float4 temp1 = read_imagef(input, (j_2+1)*cols+i); + float4 temp2 = read_imagef(input, (j_2+2)*cols+i); + float4 temp3 = read_imagef(input, (j_2+3)*cols+i); + + write_imagef(output, (i_2+0)*rows+j, (float4)(temp0.s0, temp1.s0, temp2.s0, temp3.s0)); + write_imagef(output, (i_2+1)*rows+j, (float4)(temp0.s1, temp1.s1, temp2.s1, temp3.s1)); + write_imagef(output, (i_2+2)*rows+j, (float4)(temp0.s2, temp1.s2, temp2.s2, temp3.s2)); + write_imagef(output, (i_2+3)*rows+j, (float4)(temp0.s3, temp1.s3, temp2.s3, temp3.s3)); + +} diff --git a/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32_16.cl b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32_16.cl new file mode 100644 index 000000000..d3bd1fabb --- /dev/null +++ b/ggml/src/ggml-opencl/kernels/ggml-opencl_transpose_32_16.cl @@ -0,0 +1,35 @@ +// 32-bit transpose, loading/storing a 4x4 tile of elements +// Only used for activations +// converts to FP16 +// also adds zero padding for non multiple of 8 prompt lengths +#pragma OPENCL EXTENSION cl_khr_fp16 : enable + +kernel void kernel_transpose_32_16(__read_only image1d_buffer_t input, __write_only image1d_buffer_t output, const uint rows, const uint cols, const uint padded_rows) { + + const int i = get_global_id(0); + const int j = get_global_id(1); + const int i_2 = i<<2; + const int j_2 = j<<2; + half4 temp0 = {0,0,0,0}; // initialize outputs to 0 + half4 temp1 = {0,0,0,0}; + half4 temp2 = {0,0,0,0}; + half4 temp3 = {0,0,0,0}; + + if((j_2+0)*cols+i*4+3 < rows*cols*16){ // only load from a valid location. Otherwise keep register data as 0 + temp0 = read_imageh(input, (j_2+0)*cols+i); + } + if((j_2+1)*cols+i*4+3 < rows*cols*16){ + temp1 = read_imageh(input, (j_2+1)*cols+i); + } + if((j_2+2)*cols+i*4+3 < rows*cols*16){ + temp2 = read_imageh(input, (j_2+2)*cols+i); + } + if((j_2+3)*cols+i*4+3 < rows*cols*16){ + temp3 = read_imageh(input, (j_2+3)*cols+i); + } + + write_imageh(output, (i_2+0)*padded_rows+j, (half4)(temp0.s0, temp1.s0, temp2.s0, temp3.s0)); // no conditionals for output, includes zero padding + write_imageh(output, (i_2+1)*padded_rows+j, (half4)(temp0.s1, temp1.s1, temp2.s1, temp3.s1)); + write_imageh(output, (i_2+2)*padded_rows+j, (half4)(temp0.s2, temp1.s2, temp2.s2, temp3.s2)); + write_imageh(output, (i_2+3)*padded_rows+j, (half4)(temp0.s3, temp1.s3, temp2.s3, temp3.s3)); +} diff --git a/ggml/src/ggml-opt.cpp b/ggml/src/ggml-opt.cpp new file mode 100644 index 000000000..7c3e24103 --- /dev/null +++ b/ggml/src/ggml-opt.cpp @@ -0,0 +1,854 @@ +#include "ggml-opt.h" + +#include "ggml.h" +#include "ggml-alloc.h" +#include "ggml-backend.h" +#include "ggml-impl.h" + +#include +#include +#include +#include +#include +#include +#include + +struct ggml_opt_dataset { + struct ggml_context * ctx = nullptr; + ggml_backend_buffer_t buf = nullptr; + struct ggml_tensor * data = nullptr; + struct ggml_tensor * labels = nullptr; + + int64_t ndata = -1; + int64_t ndata_shard = -1; + size_t nbs_data = -1; + size_t nbs_labels = -1; + + std::vector permutation; +}; + +struct ggml_opt_context { + ggml_backend_sched_t backend_sched = nullptr; + ggml_cgraph * allocated_graph = nullptr; + ggml_cgraph * allocated_graph_copy = nullptr; + struct ggml_context * ctx_static = nullptr; + struct ggml_context * ctx_static_cpu = nullptr; + struct ggml_context * ctx_compute = nullptr; + struct ggml_context * ctx_copy = nullptr; + ggml_backend_buffer_t buf_static = nullptr; + ggml_backend_buffer_t buf_static_cpu = nullptr; + std::mt19937 rng; + + struct ggml_tensor * inputs = nullptr; + struct ggml_tensor * outputs = nullptr; + struct ggml_tensor * labels = nullptr; + + struct ggml_tensor * loss = nullptr; + struct ggml_tensor * pred = nullptr; + struct ggml_tensor * ncorrect = nullptr; + + struct ggml_cgraph * gf = nullptr; + struct ggml_cgraph * gb_grad = nullptr; + struct ggml_cgraph * gb_opt = nullptr; + + int64_t iter = 1; + int32_t opt_period = 1; + int32_t opt_i = 0; + bool loss_per_datapoint = false; + + ggml_opt_get_optimizer_params get_opt_pars = nullptr; + void * get_opt_pars_ud = nullptr; + struct ggml_tensor * adamw_params = nullptr; +}; + +struct ggml_opt_result { + int64_t ndata = 0; + std::vector loss; + std::vector pred; + int64_t ncorrect = 0; + + int64_t opt_period = -1; + bool loss_per_datapoint = false; +}; + +// ====== Dataset ====== + +ggml_opt_dataset_t ggml_opt_dataset_init(int64_t ne_datapoint, int64_t ne_label, int64_t ndata, int64_t ndata_shard) { + GGML_ASSERT(ne_datapoint > 0); + GGML_ASSERT(ne_label >= 0); + GGML_ASSERT(ndata > 0); + GGML_ASSERT(ndata_shard > 0); + + ggml_opt_dataset_t result = new ggml_opt_dataset; + result->ndata = ndata; + result->ndata_shard = ndata_shard; + + { + struct ggml_init_params params = { + /*.mem_size =*/ 2*ggml_tensor_overhead(), + /*.mem_buffer =*/ nullptr, + /*.no_alloc =*/ true, + }; + result->ctx = ggml_init(params); + } + + result->data = ggml_new_tensor_2d(result->ctx, GGML_TYPE_F32, ne_datapoint, ndata); + result->nbs_data = ggml_nbytes(result->data) * ndata_shard/ndata; + + if (ne_label > 0) { + result->labels = ggml_new_tensor_2d(result->ctx, GGML_TYPE_F32, ne_label, ndata); + result->nbs_labels = ggml_nbytes(result->labels) * ndata_shard/ndata; + } else { + result->labels = nullptr; + result->nbs_labels = 0; + } + + result->buf = ggml_backend_alloc_ctx_tensors_from_buft(result->ctx, ggml_backend_cpu_buffer_type()); + + const int64_t nshards = ndata/ndata_shard; + result->permutation.resize(nshards); + for (int64_t i = 0; i < nshards; ++i) { + result->permutation[i] = i; + } + return result; +} + +void ggml_opt_dataset_free(ggml_opt_dataset_t dataset) { + ggml_backend_buffer_free(dataset->buf); + ggml_free(dataset->ctx); + delete dataset; +} + +struct ggml_tensor * ggml_opt_dataset_data(ggml_opt_dataset_t dataset) { + return dataset->data; +} + +struct ggml_tensor * ggml_opt_dataset_labels(ggml_opt_dataset_t dataset) { + return dataset->labels; +} + +void ggml_opt_dataset_shuffle(ggml_opt_context_t opt_ctx, ggml_opt_dataset_t dataset, int64_t idata) { + GGML_ASSERT(idata <= dataset->ndata); + + if (idata < 0) { + std::shuffle(dataset->permutation.begin(), dataset->permutation.end(), opt_ctx->rng); + return; + } + + GGML_ASSERT(idata % dataset->ndata_shard == 0); + const int64_t ishard_max = idata / dataset->ndata_shard; + std::shuffle(dataset->permutation.begin(), dataset->permutation.begin() + ishard_max, opt_ctx->rng); +} + +void ggml_opt_dataset_get_batch(ggml_opt_dataset_t dataset, struct ggml_tensor * data_batch, struct ggml_tensor * labels_batch, int64_t ibatch) { + GGML_ASSERT( data_batch && ggml_is_contiguous(data_batch)); + GGML_ASSERT(!labels_batch || ggml_is_contiguous(labels_batch)); + GGML_ASSERT((labels_batch == nullptr) == (dataset->labels == nullptr)); + + const size_t nb_data_batch = ggml_nbytes(data_batch); + GGML_ASSERT(nb_data_batch % dataset->nbs_data == 0); + const int64_t shards_per_batch = nb_data_batch / dataset->nbs_data; + + if (labels_batch) { + const size_t nb_labels_batch = ggml_nbytes(labels_batch); + GGML_ASSERT(nb_labels_batch == shards_per_batch*dataset->nbs_labels); + } + + GGML_ASSERT((ibatch + 1)*shards_per_batch <= int64_t(dataset->permutation.size())); + + for (int64_t ishard_batch = 0; ishard_batch < shards_per_batch; ++ishard_batch) { + const int64_t ishard = dataset->permutation[ibatch*shards_per_batch + ishard_batch]; + + const char * ptr_data = (const char *) dataset->data->data + ishard*dataset->nbs_data; + ggml_backend_tensor_set(data_batch, ptr_data, ishard_batch*dataset->nbs_data, dataset->nbs_data); + + if (!labels_batch) { + continue; + } + + const char * ptr_labels = (const char *) dataset->labels->data + ishard*dataset->nbs_labels; + ggml_backend_tensor_set(labels_batch, ptr_labels, ishard_batch*dataset->nbs_labels, dataset->nbs_labels); + } +} + +// ====== Model / Context ====== + +struct ggml_opt_optimizer_params ggml_opt_get_default_optimizer_params(void * userdata) { + GGML_UNUSED(userdata); + + ggml_opt_optimizer_params result; + + result.adamw.alpha = 0.001f; + result.adamw.beta1 = 0.9f; + result.adamw.beta2 = 0.999f; + result.adamw.eps = 1e-8f; + result.adamw.wd = 0.0f; + + return result; +} + +struct ggml_opt_params ggml_opt_default_params( + ggml_backend_sched_t backend_sched, + struct ggml_context * ctx_compute, + struct ggml_tensor * inputs, + struct ggml_tensor * outputs, + enum ggml_opt_loss_type loss_type) { + return { + /*backend_sched =*/ backend_sched, + /*ctx_compute =*/ ctx_compute, + /*inputs =*/ inputs, + /*logits =*/ outputs, + /*loss_type =*/ loss_type, + /*build_type =*/ GGML_OPT_BUILD_TYPE_OPT, + /*opt_period =*/ 1, + /*get_opt_pars =*/ ggml_opt_get_default_optimizer_params, + /*get_opt_pars_ud =*/ nullptr, + }; +} + +static ggml_tensor * map_tensor(std::map & tensor_map, ggml_context * ctx, ggml_tensor * tensor) { + if (!tensor) { + return nullptr; + } + + if (tensor_map.find(tensor) != tensor_map.end()) { + return tensor_map[tensor]; + } + + ggml_tensor * new_tensor = ggml_dup_tensor(ctx, tensor); + tensor_map[tensor] = new_tensor; + + new_tensor->op = tensor->op; + for (int i = 0; i < GGML_MAX_DIMS; i++) { + new_tensor->nb[i] = tensor->nb[i]; + } + new_tensor->flags = tensor->flags; + memcpy(new_tensor->op_params, tensor->op_params, sizeof(tensor->op_params)); + strcpy(new_tensor->name, tensor->name); + new_tensor->data = tensor->data; + new_tensor->buffer = tensor->buffer; + new_tensor->extra = tensor->extra; + new_tensor->view_offs = tensor->view_offs; + new_tensor->view_src = map_tensor(tensor_map, ctx, tensor->view_src); + for (int i = 0; i < GGML_MAX_SRC; i++) { + new_tensor->src[i] = map_tensor(tensor_map, ctx, tensor->src[i]); + } + + return new_tensor; +} + +static ggml_cgraph * dup_graph(ggml_context * ctx, ggml_cgraph * src) { + std::map tensor_map; + + ggml_cgraph * dst = ggml_new_graph_custom(ctx, src->size, /*grads =*/ true); + + for (int i = 0; i < src->n_leafs; i++) { + ggml_build_forward_expand(dst, map_tensor(tensor_map, ctx, src->leafs[i])); + } + GGML_ASSERT(dst->n_leafs == src->n_leafs); + for (int i = 0; i < src->n_nodes; i++) { + ggml_build_forward_expand(dst, map_tensor(tensor_map, ctx, src->nodes[i])); + } + GGML_ASSERT(dst->n_nodes == src->n_nodes); + for (int i = 0; i < src->n_nodes; ++i) { + const size_t igrad_src = ggml_hash_find(&src->visited_hash_set, src->nodes[i]); + const size_t igrad_dst = ggml_hash_find(&dst->visited_hash_set, dst->nodes[i]); + + GGML_ASSERT(igrad_src != GGML_HASHSET_FULL); + GGML_ASSERT(ggml_bitset_get(src->visited_hash_set.used, igrad_src)); + GGML_ASSERT(igrad_dst != GGML_HASHSET_FULL); + GGML_ASSERT(ggml_bitset_get(dst->visited_hash_set.used, igrad_dst)); + + dst->grads[igrad_dst] = src->grads[igrad_src]; + dst->grad_accs[igrad_dst] = src->grad_accs[igrad_src]; + } + + return dst; +} + +static void ggml_opt_alloc_graph(ggml_opt_context_t opt_ctx, ggml_cgraph * graph) { + GGML_ASSERT(graph); + if (opt_ctx->allocated_graph == graph) { + return; + } + + ggml_backend_sched_reset(opt_ctx->backend_sched); // clear allocation of previous graph + + { + ggml_init_params params = { + /*.mem_size =*/ ggml_tensor_overhead() * GGML_DEFAULT_GRAPH_SIZE, + /*.mem_buffer =*/ nullptr, + /*.no_alloc =*/ true, + }; + ggml_free(opt_ctx->ctx_copy); + opt_ctx->ctx_copy = ggml_init(params); + } + + opt_ctx->allocated_graph_copy = dup_graph(opt_ctx->ctx_copy, graph); + + ggml_backend_sched_alloc_graph(opt_ctx->backend_sched, opt_ctx->allocated_graph_copy); + opt_ctx->allocated_graph = graph; +} + +ggml_opt_context_t ggml_opt_init(struct ggml_opt_params params) { + ggml_opt_context_t result = new struct ggml_opt_context; + result->backend_sched = params.backend_sched; + result->ctx_compute = params.ctx_compute; + result->inputs = params.inputs; + result->outputs = params.outputs; + result->opt_period = params.opt_period; + result->get_opt_pars = params.get_opt_pars; + result->get_opt_pars_ud = params.get_opt_pars_ud; + + GGML_ASSERT(result->inputs->data && "the inputs must be allocated statically"); + GGML_ASSERT(result->opt_period >= 1); + + const bool accumulate = params.build_type == GGML_OPT_BUILD_TYPE_GRAD || + (params.build_type == GGML_OPT_BUILD_TYPE_OPT && result->opt_period > 1); + + ggml_set_input(result->inputs); + ggml_set_output(result->outputs); + + result->gf = ggml_new_graph_custom(result->ctx_compute, GGML_DEFAULT_GRAPH_SIZE, /*grads =*/ true); // Forward pass. + ggml_build_forward_expand(result->gf, result->outputs); + + int n_param = 0; + for (int i = 0; i < result->gf->n_nodes; ++i) { + if (result->gf->nodes[i]->flags & GGML_TENSOR_FLAG_PARAM) { + n_param++; + } + } + + { + // The static context is used for: + // - gradients (1 tensor per param if using gradient accumulation) + // - optimizer momenta (2 tensors per param) + // - labels + // - loss + its gradient (up to 5 tensors) + // - pred + // - ncorrect (2 tensors). + const size_t tensors_per_param = (accumulate ? 1 : 0) + (params.build_type == GGML_OPT_BUILD_TYPE_OPT ? 2 : 0); + const size_t size_meta = (tensors_per_param*n_param + 9) * ggml_tensor_overhead(); + struct ggml_init_params params = { + /*.mem_size =*/ size_meta, + /*.mem_buffer =*/ nullptr, + /*.no_alloc =*/ true, + }; + result->ctx_static = ggml_init(params); + } + { + // The static cpu context is used for: + // - optimizer parameters (1 for the entire context) + const size_t size_meta = 1 * ggml_tensor_overhead(); + struct ggml_init_params params = { + /*.mem_size =*/ size_meta, + /*.mem_buffer =*/ nullptr, + /*.no_alloc =*/ true, + }; + result->ctx_static_cpu = ggml_init(params); + } + + + switch (params.loss_type) { + case GGML_OPT_LOSS_TYPE_MEAN: { + result->loss = ggml_sum(result->ctx_static, result->outputs); + ggml_set_name(result->loss, "loss_sum"); + const float scale = 1.0f / (result->opt_period * ggml_nelements(result->outputs)); + result->loss = ggml_scale(result->ctx_static, result->loss, scale); + ggml_set_name(result->loss, "loss_mean"); + result->loss_per_datapoint = true; + break; + } + case GGML_OPT_LOSS_TYPE_SUM: { + result->loss = ggml_sum(result->ctx_static, result->outputs); + ggml_set_name(result->loss, "loss_sum"); + result->loss_per_datapoint = false; + break; + } + case GGML_OPT_LOSS_TYPE_CROSS_ENTROPY: { + result->labels = ggml_dup_tensor(result->ctx_static, result->outputs); + ggml_set_input(result->labels); + ggml_set_name(result->labels, "labels"); + result->loss = ggml_cross_entropy_loss(result->ctx_static, result->outputs, result->labels); + ggml_set_name(result->loss, "loss_cross_entropy"); + if (result->opt_period > 1) { + result->loss = ggml_scale(result->ctx_static, result->loss, 1.0f / result->opt_period); + ggml_set_name(result->loss, "loss_cross_entropy_scaled"); + } + result->loss_per_datapoint = true; + break; + } + case GGML_OPT_LOSS_TYPE_MEAN_SQUARED_ERROR: { + result->labels = ggml_dup_tensor(result->ctx_static, result->outputs); + ggml_set_input(result->labels); + ggml_set_name(result->labels, "labels"); + result->loss = ggml_sub(result->ctx_static, result->outputs, result->labels); + ggml_set_name(result->loss, "loss_error"); + result->loss = ggml_sqr(result->ctx_static, result->loss); + ggml_set_name(result->loss, "loss_squared_error"); + result->loss = ggml_sum(result->ctx_static, result->loss); + ggml_set_name(result->loss, "loss_sum_squared_error"); + const float scale = 1.0f / (result->opt_period * ggml_nelements(result->outputs)); + result->loss = ggml_scale(result->ctx_static, result->loss, scale); + ggml_set_name(result->loss, "loss_mean_squared_error"); + result->loss_per_datapoint = true; + break; + } + } + ggml_set_output(result->loss); + ggml_set_loss(result->loss); + ggml_build_forward_expand(result->gf, result->loss); + + result->pred = ggml_argmax(result->ctx_static, result->outputs); + ggml_set_name(result->pred, "pred"); + ggml_set_output(result->pred); + ggml_build_forward_expand(result->gf, result->pred); + + if (result->labels) { + result->ncorrect = ggml_count_equal(result->ctx_static, result->pred, ggml_argmax(result->ctx_static, result->labels)); + ggml_set_name(result->ncorrect, "ncorrect"); + ggml_set_output(result->ncorrect); + ggml_build_forward_expand(result->gf, result->ncorrect); + } else { + result->ncorrect = nullptr; + } + + if (params.build_type == GGML_OPT_BUILD_TYPE_FORWARD) { + result->buf_static = ggml_backend_alloc_ctx_tensors(result->ctx_static, ggml_backend_sched_get_backend(result->backend_sched, 0)); + return result; + } + + // gb_grad == graph backward gradients, forward pass, then backward pass to calculate gradients. + result->gb_grad = ggml_graph_dup(result->ctx_compute, result->gf); + ggml_build_backward_expand(result->ctx_static, result->ctx_compute, result->gb_grad, accumulate); + + if (params.build_type == GGML_OPT_BUILD_TYPE_GRAD) { + result->buf_static = ggml_backend_alloc_ctx_tensors(result->ctx_static, ggml_backend_sched_get_backend(result->backend_sched, 0)); + ggml_graph_reset(result->gb_grad); + return result; + } + + GGML_ASSERT(params.build_type == GGML_OPT_BUILD_TYPE_OPT); + + // gb_opt == graph backward optimize, forward pass, then backward pass to calculate gradients, then optimizer step. + result->gb_opt = ggml_graph_dup(result->ctx_compute, result->gb_grad); + + result->adamw_params = ggml_new_tensor_1d(result->ctx_static_cpu, GGML_TYPE_F32, 7); + ggml_set_input(result->adamw_params); + ggml_set_name(result->adamw_params, "adamw_params"); + + for (int i = result->gf->n_nodes-1; i >= 0; --i) { + struct ggml_tensor * node = result->gb_opt->nodes[i]; + struct ggml_tensor * grad = ggml_graph_get_grad(result->gb_opt, node); + + if (node->flags & GGML_TENSOR_FLAG_PARAM) { + struct ggml_tensor * m = ggml_dup_tensor(result->ctx_static, node); + struct ggml_tensor * v = ggml_dup_tensor(result->ctx_static, node); + struct ggml_tensor * opt_step = ggml_opt_step_adamw(result->ctx_compute, node, grad, m, v, result->adamw_params); + ggml_build_forward_expand(result->gb_opt, opt_step); + } + } + + result->buf_static = ggml_backend_alloc_ctx_tensors( + result->ctx_static, ggml_backend_sched_get_backend(result->backend_sched, 0)); + + result->buf_static_cpu = ggml_backend_alloc_ctx_tensors_from_buft(result->ctx_static_cpu, ggml_backend_cpu_buffer_type()); + + ggml_graph_reset(result->gb_opt); + + return result; +} + +void ggml_opt_free(ggml_opt_context_t opt_ctx) { + if (opt_ctx == nullptr) { + return; + } + ggml_backend_buffer_free(opt_ctx->buf_static); + ggml_backend_buffer_free(opt_ctx->buf_static_cpu); + ggml_free(opt_ctx->ctx_static); + ggml_free(opt_ctx->ctx_static_cpu); + delete opt_ctx; +} + +void ggml_opt_reset(ggml_opt_context_t opt_ctx, bool optimizer) { + if (optimizer) { + ggml_graph_reset(opt_ctx->gb_opt); + opt_ctx->iter = 1; + } else { + ggml_graph_reset(opt_ctx->gb_grad); + } +} + +struct ggml_tensor * ggml_opt_inputs(ggml_opt_context_t opt_ctx) { + return opt_ctx->inputs; +} + +struct ggml_tensor * ggml_opt_outputs(ggml_opt_context_t opt_ctx) { + return opt_ctx->outputs; +} + +struct ggml_tensor * ggml_opt_labels(ggml_opt_context_t opt_ctx) { + return opt_ctx->labels; +} + +struct ggml_tensor * ggml_opt_loss(ggml_opt_context_t opt_ctx) { + return opt_ctx->loss; +} + +struct ggml_tensor * ggml_opt_pred(ggml_opt_context_t opt_ctx) { + return opt_ctx->pred; +} + +struct ggml_tensor * ggml_opt_ncorrect(ggml_opt_context_t opt_ctx) { + return opt_ctx->ncorrect; +} + +struct ggml_tensor * ggml_opt_grad_acc(ggml_opt_context_t opt_ctx, struct ggml_tensor * node) { + return ggml_graph_get_grad_acc(opt_ctx->gb_opt, node); +} + +// ====== Optimization Result ====== + +ggml_opt_result_t ggml_opt_result_init() { + return new ggml_opt_result; +} + +void ggml_opt_result_free(ggml_opt_result_t result) { + delete result; +} + +void ggml_opt_result_reset(ggml_opt_result_t result) { + result->ndata = 0; + result->loss.clear(); + result->pred.clear(); + result->ncorrect = 0; +} + +void ggml_opt_result_ndata(ggml_opt_result_t result, int64_t * ndata) { + *ndata = result->ndata; +} + +void ggml_opt_result_loss(ggml_opt_result_t result, double * loss, double * unc) { + const int64_t nbatches = result->loss.size(); // Number of physical batches. + + if (nbatches == 0) { + *loss = 0.0; + *unc = NAN; + return; + } + + double sum = 0.0; + double sum_squared = 0.0; + + for (const float & loss : result->loss) { + // If the loss is per datapoint it was scaled by 1.0f/opt_period for each physical batch. + const float loss_scaled = result->loss_per_datapoint ? loss*result->opt_period : loss; + sum += loss_scaled; + sum_squared += loss_scaled*loss_scaled; + } + + const double mean = sum/nbatches; + *loss = result->loss_per_datapoint ? mean : sum; + + if (!unc) { + return; + } + + if (nbatches < 2) { + *unc = NAN; + return; + } + + const double var_sum = sum_squared/nbatches - mean*mean; // variance without Bessel's correction, i.e. nbatches/(nbatches-1) + *unc = result->loss_per_datapoint ? sqrt(var_sum / (nbatches - 1)) : sqrt(var_sum * nbatches/(nbatches - 1)); +} + +void ggml_opt_result_pred(ggml_opt_result_t result, int32_t * pred) { + for (size_t i = 0; i < result->pred.size(); ++i) { + pred[i] = result->pred[i]; + } +} + +void ggml_opt_result_accuracy(ggml_opt_result_t result, double * accuracy, double * unc) { + *accuracy = result->ncorrect >= 0 ? double(result->ncorrect) / double(result->ndata) : NAN; + + if (!unc) { + return; + } + + *unc = result->ncorrect >= 0 && result->ndata >= 2 ? + sqrt((*accuracy) * (1.0 - (*accuracy)) / double(result->ndata - 1)) : NAN; +} + +// ====== Computation ====== + +static void ggml_opt_eval_graph(ggml_opt_context_t opt_ctx, ggml_cgraph * graph, ggml_opt_result * result) { + if (graph != opt_ctx->gf) { + struct ggml_opt_optimizer_params opt_pars = opt_ctx->get_opt_pars(opt_ctx->get_opt_pars_ud); + + GGML_ASSERT(opt_pars.adamw.alpha > 0.0f); + GGML_ASSERT(opt_pars.adamw.beta1 >= 0.0f); + GGML_ASSERT(opt_pars.adamw.beta1 <= 1.0f); + GGML_ASSERT(opt_pars.adamw.beta2 >= 0.0f); + GGML_ASSERT(opt_pars.adamw.beta2 <= 1.0f); + GGML_ASSERT(opt_pars.adamw.eps >= 0.0f); + GGML_ASSERT(opt_pars.adamw.wd >= 0.0f); + GGML_ASSERT(opt_pars.adamw.wd <= 1.0f); + + // beta1, beta2 after applying warmup + const float beta1h = 1.0f/(1.0f - powf(opt_pars.adamw.beta1, opt_ctx->iter)); + const float beta2h = 1.0f/(1.0f - powf(opt_pars.adamw.beta2, opt_ctx->iter)); + + float * adamw_par_data = ggml_get_data_f32(opt_ctx->adamw_params); + adamw_par_data[0] = opt_pars.adamw.alpha; + adamw_par_data[1] = opt_pars.adamw.beta1; + adamw_par_data[2] = opt_pars.adamw.beta2; + adamw_par_data[3] = opt_pars.adamw.eps; + adamw_par_data[4] = opt_pars.adamw.wd; + adamw_par_data[5] = beta1h; + adamw_par_data[6] = beta2h; + } + + ggml_opt_alloc_graph(opt_ctx, graph); + ggml_backend_sched_graph_compute(opt_ctx->backend_sched, opt_ctx->allocated_graph_copy); + opt_ctx->iter += opt_ctx->allocated_graph == opt_ctx->gb_opt; + + if (!result) { + return; + } + + if (result->ndata == 0) { + result->loss_per_datapoint = opt_ctx->loss_per_datapoint; + result->opt_period = opt_ctx->opt_period; + } else { + GGML_ASSERT(result->loss_per_datapoint == opt_ctx->loss_per_datapoint); + GGML_ASSERT(result->opt_period == opt_ctx->opt_period); + } + + const int64_t ndata = opt_ctx->outputs->ne[1]; + GGML_ASSERT(result->ndata == ndata*int64_t(result->loss.size()) && "varying batch size not supported"); + result->ndata += ndata; + + GGML_ASSERT(ggml_is_scalar(opt_ctx->loss)); + GGML_ASSERT(opt_ctx->loss->type == GGML_TYPE_F32); + float loss; + ggml_backend_tensor_get(opt_ctx->loss, &loss, 0, ggml_nbytes(opt_ctx->loss)); + result->loss.push_back(loss); + + GGML_ASSERT(opt_ctx->pred->type == GGML_TYPE_I32); + std::vector pred(ndata); + ggml_backend_tensor_get(opt_ctx->pred, pred.data(), 0, ggml_nbytes(opt_ctx->pred)); + result->pred.insert(result->pred.end(), pred.begin(), pred.end()); + + if (!opt_ctx->labels || result->ncorrect < 0) { + result->ncorrect = -1; + return; + } + + GGML_ASSERT(ggml_is_scalar(opt_ctx->ncorrect)); + GGML_ASSERT(opt_ctx->ncorrect->type == GGML_TYPE_I64); + int64_t ncorrect; + ggml_backend_tensor_get(opt_ctx->ncorrect, &ncorrect, 0, ggml_nbytes(opt_ctx->ncorrect)); + result->ncorrect += ncorrect; +} + +void ggml_opt_forward(ggml_opt_context_t opt_ctx, ggml_opt_result * result) { + ggml_opt_eval_graph(opt_ctx, opt_ctx->gf, result); +} + +void ggml_opt_forward_backward(ggml_opt_context_t opt_ctx, ggml_opt_result * result) { + if (opt_ctx->opt_period == 1) { + ggml_opt_eval_graph(opt_ctx, opt_ctx->gb_opt, result); + return; + } + + const int32_t opt_i_next = (opt_ctx->opt_i + 1) % opt_ctx->opt_period; + if (opt_i_next == 0) { + ggml_opt_eval_graph(opt_ctx, opt_ctx->gb_opt, result); + ggml_opt_reset(opt_ctx, /*optimizer =*/ false); + } else { + ggml_opt_eval_graph(opt_ctx, opt_ctx->gb_grad, result); + } + opt_ctx->opt_i = opt_i_next; +} + +// ====== High-Level Functions ====== + +void ggml_opt_epoch( + ggml_opt_context_t opt_ctx, + ggml_opt_dataset_t dataset, + ggml_opt_result_t result_train, + ggml_opt_result_t result_eval, + int64_t idata_split, + ggml_opt_epoch_callback callback_train, + ggml_opt_epoch_callback callback_eval) { + struct ggml_tensor * inputs = ggml_opt_inputs(opt_ctx); + struct ggml_tensor * labels = ggml_opt_labels(opt_ctx); + struct ggml_tensor * data = ggml_opt_dataset_data(dataset); + GGML_ASSERT(data->ne[0] == inputs->ne[0]); + + const int64_t ndata = data->ne[1]; + const int64_t ndata_batch = inputs->ne[1]; + + GGML_ASSERT(data->ne[1] % inputs->ne[1] == 0); + const int64_t nbatches = ndata/ndata_batch; + + idata_split = idata_split < 0 ? ndata : idata_split; + GGML_ASSERT(idata_split % ndata_batch == 0); + const int64_t ibatch_split = idata_split / ndata_batch; + + int64_t ibatch = 0; + int64_t t_loop_start = ggml_time_us(); + for (; ibatch < ibatch_split; ++ibatch) { + ggml_opt_dataset_get_batch(dataset, inputs, labels, ibatch); + ggml_opt_forward_backward(opt_ctx, result_train); + if (callback_train) { + callback_train(true, opt_ctx, dataset, result_train, ibatch+1, ibatch_split, t_loop_start); + } + } + t_loop_start = ggml_time_us(); + for (; ibatch < nbatches; ++ibatch) { + ggml_opt_dataset_get_batch(dataset, inputs, labels, ibatch); + ggml_opt_forward(opt_ctx, result_eval); + if (callback_eval) { + callback_eval(false, opt_ctx, dataset, result_eval, ibatch+1-ibatch_split, nbatches-ibatch_split, t_loop_start); + } + } +} + +void ggml_opt_epoch_callback_progress_bar( + bool train, + ggml_opt_context_t opt_ctx, + ggml_opt_dataset_t dataset, + ggml_opt_result_t result, + int64_t ibatch, + int64_t ibatch_max, + int64_t t_start_us) { + fprintf(stderr, "%s[", train ? "train: " : "val: "); + + constexpr int64_t bar_length = 25; + for (int64_t j = 0; j < bar_length; ++j) { + const int64_t ibatch_j = ibatch_max * j/bar_length; + if (ibatch_j < ibatch) { + fprintf(stderr, "="); + } else if (ibatch_max * (j - 1)/bar_length < ibatch) { + fprintf(stderr, ">"); + } else { + fprintf(stderr, " "); + } + } + + const int64_t batch_size = ggml_opt_inputs(opt_ctx)->ne[1]; + const int64_t idata = ibatch*batch_size; + const int64_t idata_max = ibatch_max*batch_size; + + double loss; + double loss_unc; + ggml_opt_result_loss(result, &loss, &loss_unc); + + double accuracy; + double accuracy_unc; + ggml_opt_result_accuracy(result, &accuracy, &accuracy_unc); + + const int64_t t_ibatch_us = ggml_time_us() - t_start_us; + int64_t t_ibatch_s = t_ibatch_us / 1000000; + const int64_t t_ibatch_h = t_ibatch_s / 3600; + t_ibatch_s -= t_ibatch_h * 3600; + const int64_t t_ibatch_m = t_ibatch_s / 60; + t_ibatch_s -= t_ibatch_m * 60; + + const int64_t t_eta_us = t_ibatch_us * (ibatch_max - ibatch)/ibatch; + int64_t t_eta_s = t_eta_us / 1000000; + const int64_t t_eta_h = t_eta_s / 3600; + t_eta_s -= t_eta_h * 3600; + const int64_t t_eta_m = t_eta_s / 60; + t_eta_s -= t_eta_m * 60; + + fprintf(stderr, "| data=%06" PRId64 "/%06" PRId64 ", loss=%.6lf+-%.6lf, accuracy=%.2lf+-%.2lf%%, " + "t=%02" PRId64 ":%02" PRId64 ":%02" PRId64 ", ETA=%02" PRId64 ":%02" PRId64 ":%02" PRId64 "]\r", + idata, idata_max, loss, loss_unc, 100.0*accuracy, 100.0*accuracy_unc, + t_ibatch_h, t_ibatch_m, t_ibatch_s, t_eta_h, t_eta_m, t_eta_s); + if (ibatch == ibatch_max) { + fprintf(stderr, "\n"); + } + fflush(stderr); + + GGML_UNUSED(dataset); +} + +void ggml_opt_fit( + ggml_backend_sched_t backend_sched, + ggml_context * ctx_compute, + ggml_tensor * inputs, + ggml_tensor * outputs, + ggml_opt_dataset_t dataset, + enum ggml_opt_loss_type loss_type, + ggml_opt_get_optimizer_params get_opt_pars, + int64_t nepoch, + int64_t nbatch_logical, + float val_split, + bool silent) { + ggml_time_init(); + const int64_t t_start_us = ggml_time_us(); + + const int64_t ndata = ggml_opt_dataset_data(dataset)->ne[1]; + const int64_t nbatch_physical = inputs->ne[1]; + GGML_ASSERT(ndata % nbatch_logical == 0); + GGML_ASSERT(nbatch_logical % nbatch_physical == 0); + + const int64_t opt_period = nbatch_logical / nbatch_physical; + const int64_t nbatches_logical = ndata / nbatch_logical; + + GGML_ASSERT(val_split >= 0.0f); + GGML_ASSERT(val_split < 1.0f); + const int64_t ibatch_split = int64_t(((1.0f - val_split) * nbatches_logical)) * opt_period; // train <-> val split index (physical) + const int64_t idata_split = ibatch_split * nbatch_physical; + + int64_t epoch = 1; + + ggml_opt_params params = ggml_opt_default_params(backend_sched, ctx_compute, inputs, outputs, loss_type); + params.opt_period = opt_period; + params.get_opt_pars = get_opt_pars; + params.get_opt_pars_ud = &epoch; + ggml_opt_context_t opt_ctx = ggml_opt_init(params); + + // Shuffling the data is generally useful but there is only a point if not all data is used in a single batch. + if (nbatch_logical < ndata) { + ggml_opt_dataset_shuffle(opt_ctx, dataset, -1); // Shuffle all data (train + validation). + } + + ggml_opt_result_t result_train = ggml_opt_result_init(); + ggml_opt_result_t result_val = ggml_opt_result_init(); + + ggml_opt_epoch_callback epoch_callback = silent ? nullptr : ggml_opt_epoch_callback_progress_bar; + + for (; epoch <= nepoch; ++epoch) { + if (nbatch_logical < idata_split) { + ggml_opt_dataset_shuffle(opt_ctx, dataset, idata_split); + } + + ggml_opt_result_reset(result_train); + ggml_opt_result_reset(result_val); + + if (!silent) { + fprintf(stderr, "%s: epoch %04" PRId64 "/%04" PRId64 ":\n", __func__, epoch, nepoch); + } + ggml_opt_epoch(opt_ctx, dataset, result_train, result_val, idata_split, epoch_callback, epoch_callback); + if (!silent) { + fprintf(stderr, "\n"); + } + } + + if (!silent) { + int64_t t_total_s = (ggml_time_us() - t_start_us) / 1000000; + const int64_t t_total_h = t_total_s / 3600; + t_total_s -= t_total_h * 3600; + const int64_t t_total_m = t_total_s / 60; + t_total_s -= t_total_m * 60; + fprintf(stderr, "%s: training took %02" PRId64 ":%02" PRId64 ":%02" PRId64 "\n", __func__, t_total_h, t_total_m, t_total_s); + } + + ggml_opt_free(opt_ctx); + ggml_opt_result_free(result_train); + ggml_opt_result_free(result_val); +} diff --git a/ggml/src/ggml-quants.c b/ggml/src/ggml-quants.c index 82a463f27..7918388ae 100644 --- a/ggml/src/ggml-quants.c +++ b/ggml/src/ggml-quants.c @@ -3,7 +3,7 @@ #include "ggml-quants.h" #include "ggml-impl.h" -#include "ggml-cpu-impl.h" +#include "ggml-cpu/ggml-cpu-impl.h" #include "ggml-cpu.h" #include @@ -27,643 +27,6 @@ #define UNUSED GGML_UNUSED -// some compilers don't provide _mm256_set_m128i, e.g. gcc 7 -#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1) - -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) -// multiply int8_t, add results pairwise twice -static inline __m128i mul_sum_i8_pairs(const __m128i x, const __m128i y) { - // Get absolute values of x vectors - const __m128i ax = _mm_sign_epi8(x, x); - // Sign the values of the y vectors - const __m128i sy = _mm_sign_epi8(y, x); - // Perform multiplication and create 16-bit values - const __m128i dot = _mm_maddubs_epi16(ax, sy); - const __m128i ones = _mm_set1_epi16(1); - return _mm_madd_epi16(ones, dot); -} - -#if __AVX__ || __AVX2__ || __AVX512F__ -// horizontally add 8 floats -static inline float hsum_float_8(const __m256 x) { - __m128 res = _mm256_extractf128_ps(x, 1); - res = _mm_add_ps(res, _mm256_castps256_ps128(x)); - res = _mm_add_ps(res, _mm_movehl_ps(res, res)); - res = _mm_add_ss(res, _mm_movehdup_ps(res)); - return _mm_cvtss_f32(res); -} - -// horizontally add 8 int32_t -static inline int hsum_i32_8(const __m256i a) { - const __m128i sum128 = _mm_add_epi32(_mm256_castsi256_si128(a), _mm256_extractf128_si256(a, 1)); - const __m128i hi64 = _mm_unpackhi_epi64(sum128, sum128); - const __m128i sum64 = _mm_add_epi32(hi64, sum128); - const __m128i hi32 = _mm_shuffle_epi32(sum64, _MM_SHUFFLE(2, 3, 0, 1)); - return _mm_cvtsi128_si32(_mm_add_epi32(sum64, hi32)); -} - -// horizontally add 4 int32_t -static inline int hsum_i32_4(const __m128i a) { - const __m128i hi64 = _mm_unpackhi_epi64(a, a); - const __m128i sum64 = _mm_add_epi32(hi64, a); - const __m128i hi32 = _mm_shuffle_epi32(sum64, _MM_SHUFFLE(2, 3, 0, 1)); - return _mm_cvtsi128_si32(_mm_add_epi32(sum64, hi32)); -} - -#if defined(__AVX2__) || defined(__AVX512F__) -// spread 32 bits to 32 bytes { 0x00, 0xFF } -static inline __m256i bytes_from_bits_32(const uint8_t * x) { - uint32_t x32; - memcpy(&x32, x, sizeof(uint32_t)); - const __m256i shuf_mask = _mm256_set_epi64x( - 0x0303030303030303, 0x0202020202020202, - 0x0101010101010101, 0x0000000000000000); - __m256i bytes = _mm256_shuffle_epi8(_mm256_set1_epi32(x32), shuf_mask); - const __m256i bit_mask = _mm256_set1_epi64x(0x7fbfdfeff7fbfdfe); - bytes = _mm256_or_si256(bytes, bit_mask); - return _mm256_cmpeq_epi8(bytes, _mm256_set1_epi64x(-1)); -} - -// Unpack 32 4-bit fields into 32 bytes -// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval -static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) -{ - const __m128i tmp = _mm_loadu_si128((const __m128i *)rsi); - const __m256i bytes = MM256_SET_M128I(_mm_srli_epi16(tmp, 4), tmp); - const __m256i lowMask = _mm256_set1_epi8( 0xF ); - return _mm256_and_si256(lowMask, bytes); -} - -// add int16_t pairwise and return as float vector -static inline __m256 sum_i16_pairs_float(const __m256i x) { - const __m256i ones = _mm256_set1_epi16(1); - const __m256i summed_pairs = _mm256_madd_epi16(ones, x); - return _mm256_cvtepi32_ps(summed_pairs); -} - -static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { -#if defined(__AVXVNNI__) || (defined(__AVX512VNNI__) && defined(__AVX512VL__)) - const __m256i zero = _mm256_setzero_si256(); - const __m256i summed_pairs = _mm256_dpbusd_epi32(zero, ax, sy); - return _mm256_cvtepi32_ps(summed_pairs); -#else - // Perform multiplication and create 16-bit values - const __m256i dot = _mm256_maddubs_epi16(ax, sy); - return sum_i16_pairs_float(dot); -#endif -} - -// multiply int8_t, add results pairwise twice and return as float vector -static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { -#if __AVXVNNIINT8__ - const __m256i zero = _mm256_setzero_si256(); - const __m256i summed_pairs = _mm256_dpbssd_epi32(zero, x, y); - return _mm256_cvtepi32_ps(summed_pairs); -#else - // Get absolute values of x vectors - const __m256i ax = _mm256_sign_epi8(x, x); - // Sign the values of the y vectors - const __m256i sy = _mm256_sign_epi8(y, x); - return mul_sum_us8_pairs_float(ax, sy); -#endif -} - -static inline __m128i packNibbles( __m256i bytes ) -{ - // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh -#if __AVX512F__ - const __m256i bytes_srli_4 = _mm256_srli_epi16(bytes, 4); // 0000_0000_abcd_0000 - bytes = _mm256_or_si256(bytes, bytes_srli_4); // 0000_abcd_abcd_efgh - return _mm256_cvtepi16_epi8(bytes); // abcd_efgh -#else - const __m256i lowByte = _mm256_set1_epi16( 0xFF ); - __m256i high = _mm256_andnot_si256( lowByte, bytes ); - __m256i low = _mm256_and_si256( lowByte, bytes ); - high = _mm256_srli_epi16( high, 4 ); - bytes = _mm256_or_si256( low, high ); - - // Compress uint16_t lanes into bytes - __m128i r0 = _mm256_castsi256_si128( bytes ); - __m128i r1 = _mm256_extracti128_si256( bytes, 1 ); - return _mm_packus_epi16( r0, r1 ); -#endif -} -#elif defined(__AVX__) -// spread 32 bits to 32 bytes { 0x00, 0xFF } -static inline __m256i bytes_from_bits_32(const uint8_t * x) { - uint32_t x32; - memcpy(&x32, x, sizeof(uint32_t)); - const __m128i shuf_maskl = _mm_set_epi64x(0x0101010101010101, 0x0000000000000000); - const __m128i shuf_maskh = _mm_set_epi64x(0x0303030303030303, 0x0202020202020202); - __m128i bytesl = _mm_shuffle_epi8(_mm_set1_epi32(x32), shuf_maskl); - __m128i bytesh = _mm_shuffle_epi8(_mm_set1_epi32(x32), shuf_maskh); - const __m128i bit_mask = _mm_set1_epi64x(0x7fbfdfeff7fbfdfe); - bytesl = _mm_or_si128(bytesl, bit_mask); - bytesh = _mm_or_si128(bytesh, bit_mask); - bytesl = _mm_cmpeq_epi8(bytesl, _mm_set1_epi64x(-1)); - bytesh = _mm_cmpeq_epi8(bytesh, _mm_set1_epi64x(-1)); - return MM256_SET_M128I(bytesh, bytesl); -} - -// Unpack 32 4-bit fields into 32 bytes -// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval -static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) -{ - // Load 16 bytes from memory - __m128i tmpl = _mm_loadu_si128((const __m128i *)rsi); - __m128i tmph = _mm_srli_epi16(tmpl, 4); - const __m128i lowMask = _mm_set1_epi8(0xF); - tmpl = _mm_and_si128(lowMask, tmpl); - tmph = _mm_and_si128(lowMask, tmph); - return MM256_SET_M128I(tmph, tmpl); -} - -// add int16_t pairwise and return as float vector -static inline __m256 sum_i16_pairs_float(const __m128i xh, const __m128i xl) { - const __m128i ones = _mm_set1_epi16(1); - const __m128i summed_pairsl = _mm_madd_epi16(ones, xl); - const __m128i summed_pairsh = _mm_madd_epi16(ones, xh); - const __m256i summed_pairs = MM256_SET_M128I(summed_pairsh, summed_pairsl); - return _mm256_cvtepi32_ps(summed_pairs); -} - -static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { - const __m128i axl = _mm256_castsi256_si128(ax); - const __m128i axh = _mm256_extractf128_si256(ax, 1); - const __m128i syl = _mm256_castsi256_si128(sy); - const __m128i syh = _mm256_extractf128_si256(sy, 1); - // Perform multiplication and create 16-bit values - const __m128i dotl = _mm_maddubs_epi16(axl, syl); - const __m128i doth = _mm_maddubs_epi16(axh, syh); - return sum_i16_pairs_float(doth, dotl); -} - -// multiply int8_t, add results pairwise twice and return as float vector -static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { - const __m128i xl = _mm256_castsi256_si128(x); - const __m128i xh = _mm256_extractf128_si256(x, 1); - const __m128i yl = _mm256_castsi256_si128(y); - const __m128i yh = _mm256_extractf128_si256(y, 1); - // Get absolute values of x vectors - const __m128i axl = _mm_sign_epi8(xl, xl); - const __m128i axh = _mm_sign_epi8(xh, xh); - // Sign the values of the y vectors - const __m128i syl = _mm_sign_epi8(yl, xl); - const __m128i syh = _mm_sign_epi8(yh, xh); - // Perform multiplication and create 16-bit values - const __m128i dotl = _mm_maddubs_epi16(axl, syl); - const __m128i doth = _mm_maddubs_epi16(axh, syh); - return sum_i16_pairs_float(doth, dotl); -} - -static inline __m128i packNibbles( __m128i bytes1, __m128i bytes2 ) -{ - // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh - const __m128i lowByte = _mm_set1_epi16( 0xFF ); - __m128i high = _mm_andnot_si128( lowByte, bytes1 ); - __m128i low = _mm_and_si128( lowByte, bytes1 ); - high = _mm_srli_epi16( high, 4 ); - bytes1 = _mm_or_si128( low, high ); - high = _mm_andnot_si128( lowByte, bytes2 ); - low = _mm_and_si128( lowByte, bytes2 ); - high = _mm_srli_epi16( high, 4 ); - bytes2 = _mm_or_si128( low, high ); - - return _mm_packus_epi16( bytes1, bytes2); -} - -static inline __m128i mul_add_epi8_sse(const __m128i x, const __m128i y) { - const __m128i ax = _mm_sign_epi8(x, x); - const __m128i sy = _mm_sign_epi8(y, x); - return _mm_maddubs_epi16(ax, sy); -} -#endif -#elif defined(__SSSE3__) -// horizontally add 4x4 floats -static inline float hsum_float_4x4(const __m128 a, const __m128 b, const __m128 c, const __m128 d) { - __m128 res_0 =_mm_hadd_ps(a, b); - __m128 res_1 =_mm_hadd_ps(c, d); - __m128 res =_mm_hadd_ps(res_0, res_1); - res =_mm_hadd_ps(res, res); - res =_mm_hadd_ps(res, res); - - return _mm_cvtss_f32(res); -} -#endif // __AVX__ || __AVX2__ || __AVX512F__ -#endif // defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) - -#if defined(__ARM_NEON) || defined(__wasm_simd128__) || defined(__POWER9_VECTOR__) -#define B1(c,s,n) 0x ## n ## c , 0x ## n ## s -#define B2(c,s,n) B1(c,s,n ## c), B1(c,s,n ## s) -#define B3(c,s,n) B2(c,s,n ## c), B2(c,s,n ## s) -#define B4(c,s,n) B3(c,s,n ## c), B3(c,s,n ## s) -#define B5(c,s,n) B4(c,s,n ## c), B4(c,s,n ## s) -#define B6(c,s,n) B5(c,s,n ## c), B5(c,s,n ## s) -#define B7(c,s,n) B6(c,s,n ## c), B6(c,s,n ## s) -#define B8(c,s ) B7(c,s, c), B7(c,s, s) - -// precomputed tables for expanding 8bits to 8 bytes: -static const uint64_t table_b2b_0[1 << 8] = { B8(00, 10) }; // ( b) << 4 -static const uint64_t table_b2b_1[1 << 8] = { B8(10, 00) }; // (!b) << 4 -#endif - -#if defined(__loongarch_asx) - -#ifdef __clang__ -#define VREGS_PREFIX "$vr" -#define XREGS_PREFIX "$xr" -#else // GCC -#define VREGS_PREFIX "$f" -#define XREGS_PREFIX "$f" -#endif -#define __ALL_REGS "0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31" -// Convert __m128i to __m256i -static inline __m256i ____m256i(__m128i in) { - __m256i out = __lasx_xvldi(0); - __asm__ volatile ( - ".irp i," __ALL_REGS "\n\t" - " .ifc %[out], " XREGS_PREFIX"\\i \n\t" - " .irp j," __ALL_REGS "\n\t" - " .ifc %[in], " VREGS_PREFIX "\\j \n\t" - " xvpermi.q $xr\\i, $xr\\j, 0x20 \n\t" - " .endif \n\t" - " .endr \n\t" - " .endif \n\t" - ".endr \n\t" - : [out] "+f" (out) : [in] "f" (in) - ); - return out; -} -// Convert two __m128i to __m256i -static inline __m256i lasx_set_q(__m128i inhi, __m128i inlo) { - __m256i out; - __asm__ volatile ( - ".irp i," __ALL_REGS "\n\t" - " .ifc %[hi], " VREGS_PREFIX "\\i \n\t" - " .irp j," __ALL_REGS "\n\t" - " .ifc %[lo], " VREGS_PREFIX "\\j \n\t" - " xvpermi.q $xr\\i, $xr\\j, 0x20 \n\t" - " .endif \n\t" - " .endr \n\t" - " .endif \n\t" - ".endr \n\t" - ".ifnc %[out], %[hi] \n\t" - ".irp i," __ALL_REGS "\n\t" - " .ifc %[out], " XREGS_PREFIX "\\i \n\t" - " .irp j," __ALL_REGS "\n\t" - " .ifc %[hi], " VREGS_PREFIX "\\j \n\t" - " xvori.b $xr\\i, $xr\\j, 0 \n\t" - " .endif \n\t" - " .endr \n\t" - " .endif \n\t" - ".endr \n\t" - ".endif \n\t" - : [out] "=f" (out), [hi] "+f" (inhi) - : [lo] "f" (inlo) - ); - return out; -} -// Convert __m256i low part to __m128i -static inline __m128i lasx_extracti128_lo(__m256i in) { - __m128i out; - __asm__ volatile ( - ".ifnc %[out], %[in] \n\t" - ".irp i," __ALL_REGS "\n\t" - " .ifc %[out], " VREGS_PREFIX "\\i \n\t" - " .irp j," __ALL_REGS "\n\t" - " .ifc %[in], " XREGS_PREFIX "\\j \n\t" - " vori.b $vr\\i, $vr\\j, 0 \n\t" - " .endif \n\t" - " .endr \n\t" - " .endif \n\t" - ".endr \n\t" - ".endif \n\t" - : [out] "=f" (out) : [in] "f" (in) - ); - return out; -} -// Convert __m256i high part to __m128i -static inline __m128i lasx_extracti128_hi(__m256i in) { - __m128i out; - __asm__ volatile ( - ".irp i," __ALL_REGS "\n\t" - " .ifc %[out], " VREGS_PREFIX "\\i \n\t" - " .irp j," __ALL_REGS "\n\t" - " .ifc %[in], " XREGS_PREFIX "\\j \n\t" - " xvpermi.q $xr\\i, $xr\\j, 0x11 \n\t" - " .endif \n\t" - " .endr \n\t" - " .endif \n\t" - ".endr \n\t" - : [out] "=f" (out) : [in] "f" (in) - ); - return out; -} - -static __m256i lasx_set_w(int e7, int e6, int e5, int e4, int e3, int e2, int e1, int e0) { - v8i32 __ret = {e0, e1, e2, e3, e4, e5, e6, e7}; - return (__m256i)__ret; -} - -static __m128i lsx_set_w(int32_t a, int32_t b, int32_t c, int32_t d) { - v4i32 __ret = {d, c, b, a}; - return (__m128i)__ret; -} - -static __m256i lasx_set_d(int64_t a, int64_t b, int64_t c, int64_t d) { - v4i64 __ret = {d, c, b, a}; - return (__m256i)__ret; -} - -static __m256i lasx_insertf128( __m128i x, __m128i y) { - return lasx_set_q(x, y); -} - -static __m128i lsx_shuffle_b(__m128i a, __m128i b) { - __m128i mask_f, zero, tmp0, tmp2, mask; - int f = 0x8f; - mask_f = __lsx_vreplgr2vr_b(f); - zero = __lsx_vldi(0); - tmp0 = __lsx_vand_v(b, mask_f); // get mask with low 4 bit and sign bits - tmp0 = __lsx_vori_b(tmp0, 0x10); // make each mask or with 0x10 prepare for positive - mask = __lsx_vsle_b(zero, tmp0); // if mask >= 0, set mask - tmp2 = __lsx_vand_v(tmp0, mask); // maskout the in2 < ones - return __lsx_vshuf_b(a, zero, tmp2); -} - -static __m256i lasx_shuffle_b(__m256i a, __m256i b) { - __m256i mask_f, zero, tmp0, tmp2, mask; - int f = 0x8f; - mask_f = __lasx_xvreplgr2vr_b(f); - zero = __lasx_xvldi(0); - tmp0 = __lasx_xvand_v(b, mask_f); // get mask with low 4 bit and sign bits - tmp0 = __lasx_xvori_b(tmp0, 0x10); // make each mask or with 0x10 prepare for positive - mask = __lasx_xvsle_b(zero, tmp0); // if mask >= 0, set mask - tmp2 = __lasx_xvand_v(tmp0, mask); // maskout the in2 < ones - return __lasx_xvshuf_b(a, zero, tmp2); -} - -static __m256i lasx_extu8_16(__m128i a) { - __m128i zero = __lsx_vldi(0); - __m128i vlo = __lsx_vilvl_b(zero, a); - __m128i vhi = __lsx_vilvh_b(zero, a); - return lasx_set_q(vhi, vlo); -} - -static __m256i lasx_ext8_16(__m128i a) { - __m128i sign = __lsx_vslti_b(a, 0); - __m128i vlo = __lsx_vilvl_b(sign, a); - __m128i vhi = __lsx_vilvh_b(sign, a); - return lasx_set_q(vhi, vlo); -} - -static __m256i lasx_ext16_32(__m128i a) { - __m256i tmp1; - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 0), 0); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 1), 1); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 2), 2); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 3), 3); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 4), 4); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 5), 5); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 6), 6); - tmp1 = __lasx_xvinsgr2vr_w(tmp1, __lsx_vpickve2gr_h(a, 7), 7); - return tmp1; -} - -static __m128i lasx_extracti128( __m256i a, int pos) { - __m128i ret; - if( pos == 0) - { - ret = lasx_extracti128_lo(a); - } else { - ret = lasx_extracti128_hi(a); - } - return ret; -} - -static __m128 lasx_extractf128( __m256 a, int pos) { - __m128 ret; - if( pos == 0) - { - ret = (__m128)lasx_extracti128_lo((__m256i)a); - } else { - ret = (__m128)lasx_extracti128_hi((__m256i)a); - } - return ret; -} - -static __m128i lsx_hadd_h(__m128i a, __m128i b) { - __m128i tmp1 = __lsx_vpickev_h(b, a); - __m128i tmp2 = __lsx_vpickod_h(b, a); - return __lsx_vadd_h(tmp1, tmp2); -} - -static __m128i lsx_hadd_w(__m128i a, __m128i b) { - __m128i tmp1 = __lsx_vpickev_w(b, a); - __m128i tmp2 = __lsx_vpickod_w(b, a); - return __lsx_vadd_w(tmp1, tmp2); -} - -static __m128 lsx_hadd_s(__m128 a, __m128 b) { - __m128 tmp1 = (__m128)__lsx_vpickev_w((__m128i)b, (__m128i)a); - __m128 tmp2 = (__m128)__lsx_vpickod_w((__m128i)b, (__m128i)a); - - return __lsx_vfadd_s(tmp1, tmp2); -} - -static __m256i lasx_maddubs_h(__m256i a, __m256i b) { - __m256i tmp1, tmp2; - tmp1 = __lasx_xvmulwev_h_b(a, b); - tmp2 = __lasx_xvmulwod_h_b(a, b); - return __lasx_xvsadd_h(tmp1, tmp2); -} - -static __m256i lasx_madd_h(__m256i a, __m256i b) { - __m256i tmp1, tmp2; - tmp1 = __lasx_xvmulwev_w_h(a, b); - tmp2 = __lasx_xvmulwod_w_h(a, b); - return __lasx_xvadd_w(tmp1, tmp2); -} - -static __m256i lasx_packs_w(__m256i a, __m256i b) { - __m256i tmp, tmp1; - tmp = __lasx_xvsat_w(a, 15); - tmp1 = __lasx_xvsat_w(b, 15); - return __lasx_xvpickev_h(tmp1, tmp); -} - -static __m256i lasx_packs_h(__m256i a, __m256i b) { - __m256i tmp, tmp1; - tmp = __lasx_xvsat_h(a, 7); - tmp1 = __lasx_xvsat_h(b, 7); - return __lasx_xvpickev_b(tmp1, tmp); -} - -static __m128i lsx_packs_w(__m128i a, __m128i b) { - __m128i tmp, tmp1; - tmp = __lsx_vsat_w(a, 15); - tmp1 = __lsx_vsat_w(b, 15); - return __lsx_vpickev_h(tmp1, tmp); -} - -static __m128i lsx_packs_h(__m128i a, __m128i b) { - __m128i tmp, tmp1; - tmp = __lsx_vsat_h(a, 7); - tmp1 = __lsx_vsat_h(b, 7); - return __lsx_vpickev_b(tmp1, tmp); -} - -static __m128i lsx_packus_h(__m128i a, __m128i b) { - __m128i tmp, tmp1; - tmp = __lsx_vsat_hu(a, 7); - tmp1 = __lsx_vsat_hu(b, 7); - return __lsx_vpickev_b(tmp1, tmp); -} - - -static __m128i lsx_maddubs_h(__m128i a, __m128i b) { - __m128i tmp1, tmp2; - tmp1 = __lsx_vmulwev_h_b(a, b); - tmp2 = __lsx_vmulwod_h_b(a, b); - return __lsx_vsadd_h(tmp1, tmp2); -} - -static __m128i lsx_madd_h(__m128i a, __m128i b) { - __m128i tmp1, tmp2; - tmp1 = __lsx_vmulwev_w_h(a, b); - tmp2 = __lsx_vmulwod_w_h(a, b); - return __lsx_vadd_w(tmp1, tmp2); -} - -// multiply int8_t, add results pairwise twice -static inline __m128i mul_sum_i8_pairs(const __m128i x, const __m128i y) { - // Get absolute values of x vectors - const __m128i ax = __lsx_vsigncov_b(x, x); - // Sign the values of the y vectors - const __m128i sy = __lsx_vsigncov_b(x, y); - // Perform multiplication and create 16-bit values - const __m128i dot = lsx_maddubs_h(ax, sy); - const __m128i ones = __lsx_vreplgr2vr_h(1); - return lsx_madd_h(ones, dot); -} - -// horizontally add 8 floats -static inline float hsum_float_8(const __m256 x) { - __m128 res = lasx_extractf128(x, 1); - ft_union tmp; - res = __lsx_vfadd_s(res, lasx_extractf128(x, 0)); - res = __lsx_vfadd_s(res, (__m128)__lsx_vpickod_d((__m128i)res, (__m128i)res)); - res = __lsx_vfadd_s(res, (__m128)__lsx_vinsgr2vr_w(__lsx_vldi(0), __lsx_vpickve2gr_w(res, 1), 0)); - tmp.i = __lsx_vpickve2gr_w(res, 0); - return tmp.f; -} - -// horizontally add 8 int32_t -static inline int hsum_i32_8(const __m256i a) { - - __m256i tmp1 = __lasx_xvpermi_q(a, a, 0x11); - __m256i tmp2 = __lasx_xvpermi_q(a, a, 0x00); - - __m128i tmp1_128 = lasx_extracti128_lo(tmp1); - __m128i tmp2_128 = lasx_extracti128_lo(tmp2); - - __m128i sum128 = __lsx_vadd_w(tmp1_128, tmp2_128); - - __m128i ev = __lsx_vpickev_w(sum128, sum128); - __m128i od = __lsx_vpickod_w(sum128, sum128); - __m128i sum64 = __lsx_vadd_w(ev, od); - - int sum64_1, sum64_2; - sum64_1 = __lsx_vpickve2gr_w(sum64, 0); - sum64_2 = __lsx_vpickve2gr_w(sum64, 1); - - return sum64_1 + sum64_2; -} - -// horizontally add 4 int32_t -static inline int hsum_i32_4(const __m128i a) { - __m128i ev = __lsx_vpickev_w(a, a); - __m128i od = __lsx_vpickod_w(a, a); - __m128i sum64 = __lsx_vadd_w(ev, od); - - int sum64_1, sum64_2; - sum64_1 = __lsx_vpickve2gr_w(sum64, 0); - sum64_2 = __lsx_vpickve2gr_w(sum64, 1); - - return sum64_1 + sum64_2; -} - -// spread 32 bits to 32 bytes { 0x00, 0xFF } -static inline __m256i bytes_from_bits_32(const uint8_t * x) { - - uint32_t x32; - memcpy(&x32, x, sizeof(uint32_t)); - const __m256i shuf_mask = lasx_set_d( - 0x0303030303030303, 0x0202020202020202, - 0x0101010101010101, 0x0000000000000000); - - __m256i bytes = lasx_shuffle_b(__lasx_xvreplgr2vr_w(x32), shuf_mask); - const __m256i bit_mask = __lasx_xvreplgr2vr_d(0x7fbfdfeff7fbfdfe); - bytes = __lasx_xvor_v(bytes, bit_mask); - return __lasx_xvseq_b(bytes, __lasx_xvreplgr2vr_d(-1)); -} - -// Unpack 32 4-bit fields into 32 bytes -// The output vector contains 32 bytes, each one in [ 0 .. 15 ] interval -static inline __m256i bytes_from_nibbles_32(const uint8_t * rsi) { - const __m128i lo = __lsx_vld((const __m128i *)rsi, 0); - __m128i hi = __lsx_vsrli_h(lo, 4); - return __lasx_xvandi_b(lasx_insertf128(hi, lo), 0xf); -} - -// add int16_t pairwise and return as float vector -static inline __m256 sum_i16_pairs_float(const __m256i x) { - __m256i v = __lasx_xvpackod_h(x, x); - __m256i summed_pairs = __lasx_xvaddwev_w_h(x, v); - return __lasx_xvffint_s_w(summed_pairs); -} - -static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { - // Perform multiplication and create 16-bit values - const __m256i dot = lasx_maddubs_h(ax, sy); - return sum_i16_pairs_float(dot); -} - -// multiply int8_t, add results pairwise twice and return as float vector -static inline __m256 mul_sum_i8_pairs_float(const __m256i x, const __m256i y) { - - // Get absolute values of x vectors - const __m256i ax = __lasx_xvsigncov_b(x, x); - // Sign the values of the y vectors - const __m256i sy = __lasx_xvsigncov_b(x, y); - - return mul_sum_us8_pairs_float(ax, sy); -} - -static inline __m128i packNibbles( __m256i bytes ) { - // Move bits within 16-bit lanes from 0000_abcd_0000_efgh into 0000_0000_abcd_efgh - const __m256i lowByte = __lasx_xvreplgr2vr_h(0xFF); - __m256i high = __lasx_xvandn_v(lowByte, bytes); - __m256i low = __lasx_xvand_v(lowByte, bytes); - high = __lasx_xvsrli_h(high, 4); - bytes = __lasx_xvor_v(low, high); - // Compress uint16_t lanes into bytes - __m128i *r0 = (__m128i *)&bytes; - __m256i tmp_h128 = __lasx_xvpermi_q(bytes, bytes, 0x11); - __m128i *r1 = (__m128i *)&tmp_h128; - - __m128i zero = __lsx_vldi(0); - __m128i tmp, tmp2, tmp3; - - tmp = __lsx_vmax_h(zero, *r0); - tmp2 = __lsx_vsat_hu(tmp, 7); - - tmp = __lsx_vmax_h(zero, *r1); - tmp3 = __lsx_vsat_hu(tmp, 7); - return __lsx_vpickev_b(tmp3, tmp2); -} -#endif //__loongarch_asx - // reference implementation for deterministic creation of model files void quantize_row_q4_0_ref(const float * restrict x, block_q4_0 * restrict y, int64_t k) { static const int qk = QK4_0; @@ -702,11 +65,6 @@ void quantize_row_q4_0_ref(const float * restrict x, block_q4_0 * restrict y, in } } -void quantize_row_q4_0(const float * restrict x, void * restrict y, int64_t k) { - quantize_row_q4_0_ref(x, y, k); -} - - void quantize_row_q4_1_ref(const float * restrict x, block_q4_1 * restrict y, int64_t k) { const int qk = QK4_1; @@ -744,10 +102,6 @@ void quantize_row_q4_1_ref(const float * restrict x, block_q4_1 * restrict y, in } } -void quantize_row_q4_1(const float * restrict x, void * restrict y, int64_t k) { - quantize_row_q4_1_ref(x, y, k); -} - void quantize_row_q5_0_ref(const float * restrict x, block_q5_0 * restrict y, int64_t k) { static const int qk = QK5_0; @@ -792,10 +146,6 @@ void quantize_row_q5_0_ref(const float * restrict x, block_q5_0 * restrict y, in } } -void quantize_row_q5_0(const float * restrict x, void * restrict y, int64_t k) { - quantize_row_q5_0_ref(x, y, k); -} - void quantize_row_q5_1_ref(const float * restrict x, block_q5_1 * restrict y, int64_t k) { const int qk = QK5_1; @@ -840,10 +190,6 @@ void quantize_row_q5_1_ref(const float * restrict x, block_q5_1 * restrict y, in } } -void quantize_row_q5_1(const float * restrict x, void * restrict y, int64_t k) { - quantize_row_q5_1_ref(x, y, k); -} - // reference implementation for deterministic creation of model files void quantize_row_q8_0_ref(const float * restrict x, block_q8_0 * restrict y, int64_t k) { assert(k % QK8_0 == 0); @@ -870,291 +216,6 @@ void quantize_row_q8_0_ref(const float * restrict x, block_q8_0 * restrict y, in } } -void quantize_row_q8_0(const float * restrict x, void * restrict vy, int64_t k) { - assert(QK8_0 == 32); - assert(k % QK8_0 == 0); - const int nb = k / QK8_0; - - block_q8_0 * restrict y = vy; - -#if defined(__ARM_NEON) - for (int i = 0; i < nb; i++) { - float32x4_t srcv [8]; - float32x4_t asrcv[8]; - float32x4_t amaxv[8]; - - for (int j = 0; j < 8; j++) srcv[j] = vld1q_f32(x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = vabsq_f32(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = vmaxq_f32(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = vmaxq_f32(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = vmaxq_f32(amaxv[8*j], amaxv[8*j+4]); - - const float amax = vmaxvq_f32(amaxv[0]); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - for (int j = 0; j < 8; j++) { - const float32x4_t v = vmulq_n_f32(srcv[j], id); - const int32x4_t vi = vcvtnq_s32_f32(v); - - y[i].qs[4*j + 0] = vgetq_lane_s32(vi, 0); - y[i].qs[4*j + 1] = vgetq_lane_s32(vi, 1); - y[i].qs[4*j + 2] = vgetq_lane_s32(vi, 2); - y[i].qs[4*j + 3] = vgetq_lane_s32(vi, 3); - } - } -#elif defined(__wasm_simd128__) - for (int i = 0; i < nb; i++) { - v128_t srcv [8]; - v128_t asrcv[8]; - v128_t amaxv[8]; - - for (int j = 0; j < 8; j++) srcv[j] = wasm_v128_load(x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = wasm_f32x4_abs(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = wasm_f32x4_max(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = wasm_f32x4_max(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = wasm_f32x4_max(amaxv[8*j], amaxv[8*j+4]); - - const float amax = MAX(MAX(wasm_f32x4_extract_lane(amaxv[0], 0), - wasm_f32x4_extract_lane(amaxv[0], 1)), - MAX(wasm_f32x4_extract_lane(amaxv[0], 2), - wasm_f32x4_extract_lane(amaxv[0], 3))); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - for (int j = 0; j < 8; j++) { - const v128_t v = wasm_f32x4_mul(srcv[j], wasm_f32x4_splat(id)); - const v128_t vi = wasm_i32x4_trunc_sat_f32x4(v); - - y[i].qs[4*j + 0] = wasm_i32x4_extract_lane(vi, 0); - y[i].qs[4*j + 1] = wasm_i32x4_extract_lane(vi, 1); - y[i].qs[4*j + 2] = wasm_i32x4_extract_lane(vi, 2); - y[i].qs[4*j + 3] = wasm_i32x4_extract_lane(vi, 3); - } - } -#elif defined(__AVX2__) || defined(__AVX__) - for (int i = 0; i < nb; i++) { - // Load elements into 4 AVX vectors - __m256 v0 = _mm256_loadu_ps( x ); - __m256 v1 = _mm256_loadu_ps( x + 8 ); - __m256 v2 = _mm256_loadu_ps( x + 16 ); - __m256 v3 = _mm256_loadu_ps( x + 24 ); - x += 32; - - // Compute max(abs(e)) for the block - const __m256 signBit = _mm256_set1_ps( -0.0f ); - __m256 maxAbs = _mm256_andnot_ps( signBit, v0 ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v1 ) ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v2 ) ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v3 ) ); - - __m128 max4 = _mm_max_ps( _mm256_extractf128_ps( maxAbs, 1 ), _mm256_castps256_ps128( maxAbs ) ); - max4 = _mm_max_ps( max4, _mm_movehl_ps( max4, max4 ) ); - max4 = _mm_max_ss( max4, _mm_movehdup_ps( max4 ) ); - const float maxScalar = _mm_cvtss_f32( max4 ); - - // Quantize these floats - const float d = maxScalar / 127.f; - y[i].d = GGML_FP32_TO_FP16(d); - const float id = ( maxScalar != 0.0f ) ? 127.f / maxScalar : 0.0f; - const __m256 mul = _mm256_set1_ps( id ); - - // Apply the multiplier - v0 = _mm256_mul_ps( v0, mul ); - v1 = _mm256_mul_ps( v1, mul ); - v2 = _mm256_mul_ps( v2, mul ); - v3 = _mm256_mul_ps( v3, mul ); - - // Round to nearest integer - v0 = _mm256_round_ps( v0, _MM_ROUND_NEAREST ); - v1 = _mm256_round_ps( v1, _MM_ROUND_NEAREST ); - v2 = _mm256_round_ps( v2, _MM_ROUND_NEAREST ); - v3 = _mm256_round_ps( v3, _MM_ROUND_NEAREST ); - - // Convert floats to integers - __m256i i0 = _mm256_cvtps_epi32( v0 ); - __m256i i1 = _mm256_cvtps_epi32( v1 ); - __m256i i2 = _mm256_cvtps_epi32( v2 ); - __m256i i3 = _mm256_cvtps_epi32( v3 ); - -#if defined(__AVX2__) - // Convert int32 to int16 - i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 - i2 = _mm256_packs_epi32( i2, i3 ); // 16, 17, 18, 19, 24, 25, 26, 27, 20, 21, 22, 23, 28, 29, 30, 31 - // Convert int16 to int8 - i0 = _mm256_packs_epi16( i0, i2 ); // 0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27, 4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 - - // We got our precious signed bytes, but the order is now wrong - // These AVX2 pack instructions process 16-byte pieces independently - // The following instruction is fixing the order - const __m256i perm = _mm256_setr_epi32( 0, 4, 1, 5, 2, 6, 3, 7 ); - i0 = _mm256_permutevar8x32_epi32( i0, perm ); - - _mm256_storeu_si256((__m256i *)y[i].qs, i0); -#else - // Since we don't have in AVX some necessary functions, - // we split the registers in half and call AVX2 analogs from SSE - __m128i ni0 = _mm256_castsi256_si128( i0 ); - __m128i ni1 = _mm256_extractf128_si256( i0, 1); - __m128i ni2 = _mm256_castsi256_si128( i1 ); - __m128i ni3 = _mm256_extractf128_si256( i1, 1); - __m128i ni4 = _mm256_castsi256_si128( i2 ); - __m128i ni5 = _mm256_extractf128_si256( i2, 1); - __m128i ni6 = _mm256_castsi256_si128( i3 ); - __m128i ni7 = _mm256_extractf128_si256( i3, 1); - - // Convert int32 to int16 - ni0 = _mm_packs_epi32( ni0, ni1 ); - ni2 = _mm_packs_epi32( ni2, ni3 ); - ni4 = _mm_packs_epi32( ni4, ni5 ); - ni6 = _mm_packs_epi32( ni6, ni7 ); - // Convert int16 to int8 - ni0 = _mm_packs_epi16( ni0, ni2 ); - ni4 = _mm_packs_epi16( ni4, ni6 ); - - _mm_storeu_si128((__m128i *)(y[i].qs + 0), ni0); - _mm_storeu_si128((__m128i *)(y[i].qs + 16), ni4); -#endif - } -#elif defined(__riscv_v_intrinsic) - - size_t vl = __riscv_vsetvl_e32m4(QK8_0); - - for (int i = 0; i < nb; i++) { - // load elements - vfloat32m4_t v_x = __riscv_vle32_v_f32m4(x+i*QK8_0, vl); - - vfloat32m4_t vfabs = __riscv_vfabs_v_f32m4(v_x, vl); - vfloat32m1_t tmp = __riscv_vfmv_v_f_f32m1(0.0f, vl); - vfloat32m1_t vmax = __riscv_vfredmax_vs_f32m4_f32m1(vfabs, tmp, vl); - float amax = __riscv_vfmv_f_s_f32m1_f32(vmax); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - vfloat32m4_t x0 = __riscv_vfmul_vf_f32m4(v_x, id, vl); - - // convert to integer - vint16m2_t vi = __riscv_vfncvt_x_f_w_i16m2(x0, vl); - vint8m1_t vs = __riscv_vncvt_x_x_w_i8m1(vi, vl); - - // store result - __riscv_vse8_v_i8m1(y[i].qs , vs, vl); - } - -#elif defined(__POWER9_VECTOR__) - for (int i = 0; i < nb; i++) { - vector float srcv [8]; - vector float asrcv[8]; - vector float amaxv[8]; - vector signed int vi[8]; - - for (int j = 0; j < 8; j++) srcv[j] = vec_xl(0, x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = vec_abs(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = vec_max(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = vec_max(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = vec_max(amaxv[8*j], amaxv[8*j+4]); - - const float amax = MAX(MAX(vec_extract(amaxv[0], 0), - vec_extract(amaxv[0], 1)), - MAX(vec_extract(amaxv[0], 2), - vec_extract(amaxv[0], 3))); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - const vector float vid = vec_splats(id); - - y[i].d = GGML_FP32_TO_FP16(d); - - for (int j = 0; j < 8; j++) { - const vector float v = vec_round(vec_mul(srcv[j], vid)); - vi[j] = vec_cts(v, 0); - } - vec_xst(vec_pack(vec_pack(vi[0], vi[1]), vec_pack(vi[2], vi[3])), 0, &y[i].qs[0]); - vec_xst(vec_pack(vec_pack(vi[4], vi[5]), vec_pack(vi[6], vi[7])), 16, &y[i].qs[0]); - } - -#elif defined(__loongarch_asx) - for (int i = 0; i < nb; i++) { - ft_union fi; - __m256 v0 = (__m256)__lasx_xvld( x , 0); - __m256 v1 = (__m256)__lasx_xvld( x , 32); - __m256 v2 = (__m256)__lasx_xvld( x , 64); - __m256 v3 = (__m256)__lasx_xvld( x , 96); - x += 32; - - // Compute max(abs(e)) for the block - const __m256 sign_bit = __lasx_xvreplfr2vr_s( -0.0f ); - __m256 max_abs = (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v0 ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v1 ) ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v2 ) ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v3 ) ); - - __m128 max4 = __lsx_vfmax_s( lasx_extractf128( max_abs, 1 ), lasx_extractf128( max_abs , 0) ); - max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vpickod_d((__m128i) max4, (__m128i)max4 ) ); - __m128 tmp = max4; - max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vinsgr2vr_w(tmp, __lsx_vpickve2gr_w( max4, 1 ), 0 )); - fi.i = __lsx_vpickve2gr_w( (__m128i)max4, 0 ); - const float max_scalar = fi.f; - - // Quantize these floats - const float d = max_scalar / 127.f; - y[i].d = GGML_FP32_TO_FP16(d); - const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; - const __m256 mul = (__m256)__lasx_xvreplfr2vr_s( id ); - - // Apply the multiplier - v0 = __lasx_xvfmul_s( v0, mul ); - v1 = __lasx_xvfmul_s( v1, mul ); - v2 = __lasx_xvfmul_s( v2, mul ); - v3 = __lasx_xvfmul_s( v3, mul ); - - // Round to nearest integer - __m256i i0 = __lasx_xvftintrne_w_s( v0 ); - __m256i i1 = __lasx_xvftintrne_w_s( v1 ); - __m256i i2 = __lasx_xvftintrne_w_s( v2 ); - __m256i i3 = __lasx_xvftintrne_w_s( v3 ); - - __m128i ni0 = lasx_extracti128( i0, 0 ); - __m128i ni1 = lasx_extracti128( i0, 1); - __m128i ni2 = lasx_extracti128( i1, 0); - __m128i ni3 = lasx_extracti128( i1, 1); - __m128i ni4 = lasx_extracti128( i2, 0); - __m128i ni5 = lasx_extracti128( i2, 1); - __m128i ni6 = lasx_extracti128( i3, 0); - __m128i ni7 = lasx_extracti128( i3, 1); - - // Convert int32 to int16 - ni0 = lsx_packs_w( ni0, ni1 ); - ni2 = lsx_packs_w( ni2, ni3 ); - ni4 = lsx_packs_w( ni4, ni5 ); - ni6 = lsx_packs_w( ni6, ni7 ); - // Convert int16 to int8 - ni0 = lsx_packs_h( ni0, ni2 ); - ni4 = lsx_packs_h( ni4, ni6 ); - - __lsx_vst(ni0, (__m128i *)(y[i].qs + 0), 0); - __lsx_vst(ni4, (__m128i *)(y[i].qs + 16), 0); - - } -#else - GGML_UNUSED(nb); - // scalar - quantize_row_q8_0_ref(x, y, k); -#endif -} - // reference implementation for deterministic creation of model files void quantize_row_q8_1_ref(const float * restrict x, block_q8_1 * restrict y, int64_t k) { assert(QK8_1 == 32); @@ -1191,334 +252,6 @@ void quantize_row_q8_1_ref(const float * restrict x, block_q8_1 * restrict y, in } } -void quantize_row_q8_1(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK8_1 == 0); - const int nb = k / QK8_1; - - block_q8_1 * restrict y = vy; - -#if defined(__ARM_NEON) - for (int i = 0; i < nb; i++) { - float32x4_t srcv [8]; - float32x4_t asrcv[8]; - float32x4_t amaxv[8]; - - for (int j = 0; j < 8; j++) srcv[j] = vld1q_f32(x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = vabsq_f32(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = vmaxq_f32(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = vmaxq_f32(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = vmaxq_f32(amaxv[8*j], amaxv[8*j+4]); - - const float amax = vmaxvq_f32(amaxv[0]); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - int32x4_t accv = vdupq_n_s32(0); - - for (int j = 0; j < 8; j++) { - const float32x4_t v = vmulq_n_f32(srcv[j], id); - const int32x4_t vi = vcvtnq_s32_f32(v); - - y[i].qs[4*j + 0] = vgetq_lane_s32(vi, 0); - y[i].qs[4*j + 1] = vgetq_lane_s32(vi, 1); - y[i].qs[4*j + 2] = vgetq_lane_s32(vi, 2); - y[i].qs[4*j + 3] = vgetq_lane_s32(vi, 3); - - accv = vaddq_s32(accv, vi); - } - - y[i].s = GGML_FP32_TO_FP16(d * vaddvq_s32(accv)); - } -#elif defined(__wasm_simd128__) - for (int i = 0; i < nb; i++) { - v128_t srcv [8]; - v128_t asrcv[8]; - v128_t amaxv[8]; - - for (int j = 0; j < 8; j++) srcv[j] = wasm_v128_load(x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = wasm_f32x4_abs(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = wasm_f32x4_max(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = wasm_f32x4_max(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = wasm_f32x4_max(amaxv[8*j], amaxv[8*j+4]); - - const float amax = MAX(MAX(wasm_f32x4_extract_lane(amaxv[0], 0), - wasm_f32x4_extract_lane(amaxv[0], 1)), - MAX(wasm_f32x4_extract_lane(amaxv[0], 2), - wasm_f32x4_extract_lane(amaxv[0], 3))); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - v128_t accv = wasm_i32x4_splat(0); - - for (int j = 0; j < 8; j++) { - const v128_t v = wasm_f32x4_mul(srcv[j], wasm_f32x4_splat(id)); - const v128_t vi = wasm_i32x4_trunc_sat_f32x4(v); - - y[i].qs[4*j + 0] = wasm_i32x4_extract_lane(vi, 0); - y[i].qs[4*j + 1] = wasm_i32x4_extract_lane(vi, 1); - y[i].qs[4*j + 2] = wasm_i32x4_extract_lane(vi, 2); - y[i].qs[4*j + 3] = wasm_i32x4_extract_lane(vi, 3); - - accv = wasm_i32x4_add(accv, vi); - } - - y[i].s = GGML_FP32_TO_FP16( - d * (wasm_i32x4_extract_lane(accv, 0) + - wasm_i32x4_extract_lane(accv, 1) + - wasm_i32x4_extract_lane(accv, 2) + - wasm_i32x4_extract_lane(accv, 3))); - } -#elif defined(__AVX2__) || defined(__AVX__) - for (int i = 0; i < nb; i++) { - // Load elements into 4 AVX vectors - __m256 v0 = _mm256_loadu_ps( x ); - __m256 v1 = _mm256_loadu_ps( x + 8 ); - __m256 v2 = _mm256_loadu_ps( x + 16 ); - __m256 v3 = _mm256_loadu_ps( x + 24 ); - x += 32; - - // Compute max(abs(e)) for the block - const __m256 signBit = _mm256_set1_ps( -0.0f ); - __m256 maxAbs = _mm256_andnot_ps( signBit, v0 ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v1 ) ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v2 ) ); - maxAbs = _mm256_max_ps( maxAbs, _mm256_andnot_ps( signBit, v3 ) ); - - __m128 max4 = _mm_max_ps( _mm256_extractf128_ps( maxAbs, 1 ), _mm256_castps256_ps128( maxAbs ) ); - max4 = _mm_max_ps( max4, _mm_movehl_ps( max4, max4 ) ); - max4 = _mm_max_ss( max4, _mm_movehdup_ps( max4 ) ); - const float max_scalar = _mm_cvtss_f32( max4 ); - - // Quantize these floats - const float d = max_scalar / 127.f; - y[i].d = GGML_FP32_TO_FP16(d); - const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; - const __m256 mul = _mm256_set1_ps( id ); - - // Apply the multiplier - v0 = _mm256_mul_ps( v0, mul ); - v1 = _mm256_mul_ps( v1, mul ); - v2 = _mm256_mul_ps( v2, mul ); - v3 = _mm256_mul_ps( v3, mul ); - - // Round to nearest integer - v0 = _mm256_round_ps( v0, _MM_ROUND_NEAREST ); - v1 = _mm256_round_ps( v1, _MM_ROUND_NEAREST ); - v2 = _mm256_round_ps( v2, _MM_ROUND_NEAREST ); - v3 = _mm256_round_ps( v3, _MM_ROUND_NEAREST ); - - // Convert floats to integers - __m256i i0 = _mm256_cvtps_epi32( v0 ); - __m256i i1 = _mm256_cvtps_epi32( v1 ); - __m256i i2 = _mm256_cvtps_epi32( v2 ); - __m256i i3 = _mm256_cvtps_epi32( v3 ); - -#if defined(__AVX2__) - // Compute the sum of the quants and set y[i].s - y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_8(_mm256_add_epi32(_mm256_add_epi32(i0, i1), _mm256_add_epi32(i2, i3)))); - - // Convert int32 to int16 - i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 - i2 = _mm256_packs_epi32( i2, i3 ); // 16, 17, 18, 19, 24, 25, 26, 27, 20, 21, 22, 23, 28, 29, 30, 31 - // Convert int16 to int8 - i0 = _mm256_packs_epi16( i0, i2 ); // 0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27, 4, 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 - - // We got our precious signed bytes, but the order is now wrong - // These AVX2 pack instructions process 16-byte pieces independently - // The following instruction is fixing the order - const __m256i perm = _mm256_setr_epi32( 0, 4, 1, 5, 2, 6, 3, 7 ); - i0 = _mm256_permutevar8x32_epi32( i0, perm ); - - _mm256_storeu_si256((__m256i *)y[i].qs, i0); -#else - // Since we don't have in AVX some necessary functions, - // we split the registers in half and call AVX2 analogs from SSE - __m128i ni0 = _mm256_castsi256_si128( i0 ); - __m128i ni1 = _mm256_extractf128_si256( i0, 1); - __m128i ni2 = _mm256_castsi256_si128( i1 ); - __m128i ni3 = _mm256_extractf128_si256( i1, 1); - __m128i ni4 = _mm256_castsi256_si128( i2 ); - __m128i ni5 = _mm256_extractf128_si256( i2, 1); - __m128i ni6 = _mm256_castsi256_si128( i3 ); - __m128i ni7 = _mm256_extractf128_si256( i3, 1); - - // Compute the sum of the quants and set y[i].s - const __m128i s0 = _mm_add_epi32(_mm_add_epi32(ni0, ni1), _mm_add_epi32(ni2, ni3)); - const __m128i s1 = _mm_add_epi32(_mm_add_epi32(ni4, ni5), _mm_add_epi32(ni6, ni7)); - y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_4(_mm_add_epi32(s0, s1))); - - // Convert int32 to int16 - ni0 = _mm_packs_epi32( ni0, ni1 ); - ni2 = _mm_packs_epi32( ni2, ni3 ); - ni4 = _mm_packs_epi32( ni4, ni5 ); - ni6 = _mm_packs_epi32( ni6, ni7 ); - // Convert int16 to int8 - ni0 = _mm_packs_epi16( ni0, ni2 ); - ni4 = _mm_packs_epi16( ni4, ni6 ); - - _mm_storeu_si128((__m128i *)(y[i].qs + 0), ni0); - _mm_storeu_si128((__m128i *)(y[i].qs + 16), ni4); -#endif - } -#elif defined(__riscv_v_intrinsic) - - size_t vl = __riscv_vsetvl_e32m4(QK8_1); - - for (int i = 0; i < nb; i++) { - // load elements - vfloat32m4_t v_x = __riscv_vle32_v_f32m4(x+i*QK8_1, vl); - - vfloat32m4_t vfabs = __riscv_vfabs_v_f32m4(v_x, vl); - vfloat32m1_t tmp = __riscv_vfmv_v_f_f32m1(0.0, vl); - vfloat32m1_t vmax = __riscv_vfredmax_vs_f32m4_f32m1(vfabs, tmp, vl); - float amax = __riscv_vfmv_f_s_f32m1_f32(vmax); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - - y[i].d = GGML_FP32_TO_FP16(d); - - vfloat32m4_t x0 = __riscv_vfmul_vf_f32m4(v_x, id, vl); - - // convert to integer - vint16m2_t vi = __riscv_vfncvt_x_f_w_i16m2(x0, vl); - vint8m1_t vs = __riscv_vncvt_x_x_w_i8m1(vi, vl); - - // store result - __riscv_vse8_v_i8m1(y[i].qs , vs, vl); - - // compute sum for y[i].s - vint16m1_t tmp2 = __riscv_vmv_v_x_i16m1(0, vl); - vint16m1_t vwrs = __riscv_vwredsum_vs_i8m1_i16m1(vs, tmp2, vl); - - // set y[i].s - int sum = __riscv_vmv_x_s_i16m1_i16(vwrs); - y[i].s = GGML_FP32_TO_FP16(sum*d); - } - -#elif defined(__POWER9_VECTOR__) - for (int i = 0; i < nb; i++) { - vector float srcv [8]; - vector float asrcv[8]; - vector float amaxv[8]; - vector signed int vi[8]; - - for (int j = 0; j < 8; j++) srcv[j] = vec_xl(0, x + i*32 + 4*j); - for (int j = 0; j < 8; j++) asrcv[j] = vec_abs(srcv[j]); - - for (int j = 0; j < 4; j++) amaxv[2*j] = vec_max(asrcv[2*j], asrcv[2*j+1]); - for (int j = 0; j < 2; j++) amaxv[4*j] = vec_max(amaxv[4*j], amaxv[4*j+2]); - for (int j = 0; j < 1; j++) amaxv[8*j] = vec_max(amaxv[8*j], amaxv[8*j+4]); - - const float amax = MAX(MAX(vec_extract(amaxv[0], 0), - vec_extract(amaxv[0], 1)), - MAX(vec_extract(amaxv[0], 2), - vec_extract(amaxv[0], 3))); - - const float d = amax / ((1 << 7) - 1); - const float id = d ? 1.0f/d : 0.0f; - const vector float vid = vec_splats(id); - - y[i].d = GGML_FP32_TO_FP16(d); - - vector int accv = vec_splats(0); - - for (int j = 0; j < 8; j++) { - const vector float v = vec_round(vec_mul(srcv[j], vid)); - vi[j] = vec_cts(v, 0); - - accv = vec_add(accv, vi[j]); - } - vec_xst(vec_pack(vec_pack(vi[0], vi[1]), vec_pack(vi[2], vi[3])), 0, &y[i].qs[0]); - vec_xst(vec_pack(vec_pack(vi[4], vi[5]), vec_pack(vi[6], vi[7])), 16, &y[i].qs[0]); - - accv = vec_add(accv, vec_sld(accv, accv, 4)); - accv = vec_add(accv, vec_sld(accv, accv, 8)); - y[i].s = GGML_FP32_TO_FP16(d * vec_extract(accv, 0)); - } - -#elif defined(__loongarch_asx) - for (int i = 0; i < nb; i++) { - ft_union ft; - __m256 v0 = (__m256)__lasx_xvld( x , 0 ); - __m256 v1 = (__m256)__lasx_xvld( x , 32 ); - __m256 v2 = (__m256)__lasx_xvld( x , 64 ); - __m256 v3 = (__m256)__lasx_xvld( x , 96 ); - x += 32; - - // Compute max(abs(e)) for the block - const __m256 sign_bit = __lasx_xvreplfr2vr_s( -0.0f ); - __m256 max_abs = (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v0 ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v1 ) ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v2 ) ); - max_abs = __lasx_xvfmax_s( max_abs, (__m256)__lasx_xvandn_v( (__m256i)sign_bit, (__m256i)v3 ) ); - - __m128 max4 = __lsx_vfmax_s( lasx_extractf128( max_abs, 1 ), lasx_extractf128( max_abs, 0) ); - max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vpickod_d((__m128i) max4, (__m128i)max4 ) ); - __m128 tmp = max4; - max4 = __lsx_vfmax_s( max4, (__m128)__lsx_vextrins_w((__m128i)tmp, (__m128i)max4, 0x10 )); - ft.i = __lsx_vpickve2gr_w( (__m128i)max4, 0 ); - const float max_scalar = ft.f; - - // Quantize these floats - const float d = max_scalar / 127.f; - y[i].d = GGML_FP32_TO_FP16(d); - const float id = ( max_scalar != 0.0f ) ? 127.f / max_scalar : 0.0f; - const __m256 mul = __lasx_xvreplfr2vr_s( id ); - - // Apply the multiplier - v0 = __lasx_xvfmul_s( v0, mul ); - v1 = __lasx_xvfmul_s( v1, mul ); - v2 = __lasx_xvfmul_s( v2, mul ); - v3 = __lasx_xvfmul_s( v3, mul ); - - // Round to nearest integer - __m256i i0 = __lasx_xvftintrne_w_s( v0 ); - __m256i i1 = __lasx_xvftintrne_w_s( v1 ); - __m256i i2 = __lasx_xvftintrne_w_s( v2 ); - __m256i i3 = __lasx_xvftintrne_w_s( v3 ); - - __m128i ni0 = lasx_extracti128(i0, 0); - __m128i ni1 = lasx_extracti128( i0, 1); - __m128i ni2 = lasx_extracti128( i1, 0); - __m128i ni3 = lasx_extracti128( i1, 1); - __m128i ni4 = lasx_extracti128( i2, 0 ); - __m128i ni5 = lasx_extracti128( i2, 1); - __m128i ni6 = lasx_extracti128( i3, 0); - __m128i ni7 = lasx_extracti128( i3, 1); - - // Compute the sum of the quants and set y[i].s - const __m128i s0 = __lsx_vadd_w(__lsx_vadd_w(ni0, ni1), __lsx_vadd_w(ni2, ni3)); - const __m128i s1 = __lsx_vadd_w(__lsx_vadd_w(ni4, ni5), __lsx_vadd_w(ni6, ni7)); - y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_4(__lsx_vadd_w(s0, s1))); - - // Convert int32 to int16 - ni0 = lsx_packs_w( ni0, ni1 ); - ni2 = lsx_packs_w( ni2, ni3 ); - ni4 = lsx_packs_w( ni4, ni5 ); - ni6 = lsx_packs_w( ni6, ni7 ); - // Convert int16 to int8 - ni0 = lsx_packs_h( ni0, ni2 ); - ni4 = lsx_packs_h( ni4, ni6 ); - - __lsx_vst(ni0, (__m128i *)(y[i].qs + 0), 0); - __lsx_vst(ni4, (__m128i *)(y[i].qs + 16), 0); - } -#else - GGML_UNUSED(nb); - // scalar - quantize_row_q8_1_ref(x, y, k); -#endif -} - void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int64_t k) { static const int qk = QK4_0; @@ -2008,10 +741,6 @@ void dequantize_row_q2_K(const block_q2_K * restrict x, float * restrict y, int6 } } -void quantize_row_q2_K(const float * restrict x, void * restrict vy, int64_t k) { - quantize_row_q2_K_ref(x, vy, k); -} - static float make_qkx3_quants(int n, int nmax, const float * restrict x, const float * restrict weights, uint8_t * restrict L, float * restrict the_min, uint8_t * restrict Laux, float rmin, float rdelta, int nstep, bool use_mad) { @@ -2374,10 +1103,6 @@ void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int6 } } -void quantize_row_q3_K(const float * restrict x, void * restrict vy, int64_t k) { - quantize_row_q3_K_ref(x, vy, k); -} - static void quantize_row_q3_K_impl(const float * restrict x, block_q3_K * restrict y, int64_t n_per_row, const float * restrict quant_weights) { assert(n_per_row % QK_K == 0); const int nb = n_per_row / QK_K; @@ -2576,12 +1301,6 @@ void dequantize_row_q4_K(const block_q4_K * restrict x, float * restrict y, int6 } } -void quantize_row_q4_K(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_q4_K * restrict y = vy; - quantize_row_q4_K_ref(x, y, k); -} - static void quantize_row_q4_K_impl(const float * restrict x, block_q4_K * restrict y, int64_t n_per_row, const float * quant_weights) { assert(n_per_row % QK_K == 0); const int64_t nb = n_per_row / QK_K; @@ -2787,12 +1506,6 @@ void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int6 } } -void quantize_row_q5_K(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_q5_K * restrict y = vy; - quantize_row_q5_K_ref(x, y, k); -} - static void quantize_row_q5_K_impl(const float * restrict x, block_q5_K * restrict y, int64_t n_per_row, const float * quant_weights) { assert(n_per_row % QK_K == 0); const int64_t nb = n_per_row / QK_K; @@ -3005,12 +1718,6 @@ void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int6 } } -void quantize_row_q6_K(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_q6_K * restrict y = vy; - quantize_row_q6_K_ref(x, y, k); -} - static void quantize_row_q6_K_impl(const float * restrict x, block_q6_K * restrict y, int64_t n_per_row, const float * quant_weights) { assert(n_per_row % QK_K == 0); const int64_t nb = n_per_row / QK_K; @@ -3413,33 +2120,20 @@ void quantize_row_tq2_0_ref(const float * restrict x, block_tq2_0 * restrict y, } } -void quantize_row_tq1_0(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_tq1_0 * restrict y = vy; - quantize_row_tq1_0_ref(x, y, k); -} - -void quantize_row_tq2_0(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_tq2_0 * restrict y = vy; - quantize_row_tq2_0_ref(x, y, k); -} - size_t quantize_tq1_0(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { (void)quant_weights; // not used const size_t row_size = ggml_row_size(GGML_TYPE_TQ1_0, n_per_row); - quantize_row_tq1_0(src, dst, (int64_t)nrow*n_per_row); + quantize_row_tq1_0_ref(src, dst, (int64_t)nrow*n_per_row); return nrow * row_size; } size_t quantize_tq2_0(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { (void)quant_weights; // not used const size_t row_size = ggml_row_size(GGML_TYPE_TQ2_0, n_per_row); - quantize_row_tq2_0(src, dst, (int64_t)nrow*n_per_row); + quantize_row_tq2_0_ref(src, dst, (int64_t)nrow*n_per_row); return nrow * row_size; } - void dequantize_row_tq1_0(const block_tq1_0 * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); const int64_t nb = k / QK_K; @@ -3832,9166 +2526,6 @@ void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int6 } } -void quantize_row_q8_K(const float * restrict x, void * restrict y, int64_t k) { - quantize_row_q8_K_ref(x, y, k); -} - -//===================================== Dot products ================================= - -// -// Helper functions -// -#if __AVX__ || __AVX2__ || __AVX512F__ - -// shuffles to pick the required scales in dot products -static inline __m256i get_scale_shuffle_q3k(int i) { - static const uint8_t k_shuffle[128] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, - }; - return _mm256_loadu_si256((const __m256i*)k_shuffle + i); -} -static inline __m256i get_scale_shuffle_k4(int i) { - static const uint8_t k_shuffle[256] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, - 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, - 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, - 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, - 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 - }; - return _mm256_loadu_si256((const __m256i*)k_shuffle + i); -} -static inline __m128i get_scale_shuffle(int i) { - static const uint8_t k_shuffle[128] = { - 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, - 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, - 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, - 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, - 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, - 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, - 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, - 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 - }; - return _mm_loadu_si128((const __m128i*)k_shuffle + i); -} -#elif defined(__loongarch_asx) -// shuffles to pick the required scales in dot products -static inline __m256i get_scale_shuffle_q3k(int i) { - static const uint8_t k_shuffle[128] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, - }; - return __lasx_xvld((const __m256i*)k_shuffle + i, 0); -} -static inline __m256i get_scale_shuffle_k4(int i) { - static const uint8_t k_shuffle[256] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, - 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, - 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, - 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, - 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 - }; - return __lasx_xvld((const __m256i*)k_shuffle + i, 0); -} -static inline __m128i get_scale_shuffle(int i) { - static const uint8_t k_shuffle[128] = { - 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, - 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, - 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, - 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, - 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, - 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, - 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, - 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 - }; - return __lsx_vld((const __m128i*)k_shuffle + i, 0); -} -#endif - -void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - const int qk = QK8_0; - const int nb = n / qk; - - assert(n % qk == 0); -#if defined(__ARM_FEATURE_MATMUL_INT8) - assert((nrc == 2) || (nrc == 1)); -#else - assert(nrc == 1); -#endif - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q4_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; - -#if defined(__ARM_FEATURE_MATMUL_INT8) - if (nrc == 2) { - const block_q4_0 * restrict vx0 = vx; - const block_q4_0 * restrict vx1 = (const block_q4_0 *) ((const uint8_t*)vx + bx); - const block_q8_0 * restrict vy0 = vy; - const block_q8_0 * restrict vy1 = (const block_q8_0 *) ((const uint8_t*)vy + by); - - float32x4_t sumv0 = vdupq_n_f32(0.0f); - - for (int i = 0; i < nb; i++) { - const block_q4_0 * restrict b_x0 = &vx0[i]; - const block_q4_0 * restrict b_x1 = &vx1[i]; - const block_q8_0 * restrict b_y0 = &vy0[i]; - const block_q8_0 * restrict b_y1 = &vy1[i]; - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - const int8x16_t s8b = vdupq_n_s8(0x8); - - const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); - const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); - - // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // sub 8 - const int8x16_t x0_l = vsubq_s8(v0_0l, s8b); - const int8x16_t x0_h = vsubq_s8(v0_0h, s8b); - const int8x16_t x1_l = vsubq_s8(v0_1l, s8b); - const int8x16_t x1_h = vsubq_s8(v0_1h, s8b); - - // load y - const int8x16_t y0_l = vld1q_s8(b_y0->qs); - const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); - const int8x16_t y1_l = vld1q_s8(b_y1->qs); - const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); - - float32_t _scale[4] = { GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), - GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), - GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), - GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d)}; - - float32x4_t scale = vld1q_f32(_scale); - - int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - - int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - - int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - - int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - - sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), - l1, r1)), l2, r2)), l3, r3))), scale); - } - float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); - float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); - - vst1_f32(s, vget_low_f32(sumv2)); - vst1_f32(s + bs, vget_high_f32(sumv2)); - return; - } -#endif - - int ib = 0; - float sumf = 0; - -#if defined(__ARM_FEATURE_SVE) - svfloat32_t sumv0 = svdup_n_f32(0.0f); - svfloat32_t sumv1 = svdup_n_f32(0.0f); - - const int vector_length = ggml_cpu_get_sve_cnt()*8; - - // VLA Implementation using switch case - switch (vector_length) { - case 128: - { - // predicate for activating higher lanes for 4 float32 elements - const svbool_t ph4 = svptrue_pat_b32(SV_VL4); - - for (; ib + 1 < nb; ib += 2) { - const block_q4_0 * restrict x0 = &x[ib + 0]; - const block_q4_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - // load x - const svuint8_t qx0r = svld1rq_u8(svptrue_b8(), x0->qs); - const svuint8_t qx1r = svld1rq_u8(svptrue_b8(), x1->qs); - - // 4-bit -> 8-bit - const svint8_t qx0l = svreinterpret_s8_u8(svand_n_u8_m(svptrue_b8(), qx0r, 0x0F)); - const svint8_t qx0h = svreinterpret_s8_u8(svlsr_n_u8_m(svptrue_b8(), qx0r, 0x04)); - const svint8_t qx1l = svreinterpret_s8_u8(svand_n_u8_m(svptrue_b8(), qx1r, 0x0F)); - const svint8_t qx1h = svreinterpret_s8_u8(svlsr_n_u8_m(svptrue_b8(), qx1r, 0x04)); - - // sub 8 - const svint8_t qx0ls = svsub_n_s8_x(svptrue_b8(), qx0h, 8); - const svint8_t qx0hs = svsub_n_s8_x(svptrue_b8(), qx0l, 8); - const svint8_t qx1ls = svsub_n_s8_x(svptrue_b8(), qx1h, 8); - const svint8_t qx1hs = svsub_n_s8_x(svptrue_b8(), qx1l, 8); - - // load y - const svint8_t qy0h = svld1_s8(svptrue_b8(), y0->qs); - const svint8_t qy0l = svld1_s8(svptrue_b8(), y0->qs + 16); - const svint8_t qy1h = svld1_s8(svptrue_b8(), y1->qs); - const svint8_t qy1l = svld1_s8(svptrue_b8(), y1->qs + 16); - - // dot product - sumv0 = svmla_n_f32_x(ph4, sumv0, svcvt_f32_s32_x(ph4, svadd_x(ph4, - svdot_s32(svdup_n_s32(0), qx0ls, qy0l), - svdot_s32(svdup_n_s32(0), qx0hs, qy0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = svmla_n_f32_x(ph4, sumv1, svcvt_f32_s32_x(ph4, svadd_x(ph4, - svdot_s32(svdup_n_s32(0), qx1ls, qy1l), - svdot_s32(svdup_n_s32(0), qx1hs, qy1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); - } break; - case 256: - { - // predicate for activating higher lanes for 16 int8 elements - const svbool_t ph16 = svptrue_pat_b8(SV_VL16); - // predicate for activating lower lanes for 16 int8 elements - const svbool_t pl16 = svnot_b_z(svptrue_b8(), ph16); - - for (; ib + 1 < nb; ib += 2) { - const block_q4_0 * restrict x0 = &x[ib + 0]; - const block_q4_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - // load x - const svuint8_t qx0r = svld1rq_u8(svptrue_b8(), x0->qs); - const svuint8_t qx1r = svld1rq_u8(svptrue_b8(), x1->qs); - - // 4-bit -> 8-bit - const svint8_t qx0 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx0r, 0x0F), 0x04)); - const svint8_t qx1 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx1r, 0x0F), 0x04)); - - // sub 8 - const svint8_t qx0s = svsub_n_s8_x(svptrue_b8(), qx0, 8); - const svint8_t qx1s = svsub_n_s8_x(svptrue_b8(), qx1, 8); - - // load y - const svint8_t qy0 = svld1_s8(svptrue_b8(), y0->qs); - const svint8_t qy1 = svld1_s8(svptrue_b8(), y1->qs); - - // dot product - sumv0 = svmla_n_f32_x(svptrue_b32(), sumv0, svcvt_f32_s32_x(svptrue_b32(), - svdot_s32(svdup_n_s32(0), qx0s, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = svmla_n_f32_x(svptrue_b32(), sumv1, svcvt_f32_s32_x(svptrue_b32(), - svdot_s32(svdup_n_s32(0), qx1s, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); - } break; - case 512: - { - // predicate for activating higher lanes for 32 int8 elements - const svbool_t ph32 = svptrue_pat_b8(SV_VL32); - - // predicate for activating higher lanes for 16 int8 elements - const svbool_t ph16 = svptrue_pat_b8(SV_VL16); - // predicate for activating lower lanes for 16 int8 elements from first 32 int8 activated lanes - const svbool_t pl16 = svnot_b_z(ph32, ph16); - - for (; ib + 1 < nb; ib += 2) { - const block_q4_0 * restrict x0 = &x[ib + 0]; - const block_q4_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - // load x - const svuint8_t qx0r = svld1rq_u8(ph32, x0->qs); - const svuint8_t qx1r = svld1rq_u8(ph32, x1->qs); - - // 4-bit -> 8-bit - const svint8_t qx0 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx0r, 0x0F), 0x04)); - const svint8_t qx1 = svreinterpret_s8_u8(svlsr_n_u8_m(pl16, svand_n_u8_m(ph16, qx1r, 0x0F), 0x04)); - - // sub 8 - const svint8_t qx0s = svsub_n_s8_x(ph32, qx0, 8); - const svint8_t qx1s = svsub_n_s8_x(ph32, qx1, 8); - - // load y - const svint8_t qy0 = svld1_s8(ph32, y0->qs); - const svint8_t qy1 = svld1_s8(ph32, y1->qs); - - // dot product - sumv0 = svmla_n_f32_x(ph32, sumv0, svcvt_f32_s32_x(ph32, - svdot_s32(svdup_n_s32(0), qx0s, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = svmla_n_f32_x(ph32, sumv1, svcvt_f32_s32_x(ph32, - svdot_s32(svdup_n_s32(0), qx1s, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = svaddv_f32(ph32, svadd_f32_x(ph32, sumv0, sumv1)); - } break; - default: - assert(false && "Unsupported vector length"); - break; - } - -#elif defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); - - for (; ib + 1 < nb; ib += 2) { - const block_q4_0 * restrict x0 = &x[ib + 0]; - const block_q4_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - const int8x16_t s8b = vdupq_n_s8(0x8); - - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); - - // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // sub 8 - const int8x16_t v0_0ls = vsubq_s8(v0_0l, s8b); - const int8x16_t v0_0hs = vsubq_s8(v0_0h, s8b); - const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b); - const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b); - - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); - - // dot product into int32x4_t - const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0l), v0_0hs, v1_0h); - const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1l), v0_1hs, v1_1h); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__AVX2__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - - // Main loop - for (; ib < nb; ++ib) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - - // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. - const __m256i off = _mm256_set1_epi8( 8 ); - qx = _mm256_sub_epi8( qx, off ); - - __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - /* Multiply q with scale and accumulate */ - acc = _mm256_fmadd_ps( d, q, acc ); - } - - sumf = hsum_float_8(acc); -#elif defined(__AVX__) - const __m128i mone = _mm_set1_epi16(1); - - __m256 accum1 = _mm256_setzero_ps(); - __m256 accum2 = _mm256_setzero_ps(); - for (; ib + 1 < nb; ib += 2) { - const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)x[ib + 0].qs); - const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); - const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs); - const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs + 1); - const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); - const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs + 1); - - const __m128i q4b_1_0 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), q4bits_1), _mm_set1_epi8(8)); - const __m128i q4b_1_1 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(q4bits_1, 4)), _mm_set1_epi8(8)); - const __m128i q4b_2_0 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), q4bits_2), _mm_set1_epi8(8)); - const __m128i q4b_2_1 = _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(q4bits_2, 4)), _mm_set1_epi8(8)); - const __m128i p16_1_0 = mul_add_epi8_sse(q4b_1_0, q8b_1_0); - const __m128i p16_1_1 = mul_add_epi8_sse(q4b_1_1, q8b_1_1); - const __m128i p16_2_0 = mul_add_epi8_sse(q4b_2_0, q8b_2_0); - const __m128i p16_2_1 = mul_add_epi8_sse(q4b_2_1, q8b_2_1); - const __m128i p_1_0 = _mm_madd_epi16(p16_1_0, mone); - const __m128i p_1_1 = _mm_madd_epi16(p16_1_1, mone); - const __m128i p_2_0 = _mm_madd_epi16(p16_2_0, mone); - const __m128i p_2_1 = _mm_madd_epi16(p16_2_1, mone); - accum1 = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), - _mm256_cvtepi32_ps(MM256_SET_M128I(p_1_1, p_1_0))), accum1); - accum2 = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), - _mm256_cvtepi32_ps(MM256_SET_M128I(p_2_1, p_2_0))), accum2); - } - - sumf = hsum_float_8(_mm256_add_ps(accum1, accum2)); -#elif defined(__SSSE3__) - // set constants - const __m128i lowMask = _mm_set1_epi8(0xF); - const __m128i off = _mm_set1_epi8(8); - - // Initialize accumulator with zeros - __m128 acc_0 = _mm_setzero_ps(); - __m128 acc_1 = _mm_setzero_ps(); - __m128 acc_2 = _mm_setzero_ps(); - __m128 acc_3 = _mm_setzero_ps(); - - for (; ib + 1 < nb; ib += 2) { - _mm_prefetch(&x[ib] + sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[ib] + sizeof(block_q8_0), _MM_HINT_T0); - - // Compute combined scale for the block 0 and 1 - const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); - - const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[ib].qs); - - __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); - __m128i by_0 = _mm_loadu_si128((const __m128i *)y[ib].qs); - bx_0 = _mm_sub_epi8(bx_0, off); - const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); - - __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); - __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[ib].qs + 16)); - bx_1 = _mm_sub_epi8(bx_1, off); - const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); - - _mm_prefetch(&x[ib] + 2 * sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[ib] + 2 * sizeof(block_q8_0), _MM_HINT_T0); - - // Compute combined scale for the block 2 and 3 - const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[ib + 1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) ); - - const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); - - __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); - __m128i by_2 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); - bx_2 = _mm_sub_epi8(bx_2, off); - const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); - - __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); - __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[ib + 1].qs + 16)); - bx_3 = _mm_sub_epi8(bx_3, off); - const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); - - // Convert int32_t to float - __m128 p0 = _mm_cvtepi32_ps(i32_0); - __m128 p1 = _mm_cvtepi32_ps(i32_1); - __m128 p2 = _mm_cvtepi32_ps(i32_2); - __m128 p3 = _mm_cvtepi32_ps(i32_3); - - // Apply the scale - __m128 p0_d = _mm_mul_ps( d_0_1, p0 ); - __m128 p1_d = _mm_mul_ps( d_0_1, p1 ); - __m128 p2_d = _mm_mul_ps( d_2_3, p2 ); - __m128 p3_d = _mm_mul_ps( d_2_3, p3 ); - - // Acummulate - acc_0 = _mm_add_ps(p0_d, acc_0); - acc_1 = _mm_add_ps(p1_d, acc_1); - acc_2 = _mm_add_ps(p2_d, acc_2); - acc_3 = _mm_add_ps(p3_d, acc_3); - } - - sumf = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); -#elif defined(__riscv_v_intrinsic) - size_t vl = __riscv_vsetvl_e8m1(qk/2); - - for (; ib < nb; ++ib) { - // load elements - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); - - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); - - // mask and store lower part of x, and then upper part - vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - - vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - - // subtract offset - vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 8, vl); - vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 8, vl); - - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - - sumf += sumi*GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d); - } - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed int v0 = vec_splats((int32_t)0); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - const vector signed char v8 = vec_splats((signed char)0x8); - - vector float vsumf0 = vec_splats(0.0f); - -#pragma GCC unroll 8 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl(16, y[ib].qs); - - vector signed char q4x0 = vec_and(qxs, lowMask); - vector signed char q4x1 = vec_sr(qxs, v4); - - q4x0 = vec_sub(q4x0, v8); - q4x1 = vec_sub(q4x1, v8); - - vector signed short qv0 = vec_add(vec_mule(q4x0, q8y0), vec_mulo(q4x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q4x1, q8y1), vec_mulo(q4x1, q8y1)); - - vector signed int vsumi0 = v0; - - vsumi0 = vec_sum4s(qv0, vsumi0); - vsumi0 = vec_sum4s(qv1, vsumi0); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - } - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - // Initialize accumulator with zeros - __m256 acc = (__m256)__lasx_xvldi(0); - - // Main loop - for (; ib < nb; ++ib) { - /* Compute combined scale for the block */ - const __m256 d = __lasx_xvreplfr2vr_s( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - - // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. - const __m256i off = __lasx_xvreplgr2vr_b( 8 ); - qx = __lasx_xvsub_b( qx, off ); - - __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - /* Multiply q with scale and accumulate */ - acc = __lasx_xvfmadd_s( d, q, acc ); - } - - sumf = hsum_float_8(acc); -#elif defined(__loongarch_sx) - // set constants - const __m128i low_mask = __lsx_vreplgr2vr_b(0xF); - const __m128i off = __lsx_vreplgr2vr_b(8); - - // Initialize accumulator with zeros - __m128 acc_0 = __lsx_vldi(0); - __m128 acc_1 = __lsx_vldi(0); - __m128 acc_2 = __lsx_vldi(0); - __m128 acc_3 = __lsx_vldi(0); - - for (; ib + 1 < nb; ib += 2) { - - // Compute combined scale for the block 0 and 1 - const __m128 d_0_1 = __lsx_vreplgr2vr_w( GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d) ); - - const __m128i tmp_0_1 = __lsx_vld((const __m128i *)x[ib].qs, 0); - - __m128i bx_0 = __lsx_vand_v(low_mask, tmp_0_1); - __m128i by_0 = __lsx_vld((const __m128i *)y[ib].qs, 0); - bx_0 = __lsx_vsub_b(bx_0, off); - const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); - - __m128i bx_1 = __lsx_vand_v(low_mask, __lsx_vsrli_d(tmp_0_1, 4)); - __m128i by_1 = __lsx_vld((const __m128i *)(y[ib].qs + 16), 0); - bx_1 = __lsx_vsub_b(bx_1, off); - const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); - - //_mm_prefetch(&x[ib] + 2 * sizeof(block_q4_0), _MM_HINT_T0); - //_mm_prefetch(&y[ib] + 2 * sizeof(block_q8_0), _MM_HINT_T0); - - // Compute combined scale for the block 2 and 3 - const __m128 d_2_3 = __lsx_vreplgr2vr_w( GGML_FP16_TO_FP32(x[ib + 1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) ); - - const __m128i tmp_2_3 = __lsx_vld((const __m128i *)x[ib + 1].qs, 0); - - __m128i bx_2 = __lsx_vand_v(low_mask, tmp_2_3); - __m128i by_2 = __lsx_vld((const __m128i *)y[ib + 1].qs, 0); - bx_2 = __lsx_vsub_b(bx_2, off); - const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); - - __m128i bx_3 = __lsx_vand_v(low_mask, __lsx_vsrli_d(tmp_2_3, 4)); - __m128i by_3 = __lsx_vld((const __m128i *)(y[ib + 1].qs + 16), 0); - bx_3 = __lsx_vsub_b(bx_3, off); - const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); - - // Convert int32_t to float - __m128 p0 = __lsx_vffint_s_w(i32_0); - __m128 p1 = __lsx_vffint_s_w(i32_1); - __m128 p2 = __lsx_vffint_s_w(i32_2); - __m128 p3 = __lsx_vffint_s_w(i32_3); - - // Apply the scale - __m128 p0_d = __lsx_vfmul_s( d_0_1, p0 ); - __m128 p1_d = __lsx_vfmul_s( d_0_1, p1 ); - __m128 p2_d = __lsx_vfmul_s( d_2_3, p2 ); - __m128 p3_d = __lsx_vfmul_s( d_2_3, p3 ); - - // Acummulate - acc_0 = __lsx_vfadd_s(p0_d, acc_0); - acc_1 = __lsx_vfadd_s(p1_d, acc_1); - acc_2 = __lsx_vfadd_s(p2_d, acc_2); - acc_3 = __lsx_vfadd_s(p3_d, acc_3); - } - - sumf = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); -#endif - for (; ib < nb; ++ib) { - int sumi0 = 0; - int sumi1 = 0; - - for (int j = 0; j < qk/2; ++j) { - const int v0 = (x[ib].qs[j] & 0x0F) - 8; - const int v1 = (x[ib].qs[j] >> 4) - 8; - - sumi0 += (v0 * y[ib].qs[j]); - sumi1 += (v1 * y[ib].qs[j + qk/2]); - } - - int sumi = sumi0 + sumi1; - sumf += sumi*GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d); - } - - *s = sumf; -} - -void ggml_vec_dot_q4_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - const int qk = QK8_1; - const int nb = n / qk; - - assert(n % qk == 0); -#if defined(__ARM_FEATURE_MATMUL_INT8) - assert((nrc == 2) || (nrc == 1)); -#else - assert(nrc == 1); -#endif - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q4_1 * restrict x = vx; - const block_q8_1 * restrict y = vy; - -#if defined(__ARM_FEATURE_MATMUL_INT8) - if (nrc == 2) { - const block_q4_1 * restrict vx0 = vx; - const block_q4_1 * restrict vx1 = (const block_q4_1 *) ((const uint8_t*)vx + bx); - const block_q8_1 * restrict vy0 = vy; - const block_q8_1 * restrict vy1 = (const block_q8_1 *) ((const uint8_t*)vy + by); - - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t summs0 = vdupq_n_f32(0.0f); - - for (int i = 0; i < nb; i++) { - const block_q4_1 * restrict b_x0 = &vx0[i]; - const block_q4_1 * restrict b_x1 = &vx1[i]; - const block_q8_1 * restrict b_y0 = &vy0[i]; - const block_q8_1 * restrict b_y1 = &vy1[i]; - - float32_t summs_t[4] = {GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y0->s), - GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y0->s), - GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y1->s), - GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y1->s)}; - summs0 = vaddq_f32(summs0, vld1q_f32(summs_t)); - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - - const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); - const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); - - // 4-bit -> 8-bit - const int8x16_t x0_l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t x0_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t x1_l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t x1_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // load y - const int8x16_t y0_l = vld1q_s8(b_y0->qs); - const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); - const int8x16_t y1_l = vld1q_s8(b_y1->qs); - const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); - - // mmla into int32x4_t - float32_t _scale[4] = {GGML_FP16_TO_FP32(b_x0->d)*b_y0->d, - GGML_FP16_TO_FP32(b_x0->d)*b_y1->d, - GGML_FP16_TO_FP32(b_x1->d)*b_y0->d, - GGML_FP16_TO_FP32(b_x1->d)*b_y1->d}; - float32x4_t scale = vld1q_f32(_scale); - - int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - - int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - - int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - - int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), - l1, r1)), l2, r2)), l3, r3))), scale); - } - - float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); - float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); - sumv2 = vaddq_f32(sumv2, summs0); - - vst1_f32(s, vget_low_f32 (sumv2)); - vst1_f32(s + bs, vget_high_f32(sumv2)); - return; - } -#endif - - int ib = 0; - float sumf = 0; - - // TODO: add WASM SIMD -#if defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); - - float summs = 0; - - for (; ib + 1 < nb; ib += 2) { - const block_q4_1 * restrict x0 = &x[ib + 0]; - const block_q4_1 * restrict x1 = &x[ib + 1]; - const block_q8_1 * restrict y0 = &y[ib + 0]; - const block_q8_1 * restrict y1 = &y[ib + 1]; - - summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s) + GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); - - // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); - - // dot product into int32x4_t - const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0l, v1_0l), v0_0h, v1_0h); - const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1l, v1_1l), v0_1h, v1_1h); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs; -#elif defined(__AVX2__) || defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - - float summs = 0; - - // Main loop - for (; ib < nb; ++ib) { - const float d0 = GGML_FP16_TO_FP32(x[ib].d); - const float d1 = GGML_FP16_TO_FP32(y[ib].d); - - summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); - - const __m256 d0v = _mm256_set1_ps( d0 ); - const __m256 d1v = _mm256_set1_ps( d1 ); - - // Compute combined scales - const __m256 d0d1 = _mm256_mul_ps( d0v, d1v ); - - // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes - const __m256i qx = bytes_from_nibbles_32(x[ib].qs); - const __m256i qy = _mm256_loadu_si256( (const __m256i *)y[ib].qs ); - - const __m256 xy = mul_sum_us8_pairs_float(qx, qy); - - // Accumulate d0*d1*x*y -#if defined(__AVX2__) - acc = _mm256_fmadd_ps( d0d1, xy, acc ); -#else - acc = _mm256_add_ps( _mm256_mul_ps( d0d1, xy ), acc ); -#endif - } - - sumf = hsum_float_8(acc) + summs; -#elif defined(__riscv_v_intrinsic) - size_t vl = __riscv_vsetvl_e8m1(qk/2); - - for (; ib < nb; ++ib) { - // load elements - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); - - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); - - // mask and store lower part of x, and then upper part - vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - - vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); - } - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed int v0 = vec_splats((int32_t)0); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - -#pragma GCC unroll 4 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[ib].m)); - vector float vys = {GGML_FP16_TO_FP32(y[ib].s), 0.0f, 0.0f, 0.0f}; - vsumf0 = vec_madd(vxmin, vys, vsumf0); - - vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl(16, y[ib].qs); - - vector unsigned char q4x0 = (vector unsigned char)vec_and(qxs, lowMask); - vector unsigned char q4x1 = (vector unsigned char)vec_sr(qxs, v4); - - vector signed int vsumi0 = v0; - - vsumi0 = vec_msum(q8y0, q4x0, vsumi0); - vsumi0 = vec_msum(q8y1, q4x1, vsumi0); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - } - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - // Initialize accumulator with zeros - __m256 acc = (__m256)__lasx_xvldi(0); - - float summs = 0; - - // Main loop - for (; ib < nb; ++ib) { - const float d0 = GGML_FP16_TO_FP32(x[ib].d); - const float d1 = GGML_FP16_TO_FP32(y[ib].d); - - summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); - - const __m256 d0v = __lasx_xvreplfr2vr_s( d0 ); - const __m256 d1v = __lasx_xvreplfr2vr_s( d1 ); - - // Compute combined scales - const __m256 d0d1 = __lasx_xvfmul_s( d0v, d1v ); - - // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes - const __m256i qx = bytes_from_nibbles_32(x[ib].qs); - const __m256i qy = __lasx_xvld( (const __m256i *)y[ib].qs, 0); - - const __m256 xy = mul_sum_us8_pairs_float(qx, qy); - - // Accumulate d0*d1*x*y - acc = __lasx_xvfmadd_s( d0d1, xy, acc ); - } - - sumf = hsum_float_8(acc) + summs; -#endif - for (; ib < nb; ++ib) { - int sumi0 = 0; - int sumi1 = 0; - - for (int j = 0; j < qk/2; ++j) { - const int v0 = (x[ib].qs[j] & 0x0F); - const int v1 = (x[ib].qs[j] >> 4); - - sumi0 += (v0 * y[ib].qs[j]); - sumi1 += (v1 * y[ib].qs[j + qk/2]); - } - - int sumi = sumi0 + sumi1; - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); - } - - *s = sumf; -} - -void ggml_vec_dot_q5_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - const int qk = QK8_0; - const int nb = n / qk; - - int ib = 0; - float sumf = 0; - - assert(n % qk == 0); - assert(qk == QK5_0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q5_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; - -#if defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); - - uint32_t qh0; - uint32_t qh1; - - uint64_t tmp0[4]; - uint64_t tmp1[4]; - - for (; ib + 1 < nb; ib += 2) { - const block_q5_0 * restrict x0 = &x[ib]; - const block_q5_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - - // extract the 5th bit via lookup table ((!b) << 4) - memcpy(&qh0, x0->qh, sizeof(qh0)); - memcpy(&qh1, x1->qh, sizeof(qh1)); - - tmp0[0] = table_b2b_1[(qh0 >> 0) & 0xFF]; - tmp0[1] = table_b2b_1[(qh0 >> 8) & 0xFF]; - tmp0[2] = table_b2b_1[(qh0 >> 16) & 0xFF]; - tmp0[3] = table_b2b_1[(qh0 >> 24) ]; - - tmp1[0] = table_b2b_1[(qh1 >> 0) & 0xFF]; - tmp1[1] = table_b2b_1[(qh1 >> 8) & 0xFF]; - tmp1[2] = table_b2b_1[(qh1 >> 16) & 0xFF]; - tmp1[3] = table_b2b_1[(qh1 >> 24) ]; - - const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); - const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); - const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); - const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); - - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); - - // 4-bit -> 8-bit - int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) - const int8x16_t v0_0lf = vsubq_s8(v0_0l, qhl0); - const int8x16_t v0_0hf = vsubq_s8(v0_0h, qhh0); - const int8x16_t v0_1lf = vsubq_s8(v0_1l, qhl1); - const int8x16_t v0_1hf = vsubq_s8(v0_1h, qhh1); - - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), - ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), - ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__wasm_simd128__) - v128_t sumv = wasm_f32x4_splat(0.0f); - - uint32_t qh; - uint64_t tmp[4]; - - // TODO: check if unrolling this is better - for (; ib < nb; ++ib) { - const block_q5_0 * restrict x0 = &x[ib]; - const block_q8_0 * restrict y0 = &y[ib]; - - const v128_t m4b = wasm_i8x16_splat(0x0F); - - // extract the 5th bit - memcpy(&qh, x0->qh, sizeof(qh)); - - tmp[0] = table_b2b_1[(qh >> 0) & 0xFF]; - tmp[1] = table_b2b_1[(qh >> 8) & 0xFF]; - tmp[2] = table_b2b_1[(qh >> 16) & 0xFF]; - tmp[3] = table_b2b_1[(qh >> 24) ]; - - const v128_t qhl = wasm_v128_load(tmp + 0); - const v128_t qhh = wasm_v128_load(tmp + 2); - - const v128_t v0 = wasm_v128_load(x0->qs); - - // 4-bit -> 8-bit - const v128_t v0l = wasm_v128_and (v0, m4b); - const v128_t v0h = wasm_u8x16_shr(v0, 4); - - // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) - const v128_t v0lf = wasm_i8x16_sub(v0l, qhl); - const v128_t v0hf = wasm_i8x16_sub(v0h, qhh); - - // load y - const v128_t v1l = wasm_v128_load(y0->qs); - const v128_t v1h = wasm_v128_load(y0->qs + 16); - - // int8x16 -> int16x8 - const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); - const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); - const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); - const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); - - const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); - const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); - const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); - const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); - - // dot product - sumv = wasm_f32x4_add(sumv, wasm_f32x4_mul(wasm_f32x4_convert_i32x4( - wasm_i32x4_add( - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), - wasm_i32x4_dot_i16x8(v0lfh, v1lh)), - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), - wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), - wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); - } - - sumf = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + - wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3); -#elif defined(__AVX2__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - - // Main loop - for (; ib < nb; ++ib) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - __m256i bxhi = bytes_from_bits_32(x[ib].qh); - bxhi = _mm256_andnot_si256(bxhi, _mm256_set1_epi8((char)0xF0)); - qx = _mm256_or_si256(qx, bxhi); - - __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - /* Multiply q with scale and accumulate */ - acc = _mm256_fmadd_ps(d, q, acc); - } - - sumf = hsum_float_8(acc); -#elif defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - __m128i mask = _mm_set1_epi8((char)0xF0); - - // Main loop - for (; ib < nb; ++ib) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); - - __m256i bx_0 = bytes_from_nibbles_32(x[ib].qs); - const __m256i bxhi = bytes_from_bits_32(x[ib].qh); - __m128i bxhil = _mm256_castsi256_si128(bxhi); - __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); - bxhil = _mm_andnot_si128(bxhil, mask); - bxhih = _mm_andnot_si128(bxhih, mask); - __m128i bxl = _mm256_castsi256_si128(bx_0); - __m128i bxh = _mm256_extractf128_si256(bx_0, 1); - bxl = _mm_or_si128(bxl, bxhil); - bxh = _mm_or_si128(bxh, bxhih); - bx_0 = MM256_SET_M128I(bxh, bxl); - - const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_i8_pairs_float(bx_0, by_0); - - /* Multiply q with scale and accumulate */ - acc = _mm256_add_ps(_mm256_mul_ps(d, q), acc); - } - - sumf = hsum_float_8(acc); -#elif defined(__riscv_v_intrinsic) - uint32_t qh; - - size_t vl = __riscv_vsetvl_e8m1(qk/2); - - // These temporary registers are for masking and shift operations - vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); - vuint32m2_t vt_2 = __riscv_vsll_vv_u32m2(__riscv_vmv_v_x_u32m2(1, vl), vt_1, vl); - - vuint32m2_t vt_3 = __riscv_vsll_vx_u32m2(vt_2, 16, vl); - vuint32m2_t vt_4 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); - - for (; ib < nb; ++ib) { - memcpy(&qh, x[ib].qh, sizeof(uint32_t)); - - // ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; - vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(vt_2, qh, vl); - vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(xha_0, vt_1, vl); - vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); - - // ((qh & (1u << (j + 16))) >> (j + 12)); - vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(vt_3, qh, vl); - vuint32m2_t xhl_1 = __riscv_vsrl_vv_u32m2(xha_1, vt_4, vl); - - // narrowing - vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xhl_0, vl); - vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); - - vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xhl_1, vl); - vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); - - // load - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); - - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); - - vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - - vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); - vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); - - vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - - vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 16, vl); - vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 16, vl); - - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)) * sumi; - } - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector unsigned char v4 = vec_splats((unsigned char)4); - - vector float vsumf0 = vec_splats(0.0f); - -#pragma GCC unroll 4 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector signed long long aux64x2_0 = {(uint64_t)(table_b2b_1[x[ib].qh[0]]), (uint64_t)(table_b2b_1[x[ib].qh[1]])}; - vector signed long long aux64x2_1 = {(uint64_t)(table_b2b_1[x[ib].qh[2]]), (uint64_t)(table_b2b_1[x[ib].qh[3]])}; - - vector signed char qh0 = (vector signed char)aux64x2_0; - vector signed char qh1 = (vector signed char)aux64x2_1; - - vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); - - vector signed char q5x0 = vec_sub(vec_and (qxs, lowMask), qh0); - vector signed char q5x1 = vec_sub(vec_sr(qxs, v4), qh1); - - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl( 16, y[ib].qs); - - vector signed short qv0 = vec_add(vec_mule(q5x0, q8y0), vec_mulo(q5x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q5x1, q8y1), vec_mulo(q5x1, q8y1)); - - qv0 = vec_add(qv0, qv1); - - vector signed int vsumi0 = vec_add(vec_unpackh(qv0), vec_unpackl(qv0)); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - } - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - // Initialize accumulator with zeros - __m256 acc = (__m256)__lasx_xvldi(0); - - // Main loop - for (; ib < nb; ++ib) { - /* Compute combined scale for the block */ - const __m256 d = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); //FIXME - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - __m256i bxhi = bytes_from_bits_32(x[ib].qh); - bxhi = __lasx_xvandn_v(bxhi, __lasx_xvreplgr2vr_b((char)0xF0)); - qx = __lasx_xvor_v(qx, bxhi); - - __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - /* Multiply q with scale and accumulate */ - acc = __lasx_xvfmadd_s(d, q, acc); - } - - sumf = hsum_float_8(acc); -#endif - for (; ib < nb; ++ib) { - uint32_t qh; - memcpy(&qh, x[ib].qh, sizeof(qh)); - - int sumi0 = 0; - int sumi1 = 0; - - for (int j = 0; j < qk/2; ++j) { - const uint8_t xh_0 = ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; - const uint8_t xh_1 = ((qh & (1u << (j + 16))) >> (j + 12)); - - const int32_t x0 = (int8_t)(((x[ib].qs[j] & 0x0F) | xh_0) - 16); - const int32_t x1 = (int8_t)(((x[ib].qs[j] >> 4) | xh_1) - 16); - - sumi0 += (x0 * y[ib].qs[j]); - sumi1 += (x1 * y[ib].qs[j + qk/2]); - } - - int sumi = sumi0 + sumi1; - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)) * sumi; - } - - *s = sumf; -} - -void ggml_vec_dot_q5_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - const int qk = QK8_1; - const int nb = n / qk; - - int ib = 0; - float sumf = 0; - - assert(n % qk == 0); - assert(qk == QK5_1); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q5_1 * restrict x = vx; - const block_q8_1 * restrict y = vy; - -#if defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); - - float summs0 = 0.0f; - float summs1 = 0.0f; - - uint32_t qh0; - uint32_t qh1; - - uint64_t tmp0[4]; - uint64_t tmp1[4]; - - for (; ib + 1 < nb; ib += 2) { - const block_q5_1 * restrict x0 = &x[ib]; - const block_q5_1 * restrict x1 = &x[ib + 1]; - const block_q8_1 * restrict y0 = &y[ib]; - const block_q8_1 * restrict y1 = &y[ib + 1]; - - const uint8x16_t m4b = vdupq_n_u8(0x0F); - - summs0 += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); - summs1 += GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); - - // extract the 5th bit via lookup table ((b) << 4) - memcpy(&qh0, x0->qh, sizeof(qh0)); - memcpy(&qh1, x1->qh, sizeof(qh1)); - - tmp0[0] = table_b2b_0[(qh0 >> 0) & 0xFF]; - tmp0[1] = table_b2b_0[(qh0 >> 8) & 0xFF]; - tmp0[2] = table_b2b_0[(qh0 >> 16) & 0xFF]; - tmp0[3] = table_b2b_0[(qh0 >> 24) ]; - - tmp1[0] = table_b2b_0[(qh1 >> 0) & 0xFF]; - tmp1[1] = table_b2b_0[(qh1 >> 8) & 0xFF]; - tmp1[2] = table_b2b_0[(qh1 >> 16) & 0xFF]; - tmp1[3] = table_b2b_0[(qh1 >> 24) ]; - - const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); - const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); - const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); - const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); - - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); - - // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - - // add high bit - const int8x16_t v0_0lf = vorrq_s8(v0_0l, qhl0); - const int8x16_t v0_0hf = vorrq_s8(v0_0h, qhh0); - const int8x16_t v0_1lf = vorrq_s8(v0_1l, qhl1); - const int8x16_t v0_1hf = vorrq_s8(v0_1h, qhh1); - - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), - ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), - ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs0 + summs1; -#elif defined(__wasm_simd128__) - v128_t sumv = wasm_f32x4_splat(0.0f); - - float summs = 0.0f; - - uint32_t qh; - uint64_t tmp[4]; - - // TODO: check if unrolling this is better - for (; ib < nb; ++ib) { - const block_q5_1 * restrict x0 = &x[ib]; - const block_q8_1 * restrict y0 = &y[ib]; - - summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); - - const v128_t m4b = wasm_i8x16_splat(0x0F); - - // extract the 5th bit - memcpy(&qh, x0->qh, sizeof(qh)); - - tmp[0] = table_b2b_0[(qh >> 0) & 0xFF]; - tmp[1] = table_b2b_0[(qh >> 8) & 0xFF]; - tmp[2] = table_b2b_0[(qh >> 16) & 0xFF]; - tmp[3] = table_b2b_0[(qh >> 24) ]; - - const v128_t qhl = wasm_v128_load(tmp + 0); - const v128_t qhh = wasm_v128_load(tmp + 2); - - const v128_t v0 = wasm_v128_load(x0->qs); - - // 4-bit -> 8-bit - const v128_t v0l = wasm_v128_and (v0, m4b); - const v128_t v0h = wasm_u8x16_shr(v0, 4); - - // add high bit - const v128_t v0lf = wasm_v128_or(v0l, qhl); - const v128_t v0hf = wasm_v128_or(v0h, qhh); - - // load y - const v128_t v1l = wasm_v128_load(y0->qs); - const v128_t v1h = wasm_v128_load(y0->qs + 16); - - // int8x16 -> int16x8 - const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); - const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); - const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); - const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); - - const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); - const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); - const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); - const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); - - // dot product - sumv = wasm_f32x4_add(sumv, - wasm_f32x4_mul(wasm_f32x4_convert_i32x4(wasm_i32x4_add( - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), - wasm_i32x4_dot_i16x8(v0lfh, v1lh)), - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), - wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), - wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); - } - - sumf = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + - wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3) + summs; -#elif defined(__AVX2__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - - float summs = 0.0f; - - // Main loop - for (; ib < nb; ++ib) { - const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d)); - - summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - __m256i bxhi = bytes_from_bits_32(x[ib].qh); - bxhi = _mm256_and_si256(bxhi, _mm256_set1_epi8(0x10)); - qx = _mm256_or_si256(qx, bxhi); - - const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[ib].d)); - const __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_us8_pairs_float(qx, qy); - - acc = _mm256_fmadd_ps(q, _mm256_mul_ps(dx, dy), acc); - } - - sumf = hsum_float_8(acc) + summs; -#elif defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - __m128i mask = _mm_set1_epi8(0x10); - - float summs = 0.0f; - - // Main loop - for (; ib < nb; ++ib) { - const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d)); - - summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); - - __m256i bx_0 = bytes_from_nibbles_32(x[ib].qs); - const __m256i bxhi = bytes_from_bits_32(x[ib].qh); - __m128i bxhil = _mm256_castsi256_si128(bxhi); - __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); - bxhil = _mm_and_si128(bxhil, mask); - bxhih = _mm_and_si128(bxhih, mask); - __m128i bxl = _mm256_castsi256_si128(bx_0); - __m128i bxh = _mm256_extractf128_si256(bx_0, 1); - bxl = _mm_or_si128(bxl, bxhil); - bxh = _mm_or_si128(bxh, bxhih); - bx_0 = MM256_SET_M128I(bxh, bxl); - - const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[ib].d)); - const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_us8_pairs_float(bx_0, by_0); - - acc = _mm256_add_ps(_mm256_mul_ps(q, _mm256_mul_ps(dx, dy)), acc); - } - - sumf = hsum_float_8(acc) + summs; -#elif defined(__riscv_v_intrinsic) - uint32_t qh; - - size_t vl = __riscv_vsetvl_e8m1(qk/2); - - // temporary registers for shift operations - vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); - vuint32m2_t vt_2 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); - - for (; ib < nb; ++ib) { - memcpy(&qh, x[ib].qh, sizeof(uint32_t)); - - // load qh - vuint32m2_t vqh = __riscv_vmv_v_x_u32m2(qh, vl); - - // ((qh >> (j + 0)) << 4) & 0x10; - vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(vqh, vt_1, vl); - vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); - vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(xhl_0, 0x10, vl); - - // ((qh >> (j + 12)) ) & 0x10; - vuint32m2_t xhr_1 = __riscv_vsrl_vv_u32m2(vqh, vt_2, vl); - vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(xhr_1, 0x10, vl); - - // narrowing - vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xha_0, vl); - vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); - - vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xha_1, vl); - vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); - - // load - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[ib].qs, vl); - - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[ib].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[ib].qs+16, vl); - - vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - - vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); - vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); - - vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); - } - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed int v0 = vec_splats((int32_t)0); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - -#pragma GCC unroll 4 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[ib].m)); - vector float vys = {GGML_FP16_TO_FP32(y[ib].s), 0.f, 0.f, 0.f}; - vsumf0 = vec_madd(vxmin, vys, vsumf0); - - vector unsigned long long aux64x2_0 = {(uint64_t)(table_b2b_0[x[ib].qh[0]]), (uint64_t)(table_b2b_0[x[ib].qh[1]])}; - vector unsigned long long aux64x2_1 = {(uint64_t)(table_b2b_0[x[ib].qh[2]]), (uint64_t)(table_b2b_0[x[ib].qh[3]])}; - - vector signed char qh0 = (vector signed char)aux64x2_0; - vector signed char qh1 = (vector signed char)aux64x2_1; - - vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); - - vector unsigned char q5x0 = (vector unsigned char)vec_or(vec_and(qxs, lowMask), qh0); - vector unsigned char q5x1 = (vector unsigned char)vec_or(vec_sr(qxs, v4), qh1); - - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl( 16, y[ib].qs); - - vector signed int vsumi0 = v0; - - vsumi0 = vec_msum(q8y0, q5x0, vsumi0); - vsumi0 = vec_msum(q8y1, q5x1, vsumi0); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - } - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - // Initialize accumulator with zeros - __m256 acc = (__m256)__lasx_xvldi(0); - - float summs = 0.0f; - - // Main loop - for (; ib < nb; ++ib) { - const __m256 dx = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d)); - - summs += GGML_FP16_TO_FP32(x[ib].m) * GGML_FP16_TO_FP32(y[ib].s); - - __m256i qx = bytes_from_nibbles_32(x[ib].qs); - __m256i bxhi = bytes_from_bits_32(x[ib].qh); - bxhi = __lasx_xvand_v(bxhi, __lasx_xvreplgr2vr_b(0x10)); - qx = __lasx_xvor_v(qx, bxhi); - - const __m256 dy = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib].d)); - const __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); - - const __m256 q = mul_sum_us8_pairs_float(qx, qy); - - acc = __lasx_xvfmadd_s(q, __lasx_xvfmul_s(dx, dy), acc); - } - - sumf = hsum_float_8(acc) + summs; -#endif - for (; ib < nb; ++ib) { - uint32_t qh; - memcpy(&qh, x[ib].qh, sizeof(qh)); - - int sumi0 = 0; - int sumi1 = 0; - - for (int j = 0; j < qk/2; ++j) { - const uint8_t xh_0 = ((qh >> (j + 0)) << 4) & 0x10; - const uint8_t xh_1 = ((qh >> (j + 12)) ) & 0x10; - - const int32_t x0 = (x[ib].qs[j] & 0xF) | xh_0; - const int32_t x1 = (x[ib].qs[j] >> 4) | xh_1; - - sumi0 += (x0 * y[ib].qs[j]); - sumi1 += (x1 * y[ib].qs[j + qk/2]); - } - - int sumi = sumi0 + sumi1; - sumf += (GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d))*sumi + GGML_FP16_TO_FP32(x[ib].m)*GGML_FP16_TO_FP32(y[ib].s); - } - - *s = sumf; -} - -void ggml_vec_dot_q8_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - const int qk = QK8_0; - const int nb = n / qk; - - assert(n % qk == 0); -#if defined(__ARM_FEATURE_MATMUL_INT8) - assert((nrc == 2) || (nrc == 1)); -#else - assert(nrc == 1); -#endif - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q8_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; - -#if defined(__ARM_FEATURE_MATMUL_INT8) - if (nrc == 2) { - const block_q8_0 * restrict vx0 = vx; - const block_q8_0 * restrict vx1 = (const block_q8_0 *) ((const uint8_t*)vx + bx); - const block_q8_0 * restrict vy0 = vy; - const block_q8_0 * restrict vy1 = (const block_q8_0 *) ((const uint8_t*)vy + by); - - float32x4_t sumv0 = vdupq_n_f32(0.0f); - - for (int i = 0; i < nb; i++) { - const block_q8_0 * restrict b_x0 = &vx0[i]; - const block_q8_0 * restrict b_y0 = &vy0[i]; - - const block_q8_0 * restrict b_x1 = &vx1[i]; - const block_q8_0 * restrict b_y1 = &vy1[i]; - - const int8x16_t x0_l = vld1q_s8(b_x0->qs); - const int8x16_t x0_h = vld1q_s8(b_x0->qs + 16); - const int8x16_t x1_l = vld1q_s8(b_x1->qs); - const int8x16_t x1_h = vld1q_s8(b_x1->qs + 16); - - // load y - const int8x16_t y0_l = vld1q_s8(b_y0->qs); - const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); - const int8x16_t y1_l = vld1q_s8(b_y1->qs); - const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); - - float32_t _scale[4] = {GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), - GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), - GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), - GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d)}; - float32x4_t scale = vld1q_f32(_scale); - - int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - - int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - - int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - - int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - - sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), - l1, r1)), l2, r2)), l3, r3))), scale); - } - float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); - float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); - - vst1_f32(s, vget_low_f32(sumv2)); - vst1_f32(s + bs, vget_high_f32(sumv2)); - return; - } -#endif - - int ib = 0; - float sumf = 0; - -#if defined(__ARM_FEATURE_SVE) - svfloat32_t sumv0 = svdup_n_f32(0.0f); - svfloat32_t sumv1 = svdup_n_f32(0.0f); - - const int vector_length = ggml_cpu_get_sve_cnt()*8; - - //VLA Implemenation for SVE - switch (vector_length) { - case 128: - { - // predicate for activating lanes for 16 Int8 elements - const svbool_t ph16 = svptrue_pat_b8 (SV_VL16); - const svbool_t pl16 = svptrue_pat_b32(SV_VL4); - - for (; ib + 1 < nb; ib += 2) { - const block_q8_0 * restrict x0 = &x[ib + 0]; - const block_q8_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - // load x - const svint8_t qx0_0 = svld1_s8(ph16, x0->qs); - const svint8_t qx0_1 = svld1_s8(ph16, x0->qs+16); - const svint8_t qx1_0 = svld1_s8(ph16, x1->qs); - const svint8_t qx1_1 = svld1_s8(ph16, x1->qs+16); - - // load y - const svint8_t qy0_0 = svld1_s8(ph16, y0->qs); - const svint8_t qy0_1 = svld1_s8(ph16, y0->qs+16); - const svint8_t qy1_0 = svld1_s8(ph16, y1->qs); - const svint8_t qy1_1 = svld1_s8(ph16, y1->qs+16); - - sumv0 = svmla_n_f32_x(pl16, sumv0, svcvt_f32_s32_x(pl16, svadd_x(pl16, - svdot_s32(svdup_n_s32(0), qx0_0, qy0_0), - svdot_s32(svdup_n_s32(0), qx0_1, qy0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = svmla_n_f32_x(pl16, sumv1, svcvt_f32_s32_x(pl16, svadd_x(pl16, - svdot_s32(svdup_n_s32(0), qx1_0, qy1_0), - svdot_s32(svdup_n_s32(0), qx1_1, qy1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = svaddv_f32(pl16, svadd_f32_x(pl16, sumv0, sumv1)); - } break; - case 256: - { - //printf("sve256"); - for (; ib + 1 < nb; ib += 2) { - const block_q8_0 * restrict x0 = &x[ib + 0]; - const block_q8_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - // load x - const svint8_t qx0 = svld1_s8(svptrue_b8(), x0->qs); - const svint8_t qx1 = svld1_s8(svptrue_b8(), x1->qs); - - // load y - const svint8_t qy0 = svld1_s8(svptrue_b8(), y0->qs); - const svint8_t qy1 = svld1_s8(svptrue_b8(), y1->qs); - - sumv0 = svmla_n_f32_x(svptrue_b32(), sumv0, svcvt_f32_s32_x(svptrue_b32(), - svdot_s32(svdup_n_s32(0), qx0, qy0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = svmla_n_f32_x(svptrue_b32(), sumv1, svcvt_f32_s32_x(svptrue_b32(), - svdot_s32(svdup_n_s32(0), qx1, qy1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = svaddv_f32(svptrue_b32(), svadd_f32_x(svptrue_b32(), sumv0, sumv1)); - } break; - case 512: - { - // predicate for activating high 256 bit - const svbool_t ph32 = svptrue_pat_b8(SV_VL32); - // predicate for activating low 256 bit - const svbool_t pl32 = svnot_b_z(svptrue_b8(), ph32); - - // predicate for activating high lanes for 8 float32 elements - const svbool_t ph8 = svptrue_pat_b32(SV_VL8); - // predicate for activating low lanes for 8 float32 elements - const svbool_t pl8 = svnot_b_z(svptrue_b32(), ph8); - - svfloat32_t sumv00 = svdup_n_f32(0.0f); - - for (; ib + 1 < nb; ib += 2) { - const block_q8_0 * restrict x0 = &x[ib + 0]; - const block_q8_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - //load 32 int8_t in first half of vector and put another 32 int8_t in second vector lower bits - // and add them to make one 64 element vector - // load x - const svint8_t qx_32 = svld1_s8(ph32, x0->qs); - svint8_t qx_64 = svld1_s8(pl32, x0->qs + 2); - - qx_64 = svadd_s8_x(svptrue_b8(), qx_32, qx_64); - - // load y - const svint8_t qy_32 = svld1_s8(ph32, y0->qs); - svint8_t qy_64 = svld1_s8(pl32, y0->qs + 2); - - qy_64 = svadd_s8_x(svptrue_b8(), qy_32, qy_64); - - // scale creation - const float32_t deq1 = GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d); - const float32_t deq2 = GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d); - - // duplicate deq1 in first half of vector and deq2 in second half of vector - const svfloat32_t temp = svdup_f32_m(svdup_f32_z(ph8, deq1), pl8, deq2); - - const svfloat32_t sumvt = svcvt_f32_s32_x(svptrue_b32(), svdot_s32(svdup_n_s32(0), qx_64, qy_64)); - - sumv00 = svmla_f32_m(svptrue_b32(), sumv00, sumvt, temp); - } - - sumf = svaddv_f32(svptrue_b32(), sumv00); - break; - } - default: - assert(false && "Unsupported vector length"); - break; - } -#elif defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); - - for (; ib + 1 < nb; ib += 2) { - const block_q8_0 * restrict x0 = &x[ib + 0]; - const block_q8_0 * restrict x1 = &x[ib + 1]; - const block_q8_0 * restrict y0 = &y[ib + 0]; - const block_q8_0 * restrict y1 = &y[ib + 1]; - - const int8x16_t x0_0 = vld1q_s8(x0->qs); - const int8x16_t x0_1 = vld1q_s8(x0->qs + 16); - const int8x16_t x1_0 = vld1q_s8(x1->qs); - const int8x16_t x1_1 = vld1q_s8(x1->qs + 16); - - // load y - const int8x16_t y0_0 = vld1q_s8(y0->qs); - const int8x16_t y0_1 = vld1q_s8(y0->qs + 16); - const int8x16_t y1_0 = vld1q_s8(y1->qs); - const int8x16_t y1_1 = vld1q_s8(y1->qs + 16); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), x0_0, y0_0), - ggml_vdotq_s32(vdupq_n_s32(0), x0_1, y0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - ggml_vdotq_s32(vdupq_n_s32(0), x1_0, y1_0), - ggml_vdotq_s32(vdupq_n_s32(0), x1_1, y1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); - } - - sumf = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__AVX2__) || defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - - // Main loop - for (; ib < nb; ++ib) { - // Compute combined scale for the block - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); - __m256i qx = _mm256_loadu_si256((const __m256i *)x[ib].qs); - __m256i qy = _mm256_loadu_si256((const __m256i *)y[ib].qs); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - // Multiply q with scale and accumulate -#if defined(__AVX2__) - acc = _mm256_fmadd_ps( d, q, acc ); -#else - acc = _mm256_add_ps( _mm256_mul_ps( d, q ), acc ); -#endif - } - - sumf = hsum_float_8(acc); -#elif defined(__riscv_v_intrinsic) - size_t vl = __riscv_vsetvl_e8m1(qk); - - for (; ib < nb; ++ib) { - // load elements - vint8m1_t bx_0 = __riscv_vle8_v_i8m1(x[ib].qs, vl); - vint8m1_t by_0 = __riscv_vle8_v_i8m1(y[ib].qs, vl); - - vint16m2_t vw_mul = __riscv_vwmul_vv_i16m2(bx_0, by_0, vl); - - vint32m1_t v_zero = __riscv_vmv_v_x_i32m1(0, vl); - vint32m1_t v_sum = __riscv_vwredsum_vs_i16m2_i32m1(vw_mul, v_zero, vl); - - int sumi = __riscv_vmv_x_s_i32m1_i32(v_sum); - - sumf += sumi*(GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)); - } -#elif defined(__POWER9_VECTOR__) - const vector signed int v0 = vec_splats((int32_t)0); - vector float vsumf0 = vec_splats(0.0f); - -#pragma GCC unroll 8 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector signed char q8x0 = vec_xl( 0, x[ib].qs); - vector signed char q8x1 = vec_xl(16, x[ib].qs); - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl(16, y[ib].qs); - - vector signed short qv0 = vec_mule(q8x0, q8y0); - vector signed short qv1 = vec_mulo(q8x0, q8y0); - vector signed short qv2 = vec_mule(q8x1, q8y1); - vector signed short qv3 = vec_mulo(q8x1, q8y1); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - - vsumi0 = vec_sum4s(qv0, vsumi0); - vsumi1 = vec_sum4s(qv1, vsumi1); - vsumi0 = vec_sum4s(qv2, vsumi0); - vsumi1 = vec_sum4s(qv3, vsumi1); - - vsumi0 = vec_add(vsumi0, vsumi1); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - } - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - // Initialize accumulator with zeros - __m256 acc = (__m256)__lasx_xvldi(0); - - // Main loop - for (; ib < nb; ++ib) { - // Compute combined scale for the block - const __m256 d = __lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ib].d) * GGML_FP16_TO_FP32(y[ib].d)); - __m256i qx = __lasx_xvld((const __m256i *)x[ib].qs, 0); - __m256i qy = __lasx_xvld((const __m256i *)y[ib].qs, 0); - - const __m256 q = mul_sum_i8_pairs_float(qx, qy); - - // Multiply q with scale and accumulate - acc = __lasx_xvfmadd_s( d, q, acc ); - } - - sumf = hsum_float_8(acc); -#endif - for (; ib < nb; ++ib) { - int sumi = 0; - - for (int j = 0; j < qk; j++) { - sumi += x[ib].qs[j]*y[ib].qs[j]; - } - - sumf += sumi*(GGML_FP16_TO_FP32(x[ib].d)*GGML_FP16_TO_FP32(y[ib].d)); - } - - *s = sumf; -} - -void ggml_vec_dot_tq1_0_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_tq1_0 * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - float sumf = 0.0f; - - uint8_t k_shift[16] = {1, 1, 1, 1, 3, 3, 3, 3, 9, 9, 9, 9, 27, 27, 27, 27}; - - const uint8x16_t shift = vld1q_u8(k_shift); - - for (int i = 0; i < nb; ++i) { -#if defined(__ARM_FEATURE_DOTPROD) - int32x4_t sumi0 = vdupq_n_s32(0); - int32x4_t sumi1 = vdupq_n_s32(0); -#else - int16x8_t sumi0 = vdupq_n_s16(0); - int16x8_t sumi1 = vdupq_n_s16(0); -#endif - - // first 32 bytes of 5 elements - { - uint8x16_t qx0 = vld1q_u8(x[i].qs + 0); - uint8x16_t qx1 = vld1q_u8(x[i].qs + 16); - uint8x16_t qx2 = vmulq_u8(qx0, vdupq_n_u8(3)); - uint8x16_t qx3 = vmulq_u8(qx1, vdupq_n_u8(3)); - uint8x16_t qx4 = vmulq_u8(qx0, vdupq_n_u8(9)); - uint8x16_t qx5 = vmulq_u8(qx1, vdupq_n_u8(9)); - uint8x16_t qx6 = vmulq_u8(qx0, vdupq_n_u8(27)); - uint8x16_t qx7 = vmulq_u8(qx1, vdupq_n_u8(27)); - uint8x16_t qx8 = vmulq_u8(qx0, vdupq_n_u8(81)); - uint8x16_t qx9 = vmulq_u8(qx1, vdupq_n_u8(81)); - - // multiply by 3 and keep the 2 bits above 8 bits - int8x16_t sqx0 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx0, vshrq_n_u8(qx0, 1)), 6)); - int8x16_t sqx1 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx1, vshrq_n_u8(qx1, 1)), 6)); - int8x16_t sqx2 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx2, vshrq_n_u8(qx2, 1)), 6)); - int8x16_t sqx3 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx3, vshrq_n_u8(qx3, 1)), 6)); - int8x16_t sqx4 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx4, vshrq_n_u8(qx4, 1)), 6)); - int8x16_t sqx5 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx5, vshrq_n_u8(qx5, 1)), 6)); - int8x16_t sqx6 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx6, vshrq_n_u8(qx6, 1)), 6)); - int8x16_t sqx7 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx7, vshrq_n_u8(qx7, 1)), 6)); - int8x16_t sqx8 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx8, vshrq_n_u8(qx8, 1)), 6)); - int8x16_t sqx9 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx9, vshrq_n_u8(qx9, 1)), 6)); - - const int8x16_t qy0 = vld1q_s8(y[i].qs + 0); - const int8x16_t qy1 = vld1q_s8(y[i].qs + 16); - const int8x16_t qy2 = vld1q_s8(y[i].qs + 32); - const int8x16_t qy3 = vld1q_s8(y[i].qs + 48); - const int8x16_t qy4 = vld1q_s8(y[i].qs + 64); - const int8x16_t qy5 = vld1q_s8(y[i].qs + 80); - const int8x16_t qy6 = vld1q_s8(y[i].qs + 96); - const int8x16_t qy7 = vld1q_s8(y[i].qs + 112); - const int8x16_t qy8 = vld1q_s8(y[i].qs + 128); - const int8x16_t qy9 = vld1q_s8(y[i].qs + 144); - -#if defined(__ARM_FEATURE_DOTPROD) - sumi0 = vdotq_s32(sumi0, sqx0, qy0); - sumi1 = vdotq_s32(sumi1, sqx1, qy1); - sumi0 = vdotq_s32(sumi0, sqx2, qy2); - sumi1 = vdotq_s32(sumi1, sqx3, qy3); - sumi0 = vdotq_s32(sumi0, sqx4, qy4); - sumi1 = vdotq_s32(sumi1, sqx5, qy5); - sumi0 = vdotq_s32(sumi0, sqx6, qy6); - sumi1 = vdotq_s32(sumi1, sqx7, qy7); - sumi0 = vdotq_s32(sumi0, sqx8, qy8); - sumi1 = vdotq_s32(sumi1, sqx9, qy9); -#else - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx6), vget_low_s8(qy6)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx6), vget_high_s8(qy6)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx7), vget_low_s8(qy7)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx7), vget_high_s8(qy7)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx8), vget_low_s8(qy8)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx8), vget_high_s8(qy8)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx9), vget_low_s8(qy9)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx9), vget_high_s8(qy9)); -#endif - } - - // last 16 bytes of 5-element, along with the 4 bytes of 4 elements - { - uint8x16_t qx0 = vld1q_u8(x[i].qs + 32); - uint8x16_t qx1 = vmulq_u8(qx0, vdupq_n_u8(3)); - uint8x16_t qx2 = vmulq_u8(qx0, vdupq_n_u8(9)); - uint8x16_t qx3 = vmulq_u8(qx0, vdupq_n_u8(27)); - uint8x16_t qx4 = vmulq_u8(qx0, vdupq_n_u8(81)); - uint32_t qh; - memcpy(&qh, x[i].qh, sizeof(qh)); // potentially unaligned - uint8x16_t qx5 = vreinterpretq_u8_u32(vdupq_n_u32(qh)); - qx5 = vmulq_u8(qx5, shift); - - // multiply by 3 and keep the 2 bits above 8 bits - int8x16_t sqx0 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx0, vshrq_n_u8(qx0, 1)), 6)); - int8x16_t sqx1 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx1, vshrq_n_u8(qx1, 1)), 6)); - int8x16_t sqx2 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx2, vshrq_n_u8(qx2, 1)), 6)); - int8x16_t sqx3 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx3, vshrq_n_u8(qx3, 1)), 6)); - int8x16_t sqx4 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx4, vshrq_n_u8(qx4, 1)), 6)); - int8x16_t sqx5 = vreinterpretq_s8_u8(vshrq_n_u8(vhaddq_u8(qx5, vshrq_n_u8(qx5, 1)), 6)); - - const int8x16_t qy0 = vld1q_s8(y[i].qs + 160); - const int8x16_t qy1 = vld1q_s8(y[i].qs + 176); - const int8x16_t qy2 = vld1q_s8(y[i].qs + 192); - const int8x16_t qy3 = vld1q_s8(y[i].qs + 208); - const int8x16_t qy4 = vld1q_s8(y[i].qs + 224); - const int8x16_t qy5 = vld1q_s8(y[i].qs + 240); - -#if defined(__ARM_FEATURE_DOTPROD) - sumi0 = vdotq_s32(sumi0, sqx0, qy0); - sumi1 = vdotq_s32(sumi1, sqx1, qy1); - sumi0 = vdotq_s32(sumi0, sqx2, qy2); - sumi1 = vdotq_s32(sumi1, sqx3, qy3); - sumi0 = vdotq_s32(sumi0, sqx4, qy4); - sumi1 = vdotq_s32(sumi1, sqx5, qy5); -#else - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); -#endif - } - - const int16x8_t ysum0 = vld1q_s16(y[i].bsums); - const int16x8_t ysum1 = vld1q_s16(y[i].bsums + 8); - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - -#if defined(__ARM_FEATURE_DOTPROD) - sumi0 = vaddq_s32(sumi0, sumi1); - sumi0 = vsubq_s32(sumi0, vpaddlq_s16(vaddq_s16(ysum0, ysum1))); - - sumf += d * (float) vaddvq_s32(sumi0); -#else - sumi0 = vaddq_s16(sumi0, sumi1); - sumi0 = vsubq_s16(sumi0, vaddq_s16(ysum0, ysum1)); - - sumf += d * (float) vaddlvq_s16(sumi0); -#endif - } - - *s = sumf; - -#elif defined(__AVX2__) - __m256 sumf = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - // 16-bit sums - __m256i sumi0 = _mm256_setzero_si256(); - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - - // first 32 bytes of 5 elements - { - __m256i qx0 = _mm256_loadu_si256((const __m256i *) (x[i].qs)); - // 8-bit multiplies with shifts, masks and adds - __m256i qx1 = _mm256_add_epi8(qx0, _mm256_add_epi8(qx0, qx0)); // 1 * 3 - __m256i qx2 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx0, 3), _mm256_set1_epi8(-8)), qx0); // 1 * 9 - __m256i qx3 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx1, 3), _mm256_set1_epi8(-8)), qx1); // 3 * 9 - __m256i qx4 = _mm256_add_epi8(_mm256_and_si256(_mm256_slli_epi16(qx2, 3), _mm256_set1_epi8(-8)), qx2); // 9 * 9 - - // TODO: can _mm256_mulhi_epu16 be faster even if 16-bits? - - // Cancel the +1 from avg so that it behaves like a halving add - qx0 = _mm256_subs_epu8(qx0, _mm256_set1_epi8(1)); - qx1 = _mm256_subs_epu8(qx1, _mm256_set1_epi8(1)); - qx2 = _mm256_subs_epu8(qx2, _mm256_set1_epi8(1)); - qx3 = _mm256_subs_epu8(qx3, _mm256_set1_epi8(1)); - qx4 = _mm256_subs_epu8(qx4, _mm256_set1_epi8(1)); - // Multiply by 3 and get the top 2 bits - qx0 = _mm256_avg_epu8(qx0, _mm256_avg_epu8(qx0, _mm256_setzero_si256())); - qx1 = _mm256_avg_epu8(qx1, _mm256_avg_epu8(qx1, _mm256_setzero_si256())); - qx2 = _mm256_avg_epu8(qx2, _mm256_avg_epu8(qx2, _mm256_setzero_si256())); - qx3 = _mm256_avg_epu8(qx3, _mm256_avg_epu8(qx3, _mm256_setzero_si256())); - qx4 = _mm256_avg_epu8(qx4, _mm256_avg_epu8(qx4, _mm256_setzero_si256())); - qx0 = _mm256_and_si256(_mm256_srli_epi16(qx0, 6), _mm256_set1_epi8(3)); - qx1 = _mm256_and_si256(_mm256_srli_epi16(qx1, 6), _mm256_set1_epi8(3)); - qx2 = _mm256_and_si256(_mm256_srli_epi16(qx2, 6), _mm256_set1_epi8(3)); - qx3 = _mm256_and_si256(_mm256_srli_epi16(qx3, 6), _mm256_set1_epi8(3)); - qx4 = _mm256_and_si256(_mm256_srli_epi16(qx4, 6), _mm256_set1_epi8(3)); - - const __m256i qy0 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 0)); - const __m256i qy1 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 32)); - const __m256i qy2 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 64)); - const __m256i qy3 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 96)); - const __m256i qy4 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 128)); - - qx0 = _mm256_maddubs_epi16(qx0, qy0); - qx1 = _mm256_maddubs_epi16(qx1, qy1); - qx2 = _mm256_maddubs_epi16(qx2, qy2); - qx3 = _mm256_maddubs_epi16(qx3, qy3); - qx4 = _mm256_maddubs_epi16(qx4, qy4); - - sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(qx0, qx1)); - sumi1 = _mm256_add_epi16(sumi1, _mm256_add_epi16(qx2, qx3)); - sumi2 = _mm256_add_epi16(sumi2, qx4); - } - - // last 16 bytes of 5-element, along with the 4 bytes of 4 elements - { - __m128i qx0 = _mm_loadu_si128((const __m128i *) (x[i].qs + 32)); - uint32_t qh; - memcpy(&qh, x[i].qh, sizeof(qh)); // potentially unaligned - __m256i qx5_l = _mm256_cvtepu8_epi16(_mm_set1_epi32(qh)); - __m128i qx1 = _mm_add_epi8(qx0, _mm_add_epi8(qx0, qx0)); // 1 * 3 - __m128i qx2 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx0, 3), _mm_set1_epi8(-8)), qx0); // 1 * 9 - __m128i qx3 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx1, 3), _mm_set1_epi8(-8)), qx1); // 3 * 9 - __m128i qx4 = _mm_add_epi8(_mm_and_si128(_mm_slli_epi16(qx2, 3), _mm_set1_epi8(-8)), qx2); // 9 * 9 - __m256i qx01 = MM256_SET_M128I(qx1, qx0); - __m256i qx23 = MM256_SET_M128I(qx3, qx2); - - // avx2 does not have 8-bit multiplies, so 16-bit it is. - qx5_l = _mm256_mullo_epi16(qx5_l, _mm256_set_epi16(27, 27, 27, 27, 9, 9, 9, 9, 3, 3, 3, 3, 1, 1, 1, 1)); - qx5_l = _mm256_and_si256(qx5_l, _mm256_set1_epi16(0xFF)); - __m128i qx5 = _mm_packus_epi16(_mm256_castsi256_si128(qx5_l), _mm256_extracti128_si256(qx5_l, 1)); - - __m256i qx45 = MM256_SET_M128I(qx5, qx4); - - // Cancel the +1 from avg so that it behaves like a halving add - qx01 = _mm256_subs_epu8(qx01, _mm256_set1_epi8(1)); - qx23 = _mm256_subs_epu8(qx23, _mm256_set1_epi8(1)); - qx45 = _mm256_subs_epu8(qx45, _mm256_set1_epi8(1)); - // Multiply by 3 and get the top 2 bits - qx01 = _mm256_avg_epu8(qx01, _mm256_avg_epu8(qx01, _mm256_setzero_si256())); - qx23 = _mm256_avg_epu8(qx23, _mm256_avg_epu8(qx23, _mm256_setzero_si256())); - qx45 = _mm256_avg_epu8(qx45, _mm256_avg_epu8(qx45, _mm256_setzero_si256())); - qx01 = _mm256_and_si256(_mm256_srli_epi16(qx01, 6), _mm256_set1_epi8(3)); - qx23 = _mm256_and_si256(_mm256_srli_epi16(qx23, 6), _mm256_set1_epi8(3)); - qx45 = _mm256_and_si256(_mm256_srli_epi16(qx45, 6), _mm256_set1_epi8(3)); - - const __m256i qy01 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 160)); - const __m256i qy23 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 192)); - const __m256i qy45 = _mm256_loadu_si256((const __m256i *) (y[i].qs + 224)); - - qx01 = _mm256_maddubs_epi16(qx01, qy01); - qx23 = _mm256_maddubs_epi16(qx23, qy23); - qx45 = _mm256_maddubs_epi16(qx45, qy45); - - sumi0 = _mm256_add_epi16(sumi0, qx01); - sumi1 = _mm256_add_epi16(sumi1, qx23); - sumi2 = _mm256_add_epi16(sumi2, qx45); - } - - const __m256i ysum = _mm256_loadu_si256((const __m256i *) y[i].bsums); - const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(x[i].d)); - - sumi0 = _mm256_sub_epi16(sumi0, ysum); - sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(sumi1, sumi2)); - sumi0 = _mm256_madd_epi16(sumi0, _mm256_set1_epi16(1)); - - sumf = _mm256_add_ps(_mm256_mul_ps(_mm256_cvtepi32_ps(sumi0), d), sumf); - } - - *s = hsum_float_8(sumf); - -#else - const uint8_t pow3[6] = {1, 3, 9, 27, 81, 243}; - - float sumf = 0.0f; - - for (int i = 0; i < nb; ++i) { - int sum = 0; - - for (size_t j = 0; j < sizeof(x->qs) - sizeof(x->qs) % 32; j += 32) { - for (size_t l = 0; l < 5; ++l) { - for (size_t m = 0; m < 32; ++m) { - uint8_t q = x[i].qs[j + m] * pow3[l]; - uint16_t xi = ((uint16_t) q * 3) >> 8; - sum += (xi - 1) * y[i].qs[j*5 + l*32 + m]; - } - } - } - for (size_t j = sizeof(x->qs) - sizeof(x->qs) % 32; j < sizeof(x->qs); j += 16) { - for (size_t l = 0; l < 5; ++l) { - for (size_t m = 0; m < 16; ++m) { - uint8_t q = x[i].qs[j + m] * pow3[l]; - uint16_t xi = ((uint16_t) q * 3) >> 8; - sum += (xi - 1) * y[i].qs[j*5 + l*16 + m]; - } - } - } - - for (size_t l = 0; l < 4; ++l) { - for (size_t j = 0; j < sizeof(x->qh); ++j) { - uint8_t q = x[i].qh[j] * pow3[l]; - uint16_t xi = ((uint16_t) q * 3) >> 8; - sum += (xi - 1) * y[i].qs[sizeof(x->qs)*5 + l*sizeof(x->qh) + j]; - } - } - - sumf += (float) sum * (GGML_FP16_TO_FP32(x[i].d) * y[i].d); - } - - *s = sumf; -#endif -} - -void ggml_vec_dot_tq2_0_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_tq2_0 * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - float sumf = 0.0f; - - const uint8x16_t m3 = vdupq_n_u8(3); - - for (int i = 0; i < nb; ++i) { -#if defined(__ARM_FEATURE_DOTPROD) - int32x4_t sumi0 = vdupq_n_s32(0); - int32x4_t sumi1 = vdupq_n_s32(0); -#else - int16x8_t sumi0 = vdupq_n_s16(0); - int16x8_t sumi1 = vdupq_n_s16(0); -#endif - - for (size_t j = 0; j < sizeof(x->qs); j += 32) { - uint8x16_t qx0 = vld1q_u8(x[i].qs + j); - uint8x16_t qx1 = vld1q_u8(x[i].qs + j + 16); - uint8x16_t qx2 = vshrq_n_u8(qx0, 2); - uint8x16_t qx3 = vshrq_n_u8(qx1, 2); - uint8x16_t qx4 = vshrq_n_u8(qx0, 4); - uint8x16_t qx5 = vshrq_n_u8(qx1, 4); - uint8x16_t qx6 = vshrq_n_u8(qx0, 6); - uint8x16_t qx7 = vshrq_n_u8(qx1, 6); - - int8x16_t sqx0 = vreinterpretq_s8_u8(vandq_u8(qx0, m3)); - int8x16_t sqx1 = vreinterpretq_s8_u8(vandq_u8(qx1, m3)); - int8x16_t sqx2 = vreinterpretq_s8_u8(vandq_u8(qx2, m3)); - int8x16_t sqx3 = vreinterpretq_s8_u8(vandq_u8(qx3, m3)); - int8x16_t sqx4 = vreinterpretq_s8_u8(vandq_u8(qx4, m3)); - int8x16_t sqx5 = vreinterpretq_s8_u8(vandq_u8(qx5, m3)); - int8x16_t sqx6 = vreinterpretq_s8_u8(vandq_u8(qx6, m3)); - int8x16_t sqx7 = vreinterpretq_s8_u8(vandq_u8(qx7, m3)); - - const int8x16_t qy0 = vld1q_s8(y[i].qs + j*4 + 0); - const int8x16_t qy1 = vld1q_s8(y[i].qs + j*4 + 16); - const int8x16_t qy2 = vld1q_s8(y[i].qs + j*4 + 32); - const int8x16_t qy3 = vld1q_s8(y[i].qs + j*4 + 48); - const int8x16_t qy4 = vld1q_s8(y[i].qs + j*4 + 64); - const int8x16_t qy5 = vld1q_s8(y[i].qs + j*4 + 80); - const int8x16_t qy6 = vld1q_s8(y[i].qs + j*4 + 96); - const int8x16_t qy7 = vld1q_s8(y[i].qs + j*4 + 112); - -#if defined(__ARM_FEATURE_DOTPROD) - sumi0 = vdotq_s32(sumi0, sqx0, qy0); - sumi1 = vdotq_s32(sumi1, sqx1, qy1); - sumi0 = vdotq_s32(sumi0, sqx2, qy2); - sumi1 = vdotq_s32(sumi1, sqx3, qy3); - sumi0 = vdotq_s32(sumi0, sqx4, qy4); - sumi1 = vdotq_s32(sumi1, sqx5, qy5); - sumi0 = vdotq_s32(sumi0, sqx6, qy6); - sumi1 = vdotq_s32(sumi1, sqx7, qy7); -#else - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx0), vget_low_s8(qy0)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx0), vget_high_s8(qy0)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx1), vget_low_s8(qy1)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx1), vget_high_s8(qy1)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx2), vget_low_s8(qy2)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx2), vget_high_s8(qy2)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx3), vget_low_s8(qy3)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx3), vget_high_s8(qy3)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx4), vget_low_s8(qy4)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx4), vget_high_s8(qy4)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx5), vget_low_s8(qy5)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx5), vget_high_s8(qy5)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx6), vget_low_s8(qy6)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx6), vget_high_s8(qy6)); - sumi0 = vmlal_s8(sumi0, vget_low_s8(sqx7), vget_low_s8(qy7)); - sumi1 = vmlal_s8(sumi1, vget_high_s8(sqx7), vget_high_s8(qy7)); -#endif - } - - const int16x8_t ysum0 = vld1q_s16(y[i].bsums); - const int16x8_t ysum1 = vld1q_s16(y[i].bsums + 8); - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - -#if defined(__ARM_FEATURE_DOTPROD) - sumi0 = vaddq_s32(sumi0, sumi1); - sumi0 = vsubq_s32(sumi0, vpaddlq_s16(vaddq_s16(ysum0, ysum1))); - - sumf += d * (float) vaddvq_s32(sumi0); -#else - sumi0 = vaddq_s16(sumi0, sumi1); - sumi0 = vsubq_s16(sumi0, vaddq_s16(ysum0, ysum1)); - - sumf += d * (float) vaddlvq_s16(sumi0); -#endif - } - - *s = sumf; - -#elif defined(__AVX2__) - __m256 sumf = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - // 16-bit sums, because 256*127 still fits - __m256i sumi0 = _mm256_setzero_si256(); - __m256i sumi1 = _mm256_setzero_si256(); - - for (size_t j = 0; j < sizeof(x->qs); j += 32) { - __m256i qx0 = _mm256_loadu_si256((const __m256i *) (x[i].qs + j)); - __m256i qx1 = _mm256_srli_epi16(qx0, 2); - __m256i qx2 = _mm256_srli_epi16(qx0, 4); - __m256i qx3 = _mm256_srli_epi16(qx0, 6); - - // 0, 1, 2 (should not be 3) - qx0 = _mm256_and_si256(qx0, _mm256_set1_epi8(3)); - qx1 = _mm256_and_si256(qx1, _mm256_set1_epi8(3)); - qx2 = _mm256_and_si256(qx2, _mm256_set1_epi8(3)); - qx3 = _mm256_and_si256(qx3, _mm256_set1_epi8(3)); - - const __m256i qy0 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 0)); - const __m256i qy1 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 32)); - const __m256i qy2 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 64)); - const __m256i qy3 = _mm256_loadu_si256((const __m256i *) (y[i].qs + j*4 + 96)); - - qx0 = _mm256_maddubs_epi16(qx0, qy0); - qx1 = _mm256_maddubs_epi16(qx1, qy1); - qx2 = _mm256_maddubs_epi16(qx2, qy2); - qx3 = _mm256_maddubs_epi16(qx3, qy3); - - sumi0 = _mm256_add_epi16(sumi0, _mm256_add_epi16(qx0, qx1)); - sumi1 = _mm256_add_epi16(sumi1, _mm256_add_epi16(qx2, qx3)); - } - - const __m256i ysum = _mm256_loadu_si256((const __m256i *) y[i].bsums); - const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(x[i].d)); - - sumi0 = _mm256_add_epi16(sumi0, sumi1); - sumi0 = _mm256_sub_epi16(sumi0, ysum); - sumi0 = _mm256_madd_epi16(sumi0, _mm256_set1_epi16(1)); - - sumf = _mm256_add_ps(_mm256_mul_ps(_mm256_cvtepi32_ps(sumi0), d), sumf); - } - - *s = hsum_float_8(sumf); - -#else - float sumf = 0.0f; - - for (int i = 0; i < nb; ++i) { - int32_t sumi = 0; - - for (size_t j = 0; j < sizeof(x->qs); j += 32) { - for (size_t l = 0; l < 4; ++l) { - for (size_t k = 0; k < 32; ++k) { - sumi += y[i].qs[j*4 + l*32 + k] * (((x[i].qs[j + k] >> (l*2)) & 3) - 1); - } - } - } - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - sumf += (float) sumi * d; - } - - *s = sumf; -#endif -} - -void ggml_vec_dot_q2_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q2_K * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#ifdef __ARM_NEON - const uint8x16_t m3 = vdupq_n_u8(0x3); - const uint8x16_t m4 = vdupq_n_u8(0xF); - - const int32x4_t vzero = vdupq_n_s32(0); - - ggml_int8x16x2_t q2bytes; - uint8_t aux[16]; - - float sum = 0; - - for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - const uint8_t * restrict sc = x[i].scales; - - const uint8x16_t mins_and_scales = vld1q_u8(sc); - const uint8x16_t scales = vandq_u8(mins_and_scales, m4); - vst1q_u8(aux, scales); - - const uint8x16_t mins = vshrq_n_u8(mins_and_scales, 4); - const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); - const ggml_int16x8x2_t mins16 = {{vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(mins))), vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(mins)))}}; - const int32x4_t s0 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[0]), vget_low_s16 (q8sums.val[0])), - vmull_s16(vget_high_s16(mins16.val[0]), vget_high_s16(q8sums.val[0]))); - const int32x4_t s1 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[1]), vget_low_s16 (q8sums.val[1])), - vmull_s16(vget_high_s16(mins16.val[1]), vget_high_s16(q8sums.val[1]))); - sum += dmin * vaddvq_s32(vaddq_s32(s0, s1)); - - int isum = 0; - int is = 0; - -// We use this macro instead of a function call because for some reason -// the code runs 2-3% slower, even if the function is declared inline -#define MULTIPLY_ACCUM_WITH_SCALE(index)\ - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * aux[is+(index)];\ - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * aux[is+1+(index)]; - -#define SHIFT_MULTIPLY_ACCUM_WITH_SCALE(shift, index)\ - q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32;\ - q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[0], (shift)), m3));\ - q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[1], (shift)), m3));\ - MULTIPLY_ACCUM_WITH_SCALE((index)); - - for (int j = 0; j < QK_K/128; ++j) { - const ggml_uint8x16x2_t q2bits = ggml_vld1q_u8_x2(q2); q2 += 32; - - ggml_int8x16x2_t q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; - q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[0], m3)); - q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[1], m3)); - - MULTIPLY_ACCUM_WITH_SCALE(0); - - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(2, 2); - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(4, 4); - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(6, 6); - - is += 8; - } - - sum += d * isum; - } - - *s = sum; - -#elif defined __AVX2__ - - const __m256i m3 = _mm256_set1_epi8(3); - const __m128i m4 = _mm_set1_epi8(0xF); - - __m256 acc = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); - const __m128i scales8 = _mm_and_si128(mins_and_scales, m4); - const __m128i mins8 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); - const __m256i mins = _mm256_cvtepi8_epi16(mins8); - const __m256i prod = _mm256_madd_epi16(mins, _mm256_loadu_si256((const __m256i*)y[i].bsums)); - - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(prod), acc); - - const __m256i all_scales = _mm256_cvtepi8_epi16(scales8); - const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); - const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); - const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; - - __m256i sumi = _mm256_setzero_si256(); - - for (int j = 0; j < QK_K/128; ++j) { - - const __m256i q2bits = _mm256_loadu_si256((const __m256i*)q2); q2 += 32; - - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - const __m256i q2_0 = _mm256_and_si256(q2bits, m3); - const __m256i q2_1 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 2), m3); - const __m256i q2_2 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 4), m3); - const __m256i q2_3 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 6), m3); - - __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); - __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); - __m256i p2 = _mm256_maddubs_epi16(q2_2, q8_2); - __m256i p3 = _mm256_maddubs_epi16(q2_3, q8_3); - - p0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(0)), p0); - p1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(1)), p1); - p2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(2)), p2); - p3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(3)), p3); - - p0 = _mm256_add_epi32(p0, p1); - p2 = _mm256_add_epi32(p2, p3); - - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p0, p2)); - } - - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); - - } - - *s = hsum_float_8(acc); - -#elif defined __AVX__ - - const __m128i m3 = _mm_set1_epi8(0x3); - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m2 = _mm_set1_epi8(0x2); - - __m256 acc = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - // load mins and scales from block_q2_K.scales[QK_K/16] - const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); - const __m128i scales16 = _mm_and_si128(mins_and_scales, m4); - const __m128i mins16 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); - const __m128i mins_0 = _mm_cvtepi8_epi16(mins16); - const __m128i mins_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(mins16, mins16)); - - // summs = y[i].bsums * (x[i].scales >> 4) in 16bits*8*2 to 32bits*4*2 - const __m128i summs_0 = _mm_madd_epi16(mins_0, _mm_loadu_si128((const __m128i*)&y[i].bsums[0])); - const __m128i summs_1 = _mm_madd_epi16(mins_1, _mm_loadu_si128((const __m128i*)&y[i].bsums[8])); - - // sumf += -dmin * summs in 32bits*8 - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(MM256_SET_M128I(summs_1, summs_0))), acc); - - const __m128i scales_0 = _mm_cvtepi8_epi16(scales16); - const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales16, scales16)); - const __m128i scales[2] = { scales_0, scales_1 }; - - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); - - for (int j = 0; j < QK_K/128; ++j) { - - // load Q8 quants int8*16*8 from block_q8_K.qs[QK_K] - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - - // load 2bits*16*8 from block_q2_K.qs[QK_K/4] - __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; - const __m128i q2_0 = _mm_and_si128(q2bits, m3); - const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); - const __m128i q2_4 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); - const __m128i q2_6 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); - q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; - const __m128i q2_1 = _mm_and_si128(q2bits, m3); - const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); - const __m128i q2_5 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); - const __m128i q2_7 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); - - // isuml = q8[l] * ((q2[l] >> shift) & 3) in 8bits*16*8 to 16bits*8*8 - __m128i p0 = _mm_maddubs_epi16(q2_0, q8_0); - __m128i p1 = _mm_maddubs_epi16(q2_1, q8_1); - __m128i p2 = _mm_maddubs_epi16(q2_2, q8_2); - __m128i p3 = _mm_maddubs_epi16(q2_3, q8_3); - __m128i p4 = _mm_maddubs_epi16(q2_4, q8_4); - __m128i p5 = _mm_maddubs_epi16(q2_5, q8_5); - __m128i p6 = _mm_maddubs_epi16(q2_6, q8_6); - __m128i p7 = _mm_maddubs_epi16(q2_7, q8_7); - - // isum += (x[i].scales[is++] & 0xF) * isuml in 16bits*8*8 to 32bits*4*8 - __m128i shuffle = _mm_set1_epi16(0x0100); - p0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p0); - shuffle = _mm_add_epi16(shuffle, m2); - p1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p1); - shuffle = _mm_add_epi16(shuffle, m2); - p2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p2); - shuffle = _mm_add_epi16(shuffle, m2); - p3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p3); - shuffle = _mm_add_epi16(shuffle, m2); - p4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p4); - shuffle = _mm_add_epi16(shuffle, m2); - p5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p5); - shuffle = _mm_add_epi16(shuffle, m2); - p6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p6); - shuffle = _mm_add_epi16(shuffle, m2); - p7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p7); - - p0 = _mm_add_epi32(p0, p1); - p2 = _mm_add_epi32(p2, p3); - p4 = _mm_add_epi32(p4, p5); - p6 = _mm_add_epi32(p6, p7); - - // isum in 32bits*4*2 - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p0, p2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p4, p6)); - } - - // sumf += dall * isum - dmin * summs in 32bits - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dall), _mm256_cvtepi32_ps(sumi)), acc); - } - - *s = hsum_float_8(acc); - -#elif defined __riscv_v_intrinsic - - float sumf = 0; - uint8_t temp_01[32] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; - - for (int i = 0; i < nb; ++i) { - - const uint8_t * q2 = x[i].qs; - const int8_t * q8 = y[i].qs; - const uint8_t * sc = x[i].scales; - - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - size_t vl = 16; - - vuint8m1_t scales = __riscv_vle8_v_u8m1(sc, vl); - vuint8m1_t aux = __riscv_vand_vx_u8m1(scales, 0x0F, vl); - - vint16m1_t q8sums = __riscv_vle16_v_i16m1(y[i].bsums, vl); - - vuint8mf2_t scales_2 = __riscv_vle8_v_u8mf2(sc, vl); - vuint8mf2_t mins8 = __riscv_vsrl_vx_u8mf2(scales_2, 0x4, vl); - vint16m1_t mins = __riscv_vreinterpret_v_u16m1_i16m1(__riscv_vzext_vf2_u16m1(mins8, vl)); - vint32m2_t prod = __riscv_vwmul_vv_i32m2(q8sums, mins, vl); - vint32m1_t vsums = __riscv_vredsum_vs_i32m2_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - - sumf += dmin * __riscv_vmv_x_s_i32m1_i32(vsums); - - vl = 32; - - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t v_b = __riscv_vle8_v_u8m1(temp_01, vl); - - uint8_t is=0; - int isum=0; - - for (int j = 0; j < QK_K/128; ++j) { - // load Q2 - vuint8m1_t q2_x = __riscv_vle8_v_u8m1(q2, vl); - - vuint8m1_t q2_0 = __riscv_vand_vx_u8m1(q2_x, 0x03, vl); - vuint8m1_t q2_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x2, vl), 0x03 , vl); - vuint8m1_t q2_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x4, vl), 0x03 , vl); - vuint8m1_t q2_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x6, vl), 0x03 , vl); - - // duplicate scale elements for product - vuint8m1_t sc0 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 0+is, vl), vl); - vuint8m1_t sc1 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 2+is, vl), vl); - vuint8m1_t sc2 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 4+is, vl), vl); - vuint8m1_t sc3 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 6+is, vl), vl); - - vint16m2_t p0 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_0, sc0, vl)); - vint16m2_t p1 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_1, sc1, vl)); - vint16m2_t p2 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_2, sc2, vl)); - vint16m2_t p3 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_3, sc3, vl)); - - // load Q8 - vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); - vint8m1_t q8_2 = __riscv_vle8_v_i8m1(q8+64, vl); - vint8m1_t q8_3 = __riscv_vle8_v_i8m1(q8+96, vl); - - vint32m4_t s0 = __riscv_vwmul_vv_i32m4(p0, __riscv_vwcvt_x_x_v_i16m2(q8_0, vl), vl); - vint32m4_t s1 = __riscv_vwmul_vv_i32m4(p1, __riscv_vwcvt_x_x_v_i16m2(q8_1, vl), vl); - vint32m4_t s2 = __riscv_vwmul_vv_i32m4(p2, __riscv_vwcvt_x_x_v_i16m2(q8_2, vl), vl); - vint32m4_t s3 = __riscv_vwmul_vv_i32m4(p3, __riscv_vwcvt_x_x_v_i16m2(q8_3, vl), vl); - - vint32m1_t isum0 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s0, s1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s2, s3, vl), isum0, vl); - - isum += __riscv_vmv_x_s_i32m1_i32(isum1); - - q2+=32; q8+=128; is=8; - - } - - sumf += dall * isum; - - } - - *s = sumf; - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0x3); - const vector signed char lowScaleMask = vec_splats((signed char)0xF); - const vector int v0 = vec_splats((int32_t)0); - const vector unsigned char v2 = vec_splats((unsigned char)0x2); - const vector unsigned char v6 = vec_splats((unsigned char)0x6); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); - vector float vdmin = vec_mul(vxmin, vyd); - - vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); - vector signed short q8ysums1 = vec_xl(16, y[i].bsums); - - vector signed char q2xmins = (vector signed char)vec_xl( 0, x[i].scales); - vector signed char vscales = vec_and(q2xmins, lowScaleMask); - - q2xmins = vec_sr(q2xmins, v4); - vector signed short q2xmins0 = vec_unpackh(q2xmins); - vector signed short q2xmins1 = vec_unpackl(q2xmins); - - vector signed int prod0 = vec_mule(q2xmins0, q8ysums0); - vector signed int prod1 = vec_mulo(q2xmins0, q8ysums0); - vector signed int prod2 = vec_mule(q2xmins1, q8ysums1); - vector signed int prod3 = vec_mulo(q2xmins1, q8ysums1); - - vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); - vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); - vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); - vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - vector signed int vsumi4 = v0; - vector signed int vsumi5 = v0; - vector signed int vsumi6 = v0; - vector signed int vsumi7 = v0; - - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/128; ++j) { - __builtin_prefetch(q2, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q2); - vector signed char qxs1 = (vector signed char)vec_xl(16, q2); - q2 += 32; - - vector unsigned char q2x00 = (vector unsigned char)vec_and(qxs0, lowMask); - vector unsigned char q2x01 = (vector unsigned char)vec_and(vec_sr(qxs0, v2), lowMask); - vector unsigned char q2x02 = (vector unsigned char)vec_and(vec_sr(qxs0, v4), lowMask); - vector unsigned char q2x03 = (vector unsigned char)vec_and(vec_sr(qxs0, v6), lowMask); - vector unsigned char q2x10 = (vector unsigned char)vec_and(qxs1, lowMask); - vector unsigned char q2x11 = (vector unsigned char)vec_and(vec_sr(qxs1, v2), lowMask); - vector unsigned char q2x12 = (vector unsigned char)vec_and(vec_sr(qxs1, v4), lowMask); - vector unsigned char q2x13 = (vector unsigned char)vec_and(vec_sr(qxs1, v6), lowMask); - - vector signed char q8y00 = vec_xl( 0, q8); - vector signed char q8y10 = vec_xl( 16, q8); - vector signed char q8y01 = vec_xl( 32, q8); - vector signed char q8y11 = vec_xl( 48, q8); - vector signed char q8y02 = vec_xl( 64, q8); - vector signed char q8y12 = vec_xl( 80, q8); - vector signed char q8y03 = vec_xl( 96, q8); - vector signed char q8y13 = vec_xl(112, q8); - q8 += 128; - - vector signed int qv0 = vec_msum(q8y00, q2x00, v0); - vector signed int qv1 = vec_msum(q8y01, q2x01, v0); - vector signed int qv2 = vec_msum(q8y02, q2x02, v0); - vector signed int qv3 = vec_msum(q8y03, q2x03, v0); - vector signed int qv4 = vec_msum(q8y10, q2x10, v0); - vector signed int qv5 = vec_msum(q8y11, q2x11, v0); - vector signed int qv6 = vec_msum(q8y12, q2x12, v0); - vector signed int qv7 = vec_msum(q8y13, q2x13, v0); - - vector signed short vscales_07 = vec_unpackh(vscales); - vector signed int vscales_03 = vec_unpackh(vscales_07); - vector signed int vscales_47 = vec_unpackl(vscales_07); - vector signed int vs0 = vec_splat(vscales_03, 0); - vector signed int vs1 = vec_splat(vscales_03, 1); - vector signed int vs2 = vec_splat(vscales_03, 2); - vector signed int vs3 = vec_splat(vscales_03, 3); - vector signed int vs4 = vec_splat(vscales_47, 0); - vector signed int vs5 = vec_splat(vscales_47, 1); - vector signed int vs6 = vec_splat(vscales_47, 2); - vector signed int vs7 = vec_splat(vscales_47, 3); - vscales = vec_sld(vscales, vscales, 8); - - vsumi0 = vec_add(vec_mul(qv0, vs0), vsumi0); - vsumi1 = vec_add(vec_mul(qv1, vs2), vsumi1); - vsumi2 = vec_add(vec_mul(qv2, vs4), vsumi2); - vsumi3 = vec_add(vec_mul(qv3, vs6), vsumi3); - vsumi4 = vec_add(vec_mul(qv4, vs1), vsumi4); - vsumi5 = vec_add(vec_mul(qv5, vs3), vsumi5); - vsumi6 = vec_add(vec_mul(qv6, vs5), vsumi6); - vsumi7 = vec_add(vec_mul(qv7, vs7), vsumi7); - } - - vsumi0 = vec_add(vsumi0, vsumi4); - vsumi1 = vec_add(vsumi1, vsumi5); - vsumi2 = vec_add(vsumi2, vsumi6); - vsumi3 = vec_add(vsumi3, vsumi7); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined __loongarch_asx - - const __m256i m3 = __lasx_xvreplgr2vr_b(3); - const __m128i m4 = __lsx_vreplgr2vr_b(0xF); - - __m256 acc = (__m256)__lasx_xvldi(0); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const __m128i mins_and_scales = __lsx_vld((const __m128i*)x[i].scales, 0); - const __m128i scales8 = __lsx_vand_v(mins_and_scales, m4); - const __m128i mins8 = __lsx_vand_v(__lsx_vsrli_h(mins_and_scales, 4), m4); - const __m256i mins = lasx_ext8_16(mins8); - const __m256i prod = lasx_madd_h(mins, __lasx_xvld((const __m256i*)y[i].bsums, 0)); - - acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(dmin), __lasx_xvffint_s_w(prod), acc); - - const __m256i all_scales = lasx_ext8_16(scales8); - const __m128i l_scales = lasx_extracti128(all_scales, 0); - const __m128i h_scales = lasx_extracti128(all_scales, 1); - const __m256i scales[2] = {lasx_insertf128(l_scales, l_scales), lasx_insertf128(h_scales, h_scales)}; - - __m256i sumi = __lasx_xvldi(0); - - for (int j = 0; j < QK_K/128; ++j) { - - const __m256i q2bits = __lasx_xvld((const __m256i*)q2, 0); q2 += 32; - - const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - - const __m256i q2_0 = __lasx_xvand_v(q2bits, m3); - const __m256i q2_1 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 2), m3); - const __m256i q2_2 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 4), m3); - const __m256i q2_3 = __lasx_xvand_v(__lasx_xvsrli_h(q2bits, 6), m3); - - __m256i p0 = lasx_maddubs_h(q2_0, q8_0); - __m256i p1 = lasx_maddubs_h(q2_1, q8_1); - __m256i p2 = lasx_maddubs_h(q2_2, q8_2); - __m256i p3 = lasx_maddubs_h(q2_3, q8_3); - - p0 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(0)), p0); - p1 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(1)), p1); - p2 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(2)), p2); - p3 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(3)), p3); - - p0 = __lasx_xvadd_w(p0, p1); - p2 = __lasx_xvadd_w(p2, p3); - - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p0, p2)); - } - - acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc); - - } - - *s = hsum_float_8(acc); - -#else - - float sumf = 0; - - for (int i = 0; i < nb; ++i) { - - const uint8_t * q2 = x[i].qs; - const int8_t * q8 = y[i].qs; - const uint8_t * sc = x[i].scales; - - int summs = 0; - for (int j = 0; j < 16; ++j) { - summs += y[i].bsums[j] * (sc[j] >> 4); - } - - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - int isum = 0; - int is = 0; - int d; - for (int k = 0; k < QK_K/128; ++k) { - int shift = 0; - for (int j = 0; j < 4; ++j) { - d = sc[is++] & 0xF; - int isuml = 0; - for (int l = 0; l < 16; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); - isum += d * isuml; - d = sc[is++] & 0xF; - isuml = 0; - for (int l = 16; l < 32; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); - isum += d * isuml; - shift += 2; - q8 += 32; - } - q2 += 32; - } - sumf += dall * isum - dmin * summs; - } - *s = sumf; -#endif -} - -void ggml_vec_dot_q3_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const uint32_t kmask1 = 0x03030303; - const uint32_t kmask2 = 0x0f0f0f0f; - - const block_q3_K * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#ifdef __ARM_NEON - - uint32_t aux[3]; - uint32_t utmp[4]; - - const uint8x16_t m3b = vdupq_n_u8(0x3); - const int32x4_t vzero = vdupq_n_s32(0); - - const uint8x16_t m0 = vdupq_n_u8(1); - const uint8x16_t m1 = vshlq_n_u8(m0, 1); - const uint8x16_t m2 = vshlq_n_u8(m0, 2); - const uint8x16_t m3 = vshlq_n_u8(m0, 3); - const int8_t m32 = 32; - - ggml_int8x16x4_t q3bytes; - - float sum = 0; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict qh = x[i].hmask; - const int8_t * restrict q8 = y[i].qs; - - ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); - - ggml_uint8x16x4_t q3h; - - int32_t isum = 0; - - // Set up scales - memcpy(aux, x[i].scales, 12); - utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); - utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); - utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); - utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); - - int8_t * scale = (int8_t *)utmp; - for (int j = 0; j < 16; ++j) scale[j] -= m32; - - for (int j = 0; j < QK_K/128; ++j) { - - const ggml_uint8x16x2_t q3bits = ggml_vld1q_u8_x2(q3); q3 += 32; - const ggml_int8x16x4_t q8bytes_1 = ggml_vld1q_s8_x4(q8); q8 += 64; - const ggml_int8x16x4_t q8bytes_2 = ggml_vld1q_s8_x4(q8); q8 += 64; - - q3h.val[0] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[0]), 2); - q3h.val[1] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[1]), 2); - q3h.val[2] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[0]), 1); - q3h.val[3] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[1]), 1); - - q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[0], m3b)), vreinterpretq_s8_u8(q3h.val[0])); - q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[1], m3b)), vreinterpretq_s8_u8(q3h.val[1])); - q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 2), m3b)), vreinterpretq_s8_u8(q3h.val[2])); - q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 2), m3b)), vreinterpretq_s8_u8(q3h.val[3])); - - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_1.val[0])) * scale[0]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_1.val[1])) * scale[1]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_1.val[2])) * scale[2]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_1.val[3])) * scale[3]; - - scale += 4; - - q3h.val[0] = vbicq_u8(m2, qhbits.val[0]); - q3h.val[1] = vbicq_u8(m2, qhbits.val[1]); - q3h.val[2] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[0]), 1); - q3h.val[3] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[1]), 1); - - q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 4), m3b)), vreinterpretq_s8_u8(q3h.val[0])); - q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 4), m3b)), vreinterpretq_s8_u8(q3h.val[1])); - q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 6), m3b)), vreinterpretq_s8_u8(q3h.val[2])); - q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 6), m3b)), vreinterpretq_s8_u8(q3h.val[3])); - - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_2.val[0])) * scale[0]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_2.val[1])) * scale[1]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_2.val[2])) * scale[2]; - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_2.val[3])) * scale[3]; - - scale += 4; - - if (j == 0) { - qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 4); - qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 4); - } - - } - sum += d * isum; - - } - - *s = sum; - -#elif defined __AVX2__ - - const __m256i m3 = _mm256_set1_epi8(3); - const __m256i mone = _mm256_set1_epi8(1); - const __m128i m32 = _mm_set1_epi8(32); - - __m256 acc = _mm256_setzero_ps(); - - uint32_t aux[3]; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q3 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - // Set up scales - memcpy(aux, x[i].scales, 12); - __m128i scales128 = _mm_set_epi32( - ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), - ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), - (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), - (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); - scales128 = _mm_sub_epi8(scales128, m32); - const __m256i all_scales = _mm256_cvtepi8_epi16(scales128); - const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); - const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); - const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; - - // high bit - const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].hmask); - - // integer accumulator - __m256i sumi = _mm256_setzero_si256(); - - int bit = 0; - int is = 0; - - for (int j = 0; j < QK_K/128; ++j) { - // load low 2 bits - const __m256i q3bits = _mm256_loadu_si256((const __m256i*)q3); q3 += 32; - - // prepare low and high bits - const __m256i q3l_0 = _mm256_and_si256(q3bits, m3); - const __m256i q3h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; - - const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 2), m3); - const __m256i q3h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; - - const __m256i q3l_2 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 4), m3); - const __m256i q3h_2 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; - - const __m256i q3l_3 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 6), m3); - const __m256i q3h_3 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; - - // load Q8 quants - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); - __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); - __m256i q8s_2 = _mm256_maddubs_epi16(q3h_2, q8_2); - __m256i q8s_3 = _mm256_maddubs_epi16(q3h_3, q8_3); - - __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); - __m256i p16_2 = _mm256_maddubs_epi16(q3l_2, q8_2); - __m256i p16_3 = _mm256_maddubs_epi16(q3l_3, q8_3); - - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - p16_2 = _mm256_sub_epi16(p16_2, q8s_2); - p16_3 = _mm256_sub_epi16(p16_3, q8s_3); - - // multiply with scales - p16_0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); - p16_1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); - p16_2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); - p16_3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); - - // accumulate - p16_0 = _mm256_add_epi32(p16_0, p16_1); - p16_2 = _mm256_add_epi32(p16_2, p16_3); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_2)); - - } - - // multiply with block scale and accumulate - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); - - } - - *s = hsum_float_8(acc); - -#elif defined __AVX__ - - const __m128i m3 = _mm_set1_epi8(3); - const __m128i mone = _mm_set1_epi8(1); - const __m128i m32 = _mm_set1_epi8(32); - const __m128i m2 = _mm_set1_epi8(2); - - __m256 acc = _mm256_setzero_ps(); - - const uint32_t *aux; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q3 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - // Set up scales - aux = (const uint32_t *)x[i].scales; - __m128i scales128 = _mm_set_epi32( - ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), - ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), - (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), - (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); - scales128 = _mm_sub_epi8(scales128, m32); - const __m128i scales_0 = _mm_cvtepi8_epi16(scales128); - const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales128, scales128)); - const __m128i scales[2] = { scales_0, scales_1 }; - - // high bit *128*2 from block_q3_K.hmask[QK_K/8] - const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].hmask[0]); - const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].hmask[16]); - - // integer accumulator - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); - - for (int j = 0; j < QK_K/128; ++j) { - // load low 2 bits *64*2 from block_q3_K.qs[QK_K/4] - const __m128i q3bits_0 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; - const __m128i q3bits_1 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; - - // prepare low and high bits - const int bit = j << 2; - - const __m128i q3l_0 = _mm_and_si128(q3bits_0, m3); - const __m128i q3l_1 = _mm_and_si128(q3bits_1, m3); - const __m128i q3h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit)), bit), 2); - const __m128i q3h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit)), bit), 2); - - const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 2), m3); - const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 2), m3); - const __m128i q3h_2 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+1)), bit+1), 2); - const __m128i q3h_3 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+1)), bit+1), 2); - - const __m128i q3l_4 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 4), m3); - const __m128i q3l_5 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 4), m3); - const __m128i q3h_4 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+2)), bit+2), 2); - const __m128i q3h_5 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+2)), bit+2), 2); - - const __m128i q3l_6 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 6), m3); - const __m128i q3l_7 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 6), m3); - const __m128i q3h_6 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+3)), bit+3), 2); - const __m128i q3h_7 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+3)), bit+3), 2); - - // load Q8 quants from block_q8_K.qs[QK_K] - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, q8_0); - __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, q8_1); - __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, q8_2); - __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, q8_3); - __m128i q8s_4 = _mm_maddubs_epi16(q3h_4, q8_4); - __m128i q8s_5 = _mm_maddubs_epi16(q3h_5, q8_5); - __m128i q8s_6 = _mm_maddubs_epi16(q3h_6, q8_6); - __m128i q8s_7 = _mm_maddubs_epi16(q3h_7, q8_7); - - __m128i p16_0 = _mm_maddubs_epi16(q3l_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q3l_1, q8_1); - __m128i p16_2 = _mm_maddubs_epi16(q3l_2, q8_2); - __m128i p16_3 = _mm_maddubs_epi16(q3l_3, q8_3); - __m128i p16_4 = _mm_maddubs_epi16(q3l_4, q8_4); - __m128i p16_5 = _mm_maddubs_epi16(q3l_5, q8_5); - __m128i p16_6 = _mm_maddubs_epi16(q3l_6, q8_6); - __m128i p16_7 = _mm_maddubs_epi16(q3l_7, q8_7); - - p16_0 = _mm_sub_epi16(p16_0, q8s_0); - p16_1 = _mm_sub_epi16(p16_1, q8s_1); - p16_2 = _mm_sub_epi16(p16_2, q8s_2); - p16_3 = _mm_sub_epi16(p16_3, q8s_3); - p16_4 = _mm_sub_epi16(p16_4, q8s_4); - p16_5 = _mm_sub_epi16(p16_5, q8s_5); - p16_6 = _mm_sub_epi16(p16_6, q8s_6); - p16_7 = _mm_sub_epi16(p16_7, q8s_7); - - // multiply with scales - __m128i shuffle = _mm_set1_epi16(0x0100); - p16_0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_0); - shuffle = _mm_add_epi16(shuffle, m2); - p16_1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_1); - shuffle = _mm_add_epi16(shuffle, m2); - p16_2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_2); - shuffle = _mm_add_epi16(shuffle, m2); - p16_3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_3); - shuffle = _mm_add_epi16(shuffle, m2); - p16_4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_4); - shuffle = _mm_add_epi16(shuffle, m2); - p16_5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_5); - shuffle = _mm_add_epi16(shuffle, m2); - p16_6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_6); - shuffle = _mm_add_epi16(shuffle, m2); - p16_7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_7); - - // accumulate - p16_0 = _mm_add_epi32(p16_0, p16_1); - p16_2 = _mm_add_epi32(p16_2, p16_3); - p16_4 = _mm_add_epi32(p16_4, p16_5); - p16_6 = _mm_add_epi32(p16_6, p16_7); - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_4, p16_6)); - - } - - // multiply with block scale and accumulate - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); - - } - - *s = hsum_float_8(acc); - -#elif defined __riscv_v_intrinsic - - uint32_t aux[3]; - uint32_t utmp[4]; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict qh = x[i].hmask; - const int8_t * restrict q8 = y[i].qs; - - memcpy(aux, x[i].scales, 12); - utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); - utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); - utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); - utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); - - int8_t * scale = (int8_t *)utmp; - for (int j = 0; j < 16; ++j) scale[j] -= 32; - - - size_t vl = 32; - uint8_t m = 1; - - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t vqh = __riscv_vle8_v_u8m1(qh, vl); - - int sum_t = 0; - - for (int j = 0; j < QK_K; j += 128) { - - vl = 32; - - // load Q3 - vuint8m1_t q3_x = __riscv_vle8_v_u8m1(q3, vl); - - vint8m1_t q3_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q3_x, 0x03, vl)); - vint8m1_t q3_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x2, vl), 0x03 , vl)); - vint8m1_t q3_2 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x4, vl), 0x03 , vl)); - vint8m1_t q3_3 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x6, vl), 0x03 , vl)); - - // compute mask for subtraction - vuint8m1_t qh_m0 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_0 = __riscv_vmseq_vx_u8m1_b8(qh_m0, 0, vl); - vint8m1_t q3_m0 = __riscv_vsub_vx_i8m1_mu(vmask_0, q3_0, q3_0, 0x4, vl); - m <<= 1; - - vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_1 = __riscv_vmseq_vx_u8m1_b8(qh_m1, 0, vl); - vint8m1_t q3_m1 = __riscv_vsub_vx_i8m1_mu(vmask_1, q3_1, q3_1, 0x4, vl); - m <<= 1; - - vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_2 = __riscv_vmseq_vx_u8m1_b8(qh_m2, 0, vl); - vint8m1_t q3_m2 = __riscv_vsub_vx_i8m1_mu(vmask_2, q3_2, q3_2, 0x4, vl); - m <<= 1; - - vuint8m1_t qh_m3 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_3 = __riscv_vmseq_vx_u8m1_b8(qh_m3, 0, vl); - vint8m1_t q3_m3 = __riscv_vsub_vx_i8m1_mu(vmask_3, q3_3, q3_3, 0x4, vl); - m <<= 1; - - // load Q8 and take product with Q3 - vint16m2_t a0 = __riscv_vwmul_vv_i16m2(q3_m0, __riscv_vle8_v_i8m1(q8, vl), vl); - vint16m2_t a1 = __riscv_vwmul_vv_i16m2(q3_m1, __riscv_vle8_v_i8m1(q8+32, vl), vl); - vint16m2_t a2 = __riscv_vwmul_vv_i16m2(q3_m2, __riscv_vle8_v_i8m1(q8+64, vl), vl); - vint16m2_t a3 = __riscv_vwmul_vv_i16m2(q3_m3, __riscv_vle8_v_i8m1(q8+96, vl), vl); - - vl = 16; - - // retrieve lane to multiply with scale - vint32m2_t aux0_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 0), (scale[0]), vl); - vint32m2_t aux0_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 1), (scale[1]), vl); - vint32m2_t aux1_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 0), (scale[2]), vl); - vint32m2_t aux1_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 1), (scale[3]), vl); - vint32m2_t aux2_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 0), (scale[4]), vl); - vint32m2_t aux2_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 1), (scale[5]), vl); - vint32m2_t aux3_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 0), (scale[6]), vl); - vint32m2_t aux3_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 1), (scale[7]), vl); - - vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux0_0, aux0_1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux1_0, aux1_1, vl), isum0, vl); - vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux2_0, aux2_1, vl), isum1, vl); - vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux3_0, aux3_1, vl), isum2, vl); - - sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); - - q3 += 32; q8 += 128; scale += 8; - - } - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - - sumf += d*sum_t; - - } - - *s = sumf; - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0x3); - const vector signed char lowMask1 = vec_splats((int8_t)0xf); - const vector signed char lowMask2 = vec_splats((int8_t)0x30); - const vector int v0 = vec_splats((int32_t)0); - const vector signed char v1 = vec_splats((signed char)0x1); - const vector unsigned char v2 = vec_splats((unsigned char)0x2); - const vector unsigned char v3 = vec_splats((unsigned char)0x3); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - const vector unsigned char v6 = vec_splats((unsigned char)0x6); - const vector signed char off = vec_splats((signed char)0x20); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - UNUSED(kmask1); - UNUSED(kmask2); - - vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); - vector signed char u1 = vec_and(u0, lowMask1); - vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); - vector signed char u3 = (vector signed char)vec_mergeh((vector signed int)u2, (vector signed int)vec_sr(u2, v2)); - vector signed char u30 = vec_sl(vec_and(u3, lowMask), v4); - vector signed char u31 = vec_and(u3, lowMask2); - - u1 = vec_or(u1, u30); - u2 = vec_or(vec_sr(u0, v4), u31); - - vector signed char vscales = (vector signed char)vec_mergeh((vector signed long long)u1, (vector signed long long)u2); - vector signed char qxhs0 = (vector signed char)vec_xl( 0, x[i].hmask); - vector signed char qxhs1 = (vector signed char)vec_xl(16, x[i].hmask); - - vscales = vec_sub(vscales, off); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - vector signed int vsumi4 = v0; - vector signed int vsumi5 = v0; - vector signed int vsumi6 = v0; - vector signed int vsumi7 = v0; - - const uint8_t * restrict q3 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/128; ++j) { - __builtin_prefetch(q3, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q3); - vector signed char qxs1 = (vector signed char)vec_xl(16, q3); - q3 += 32; - - //the low 2 bits - vector signed char qxs00 = vec_and(qxs0, lowMask); - vector signed char qxs01 = vec_and(vec_sr(qxs0, v2), lowMask); - vector signed char qxs02 = vec_and(vec_sr(qxs0, v4), lowMask); - vector signed char qxs03 = vec_and(vec_sr(qxs0, v6), lowMask); - vector signed char qxs10 = vec_and(qxs1, lowMask); - vector signed char qxs11 = vec_and(vec_sr(qxs1, v2), lowMask); - vector signed char qxs12 = vec_and(vec_sr(qxs1, v4), lowMask); - vector signed char qxs13 = vec_and(vec_sr(qxs1, v6), lowMask); - - //the 3rd bit - vector signed char qxh00 = vec_sl(vec_andc(v1, qxhs0), v2); - vector signed char qxh01 = vec_sl(vec_andc(v1, vec_sr(qxhs0, (vector unsigned char)v1)), v2); - vector signed char qxh02 = vec_sl(vec_andc(v1, vec_sr(qxhs0, v2)), v2); - vector signed char qxh03 = vec_sl(vec_andc(v1, vec_sr(qxhs0, v3)), v2); - vector signed char qxh10 = vec_sl(vec_andc(v1, qxhs1), v2); - vector signed char qxh11 = vec_sl(vec_andc(v1, vec_sr(qxhs1, (vector unsigned char)v1)), v2); - vector signed char qxh12 = vec_sl(vec_andc(v1, vec_sr(qxhs1, v2)), v2); - vector signed char qxh13 = vec_sl(vec_andc(v1, vec_sr(qxhs1, v3)), v2); - qxhs0 = vec_sr(qxhs0, v4); - qxhs1 = vec_sr(qxhs1, v4); - - vector signed char q3x00 = vec_sub(qxs00, qxh00); - vector signed char q3x01 = vec_sub(qxs01, qxh01); - vector signed char q3x02 = vec_sub(qxs02, qxh02); - vector signed char q3x03 = vec_sub(qxs03, qxh03); - vector signed char q3x10 = vec_sub(qxs10, qxh10); - vector signed char q3x11 = vec_sub(qxs11, qxh11); - vector signed char q3x12 = vec_sub(qxs12, qxh12); - vector signed char q3x13 = vec_sub(qxs13, qxh13); - - vector signed char q8y00 = vec_xl( 0, q8); - vector signed char q8y10 = vec_xl( 16, q8); - vector signed char q8y01 = vec_xl( 32, q8); - vector signed char q8y11 = vec_xl( 48, q8); - vector signed char q8y02 = vec_xl( 64, q8); - vector signed char q8y12 = vec_xl( 80, q8); - vector signed char q8y03 = vec_xl( 96, q8); - vector signed char q8y13 = vec_xl(112, q8); - q8 += 128; - - vector signed short vscales_h = vec_unpackh(vscales); - vector signed short vs0 = vec_splat(vscales_h, 0); - vector signed short vs1 = vec_splat(vscales_h, 1); - vector signed short vs2 = vec_splat(vscales_h, 2); - vector signed short vs3 = vec_splat(vscales_h, 3); - vector signed short vs4 = vec_splat(vscales_h, 4); - vector signed short vs5 = vec_splat(vscales_h, 5); - vector signed short vs6 = vec_splat(vscales_h, 6); - vector signed short vs7 = vec_splat(vscales_h, 7); - vscales = vec_sld(vscales, vscales, 8); - - vector signed short qv00 = vec_add(vec_mule(q3x00, q8y00), vec_mulo(q3x00, q8y00)); - vector signed short qv01 = vec_add(vec_mule(q3x01, q8y01), vec_mulo(q3x01, q8y01)); - vector signed short qv02 = vec_add(vec_mule(q3x02, q8y02), vec_mulo(q3x02, q8y02)); - vector signed short qv03 = vec_add(vec_mule(q3x03, q8y03), vec_mulo(q3x03, q8y03)); - vector signed short qv10 = vec_add(vec_mule(q3x10, q8y10), vec_mulo(q3x10, q8y10)); - vector signed short qv11 = vec_add(vec_mule(q3x11, q8y11), vec_mulo(q3x11, q8y11)); - vector signed short qv12 = vec_add(vec_mule(q3x12, q8y12), vec_mulo(q3x12, q8y12)); - vector signed short qv13 = vec_add(vec_mule(q3x13, q8y13), vec_mulo(q3x13, q8y13)); - - vsumi0 = vec_msum(qv00, vs0, vsumi0); - vsumi1 = vec_msum(qv01, vs2, vsumi1); - vsumi2 = vec_msum(qv02, vs4, vsumi2); - vsumi3 = vec_msum(qv03, vs6, vsumi3); - vsumi4 = vec_msum(qv10, vs1, vsumi4); - vsumi5 = vec_msum(qv11, vs3, vsumi5); - vsumi6 = vec_msum(qv12, vs5, vsumi6); - vsumi7 = vec_msum(qv13, vs7, vsumi7); - } - - vsumi0 = vec_add(vsumi0, vsumi4); - vsumi1 = vec_add(vsumi1, vsumi5); - vsumi2 = vec_add(vsumi2, vsumi6); - vsumi3 = vec_add(vsumi3, vsumi7); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined __loongarch_asx - - const __m256i m3 = __lasx_xvreplgr2vr_b(3); - const __m256i mone = __lasx_xvreplgr2vr_b(1); - const __m128i m32 = __lsx_vreplgr2vr_b(32); - - __m256 acc = (__m256)__lasx_xvldi(0); - - uint32_t aux[3]; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const uint8_t * restrict q3 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - // Set up scales - memcpy(aux, x[i].scales, 12); - __m128i scales128 = lsx_set_w( - ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), - ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), - (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), - (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); - scales128 = __lsx_vsub_b(scales128, m32); - const __m256i all_scales = lasx_ext8_16(scales128); - const __m128i l_scales = lasx_extracti128(all_scales, 0); - const __m128i h_scales = lasx_extracti128(all_scales, 1); - const __m256i scales[2] = {lasx_insertf128(l_scales, l_scales), lasx_insertf128(h_scales, h_scales)}; - - // high bit - const __m256i hbits = __lasx_xvld((const __m256i*)x[i].hmask, 0); - - // integer accumulator - __m256i sumi = __lasx_xvldi(0); - - int bit = 0; - int is = 0; - __m256i xvbit; - - - for (int j = 0; j < QK_K/128; ++j) { - // load low 2 bits - const __m256i q3bits = __lasx_xvld((const __m256i*)q3, 0); q3 += 32; - - xvbit = __lasx_xvreplgr2vr_h(bit); - // prepare low and high bits - const __m256i q3l_0 = __lasx_xvand_v(q3bits, m3); - const __m256i q3h_0 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); - ++bit; - - xvbit = __lasx_xvreplgr2vr_h(bit); - const __m256i q3l_1 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 2), m3); - const __m256i q3h_1 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); - ++bit; - - xvbit = __lasx_xvreplgr2vr_h(bit); - const __m256i q3l_2 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 4), m3); - const __m256i q3h_2 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); - ++bit; - - xvbit = __lasx_xvreplgr2vr_h(bit); - const __m256i q3l_3 = __lasx_xvand_v(__lasx_xvsrli_h(q3bits, 6), m3); - const __m256i q3h_3 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvandn_v(hbits, __lasx_xvsll_h(mone, xvbit)), xvbit), 2); - ++bit; - - // load Q8 quants - const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use lasx_maddubs_h, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - __m256i q8s_0 = lasx_maddubs_h(q3h_0, q8_0); - __m256i q8s_1 = lasx_maddubs_h(q3h_1, q8_1); - __m256i q8s_2 = lasx_maddubs_h(q3h_2, q8_2); - __m256i q8s_3 = lasx_maddubs_h(q3h_3, q8_3); - - __m256i p16_0 = lasx_maddubs_h(q3l_0, q8_0); - __m256i p16_1 = lasx_maddubs_h(q3l_1, q8_1); - __m256i p16_2 = lasx_maddubs_h(q3l_2, q8_2); - __m256i p16_3 = lasx_maddubs_h(q3l_3, q8_3); - - p16_0 = __lasx_xvsub_h(p16_0, q8s_0); - p16_1 = __lasx_xvsub_h(p16_1, q8s_1); - p16_2 = __lasx_xvsub_h(p16_2, q8s_2); - p16_3 = __lasx_xvsub_h(p16_3, q8s_3); - - // multiply with scales - p16_0 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); - p16_1 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); - p16_2 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); - p16_3 = lasx_madd_h(lasx_shuffle_b(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); - - // accumulate - p16_0 = __lasx_xvadd_w(p16_0, p16_1); - p16_2 = __lasx_xvadd_w(p16_2, p16_3); - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_2)); - } - // multiply with block scale and accumulate - acc = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc);//FIXME - } - - *s = hsum_float_8(acc); - -#else - // scalar version - // This function is written like this so the compiler can manage to vectorize most of it - // Using -Ofast, GCC and clang manage to produce code that is within a factor of 2 or so from the - // manually vectorized version above. Every other version I tried would run at least 4 times slower. - // The ideal situation would be if we could just write the code once, and the compiler would - // automatically produce the best possible set of machine instructions, instead of us having to manually - // write vectorized versions for AVX, ARM_NEON, etc. - - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); - - uint32_t auxs[4]; - const int8_t * scales = (const int8_t*)auxs; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict hm = x[i].hmask; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - uint8_t m = 1; - for (int j = 0; j < QK_K; j += 128) { - for (int l = 0; l < 32; ++l) a[l] = q3[l] & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 2) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 4) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 6) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - q3 += 32; - } - a = aux8; - - memcpy(auxs, x[i].scales, 12); - uint32_t tmp = auxs[2]; - auxs[2] = ((auxs[0] >> 4) & kmask2) | (((tmp >> 4) & kmask1) << 4); - auxs[3] = ((auxs[1] >> 4) & kmask2) | (((tmp >> 6) & kmask1) << 4); - auxs[0] = (auxs[0] & kmask2) | (((tmp >> 0) & kmask1) << 4); - auxs[1] = (auxs[1] & kmask2) | (((tmp >> 2) & kmask1) << 4); - for (int j = 0; j < QK_K/16; ++j) { - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; - q8 += 8; a += 8; - } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; - -#endif - -} - -void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q4_K * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - - static const uint32_t kmask1 = 0x3f3f3f3f; - static const uint32_t kmask2 = 0x0f0f0f0f; - static const uint32_t kmask3 = 0x03030303; - - uint32_t utmp[4]; - -#ifdef __ARM_NEON - const uint8x16_t m4b = vdupq_n_u8(0xf); - const int32x4_t mzero = vdupq_n_s32(0); - - ggml_int8x16x2_t q4bytes; - ggml_int8x16x2_t q8bytes; - - float sumf = 0; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); - - memcpy(utmp, x[i].scales, 12); - - uint32x2_t mins8 = { 0 }; - mins8 = vset_lane_u32(utmp[1] & kmask1, mins8, 0); - mins8 = vset_lane_u32(((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4), mins8, 1); - - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[0] &= kmask1; - - const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(vreinterpret_u8_u32(mins8))); - const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), - vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); - sumf -= dmin * vaddvq_s32(prod); - - const uint8_t * scales = (const uint8_t *)utmp; - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - int32_t sumi1 = 0; - int32_t sumi2 = 0; - - for (int j = 0; j < QK_K/64; ++j) { - const ggml_uint8x16x2_t q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; - - q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); - q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); - - const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); - sumi1 += vaddvq_s32(p1) * scales[2*j+0]; - - q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); - q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); - - const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); - - sumi2 += vaddvq_s32(p2) * scales[2*j+1]; - } - - sumf += d * (sumi1 + sumi2); - - } - - *s = sumf; - -#elif defined __AVX2__ - - const __m256i m4 = _mm256_set1_epi8(0xF); - - __m256 acc = _mm256_setzero_ps(); - __m128 acc_m = _mm_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); - - const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); - const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); - const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); - acc_m = _mm_fmadd_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod), acc_m); - - const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); - const __m256i scales = MM256_SET_M128I(sc128, sc128); - - __m256i sumi = _mm256_setzero_si256(); - - for (int j = 0; j < QK_K/64; ++j) { - - const __m256i scale_l = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_h = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); - - const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4l = _mm256_and_si256(q4bits, m4); - const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); - - const __m256i q8l = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); - p16l = _mm256_madd_epi16(scale_l, p16l); - - const __m256i q8h = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); - p16h = _mm256_madd_epi16(scale_h, p16h); - const __m256i sumj = _mm256_add_epi32(p16l, p16h); - - sumi = _mm256_add_epi32(sumi, sumj); - } - - __m256 vd = _mm256_set1_ps(d); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); - - } - - acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); - acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); - - *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); - -#elif defined __AVX__ - - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m2 = _mm_set1_epi8(0x2); - - __m256 acc = _mm256_setzero_ps(); - __m128 acc_m = _mm_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); - const __m128i scales = _mm_cvtepu8_epi16(utmps); - const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); - - const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); - const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); - const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); - const __m128i prod = _mm_madd_epi16(mins, q8s); - acc_m = _mm_add_ps(_mm_mul_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod)), acc_m); - - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); - - __m128i shuffle = _mm_set1_epi16(0x0100); - for (int j = 0; j < QK_K/64; ++j) { - - const __m128i scale_l = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - const __m128i scale_h = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - - __m128i q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4l_0 = _mm_and_si128(q4bits, m4); - const __m128i q4h_0 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); - q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4l_1 = _mm_and_si128(q4bits, m4); - const __m128i q4h_1 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); - - const __m128i q8l_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16l = _mm_maddubs_epi16(q4l_0, q8l_0); - p16l = _mm_madd_epi16(scale_l, p16l); - sumi_0 = _mm_add_epi32(sumi_0, p16l); - const __m128i q8l_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - p16l = _mm_maddubs_epi16(q4l_1, q8l_1); - p16l = _mm_madd_epi16(scale_l, p16l); - sumi_1 = _mm_add_epi32(sumi_1, p16l); - - const __m128i q8h_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16h = _mm_maddubs_epi16(q4h_0, q8h_0); - p16h = _mm_madd_epi16(scale_h, p16h); - sumi_0 = _mm_add_epi32(sumi_0, p16h); - const __m128i q8h_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - p16h = _mm_maddubs_epi16(q4h_1, q8h_1); - p16h = _mm_madd_epi16(scale_h, p16h); - sumi_1 = _mm_add_epi32(sumi_1, p16h); - - } - - __m256 vd = _mm256_set1_ps(d); - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); - - } - - acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); - acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); - - *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); - -#elif defined __riscv_v_intrinsic - - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - - float sumf = 0; - - for (int i = 0; i < nb; ++i) { - - size_t vl = 8; - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); - vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); - vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); - vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); - vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); - - vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - vl = 32; - - int32_t sum_1 = 0; - int32_t sum_2 = 0; - - vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); - - for (int j = 0; j < QK_K/64; ++j) { - // load Q4 - vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); - - // load Q8 and multiply it with lower Q4 nibble - vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q4_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); - vint16m2_t qv_0 = __riscv_vwmul_vv_i16m2(q4_0, q8_0, vl); - vint16m1_t vs_0 = __riscv_vredsum_vs_i16m2_i16m1(qv_0, vzero, vl); - - sum_1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[2*j+0]; - - // load Q8 and multiply it with upper Q4 nibble - vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); - vint8m1_t q4_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); - vint16m2_t qv_1 = __riscv_vwmul_vv_i16m2(q4_1, q8_1, vl); - vint16m1_t vs_1 = __riscv_vredsum_vs_i16m2_i16m1(qv_1, vzero, vl); - - sum_2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[2*j+1]; - - q4 += 32; q8 += 64; - - } - - sumf += d*(sum_1 + sum_2); - - } - - *s = sumf; - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed char lowMask1 = vec_splats((int8_t)0x3f); - const vector signed char lowMask2 = vec_splats((int8_t)0x30); - const vector int v0 = vec_splats((int32_t)0); - const vector unsigned char v2 = vec_splats((uint8_t)2); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); - vector float vdmin = vec_mul(vxmin, vyd); - - vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); - vector signed short q8ysums1 = vec_xl(16, y[i].bsums); - - UNUSED(kmask1); - UNUSED(kmask2); - UNUSED(kmask3); - UNUSED(utmp); - - vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); - vector signed char u1 = vec_and(vec_sr(u0, v2), lowMask2); - vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); - vector signed char u3 = vec_sr(u2, v4); - - vector signed char u30 = u1; - vector signed char u31 = (vector signed char)vec_mergeh((vector signed int)vec_and(u2, lowMask), (vector signed int)u3); - - u1 = vec_and(u0, lowMask1); - u2 = vec_or(u30, u31); - - vector signed char utmps = (vector signed char)vec_mergeh((vector signed int)u1, (vector signed int)u2); - - vector signed short vscales = vec_unpackh(utmps); - vector signed short q4xmins = vec_unpackl(utmps); - vector signed short q4xmins0 = vec_mergeh(q4xmins, q4xmins); - vector signed short q4xmins1 = vec_mergel(q4xmins, q4xmins); - - vector signed int prod0 = vec_mule(q4xmins0, q8ysums0); - vector signed int prod1 = vec_mule(q4xmins1, q8ysums1); - vector signed int prod2 = vec_mulo(q4xmins0, q8ysums0); - vector signed int prod3 = vec_mulo(q4xmins1, q8ysums1); - - vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); - vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); - vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); - vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/64; j+=2) { - __builtin_prefetch(q4, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q4); - vector signed char qxs1 = (vector signed char)vec_xl(16, q4); - vector signed char qxs2 = (vector signed char)vec_xl(32, q4); - vector signed char qxs3 = (vector signed char)vec_xl(48, q4); - q4 += 64; - - vector unsigned char q4x00 = (vector unsigned char)vec_and(qxs0, lowMask); - vector unsigned char q4x01 = (vector unsigned char)vec_sr(qxs0, v4); - vector unsigned char q4x10 = (vector unsigned char)vec_and(qxs1, lowMask); - vector unsigned char q4x11 = (vector unsigned char)vec_sr(qxs1, v4); - vector unsigned char q4x20 = (vector unsigned char)vec_and(qxs2, lowMask); - vector unsigned char q4x21 = (vector unsigned char)vec_sr(qxs2, v4); - vector unsigned char q4x30 = (vector unsigned char)vec_and(qxs3, lowMask); - vector unsigned char q4x31 = (vector unsigned char)vec_sr(qxs3, v4); - - vector signed char q8y00 = vec_xl( 0, q8); - vector signed char q8y10 = vec_xl( 16, q8); - vector signed char q8y01 = vec_xl( 32, q8); - vector signed char q8y11 = vec_xl( 48, q8); - vector signed char q8y20 = vec_xl( 64, q8); - vector signed char q8y30 = vec_xl( 80, q8); - vector signed char q8y21 = vec_xl( 96, q8); - vector signed char q8y31 = vec_xl(112, q8); - q8 += 128; - - vector signed int qv00 = vec_msum(q8y00, q4x00, v0); - vector signed int qv01 = vec_msum(q8y01, q4x01, v0); - vector signed int qv10 = vec_msum(q8y10, q4x10, v0); - vector signed int qv11 = vec_msum(q8y11, q4x11, v0); - vector signed int qv20 = vec_msum(q8y20, q4x20, v0); - vector signed int qv21 = vec_msum(q8y21, q4x21, v0); - vector signed int qv30 = vec_msum(q8y30, q4x30, v0); - vector signed int qv31 = vec_msum(q8y31, q4x31, v0); - - vector signed int vscales_h = vec_unpackh(vscales); - vector signed int vs0 = vec_splat(vscales_h, 0); - vector signed int vs1 = vec_splat(vscales_h, 1); - vector signed int vs2 = vec_splat(vscales_h, 2); - vector signed int vs3 = vec_splat(vscales_h, 3); - vscales = vec_sld(vscales, vscales, 8); - - vsumi0 = vec_add(vec_mul(qv00, vs0), vsumi0); - vsumi1 = vec_add(vec_mul(qv01, vs1), vsumi1); - vsumi2 = vec_add(vec_mul(qv20, vs2), vsumi2); - vsumi3 = vec_add(vec_mul(qv21, vs3), vsumi3); - - vsumi0 = vec_add(vec_mul(qv10, vs0), vsumi0); - vsumi1 = vec_add(vec_mul(qv11, vs1), vsumi1); - vsumi2 = vec_add(vec_mul(qv30, vs2), vsumi2); - vsumi3 = vec_add(vec_mul(qv31, vs3), vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined __loongarch_asx - GGML_UNUSED(kmask1); - GGML_UNUSED(kmask2); - GGML_UNUSED(kmask3); - - const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); - - __m256 acc = (__m256)__lasx_xvldi(0); - __m128 acc_m = (__m128)__lsx_vldi(0); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const __m256i mins_and_scales = lasx_extu8_16(lsx_set_w(utmp[3], utmp[2], utmp[1], utmp[0])); - - const __m256i q8sums = __lasx_xvld((const __m256i*)y[i].bsums, 0); - const __m128i q8s = lsx_hadd_h(lasx_extracti128(q8sums, 0), lasx_extracti128(q8sums, 1)); - const __m128i prod = lsx_madd_h(lasx_extracti128(mins_and_scales, 1), q8s); - acc_m = __lsx_vfmadd_s(__lsx_vreplfr2vr_s(dmin), __lsx_vffint_s_w(prod), acc_m); - - const __m128i sc128 = lasx_extracti128(mins_and_scales, 0); - const __m256i scales = lasx_insertf128(sc128, sc128); - - __m256i sumi = __lasx_xvldi(0); - - for (int j = 0; j < QK_K/64; ++j) { - - const __m256i scale_l = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_h = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+1)); - - const __m256i q4bits = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; - const __m256i q4l = __lasx_xvand_v(q4bits, m4); - const __m256i q4h = __lasx_xvand_v(__lasx_xvsrli_h(q4bits, 4), m4); - - const __m256i q8l = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - __m256i p16l = lasx_maddubs_h(q4l, q8l); - p16l = lasx_madd_h(scale_l, p16l); - - const __m256i q8h = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - __m256i p16h = lasx_maddubs_h(q4h, q8h); - p16h = lasx_madd_h(scale_h, p16h); - const __m256i sumj = __lasx_xvadd_w(p16l, p16h); - - sumi = __lasx_xvadd_w(sumi, sumj); - } - - __m256 vd = __lasx_xvreplfr2vr_s(d); - acc = __lasx_xvfmadd_s(vd, __lasx_xvffint_s_w(sumi), acc); - - } - - acc_m = __lsx_vfadd_s(acc_m, (__m128)__lsx_vpermi_w((__m128i)acc_m, (__m128i)acc_m, 0xee)); - __m128i tmp1 = __lsx_vinsgr2vr_w(__lsx_vldi(0), __lsx_vpickve2gr_w((__m128i)acc_m, 1), 0); - acc_m = __lsx_vfadd_s(acc_m, (__m128)tmp1); - - - ft_union fi; - fi.i = __lsx_vpickve2gr_w(acc_m, 0); - *s = hsum_float_8(acc) + fi.f ; -#else - - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - for (int j = 0; j < QK_K/64; ++j) { - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); - a += 32; - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); - a += 32; q4 += 32; - } - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - int sumi = 0; - for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; - a = aux8; - int is = 0; - for (int j = 0; j < QK_K/32; ++j) { - int32_t scale = scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; - sumf -= dmin * sumi; - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif -} - -void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q5_K * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - - static const uint32_t kmask1 = 0x3f3f3f3f; - static const uint32_t kmask2 = 0x0f0f0f0f; - static const uint32_t kmask3 = 0x03030303; - - uint32_t utmp[4]; - -#ifdef __ARM_NEON - const uint8x16_t m4b = vdupq_n_u8(0xf); - const uint8x16_t mone = vdupq_n_u8(1); - const uint8x16_t mtwo = vdupq_n_u8(2); - const int32x4_t mzero = vdupq_n_s32(0); - - ggml_int8x16x4_t q5bytes; - - float sumf = 0; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const uint8x8_t mins8 = vld1_u8((const uint8_t*)utmp + 8); - const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(mins8)); - const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), - vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); - int32_t sumi_mins = vaddvq_s32(prod); - - const uint8_t * scales = (const uint8_t *)utmp; - - const uint8_t * restrict q5 = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); - - ggml_uint8x16x4_t q5h; - - int32_t sumi = 0; - - for (int j = 0; j < QK_K/64; ++j) { - - const ggml_uint8x16x2_t q5bits = ggml_vld1q_u8_x2(q5); q5 += 32; - const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; - - q5h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); - q5h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); - q5h.val[2] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[0]), 3); - q5h.val[3] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[1]), 3); - qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 2); - qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 2); - - q5bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[0], m4b), q5h.val[0])); - q5bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[1], m4b), q5h.val[1])); - q5bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[0], 4), q5h.val[2])); - q5bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[1], 4), q5h.val[3])); - - sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0]), q5bytes.val[1], q8bytes.val[1])) * *scales++; - sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2]), q5bytes.val[3], q8bytes.val[3])) * *scales++; - } - - sumf += d * sumi - dmin * sumi_mins; - } - - *s = sumf; - -#elif defined __AVX2__ - - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m128i mzero = _mm_setzero_si128(); - const __m256i mone = _mm256_set1_epi8(1); - - __m256 acc = _mm256_setzero_ps(); - - float summs = 0.f; - - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); - - const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); - const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); - const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); - const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); - summs += dmin * _mm_extract_epi32(hsum, 0); - - const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); - const __m256i scales = MM256_SET_M128I(sc128, sc128); - - const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].qh); - __m256i hmask = mone; - - __m256i sumi = _mm256_setzero_si256(); - - int bit = 0; - - for (int j = 0; j < QK_K/64; ++j) { - - const __m256i scale_0 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_1 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); - - const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); q5 += 32; - - const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); - const __m256i q5h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); - const __m256i q5_0 = _mm256_add_epi8(q5l_0, q5h_0); - hmask = _mm256_slli_epi16(hmask, 1); - - const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); - const __m256i q5h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); - const __m256i q5_1 = _mm256_add_epi8(q5l_1, q5h_1); - hmask = _mm256_slli_epi16(hmask, 1); - - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - __m256i p16_0 = _mm256_maddubs_epi16(q5_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q5_1, q8_1); - - p16_0 = _mm256_madd_epi16(scale_0, p16_0); - p16_1 = _mm256_madd_epi16(scale_1, p16_1); - - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); - - } - - __m256 vd = _mm256_set1_ps(d); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); - - } - - *s = hsum_float_8(acc) + summs; - -#elif defined __AVX__ - - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i mzero = _mm_setzero_si128(); - const __m128i mone = _mm_set1_epi8(1); - const __m128i m2 = _mm_set1_epi8(2); - - __m256 acc = _mm256_setzero_ps(); - - float summs = 0.f; - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); - const __m128i scales = _mm_cvtepu8_epi16(utmps); - const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); - - const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); - const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); - const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); - const __m128i prod = _mm_madd_epi16(mins, q8s); - const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); - summs += dmin * _mm_extract_epi32(hsum, 0); - - const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].qh[0]); - const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].qh[16]); - __m128i hmask = mone; - - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); - - int bit = 0; - - __m128i shuffle = _mm_set1_epi16(0x0100); - for (int j = 0; j < QK_K/64; ++j) { - - const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - - const __m128i q5bits_0 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; - const __m128i q5bits_1 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; - - __m128i q5l_0 = _mm_and_si128(q5bits_0, m4); - __m128i q5l_1 = _mm_and_si128(q5bits_1, m4); - __m128i q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); - __m128i q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); - __m128i q5_0 = _mm_add_epi8(q5l_0, q5h_0); - __m128i q5_1 = _mm_add_epi8(q5l_1, q5h_1); - hmask = _mm_slli_epi16(hmask, 1); - - __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16_0 = _mm_maddubs_epi16(q5_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q5_1, q8_1); - p16_0 = _mm_madd_epi16(scale_0, p16_0); - p16_1 = _mm_madd_epi16(scale_0, p16_1); - - q5l_0 = _mm_and_si128(_mm_srli_epi16(q5bits_0, 4), m4); - q5l_1 = _mm_and_si128(_mm_srli_epi16(q5bits_1, 4), m4); - q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); - q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); - q5_0 = _mm_add_epi8(q5l_0, q5h_0); - q5_1 = _mm_add_epi8(q5l_1, q5h_1); - hmask = _mm_slli_epi16(hmask, 1); - - q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16_2 = _mm_maddubs_epi16(q5_0, q8_0); - __m128i p16_3 = _mm_maddubs_epi16(q5_1, q8_1); - p16_2 = _mm_madd_epi16(scale_1, p16_2); - p16_3 = _mm_madd_epi16(scale_1, p16_3); - - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); - - } - - __m256 vd = _mm256_set1_ps(d); - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); - - } - - *s = hsum_float_8(acc) + summs; - -#elif defined __riscv_v_intrinsic - - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - - float sumf = 0; - float sums = 0.0; - - size_t vl; - - for (int i = 0; i < nb; ++i) { - - vl = 8; - - const uint8_t * restrict q5 = x[i].qs; - const uint8_t * restrict hm = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; - - vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); - vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); - vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); - vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); - vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); - - vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); - - vl = 32; - int32_t aux32 = 0; - int is = 0; - - uint8_t m = 1; - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t vqh = __riscv_vle8_v_u8m1(hm, vl); - - for (int j = 0; j < QK_K/64; ++j) { - // load Q5 and Q8 - vuint8m1_t q5_x = __riscv_vle8_v_u8m1(q5, vl); - vint8m1_t q8_y1 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q8_y2 = __riscv_vle8_v_i8m1(q8+32, vl); - - // compute mask for addition - vint8m1_t q5_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q5_x, 0x0F, vl)); - vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_1 = __riscv_vmsne_vx_u8m1_b8(qh_m1, 0, vl); - vint8m1_t q5_m1 = __riscv_vadd_vx_i8m1_mu(vmask_1, q5_a, q5_a, 16, vl); - m <<= 1; - - vint8m1_t q5_l = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q5_x, 0x04, vl)); - vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_2 = __riscv_vmsne_vx_u8m1_b8(qh_m2, 0, vl); - vint8m1_t q5_m2 = __riscv_vadd_vx_i8m1_mu(vmask_2, q5_l, q5_l, 16, vl); - m <<= 1; - - vint16m2_t v0 = __riscv_vwmul_vv_i16m2(q5_m1, q8_y1, vl); - vint16m2_t v1 = __riscv_vwmul_vv_i16m2(q5_m2, q8_y2, vl); - - vint32m4_t vs1 = __riscv_vwmul_vx_i32m4(v0, scales[is++], vl); - vint32m4_t vs2 = __riscv_vwmul_vx_i32m4(v1, scales[is++], vl); - - vint32m1_t vacc1 = __riscv_vredsum_vs_i32m4_i32m1(vs1, vzero, vl); - vint32m1_t vacc2 = __riscv_vredsum_vs_i32m4_i32m1(vs2, vzero, vl); - - aux32 += __riscv_vmv_x_s_i32m1_i32(vacc1) + __riscv_vmv_x_s_i32m1_i32(vacc2); - q5 += 32; q8 += 64; - - } - - vfloat32m1_t vaux = __riscv_vfmul_vf_f32m1(__riscv_vfmv_v_f_f32m1(aux32, 1), d, 1); - sums += __riscv_vfmv_f_s_f32m1_f32(vaux); - - } - - *s = sumf+sums; - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed char lowMask1 = vec_splats((int8_t)0x3f); - const vector signed char lowMask2 = vec_splats((int8_t)0x30); - const vector int v0 = vec_splats((int32_t)0); - const vector unsigned char v1 = vec_splats((unsigned char)0x1); - const vector unsigned char v2 = vec_splats((unsigned char)0x2); - const vector unsigned char v3 = vec_splats((unsigned char)0x3); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector float vxmin = vec_splats(GGML_FP16_TO_FP32(x[i].dmin)); - vector float vdmin = vec_mul(vxmin, vyd); - - UNUSED(kmask1); - UNUSED(kmask2); - UNUSED(kmask3); - UNUSED(utmp); - - vector signed char u0 = (vector signed char)vec_xl_len(x[i].scales, 8); - vector signed char u1 = vec_and(vec_sr(u0, v2), lowMask2); - vector signed char u2 = (vector signed char)vec_xl_len(x[i].scales + 8, 4); - vector signed char u3 = vec_sr(u2, v4); - - vector signed char u30 = u1; - vector signed char u31 = (vector signed char)vec_mergeh((vector signed int)vec_and(u2, lowMask), (vector signed int)u3); - - u1 = vec_and(u0, lowMask1); - u2 = vec_or(u30, u31); - - vector signed char utmps = (vector signed char)vec_mergeh((vector signed int)u1, (vector signed int)u2); - - vector signed short q8ysums0 = vec_xl( 0, y[i].bsums); - vector signed short q8ysums1 = vec_xl(16, y[i].bsums); - - vector signed short vscales = vec_unpackh(utmps); - - vector signed short q5xmins = vec_unpackl(utmps); - vector signed short q5xmins0 = vec_mergeh(q5xmins, q5xmins); - vector signed short q5xmins1 = vec_mergel(q5xmins, q5xmins); - - vector signed int prod0 = vec_mule(q5xmins0, q8ysums0); - vector signed int prod1 = vec_mule(q5xmins1, q8ysums1); - vector signed int prod2 = vec_mulo(q5xmins0, q8ysums0); - vector signed int prod3 = vec_mulo(q5xmins1, q8ysums1); - - vsumf0 = vec_nmsub(vec_ctf(prod0, 0), vdmin, vsumf0); - vsumf1 = vec_nmsub(vec_ctf(prod1, 0), vdmin, vsumf1); - vsumf2 = vec_nmsub(vec_ctf(prod2, 0), vdmin, vsumf2); - vsumf3 = vec_nmsub(vec_ctf(prod3, 0), vdmin, vsumf3); - - vector signed char qxhs0 = (vector signed char)vec_xl( 0, x[i].qh); - vector signed char qxhs1 = (vector signed char)vec_xl(16, x[i].qh); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/64; ++j) { - __builtin_prefetch(q5, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q5); - vector signed char qxs1 = (vector signed char)vec_xl(16, q5); - q5 += 32; - - vector signed char qxs00 = vec_and(qxs0, lowMask); - vector signed char qxs01 = vec_sr(qxs0, v4); - vector signed char qxs10 = vec_and(qxs1, lowMask); - vector signed char qxs11 = vec_sr(qxs1, v4); - - vector signed char q5h00 = vec_sl(vec_and((vector signed char)v1, qxhs0), v4); - vector signed char q5h01 = vec_sl(vec_and((vector signed char)v2, qxhs0), v3); - vector signed char q5h10 = vec_sl(vec_and((vector signed char)v1, qxhs1), v4); - vector signed char q5h11 = vec_sl(vec_and((vector signed char)v2, qxhs1), v3); - qxhs0 = vec_sr(qxhs0, v2); - qxhs1 = vec_sr(qxhs1, v2); - - vector unsigned char q5x00 = (vector unsigned char)vec_or(q5h00, qxs00); - vector unsigned char q5x01 = (vector unsigned char)vec_or(q5h01, qxs01); - vector unsigned char q5x10 = (vector unsigned char)vec_or(q5h10, qxs10); - vector unsigned char q5x11 = (vector unsigned char)vec_or(q5h11, qxs11); - - vector signed char q8y00 = vec_xl( 0, q8); - vector signed char q8y10 = vec_xl(16, q8); - vector signed char q8y01 = vec_xl(32, q8); - vector signed char q8y11 = vec_xl(48, q8); - q8 += 64; - - vector signed int qv00 = vec_msum(q8y00, q5x00, v0); - vector signed int qv01 = vec_msum(q8y01, q5x01, v0); - vector signed int qv10 = vec_msum(q8y10, q5x10, v0); - vector signed int qv11 = vec_msum(q8y11, q5x11, v0); - - vector signed int vscales_h = vec_unpackh(vscales); - vector signed int vs0 = vec_splat(vscales_h, 0); - vector signed int vs1 = vec_splat(vscales_h, 1); - vscales = vec_sld(vscales, vscales, 12); - - vsumi0 = vec_add(vec_mul(qv00, vs0), vsumi0); - vsumi1 = vec_add(vec_mul(qv10, vs0), vsumi1); - vsumi2 = vec_add(vec_mul(qv01, vs1), vsumi2); - vsumi3 = vec_add(vec_mul(qv11, vs1), vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined __loongarch_asx - GGML_UNUSED(kmask1); - GGML_UNUSED(kmask2); - GGML_UNUSED(kmask3); - - const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); - const __m128i mzero = __lsx_vldi(0); - const __m256i mone = __lasx_xvreplgr2vr_b(1); - - __m256 acc = (__m256)__lasx_xvldi(0); - - float summs = 0.f; - - for (int i = 0; i < nb; ++i) { - - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const __m256i mins_and_scales = lasx_extu8_16(lsx_set_w(utmp[3], utmp[2], utmp[1], utmp[0])); - - const __m256i q8sums = __lasx_xvld((const __m256i*)y[i].bsums, 0); - const __m128i q8s = lsx_hadd_h(lasx_extracti128(q8sums, 0), lasx_extracti128(q8sums, 1)); - const __m128i prod = lsx_madd_h(lasx_extracti128(mins_and_scales, 1), q8s); - const __m128i hsum = lsx_hadd_w(lsx_hadd_w(prod, mzero), mzero); - summs += dmin * __lsx_vpickve2gr_w(hsum, 0); //TODO check - - const __m128i sc128 = lasx_extracti128(mins_and_scales, 0); - const __m256i scales = lasx_insertf128(sc128, sc128); - - const __m256i hbits = __lasx_xvld((const __m256i*)x[i].qh, 0); - __m256i hmask = mone; - - __m256i sumi = __lasx_xvldi(0); - - int bit = 0; - __m256i xvbit; - - for (int j = 0; j < QK_K/64; ++j) { - - const __m256i scale_0 = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_1 = lasx_shuffle_b(scales, get_scale_shuffle_k4(2*j+1)); - - const __m256i q5bits = __lasx_xvld((const __m256i*)q5, 0); q5 += 32; - - xvbit = __lasx_xvreplgr2vr_h(bit++); - const __m256i q5l_0 = __lasx_xvand_v(q5bits, m4); - const __m256i q5h_0 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvand_v(hbits, hmask), xvbit), 4); - const __m256i q5_0 = __lasx_xvadd_b(q5l_0, q5h_0); - hmask = __lasx_xvslli_h(hmask, 1); - - xvbit = __lasx_xvreplgr2vr_h(bit++); - const __m256i q5l_1 = __lasx_xvand_v(__lasx_xvsrli_h(q5bits, 4), m4); - const __m256i q5h_1 = __lasx_xvslli_h(__lasx_xvsrl_h(__lasx_xvand_v(hbits, hmask), xvbit), 4); - const __m256i q5_1 = __lasx_xvadd_b(q5l_1, q5h_1); - hmask = __lasx_xvslli_h(hmask, 1); - - const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - - __m256i p16_0 = lasx_maddubs_h(q5_0, q8_0); - __m256i p16_1 = lasx_maddubs_h(q5_1, q8_1); - - p16_0 = lasx_madd_h(scale_0, p16_0); - p16_1 = lasx_madd_h(scale_1, p16_1); - - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_1)); - - } - - __m256 vd = __lasx_xvreplfr2vr_s(d); - acc = __lasx_xvfmadd_s(vd, __lasx_xvffint_s_w(sumi), acc); - - } - - *s = hsum_float_8(acc) + summs; - -#else - - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const uint8_t * restrict hm = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - uint8_t m = 1; - for (int j = 0; j < QK_K/64; ++j) { - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); - for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); - for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); - a += 32; m <<= 1; - q4 += 32; - } - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - int sumi = 0; - for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; - a = aux8; - int is = 0; - for (int j = 0; j < QK_K/32; ++j) { - int32_t scale = scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; - sumf -= dmin * sumi; - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif -} - -void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_q6_K * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#ifdef __ARM_NEON - float sum = 0; - - const uint8x16_t m4b = vdupq_n_u8(0xF); - const int32x4_t vzero = vdupq_n_s32(0); - //const int8x16_t m32s = vdupq_n_s8(32); - - const uint8x16_t mone = vdupq_n_u8(3); - - ggml_int8x16x4_t q6bytes; - ggml_uint8x16x4_t q6h; - - for (int i = 0; i < nb; ++i) { - - const float d_all = GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - const int8_t * restrict scale = x[i].scales; - - const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); - const int8x16_t scales = vld1q_s8(scale); - const ggml_int16x8x2_t q6scales = {{vmovl_s8(vget_low_s8(scales)), vmovl_s8(vget_high_s8(scales))}}; - - const int32x4_t prod = vaddq_s32(vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[0]), vget_low_s16 (q6scales.val[0])), - vmull_s16(vget_high_s16(q8sums.val[0]), vget_high_s16(q6scales.val[0]))), - vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[1]), vget_low_s16 (q6scales.val[1])), - vmull_s16(vget_high_s16(q8sums.val[1]), vget_high_s16(q6scales.val[1])))); - int32_t isum_mins = vaddvq_s32(prod); - - int32_t isum = 0; - - for (int j = 0; j < QK_K/128; ++j) { - - ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); qh += 32; - ggml_uint8x16x4_t q6bits = ggml_vld1q_u8_x4(q6); q6 += 64; - ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; - - q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); - q6h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); - uint8x16_t shifted = vshrq_n_u8(qhbits.val[0], 2); - q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 2); - q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - - //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); - //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); - //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])), m32s); - //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])), m32s); - q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])); - q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])); - q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])); - q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])); - - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; - - scale += 4; - - q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; - - shifted = vshrq_n_u8(qhbits.val[0], 4); - q6h.val[0] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 4); - q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[0], 6); - q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 6); - q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - - //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])), m32s); - //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])), m32s); - //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])), m32s); - //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])), m32s); - q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])); - q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])); - q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])); - q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])); - - isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + - vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; - scale += 4; - } - //sum += isum * d_all * y[i].d; - sum += d_all * y[i].d * (isum - 32 * isum_mins); - - } - *s = sum; - -#elif defined __AVX2__ - - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m256i m2 = _mm256_set1_epi8(3); - const __m256i m32s = _mm256_set1_epi8(32); - - __m256 acc = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); - - __m256i sumi = _mm256_setzero_si256(); - - int is = 0; - - for (int j = 0; j < QK_K/128; ++j) { - - const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); - const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); - const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); - const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); - is += 4; - - const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4bits2 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4bitsH = _mm256_loadu_si256((const __m256i*)qh); qh += 32; - - const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(q4bitsH, m2), 4); - const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 2), m2), 4); - const __m256i q4h_2 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 4), m2), 4); - const __m256i q4h_3 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 6), m2), 4); - - const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); - const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(q4bits2, m4), q4h_1); - const __m256i q4_2 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_2); - const __m256i q4_3 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits2, 4), m4), q4h_3); - - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); - __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); - __m256i q8s_2 = _mm256_maddubs_epi16(m32s, q8_2); - __m256i q8s_3 = _mm256_maddubs_epi16(m32s, q8_3); - - __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); - __m256i p16_2 = _mm256_maddubs_epi16(q4_2, q8_2); - __m256i p16_3 = _mm256_maddubs_epi16(q4_3, q8_3); - - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - p16_2 = _mm256_sub_epi16(p16_2, q8s_2); - p16_3 = _mm256_sub_epi16(p16_3, q8s_3); - - p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); - p16_2 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_2), p16_2); - p16_3 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_3), p16_3); - - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_2, p16_3)); - - } - - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); - } - - *s = hsum_float_8(acc); - -#elif defined __AVX__ - - const __m128i m3 = _mm_set1_epi8(3); - const __m128i m15 = _mm_set1_epi8(15); - - __m256 acc = _mm256_setzero_ps(); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - // handle the q6_k -32 offset separately using bsums - const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)y[i].bsums); - const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)y[i].bsums + 1); - const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); - const __m128i scales_16_0 = _mm_cvtepi8_epi16(scales); - const __m128i scales_16_1 = _mm_cvtepi8_epi16(_mm_bsrli_si128(scales, 8)); - const __m128i q8sclsub_0 = _mm_slli_epi32(_mm_madd_epi16(q8sums_0, scales_16_0), 5); - const __m128i q8sclsub_1 = _mm_slli_epi32(_mm_madd_epi16(q8sums_1, scales_16_1), 5); - - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); - - int is = 0; - - for (int j = 0; j < QK_K/128; ++j) { - - const __m128i q4bitsH_0 = _mm_loadu_si128((const __m128i*)qh); qh += 16; - const __m128i q4bitsH_1 = _mm_loadu_si128((const __m128i*)qh); qh += 16; - - const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, m3), 4); - const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, m3), 4); - const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, _mm_set1_epi8(12)), 2); - const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, _mm_set1_epi8(12)), 2); - const __m128i q4h_4 = _mm_and_si128(q4bitsH_0, _mm_set1_epi8(48)); - const __m128i q4h_5 = _mm_and_si128(q4bitsH_1, _mm_set1_epi8(48)); - const __m128i q4h_6 = _mm_srli_epi16(_mm_and_si128(q4bitsH_0, _mm_set1_epi8(-64)), 2); - const __m128i q4h_7 = _mm_srli_epi16(_mm_and_si128(q4bitsH_1, _mm_set1_epi8(-64)), 2); - - const __m128i q4bits1_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits1_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits2_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits2_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - - const __m128i q4_0 = _mm_or_si128(_mm_and_si128(q4bits1_0, m15), q4h_0); - const __m128i q4_1 = _mm_or_si128(_mm_and_si128(q4bits1_1, m15), q4h_1); - const __m128i q4_2 = _mm_or_si128(_mm_and_si128(q4bits2_0, m15), q4h_2); - const __m128i q4_3 = _mm_or_si128(_mm_and_si128(q4bits2_1, m15), q4h_3); - const __m128i q4_4 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_0, 4), m15), q4h_4); - const __m128i q4_5 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_1, 4), m15), q4h_5); - const __m128i q4_6 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_0, 4), m15), q4h_6); - const __m128i q4_7 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_1, 4), m15), q4h_7); - - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - - __m128i p16_0 = _mm_maddubs_epi16(q4_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q4_1, q8_1); - __m128i p16_2 = _mm_maddubs_epi16(q4_2, q8_2); - __m128i p16_3 = _mm_maddubs_epi16(q4_3, q8_3); - __m128i p16_4 = _mm_maddubs_epi16(q4_4, q8_4); - __m128i p16_5 = _mm_maddubs_epi16(q4_5, q8_5); - __m128i p16_6 = _mm_maddubs_epi16(q4_6, q8_6); - __m128i p16_7 = _mm_maddubs_epi16(q4_7, q8_7); - - const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); - const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); - const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); - const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); - is += 4; - - p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_0, 8)), p16_1); - p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); - p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_1, 8)), p16_3); - p16_4 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_2), p16_4); - p16_5 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_2, 8)), p16_5); - p16_6 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_3), p16_6); - p16_7 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_bsrli_si128(scale_3, 8)), p16_7); - - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_4, p16_6)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_5, p16_7)); - - } - - sumi_0 = _mm_sub_epi32(sumi_0, q8sclsub_0); - sumi_1 = _mm_sub_epi32(sumi_1, q8sclsub_1); - const __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sumi)), acc); - } - - *s = hsum_float_8(acc); - -#elif defined __riscv_v_intrinsic - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - const int8_t * restrict scale = x[i].scales; - - size_t vl; - - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - - int sum_t = 0; - int is = 0; - - for (int j = 0; j < QK_K/128; ++j) { - - vl = 32; - - // load qh - vuint8m1_t qh_x = __riscv_vle8_v_u8m1(qh, vl); - - // load Q6 - vuint8m1_t q6_0 = __riscv_vle8_v_u8m1(q6, vl); - vuint8m1_t q6_1 = __riscv_vle8_v_u8m1(q6+32, vl); - - vuint8m1_t q6a_0 = __riscv_vand_vx_u8m1(q6_0, 0x0F, vl); - vuint8m1_t q6a_1 = __riscv_vand_vx_u8m1(q6_1, 0x0F, vl); - vuint8m1_t q6s_0 = __riscv_vsrl_vx_u8m1(q6_0, 0x04, vl); - vuint8m1_t q6s_1 = __riscv_vsrl_vx_u8m1(q6_1, 0x04, vl); - - vuint8m1_t qh_0 = __riscv_vand_vx_u8m1(qh_x, 0x03, vl); - vuint8m1_t qh_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x2, vl), 0x03 , vl); - vuint8m1_t qh_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x4, vl), 0x03 , vl); - vuint8m1_t qh_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x6, vl), 0x03 , vl); - - vuint8m1_t qhi_0 = __riscv_vor_vv_u8m1(q6a_0, __riscv_vsll_vx_u8m1(qh_0, 0x04, vl), vl); - vuint8m1_t qhi_1 = __riscv_vor_vv_u8m1(q6a_1, __riscv_vsll_vx_u8m1(qh_1, 0x04, vl), vl); - vuint8m1_t qhi_2 = __riscv_vor_vv_u8m1(q6s_0, __riscv_vsll_vx_u8m1(qh_2, 0x04, vl), vl); - vuint8m1_t qhi_3 = __riscv_vor_vv_u8m1(q6s_1, __riscv_vsll_vx_u8m1(qh_3, 0x04, vl), vl); - - vint8m1_t a_0 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_0), 32, vl); - vint8m1_t a_1 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_1), 32, vl); - vint8m1_t a_2 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_2), 32, vl); - vint8m1_t a_3 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_3), 32, vl); - - // load Q8 and take product - vint16m2_t va_q_0 = __riscv_vwmul_vv_i16m2(a_0, __riscv_vle8_v_i8m1(q8, vl), vl); - vint16m2_t va_q_1 = __riscv_vwmul_vv_i16m2(a_1, __riscv_vle8_v_i8m1(q8+32, vl), vl); - vint16m2_t va_q_2 = __riscv_vwmul_vv_i16m2(a_2, __riscv_vle8_v_i8m1(q8+64, vl), vl); - vint16m2_t va_q_3 = __riscv_vwmul_vv_i16m2(a_3, __riscv_vle8_v_i8m1(q8+96, vl), vl); - - vl = 16; - - vint32m2_t vaux_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 0), scale[is+0], vl); - vint32m2_t vaux_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 1), scale[is+1], vl); - vint32m2_t vaux_2 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 0), scale[is+2], vl); - vint32m2_t vaux_3 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 1), scale[is+3], vl); - vint32m2_t vaux_4 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 0), scale[is+4], vl); - vint32m2_t vaux_5 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 1), scale[is+5], vl); - vint32m2_t vaux_6 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 0), scale[is+6], vl); - vint32m2_t vaux_7 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 1), scale[is+7], vl); - - vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_0, vaux_1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_2, vaux_3, vl), isum0, vl); - vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_4, vaux_5, vl), isum1, vl); - vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_6, vaux_7, vl), isum2, vl); - - sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); - - q6 += 64; qh += 32; q8 += 128; is=8; - - } - - sumf += d * sum_t; - - } - - *s = sumf; - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector int v0 = vec_splats((int32_t)0); - const vector unsigned char v2 = vec_splats((unsigned char)0x2); - const vector unsigned char v3 = vec_splats((unsigned char)0x3); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - const vector unsigned char v6 = vec_splats((unsigned char)0x6); - const vector signed char off = vec_splats((signed char)0x20); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - vector signed int vsumi4 = v0; - vector signed int vsumi5 = v0; - vector signed int vsumi6 = v0; - vector signed int vsumi7 = v0; - - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict qs = x[i].scales; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/128; ++j) { - __builtin_prefetch(q6, 0, 0); - __builtin_prefetch(qh, 0, 0); - __builtin_prefetch(q8, 0, 0); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q6); - vector signed char qxs1 = (vector signed char)vec_xl(16, q6); - vector signed char qxs2 = (vector signed char)vec_xl(32, q6); - vector signed char qxs3 = (vector signed char)vec_xl(48, q6); - q6 += 64; - - vector signed char qxs00 = vec_and(qxs0, lowMask); - vector signed char qxs01 = vec_sr(qxs0, v4); - vector signed char qxs10 = vec_and(qxs1, lowMask); - vector signed char qxs11 = vec_sr(qxs1, v4); - vector signed char qxs20 = vec_and(qxs2, lowMask); - vector signed char qxs21 = vec_sr(qxs2, v4); - vector signed char qxs30 = vec_and(qxs3, lowMask); - vector signed char qxs31 = vec_sr(qxs3, v4); - - vector signed char qxhs0 = (vector signed char)vec_xl( 0, qh); - vector signed char qxhs1 = (vector signed char)vec_xl(16, qh); - qh += 32; - - vector signed char qxh00 = vec_sl(vec_and((vector signed char)v3, qxhs0), v4); - vector signed char qxh01 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v4)), v4); - vector signed char qxh10 = vec_sl(vec_and((vector signed char)v3, qxhs1), v4); - vector signed char qxh11 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v4)), v4); - vector signed char qxh20 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v2)), v4); - vector signed char qxh21 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs0, v6)), v4); - vector signed char qxh30 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v2)), v4); - vector signed char qxh31 = vec_sl(vec_and((vector signed char)v3, vec_sr(qxhs1, v6)), v4); - - vector signed char q6x00 = vec_sub(vec_or(qxh00, qxs00), off); - vector signed char q6x01 = vec_sub(vec_or(qxh01, qxs01), off); - vector signed char q6x10 = vec_sub(vec_or(qxh10, qxs10), off); - vector signed char q6x11 = vec_sub(vec_or(qxh11, qxs11), off); - vector signed char q6x20 = vec_sub(vec_or(qxh20, qxs20), off); - vector signed char q6x21 = vec_sub(vec_or(qxh21, qxs21), off); - vector signed char q6x30 = vec_sub(vec_or(qxh30, qxs30), off); - vector signed char q6x31 = vec_sub(vec_or(qxh31, qxs31), off); - - vector signed char q8y00 = vec_xl( 0, q8); - vector signed char q8y10 = vec_xl( 16, q8); - vector signed char q8y20 = vec_xl( 32, q8); - vector signed char q8y30 = vec_xl( 48, q8); - vector signed char q8y01 = vec_xl( 64, q8); - vector signed char q8y11 = vec_xl( 80, q8); - vector signed char q8y21 = vec_xl( 96, q8); - vector signed char q8y31 = vec_xl(112, q8); - q8 += 128; - - vector signed short qv00 = vec_add(vec_mule(q6x00, q8y00), vec_mulo(q6x00, q8y00)); - vector signed short qv10 = vec_add(vec_mule(q6x10, q8y10), vec_mulo(q6x10, q8y10)); - vector signed short qv20 = vec_add(vec_mule(q6x20, q8y20), vec_mulo(q6x20, q8y20)); - vector signed short qv30 = vec_add(vec_mule(q6x30, q8y30), vec_mulo(q6x30, q8y30)); - vector signed short qv01 = vec_add(vec_mule(q6x01, q8y01), vec_mulo(q6x01, q8y01)); - vector signed short qv11 = vec_add(vec_mule(q6x11, q8y11), vec_mulo(q6x11, q8y11)); - vector signed short qv21 = vec_add(vec_mule(q6x21, q8y21), vec_mulo(q6x21, q8y21)); - vector signed short qv31 = vec_add(vec_mule(q6x31, q8y31), vec_mulo(q6x31, q8y31)); - - vector signed short vscales = vec_unpackh(vec_xl_len(qs, 8)); - qs += 8; - - vector signed short vs0 = vec_splat(vscales, 0); - vector signed short vs1 = vec_splat(vscales, 1); - vector signed short vs2 = vec_splat(vscales, 2); - vector signed short vs3 = vec_splat(vscales, 3); - vector signed short vs4 = vec_splat(vscales, 4); - vector signed short vs5 = vec_splat(vscales, 5); - vector signed short vs6 = vec_splat(vscales, 6); - vector signed short vs7 = vec_splat(vscales, 7); - - vsumi0 = vec_msum(qv00, vs0, vsumi0); - vsumi1 = vec_msum(qv01, vs4, vsumi1); - vsumi2 = vec_msum(qv10, vs1, vsumi2); - vsumi3 = vec_msum(qv11, vs5, vsumi3); - vsumi4 = vec_msum(qv20, vs2, vsumi4); - vsumi5 = vec_msum(qv21, vs6, vsumi5); - vsumi6 = vec_msum(qv30, vs3, vsumi6); - vsumi7 = vec_msum(qv31, vs7, vsumi7); - } - - vsumi0 = vec_add(vsumi0, vsumi4); - vsumi1 = vec_add(vsumi1, vsumi5); - vsumi2 = vec_add(vsumi2, vsumi6); - vsumi3 = vec_add(vsumi3, vsumi7); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined __loongarch_asx - - const __m256i m4 = __lasx_xvreplgr2vr_b(0xF); - const __m256i m2 = __lasx_xvreplgr2vr_b(3); - const __m256i m32s = __lasx_xvreplgr2vr_b(32); - - __m256 acc = (__m256)__lasx_xvldi(0); - - for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - - const __m128i scales = __lsx_vld((const __m128i*)x[i].scales, 0); - - __m256i sumi = __lasx_xvldi(0); - - int is = 0; - - for (int j = 0; j < QK_K/128; ++j) { - - const __m128i scale_0 = lsx_shuffle_b(scales, get_scale_shuffle(is + 0)); - const __m128i scale_1 = lsx_shuffle_b(scales, get_scale_shuffle(is + 1)); - const __m128i scale_2 = lsx_shuffle_b(scales, get_scale_shuffle(is + 2)); - const __m128i scale_3 = lsx_shuffle_b(scales, get_scale_shuffle(is + 3)); - is += 4; - - const __m256i q4bits1 = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; - const __m256i q4bits2 = __lasx_xvld((const __m256i*)q4, 0); q4 += 32; - const __m256i q4bitsH = __lasx_xvld((const __m256i*)qh, 0); qh += 32; - - const __m256i q4h_0 = __lasx_xvslli_h(__lasx_xvand_v(q4bitsH, m2), 4); - const __m256i q4h_1 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 2), m2), 4); - const __m256i q4h_2 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 4), m2), 4); - const __m256i q4h_3 = __lasx_xvslli_h(__lasx_xvand_v(__lasx_xvsrli_h(q4bitsH, 6), m2), 4); - - const __m256i q4_0 = __lasx_xvor_v(__lasx_xvand_v(q4bits1, m4), q4h_0); - const __m256i q4_1 = __lasx_xvor_v(__lasx_xvand_v(q4bits2, m4), q4h_1); - const __m256i q4_2 = __lasx_xvor_v(__lasx_xvand_v(__lasx_xvsrli_h(q4bits1, 4), m4), q4h_2); - const __m256i q4_3 = __lasx_xvor_v(__lasx_xvand_v(__lasx_xvsrli_h(q4bits2, 4), m4), q4h_3); - - const __m256i q8_0 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8_3 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - - __m256i q8s_0 = lasx_maddubs_h(m32s, q8_0); - __m256i q8s_1 = lasx_maddubs_h(m32s, q8_1); - __m256i q8s_2 = lasx_maddubs_h(m32s, q8_2); - __m256i q8s_3 = lasx_maddubs_h(m32s, q8_3); - - __m256i p16_0 = lasx_maddubs_h(q4_0, q8_0); - __m256i p16_1 = lasx_maddubs_h(q4_1, q8_1); - __m256i p16_2 = lasx_maddubs_h(q4_2, q8_2); - __m256i p16_3 = lasx_maddubs_h(q4_3, q8_3); - - p16_0 = __lasx_xvsub_h(p16_0, q8s_0); - p16_1 = __lasx_xvsub_h(p16_1, q8s_1); - p16_2 = __lasx_xvsub_h(p16_2, q8s_2); - p16_3 = __lasx_xvsub_h(p16_3, q8s_3); - - p16_0 = lasx_madd_h(lasx_ext8_16(scale_0), p16_0); - p16_1 = lasx_madd_h(lasx_ext8_16(scale_1), p16_1); - p16_2 = lasx_madd_h(lasx_ext8_16(scale_2), p16_2); - p16_3 = lasx_madd_h(lasx_ext8_16(scale_3), p16_3); - - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_0, p16_1)); - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p16_2, p16_3)); - } - - acc = __lasx_xvfmadd_s((__m256)__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), acc); - } - - *s = hsum_float_8(acc); - -#else - - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - for (int j = 0; j < QK_K; j += 128) { - for (int l = 0; l < 32; ++l) { - a[l + 0] = (int8_t)((q4[l + 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; - a[l + 32] = (int8_t)((q4[l + 32] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; - a[l + 64] = (int8_t)((q4[l + 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; - a[l + 96] = (int8_t)((q4[l + 32] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; - } - a += 128; - q4 += 64; - qh += 32; - } - a = aux8; - int is = 0; - for (int j = 0; j < QK_K/16; ++j) { - int scale = x[i].scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif -} - -#if defined (__AVX__) || defined (__AVX2__) || defined (__ARM_NEON) || defined (__POWER9_VECTOR__) || defined(__loongarch_asx) -static const int8_t keven_signs_q2xs[1024] = { - 1, 1, 1, 1, 1, 1, 1, 1, -1, 1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, 1, - 1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, 1, 1, -1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, -1, - 1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, -1, - 1, 1, -1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, 1, - 1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, -1, - 1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, 1, - 1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, 1, - 1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, 1, 1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, -1, - 1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, -1, - 1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, 1, - 1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, 1, - 1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, -1, - 1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, 1, - 1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, -1, - 1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, -1, - 1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, 1, - 1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, -1, - 1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, 1, - 1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, 1, - 1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, -1, - 1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, 1, - 1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, -1, - 1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, -1, - 1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, 1, - 1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, 1, - 1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, -1, - 1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, -1, - 1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, 1, - 1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, -1, - 1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, 1, - 1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, 1, - 1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, 1, 1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -}; -#endif - -void ggml_vec_dot_iq2_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq2_xxs * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[4]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - ggml_int8x16x4_t q2u; - ggml_int8x16x4_t q2s; - ggml_int8x16x4_t q8b; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - float sumf1 = 0, sumf2 = 0; - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; - q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 0])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 1]))); - q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 2])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 3]))); - q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 8])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 9]))); - q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[10])), vld1_s8((const void *)(iq2xxs_grid + aux8[11]))); - q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); - q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); - q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 7) & 127)))); - q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 21) & 127)))); - q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); - q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); - q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); - q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); - const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]), q2u.val[1], q8b.val[1]); - const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]), q2u.val[3], q8b.val[3]); - sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[1] >> 28)); - sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[3] >> 28)); - } - sumf += d*(sumf1 + sumf2); - } - *s = 0.25f * sumf; - -#elif defined(__AVX2__) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[4]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; - const __m256i q2_1 = _mm256_set_epi64x(iq2xxs_grid[aux8[ 3]], iq2xxs_grid[aux8[ 2]], iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); - const __m256i q2_2 = _mm256_set_epi64x(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]], iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); - const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], - signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127], - signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); - const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); - const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); - const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); - const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); - const uint16_t ls1 = aux32[1] >> 28; - const uint16_t ls2 = aux32[3] >> 28; - const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); - const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); - sumi1 = _mm256_add_epi32(sumi1, p1); - sumi2 = _mm256_add_epi32(sumi2, p2); - } - - accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__AVX__) - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[4]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; - const __m128i q2_1_0 = _mm_set_epi64x(iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); - const __m128i q2_1_1 = _mm_set_epi64x(iq2xxs_grid[aux8[3]], iq2xxs_grid[aux8[2]]); - const __m128i q2_2_0 = _mm_set_epi64x(iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); - const __m128i q2_2_1 = _mm_set_epi64x(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]]); - const __m128i s2_1_0 = _mm_set_epi64x(signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m128i s2_1_1 = _mm_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127]); - const __m128i s2_2_0 = _mm_set_epi64x(signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); - const __m128i s2_2_1 = _mm_set_epi64x(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127]); - const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, s2_1_0); - const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, s2_1_1); - const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, s2_2_0); - const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, s2_2_1); - const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); - const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); - const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); - const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); - const uint16_t ls1 = aux32[1] >> 28; - const uint16_t ls2 = aux32[3] >> 28; - const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); - sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); - sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); - sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); - sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); - } - - accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__POWER9_VECTOR__) - const vector int v0 = vec_splats((int32_t)0); - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/32; j += 2) { - __builtin_prefetch(q2, 0, 1); - __builtin_prefetch(q8, 0, 1); - - uint32_t aux32[4]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - memcpy(aux32, q2, 4*sizeof(uint32_t)); - q2 += 8; - - vector signed long long aux64x2_0 = {*(const int64_t *)(iq2xxs_grid + aux8[ 0]), *(const int64_t *)(iq2xxs_grid + aux8[ 1])}; - vector signed long long aux64x2_1 = {*(const int64_t *)(iq2xxs_grid + aux8[ 2]), *(const int64_t *)(iq2xxs_grid + aux8[ 3])}; - vector signed long long aux64x2_2 = {*(const int64_t *)(iq2xxs_grid + aux8[ 8]), *(const int64_t *)(iq2xxs_grid + aux8[ 9])}; - vector signed long long aux64x2_3 = {*(const int64_t *)(iq2xxs_grid + aux8[10]), *(const int64_t *)(iq2xxs_grid + aux8[11])}; - - vector signed long long vsigns0 = {*(const int64_t *)(signs64 + ((aux32[1] >> 0) & 127)), *(const int64_t *)(signs64 + ((aux32[1] >> 7) & 127))}; - vector signed long long vsigns1 = {*(const int64_t *)(signs64 + ((aux32[1] >> 14) & 127)), *(const int64_t *)(signs64 + ((aux32[1] >> 21) & 127))}; - vector signed long long vsigns2 = {*(const int64_t *)(signs64 + ((aux32[3] >> 0) & 127)), *(const int64_t *)(signs64 + ((aux32[3] >> 7) & 127))}; - vector signed long long vsigns3 = {*(const int64_t *)(signs64 + ((aux32[3] >> 14) & 127)), *(const int64_t *)(signs64 + ((aux32[3] >> 21) & 127))}; - - vector signed char q2x0 = (vector signed char)vec_mul((vector signed char)vsigns0, (vector signed char)aux64x2_0); - vector signed char q2x1 = (vector signed char)vec_mul((vector signed char)vsigns1, (vector signed char)aux64x2_1); - vector signed char q2x2 = (vector signed char)vec_mul((vector signed char)vsigns2, (vector signed char)aux64x2_2); - vector signed char q2x3 = (vector signed char)vec_mul((vector signed char)vsigns3, (vector signed char)aux64x2_3); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); - - const uint16_t ls0 = aux32[1] >> 28; - const uint16_t ls1 = aux32[3] >> 28; - - vector signed short vscales01 = vec_splats((int16_t)(2*ls0+1)); - vector signed short vscales23 = vec_splats((int16_t)(2*ls1+1)); - - vsumi0 = vec_msum(qv0, vscales01, vsumi0); - vsumi1 = vec_msum(qv1, vscales01, vsumi1); - vsumi2 = vec_msum(qv2, vscales23, vsumi2); - vsumi3 = vec_msum(qv3, vscales23, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = 0.125f * vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[4]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - __m256 accumf = (__m256)__lasx_xvldi(0); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; - - const __m256i q2_1 = lasx_set_d(iq2xxs_grid[aux8[ 3]], iq2xxs_grid[aux8[ 2]], iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); - const __m256i q2_2 = lasx_set_d(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]], iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); - const __m256i s2_1 = lasx_set_d(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], - signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m256i s2_2 = lasx_set_d(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127], - signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); - const __m256i q8s_1 = __lasx_xvsigncov_b(s2_1, q8_1); - const __m256i q8s_2 = __lasx_xvsigncov_b(s2_2, q8_2); - const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); - const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); - const uint16_t ls1 = aux32[1] >> 28; - const uint16_t ls2 = aux32[3] >> 28; - const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); - const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); - sumi1 = __lasx_xvadd_w(sumi1, p1); - sumi2 = __lasx_xvadd_w(sumi2, p2); - } - - accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); - } - - *s = 0.125f * hsum_float_8(accumf); - -#else - - uint32_t aux32[2]; - const uint8_t * aux8 = (const uint8_t *)aux32; - - float sumf = 0.f; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - int32_t bsum = 0; - for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { - memcpy(aux32, q2, 2*sizeof(uint32_t)); - q2 += 4; - const uint32_t ls = 2*(aux32[1] >> 28) + 1; - int32_t sumi = 0; - for (int l = 0; l < 4; ++l) { - const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[l]); - const uint8_t signs = ksigns_iq2xs[(aux32[1] >> 7*l) & 127]; - for (int j = 0; j < 8; ++j) { - sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); - } - q8 += 8; - } - bsum += sumi * ls; - } - sumf += d * bsum; - } - *s = 0.125f * sumf; -#endif -} - -void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq2_xs * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - ggml_int8x16x4_t q2u; - ggml_int8x16x4_t q2s; - ggml_int8x16x4_t q8b; - - int32x4x4_t scales32; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - const uint8x8_t scales8 = vld1_u8(x[i].scales); - const uint8x8_t scales_l = vand_u8(scales8, vdup_n_u8(0xf)); - const uint8x8_t scales_h = vshr_n_u8(scales8, 4); - uint8x16_t scales = vcombine_u8(vzip1_u8(scales_l, scales_h), vzip2_u8(scales_l, scales_h)); - scales = vaddq_u8(vshlq_n_u8(scales, 1), vdupq_n_u8(1)); - const uint16x8_t scales1 = vmovl_u8(vget_low_u8(scales)); - const uint16x8_t scales2 = vmovl_u8(vget_high_u8(scales)); - scales32.val[0] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales1))); - scales32.val[1] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales1))); - scales32.val[2] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales2))); - scales32.val[3] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales2))); - int32x4_t sumi = vdupq_n_s32(0); - for (int ib64 = 0; ib64 < QK_K/64; ++ib64) { - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[0] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[1] & 511)))); - q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[2] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[3] & 511)))); - q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[4] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[5] & 511)))); - q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[6] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[7] & 511)))); - q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[0] >> 9))), vld1_s8((const void *)(signs64 + (q2[1] >> 9)))); - q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[2] >> 9))), vld1_s8((const void *)(signs64 + (q2[3] >> 9)))); - q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[4] >> 9))), vld1_s8((const void *)(signs64 + (q2[5] >> 9)))); - q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[6] >> 9))), vld1_s8((const void *)(signs64 + (q2[7] >> 9)))); - q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); - q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); - q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); - q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); - const int32x4_t p1 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]); - const int32x4_t p2 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[1], q8b.val[1]); - const int32x4_t p3 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]); - const int32x4_t p4 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[3], q8b.val[3]); - const int32x4_t p = vpaddq_s32(vpaddq_s32(p1, p2), vpaddq_s32(p3, p4)); - sumi = vmlaq_s32(sumi, p, scales32.val[ib64]); - q2 += 8; - } - sumf += d*vaddvq_s32(sumi); - } - *s = 0.125f * sumf; - -#elif defined(__AVX2__) - - const __m256i mone = _mm256_set1_epi8(1); - static const char block_sign_shuffle_mask_1[32] = { - 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, - 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, - }; - static const char block_sign_shuffle_mask_2[32] = { - 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, - 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, - }; - static const uint8_t bit_selector_mask_bytes[32] = { - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m256i bit_selector_mask = _mm256_loadu_si256((const __m256i*)bit_selector_mask_bytes); - const __m256i block_sign_shuffle_1 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_1); - const __m256i block_sign_shuffle_2 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_2); - - static const uint8_t k_bit_helper[32] = { - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - }; - const __m256i bit_helper = _mm256_loadu_si256((const __m256i*)k_bit_helper); - const __m256i m511 = _mm256_set1_epi16(511); - const __m128i m4 = _mm_set1_epi8(0xf); - const __m128i m1 = _mm_set1_epi8(1); - - uint64_t aux64; - - // somewhat hacky, but gives a significant boost in performance - __m256i aux_gindex; - const uint16_t * gindex = (const uint16_t *)&aux_gindex; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(&aux64, x[i].scales, 8); - __m128i stmp = _mm_set1_epi64x(aux64); - stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4)); - const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1); - - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { - - const __m256i q2_data = _mm256_loadu_si256((const __m256i*)q2); q2 += 16; - aux_gindex = _mm256_and_si256(q2_data, m511); - - const __m256i partial_sign_bits = _mm256_srli_epi16(q2_data, 9); - const __m256i partial_sign_bits_upper = _mm256_srli_epi16(q2_data, 13); - const __m256i partial_sign_bits_for_counting = _mm256_xor_si256(partial_sign_bits, partial_sign_bits_upper); - - const __m256i odd_bits = _mm256_shuffle_epi8(bit_helper, partial_sign_bits_for_counting); - const __m256i full_sign_bits = _mm256_or_si256(partial_sign_bits, odd_bits); - - const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_4 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - - const __m256i q2_1 = _mm256_set_epi64x(iq2xs_grid[gindex[ 3]], iq2xs_grid[gindex[ 2]], - iq2xs_grid[gindex[ 1]], iq2xs_grid[gindex[ 0]]); - const __m256i q2_2 = _mm256_set_epi64x(iq2xs_grid[gindex[ 7]], iq2xs_grid[gindex[ 6]], - iq2xs_grid[gindex[ 5]], iq2xs_grid[gindex[ 4]]); - const __m256i q2_3 = _mm256_set_epi64x(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]], - iq2xs_grid[gindex[ 9]], iq2xs_grid[gindex[ 8]]); - const __m256i q2_4 = _mm256_set_epi64x(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]], - iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); - - const __m128i full_signs_l = _mm256_castsi256_si128(full_sign_bits); - const __m128i full_signs_h = _mm256_extractf128_si256(full_sign_bits, 1); - const __m256i full_signs_1 = MM256_SET_M128I(full_signs_l, full_signs_l); - const __m256i full_signs_2 = MM256_SET_M128I(full_signs_h, full_signs_h); - - __m256i signs; - signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_1); - signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_1 = _mm256_sign_epi8(q8_1, _mm256_or_si256(signs, mone)); - - signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_2); - signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_2 = _mm256_sign_epi8(q8_2, _mm256_or_si256(signs, mone)); - - signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_1); - signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_3 = _mm256_sign_epi8(q8_3, _mm256_or_si256(signs, mone)); - - signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_2); - signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_4 = _mm256_sign_epi8(q8_4, _mm256_or_si256(signs, mone)); - - const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); - const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); - const __m256i dot3 = _mm256_maddubs_epi16(q2_3, q8s_3); - const __m256i dot4 = _mm256_maddubs_epi16(q2_4, q8s_4); - - const __m256i sc1 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0))); - const __m256i sc2 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1))); - const __m256i sc3 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+2))); - const __m256i sc4 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+3))); - - sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot1, sc1)); - sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot2, sc2)); - sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot3, sc3)); - sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot4, sc4)); - } - - accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__AVX__) - const __m128i mone = _mm_set1_epi8(1); - static const char block_sign_shuffle_mask_1[32] = { - 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, - 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, - }; - static const char block_sign_shuffle_mask_2[32] = { - 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, - 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, - }; - static const uint8_t bit_selector_mask_bytes[32] = { - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m128i bit_selector_mask_0 = _mm_loadu_si128((const __m128i*)bit_selector_mask_bytes); - const __m128i bit_selector_mask_1 = _mm_loadu_si128((const __m128i*)bit_selector_mask_bytes + 1); - const __m128i block_sign_shuffle_1_0 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_1); - const __m128i block_sign_shuffle_1_1 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_1 + 1); - const __m128i block_sign_shuffle_2_0 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_2); - const __m128i block_sign_shuffle_2_1 = _mm_loadu_si128((const __m128i*)block_sign_shuffle_mask_2 + 1); - - static const uint8_t k_bit_helper[32] = { - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - }; - const __m128i bit_helper_0 = _mm_loadu_si128((const __m128i*)k_bit_helper); - const __m128i bit_helper_1 = _mm_loadu_si128((const __m128i*)k_bit_helper + 1); - const __m128i m511 = _mm_set1_epi16(511); - const __m128i m4 = _mm_set1_epi8(0xf); - const __m128i m1 = _mm_set1_epi8(1); - - uint64_t aux64; - - // somewhat hacky, but gives a significant boost in performance - __m256i aux_gindex; - const uint16_t * gindex = (const uint16_t *)&aux_gindex; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(&aux64, x[i].scales, 8); - __m128i stmp = _mm_set1_epi64x(aux64); - stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4)); - const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1); - - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { - - const __m128i q2_data_0 = _mm_loadu_si128((const __m128i*)q2); - const __m128i q2_data_1 = _mm_loadu_si128((const __m128i*)q2 + 1); q2 += 16; - aux_gindex = MM256_SET_M128I(_mm_and_si128(q2_data_1, m511), _mm_and_si128(q2_data_0, m511)); - - const __m128i partial_sign_bits_0 = _mm_srli_epi16(q2_data_0, 9); - const __m128i partial_sign_bits_1 = _mm_srli_epi16(q2_data_1, 9); - const __m128i partial_sign_bits_upper_0 = _mm_srli_epi16(q2_data_0, 13); - const __m128i partial_sign_bits_upper_1 = _mm_srli_epi16(q2_data_1, 13); - const __m128i partial_sign_bits_for_counting_0 = _mm_xor_si128(partial_sign_bits_0, partial_sign_bits_upper_0); - const __m128i partial_sign_bits_for_counting_1 = _mm_xor_si128(partial_sign_bits_1, partial_sign_bits_upper_1); - - const __m128i odd_bits_0 = _mm_shuffle_epi8(bit_helper_0, partial_sign_bits_for_counting_0); - const __m128i odd_bits_1 = _mm_shuffle_epi8(bit_helper_1, partial_sign_bits_for_counting_1); - const __m128i full_sign_bits_0 = _mm_or_si128(partial_sign_bits_0, odd_bits_0); - const __m128i full_sign_bits_1 = _mm_or_si128(partial_sign_bits_1, odd_bits_1); - - const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_3_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_3_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_4_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_4_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - - const __m128i q2_1_0 = _mm_set_epi64x(iq2xs_grid[gindex[1]], iq2xs_grid[gindex[0]]); - const __m128i q2_1_1 = _mm_set_epi64x(iq2xs_grid[gindex[3]], iq2xs_grid[gindex[2]]); - const __m128i q2_2_0 = _mm_set_epi64x(iq2xs_grid[gindex[5]], iq2xs_grid[gindex[4]]); - const __m128i q2_2_1 = _mm_set_epi64x(iq2xs_grid[gindex[7]], iq2xs_grid[gindex[6]]); - const __m128i q2_3_0 = _mm_set_epi64x(iq2xs_grid[gindex[9]], iq2xs_grid[gindex[8]]); - const __m128i q2_3_1 = _mm_set_epi64x(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]]); - const __m128i q2_4_0 = _mm_set_epi64x(iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); - const __m128i q2_4_1 = _mm_set_epi64x(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]]); - - // AVX2 full_signs_1 is full_sign_bits_0 here - // AVX2 full_signs_2 is full_sign_bits_1 here - __m128i signs_0, signs_1; - signs_0 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_1_0); - signs_1 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_1_1); - signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); - signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); - const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, _mm_or_si128(signs_0, mone)); - const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, _mm_or_si128(signs_1, mone)); - - signs_0 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_2_0); - signs_1 = _mm_shuffle_epi8(full_sign_bits_0, block_sign_shuffle_2_1); - signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); - signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); - const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, _mm_or_si128(signs_0, mone)); - const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, _mm_or_si128(signs_1, mone)); - - signs_0 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_1_0); - signs_1 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_1_1); - signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); - signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); - const __m128i q8s_3_0 = _mm_sign_epi8(q8_3_0, _mm_or_si128(signs_0, mone)); - const __m128i q8s_3_1 = _mm_sign_epi8(q8_3_1, _mm_or_si128(signs_1, mone)); - - signs_0 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_2_0); - signs_1 = _mm_shuffle_epi8(full_sign_bits_1, block_sign_shuffle_2_1); - signs_0 = _mm_cmpeq_epi8(_mm_and_si128(signs_0, bit_selector_mask_0), bit_selector_mask_0); - signs_1 = _mm_cmpeq_epi8(_mm_and_si128(signs_1, bit_selector_mask_1), bit_selector_mask_1); - const __m128i q8s_4_0 = _mm_sign_epi8(q8_4_0, _mm_or_si128(signs_0, mone)); - const __m128i q8s_4_1 = _mm_sign_epi8(q8_4_1, _mm_or_si128(signs_1, mone)); - - const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); - const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); - const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); - const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); - const __m128i dot3_0 = _mm_maddubs_epi16(q2_3_0, q8s_3_0); - const __m128i dot3_1 = _mm_maddubs_epi16(q2_3_1, q8s_3_1); - const __m128i dot4_0 = _mm_maddubs_epi16(q2_4_0, q8s_4_0); - const __m128i dot4_1 = _mm_maddubs_epi16(q2_4_1, q8s_4_1); - - __m128i sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0)); - const __m128i sc1_0 = _mm_cvtepi8_epi16(sc_tmp); - const __m128i sc1_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); - sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1)); - const __m128i sc2_0 = _mm_cvtepi8_epi16(sc_tmp); - const __m128i sc2_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); - sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+2)); - const __m128i sc3_0 = _mm_cvtepi8_epi16(sc_tmp); - const __m128i sc3_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); - sc_tmp = _mm_shuffle_epi8(scales, get_scale_shuffle(ib32+3)); - const __m128i sc4_0 = _mm_cvtepi8_epi16(sc_tmp); - const __m128i sc4_1 = _mm_cvtepi8_epi16(_mm_srli_si128(sc_tmp, 8)); - - sumi1_0 = _mm_add_epi32(sumi1_0, _mm_madd_epi16(dot1_0, sc1_0)); - sumi1_1 = _mm_add_epi32(sumi1_1, _mm_madd_epi16(dot1_1, sc1_1)); - sumi2_0 = _mm_add_epi32(sumi2_0, _mm_madd_epi16(dot2_0, sc2_0)); - sumi2_1 = _mm_add_epi32(sumi2_1, _mm_madd_epi16(dot2_1, sc2_1)); - sumi1_0 = _mm_add_epi32(sumi1_0, _mm_madd_epi16(dot3_0, sc3_0)); - sumi1_1 = _mm_add_epi32(sumi1_1, _mm_madd_epi16(dot3_1, sc3_1)); - sumi2_0 = _mm_add_epi32(sumi2_0, _mm_madd_epi16(dot4_0, sc4_0)); - sumi2_1 = _mm_add_epi32(sumi2_1, _mm_madd_epi16(dot4_1, sc4_1)); - } - - accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__loongarch_asx) - - const __m256i mone = __lasx_xvreplgr2vr_b(1); - static const char block_sign_shuffle_mask_1[32] = { - 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, - 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, - }; - static const char block_sign_shuffle_mask_2[32] = { - 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, - 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, - }; - static const uint8_t bit_selector_mask_bytes[32] = { - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m256i bit_selector_mask = __lasx_xvld((const __m256i*)bit_selector_mask_bytes, 0); - const __m256i block_sign_shuffle_1 = __lasx_xvld((const __m256i*)block_sign_shuffle_mask_1, 0); - const __m256i block_sign_shuffle_2 = __lasx_xvld((const __m256i*)block_sign_shuffle_mask_2, 0); - - static const uint8_t k_bit_helper[32] = { - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, - }; - const __m256i bit_helper = __lasx_xvld((const __m256i*)k_bit_helper, 0); - const __m256i m511 = __lasx_xvreplgr2vr_h(511); - const __m128i m4 = __lsx_vreplgr2vr_b(0xf); - const __m128i m1 = __lsx_vreplgr2vr_b(1); - - uint64_t aux64; - - // somewhat hacky, but gives a significant boost in performance - __m256i aux_gindex; - const uint16_t * gindex = (const uint16_t *)&aux_gindex; - - __m256 accumf = (__m256)__lasx_xvldi(0); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(&aux64, x[i].scales, 8); - __m128i stmp = __lsx_vreplgr2vr_d(aux64); - stmp = __lsx_vilvl_b( __lsx_vand_v(__lsx_vsrli_h(stmp, 4), m4), __lsx_vand_v(stmp, m4)); - const __m128i scales = __lsx_vadd_b(__lsx_vslli_h(stmp, 1), m1); - - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { - - const __m256i q2_data = __lasx_xvld((const __m256i*)q2, 0); q2 += 16; - aux_gindex = __lasx_xvand_v(q2_data, m511); - - const __m256i partial_sign_bits = __lasx_xvsrli_h(q2_data, 9); - const __m256i partial_sign_bits_upper = __lasx_xvsrli_h(q2_data, 13); - const __m256i partial_sign_bits_for_counting = __lasx_xvxor_v(partial_sign_bits, partial_sign_bits_upper); - - const __m256i odd_bits = lasx_shuffle_b(bit_helper, partial_sign_bits_for_counting); - const __m256i full_sign_bits = __lasx_xvor_v(partial_sign_bits, odd_bits); - - const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_3 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_4 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - - const __m256i q2_1 = lasx_set_d(iq2xs_grid[gindex[ 3]], iq2xs_grid[gindex[ 2]], - iq2xs_grid[gindex[ 1]], iq2xs_grid[gindex[ 0]]); - const __m256i q2_2 = lasx_set_d(iq2xs_grid[gindex[ 7]], iq2xs_grid[gindex[ 6]], - iq2xs_grid[gindex[ 5]], iq2xs_grid[gindex[ 4]]); - const __m256i q2_3 = lasx_set_d(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]], - iq2xs_grid[gindex[ 9]], iq2xs_grid[gindex[ 8]]); - const __m256i q2_4 = lasx_set_d(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]], - iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); - - const __m128i full_signs_l = lasx_extracti128(full_sign_bits, 0); - const __m128i full_signs_h = lasx_extracti128(full_sign_bits, 1); - const __m256i full_signs_1 = lasx_insertf128(full_signs_l, full_signs_l); - const __m256i full_signs_2 = lasx_insertf128(full_signs_h, full_signs_h); - - __m256i signs; - signs = lasx_shuffle_b(full_signs_1, block_sign_shuffle_1); - signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_1 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_1); - - signs = lasx_shuffle_b(full_signs_1, block_sign_shuffle_2); - signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_2 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_2); - - signs = lasx_shuffle_b(full_signs_2, block_sign_shuffle_1); - signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_3 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_3); - - signs = lasx_shuffle_b(full_signs_2, block_sign_shuffle_2); - signs = __lasx_xvseq_b(__lasx_xvand_v(signs, bit_selector_mask), bit_selector_mask); - const __m256i q8s_4 = __lasx_xvsigncov_b(__lasx_xvor_v(signs, mone), q8_4); - - const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); - const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); - const __m256i dot3 = lasx_maddubs_h(q2_3, q8s_3); - const __m256i dot4 = lasx_maddubs_h(q2_4, q8s_4); - - const __m256i sc1 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+0))); - const __m256i sc2 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+1))); - const __m256i sc3 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+2))); - const __m256i sc4 = lasx_ext8_16(lsx_shuffle_b(scales, get_scale_shuffle(ib32+3))); - - sumi1 = __lasx_xvadd_w(sumi1, lasx_madd_h(dot1, sc1)); - sumi2 = __lasx_xvadd_w(sumi2, lasx_madd_h(dot2, sc2)); - sumi1 = __lasx_xvadd_w(sumi1, lasx_madd_h(dot3, sc3)); - sumi2 = __lasx_xvadd_w(sumi2, lasx_madd_h(dot4, sc4)); - } - - accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); -#elif defined(__POWER9_VECTOR__) - const vector int v0 = vec_splats((int32_t)0); - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint16_t * restrict q2 = x[i].qs; - const uint8_t * restrict sc = x[i].scales; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/64; ++j) { - __builtin_prefetch(q2, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed long long aux64x2_0 = {*(const int64_t *)(iq2xs_grid + (q2[0] & 511)), *(const int64_t *)(iq2xs_grid + (q2[1] & 511))}; - vector signed long long aux64x2_1 = {*(const int64_t *)(iq2xs_grid + (q2[2] & 511)), *(const int64_t *)(iq2xs_grid + (q2[3] & 511))}; - vector signed long long aux64x2_2 = {*(const int64_t *)(iq2xs_grid + (q2[4] & 511)), *(const int64_t *)(iq2xs_grid + (q2[5] & 511))}; - vector signed long long aux64x2_3 = {*(const int64_t *)(iq2xs_grid + (q2[6] & 511)), *(const int64_t *)(iq2xs_grid + (q2[7] & 511))}; - - vector signed long long vsigns0 = {*(const int64_t *)(signs64 + ((q2[0] >> 9))), *(const int64_t *)(signs64 + ((q2[1] >> 9)))}; - vector signed long long vsigns1 = {*(const int64_t *)(signs64 + ((q2[2] >> 9))), *(const int64_t *)(signs64 + ((q2[3] >> 9)))}; - vector signed long long vsigns2 = {*(const int64_t *)(signs64 + ((q2[4] >> 9))), *(const int64_t *)(signs64 + ((q2[5] >> 9)))}; - vector signed long long vsigns3 = {*(const int64_t *)(signs64 + ((q2[6] >> 9))), *(const int64_t *)(signs64 + ((q2[7] >> 9)))}; - q2 += 8; - - vector signed char q2x0 = (vector signed char)vec_mul((vector signed char)vsigns0, (vector signed char)aux64x2_0); - vector signed char q2x1 = (vector signed char)vec_mul((vector signed char)vsigns1, (vector signed char)aux64x2_1); - vector signed char q2x2 = (vector signed char)vec_mul((vector signed char)vsigns2, (vector signed char)aux64x2_2); - vector signed char q2x3 = (vector signed char)vec_mul((vector signed char)vsigns3, (vector signed char)aux64x2_3); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); - - const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); - const uint16_t ls1 = (uint16_t)(sc[0] >> 4); - const uint16_t ls2 = (uint16_t)(sc[1] & 0xf); - const uint16_t ls3 = (uint16_t)(sc[1] >> 4); - sc += 2; - - vector signed short vscales0 = vec_splats((int16_t)(2*ls0+1)); - vector signed short vscales1 = vec_splats((int16_t)(2*ls1+1)); - vector signed short vscales2 = vec_splats((int16_t)(2*ls2+1)); - vector signed short vscales3 = vec_splats((int16_t)(2*ls3+1)); - - vsumi0 = vec_msum(qv0, vscales0, vsumi0); - vsumi1 = vec_msum(qv1, vscales1, vsumi1); - vsumi2 = vec_msum(qv2, vscales2, vsumi2); - vsumi3 = vec_msum(qv3, vscales3, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = 0.125f * vec_extract(vsumf0, 0); -#else - - float sumf = 0.f; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint16_t * restrict q2 = x[i].qs; - const uint8_t * restrict sc = x[i].scales; - const int8_t * restrict q8 = y[i].qs; - int32_t bsum = 0; - for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { - const uint16_t ls1 = 2*(sc[ib32] & 0xf) + 1; - const uint16_t ls2 = 2*(sc[ib32] >> 4) + 1; - int32_t sumi = 0; - for (int l = 0; l < 2; ++l) { - const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); - const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; - for (int j = 0; j < 8; ++j) { - sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); - } - q8 += 8; - } - bsum += sumi * ls1; - sumi = 0; - for (int l = 2; l < 4; ++l) { - const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); - const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; - for (int j = 0; j < 8; ++j) { - sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); - } - q8 += 8; - } - bsum += sumi * ls2; - q2 += 4; - } - sumf += d * bsum; - } - *s = 0.125f * sumf; -#endif -} - -void ggml_vec_dot_iq2_s_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq2_s * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - - const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); - const uint8x16_t mask2 = vld1q_u8(k_mask2); - const uint8x16_t m1 = vdupq_n_u8(1); - const int32x4_t vzero = vdupq_n_s32(0); - - uint8x16x2_t vs; - ggml_int8x16x4_t q2s; - ggml_int8x16x4_t q8b; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); - const int8_t * restrict q8 = y[i].qs; - - int sumi1 = 0, sumi2 = 0; - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - q2s.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[0] | ((qh[ib32+0] << 8) & 0x300)))), - vld1_s8((const int8_t *)(iq2s_grid + (qs[1] | ((qh[ib32+0] << 6) & 0x300))))); - q2s.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[2] | ((qh[ib32+0] << 4) & 0x300)))), - vld1_s8((const int8_t *)(iq2s_grid + (qs[3] | ((qh[ib32+0] << 2) & 0x300))))); - q2s.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[4] | ((qh[ib32+1] << 8) & 0x300)))), - vld1_s8((const int8_t *)(iq2s_grid + (qs[5] | ((qh[ib32+1] << 6) & 0x300))))); - q2s.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[6] | ((qh[ib32+1] << 4) & 0x300)))), - vld1_s8((const int8_t *)(iq2s_grid + (qs[7] | ((qh[ib32+1] << 2) & 0x300))))); - qs += 8; - - vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); - vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); - vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); - vs.val[0] = vceqq_u8(vs.val[0], mask2); - vs.val[1] = vceqq_u8(vs.val[1], mask2); - - q2s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[0]); - q2s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[1]); - - vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); - vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); - vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); - vs.val[0] = vceqq_u8(vs.val[0], mask2); - vs.val[1] = vceqq_u8(vs.val[1], mask2); - - signs += 4; - - q2s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[2]); - q2s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[3]); - - const int32x4_t p1 = ggml_vdotq_s32(vzero, q2s.val[0], q8b.val[0]); - const int32x4_t p2 = ggml_vdotq_s32(vzero, q2s.val[1], q8b.val[1]); - const int32x4_t p3 = ggml_vdotq_s32(vzero, q2s.val[2], q8b.val[2]); - const int32x4_t p4 = ggml_vdotq_s32(vzero, q2s.val[3], q8b.val[3]); - - sumi1 += vaddvq_s32(p1) * (1 + 2*(x[i].scales[ib32+0] & 0xf)); - sumi2 += vaddvq_s32(p2) * (1 + 2*(x[i].scales[ib32+0] >> 4)); - sumi1 += vaddvq_s32(p3) * (1 + 2*(x[i].scales[ib32+1] & 0xf)); - sumi2 += vaddvq_s32(p4) * (1 + 2*(x[i].scales[ib32+1] >> 4)); - } - sumf += d*(sumi1 + sumi2); - } - - *s = 0.125f * sumf; - -#elif defined(__AVX2__) - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m128i m4 = _mm_set1_epi8(0xf); - const __m128i m1 = _mm_set1_epi8(1); - - const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); - const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); - - uint64_t aux64; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); - const int8_t * restrict q8 = y[i].qs; - - memcpy(&aux64, x[i].scales, 8); - const __m128i scales8 = _mm_add_epi8(_mm_slli_epi16(_mm_and_si128(_mm_set_epi64x(aux64 >> 4, aux64), m4), 1), m1); - const __m256i scales16 = _mm256_cvtepi8_epi16(scales8); // 0 2 4 6 8 10 12 14 1 3 5 7 9 11 13 15 - - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q2_1 = _mm256_set_epi64x(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], - iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)], - iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], - iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); - const __m256i q2_2 = _mm256_set_epi64x(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], - iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)], - iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], - iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); - qs += 8; - - __m256i aux256 = _mm256_set1_epi32(signs[0] | ((uint32_t) signs[1] << 16)); - aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); - const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); - const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); - - aux256 = _mm256_set1_epi32(signs[2] | ((uint32_t) signs[3] << 16)); - aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); - const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); - const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); - - signs += 4; - - const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); // blocks 2*ib32+0, 2*ib32+1 - const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); // blocks 2*ib32+2, 2*ib32+3 - - const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+0))); - const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+1))); - sumi1 = _mm256_add_epi32(sumi1, p1); - sumi2 = _mm256_add_epi32(sumi2, p2); - } - - accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__AVX__) - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m128i m4 = _mm_set1_epi8(0xf); - const __m128i m1 = _mm_set1_epi8(1); - - const __m128i mask1_0 = _mm_loadu_si128((const __m128i*)k_mask1); - const __m128i mask1_1 = _mm_loadu_si128((const __m128i*)k_mask1 + 1); - const __m128i mask2_0 = _mm_loadu_si128((const __m128i*)k_mask2); - const __m128i mask2_1 = _mm_loadu_si128((const __m128i*)k_mask2 + 1); - - uint64_t aux64; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); - const int8_t * restrict q8 = y[i].qs; - - memcpy(&aux64, x[i].scales, 8); - const __m128i scales8 = _mm_add_epi8(_mm_slli_epi16(_mm_and_si128(_mm_set_epi64x(aux64 >> 4, aux64), m4), 1), m1); - const __m128i scales16_0 = _mm_cvtepi8_epi16(scales8); - const __m128i scales16_1 = _mm_cvtepi8_epi16(_mm_srli_si128(scales8, 8)); - - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q2_1_0 = _mm_set_epi64x(iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], - iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); - const __m128i q2_1_1 = _mm_set_epi64x(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], - iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)]); - const __m128i q2_2_0 = _mm_set_epi64x(iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], - iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); - const __m128i q2_2_1 = _mm_set_epi64x(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], - iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)]); - qs += 8; - - __m128i aux128_0 = _mm_set1_epi32(signs[0] | ((uint32_t) signs[1] << 16)); - __m128i aux128_1 = aux128_0; - aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); - aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); - const __m128i s2_1_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); - const __m128i s2_1_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); - const __m128i q8s_1_0 = _mm_sub_epi8(_mm_xor_si128(s2_1_0, q8_1_0), s2_1_0); - const __m128i q8s_1_1 = _mm_sub_epi8(_mm_xor_si128(s2_1_1, q8_1_1), s2_1_1); - - aux128_0 = _mm_set1_epi32(signs[2] | ((uint32_t) signs[3] << 16)); - aux128_1 = aux128_0; - aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); - aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); - const __m128i s2_2_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); - const __m128i s2_2_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); - const __m128i q8s_2_0 = _mm_sub_epi8(_mm_xor_si128(s2_2_0, q8_2_0), s2_2_0); - const __m128i q8s_2_1 = _mm_sub_epi8(_mm_xor_si128(s2_2_1, q8_2_1), s2_2_1); - - signs += 4; - - const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); - const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); - const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); - const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); - - const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_shuffle_epi8(scales16_0, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+0), 0))); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_shuffle_epi8(scales16_1, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+0), 1))); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_shuffle_epi8(scales16_0, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+1), 0))); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_shuffle_epi8(scales16_1, _mm256_extractf128_si256(get_scale_shuffle_k4(ib32+1), 1))); - sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); - sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); - sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); - sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); - } - - accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); - - } - - *s = 0.125f * hsum_float_8(accumf); - -#elif defined(__POWER9_VECTOR__) - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - - const vector int v0 = vec_splats((int32_t)0); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - const vector unsigned char mask0 = vec_xl( 0, k_mask1); - const vector unsigned char mask1 = vec_xl(16, k_mask1); - const vector signed char mask2 = (vector signed char)vec_xl( 0, k_mask2); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint8_t * restrict q2 = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); - const uint8_t * restrict sc = x[i].scales; - const int8_t * restrict q8 = y[i].qs; - - for (int j = 0; j < QK_K/32; j += 2) { - __builtin_prefetch(q2, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed long long aux64x2_0 = {*(const int64_t *)(iq2s_grid + (q2[0] | ((qh[0] << 8) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[1] | ((qh[0] << 6) & 0x300)))}; - vector signed long long aux64x2_1 = {*(const int64_t *)(iq2s_grid + (q2[2] | ((qh[0] << 4) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[3] | ((qh[0] << 2) & 0x300)))}; - vector signed long long aux64x2_2 = {*(const int64_t *)(iq2s_grid + (q2[4] | ((qh[1] << 8) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[5] | ((qh[1] << 6) & 0x300)))}; - vector signed long long aux64x2_3 = {*(const int64_t *)(iq2s_grid + (q2[6] | ((qh[1] << 4) & 0x300))), *(const int64_t *)(iq2s_grid + (q2[7] | ((qh[1] << 2) & 0x300)))}; - q2 += 8; - qh += 2; - - vector signed char vsigns01 = (vector signed char)vec_splats(*(const uint32_t *)&signs[0]); - vector signed char vsigns23 = (vector signed char)vec_splats(*(const uint32_t *)&signs[2]); - signs += 4; - - vector signed char vsigns0 = vec_perm(vsigns01, vsigns01, mask0); - vector signed char vsigns1 = vec_perm(vsigns01, vsigns01, mask1); - vector signed char vsigns2 = vec_perm(vsigns23, vsigns23, mask0); - vector signed char vsigns3 = vec_perm(vsigns23, vsigns23, mask1); - - vsigns0 = (vector signed char)vec_cmpeq(vec_and(vsigns0, mask2), mask2); - vsigns1 = (vector signed char)vec_cmpeq(vec_and(vsigns1, mask2), mask2); - vsigns2 = (vector signed char)vec_cmpeq(vec_and(vsigns2, mask2), mask2); - vsigns3 = (vector signed char)vec_cmpeq(vec_and(vsigns3, mask2), mask2); - - vector signed char q2x0 = vec_sub(vec_xor(vsigns0, (vector signed char)aux64x2_0), vsigns0); - vector signed char q2x1 = vec_sub(vec_xor(vsigns1, (vector signed char)aux64x2_1), vsigns1); - vector signed char q2x2 = vec_sub(vec_xor(vsigns2, (vector signed char)aux64x2_2), vsigns2); - vector signed char q2x3 = vec_sub(vec_xor(vsigns3, (vector signed char)aux64x2_3), vsigns3); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q2x0, q8y0), vec_mulo(q2x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q2x1, q8y1), vec_mulo(q2x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q2x2, q8y2), vec_mulo(q2x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q2x3, q8y3), vec_mulo(q2x3, q8y3)); - - const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); - const uint16_t ls1 = (uint16_t)(sc[0] >> 4); - const uint16_t ls2 = (uint16_t)(sc[1] & 0xf); - const uint16_t ls3 = (uint16_t)(sc[1] >> 4); - sc += 2; - - vector signed short vscales0 = vec_splats((int16_t)(2*ls0+1)); - vector signed short vscales1 = vec_splats((int16_t)(2*ls1+1)); - vector signed short vscales2 = vec_splats((int16_t)(2*ls2+1)); - vector signed short vscales3 = vec_splats((int16_t)(2*ls3+1)); - - vsumi0 = vec_msum(qv0, vscales0, vsumi0); - vsumi1 = vec_msum(qv1, vscales1, vsumi1); - vsumi2 = vec_msum(qv2, vscales2, vsumi2); - vsumi3 = vec_msum(qv3, vscales3, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = 0.125f * vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - - const __m128i m4 = __lsx_vreplgr2vr_b(0xf); - const __m128i m1 = __lsx_vreplgr2vr_b(1); - - const __m256i mask1 = __lasx_xvld((const __m256i*)k_mask1, 0); - const __m256i mask2 = __lasx_xvld((const __m256i*)k_mask2, 0); - uint64_t aux64; - - __m256 accumf = (__m256)__lasx_xvldi(0); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); - const int8_t * restrict q8 = y[i].qs; - - __m128i tmp1; - memcpy(&aux64, x[i].scales, 8); - tmp1 = __lsx_vinsgr2vr_d(tmp1, aux64, 0); - tmp1 = __lsx_vinsgr2vr_d(tmp1, aux64 >> 4, 1); - const __m128i scales8 = __lsx_vadd_b(__lsx_vslli_h(__lsx_vand_v(tmp1, m4), 1), m1); - const __m256i scales16 = lasx_ext8_16(scales8); // 0 2 4 6 8 10 12 14 1 3 5 7 9 11 13 15 - - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q2_1 = lasx_set_d(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], - iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)], - iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], - iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); - const __m256i q2_2 = lasx_set_d(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], - iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)], - iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], - iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); - qs += 8; - - __m256i aux256 = __lasx_xvreplgr2vr_w(signs[0] | ((uint32_t) signs[1] << 16)); - aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); - const __m256i s2_1 = __lasx_xvseq_b(aux256, mask2); - const __m256i q8s_1 = __lasx_xvsub_b(__lasx_xvxor_v(s2_1, q8_1), s2_1); - - aux256 = __lasx_xvreplgr2vr_w(signs[2] | ((uint32_t) signs[3] << 16)); - aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); - const __m256i s2_2 = __lasx_xvseq_b(aux256, mask2); - const __m256i q8s_2 = __lasx_xvsub_b(__lasx_xvxor_v(s2_2, q8_2), s2_2); - - signs += 4; - - const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); // blocks 2*ib32+0, 2*ib32+1 - const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); // blocks 2*ib32+2, 2*ib32+3 - - const __m256i p1 = lasx_madd_h(dot1, lasx_shuffle_b(scales16, get_scale_shuffle_k4(ib32+0))); - const __m256i p2 = lasx_madd_h(dot2, lasx_shuffle_b(scales16, get_scale_shuffle_k4(ib32+1))); - sumi1 = __lasx_xvadd_w(sumi1, p1); - sumi2 = __lasx_xvadd_w(sumi2, p2); - } - - accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); - } - - *s = 0.125f * hsum_float_8(accumf); - -#else - - float sumf = 0; - for (int i = 0; i < nb; i++) { - - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint8_t * qh = x[i].qh; - const uint8_t * signs = qs + QK_K/8; - - int bsum = 0; - for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { - int ls1 = 1 + 2*(x[i].scales[ib32] & 0xf); - int ls2 = 1 + 2*(x[i].scales[ib32] >> 4); - int sumi1 = 0, sumi2 = 0; - for (int l = 0; l < 2; ++l) { - const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); - for (int j = 0; j < 8; ++j) { - sumi1 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); - } - q8 += 8; - } - for (int l = 2; l < 4; ++l) { - const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); - for (int j = 0; j < 8; ++j) { - sumi2 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); - } - q8 += 8; - } - bsum += ls1 * sumi1 + ls2 * sumi2; - qs += 4; - signs += 4; - } - - sumf += d * bsum; - } - - *s = 0.125f * sumf; - -#endif - -} - -void ggml_vec_dot_iq3_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq3_xxs * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[2]; - - ggml_int8x16x4_t q3s; - ggml_int8x16x4_t q8b; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict gas = x[i].qs + QK_K/4; - const int8_t * restrict q8 = y[i].qs; - float sumf1 = 0, sumf2 = 0; - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - memcpy(aux32, gas, 2*sizeof(uint32_t)); gas += 2*sizeof(uint32_t); - const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3xxs_grid[q3[ 0]], iq3xxs_grid[q3[ 1]], iq3xxs_grid[q3[ 2]], iq3xxs_grid[q3[ 3]]); - const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3xxs_grid[q3[ 4]], iq3xxs_grid[q3[ 5]], iq3xxs_grid[q3[ 6]], iq3xxs_grid[q3[ 7]]); - const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3xxs_grid[q3[ 8]], iq3xxs_grid[q3[ 9]], iq3xxs_grid[q3[10]], iq3xxs_grid[q3[11]]); - const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3xxs_grid[q3[12]], iq3xxs_grid[q3[13]], iq3xxs_grid[q3[14]], iq3xxs_grid[q3[15]]); - q3 += 16; - q3s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 7) & 127)))); - q3s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 21) & 127)))); - q3s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); - q3s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); - q3s.val[0] = vmulq_s8(q3s.val[0], vreinterpretq_s8_u32(aux32x4_0)); - q3s.val[1] = vmulq_s8(q3s.val[1], vreinterpretq_s8_u32(aux32x4_1)); - q3s.val[2] = vmulq_s8(q3s.val[2], vreinterpretq_s8_u32(aux32x4_2)); - q3s.val[3] = vmulq_s8(q3s.val[3], vreinterpretq_s8_u32(aux32x4_3)); - const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); - const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); - sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[0] >> 28)); - sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[1] >> 28)); - } - sumf += d*(sumf1 + sumf2); - } - *s = 0.5f * sumf; - -#elif defined(__AVX2__) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[2]; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict gas = x[i].qs + QK_K/4; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q2_1 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], - iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - q3 += 8; - const __m256i q2_2 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], - iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - q3 += 8; - memcpy(aux32, gas, 8); gas += 8; - const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127], - signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); - const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], - signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); - const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); - const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); - const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); - const uint16_t ls1 = aux32[0] >> 28; - const uint16_t ls2 = aux32[1] >> 28; - const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); - const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); - sumi1 = _mm256_add_epi32(sumi1, p1); - sumi2 = _mm256_add_epi32(sumi2, p2); - } - - accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - - } - - *s = 0.25f * hsum_float_8(accumf); - -#elif defined(__AVX__) - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[2]; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict gas = x[i].qs + QK_K/4; - const int8_t * restrict q8 = y[i].qs; - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q2_1_0 = _mm_set_epi32(iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - const __m128i q2_1_1 = _mm_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]]); - q3 += 8; - const __m128i q2_2_0 = _mm_set_epi32(iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - const __m128i q2_2_1 = _mm_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]]); - q3 += 8; - memcpy(aux32, gas, 8); gas += 8; - const __m128i s2_1_0 = _mm_set_epi64x(signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); - const __m128i s2_1_1 = _mm_set_epi64x(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127]); - const __m128i s2_2_0 = _mm_set_epi64x(signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m128i s2_2_1 = _mm_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127]); - const __m128i q8s_1_0 = _mm_sign_epi8(q8_1_0, s2_1_0); - const __m128i q8s_1_1 = _mm_sign_epi8(q8_1_1, s2_1_1); - const __m128i q8s_2_0 = _mm_sign_epi8(q8_2_0, s2_2_0); - const __m128i q8s_2_1 = _mm_sign_epi8(q8_2_1, s2_2_1); - const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); - const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); - const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); - const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); - const uint16_t ls1 = aux32[0] >> 28; - const uint16_t ls2 = aux32[1] >> 28; - const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); - sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); - sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); - sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); - sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); - } - - accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); - - } - - *s = 0.25f * hsum_float_8(accumf); - -#elif defined(__POWER9_VECTOR__) - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - const vector int v0 = vec_splats((int32_t)0); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - const uint8_t * restrict q3 = x[i].qs; - const uint32_t * restrict signs = (const uint32_t *)(x[i].qs + QK_K/4); - const int8_t * restrict q8 = y[i].qs; - -#pragma GCC unroll 1 - for (int j = 0; j < QK_K/32; j += 2) { - __builtin_prefetch(q3, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector unsigned int aux32x4_0 = {iq3xxs_grid[q3[ 0]], iq3xxs_grid[q3[ 1]], iq3xxs_grid[q3[ 2]], iq3xxs_grid[q3[ 3]]}; - vector unsigned int aux32x4_1 = {iq3xxs_grid[q3[ 4]], iq3xxs_grid[q3[ 5]], iq3xxs_grid[q3[ 6]], iq3xxs_grid[q3[ 7]]}; - vector unsigned int aux32x4_2 = {iq3xxs_grid[q3[ 8]], iq3xxs_grid[q3[ 9]], iq3xxs_grid[q3[10]], iq3xxs_grid[q3[11]]}; - vector unsigned int aux32x4_3 = {iq3xxs_grid[q3[12]], iq3xxs_grid[q3[13]], iq3xxs_grid[q3[14]], iq3xxs_grid[q3[15]]}; - q3 += 16; - - vector unsigned long long aux64x2_0 = {(uint64_t)(signs64[(signs[0] >> 0) & 127]), (uint64_t)(signs64[(signs[0] >> 7) & 127])}; - vector unsigned long long aux64x2_1 = {(uint64_t)(signs64[(signs[0] >> 14) & 127]), (uint64_t)(signs64[(signs[0] >> 21) & 127])}; - vector unsigned long long aux64x2_2 = {(uint64_t)(signs64[(signs[1] >> 0) & 127]), (uint64_t)(signs64[(signs[1] >> 7) & 127])}; - vector unsigned long long aux64x2_3 = {(uint64_t)(signs64[(signs[1] >> 14) & 127]), (uint64_t)(signs64[(signs[1] >> 21) & 127])}; - - vector signed char q3x0 = vec_mul((vector signed char)aux64x2_0, (vector signed char)aux32x4_0); - vector signed char q3x1 = vec_mul((vector signed char)aux64x2_1, (vector signed char)aux32x4_1); - vector signed char q3x2 = vec_mul((vector signed char)aux64x2_2, (vector signed char)aux32x4_2); - vector signed char q3x3 = vec_mul((vector signed char)aux64x2_3, (vector signed char)aux32x4_3); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q3x0, q8y0), vec_mulo(q3x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q3x1, q8y1), vec_mulo(q3x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q3x2, q8y2), vec_mulo(q3x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q3x3, q8y3), vec_mulo(q3x3, q8y3)); - - const uint16_t ls0 = (uint16_t)(signs[0] >> 28); - const uint16_t ls1 = (uint16_t)(signs[1] >> 28); - signs += 2; - - vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); - vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); - - vsumi0 = vec_msum(qv0, vscales01, vsumi0); - vsumi1 = vec_msum(qv1, vscales01, vsumi1); - vsumi2 = vec_msum(qv2, vscales23, vsumi2); - vsumi3 = vec_msum(qv3, vscales23, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = 0.25f * vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - - uint32_t aux32[2]; - - __m256 accumf = (__m256)__lasx_xvldi(0); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict gas = x[i].qs + QK_K/4; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q2_1 = lasx_set_w(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], - iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - q3 += 8; - const __m256i q2_2 = lasx_set_w(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], - iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); - q3 += 8; - memcpy(aux32, gas, 8); gas += 8; - - const __m256i s2_1 = lasx_set_d(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127], - signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); - const __m256i s2_2 = lasx_set_d(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], - signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); - const __m256i q8s_1 = __lasx_xvsigncov_b(s2_1, q8_1); - const __m256i q8s_2 = __lasx_xvsigncov_b(s2_2, q8_2); - const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); - const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); - const uint16_t ls1 = aux32[0] >> 28; - const uint16_t ls2 = aux32[1] >> 28; - - const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); - const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); - sumi1 = __lasx_xvadd_w(sumi1, p1); - sumi2 = __lasx_xvadd_w(sumi2, p2); - } - - accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); - } - - *s = 0.25f * hsum_float_8(accumf); - -#else - - uint32_t aux32; - - float sumf = 0.f; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict gas = x[i].qs + QK_K/4; - const int8_t * restrict q8 = y[i].qs; - int32_t bsum = 0; - for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { - memcpy(&aux32, gas, sizeof(uint32_t)); gas += sizeof(uint32_t); - const uint32_t ls = 2*(aux32 >> 28) + 1; - int32_t sumi = 0; - for (int l = 0; l < 4; ++l) { - const uint8_t * grid1 = (const uint8_t *)(iq3xxs_grid + q3[2*l+0]); - const uint8_t * grid2 = (const uint8_t *)(iq3xxs_grid + q3[2*l+1]); - const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*l) & 127]; - for (int j = 0; j < 4; ++j) { - sumi += grid1[j] * q8[j+0] * (signs & kmask_iq2xs[j+0] ? -1 : 1); - sumi += grid2[j] * q8[j+4] * (signs & kmask_iq2xs[j+4] ? -1 : 1); - } - q8 += 8; - } - q3 += 8; - bsum += sumi * ls; - } - sumf += d * bsum; - } - *s = 0.25f * sumf; -#endif -} - -void ggml_vec_dot_iq3_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq3_s * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined(__ARM_NEON) - - typedef union { - uint16x8_t vec_index; - uint16_t index[8]; - } vec_index_t; - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - - static const int16_t k_shift[8] = {8, 7, 6, 5, 4, 3, 2, 1}; - - const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); - const uint8x16_t mask2 = vld1q_u8(k_mask2); - - const int16x8_t hshift = vld1q_s16(k_shift); - const uint16x8_t m256 = vdupq_n_u16(256); - const uint8x16_t m1 = vdupq_n_u8(1); - - uint8x16x2_t vs; - ggml_int8x16x4_t q3s; - ggml_int8x16x4_t q8b; - vec_index_t idx; - - uint32_t scales32[2]; - const uint8_t * scales8 = (const uint8_t *)scales32; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)x[i].signs; - const int8_t * restrict q8 = y[i].qs; - - memcpy(scales32, x[i].scales, 4); - scales32[1] = (((scales32[0] >> 4) & 0x0f0f0f0f) << 1) | 0x01010101; - scales32[0] = ((scales32[0] & 0x0f0f0f0f) << 1) | 0x01010101; - - int sumi1 = 0, sumi2 = 0; - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - - const uint8x16_t idx_l = vld1q_u8(qs); qs += 16; - idx.vec_index = vorrq_u16(vmovl_u8(vget_low_u8 (idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+0]), hshift), m256)); - const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], - iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); - const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], - iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); - idx.vec_index = vorrq_u16(vmovl_u8(vget_high_u8(idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+1]), hshift), m256)); - const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], - iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); - const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], - iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); - - - vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); - vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); - vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); - vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); - vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); - - q3s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_0)); - q3s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_1)); - - vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); - vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); - vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); - vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); - vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); - - signs += 4; - - q3s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_2)); - q3s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_3)); - - const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); - const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); - - sumi1 += vaddvq_s32(p1) * scales8[ib32/2+0]; - sumi2 += vaddvq_s32(p2) * scales8[ib32/2+4]; - } - sumf += d*(sumi1 + sumi2); - } - *s = sumf; - -#elif defined(__AVX2__) - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); - const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); - - const __m256i idx_shift = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); - const __m256i idx_mask = _mm256_set1_epi32(256); - - typedef union { - __m256i vec[2]; - uint32_t index[16]; - } index_t; - - index_t idx; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)x[i].signs; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i idx_l = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i *)qs)); qs += 16; - idx.vec[0] = _mm256_set1_epi32(qh[ib32+0]); - idx.vec[1] = _mm256_set1_epi32(qh[ib32+1]); - idx.vec[0] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[0], idx_shift), idx_mask); - idx.vec[1] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[1], idx_shift), idx_mask); - idx.vec[0] = _mm256_or_si256(idx.vec[0], _mm256_cvtepi16_epi32(_mm256_castsi256_si128(idx_l))); - idx.vec[1] = _mm256_or_si256(idx.vec[1], _mm256_cvtepi16_epi32(_mm256_extractf128_si256(idx_l, 1))); - - // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange. - //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4); - //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4); - const __m256i q2_1 = _mm256_set_epi32( - iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]], - iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]] - ); - const __m256i q2_2 = _mm256_set_epi32( - iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]], - iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]] - ); - - __m256i aux256 = _mm256_set1_epi32(signs[0] | (signs[1] << 16)); - aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); - const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); - const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); - - aux256 = _mm256_set1_epi32(signs[2] | (signs[3] << 16)); - aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); - const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); - const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); - - signs += 4; - - const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); - const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); - const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; - const uint16_t ls2 = x[i].scales[ib32/2] >> 4; - const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); - const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); - sumi1 = _mm256_add_epi32(sumi1, p1); - sumi2 = _mm256_add_epi32(sumi2, p2); - } - - accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - - } - - *s = hsum_float_8(accumf); - -#elif defined(__AVX__) - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m128i mask1_0 = _mm_loadu_si128((const __m128i*)k_mask1); - const __m128i mask1_1 = _mm_loadu_si128((const __m128i*)k_mask1 + 1); - const __m128i mask2_0 = _mm_loadu_si128((const __m128i*)k_mask2); - const __m128i mask2_1 = _mm_loadu_si128((const __m128i*)k_mask2 + 1); - - const __m128i idx_mul_0 = _mm_set_epi32(32, 64, 128, 256); - const __m128i idx_mul_1 = _mm_set_epi32(2, 4, 8, 16); - const __m128i idx_mask = _mm_set1_epi32(256); - - typedef union { - __m128i vec[4]; - uint32_t index[16]; - } index_t; - - index_t idx; - - __m256 accumf = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)x[i].signs; - const int8_t * restrict q8 = y[i].qs; - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m128i q8_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i qs_tmp = _mm_loadu_si128((const __m128i *)qs); - const __m128i idx_l_0 = _mm_cvtepu8_epi16(qs_tmp); - const __m128i idx_l_1 = _mm_cvtepu8_epi16(_mm_srli_si128(qs_tmp, 8)); qs += 16; - idx.vec[0] = _mm_set1_epi32(qh[ib32+0]); - idx.vec[1] = idx.vec[0]; - idx.vec[2] = _mm_set1_epi32(qh[ib32+1]); - idx.vec[3] = idx.vec[2]; - - idx.vec[0] = _mm_and_si128(_mm_mullo_epi32(idx.vec[0], idx_mul_0), idx_mask); - idx.vec[1] = _mm_and_si128(_mm_mullo_epi32(idx.vec[1], idx_mul_1), idx_mask); - idx.vec[2] = _mm_and_si128(_mm_mullo_epi32(idx.vec[2], idx_mul_0), idx_mask); - idx.vec[3] = _mm_and_si128(_mm_mullo_epi32(idx.vec[3], idx_mul_1), idx_mask); - - idx.vec[0] = _mm_or_si128(idx.vec[0], _mm_cvtepi16_epi32(idx_l_0)); - idx.vec[1] = _mm_or_si128(idx.vec[1], _mm_cvtepi16_epi32(_mm_srli_si128(idx_l_0, 8))); - idx.vec[2] = _mm_or_si128(idx.vec[2], _mm_cvtepi16_epi32(idx_l_1)); - idx.vec[3] = _mm_or_si128(idx.vec[3], _mm_cvtepi16_epi32(_mm_srli_si128(idx_l_1, 8))); - - const __m128i q2_1_0 = _mm_set_epi32(iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]]); - const __m128i q2_1_1 = _mm_set_epi32(iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]]); - const __m128i q2_2_0 = _mm_set_epi32(iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[9]], iq3s_grid[idx.index[8]]); - const __m128i q2_2_1 = _mm_set_epi32(iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]]); - - __m128i aux128_0 = _mm_set1_epi32(signs[0] | (signs[1] << 16)); - __m128i aux128_1 = aux128_0; - aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); - aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); - const __m128i s2_1_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); - const __m128i s2_1_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); - const __m128i q8s_1_0 = _mm_sub_epi8(_mm_xor_si128(s2_1_0, q8_1_0), s2_1_0); - const __m128i q8s_1_1 = _mm_sub_epi8(_mm_xor_si128(s2_1_1, q8_1_1), s2_1_1); - - aux128_0 = _mm_set1_epi32(signs[2] | (signs[3] << 16)); - aux128_1 = aux128_0; - aux128_0 = _mm_and_si128(_mm_shuffle_epi8(aux128_0,mask1_0), mask2_0); - aux128_1 = _mm_and_si128(_mm_shuffle_epi8(aux128_1,mask1_1), mask2_1); - const __m128i s2_2_0 = _mm_cmpeq_epi8(aux128_0, mask2_0); - const __m128i s2_2_1 = _mm_cmpeq_epi8(aux128_1, mask2_1); - const __m128i q8s_2_0 = _mm_sub_epi8(_mm_xor_si128(s2_2_0, q8_2_0), s2_2_0); - const __m128i q8s_2_1 = _mm_sub_epi8(_mm_xor_si128(s2_2_1, q8_2_1), s2_2_1); - - signs += 4; - - const __m128i dot1_0 = _mm_maddubs_epi16(q2_1_0, q8s_1_0); - const __m128i dot1_1 = _mm_maddubs_epi16(q2_1_1, q8s_1_1); - const __m128i dot2_0 = _mm_maddubs_epi16(q2_2_0, q8s_2_0); - const __m128i dot2_1 = _mm_maddubs_epi16(q2_2_1, q8s_2_1); - const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; - const uint16_t ls2 = x[i].scales[ib32/2] >> 4; - const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(2*ls1+1)); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(2*ls1+1)); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(2*ls2+1)); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(2*ls2+1)); - sumi1_0 = _mm_add_epi32(sumi1_0, p1_0); - sumi1_1 = _mm_add_epi32(sumi1_1, p1_1); - sumi2_0 = _mm_add_epi32(sumi2_0, p2_0); - sumi2_1 = _mm_add_epi32(sumi2_1, p2_1); - } - - accumf = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_add_epi32(sumi1_1, sumi2_1), _mm_add_epi32(sumi1_0, sumi2_0)))), accumf); - - } - - *s = hsum_float_8(accumf); - -#elif defined(__POWER9_VECTOR__) - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - - const vector int v0 = vec_splats((int32_t)0); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - const vector unsigned char mask0 = vec_xl( 0, k_mask1); - const vector unsigned char mask1 = vec_xl(16, k_mask1); - const vector signed char mask2 = (vector signed char)vec_xl( 0, k_mask2); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)(x[i].signs); - const uint8_t * restrict sc = x[i].scales; - const int8_t * restrict q8 = y[i].qs; - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - for (int j = 0; j < QK_K/32; j += 2) { - __builtin_prefetch(q3, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector unsigned int aux32x4_0 = {iq3s_grid[q3[ 0] | ((qh[0] << 8) & 256)], iq3s_grid[q3[ 1] | ((qh[0] << 7) & 256)], - iq3s_grid[q3[ 2] | ((qh[0] << 6) & 256)], iq3s_grid[q3[ 3] | ((qh[0] << 5) & 256)]}; - vector unsigned int aux32x4_1 = {iq3s_grid[q3[ 4] | ((qh[0] << 4) & 256)], iq3s_grid[q3[ 5] | ((qh[0] << 3) & 256)], - iq3s_grid[q3[ 6] | ((qh[0] << 2) & 256)], iq3s_grid[q3[ 7] | ((qh[0] << 1) & 256)]}; - vector unsigned int aux32x4_2 = {iq3s_grid[q3[ 8] | ((qh[1] << 8) & 256)], iq3s_grid[q3[ 9] | ((qh[1] << 7) & 256)], - iq3s_grid[q3[10] | ((qh[1] << 6) & 256)], iq3s_grid[q3[11] | ((qh[1] << 5) & 256)]}; - vector unsigned int aux32x4_3 = {iq3s_grid[q3[12] | ((qh[1] << 4) & 256)], iq3s_grid[q3[13] | ((qh[1] << 3) & 256)], - iq3s_grid[q3[14] | ((qh[1] << 2) & 256)], iq3s_grid[q3[15] | ((qh[1] << 1) & 256)]}; - q3 += 16; - qh += 2; - - vector signed char vsigns01 = (vector signed char)vec_splats(*(const uint32_t *)&signs[0]); - vector signed char vsigns02 = (vector signed char)vec_splats(*(const uint32_t *)&signs[2]); - signs += 4; - - vector signed char vsigns0 = vec_perm(vsigns01, vsigns01, mask0); - vector signed char vsigns1 = vec_perm(vsigns01, vsigns01, mask1); - vector signed char vsigns2 = vec_perm(vsigns02, vsigns02, mask0); - vector signed char vsigns3 = vec_perm(vsigns02, vsigns02, mask1); - - vsigns0 = (vector signed char)vec_cmpeq(vec_and(vsigns0, mask2), mask2); - vsigns1 = (vector signed char)vec_cmpeq(vec_and(vsigns1, mask2), mask2); - vsigns2 = (vector signed char)vec_cmpeq(vec_and(vsigns2, mask2), mask2); - vsigns3 = (vector signed char)vec_cmpeq(vec_and(vsigns3, mask2), mask2); - - vector signed char q3x0 = vec_sub(vec_xor(vsigns0, (vector signed char)aux32x4_0), vsigns0); - vector signed char q3x1 = vec_sub(vec_xor(vsigns1, (vector signed char)aux32x4_1), vsigns1); - vector signed char q3x2 = vec_sub(vec_xor(vsigns2, (vector signed char)aux32x4_2), vsigns2); - vector signed char q3x3 = vec_sub(vec_xor(vsigns3, (vector signed char)aux32x4_3), vsigns3); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q3x0, q8y0), vec_mulo(q3x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q3x1, q8y1), vec_mulo(q3x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q3x2, q8y2), vec_mulo(q3x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q3x3, q8y3), vec_mulo(q3x3, q8y3)); - - const uint16_t ls0 = (uint16_t)(sc[0] & 0xf); - const uint16_t ls1 = (uint16_t)(sc[0] >> 4); - sc ++; - - vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); - vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); - - vsumi0 = vec_msum(qv0, vscales01, vsumi0); - vsumi1 = vec_msum(qv1, vscales01, vsumi1); - vsumi2 = vec_msum(qv2, vscales23, vsumi2); - vsumi3 = vec_msum(qv3, vscales23, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, - 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 - }; - - static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, - }; - - const __m256i mask1 = __lasx_xvld((const __m256i*)k_mask1, 0); - const __m256i mask2 = __lasx_xvld((const __m256i*)k_mask2, 0); - - __m256i idx_shift = lasx_set_w(1, 2, 3, 4, 5, 6, 7, 8); - const __m256i idx_mask = __lasx_xvreplgr2vr_w(256); - - typedef union { - __m256i vec[2]; - uint32_t index[16]; - } index_t; - - index_t idx; - - __m256 accumf = (__m256)__lasx_xvldi(0); - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint16_t * restrict signs = (const uint16_t *)x[i].signs; - const int8_t * restrict q8 = y[i].qs; - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const __m256i q8_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i idx_l = lasx_extu8_16(__lsx_vld(qs, 0)); qs += 16; - idx.vec[0] = __lasx_xvreplgr2vr_w(qh[ib32+0]); - idx.vec[1] = __lasx_xvreplgr2vr_w(qh[ib32+1]); - idx.vec[0] = __lasx_xvand_v(__lasx_xvsll_w(idx.vec[0], idx_shift), idx_mask); - idx.vec[1] = __lasx_xvand_v(__lasx_xvsll_w(idx.vec[1], idx_shift), idx_mask); - idx.vec[0] = __lasx_xvor_v(idx.vec[0], lasx_ext16_32(lasx_extracti128(idx_l, 0))); - idx.vec[1] = __lasx_xvor_v(idx.vec[1], lasx_ext16_32(lasx_extracti128(idx_l, 1))); - - // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange. - //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4); - //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4); - const __m256i q2_1 = lasx_set_w( - iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]], - iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]] - ); - const __m256i q2_2 = lasx_set_w( - iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]], - iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]] - ); - - __m256i aux256 = __lasx_xvreplgr2vr_w(signs[0] | (signs[1] << 16)); - aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); - const __m256i s2_1 = __lasx_xvseq_b(aux256, mask2); - const __m256i q8s_1 = __lasx_xvsub_b(__lasx_xvxor_v(s2_1, q8_1), s2_1); - - aux256 = __lasx_xvreplgr2vr_w(signs[2] | (signs[3] << 16)); - aux256 = __lasx_xvand_v(lasx_shuffle_b(aux256,mask1), mask2); - const __m256i s2_2 = __lasx_xvseq_b(aux256, mask2); - const __m256i q8s_2 = __lasx_xvsub_b(__lasx_xvxor_v(s2_2, q8_2), s2_2); - - signs += 4; - - const __m256i dot1 = lasx_maddubs_h(q2_1, q8s_1); - const __m256i dot2 = lasx_maddubs_h(q2_2, q8s_2); - const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; - const uint16_t ls2 = x[i].scales[ib32/2] >> 4; - const __m256i p1 = lasx_madd_h(dot1, __lasx_xvreplgr2vr_h(2*ls1+1)); - const __m256i p2 = lasx_madd_h(dot2, __lasx_xvreplgr2vr_h(2*ls2+1)); - sumi1 = __lasx_xvadd_w(sumi1, p1); - sumi2 = __lasx_xvadd_w(sumi2, p2); - } - - accumf = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accumf); - } - - *s = hsum_float_8(accumf); - -#else - - float sumf = 0.f; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const uint8_t * restrict qs = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const uint8_t * restrict signs = x[i].signs; - const int8_t * restrict q8 = y[i].qs; - int32_t bsum = 0; - for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { - const uint32_t ls1 = 2*(x[i].scales[ib32/2] & 0xf) + 1; - const uint32_t ls2 = 2*(x[i].scales[ib32/2] >> 4) + 1; - int32_t sumi = 0; - for (int l = 0; l < 4; ++l) { - const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+0] << (8-2*l)) & 256))); - const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+0] << (7-2*l)) & 256))); - for (int j = 0; j < 4; ++j) { - sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); - sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); - } - q8 += 8; - } - qs += 8; - signs += 4; - bsum += sumi * ls1; - sumi = 0; - for (int l = 0; l < 4; ++l) { - const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+1] << (8-2*l)) & 256))); - const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+1] << (7-2*l)) & 256))); - for (int j = 0; j < 4; ++j) { - sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); - sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); - } - q8 += 8; - } - qs += 8; - signs += 4; - bsum += sumi * ls2; - } - sumf += d * bsum; - } - *s = sumf; -#endif -} - -#if defined(__AVX2__) -static inline __m256i mul_add_epi8(const __m256i x, const __m256i y) { - const __m256i ax = _mm256_sign_epi8(x, x); - const __m256i sy = _mm256_sign_epi8(y, x); - return _mm256_maddubs_epi16(ax, sy); -} -#elif defined(__loongarch_asx) -static inline __m256i mul_add_epi8(const __m256i x, const __m256i y) { - const __m256i ax = __lasx_xvsigncov_b(x, x); - const __m256i sy = __lasx_xvsigncov_b(x, y); - __m256i tmp1, tmp2, tmp3; - tmp1 = __lasx_xvmulwev_h_bu_b(ax, sy); - tmp2 = __lasx_xvmulwod_h_bu_b(ax, sy); - tmp3 = __lasx_xvadd_h(tmp1, tmp2); - return __lasx_xvsat_h(tmp3, 15); -} -#endif - -void ggml_vec_dot_iq1_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq1_s * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined __ARM_NEON - - ggml_int8x16x4_t q1b; - ggml_int8x16x4_t q8b; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint16_t * qh = x[i].qh; - - int sumi1 = 0, sumi2 = 0, sumi3 = 0; - - for (int ib = 0; ib < QK_K/32; ib += 2) { - - q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[ib+0] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[ib+0] << 5) & 0x700))))); - q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[ib+0] << 2) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[ib+0] >> 1) & 0x700))))); - q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[ib+1] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[ib+1] << 5) & 0x700))))); - q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[ib+1] << 2) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[ib+1] >> 1) & 0x700))))); - qs += 8; - - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - - const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[0], q8b.val[0]), q1b.val[1], q8b.val[1]); - const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[2], q8b.val[2]), q1b.val[3], q8b.val[3]); - - const int ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; - const int ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; - sumi1 += vaddvq_s32(p1) * ls1; - sumi2 += vaddvq_s32(p2) * ls2; - sumi3 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * ls1 * (qh[ib+0] & 0x8000 ? -1 : 1) - + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * ls2 * (qh[ib+1] & 0x8000 ? -1 : 1); - - } - - sumf += y[i].d * GGML_FP16_TO_FP32(x[i].d) * (sumi1 + sumi2 + IQ1S_DELTA * sumi3); - } - - *s = sumf; - -#elif defined __AVX2__ - - __m256 accum = _mm256_setzero_ps(); - float accum1 = 0; - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint16_t * qh = x[i].qh; - - __m256i sumi = _mm256_setzero_si256(); - int sumi1 = 0; - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m256i q1b_1 = _mm256_set_epi64x(iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)], - iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)]); - const __m256i q1b_2 = _mm256_set_epi64x(iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)], - iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)]); - qs += 8; - const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); - const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); - const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; - const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; - const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(ls1)); - const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(ls2)); - - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p1, p2)); - sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 - + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; - } - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - accum = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sumi), accum); - accum1 += d * sumi1; - - } - - *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; - -#elif defined __AVX__ - __m256 accum = _mm256_setzero_ps(); - float accum1 = 0; - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint16_t * qh = x[i].qh; - - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - int sumi1 = 0; - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m128i q1b_1_0 = _mm_set_epi64x(iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)]); - const __m128i q1b_1_1 = _mm_set_epi64x(iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)]); - const __m128i q1b_2_0 = _mm_set_epi64x(iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)]); - const __m128i q1b_2_1 = _mm_set_epi64x(iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)]); - qs += 8; - const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - - const __m128i dot1_0 = mul_add_epi8_sse(q1b_1_0, q8b_1_0); - const __m128i dot1_1 = mul_add_epi8_sse(q1b_1_1, q8b_1_1); - const __m128i dot2_0 = mul_add_epi8_sse(q1b_2_0, q8b_2_0); - const __m128i dot2_1 = mul_add_epi8_sse(q1b_2_1, q8b_2_1); - const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; - const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; - const __m128i p1_0 = _mm_madd_epi16(dot1_0, _mm_set1_epi16(ls1)); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, _mm_set1_epi16(ls1)); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, _mm_set1_epi16(ls2)); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, _mm_set1_epi16(ls2)); - - sumi1_0 = _mm_add_epi32(sumi1_0, _mm_add_epi32(p1_0, p2_0)); - sumi1_1 = _mm_add_epi32(sumi1_1, _mm_add_epi32(p1_1, p2_1)); - sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 - + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; - } - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - accum = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(sumi1_1, sumi1_0))), accum); - accum1 += d * sumi1; - - } - - *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; - -#elif defined(__POWER9_VECTOR__) - const vector unsigned char v0 = vec_splats((unsigned char)0x0); - const vector unsigned short vsign = vec_splats((unsigned short)0x8000); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - for (int i = 0; i < nb; ++i) { - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[i].d)); - vector float vyd = vec_splats(y[i].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = vec_splats((int32_t)0); - vector signed int vsumi1 = vec_splats((int32_t)0); - vector signed int vsumi2 = vec_splats((int32_t)0); - vector signed int vsumi3 = vec_splats((int32_t)0); - vector signed int vsumi8 = vec_splats((int32_t)0); - - const uint8_t * restrict q1 = x[i].qs; - const uint16_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - const int16_t * restrict qs = y[i].bsums; - - for (int j = 0; j < QK_K/32; j += 2) { - __builtin_prefetch(q1, 0, 1); - __builtin_prefetch(qh, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed long long aux64x2_0 = {*(const int64_t *)(iq1s_grid + (q1[0] | ((qh[0] << 8) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[1] | ((qh[0] << 5) & 0x700)))}; - vector signed long long aux64x2_1 = {*(const int64_t *)(iq1s_grid + (q1[2] | ((qh[0] << 2) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[3] | ((qh[0] >> 1) & 0x700)))}; - vector signed long long aux64x2_2 = {*(const int64_t *)(iq1s_grid + (q1[4] | ((qh[1] << 8) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[5] | ((qh[1] << 5) & 0x700)))}; - vector signed long long aux64x2_3 = {*(const int64_t *)(iq1s_grid + (q1[6] | ((qh[1] << 2) & 0x700))), *(const int64_t *)(iq1s_grid + (q1[7] | ((qh[1] >> 1) & 0x700)))}; - q1 += 8; - - vector signed char q1x0 = (vector signed char)aux64x2_0; - vector signed char q1x1 = (vector signed char)aux64x2_1; - vector signed char q1x2 = (vector signed char)aux64x2_2; - vector signed char q1x3 = (vector signed char)aux64x2_3; - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q1x0, q8y0), vec_mulo(q1x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q1x1, q8y1), vec_mulo(q1x1, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q1x2, q8y2), vec_mulo(q1x2, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q1x3, q8y3), vec_mulo(q1x3, q8y3)); - - const uint16_t ls0 = (uint16_t)((qh[0] >> 12) & 7); - const uint16_t ls1 = (uint16_t)((qh[1] >> 12) & 7); - - vector signed short vscales01 = (vector signed short)vec_splats((uint16_t)(2*ls0+1)); - vector signed short vscales23 = (vector signed short)vec_splats((uint16_t)(2*ls1+1)); - vector signed short vscales = vec_sld(vscales23, vscales01, 8); - - vsumi0 = vec_msum(qv0, vscales01, vsumi0); - vsumi1 = vec_msum(qv1, vscales01, vsumi1); - vsumi2 = vec_msum(qv2, vscales23, vsumi2); - vsumi3 = vec_msum(qv3, vscales23, vsumi3); - - vector signed short q8ysums = vec_xl_len(qs, 8); - qs += 4; - q8ysums = vec_mergeh(q8ysums, (vector signed short)v0); - - vector signed short qxh = (vector signed short)vec_sld(vec_splats(qh[1]), vec_splats(qh[0]), 8); - qh += 2; - vector __bool short vsel = vec_cmpge(qxh, (vector signed short)v0); - - vector signed short q8ysum = vec_sel((vector signed short)vec_xor((vector unsigned short)q8ysums, vsign), q8ysums, vsel); - - vsumi8 = vec_add(vec_mule(q8ysum, vscales), vsumi8); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - - vsumf0 = vec_madd(vec_ctf(vsumi8, 0), vec_mul(vd, vec_splats(IQ1S_DELTA)), vsumf0); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - __m256 accum = (__m256)__lasx_xvldi(0); - float accum1 = 0; - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint16_t * qh = x[i].qh; - - __m256i sumi = __lasx_xvldi(0); - int sumi1 = 0; - for (int ib = 0; ib < QK_K/32; ib += 2) { - __m256i q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)], 0); - q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], 1); - q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)], 2); - q1b_1 = __lasx_xvinsgr2vr_d(q1b_1, iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], 3); - - __m256i q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)], 0); - q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], 1); - q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)], 2); - q1b_2 = __lasx_xvinsgr2vr_d(q1b_2, iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], 3); - - qs += 8; - const __m256i q8b_1 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - const __m256i q8b_2 = __lasx_xvld((const __m256i*)q8, 0); q8 += 32; - - const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); - const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); - const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; - const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; - - __m256i tmp1, tmp5, tmp6; - tmp1 = __lasx_xvreplgr2vr_h(ls1); - tmp5 = __lasx_xvmulwev_w_h(dot1, tmp1); - tmp6 = __lasx_xvmulwod_w_h(dot1, tmp1); - const __m256i p1 = __lasx_xvadd_w(tmp5, tmp6); - - tmp1 = __lasx_xvreplgr2vr_h(ls2); - tmp5 = __lasx_xvmulwev_w_h(dot2, tmp1); - tmp6 = __lasx_xvmulwod_w_h(dot2, tmp1); - const __m256i p2 = __lasx_xvadd_w(tmp5, tmp6); - - sumi = __lasx_xvadd_w(sumi, __lasx_xvadd_w(p1, p2)); - sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 - + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; - } - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - accum = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(d), __lasx_xvffint_s_w(sumi), accum); - accum1 += d * sumi1; - } - - *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; - -#else - - float sumf = 0; - for (int i = 0; i < nb; i++) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint16_t * qh = x[i].qh; - - int sumi = 0, sumi1 = 0; - for (int ib = 0; ib < QK_K/32; ++ib) { - const int ls = 2*((qh[ib] >> 12) & 7) + 1; - const int delta = qh[ib] & 0x8000 ? -1 : 1; - int lsum = 0; - for (int l = 0; l < 4; ++l) { - const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((qh[ib] >> 3*l) & 7) << 8))); - for (int j = 0; j < 8; ++j) { - lsum += q8[j] * grid[j]; - } - q8 += 8; - } - sumi += ls * lsum; - sumi1 += ls * delta * (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]); - qs += 4; - } - - sumf += GGML_FP16_TO_FP32(x[i].d) * y[i].d * (sumi + IQ1S_DELTA * sumi1); - } - - *s = sumf; - -#endif -} - -void ggml_vec_dot_iq1_m_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(n % QK_K == 0); - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - - const block_iq1_m * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - - iq1m_scale_t scale; - -#if defined __ARM_NEON - const int32x4_t mask = vdupq_n_s32(0x7); - const int32x4_t mone = vdupq_n_s32(1); - const int32x4_t mzero = vdupq_n_s32(0); - - ggml_int8x16x4_t deltas; - deltas.val[0] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(+1)); - deltas.val[1] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(+1)); - deltas.val[2] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(-1)); - deltas.val[3] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(-1)); - - ggml_int8x16x4_t q1b; - ggml_int8x16x4_t q8b; - - uint32_t aux32; - const uint8_t * aux8 = (const uint8_t *)&aux32; - - float sumf = 0; - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint8_t * qh = x[i].qh; - const uint16_t * sc = (const uint16_t *)x[i].scales; - - scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); - - int32x4_t sumi1 = mzero; - int32x4_t sumi2 = mzero; - - for (int ib = 0; ib < QK_K/32; ib += 2) { - - q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[0] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[0] << 4) & 0x700))))); - q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[1] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[1] << 4) & 0x700))))); - q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[2] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[2] << 4) & 0x700))))); - q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[3] << 8) & 0x700)))), - vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[3] << 4) & 0x700))))); - - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - - const int32x4_t p1 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[0], q8b.val[0]), ggml_vdotq_s32(mzero, q1b.val[1], q8b.val[1])); - const int32x4_t p2 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[2], q8b.val[2]), ggml_vdotq_s32(mzero, q1b.val[3], q8b.val[3])); - const int32x4_t p12 = vpaddq_s32(p1, p2); - - const uint32_t * qh32 = (const uint32_t *)qh; // we are 4-byte aligned, so we can do that - aux32 = ((qh32[0] >> 3) & 0x01010101) | ((qh32[0] >> 6) & 0x02020202); - - const int32x4_t p3 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[0]], q8b.val[0]), ggml_vdotq_s32(mzero, deltas.val[aux8[1]], q8b.val[1])); - const int32x4_t p4 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[2]], q8b.val[2]), ggml_vdotq_s32(mzero, deltas.val[aux8[3]], q8b.val[3])); - const int32x4_t p34 = vpaddq_s32(p3, p4); - - int32x4_t scales_4 = ggml_vld1q_u32(sc[ib/2] >> 0, sc[ib/2] >> 3, sc[ib/2] >> 6, sc[ib/2] >> 9); - - scales_4 = vaddq_s32(vshlq_n_s32(vandq_s32(scales_4, mask), 1), mone); - - sumi1 = vmlaq_s32(sumi1, scales_4, p12); - sumi2 = vmlaq_s32(sumi2, scales_4, p34); - - qs += 8; qh += 4; - - } - - sumf += y[i].d * GGML_FP16_TO_FP32(scale.f16) * (vaddvq_s32(sumi1) + IQ1M_DELTA * vaddvq_s32(sumi2)); - } - - *s = sumf; - -#elif defined __AVX2__ - - const __m256i mask = _mm256_set1_epi16(0x7); - const __m256i mone = _mm256_set1_epi16(1); - - __m256 accum1 = _mm256_setzero_ps(); - __m256 accum2 = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint8_t * qh = x[i].qh; - const uint16_t * sc = (const uint16_t *)x[i].scales; - - scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); - - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m256i q1b_1 = _mm256_set_epi64x( - iq1s_grid[qs[3] | (((uint16_t)qh[1] << 4) & 0x700)], iq1s_grid[qs[2] | (((uint16_t)qh[1] << 8) & 0x700)], - iq1s_grid[qs[1] | (((uint16_t)qh[0] << 4) & 0x700)], iq1s_grid[qs[0] | (((uint16_t)qh[0] << 8) & 0x700)] - ); - const __m256i q1b_2 = _mm256_set_epi64x( - iq1s_grid[qs[7] | (((uint16_t)qh[3] << 4) & 0x700)], iq1s_grid[qs[6] | (((uint16_t)qh[3] << 8) & 0x700)], - iq1s_grid[qs[5] | (((uint16_t)qh[2] << 4) & 0x700)], iq1s_grid[qs[4] | (((uint16_t)qh[2] << 8) & 0x700)] - ); - const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - - const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); - const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); - - const __m256i delta1 = _mm256_set_epi64x(qh[1] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[1] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, - qh[0] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[0] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - const __m256i delta2 = _mm256_set_epi64x(qh[3] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[3] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, - qh[2] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[2] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - - const __m256i dot3 = mul_add_epi8(delta1, q8b_1); - const __m256i dot4 = mul_add_epi8(delta2, q8b_2); - - __m256i scale1 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 3), _mm_set1_epi16(sc[ib/2] >> 0)); - __m256i scale2 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 9), _mm_set1_epi16(sc[ib/2] >> 6)); - - scale1 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale1, mask), 1), mone); - scale2 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale2, mask), 1), mone); - const __m256i p1 = _mm256_madd_epi16(dot1, scale1); - const __m256i p2 = _mm256_madd_epi16(dot2, scale2); - const __m256i p3 = _mm256_madd_epi16(dot3, scale1); - const __m256i p4 = _mm256_madd_epi16(dot4, scale2); - - sumi1 = _mm256_add_epi32(sumi1, _mm256_add_epi32(p1, p2)); - sumi2 = _mm256_add_epi32(sumi2, _mm256_add_epi32(p3, p4)); - - qs += 8; qh += 4; - } - - const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(scale.f16)); - - accum1 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi1), accum1); - accum2 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi2), accum2); - } - - *s = hsum_float_8(accum1) + IQ1M_DELTA * hsum_float_8(accum2); - -#elif defined __AVX__ - const __m128i mask = _mm_set1_epi16(0x7); - const __m128i mone = _mm_set1_epi16(1); - - __m256 accum1 = _mm256_setzero_ps(); - __m256 accum2 = _mm256_setzero_ps(); - for (int i = 0; i < nb; ++i) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint8_t * qh = x[i].qh; - const uint16_t * sc = (const uint16_t *)x[i].scales; - - scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); - - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m128i q1b_1_0 = _mm_set_epi64x( - iq1s_grid[qs[1] | (((uint16_t)qh[0] << 4) & 0x700)], iq1s_grid[qs[0] | (((uint16_t)qh[0] << 8) & 0x700)]); - const __m128i q1b_1_1 = _mm_set_epi64x( - iq1s_grid[qs[3] | (((uint16_t)qh[1] << 4) & 0x700)], iq1s_grid[qs[2] | (((uint16_t)qh[1] << 8) & 0x700)]); - const __m128i q1b_2_0 = _mm_set_epi64x( - iq1s_grid[qs[5] | (((uint16_t)qh[2] << 4) & 0x700)], iq1s_grid[qs[4] | (((uint16_t)qh[2] << 8) & 0x700)]); - const __m128i q1b_2_1 = _mm_set_epi64x( - iq1s_grid[qs[7] | (((uint16_t)qh[3] << 4) & 0x700)], iq1s_grid[qs[6] | (((uint16_t)qh[3] << 8) & 0x700)]); - const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - - const __m128i dot1_0 = mul_add_epi8_sse(q1b_1_0, q8b_1_0); - const __m128i dot1_1 = mul_add_epi8_sse(q1b_1_1, q8b_1_1); - const __m128i dot2_0 = mul_add_epi8_sse(q1b_2_0, q8b_2_0); - const __m128i dot2_1 = mul_add_epi8_sse(q1b_2_1, q8b_2_1); - - const __m128i delta1_0 = _mm_set_epi64x(qh[0] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[0] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - const __m128i delta1_1 = _mm_set_epi64x(qh[1] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[1] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - const __m128i delta2_0 = _mm_set_epi64x(qh[2] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[2] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - const __m128i delta2_1 = _mm_set_epi64x(qh[3] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, - qh[3] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); - - const __m128i dot3_0 = mul_add_epi8_sse(delta1_0, q8b_1_0); - const __m128i dot3_1 = mul_add_epi8_sse(delta1_1, q8b_1_1); - const __m128i dot4_0 = mul_add_epi8_sse(delta2_0, q8b_2_0); - const __m128i dot4_1 = mul_add_epi8_sse(delta2_1, q8b_2_1); - - __m128i scale1_0 = _mm_set1_epi16(sc[ib/2] >> 0); - __m128i scale1_1 = _mm_set1_epi16(sc[ib/2] >> 3); - __m128i scale2_0 = _mm_set1_epi16(sc[ib/2] >> 6); - __m128i scale2_1 = _mm_set1_epi16(sc[ib/2] >> 9); - - scale1_0 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale1_0, mask), 1), mone); - scale1_1 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale1_1, mask), 1), mone); - scale2_0 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale2_0, mask), 1), mone); - scale2_1 = _mm_add_epi16(_mm_slli_epi16(_mm_and_si128(scale2_1, mask), 1), mone); - const __m128i p1_0 = _mm_madd_epi16(dot1_0, scale1_0); - const __m128i p1_1 = _mm_madd_epi16(dot1_1, scale1_1); - const __m128i p2_0 = _mm_madd_epi16(dot2_0, scale2_0); - const __m128i p2_1 = _mm_madd_epi16(dot2_1, scale2_1); - const __m128i p3_0 = _mm_madd_epi16(dot3_0, scale1_0); - const __m128i p3_1 = _mm_madd_epi16(dot3_1, scale1_1); - const __m128i p4_0 = _mm_madd_epi16(dot4_0, scale2_0); - const __m128i p4_1 = _mm_madd_epi16(dot4_1, scale2_1); - - sumi1_0 = _mm_add_epi32(sumi1_0, _mm_add_epi32(p1_0, p2_0)); - sumi1_1 = _mm_add_epi32(sumi1_1, _mm_add_epi32(p1_1, p2_1)); - sumi2_0 = _mm_add_epi32(sumi2_0, _mm_add_epi32(p3_0, p4_0)); - sumi2_1 = _mm_add_epi32(sumi2_1, _mm_add_epi32(p3_1, p4_1)); - - qs += 8; qh += 4; - } - - const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(scale.f16)); - - accum1 = _mm256_add_ps(_mm256_mul_ps(d, _mm256_cvtepi32_ps(MM256_SET_M128I(sumi1_1, sumi1_0))), accum1); - accum2 = _mm256_add_ps(_mm256_mul_ps(d, _mm256_cvtepi32_ps(MM256_SET_M128I(sumi2_1, sumi2_0))), accum2); - } - - *s = hsum_float_8(accum1) + IQ1M_DELTA * hsum_float_8(accum2); - -#else - - int sum1[2], sum2[2], delta[4]; - - float sumf = 0; - for (int i = 0; i < nb; i++) { - - const int8_t * q8 = y[i].qs; - const uint8_t * qs = x[i].qs; - const uint8_t * qh = x[i].qh; - const uint16_t * sc = (const uint16_t *)x[i].scales; - - scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); - - int sumi1 = 0, sumi2 = 0; - for (int ib = 0; ib < QK_K/32; ++ib) { - delta[0] = qh[0] & 0x08 ? -1 : 1; - delta[1] = qh[0] & 0x80 ? -1 : 1; - delta[2] = qh[1] & 0x08 ? -1 : 1; - delta[3] = qh[1] & 0x80 ? -1 : 1; - sum1[0] = sum1[1] = sum2[0] = sum2[1] = 0; - for (int l = 0; l < 4; ++l) { - const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((uint16_t)qh[l/2] << (8 - 4*(l%2))) & 0x700))); - int lsum1 = 0, lsum2 = 0; - for (int j = 0; j < 8; ++j) { - lsum1 += q8[j] * grid[j]; - lsum2 += q8[j]; - } - q8 += 8; - sum1[l/2] += lsum1; - sum2[l/2] += lsum2*delta[l]; - } - - const int ls1 = 2*((sc[ib/2] >> (6*(ib%2)+0)) & 0x7) + 1; - const int ls2 = 2*((sc[ib/2] >> (6*(ib%2)+3)) & 0x7) + 1; - - sumi1 += sum1[0] * ls1 + sum1[1] * ls2; - sumi2 += sum2[0] * ls1 + sum2[1] * ls2; - qs += 4; - qh += 2; - } - - sumf += GGML_FP16_TO_FP32(scale.f16) * y[i].d * (sumi1 + IQ1M_DELTA * sumi2); - } - - *s = sumf; - -#endif -} - -void ggml_vec_dot_iq4_nl_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - assert(n % QK4_NL == 0); - static_assert(QK4_NL == QK8_0, "QK4_NL and QK8_0 must be the same"); - - const block_iq4_nl * restrict x = vx; - const block_q8_0 * restrict y = vy; - - const int nb = n / QK4_NL; - - int ib = 0; - float sumf = 0; - -#if defined __ARM_NEON - const int8x16_t values = vld1q_s8(kvalues_iq4nl); - const uint8x16_t m4b = vdupq_n_u8(0x0f); - uint8x16x2_t q4bits; - int8x16x4_t q4b; - int8x16x4_t q8b; - int32x4_t prod_1, prod_2; - - for (; ib + 1 < nb; ib += 2) { - - q4bits.val[0] = vld1q_u8(x[ib + 0].qs); - q4bits.val[1] = vld1q_u8(x[ib + 1].qs); - q8b.val[0] = vld1q_s8(y[ib + 0].qs); - q8b.val[1] = vld1q_s8(y[ib + 0].qs + 16); - q8b.val[2] = vld1q_s8(y[ib + 1].qs); - q8b.val[3] = vld1q_s8(y[ib + 1].qs + 16); - - q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); - q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); - q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); - q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); - - prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); - prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); - - sumf += - GGML_FP16_TO_FP32(x[ib+0].d) * GGML_FP16_TO_FP32(y[ib + 0].d) * vaddvq_s32(prod_1) + - GGML_FP16_TO_FP32(x[ib+1].d) * GGML_FP16_TO_FP32(y[ib + 1].d) * vaddvq_s32(prod_2); - } - -#elif defined __AVX2__ - - const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); - const __m128i m4b = _mm_set1_epi8(0x0f); - const __m256i mone = _mm256_set1_epi16(1); - - __m256 accum1 = _mm256_setzero_ps(); - __m256 accum2 = _mm256_setzero_ps(); - for (; ib + 1 < nb; ib += 2) { - const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)x[ib + 0].qs); - const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)x[ib + 1].qs); - const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)y[ib + 0].qs); - const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)y[ib + 1].qs); - const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), - _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); - const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), - _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); - const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); - const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); - const __m256i p_1 = _mm256_madd_epi16(p16_1, mone); - const __m256i p_2 = _mm256_madd_epi16(p16_2, mone); - accum1 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), - _mm256_cvtepi32_ps(p_1), accum1); - accum2 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), - _mm256_cvtepi32_ps(p_2), accum2); - } - - sumf = hsum_float_8(_mm256_add_ps(accum1, accum2)); - -#elif defined __AVX__ - const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); - const __m128i m4b = _mm_set1_epi8(0x0f); - const __m128i mone = _mm_set1_epi16(1); - - __m256 accum1 = _mm256_setzero_ps(); - __m256 accum2 = _mm256_setzero_ps(); - for (; ib + 1 < nb; ib += 2) { - const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)x[ib + 0].qs); - const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)x[ib + 1].qs); - const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs); - const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)y[ib + 0].qs + 1); - const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs); - const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)y[ib + 1].qs + 1); - - const __m128i q4b_1_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)); - const __m128i q4b_1_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)); - const __m128i q4b_2_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)); - const __m128i q4b_2_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)); - const __m128i p16_1_0 = mul_add_epi8_sse(q4b_1_0, q8b_1_0); - const __m128i p16_1_1 = mul_add_epi8_sse(q4b_1_1, q8b_1_1); - const __m128i p16_2_0 = mul_add_epi8_sse(q4b_2_0, q8b_2_0); - const __m128i p16_2_1 = mul_add_epi8_sse(q4b_2_1, q8b_2_1); - const __m128i p_1_0 = _mm_madd_epi16(p16_1_0, mone); - const __m128i p_1_1 = _mm_madd_epi16(p16_1_1, mone); - const __m128i p_2_0 = _mm_madd_epi16(p16_2_0, mone); - const __m128i p_2_1 = _mm_madd_epi16(p16_2_1, mone); - accum1 = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), - _mm256_cvtepi32_ps(MM256_SET_M128I(p_1_1, p_1_0))), accum1); - accum2 = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), - _mm256_cvtepi32_ps(MM256_SET_M128I(p_2_1, p_2_0))), accum2); - } - - sumf = hsum_float_8(_mm256_add_ps(accum1, accum2)); - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector signed int v0 = vec_splats((int32_t)0); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - - const vector signed char values = vec_xl( 0, kvalues_iq4nl); - -#pragma GCC unroll 4 - for (; ib < nb; ++ib) { - __builtin_prefetch(x[ib].qs, 0, 1); - __builtin_prefetch(y[ib].qs, 0, 1); - - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ib].d)); - vector float vyd = vec_splats(GGML_FP16_TO_FP32(y[ib].d)); - vector float vd = vec_mul(vxd, vyd); - - vector signed char qxs = (vector signed char)vec_xl( 0, x[ib].qs); - vector signed char q4x0 = vec_and(qxs, lowMask); - vector signed char q4x1 = vec_sr(qxs, v4); - - q4x0 = vec_perm(values, values, (vector unsigned char)q4x0); - q4x1 = vec_perm(values, values, (vector unsigned char)q4x1); - - vector signed char q8y0 = vec_xl( 0, y[ib].qs); - vector signed char q8y1 = vec_xl(16, y[ib].qs); - - vector signed short qv0 = vec_add(vec_mule(q4x0, q8y0), vec_mulo(q4x0, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q4x1, q8y1), vec_mulo(q4x1, q8y1)); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - - vsumi0 = vec_sum4s(qv0, vsumi0); - vsumi1 = vec_sum4s(qv1, vsumi1); - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - } - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - sumf = vec_extract(vsumf0, 0); - -#elif defined (__loongarch_asx) - - const __m128i values128 = __lsx_vld((const __m128i*)kvalues_iq4nl, 0); - const __m128i m4b = __lsx_vreplgr2vr_b(0x0f); - const __m256i mone = __lasx_xvreplgr2vr_h(1); - - __m256 accum1 = (__m256)__lasx_xvldi(0); - __m256 accum2 = (__m256)__lasx_xvldi(0); - for (; ib + 1 < nb; ib += 2) { - const __m128i q4bits_1 = __lsx_vld((const __m128i*)x[ib + 0].qs, 0); - const __m128i q4bits_2 = __lsx_vld((const __m128i*)x[ib + 1].qs, 0); - const __m256i q8b_1 = __lasx_xvld((const __m256i *)y[ib + 0].qs, 0); - const __m256i q8b_2 = __lasx_xvld((const __m256i *)y[ib + 1].qs, 0); - const __m256i q4b_1 = lasx_insertf128(lsx_shuffle_b(values128, __lsx_vand_v(__lsx_vsrli_h(q4bits_1, 4), m4b)), - lsx_shuffle_b(values128, __lsx_vand_v(q4bits_1, m4b))); - const __m256i q4b_2 = lasx_insertf128(lsx_shuffle_b(values128, __lsx_vand_v(__lsx_vsrli_h(q4bits_2, 4), m4b)), - lsx_shuffle_b(values128, __lsx_vand_v(q4bits_2, m4b))); - const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); - const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); - const __m256i p_1 = lasx_madd_h(p16_1, mone); - const __m256i p_2 = lasx_madd_h(p16_2, mone); - accum1 = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib + 0].d)*GGML_FP16_TO_FP32(x[ib + 0].d)), - __lasx_xvffint_s_w(p_1), accum1); - accum2 = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(y[ib + 1].d)*GGML_FP16_TO_FP32(x[ib + 1].d)), - __lasx_xvffint_s_w(p_2), accum2); - } - - sumf = hsum_float_8(__lasx_xvfadd_s(accum1, accum2)); - -#endif - for (; ib < nb; ++ib) { - const float d = GGML_FP16_TO_FP32(y[ib].d)*GGML_FP16_TO_FP32(x[ib].d); - int sumi1 = 0, sumi2 = 0; - for (int j = 0; j < QK4_NL/2; ++j) { - sumi1 += y[ib].qs[j+ 0] * kvalues_iq4nl[x[ib].qs[j] & 0xf]; - sumi2 += y[ib].qs[j+QK4_NL/2] * kvalues_iq4nl[x[ib].qs[j] >> 4]; - } - sumf += d * (sumi1 + sumi2); - } - *s = sumf; -} - -void ggml_vec_dot_iq4_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { - assert(nrc == 1); - UNUSED(nrc); - UNUSED(bx); - UNUSED(by); - UNUSED(bs); - assert(n % QK_K == 0); - - const block_iq4_xs * restrict x = vx; - const block_q8_K * restrict y = vy; - - const int nb = n / QK_K; - -#if defined __ARM_NEON - const int8x16_t values = vld1q_s8(kvalues_iq4nl); - const uint8x16_t m4b = vdupq_n_u8(0x0f); - ggml_uint8x16x2_t q4bits; - ggml_int8x16x4_t q4b; - ggml_int8x16x4_t q8b; - int32x4_t prod_1, prod_2; - - float sumf = 0; - - for (int ibl = 0; ibl < nb; ++ibl) { - - const int8_t * q8 = y[ibl].qs; - const uint8_t * q4 = x[ibl].qs; - uint16_t h = x[ibl].scales_h; - - int sumi1 = 0, sumi2 = 0; - for (int ib = 0; ib < QK_K/64; ++ib) { - - q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; - q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - - q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); - q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); - q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); - q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); - - prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); - prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); - - int ls1 = ((x[ibl].scales_l[ib] & 0xf) | ((h << 4) & 0x30)) - 32; - int ls2 = ((x[ibl].scales_l[ib] >> 4) | ((h << 2) & 0x30)) - 32; - h >>= 4; - sumi1 += vaddvq_s32(prod_1) * ls1; - sumi2 += vaddvq_s32(prod_2) * ls2; - - } - - sumf += GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d * (sumi1 + sumi2); - } - - *s = sumf; - -#elif defined __AVX2__ - - const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); - const __m128i m4b = _mm_set1_epi8(0x0f); - - __m256 accum = _mm256_setzero_ps(); - for (int ibl = 0; ibl < nb; ++ibl) { - const uint8_t * qs = x[ibl].qs; - const int8_t * q8 = y[ibl].qs; - uint16_t sh = x[ibl].scales_h; - __m256i sumi1 = _mm256_setzero_si256(); - __m256i sumi2 = _mm256_setzero_si256(); - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)qs); qs += 16; - const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)qs); qs += 16; - const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), - _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); - const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), - _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); - const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); - const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); - const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; - const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; - sh >>= 4; - const __m256i p_1 = _mm256_madd_epi16(p16_1, _mm256_set1_epi16(ls1)); - const __m256i p_2 = _mm256_madd_epi16(p16_2, _mm256_set1_epi16(ls2)); - sumi1 = _mm256_add_epi32(p_1, sumi1); - sumi2 = _mm256_add_epi32(p_2, sumi2); - } - accum = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), - _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accum); - } - - *s = hsum_float_8(accum); - -#elif defined __AVX__ - const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); - const __m128i m4b = _mm_set1_epi8(0x0f); - - __m256 accum = _mm256_setzero_ps(); - for (int ibl = 0; ibl < nb; ++ibl) { - const uint8_t * qs = x[ibl].qs; - const int8_t * q8 = y[ibl].qs; - uint16_t sh = x[ibl].scales_h; - __m128i sumi1_0 = _mm_setzero_si128(); - __m128i sumi1_1 = _mm_setzero_si128(); - __m128i sumi2_0 = _mm_setzero_si128(); - __m128i sumi2_1 = _mm_setzero_si128(); - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m128i q4bits_1 = _mm_loadu_si128((const __m128i *)qs); qs += 16; - const __m128i q4bits_2 = _mm_loadu_si128((const __m128i *)qs); qs += 16; - const __m128i q8b_1_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_1_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_0 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q8b_2_1 = _mm_loadu_si128((const __m128i *)q8); q8 += 16; - const __m128i q4b_1_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b)); - const __m128i q4b_1_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)); - const __m128i q4b_2_0 = _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b)); - const __m128i q4b_2_1 = _mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)); - const __m128i p16_1_0 = mul_add_epi8_sse(q4b_1_0, q8b_1_0); - const __m128i p16_1_1 = mul_add_epi8_sse(q4b_1_1, q8b_1_1); - const __m128i p16_2_0 = mul_add_epi8_sse(q4b_2_0, q8b_2_0); - const __m128i p16_2_1 = mul_add_epi8_sse(q4b_2_1, q8b_2_1); - const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; - const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; - sh >>= 4; - const __m128i p_1_0 = _mm_madd_epi16(p16_1_0, _mm_set1_epi16(ls1)); - const __m128i p_1_1 = _mm_madd_epi16(p16_1_1, _mm_set1_epi16(ls1)); - const __m128i p_2_0 = _mm_madd_epi16(p16_2_0, _mm_set1_epi16(ls2)); - const __m128i p_2_1 = _mm_madd_epi16(p16_2_1, _mm_set1_epi16(ls2)); - sumi1_0 = _mm_add_epi32(p_1_0, sumi1_0); - sumi1_1 = _mm_add_epi32(p_1_1, sumi1_1); - sumi2_0 = _mm_add_epi32(p_2_0, sumi2_0); - sumi2_1 = _mm_add_epi32(p_2_1, sumi2_1); - } - __m128i sumi12_0 = _mm_add_epi32(sumi1_0, sumi2_0); - __m128i sumi12_1 = _mm_add_epi32(sumi1_1, sumi2_1); - accum = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), - _mm256_cvtepi32_ps(MM256_SET_M128I(sumi12_1, sumi12_0))), accum); - } - - *s = hsum_float_8(accum); - -#elif defined(__POWER9_VECTOR__) - const vector signed char lowMask = vec_splats((signed char)0xF); - const vector int v0 = vec_splats((int32_t)0); - const vector unsigned char v4 = vec_splats((unsigned char)0x4); - - vector float vsumf0 = vec_splats(0.0f); - vector float vsumf1 = vec_splats(0.0f); - vector float vsumf2 = vec_splats(0.0f); - vector float vsumf3 = vec_splats(0.0f); - - const vector signed char values = vec_xl( 0, kvalues_iq4nl); - - for (int ibl = 0; ibl < nb; ++ibl) { - - vector float vxd = vec_splats(GGML_FP16_TO_FP32(x[ibl].d)); - vector float vyd = vec_splats(y[ibl].d); - vector float vd = vec_mul(vxd, vyd); - - vector signed int vsumi0 = v0; - vector signed int vsumi1 = v0; - vector signed int vsumi2 = v0; - vector signed int vsumi3 = v0; - - uint16_t h = x[ibl].scales_h; - - const uint8_t * restrict q4 = x[ibl].qs; - const uint8_t * restrict sc = x[ibl].scales_l; - const int8_t * restrict q8 = y[ibl].qs; - - for (int ib = 0; ib < QK_K/64; ib ++ ) { - __builtin_prefetch(q4, 0, 1); - __builtin_prefetch(q8, 0, 1); - - vector signed char qxs0 = (vector signed char)vec_xl( 0, q4); - vector signed char qxs1 = (vector signed char)vec_xl(16, q4); - q4 += 32; - - vector signed char q4x00 = (vector signed char)vec_and(qxs0, lowMask); - vector signed char q4x01 = (vector signed char)vec_sr(qxs0, v4); - vector signed char q4x10 = (vector signed char)vec_and(qxs1, lowMask); - vector signed char q4x11 = (vector signed char)vec_sr(qxs1, v4); - - q4x00 = vec_perm(values, values, (vector unsigned char)q4x00); - q4x01 = vec_perm(values, values, (vector unsigned char)q4x01); - q4x10 = vec_perm(values, values, (vector unsigned char)q4x10); - q4x11 = vec_perm(values, values, (vector unsigned char)q4x11); - - vector signed char q8y0 = vec_xl( 0, q8); - vector signed char q8y1 = vec_xl(16, q8); - vector signed char q8y2 = vec_xl(32, q8); - vector signed char q8y3 = vec_xl(48, q8); - q8 += 64; - - vector signed short qv0 = vec_add(vec_mule(q4x00, q8y0), vec_mulo(q4x00, q8y0)); - vector signed short qv1 = vec_add(vec_mule(q4x01, q8y1), vec_mulo(q4x01, q8y1)); - vector signed short qv2 = vec_add(vec_mule(q4x10, q8y2), vec_mulo(q4x10, q8y2)); - vector signed short qv3 = vec_add(vec_mule(q4x11, q8y3), vec_mulo(q4x11, q8y3)); - - const uint16_t ls0 = (uint16_t)(((sc[0] & 0xf) | ((h << 4) & 0x30)) - 32); - const uint16_t ls1 = (uint16_t)(((sc[0] >> 4) | ((h << 2) & 0x30)) - 32); - h >>= 4; - sc ++; - - vector signed short vscales01 = vec_splats((int16_t)ls0); - vector signed short vscales23 = vec_splats((int16_t)ls1); - - vsumi0 = vec_msum(qv0, vscales01, vsumi0); - vsumi1 = vec_msum(qv1, vscales01, vsumi1); - vsumi2 = vec_msum(qv2, vscales23, vsumi2); - vsumi3 = vec_msum(qv3, vscales23, vsumi3); - } - - vsumf0 = vec_madd(vec_ctf(vsumi0, 0), vd, vsumf0); - vsumf1 = vec_madd(vec_ctf(vsumi1, 0), vd, vsumf1); - vsumf2 = vec_madd(vec_ctf(vsumi2, 0), vd, vsumf2); - vsumf3 = vec_madd(vec_ctf(vsumi3, 0), vd, vsumf3); - } - - vsumf0 = vec_add(vsumf0, vsumf2); - vsumf1 = vec_add(vsumf1, vsumf3); - - vsumf0 = vec_add(vsumf0, vsumf1); - - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 4)); - vsumf0 = vec_add(vsumf0, vec_sld(vsumf0, vsumf0, 8)); - - *s = vec_extract(vsumf0, 0); - -#elif defined(__loongarch_asx) - - const __m128i values128 = __lsx_vld((const __m128i*)kvalues_iq4nl, 0); - const __m128i m4b = __lsx_vreplgr2vr_b(0x0f); - - __m256 accum = (__m256)__lasx_xvldi(0); - __m256i tmp1; - __m128i tmp0, tmp2, tmp3, tmp4, mask_8f, mask; - - mask_8f = __lsx_vreplgr2vr_b(0x8f); - for (int ibl = 0; ibl < nb; ++ibl) { - const uint8_t * qs = x[ibl].qs; - const int8_t * q8 = y[ibl].qs; - uint16_t sh = x[ibl].scales_h; - __m256i sumi1 = __lasx_xvldi(0); - __m256i sumi2 = __lasx_xvldi(0); - __m128i zero = __lsx_vldi(0); - for (int ib = 0; ib < QK_K/32; ib += 2) { - const __m128i q4bits_1 = __lsx_vld((const __m128i*)qs, 0); qs += 16; - const __m128i q4bits_2 = __lsx_vld((const __m128i*)qs, 0); qs += 16; - const __m256i q8b_1 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - const __m256i q8b_2 = __lasx_xvld((const __m256i *)q8, 0); q8 += 32; - tmp2 = __lsx_vand_v(__lsx_vand_v(__lsx_vsrli_h(q4bits_1, 4), m4b), mask_8f); - tmp0 = __lsx_vori_b(tmp2, 0x10); - mask = __lsx_vsle_b(zero, tmp2); - tmp3 = __lsx_vand_v(tmp0, mask); - tmp3 = __lsx_vshuf_b(values128, zero, tmp3); - - tmp2 = __lsx_vand_v(__lsx_vand_v(q4bits_1, m4b), mask_8f); - tmp0 = __lsx_vori_b(tmp2, 0x10); - mask = __lsx_vsle_b(zero, tmp2); - tmp4 = __lsx_vand_v(tmp0, mask); - tmp4 = __lsx_vshuf_b(values128, zero, tmp4); - - const __m256i q4b_1 = lasx_insertf128(tmp3, tmp4); - - tmp2 = __lsx_vand_v(__lsx_vand_v(__lsx_vsrli_h(q4bits_2, 4), m4b), mask_8f); - tmp0 = __lsx_vori_b(tmp2, 0x10); - mask = __lsx_vsle_b(zero, tmp2); - tmp3 = __lsx_vand_v(tmp0, mask); - tmp3 = __lsx_vshuf_b(values128, zero, tmp3); - - tmp2 = __lsx_vand_v(__lsx_vand_v(q4bits_2, m4b), mask_8f); - tmp0 = __lsx_vori_b(tmp2, 0x10); - mask = __lsx_vsle_b(zero, tmp2); - tmp4 = __lsx_vand_v(tmp0, mask); - tmp4 = __lsx_vshuf_b(values128, zero, tmp4); - - const __m256i q4b_2 = lasx_insertf128(tmp3, tmp4); - - const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); - const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); - const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; - const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; - sh >>= 4; - __m256i tmp5, tmp6; - tmp1 = __lasx_xvreplgr2vr_h(ls1); - tmp5 = __lasx_xvmulwev_w_h(p16_1, tmp1); - tmp6 = __lasx_xvmulwod_w_h(p16_1, tmp1); - const __m256i p_1 = __lasx_xvadd_w(tmp5, tmp6); - tmp1 = __lasx_xvreplgr2vr_h(ls2); - tmp5 = __lasx_xvmulwev_w_h(p16_2, tmp1); - tmp6 = __lasx_xvmulwod_w_h(p16_2, tmp1); - const __m256i p_2 = __lasx_xvadd_w(tmp5, tmp6); - sumi1 = __lasx_xvadd_w(p_1, sumi1); - sumi2 = __lasx_xvadd_w(p_2, sumi2); - } - accum = __lasx_xvfmadd_s(__lasx_xvreplfr2vr_s(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), - __lasx_xvffint_s_w(__lasx_xvadd_w(sumi1, sumi2)), accum); - } - - *s = hsum_float_8(accum); - -#else - float sumf = 0; - for (int ibl = 0; ibl < nb; ++ibl) { - const float d4d8 = GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d; - uint16_t h = x[ibl].scales_h; - const uint8_t * qs = x[ibl].qs; - const int8_t * q8 = y[ibl].qs; - for (int ib = 0; ib < QK_K/32; ib += 2) { - const uint8_t ls1 = (x[ibl].scales_l[ib/2] & 0xf) | ((h << 4) & 0x30); - const uint8_t ls2 = (x[ibl].scales_l[ib/2] >> 4) | ((h << 2) & 0x30); - h >>= 4; - const float d1 = d4d8*(ls1 - 32); - const float d2 = d4d8*(ls2 - 32); - int sumi1 = 0, sumi2 = 0; - for (int j = 0; j < 16; ++j) { - sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; - sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; - } - sumf += d1 * (sumi1 + sumi2); - qs += 16; - q8 += 32; - sumi1 = sumi2 = 0; - for (int j = 0; j < 16; ++j) { - sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; - sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; - } - sumf += d2 * (sumi1 + sumi2); - qs += 16; - q8 += 32; - } - } - *s = sumf; -#endif -} - // ================================ IQ2 quantization ============================================= typedef struct { @@ -14236,12 +3770,6 @@ size_t quantize_iq3_xxs(const float * restrict src, void * restrict dst, int64_t return nrow * nblock * sizeof(block_iq3_xxs); } -void quantize_row_iq3_xxs(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_iq3_xxs * restrict y = vy; - quantize_row_iq3_xxs_ref(x, y, k); -} - void quantize_row_iq3_xxs_ref(const float * restrict x, block_iq3_xxs * restrict y, int64_t k) { assert(k % QK_K == 0); quantize_row_iq3_xxs_impl(256, x, y, k, NULL); @@ -14452,12 +3980,6 @@ size_t quantize_iq3_s(const float * restrict src, void * restrict dst, int64_t n return nrow * nblock * sizeof(block_iq3_s); } -void quantize_row_iq3_s(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_iq3_s * restrict y = vy; - quantize_row_iq3_s_ref(x, y, k); -} - void quantize_row_iq3_s_ref(const float * restrict x, block_iq3_s * restrict y, int64_t k) { assert(k % QK_K == 0); quantize_iq3_s(x, y, 1, k, NULL); @@ -15181,7 +4703,8 @@ size_t quantize_iq4_nl(const float * restrict src, void * restrict dst, int64_t return nrow * nblock * sizeof(block_iq4_nl); } -void quantize_row_iq4_nl(const float * restrict x, void * restrict vy, int64_t k) { +//void quantize_row_iq4_nl_ref(const float * restrict x, void * restrict vy, int64_t k) { +void quantize_row_iq4_nl_ref(const float * restrict x, block_iq4_nl * restrict y, int64_t k) { GGML_ASSERT(k%QK4_NL == 0); int64_t nblock = k/QK4_NL; uint8_t L[QK4_NL]; @@ -15189,18 +4712,13 @@ void quantize_row_iq4_nl(const float * restrict x, void * restrict vy, int64_t k uint16_t unused_h; uint8_t * unused_l = NULL; float scale; - block_iq4_nl * iq4 = (block_iq4_nl *)vy; + block_iq4_nl * iq4 = y; for (int ibl = 0; ibl < nblock; ++ibl) { quantize_row_iq4_nl_impl(QK4_NL, 32, x + QK4_NL*ibl, &iq4[ibl].d, iq4[ibl].qs, &unused_h, unused_l, &scale, weight, L, kvalues_iq4nl, NULL, -1); } } -void quantize_row_iq4_nl_ref(const float * restrict x, block_iq4_nl * restrict y, int64_t k) { - assert(k % QK4_NL == 0); - quantize_row_iq4_nl(x, y, k); -} - size_t quantize_iq4_xs(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { GGML_ASSERT(n_per_row%QK_K == 0); int64_t nblock = n_per_row/QK_K; @@ -15221,12 +4739,6 @@ size_t quantize_iq4_xs(const float * restrict src, void * restrict dst, int64_t return nrow * nblock * sizeof(block_iq4_xs); } -void quantize_row_iq4_xs(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_iq4_xs * restrict y = vy; - quantize_row_iq4_xs_ref(x, y, k); -} - void quantize_row_iq4_xs_ref(const float * restrict x, block_iq4_xs * restrict y, int64_t k) { assert(k % QK_K == 0); quantize_iq4_xs(x, y, 1, k, NULL); @@ -15419,11 +4931,7 @@ void quantize_row_iq2_s_ref(const float * restrict x, block_iq2_s * restrict y, quantize_iq2_s(x, y, 1, k, NULL); } -void quantize_row_iq2_s(const float * restrict x, void * restrict vy, int64_t k) { - assert(k % QK_K == 0); - block_iq2_s * restrict y = vy; - quantize_row_iq2_s_ref(x, y, k); -} +// =============================== data validation static bool validate_float(float f, size_t i) { if (isinf(f)) { @@ -15712,15 +5220,6 @@ bool ggml_validate_row_data(enum ggml_type type, const void * data, size_t nbyte { VALIDATE_ROW_DATA_D_F16_IMPL(block_iq4_nl, data, nb); } break; - case GGML_TYPE_Q4_0_4_4: - case GGML_TYPE_Q4_0_4_8: - { - VALIDATE_ROW_DATA_DVEC_F16_IMPL(block_q4_0x4, data, nbytes / sizeof(block_q4_0x4), 4); - } break; - case GGML_TYPE_Q4_0_8_8: - { - VALIDATE_ROW_DATA_DVEC_F16_IMPL(block_q4_0x8, data, nbytes / sizeof(block_q4_0x8), 8); - } break; case GGML_TYPE_I8: case GGML_TYPE_I16: diff --git a/ggml/src/ggml-quants.h b/ggml/src/ggml-quants.h index df9c4b24a..d09173e11 100644 --- a/ggml/src/ggml-quants.h +++ b/ggml/src/ggml-quants.h @@ -11,136 +11,89 @@ extern "C" { #endif +// NOTE: these functions are defined as GGML_API because they used by the CPU backend + // Quantization -void quantize_row_q4_0_ref(const float * GGML_RESTRICT x, block_q4_0 * GGML_RESTRICT y, int64_t k); -void quantize_row_q4_1_ref(const float * GGML_RESTRICT x, block_q4_1 * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_0_ref(const float * GGML_RESTRICT x, block_q5_0 * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_1_ref(const float * GGML_RESTRICT x, block_q5_1 * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_0_ref(const float * GGML_RESTRICT x, block_q8_0 * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_1_ref(const float * GGML_RESTRICT x, block_q8_1 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q4_0_ref(const float * GGML_RESTRICT x, block_q4_0 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q4_1_ref(const float * GGML_RESTRICT x, block_q4_1 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q5_0_ref(const float * GGML_RESTRICT x, block_q5_0 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q5_1_ref(const float * GGML_RESTRICT x, block_q5_1 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q8_0_ref(const float * GGML_RESTRICT x, block_q8_0 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q8_1_ref(const float * GGML_RESTRICT x, block_q8_1 * GGML_RESTRICT y, int64_t k); -void quantize_row_q2_K_ref(const float * GGML_RESTRICT x, block_q2_K * GGML_RESTRICT y, int64_t k); -void quantize_row_q3_K_ref(const float * GGML_RESTRICT x, block_q3_K * GGML_RESTRICT y, int64_t k); -void quantize_row_q4_K_ref(const float * GGML_RESTRICT x, block_q4_K * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_K_ref(const float * GGML_RESTRICT x, block_q5_K * GGML_RESTRICT y, int64_t k); -void quantize_row_q6_K_ref(const float * GGML_RESTRICT x, block_q6_K * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_K_ref(const float * GGML_RESTRICT x, block_q8_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q2_K_ref(const float * GGML_RESTRICT x, block_q2_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q3_K_ref(const float * GGML_RESTRICT x, block_q3_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q4_K_ref(const float * GGML_RESTRICT x, block_q4_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q5_K_ref(const float * GGML_RESTRICT x, block_q5_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q6_K_ref(const float * GGML_RESTRICT x, block_q6_K * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_q8_K_ref(const float * GGML_RESTRICT x, block_q8_K * GGML_RESTRICT y, int64_t k); -void quantize_row_tq1_0_ref(const float * GGML_RESTRICT x, block_tq1_0 * GGML_RESTRICT y, int64_t k); -void quantize_row_tq2_0_ref(const float * GGML_RESTRICT x, block_tq2_0 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_tq1_0_ref(const float * GGML_RESTRICT x, block_tq1_0 * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_tq2_0_ref(const float * GGML_RESTRICT x, block_tq2_0 * GGML_RESTRICT y, int64_t k); -void quantize_row_iq3_xxs_ref(const float * GGML_RESTRICT x, block_iq3_xxs * GGML_RESTRICT y, int64_t k); -void quantize_row_iq4_nl_ref (const float * GGML_RESTRICT x, block_iq4_nl * GGML_RESTRICT y, int64_t k); -void quantize_row_iq4_xs_ref (const float * GGML_RESTRICT x, block_iq4_xs * GGML_RESTRICT y, int64_t k); -void quantize_row_iq3_s_ref (const float * GGML_RESTRICT x, block_iq3_s * GGML_RESTRICT y, int64_t k); -void quantize_row_iq2_s_ref (const float * GGML_RESTRICT x, block_iq2_s * GGML_RESTRICT y, int64_t k); - -void quantize_row_q4_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q4_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); - -void quantize_row_q2_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q3_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q4_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q5_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q6_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_q8_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); - -void quantize_row_tq1_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_tq2_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); - -void quantize_row_iq3_xxs(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_iq4_nl (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_iq4_xs (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_iq3_s (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); -void quantize_row_iq2_s (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_iq3_xxs_ref(const float * GGML_RESTRICT x, block_iq3_xxs * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_iq4_nl_ref (const float * GGML_RESTRICT x, block_iq4_nl * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_iq4_xs_ref (const float * GGML_RESTRICT x, block_iq4_xs * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_iq3_s_ref (const float * GGML_RESTRICT x, block_iq3_s * GGML_RESTRICT y, int64_t k); +GGML_API void quantize_row_iq2_s_ref (const float * GGML_RESTRICT x, block_iq2_s * GGML_RESTRICT y, int64_t k); // Dequantization -void dequantize_row_q4_0(const block_q4_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q4_1(const block_q4_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q5_0(const block_q5_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q5_1(const block_q5_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q8_0(const block_q8_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -//void dequantize_row_q8_1(const block_q8_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q4_0(const block_q4_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q4_1(const block_q4_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q5_0(const block_q5_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q5_1(const block_q5_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q8_0(const block_q8_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +//GGML_API void dequantize_row_q8_1(const block_q8_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q2_K(const block_q2_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q3_K(const block_q3_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q4_K(const block_q4_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q5_K(const block_q5_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q6_K(const block_q6_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_q8_K(const block_q8_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q2_K(const block_q2_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q3_K(const block_q3_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q4_K(const block_q4_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q5_K(const block_q5_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q6_K(const block_q6_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_q8_K(const block_q8_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_tq1_0(const block_tq1_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_tq2_0(const block_tq2_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_tq1_0(const block_tq1_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_tq2_0(const block_tq2_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq2_xxs(const block_iq2_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq2_xs (const block_iq2_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq2_s (const block_iq2_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq3_xxs(const block_iq3_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq1_s (const block_iq1_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq1_m (const block_iq1_m * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq4_nl (const block_iq4_nl * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq4_xs (const block_iq4_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); -void dequantize_row_iq3_s (const block_iq3_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); - -// Dot product -void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q5_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q5_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); - -void ggml_vec_dot_q2_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q3_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q4_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q5_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_q6_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); - -void ggml_vec_dot_tq1_0_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_tq2_0_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); - -void ggml_vec_dot_iq2_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq2_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq2_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq3_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq1_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq1_m_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq4_nl_q8_0 (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq4_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); -void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +GGML_API void dequantize_row_iq2_xxs(const block_iq2_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq2_xs (const block_iq2_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq2_s (const block_iq2_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq3_xxs(const block_iq3_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq1_s (const block_iq1_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq1_m (const block_iq1_m * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq4_nl (const block_iq4_nl * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq4_xs (const block_iq4_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +GGML_API void dequantize_row_iq3_s (const block_iq3_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); // Quantization utilizing an importance matrix (a.k.a. "Activation aWare Quantization") -size_t quantize_iq2_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq2_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq2_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq3_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq1_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq1_m (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq4_nl (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq4_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_iq3_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq2_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq2_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq2_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq3_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq1_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq1_m (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq4_nl (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq4_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_iq3_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_tq1_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_tq2_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_tq1_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_tq2_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q2_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q3_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q4_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q5_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q6_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q4_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q4_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q5_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q5_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -size_t quantize_q8_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q2_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q3_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q4_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q5_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q6_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q4_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q4_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q5_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q5_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +GGML_API size_t quantize_q8_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); -void iq2xs_init_impl(enum ggml_type type); -void iq2xs_free_impl(enum ggml_type type); -void iq3xs_init_impl(int grid_size); -void iq3xs_free_impl(int grid_size); +GGML_API void iq2xs_init_impl(enum ggml_type type); +GGML_API void iq2xs_free_impl(enum ggml_type type); +GGML_API void iq3xs_init_impl(int grid_size); +GGML_API void iq3xs_free_impl(int grid_size); #ifdef __cplusplus } diff --git a/ggml/src/ggml-rpc/CMakeLists.txt b/ggml/src/ggml-rpc/CMakeLists.txt new file mode 100644 index 000000000..f5acb8ec2 --- /dev/null +++ b/ggml/src/ggml-rpc/CMakeLists.txt @@ -0,0 +1,9 @@ +message(STATUS "Using RPC backend") + +ggml_add_backend_library(ggml-rpc + ggml-rpc.cpp + ) + +if (WIN32) + target_link_libraries(ggml-rpc PRIVATE ws2_32) +endif() diff --git a/ggml/src/ggml-rpc.cpp b/ggml/src/ggml-rpc/ggml-rpc.cpp similarity index 89% rename from ggml/src/ggml-rpc.cpp rename to ggml/src/ggml-rpc/ggml-rpc.cpp index 8a772f224..63da2b86b 100644 --- a/ggml/src/ggml-rpc.cpp +++ b/ggml/src/ggml-rpc/ggml-rpc.cpp @@ -27,15 +27,6 @@ #endif #include -#define UNUSED GGML_UNUSED - -#define GGML_DEBUG 0 -#if (GGML_DEBUG >= 1) -#define GGML_PRINT_DEBUG(...) printf(__VA_ARGS__) -#else -#define GGML_PRINT_DEBUG(...) -#endif - #ifdef _WIN32 typedef SOCKET sockfd_t; using ssize_t = __int64; @@ -93,9 +84,23 @@ enum rpc_cmd { RPC_CMD_COPY_TENSOR, RPC_CMD_GRAPH_COMPUTE, RPC_CMD_GET_DEVICE_MEMORY, + RPC_CMD_INIT_TENSOR, + RPC_CMD_GET_ALLOC_SIZE, RPC_CMD_COUNT, }; +struct rpc_msg_get_alloc_size_req { + rpc_tensor tensor; +}; + +struct rpc_msg_get_alloc_size_rsp { + uint64_t alloc_size; +}; + +struct rpc_msg_init_tensor_req { + rpc_tensor tensor; +}; + struct rpc_msg_alloc_buffer_req { uint64_t size; }; @@ -397,7 +402,7 @@ static std::shared_ptr get_socket(const std::string & endpoint) { initialized = true; } #else - UNUSED(initialized); + GGML_UNUSED(initialized); #endif auto sock = socket_connect(host.c_str(), port); if (sock == nullptr) { @@ -461,10 +466,18 @@ static rpc_tensor serialize_tensor(const ggml_tensor * tensor) { } static void ggml_backend_rpc_buffer_init_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor) { - UNUSED(buffer); - if (ggml_is_quantized(tensor->type)) { - // TODO: this check is due to MATRIX_ROW_PADDING in CUDA and should be generalized - GGML_ASSERT(tensor->ne[0] % 512 == 0 && "unsupported quantized tensor"); + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + + // CUDA backend on the server pads everything to 512 due to CUDA limitations. + // Due to bandwidth constraints, we only call the server init tensor functions if necessary. + // In particular, only quantized tensors need padding + if (ggml_is_quantized(tensor->type) && (tensor->ne[0] % 512 != 0) && (tensor->view_src == nullptr)) { + rpc_msg_init_tensor_req request; + + request.tensor = serialize_tensor(tensor); + + bool status = send_rpc_cmd(ctx->sock, RPC_CMD_INIT_TENSOR, &request, sizeof(request), nullptr, 0); + GGML_ASSERT(status); } } @@ -577,8 +590,23 @@ static size_t ggml_backend_rpc_get_max_size(ggml_backend_buffer_type_t buft) { } static size_t ggml_backend_rpc_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { - UNUSED(buft); - return ggml_nbytes(tensor); + // See comments in init_tensor. + if (ggml_is_quantized(tensor->type) && (tensor->ne[0] % 512 != 0) && (tensor->view_src == nullptr)) { + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + auto sock = get_socket(buft_ctx->endpoint); + + rpc_msg_get_alloc_size_req request; + + request.tensor = serialize_tensor(tensor); + + rpc_msg_get_alloc_size_rsp response; + bool status = send_rpc_cmd(sock, RPC_CMD_GET_ALLOC_SIZE, &request, sizeof(request), &response, sizeof(response)); + GGML_ASSERT(status); + + return response.alloc_size; + } else { + return ggml_nbytes(tensor); + } } static ggml_backend_buffer_type_i ggml_backend_rpc_buffer_type_interface = { @@ -603,7 +631,7 @@ static void ggml_backend_rpc_free(ggml_backend_t backend) { } static void ggml_backend_rpc_synchronize(ggml_backend_t backend) { - UNUSED(backend); + GGML_UNUSED(backend); // this is no-op because we don't have any async operations } @@ -671,7 +699,7 @@ static ggml_backend_i ggml_backend_rpc_interface = { /* .event_wait = */ NULL, }; -GGML_API ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint) { +ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint) { static std::mutex mutex; std::lock_guard lock(mutex); // NOTE: buffer types are allocated and never freed; this is by design @@ -718,7 +746,7 @@ ggml_backend_t ggml_backend_rpc_init(const char * endpoint) { return backend; } -GGML_API bool ggml_backend_is_rpc(ggml_backend_t backend) { +bool ggml_backend_is_rpc(ggml_backend_t backend) { return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_rpc_guid()); } @@ -730,7 +758,7 @@ static void get_device_memory(const std::shared_ptr & sock, size_t * f *total = response.total_mem; } -GGML_API void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total) { +void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total) { auto sock = get_socket(endpoint); if (sock == nullptr) { *free = 0; @@ -757,6 +785,8 @@ public: bool get_tensor(const rpc_msg_get_tensor_req & request, std::vector & response); bool copy_tensor(const rpc_msg_copy_tensor_req & request, rpc_msg_copy_tensor_rsp & response); bool graph_compute(const std::vector & input, rpc_msg_graph_compute_rsp & response); + bool init_tensor(const rpc_msg_init_tensor_req & request); + bool get_alloc_size(const rpc_msg_get_alloc_size_req & request, rpc_msg_get_alloc_size_rsp & response); private: ggml_tensor * deserialize_tensor(struct ggml_context * ctx, const rpc_tensor * tensor); @@ -770,6 +800,36 @@ private: std::unordered_set buffers; }; +bool rpc_server::get_alloc_size(const rpc_msg_get_alloc_size_req & request, rpc_msg_get_alloc_size_rsp & response) { + ggml_backend_buffer_type_t buft; + struct ggml_init_params params { + /*.mem_size =*/ ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + + struct ggml_context * ctx = ggml_init(params); + ggml_tensor * tensor = deserialize_tensor(ctx, &request.tensor); + + if (tensor == nullptr) { + GGML_LOG_ERROR("Null tensor pointer passed to server get_alloc_size function.\n"); + ggml_free(ctx); + return false; + } + + if (tensor->buffer == nullptr) { + //No buffer allocated. + buft = ggml_backend_get_default_buffer_type(backend); + } else { + buft = tensor->buffer->buft; + } + + response.alloc_size = ggml_backend_buft_get_alloc_size(buft,tensor); + + ggml_free(ctx); + return true; +} + void rpc_server::alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_alloc_buffer_rsp & response) { ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend); ggml_backend_buffer_t buffer = ggml_backend_buft_alloc_buffer(buft, request.size); @@ -781,7 +841,7 @@ void rpc_server::alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_ GGML_PRINT_DEBUG("[%s] size: %" PRIu64 " -> remote_ptr: %" PRIx64 ", remote_size: %" PRIu64 "\n", __func__, request.size, response.remote_ptr, response.remote_size); buffers.insert(buffer); } else { - GGML_PRINT_DEBUG("[%s] size: %" PRIu64 " -> failed\n", __func__, request.size); + GGML_LOG_ERROR("[%s] size: %" PRIu64 " -> failed\n", __func__, request.size); } } @@ -803,7 +863,7 @@ bool rpc_server::buffer_get_base(const rpc_msg_buffer_get_base_req & request, rp GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 "\n", __func__, request.remote_ptr); ggml_backend_buffer_t buffer = reinterpret_cast(request.remote_ptr); if (buffers.find(buffer) == buffers.end()) { - GGML_PRINT_DEBUG("[%s] buffer not found\n", __func__); + GGML_LOG_ERROR("[%s] buffer not found\n", __func__); return false; } void * base = ggml_backend_buffer_get_base(buffer); @@ -815,7 +875,7 @@ bool rpc_server::free_buffer(const rpc_msg_free_buffer_req & request) { GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 "\n", __func__, request.remote_ptr); ggml_backend_buffer_t buffer = reinterpret_cast(request.remote_ptr); if (buffers.find(buffer) == buffers.end()) { - GGML_PRINT_DEBUG("[%s] buffer not found\n", __func__); + GGML_LOG_ERROR("[%s] buffer not found\n", __func__); return false; } ggml_backend_buffer_free(buffer); @@ -827,7 +887,7 @@ bool rpc_server::buffer_clear(const rpc_msg_buffer_clear_req & request) { GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 ", value: %u\n", __func__, request.remote_ptr, request.value); ggml_backend_buffer_t buffer = reinterpret_cast(request.remote_ptr); if (buffers.find(buffer) == buffers.end()) { - GGML_PRINT_DEBUG("[%s] buffer not found\n", __func__); + GGML_LOG_ERROR("[%s] buffer not found\n", __func__); return false; } ggml_backend_buffer_clear(buffer, request.value); @@ -883,7 +943,7 @@ bool rpc_server::set_tensor(const std::vector & input) { struct ggml_context * ctx = ggml_init(params); ggml_tensor * tensor = deserialize_tensor(ctx, in_tensor); if (tensor == nullptr) { - GGML_PRINT_DEBUG("[%s] error deserializing tensor\n", __func__); + GGML_LOG_ERROR("[%s] error deserializing tensor\n", __func__); ggml_free(ctx); return false; } @@ -905,6 +965,40 @@ bool rpc_server::set_tensor(const std::vector & input) { return true; } +bool rpc_server::init_tensor(const rpc_msg_init_tensor_req & request) { + struct ggml_init_params params { + /*.mem_size =*/ ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + ggml_tensor * tensor = deserialize_tensor(ctx, &request.tensor); + if (tensor == nullptr) { + GGML_LOG_ERROR("Null tensor pointer passed to server init_tensor function.\n"); + ggml_free(ctx); + return false; + } + + // Call the backend's buffer_init_tensor function + ggml_backend_buffer_t buffer = tensor->buffer; + if (buffer && buffer->iface.init_tensor) { + buffer->iface.init_tensor(buffer, tensor); + } else { + GGML_LOG_ERROR("Null buffer for tensor passed to init_tensor function\n"); + } + + if (tensor->extra != nullptr) { + // This pointer can either be passed around client/server, or probably better stored server-side and kept track of. + // Currently unimplemented. + GGML_LOG_ERROR("tensor->extra populated by the backend, this is currently unsupported.\n"); + ggml_free(ctx); + return false; + } + + ggml_free(ctx); + return true; +} + bool rpc_server::get_tensor(const rpc_msg_get_tensor_req & request, std::vector & response) { struct ggml_init_params params { /*.mem_size =*/ ggml_tensor_overhead(), @@ -914,7 +1008,7 @@ bool rpc_server::get_tensor(const rpc_msg_get_tensor_req & request, std::vector< struct ggml_context * ctx = ggml_init(params); ggml_tensor * tensor = deserialize_tensor(ctx, &request.tensor); if (tensor == nullptr) { - GGML_PRINT_DEBUG("[%s] error deserializing tensor\n", __func__); + GGML_LOG_ERROR("[%s] error deserializing tensor\n", __func__); ggml_free(ctx); return false; } @@ -948,7 +1042,7 @@ bool rpc_server::copy_tensor(const rpc_msg_copy_tensor_req & request, rpc_msg_co ggml_tensor * src = deserialize_tensor(ctx, &request.src); ggml_tensor * dst = deserialize_tensor(ctx, &request.dst); if (src == nullptr || dst == nullptr) { - GGML_PRINT_DEBUG("[%s] error deserializing tensors\n", __func__); + GGML_LOG_ERROR("[%s] error deserializing tensors\n", __func__); ggml_free(ctx); return false; } @@ -1058,6 +1152,18 @@ static void rpc_serve_client(ggml_backend_t backend, sockfd_t sockfd, size_t fre } break; } + case RPC_CMD_GET_ALLOC_SIZE: { + rpc_msg_get_alloc_size_req request; + if (!recv_msg(sockfd, &request, sizeof(request))) { + return; + } + rpc_msg_get_alloc_size_rsp response; + server.get_alloc_size(request, response); + if (!send_msg(sockfd, &response, sizeof(response))) { + return; + } + break; + } case RPC_CMD_GET_ALIGNMENT: { if (!recv_msg(sockfd, nullptr, 0)) { return; @@ -1133,6 +1239,19 @@ static void rpc_serve_client(ggml_backend_t backend, sockfd_t sockfd, size_t fre } break; } + case RPC_CMD_INIT_TENSOR: { + rpc_msg_init_tensor_req request; + if (!recv_msg(sockfd, &request,sizeof(request))) { + return; + } + if (!server.init_tensor(request)) { + return; + } + if (!send_msg(sockfd, nullptr, 0)) { + return; + } + break; + } case RPC_CMD_GET_TENSOR: { rpc_msg_get_tensor_req request; if (!recv_msg(sockfd, &request, sizeof(request))) { @@ -1257,14 +1376,14 @@ static void ggml_backend_rpc_device_get_memory(ggml_backend_dev_t dev, size_t * ggml_backend_rpc_get_device_memory(ctx->endpoint.c_str(), free, total); - UNUSED(dev); + GGML_UNUSED(dev); } static enum ggml_backend_dev_type ggml_backend_rpc_device_get_type(ggml_backend_dev_t dev) { // TODO: obtain value from the server return GGML_BACKEND_DEVICE_TYPE_GPU; - UNUSED(dev); + GGML_UNUSED(dev); } static void ggml_backend_rpc_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { @@ -1285,7 +1404,7 @@ static ggml_backend_t ggml_backend_rpc_device_init(ggml_backend_dev_t dev, const return ggml_backend_rpc_init(ctx->endpoint.c_str()); - UNUSED(params); + GGML_UNUSED(params); } static ggml_backend_buffer_type_t ggml_backend_rpc_device_get_buffer_type(ggml_backend_dev_t dev) { @@ -1293,12 +1412,12 @@ static ggml_backend_buffer_type_t ggml_backend_rpc_device_get_buffer_type(ggml_b return ggml_backend_rpc_buffer_type(ctx->endpoint.c_str()); - UNUSED(dev); + GGML_UNUSED(dev); } static bool ggml_backend_rpc_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { - UNUSED(dev); - UNUSED(op); + GGML_UNUSED(dev); + GGML_UNUSED(op); //TODO: call the remote backend and cache the results return true; } @@ -1335,20 +1454,20 @@ static const struct ggml_backend_device_i ggml_backend_rpc_device_i = { static const char * ggml_backend_rpc_reg_get_name(ggml_backend_reg_t reg) { return "RPC"; - UNUSED(reg); + GGML_UNUSED(reg); } static size_t ggml_backend_rpc_reg_get_device_count(ggml_backend_reg_t reg) { return 0; - UNUSED(reg); + GGML_UNUSED(reg); } static ggml_backend_dev_t ggml_backend_rpc_reg_get_device(ggml_backend_reg_t reg, size_t index) { GGML_ABORT("The RPC backend does not have enumerated devices - use ggml_backend_add_device instead"); - UNUSED(reg); - UNUSED(index); + GGML_UNUSED(reg); + GGML_UNUSED(index); } static void * ggml_backend_rpc_get_proc_address(ggml_backend_reg_t reg, const char * name) { @@ -1357,7 +1476,7 @@ static void * ggml_backend_rpc_get_proc_address(ggml_backend_reg_t reg, const ch } return NULL; - UNUSED(reg); + GGML_UNUSED(reg); } static const struct ggml_backend_reg_i ggml_backend_rpc_reg_i = { @@ -1369,8 +1488,9 @@ static const struct ggml_backend_reg_i ggml_backend_rpc_reg_i = { ggml_backend_reg_t ggml_backend_rpc_reg(void) { static struct ggml_backend_reg ggml_backend_rpc_reg = { - /* .iface = */ ggml_backend_rpc_reg_i, - /* .context = */ NULL, + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_rpc_reg_i, + /* .context = */ NULL, }; return &ggml_backend_rpc_reg; @@ -1401,3 +1521,5 @@ ggml_backend_dev_t ggml_backend_rpc_add_device(const char * endpoint) { return dev; } + +GGML_BACKEND_DL_IMPL(ggml_backend_rpc_reg) diff --git a/ggml/src/ggml-sycl/CMakeLists.txt b/ggml/src/ggml-sycl/CMakeLists.txt new file mode 100644 index 000000000..3579a311a --- /dev/null +++ b/ggml/src/ggml-sycl/CMakeLists.txt @@ -0,0 +1,84 @@ +if (NOT GGML_SYCL_TARGET MATCHES "^(INTEL|NVIDIA|AMD)$") + message(FATAL_ERROR "Invalid backend chosen, supported options are INTEL, NVIDIA, or AMD") +endif() + +check_cxx_compiler_flag("-fsycl" SUPPORTS_SYCL) + +if (DEFINED ENV{ONEAPI_ROOT}) + message(STATUS "Using oneAPI Release SYCL compiler (icpx).") +elseif(SUPPORTS_SYCL) + message(WARNING "Using open-source SYCL compiler (clang++). Didn't detect ENV {ONEAPI_ROOT}. + If you expected the oneAPI Release compiler, please install oneAPI & source it, like: + source /opt/intel/oneapi/setvars.sh") +else() + message(FATAL_ERROR, "C++ compiler lacks SYCL support.") +endif() +message(STATUS "SYCL found") +#todo: AOT + +ggml_add_backend_library(ggml-sycl + ggml-sycl.cpp + ../../include/ggml-sycl.h + ) + +if (GGML_SYCL_F16) + if (GGML_SYCL_TARGET STREQUAL "AMD") + message(WARNING "AMD target does not entirely support FP16 in the SYCL backend.") + endif() + add_compile_definitions(GGML_SYCL_F16) +endif() + +set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-narrowing -fsycl") + +if (GGML_SYCL_TARGET STREQUAL "NVIDIA") + add_compile_definitions(GGML_SYCL_WARP_SIZE=32) +elseif (GGML_SYCL_TARGET STREQUAL "AMD") + # INFO: Allowed Sub_group_sizes are not consistent through all + # hip targets. For example, 64 is used for certain models, but the backend + # does not support it. + # Target archs tested working: gfx1030, gfx1031, (Only tested sub_group_size = 32) + add_compile_definitions(GGML_SYCL_WARP_SIZE=32) +else() + add_compile_definitions(GGML_SYCL_WARP_SIZE=16) +endif() + +file(GLOB GGML_HEADERS_SYCL "*.hpp") +file(GLOB GGML_SOURCES_SYCL "*.cpp") +target_sources(ggml-sycl PRIVATE ${GGML_HEADERS_SYCL} ${GGML_SOURCES_SYCL}) + +find_package(DNNL) +message("-- DNNL found:" ${DNNL_FOUND}) + +if (GGML_SYCL_TARGET STREQUAL "INTEL") + add_compile_definitions(GGML_SYCL_DNNL=${DNNL_FOUND}) +else() + add_compile_definitions(GGML_SYCL_DNNL=0) +endif() + +if (${DNNL_FOUND} AND GGML_SYCL_TARGET STREQUAL "INTEL") + target_link_libraries(ggml-sycl PRIVATE DNNL::dnnl) +endif() + +if (WIN32) + find_package(IntelSYCL REQUIRED) + find_package(MKL REQUIRED) + target_link_libraries(ggml-sycl PRIVATE IntelSYCL::SYCL_CXX MKL::MKL MKL::MKL_SYCL) +else() + if (GGML_SYCL_TARGET STREQUAL "INTEL") + target_link_libraries(ggml-sycl PRIVATE sycl OpenCL mkl_core pthread m dl mkl_sycl_blas mkl_intel_ilp64 mkl_tbb_thread) + elseif (GGML_SYCL_TARGET STREQUAL "NVIDIA") + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsycl-targets=nvptx64-nvidia-cuda") + add_compile_definitions(GGML_SYCL_NVIDIA) + target_link_libraries(ggml-sycl PRIVATE sycl pthread m dl onemkl_blas_cublas) + elseif (GGML_SYCL_TARGET STREQUAL "AMD") + if (NOT GGML_SYCL_DEVICE_ARCH) + message(ERROR "Can't enable SYCL hip backend, GGML_SYCL_DEVICE_ARCH has not been set.") + endif() + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fsycl-targets=amdgcn-amd-amdhsa") + target_link_libraries(ggml-sycl PRIVATE sycl pthread m dl onemkl) + endif() + + if (GGML_SYCL_DEVICE_ARCH) + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Xsycl-target-backend --offload-arch=${GGML_SYCL_DEVICE_ARCH}") + endif() +endif() diff --git a/ggml/src/ggml-sycl/backend.hpp b/ggml/src/ggml-sycl/backend.hpp index 85748a5b4..b1df4e5db 100644 --- a/ggml/src/ggml-sycl/backend.hpp +++ b/ggml/src/ggml-sycl/backend.hpp @@ -29,5 +29,6 @@ #include "wkv6.hpp" #include "outprod.hpp" #include "element_wise.hpp" +#include "gla.hpp" #endif // GGML_SYCL_BACKEND_HPP diff --git a/ggml/src/ggml-sycl/common.cpp b/ggml/src/ggml-sycl/common.cpp index 97ab2003c..022e7b763 100644 --- a/ggml/src/ggml-sycl/common.cpp +++ b/ggml/src/ggml-sycl/common.cpp @@ -12,6 +12,9 @@ #include "common.hpp" +#include "ggml-backend-impl.h" +#include "ggml-impl.h" + int get_current_device_id() { return dpct::dev_mgr::instance().current_device_id(); } @@ -28,11 +31,7 @@ void* ggml_sycl_host_malloc(size_t size) try { if (err != 0) { // clear the error - fprintf( - stderr, - "WARNING: failed to allocate %.2f MB of pinned memory: %s\n", - size / 1024.0 / 1024.0, - "syclGetErrorString is not supported"); + GGML_LOG_ERROR("WARNING: failed to allocate %.2f MB of pinned memory: %s\n", size / 1024.0 / 1024.0, "syclGetErrorString is not supported"); return nullptr; } @@ -52,6 +51,10 @@ void ggml_sycl_host_free(void* ptr) try { std::exit(1); } +bool gpu_has_xmx(sycl::device &dev) { + return dev.has(sycl::aspect::ext_intel_matrix); +} + int64_t downsample_sycl_global_range(int64_t accumulate_block_num, int64_t block_size) { const int64_t max_range = std::numeric_limits::max(); int64_t sycl_down_blk_size = block_size; @@ -66,17 +69,11 @@ int64_t downsample_sycl_global_range(int64_t accumulate_block_num, int64_t block void ggml_sycl_op_flatten(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, ggml_tensor *dst, const ggml_sycl_op_flatten_t op) try { - const int64_t nrows0 = ggml_nrows(src0); const bool use_src1 = src1 != nullptr; - const int64_t nrows1 = use_src1 ? ggml_nrows(src1) : 1; - - GGML_ASSERT(!use_src1 || src1->backend != GGML_BACKEND_TYPE_GPU_SPLIT); - GGML_ASSERT( dst->backend != GGML_BACKEND_TYPE_GPU_SPLIT); - - ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra; - ggml_tensor_extra_gpu * src1_extra = use_src1 ? (ggml_tensor_extra_gpu *) src1->extra : nullptr; - ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra; + if(use_src1) + GGML_ASSERT(strcmp(src1->buffer->buft->iface.get_name(src1->buffer->buft), GGML_SYCL_NAME "_Split") != 0); + GGML_ASSERT(strcmp(dst->buffer->buft->iface.get_name(dst->buffer->buft), GGML_SYCL_NAME "_Split") != 0); // dd = data device float * src0_ddf = (float *) src0->data; diff --git a/ggml/src/ggml-sycl/common.hpp b/ggml/src/ggml-sycl/common.hpp index 4549fa5e9..e9500f3a1 100644 --- a/ggml/src/ggml-sycl/common.hpp +++ b/ggml/src/ggml-sycl/common.hpp @@ -26,7 +26,11 @@ #define GGML_COMMON_DECL_SYCL #define GGML_COMMON_IMPL_SYCL +/* suppress warning spam */ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wnested-anon-types" #include "ggml-common.h" +#pragma clang diagnostic pop void* ggml_sycl_host_malloc(size_t size); void ggml_sycl_host_free(void* ptr); @@ -626,6 +630,7 @@ struct bin_bcast_sycl { }); } } + GGML_UNUSED(ctx); } }; @@ -657,6 +662,7 @@ inline void ggml_sycl_op_bin_bcast(ggml_backend_sycl_context & ctx, const ggml_t } } +bool gpu_has_xmx(sycl::device &dev); void ggml_sycl_op_flatten(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, ggml_tensor *dst, diff --git a/ggml/src/ggml-sycl/concat.cpp b/ggml/src/ggml-sycl/concat.cpp index c90c452d8..d41cfd3a6 100644 --- a/ggml/src/ggml-sycl/concat.cpp +++ b/ggml/src/ggml-sycl/concat.cpp @@ -47,7 +47,7 @@ static void concat_f32_dim1(const float *x, const float *y, float *dst, // operation int offset_dst = nidx + item_ct1.get_group(1) * ne0 + item_ct1.get_group(0) * ne0 * item_ct1.get_group_range(1); - if (item_ct1.get_group(1) < ne01) { // src0 + if (item_ct1.get_group(1) < (size_t) ne01) { // src0 int offset_src = nidx + item_ct1.get_group(1) * ne0 + item_ct1.get_group(0) * ne0 * ne01; dst[offset_dst] = x[offset_src]; @@ -70,7 +70,7 @@ static void concat_f32_dim2(const float *x, const float *y, float *dst, // operation int offset_dst = nidx + item_ct1.get_group(1) * ne0 + item_ct1.get_group(0) * ne0 * item_ct1.get_group_range(1); - if (item_ct1.get_group(0) < ne02) { // src0 + if (item_ct1.get_group(0) < (size_t) ne02) { // src0 int offset_src = nidx + item_ct1.get_group(1) * ne0 + item_ct1.get_group(0) * ne0 * item_ct1.get_group_range(1); dst[offset_dst] = x[offset_src]; @@ -158,8 +158,9 @@ static void concat_f32_sycl_non_cont( }); } -void ggml_sycl_op_concat(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor *dst) { +void ggml_sycl_op_concat(ggml_backend_sycl_context & ctx, ggml_tensor *dst) { + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; queue_ptr stream = ctx.stream(); const int32_t dim = ((int32_t *)dst->op_params)[0]; diff --git a/ggml/src/ggml-sycl/concat.hpp b/ggml/src/ggml-sycl/concat.hpp index 5a04feaab..e5cb7314c 100644 --- a/ggml/src/ggml-sycl/concat.hpp +++ b/ggml/src/ggml-sycl/concat.hpp @@ -15,7 +15,6 @@ #include "common.hpp" -void ggml_sycl_op_concat(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor *dst); +void ggml_sycl_op_concat(ggml_backend_sycl_context & ctx, ggml_tensor *dst); #endif // GGML_SYCL_CONCAT_HPP diff --git a/ggml/src/ggml-sycl/conv.cpp b/ggml/src/ggml-sycl/conv.cpp index bc4ab1ddb..ddba601e1 100644 --- a/ggml/src/ggml-sycl/conv.cpp +++ b/ggml/src/ggml-sycl/conv.cpp @@ -71,8 +71,9 @@ static void conv_transpose_1d_f32_f32_sycl( }); } -void ggml_sycl_op_conv_transpose_1d(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor *dst) { +void ggml_sycl_op_conv_transpose_1d(ggml_backend_sycl_context & ctx, ggml_tensor *dst) { + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; const float * src0_d = (const float *)src0->data; const float * src1_d = (const float *)src1->data; diff --git a/ggml/src/ggml-sycl/conv.hpp b/ggml/src/ggml-sycl/conv.hpp index eb20730f9..f9e60dc75 100644 --- a/ggml/src/ggml-sycl/conv.hpp +++ b/ggml/src/ggml-sycl/conv.hpp @@ -15,7 +15,6 @@ #include "common.hpp" -void ggml_sycl_op_conv_transpose_1d(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor *dst); +void ggml_sycl_op_conv_transpose_1d(ggml_backend_sycl_context & ctx, ggml_tensor *dst); #endif // GGML_SYCL_CONV_HPP diff --git a/ggml/src/ggml-sycl/convert.cpp b/ggml/src/ggml-sycl/convert.cpp index 5fd15e6cd..05b01db2d 100644 --- a/ggml/src/ggml-sycl/convert.cpp +++ b/ggml/src/ggml-sycl/convert.cpp @@ -424,7 +424,7 @@ static void convert_unary(const void * __restrict__ vx, dst_t * __restrict__ y, const int64_t global_id = item_ct1.get_local_id(2) + work_group_size * item_ct1.get_group(2); // make each work-item deal with more elements since sycl global range can not exceed max int - const src_t * x = (src_t *) vx; + const src_t * x = (const src_t *) vx; for (int64_t i = global_id; i < k; i += work_group_size * item_ct1.get_group_range(2)) { y[i] = x[i]; } diff --git a/ggml/src/ggml-sycl/dmmv.cpp b/ggml/src/ggml-sycl/dmmv.cpp index 0c3dfaa37..0d097357c 100644 --- a/ggml/src/ggml-sycl/dmmv.cpp +++ b/ggml/src/ggml-sycl/dmmv.cpp @@ -1015,9 +1015,9 @@ void ggml_sycl_op_dequantize_mul_mat_vec( break; } - (void) src1; - (void) dst; - (void) src1_ddq_i; - (void) src1_ncols; - (void) src1_padded_row_size; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_ddq_i); + GGML_UNUSED(src1_ncols); + GGML_UNUSED(src1_padded_row_size); } diff --git a/ggml/src/ggml-sycl/dpct/helper.hpp b/ggml/src/ggml-sycl/dpct/helper.hpp index fe4a8f744..e167948e7 100644 --- a/ggml/src/ggml-sycl/dpct/helper.hpp +++ b/ggml/src/ggml-sycl/dpct/helper.hpp @@ -15,6 +15,7 @@ #include #include +#include #include #include @@ -1236,7 +1237,7 @@ namespace dpct std::map::iterator get_map_iterator(const void *ptr) { - auto it = m_map.upper_bound((byte_t *)ptr); + auto it = m_map.upper_bound(const_cast(reinterpret_cast(ptr))); if (it == m_map.end()) { // Not a virtual pointer. @@ -1688,9 +1689,14 @@ namespace dpct auto data_a = get_memory(a); auto data_b = get_memory(b); auto data_c = get_memory(c); - oneapi::mkl::blas::column_major::gemm( - q, a_trans, b_trans, m, n, k, alpha_value, data_a, lda, - data_b, ldb, beta_value, data_c, ldc); +#ifdef GGML_SYCL_NVIDIA + oneapi::mkl::blas::column_major::gemm(oneapi::mkl::backend_selector{ q }, + a_trans, b_trans, m, n, k, alpha_value, data_a, lda, data_b, ldb, + beta_value, data_c, ldc); +#else + oneapi::mkl::blas::column_major::gemm(q, a_trans, b_trans, m, n, k, alpha_value, data_a, lda, data_b, ldb, + beta_value, data_c, ldc); +#endif } template @@ -1753,14 +1759,22 @@ namespace dpct matrix_info->ld_info[2] = ldc; matrix_info->groupsize_info = batch_size; +#ifdef GGML_SYCL_NVIDIA sycl::event e = oneapi::mkl::blas::column_major::gemm_batch( - q, matrix_info->transpose_info, matrix_info->transpose_info + 1, - matrix_info->size_info, matrix_info->size_info + 1, - matrix_info->size_info + 2, matrix_info->value_info, - reinterpret_cast(a), matrix_info->ld_info, - reinterpret_cast(b), matrix_info->ld_info + 1, - matrix_info->value_info + 1, reinterpret_cast(c), + oneapi::mkl::backend_selector{ q }, matrix_info->transpose_info, + matrix_info->transpose_info + 1, matrix_info->size_info, matrix_info->size_info + 1, + matrix_info->size_info + 2, matrix_info->value_info, reinterpret_cast(a), + matrix_info->ld_info, reinterpret_cast(b), matrix_info->ld_info + 1, + matrix_info->value_info + 1, reinterpret_cast(c), matrix_info->ld_info + 2, 1, + &(matrix_info->groupsize_info)); +#else + sycl::event e = oneapi::mkl::blas::column_major::gemm_batch( + q, matrix_info->transpose_info, matrix_info->transpose_info + 1, matrix_info->size_info, + matrix_info->size_info + 1, matrix_info->size_info + 2, matrix_info->value_info, + reinterpret_cast(a), matrix_info->ld_info, reinterpret_cast(b), + matrix_info->ld_info + 1, matrix_info->value_info + 1, reinterpret_cast(c), matrix_info->ld_info + 2, 1, &(matrix_info->groupsize_info)); +#endif q.submit([&](sycl::handler &cgh) { @@ -1782,10 +1796,16 @@ namespace dpct auto data_a = get_memory(a); auto data_b = get_memory(b); auto data_c = get_memory(c); +#ifdef GGML_SYCL_NVIDIA oneapi::mkl::blas::column_major::gemm_batch( - q, a_trans, b_trans, m, n, k, alpha_value, data_a, lda, - stride_a, data_b, ldb, stride_b, beta_value, - data_c, ldc, stride_c, batch_size); + oneapi::mkl::backend_selector{ q }, a_trans, b_trans, m, n, k, + alpha_value, data_a, lda, stride_a, data_b, ldb, stride_b, beta_value, data_c, ldc, stride_c, + batch_size); +#else + oneapi::mkl::blas::column_major::gemm_batch(q, a_trans, b_trans, m, n, k, alpha_value, data_a, lda, + stride_a, data_b, ldb, stride_b, beta_value, data_c, ldc, + stride_c, batch_size); +#endif } } // namespace detail @@ -1830,31 +1850,10 @@ namespace dpct : id); } - template - sycl::vec extract_and_sign_or_zero_extend4(T val) - { - return sycl::vec(val) - .template as, int8_t, uint8_t>, 4>>() - .template convert(); - } - - template - using dot_product_acc_t = - std::conditional_t && std::is_unsigned_v, - uint32_t, int32_t>; - template inline auto dp4a(T1 a, T2 b, T3 c) { - dot_product_acc_t res = c; - auto va = extract_and_sign_or_zero_extend4(a); - auto vb = extract_and_sign_or_zero_extend4(b); - res += va[0] * vb[0]; - res += va[1] * vb[1]; - res += va[2] * vb[2]; - res += va[3] * vb[3]; - return res; + return syclcompat::dp4a(a, b, c); } struct sub_sat diff --git a/ggml/src/ggml-sycl/element_wise.cpp b/ggml/src/ggml-sycl/element_wise.cpp index e5cd736eb..4bcd74376 100644 --- a/ggml/src/ggml-sycl/element_wise.cpp +++ b/ggml/src/ggml-sycl/element_wise.cpp @@ -237,7 +237,7 @@ void upscale_f32(const float *x, float *dst, const int nb00, const int nb01, int i02 = i12 / sf2; int i03 = i13 / sf3; - dst[index] = *(float *)((char *)x + i03 * nb03 + i02 * nb02 + i01 * nb01 + i00 * nb00); + dst[index] = *(const float *)((const char *)x + i03 * nb03 + i02 * nb02 + i01 * nb01 + i00 * nb00); } void pad_f32(const float *x, float *dst, const int ne0, const int ne00, const int ne01, const int ne02, @@ -251,8 +251,7 @@ void pad_f32(const float *x, float *dst, const int ne0, const int ne00, const i // operation int offset_dst = nidx + item_ct1.get_group(1) * ne0 + item_ct1.get_group(0) * ne0 * item_ct1.get_group_range(1); - if (nidx < ne00 && item_ct1.get_group(1) < ne01 && - item_ct1.get_group(0) < ne02) { + if (nidx < ne00 && item_ct1.get_group(1) < (size_t) ne01 && item_ct1.get_group(0) < (size_t) ne02) { int offset_src = nidx + item_ct1.get_group(1) * ne00 + item_ct1.get_group(0) * ne00 * ne01; dst[offset_dst] = x[offset_src]; @@ -520,9 +519,10 @@ inline void ggml_sycl_op_silu(ggml_backend_sycl_context & ctx, const ggml_tensor silu_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_gelu(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -535,9 +535,10 @@ inline void ggml_sycl_op_gelu(ggml_backend_sycl_context & ctx, const ggml_tensor gelu_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_gelu_quick(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, ggml_tensor *dst, @@ -550,9 +551,10 @@ inline void ggml_sycl_op_gelu_quick(ggml_backend_sycl_context & ctx, const ggml_ gelu_quick_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_tanh(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -564,9 +566,10 @@ inline void ggml_sycl_op_tanh(ggml_backend_sycl_context & ctx, const ggml_tensor GGML_ASSERT( dst->type == GGML_TYPE_F32); tanh_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_relu(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -579,9 +582,10 @@ inline void ggml_sycl_op_relu(ggml_backend_sycl_context & ctx, const ggml_tensor relu_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_hardsigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -595,9 +599,10 @@ inline void ggml_sycl_op_hardsigmoid(ggml_backend_sycl_context & ctx, const ggml hardsigmoid_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_hardswish(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -610,9 +615,10 @@ inline void ggml_sycl_op_hardswish(ggml_backend_sycl_context & ctx, const ggml_t hardswish_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_exp(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -625,9 +631,10 @@ inline void ggml_sycl_op_exp(ggml_backend_sycl_context & ctx, const ggml_tensor exp_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_log(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -640,9 +647,10 @@ inline void ggml_sycl_op_log(ggml_backend_sycl_context & ctx, const ggml_tensor log_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -655,9 +663,10 @@ inline void ggml_sycl_op_sigmoid(ggml_backend_sycl_context & ctx, const ggml_ten sigmoid_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sqrt(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -670,9 +679,10 @@ inline void ggml_sycl_op_sqrt(ggml_backend_sycl_context & ctx, const ggml_tensor sqrt_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sin(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -685,9 +695,10 @@ inline void ggml_sycl_op_sin(ggml_backend_sycl_context & ctx, const ggml_tensor sin_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_cos(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -700,9 +711,10 @@ inline void ggml_sycl_op_cos(ggml_backend_sycl_context & ctx, const ggml_tensor cos_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_step(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -715,9 +727,10 @@ inline void ggml_sycl_op_step(ggml_backend_sycl_context & ctx, const ggml_tensor step_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_neg(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -730,9 +743,10 @@ inline void ggml_sycl_op_neg(ggml_backend_sycl_context & ctx, const ggml_tensor neg_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_leaky_relu(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -749,9 +763,10 @@ inline void ggml_sycl_op_leaky_relu(ggml_backend_sycl_context & ctx, const ggml_ leaky_relu_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), negative_slope, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sqr(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -764,9 +779,10 @@ inline void ggml_sycl_op_sqr(ggml_backend_sycl_context & ctx, const ggml_tensor sqr_f32_sycl(src0_dd, dst_dd, ggml_nelements(src0), main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_upscale(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -787,9 +803,10 @@ inline void ggml_sycl_op_upscale(ggml_backend_sycl_context & ctx, const ggml_ten dst->ne[0], dst->ne[1], dst->ne[2], dst->ne[3], sf0, sf1, sf2, sf3, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_pad(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -805,9 +822,10 @@ inline void ggml_sycl_op_pad(ggml_backend_sycl_context & ctx, const ggml_tensor src0->ne[0], src0->ne[1], src0->ne[2], dst->ne[0], dst->ne[1], dst->ne[2], main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_acc(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -827,7 +845,8 @@ inline void ggml_sycl_op_acc(ggml_backend_sycl_context & ctx, const ggml_tensor acc_f32_sycl(src0_dd, src1_dd, dst_dd, ggml_nelements(dst), src1->ne[0], src1->ne[1], src1->ne[2], nb1, nb2, offset, main_stream); - (void) dst; + GGML_UNUSED(dst); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_add(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -863,149 +882,149 @@ inline void ggml_sycl_op_div(ggml_backend_sycl_context & ctx, const ggml_tensor } -void ggml_sycl_sqrt(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_sqrt(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sqrt); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sqrt); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_sin(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_sin(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sin); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sin); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_cos(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_cos(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_cos); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_cos); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_acc(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_acc(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_acc); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_acc); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_gelu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_gelu(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_gelu); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_gelu); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_silu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_silu(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_silu); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_silu); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_gelu_quick(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_gelu_quick(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_gelu_quick); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_gelu_quick); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_tanh(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_tanh(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_tanh); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_tanh); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_relu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_relu(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_relu); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_relu); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_sigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_sigmoid(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sigmoid); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sigmoid); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_hardsigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_hardsigmoid(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_hardsigmoid); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_hardsigmoid); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_hardswish(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_hardswish(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_hardswish); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_hardswish); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_exp(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_exp(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_exp); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_exp); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_log(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_log(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_log); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_log); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_neg(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_neg(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_neg); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_neg); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_step(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_step(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_step); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_step); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_leaky_relu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_leaky_relu(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_leaky_relu); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_leaky_relu); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_sqr(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_sqr(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sqr); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sqr); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_upscale(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_upscale(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_upscale); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_upscale); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_pad(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_pad(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_pad); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_pad); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_add(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_add(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_add); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_add); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_sub(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_sub(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sub); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sub); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_mul(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_mul(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_mul); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_mul); GGML_SYCL_DEBUG("call %s done\n", __func__); } -void ggml_sycl_div(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +void ggml_sycl_div(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_div); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_div); GGML_SYCL_DEBUG("call %s done\n", __func__); } diff --git a/ggml/src/ggml-sycl/element_wise.hpp b/ggml/src/ggml-sycl/element_wise.hpp index 8152edf58..464432645 100644 --- a/ggml/src/ggml-sycl/element_wise.hpp +++ b/ggml/src/ggml-sycl/element_wise.hpp @@ -25,52 +25,52 @@ static __dpct_inline__ float op_div(const float a, const float b) { } -void ggml_sycl_sqrt(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_sqrt(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_sin(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_sin(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_cos(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_cos(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_acc(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_acc(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_gelu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_gelu(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_silu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_silu(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_gelu_quick(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_gelu_quick(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_tanh(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_tanh(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_relu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_relu(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_sigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_sigmoid(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_hardsigmoid(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_hardsigmoid(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_hardswish(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_hardswish(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_exp(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_exp(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_log(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_log(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_neg(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_neg(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_step(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_step(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_leaky_relu(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_leaky_relu(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_sqr(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_sqr(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_upscale(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_upscale(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_pad(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_pad(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_add(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_add(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_sub(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_sub(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_mul(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_mul(ggml_backend_sycl_context & ctx, ggml_tensor * dst); -void ggml_sycl_div(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); +void ggml_sycl_div(ggml_backend_sycl_context & ctx, ggml_tensor * dst); #endif // GGML_SYCL_ELEMENTWISE_HPP diff --git a/ggml/src/ggml-sycl/gemm.hpp b/ggml/src/ggml-sycl/gemm.hpp index 2ad9b36f4..3f0f34ad6 100644 --- a/ggml/src/ggml-sycl/gemm.hpp +++ b/ggml/src/ggml-sycl/gemm.hpp @@ -51,8 +51,8 @@ public: const auto a_in_md = dnnl::memory::desc(a_dims, at, a_trans ? tag::ba : tag::ab); const auto b_in_md = dnnl::memory::desc(b_dims, bt, b_trans ? tag::ba : tag::ab); const auto c_md = dnnl::memory::desc(c_dims, ct, tag::ab); - auto a_mem = dnnl::memory(a_in_md, eng, (void*)a); - auto b_mem = dnnl::memory(b_in_md, eng, (void*)b); + auto a_mem = dnnl::memory(a_in_md, eng, const_cast(a)); + auto b_mem = dnnl::memory(b_in_md, eng, const_cast(b)); auto matmul_pd = dnnl::matmul::primitive_desc(eng, a_in_md, b_in_md, c_md); auto c_mem = dnnl::memory(matmul_pd.dst_desc(), eng, c); @@ -79,8 +79,8 @@ public: const auto a_in_md = dnnl::memory::desc(a_dims, at, a_trans ? tag::ba : tag::ab); const auto b_in_md = dnnl::memory::desc(b_dims, bt, b_trans ? tag::ba : tag::ab); const auto c_md = dnnl::memory::desc(c_dims, ct, tag::ab); - auto a_mem = dnnl::memory(a_in_md, eng, (void*)a); - auto b_mem = dnnl::memory(b_in_md, eng, (void*)b); + auto a_mem = dnnl::memory(a_in_md, eng, const_cast(a)); + auto b_mem = dnnl::memory(b_in_md, eng, const_cast(b)); auto matmul_pd = dnnl::matmul::primitive_desc(eng, a_in_md, b_in_md, c_md); auto c_mem = dnnl::memory(matmul_pd.dst_desc(), eng, c); diff --git a/ggml/src/ggml-sycl.cpp b/ggml/src/ggml-sycl/ggml-sycl.cpp similarity index 92% rename from ggml/src/ggml-sycl.cpp rename to ggml/src/ggml-sycl/ggml-sycl.cpp index 255bc64c6..5272ca454 100644 --- a/ggml/src/ggml-sycl.cpp +++ b/ggml/src/ggml-sycl/ggml-sycl.cpp @@ -47,25 +47,19 @@ static ggml_sycl_device_info ggml_sycl_init() { info.device_count = dpct::dev_mgr::instance().device_count(); if (info.device_count == 0) { - fprintf(stderr, "%s: failed to initialize " GGML_SYCL_NAME ": %s\n", __func__); + GGML_LOG_ERROR("%s: failed to initialize: %s\n", GGML_SYCL_NAME, __func__); return info; } GGML_ASSERT(info.device_count <= GGML_SYCL_MAX_DEVICES); int64_t total_vram = 0; -#if defined(GGML_SYCL_FORCE_MMQ) - fprintf(stderr, "%s: GGML_SYCL_FORCE_MMQ: yes\n", __func__); -#else - fprintf(stderr, "%s: GGML_SYCL_FORCE_MMQ: no\n", __func__); -#endif -#if defined(SYCL_USE_XMX) - fprintf(stderr, "%s: SYCL_USE_XMX: yes\n", __func__); -#else - fprintf(stderr, "%s: SYCL_USE_XMX: no\n", __func__); -#endif - fprintf(stderr, "%s: found %d " GGML_SYCL_NAME " devices:\n", __func__, info.device_count); - +/* This is a bit misleading; reserved for later */ +// #if defined(SYCL_USE_XMX) +// GGML_LOG_INFO("%s: SYCL_USE_XMX: yes\n", __func__); +// #else +// GGML_LOG_INFO("%s: SYCL_USE_XMX: no\n", __func__); +// #endif for (int i = 0; i < info.device_count; ++i) { info.devices[i].vmm = 0; dpct::device_info prop; @@ -109,30 +103,40 @@ void print_device_detail(int id, sycl::device &device, std::string device_type) name = std::regex_replace(name, std::regex("\\(TM\\)"), ""); auto global_mem_size = prop.get_global_mem_size()/1000000; - - fprintf(stderr, "|%2d|%19s|%39s|%7s|%7d|%8d|%5d|%6luM|%21s|\n", id, device_type.c_str(), + std::string xmx = gpu_has_xmx(device) ? "yes" : "no"; + GGML_LOG_INFO("|%2d|%19s|%39s|%7s|%7d|%8d|%5d|%6luM|%21s|%14s|\n", id, device_type.c_str(), name.c_str(), version.c_str(), prop.get_max_compute_units(), prop.get_max_work_group_size(), prop.get_max_sub_group_size(), - global_mem_size, device.get_info().c_str()); + global_mem_size, device.get_info().c_str(), xmx.c_str()); } void ggml_backend_sycl_print_sycl_devices() { GGML_SYCL_DEBUG("[SYCL] call ggml_backend_sycl_print_sycl_devices\n"); int device_count = dpct::dev_mgr::instance().device_count(); std::map DeviceNums; - fprintf(stderr, "found %d SYCL devices:\n", device_count); - fprintf(stderr, "| | | | |Max | |Max |Global | |\n"); - fprintf(stderr, "| | | | |compute|Max work|sub |mem | |\n"); - fprintf(stderr, "|ID| Device Type| Name|Version|units |group |group|size | Driver version|\n"); - fprintf(stderr, "|--|-------------------|---------------------------------------|-------|-------|--------|-----|-------|---------------------|\n"); + GGML_LOG_INFO("Found %d SYCL devices:\n", device_count); + + GGML_LOG_INFO( + "| | | | " + " |Max | |Max |Global | | XMX |\n"); + GGML_LOG_INFO( + "| | | | " + " |compute|Max work|sub |mem | | or |\n"); + GGML_LOG_INFO( + "|ID| Device Type| " + "Name|Version|units |group |group|size | Driver version| Tensor Cores |\n"); + GGML_LOG_INFO( + "|--|-------------------|---------------------------------------|------" + "-|-------|--------|-----|-------|---------------------|--------------|\n"); + for (int id = 0; id < device_count; ++id) { - sycl::device device = dpct::dev_mgr::instance().get_device(id); - sycl::backend backend = device.get_backend(); - std::string backend_type = get_device_backend_and_type(device); - int type_id=DeviceNums[backend_type]++; - std::stringstream device_type; - device_type << "[" << backend_type << ":" << std::to_string(type_id) << "]"; - print_device_detail(id, device, device_type.str()); + sycl::device device = dpct::dev_mgr::instance().get_device(id); + std::string backend_type = get_device_backend_and_type(device); + int type_id = DeviceNums[backend_type]++; + std::stringstream device_type; + device_type << "[" << backend_type << ":" << std::to_string(type_id) + << "]"; + print_device_detail(id, device, device_type.str()); } } @@ -154,15 +158,18 @@ static void ggml_check_sycl() try { static bool initialized = false; if (!initialized) { - fprintf(stderr, "[SYCL] call ggml_check_sycl\n"); + GGML_SYCL_DEBUG("[SYCL] call ggml_check_sycl\n"); g_ggml_sycl_debug = get_sycl_env("GGML_SYCL_DEBUG", 0); - - fprintf(stderr, "%s: GGML_SYCL_DEBUG: %d\n", __func__, g_ggml_sycl_debug); - -#if defined(GGML_SYCL_F16) - fprintf(stderr, "%s: GGML_SYCL_F16: yes\n", __func__); + GGML_LOG_INFO("GGML_SYCL_DEBUG: %d\n", g_ggml_sycl_debug); +#if defined(GGML_SYCL_FORCE_MMQ) + GGML_LOG_INFO("GGML_SYCL_FORCE_MMQ: yes\n"); #else - fprintf(stderr, "%s: GGML_SYCL_F16: no\n", __func__); + GGML_LOG_INFO("GGML_SYCL_FORCE_MMQ: no\n"); +#endif +#if defined(GGML_SYCL_F16) + GGML_LOG_INFO("GGML_SYCL_F16: yes\n"); +#else + GGML_LOG_INFO("GGML_SYCL_F16: no\n"); #endif /* NOT REMOVE, keep it for next optimize for XMX. @@ -180,9 +187,10 @@ static void ggml_check_sycl() try { return; } GGML_ASSERT(g_all_sycl_device_count <= GGML_SYCL_MAX_DEVICES); - ggml_backend_sycl_print_sycl_devices(); + initialized = true; g_sycl_loaded = true; + ggml_backend_sycl_print_sycl_devices(); } } catch (sycl::exception const &exc) { @@ -205,7 +213,7 @@ inline void check_allow_gpu_index(const int device_index) { __func__, device_index, ggml_sycl_info().device_count - 1); - fprintf(stderr, "%s\n", error_buf); + GGML_LOG_ERROR("%s\n", error_buf); assert(false); } } @@ -278,10 +286,8 @@ ggml_backend_sycl_buffer_init_tensor(ggml_backend_buffer_t buffer, ggml_tensor *tensor) try { ggml_backend_sycl_buffer_context * ctx = (ggml_backend_sycl_buffer_context *)buffer->context; - if (tensor->view_src != NULL && tensor->view_offs == 0) { + if (tensor->view_src != NULL) { assert(tensor->view_src->buffer->buft == buffer->buft); - tensor->backend = tensor->view_src->backend; - tensor->extra = tensor->view_src->extra; return; } @@ -409,13 +415,11 @@ ggml_backend_sycl_buffer_cpy_tensor(ggml_backend_buffer_t buffer, return true; } return false; + GGML_UNUSED(buffer); +} catch (const sycl::exception & exc) { + std::cerr << exc.what() << "Exception caught at file:" << __FILE__ << ", line:" << __LINE__ << std::endl; + std::exit(1); } -catch (sycl::exception const &exc) { - std::cerr << exc.what() << "Exception caught at file:" << __FILE__ - << ", line:" << __LINE__ << std::endl; - std::exit(1); -} - static void ggml_backend_sycl_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) try { @@ -475,8 +479,8 @@ ggml_backend_sycl_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, SYCL_CHECK(CHECK_TRY_ERROR(dev_ptr = (void *)sycl::malloc_device( size, *stream))); if (!dev_ptr) { - fprintf(stderr, "%s: can't malloc %lu Bytes memory on device", __func__, size); - return nullptr; + GGML_LOG_ERROR("%s: can't allocate %lu Bytes of memory on device\n", __func__, size); + return nullptr; } ggml_backend_sycl_buffer_context * ctx = new ggml_backend_sycl_buffer_context(buft_ctx->device, dev_ptr, buft_ctx->stream); return ggml_backend_buffer_init(buft, ggml_backend_sycl_buffer_interface, ctx, size); @@ -531,7 +535,7 @@ ggml_backend_buffer_type_t ggml_backend_sycl_buffer_type(int device) { auto dev_count = ggml_backend_sycl_get_device_count(); if (device>=dev_count or device<0) { - printf("ggml_backend_sycl_buffer_type error: device_index:%d is out of range [0, %d], miss to call ggml_backend_sycl_set_single_device()\n", + GGML_LOG_ERROR("ggml_backend_sycl_buffer_type error: device_index:%d is out of range [0, %d], miss to call ggml_backend_sycl_set_single_device()\n", device, dev_count-1); GGML_ASSERT(devicedevice; if (device>=ggml_sycl_info().device_count or device<0) { - printf("ggml_backend_sycl_buffer_type error: device_index:%d is out of range [0, %d], miss to call ggml_backend_sycl_set_single_device()\n", + GGML_LOG_ERROR("ggml_backend_sycl_buffer_type error: device_index:%d is out of range [0, %d], miss to call ggml_backend_sycl_set_single_device()\n", device, ggml_sycl_info().device_count-1); GGML_ASSERT(devicetype, MATRIX_ROW_PADDING - ne0 % MATRIX_ROW_PADDING); } - // FIXME: do not crash if cudaMalloc fails + // FIXME: do not crash if SYCL Buffer alloc fails // currently, init_tensor cannot fail, it needs to be fixed in ggml-backend first ggml_sycl_set_device(i); const queue_ptr stream = ctx->streams[i]; @@ -752,7 +756,7 @@ ggml_backend_sycl_split_buffer_init_tensor(ggml_backend_buffer_t buffer, size, *stream))); if (!buf) { char err_buf[1024]; - snprintf(err_buf, 1023, "%s: can't malloc %lu Bytes memory on device", __func__, size); + snprintf(err_buf, 1023, "%s: can't allocate %lu Bytes of memory on device\n", __func__, size); throw std::runtime_error(err_buf); } // set padding to 0 to avoid possible NaN values @@ -780,7 +784,6 @@ ggml_backend_sycl_split_buffer_init_tensor(ggml_backend_buffer_t buffer, CHECK_TRY_ERROR(extra->events[i][is] = new sycl::event())); } } - tensor->backend = GGML_BACKEND_TYPE_GPU_SPLIT; tensor->extra = extra; } catch (sycl::exception const &exc) { @@ -1081,10 +1084,7 @@ struct ggml_sycl_pool_leg : public ggml_sycl_pool { ggml_sycl_buffer buffer_pool[MAX_SYCL_BUFFERS] = {}; size_t pool_size = 0; - explicit ggml_sycl_pool_leg(queue_ptr qptr_, int device_) : - qptr(qptr_), - device(device_) { - } + explicit ggml_sycl_pool_leg(queue_ptr qptr_, int device_) : device(device_), qptr(qptr_) {} ~ggml_sycl_pool_leg() { for (int i = 0; i < MAX_SYCL_BUFFERS; ++i) { @@ -1142,17 +1142,18 @@ struct ggml_sycl_pool_leg : public ggml_sycl_pool { CHECK_TRY_ERROR(ptr = (void *)sycl::malloc_device( look_ahead_size, *qptr))); if (!ptr) { - fprintf(stderr, "%s: can't malloc %lu Bytes memory on device", __func__, look_ahead_size); + GGML_LOG_ERROR("%s: can't allocate %lu Bytes of memory on device/GPU\n", __func__, look_ahead_size); return nullptr; } *actual_size = look_ahead_size; pool_size += look_ahead_size; - #ifdef DEBUG_SYCL_MALLOC - fprintf(stderr, "%s[%d]: %d buffers, max_size = %u MB, pool_size = %u MB, requested %u MB\n", __func__, id, nnz, +#ifdef DEBUG_SYCL_MALLOC + GGML_LOG_DEBUG("%s[%d]: %d buffers, max_size = %u MB, pool_size = %u MB, requested %u MB\n", __func__, id, nnz, (uint32_t)(max_size/1024/1024), (uint32_t)(g_sycl_pool_size[id]/1024/1024), (uint32_t)(size/1024/1024)); - #endif +#endif + // GGML_SYCL_DEBUG("ggml_sycl_pool_malloc_leg look_ahead_size=%lu, return %p\n", look_ahead_size, ptr); return ptr; } @@ -1166,7 +1167,7 @@ struct ggml_sycl_pool_leg : public ggml_sycl_pool { return; } } - fprintf(stderr, "WARNING: sycl buffer pool full, increase MAX_sycl_BUFFERS\n"); + GGML_LOG_WARN("WARNING: sycl buffer pool full, increase MAX_sycl_BUFFERS\n"); SYCL_CHECK(CHECK_TRY_ERROR(sycl::free(ptr, *qptr))); pool_size -= size; } @@ -1186,7 +1187,6 @@ std::unique_ptr ggml_backend_sycl_context::new_pool_for_device(q /// kernels typedef void (*cpy_kernel_t)(const char * cx, char * cdst); -typedef void (*ggml_sycl_func_t)(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst); typedef void (*ggml_sycl_op_mul_mat_t)( ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, ggml_tensor *dst, @@ -1226,7 +1226,7 @@ static void quantize_q8_1(const float * __restrict__ x, void * __restrict__ vy, zeros[i] = 0.f; qzeros[i] = 0; } - const TC xi = ix < kx ? *(TC *)&x[iy * kx + ix] : zeros; + const TC xi = ix < kx ? *(const TC *)&x[iy * kx + ix] : zeros; float sum = xi[0]; float amax = sycl::fabs(xi[0]); #pragma unroll @@ -1787,6 +1787,9 @@ static void pool2d_nchw_kernel( switch (op) { case GGML_OP_POOL_AVG: res = 0; break; case GGML_OP_POOL_MAX: res = -FLT_MAX; break; + default: + res = (To) sycl::nan(uint32_t(0)); + break; } for (int i = bh; i < eh; i += 1) { @@ -1805,6 +1808,9 @@ static void pool2d_nchw_kernel( switch (op) { case GGML_OP_POOL_AVG: res += (cur / (kh * kw)); break; case GGML_OP_POOL_MAX: res = sycl::max(res, (To)cur); break; + default: + res = (To) sycl::nan(uint32_t(0)); + break; } } } @@ -1843,7 +1849,8 @@ static void get_rows_sycl(ggml_backend_sycl_context & ctx, const ggml_tensor *sr s3, nb01, nb02, nb03, s10, s11, s12, item_ct1); }); - (void) dst; + GGML_UNUSED(dst); + GGML_UNUSED(ctx); } template @@ -1881,10 +1888,10 @@ static void get_rows_sycl_float(ggml_backend_sycl_context & ctx, const ggml_tens }); } - (void) dst; + GGML_UNUSED(dst); + GGML_UNUSED(ctx); } - static void quantize_row_q8_1_sycl(const float *x, void *vy, const int kx, const int ky, const int kx_padded, queue_ptr stream) { @@ -2336,12 +2343,22 @@ static dpct::err0 ggml_sycl_cpy_tensor_2d(void *dst, dpct::memcpy_direction kind; char * src_ptr; - if (src->backend == GGML_BACKEND_TYPE_CPU) { + if (ggml_backend_buffer_is_host(src->buffer)) { kind = dpct::host_to_device; + //GGML_SYCL_DEBUG("%s: Host buffer type src tensor\n", __func__); src_ptr = (char *) src->data; // GGML_SYCL_DEBUG("ggml_sycl_cpy_tensor_2d GGML_BACKEND_TYPE_CPU src_ptr %p\n", src_ptr); - } else if (src->backend == GGML_BACKEND_TYPE_GPU || src->backend == GGML_BACKEND_TYPE_GPU_SPLIT) { - GGML_ASSERT(src->backend != GGML_BACKEND_TYPE_GPU_SPLIT || (i1_low == 0 && i1_high == src->ne[1])); + } else if (ggml_backend_buffer_is_sycl(src->buffer)) { + // If buffer is a SYCL buffer + //GGML_SYCL_DEBUG("%s: SYCL buffer type src tensor\n", __func__); + kind = dpct::device_to_device; + src_ptr = (char *) src->data; + } else if (ggml_backend_buffer_is_sycl_split(src->buffer)) { + /* + If buffer is a SYCL split buffer + */ + //GGML_SYCL_DEBUG("%s: Split buffer type src tensor\n", __func__); + GGML_ASSERT(i1_low == 0 && i1_high == src->ne[1]); kind = dpct::device_to_device; ggml_tensor_extra_gpu * extra = (ggml_tensor_extra_gpu *) src->extra; int id; @@ -2437,7 +2454,7 @@ static void ggml_sycl_op_get_rows(ggml_backend_sycl_context & ctx, const ggml_te break; default: // TODO: k-quants - fprintf(stderr, "%s: unsupported type: %s\n", __func__, ggml_type_name(src0->type)); + GGML_LOG_ERROR("%s: unsupported type: %s\n", __func__, ggml_type_name(src0->type)); GGML_ABORT("fatal error"); break; } @@ -2452,8 +2469,8 @@ static void ggml_sycl_op_repeat(ggml_backend_sycl_context & ctx, const ggml_tens ggml_sycl_op_bin_bcast>(ctx, dst, src0, dst, nullptr, src0_d, dst_d, main_stream); - (void) src1; - (void) src1_d; + GGML_UNUSED(src1); + GGML_UNUSED(src1_d); } @@ -2472,17 +2489,18 @@ inline void ggml_sycl_op_mul_mat_sycl( const int64_t ne00 = src0->ne[0]; const int64_t ne10 = src1->ne[0]; - const int64_t ne0 = dst->ne[0]; const int64_t row_diff = row_high - row_low; int id; SYCL_CHECK( CHECK_TRY_ERROR(id = get_current_device_id())); - +#if !GGML_SYCL_DNNL + const int64_t ne0 = dst->ne[0]; // the main device has a larger memory buffer to hold the results from all GPUs // ldc == nrows of the matrix that cuBLAS writes into int ldc = id == ctx.device ? ne0 : row_diff; +#endif #ifdef GGML_SYCL_F16 bool use_fp16 = true; // TODO(Yu) SYCL capability check @@ -2519,9 +2537,9 @@ inline void ggml_sycl_op_mul_mat_sycl( : src1_as_f16.get(); ggml_sycl_pool_alloc dst_f16(ctx.pool(), row_diff * src1_ncols); - const sycl::half alpha_f16 = 1.0f; - const sycl::half beta_f16 = 0.0f; #if !GGML_SYCL_DNNL + const sycl::half alpha_f16 = 1.0f; + const sycl::half beta_f16 = 0.0f; SYCL_CHECK(CHECK_TRY_ERROR(dpct::gemm( *stream, oneapi::mkl::transpose::trans, oneapi::mkl::transpose::nontrans, row_diff, src1_ncols, ne10, @@ -2558,24 +2576,29 @@ inline void ggml_sycl_op_mul_mat_sycl( const float * src0_ddf_i = src0->type == GGML_TYPE_F32 ? (const float *) src0_dd_i : src0_ddq_as_f32.get(); const float * src1_ddf1_i = src1->type == GGML_TYPE_F32 ? (const float *) src1_ddf_i : src1_ddq_as_f32.get(); - const float alpha = 1.0f; - const float beta = 0.0f; #if !GGML_SYCL_DNNL + const float alpha = 1.0f; + const float beta = 0.0f; +# ifdef GGML_SYCL_NVIDIA SYCL_CHECK(CHECK_TRY_ERROR(oneapi::mkl::blas::column_major::gemm( - *stream, oneapi::mkl::transpose::trans, - oneapi::mkl::transpose::nontrans, row_diff, src1_ncols, ne10, - dpct::get_value(&alpha, *stream), src0_ddf_i, ne00, - src1_ddf1_i, ne10, dpct::get_value(&beta, *stream), + oneapi::mkl::backend_selector{ *stream }, oneapi::mkl::transpose::trans, + oneapi::mkl::transpose::nontrans, row_diff, src1_ncols, ne10, dpct::get_value(&alpha, *stream), src0_ddf_i, + ne00, src1_ddf1_i, ne10, dpct::get_value(&beta, *stream), dst_dd_i, ldc))); +# else + SYCL_CHECK(CHECK_TRY_ERROR(oneapi::mkl::blas::column_major::gemm( + *stream, oneapi::mkl::transpose::trans, oneapi::mkl::transpose::nontrans, row_diff, src1_ncols, ne10, + dpct::get_value(&alpha, *stream), src0_ddf_i, ne00, src1_ddf1_i, ne10, dpct::get_value(&beta, *stream), dst_dd_i, ldc))); +# endif #else auto dnnl_stream = ctx.stream_dnnl(stream); DnnlGemmWrapper::row_gemm(dnnl_stream, false, true, src1_ncols, row_diff, ne10, src1_ddf1_i, DnnlGemmWrapper::to_dt(), src0_ddf_i, DnnlGemmWrapper::to_dt(), dst_dd_i, DnnlGemmWrapper::to_dt()); #endif } - (void) dst; - (void) src1_ddq_i; - (void) src1_padded_row_size; + GGML_UNUSED(dst); + GGML_UNUSED(src1_ddq_i); + GGML_UNUSED(src1_padded_row_size); } catch (sycl::exception const &exc) { std::cerr << exc.what() << "Exception caught at file:" << __FILE__ @@ -2621,8 +2644,9 @@ static void ggml_sycl_op_pool2d(ggml_backend_sycl_context & ctx, const ggml_tens item_ct1); }); - (void) src1; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sum(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -2637,9 +2661,10 @@ inline void ggml_sycl_op_sum(ggml_backend_sycl_context & ctx, const ggml_tensor sum_rows_f32_sycl(src0_dd, dst_dd, ne, 1, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_sum_rows(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -2656,9 +2681,10 @@ inline void ggml_sycl_op_sum_rows(ggml_backend_sycl_context & ctx, const ggml_te sum_rows_f32_sycl(src0_dd, dst_dd, ncols, nrows, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_argsort(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -2677,9 +2703,10 @@ inline void ggml_sycl_op_argsort(ggml_backend_sycl_context & ctx, const ggml_ten argsort_f32_i32_sycl(src0_dd, (int *)dst_dd, ncols, nrows, order, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_argmax(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -2696,9 +2723,10 @@ inline void ggml_sycl_op_argmax(ggml_backend_sycl_context & ctx, const ggml_tens argmax_f32_i32_sycl(src0_dd, (int *)dst_dd, ncols, nrows, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_diag_mask_inf(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, @@ -2718,9 +2746,10 @@ inline void ggml_sycl_op_diag_mask_inf(ggml_backend_sycl_context & ctx, const gg diag_mask_inf_f32_sycl(src0_dd, dst_dd, ne00, nrows0, ne01, n_past, main_stream); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_scale(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -2741,9 +2770,10 @@ inline void ggml_sycl_op_scale(ggml_backend_sycl_context & ctx, const ggml_tenso */ SYCL_CHECK(0); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } inline void ggml_sycl_op_clamp(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -2766,9 +2796,10 @@ inline void ggml_sycl_op_clamp(ggml_backend_sycl_context & ctx, const ggml_tenso */ SYCL_CHECK(0); - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } static void ggml_sycl_set_peer_access(const int n_tokens, int main_device) { @@ -2830,8 +2861,8 @@ static void ggml_sycl_op_mul_mat(ggml_backend_sycl_context & ctx, const ggml_ten const int nb2 = dst->nb[2]; const int nb3 = dst->nb[3]; - GGML_ASSERT(dst->backend != GGML_BACKEND_TYPE_GPU_SPLIT); - GGML_ASSERT(src1->backend != GGML_BACKEND_TYPE_GPU_SPLIT); + GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(dst->buffer)); + GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(src1->buffer)); GGML_ASSERT(src1->type == GGML_TYPE_F32 || (src1->ne[2] == 1 && src1->ne[3] == 1)); GGML_ASSERT(ne12 >= ne02 && ne12 % ne02 == 0); @@ -2845,14 +2876,13 @@ static void ggml_sycl_op_mul_mat(ggml_backend_sycl_context & ctx, const ggml_ten ggml_tensor_extra_gpu * src0_extra = (ggml_tensor_extra_gpu *) src0->extra; ggml_tensor_extra_gpu * src1_extra = (ggml_tensor_extra_gpu *) src1->extra; - ggml_tensor_extra_gpu * dst_extra = (ggml_tensor_extra_gpu *) dst->extra; const bool src0_is_contiguous = ggml_is_contiguous(src0); const bool src1_is_contiguous = ggml_is_contiguous(src1); int64_t src1_padded_col_size = GGML_PAD(ne10, MATRIX_ROW_PADDING); - const bool split = src0->backend == GGML_BACKEND_TYPE_GPU_SPLIT; + const bool split = ggml_backend_buffer_is_sycl_split(src0->buffer); GGML_ASSERT(!(split && ne02 > 1)); GGML_ASSERT(!(split && ne03 > 1)); GGML_ASSERT(!(split && ne02 < ne12)); @@ -3138,33 +3168,33 @@ catch (sycl::exception const &exc) { } -static void ggml_sycl_repeat(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_repeat(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_repeat); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_repeat); GGML_SYCL_DEBUG("call %s done\n", __func__); } -static void ggml_sycl_get_rows(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_get_rows(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_get_rows); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_get_rows); GGML_SYCL_DEBUG("call %s done\n", __func__); } -static void ggml_sycl_norm(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_norm(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_norm); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_norm); GGML_SYCL_DEBUG("call %s done\n", __func__); } -static void ggml_sycl_rms_norm(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_rms_norm(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_rms_norm); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_rms_norm); GGML_SYCL_DEBUG("call %s done\n", __func__); } -static void ggml_sycl_group_norm(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_group_norm(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { GGML_SYCL_DEBUG("call %s\n", __func__); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_group_norm); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_group_norm); GGML_SYCL_DEBUG("call %s done\n", __func__); } @@ -3172,7 +3202,7 @@ static void ggml_sycl_mul_mat_vec_p021(ggml_backend_sycl_context & ctx, const gg const ggml_tensor *src1, ggml_tensor *dst) try { GGML_ASSERT(ggml_is_permuted(src0) && ggml_is_permuted(src1)); - GGML_ASSERT(src0->backend != GGML_BACKEND_TYPE_GPU_SPLIT); + GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(src0->buffer)); GGML_ASSERT(src0->nb[0] <= src0->nb[1] && src0->nb[2] <= src0->nb[3]); // 0213 permutation GGML_ASSERT(src1->nb[0] <= src1->nb[1] && src1->nb[2] <= src1->nb[3]); // 0213 permutation GGML_ASSERT(src0->type == GGML_TYPE_F16); @@ -3205,7 +3235,7 @@ static void ggml_sycl_mul_mat_vec_nc(ggml_backend_sycl_context & ctx, const ggml GGML_ASSERT(!ggml_is_transposed(src0)); GGML_ASSERT(!ggml_is_transposed(src1)); GGML_ASSERT(!ggml_is_permuted(src0)); - GGML_ASSERT(src0->backend != GGML_BACKEND_TYPE_GPU_SPLIT); + GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(src0->buffer)); GGML_ASSERT(src0->type == GGML_TYPE_F16); GGML_ASSERT(src1->type == GGML_TYPE_F32); @@ -3267,12 +3297,11 @@ static void ggml_sycl_mul_mat_batched_sycl(ggml_backend_sycl_context & ctx, ggml_tensor *dst) try { GGML_ASSERT(!ggml_is_transposed(src0)); GGML_ASSERT(!ggml_is_transposed(src1)); - GGML_ASSERT(src0->backend != GGML_BACKEND_TYPE_GPU_SPLIT); + GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(src0->buffer)); GGML_ASSERT(src0->type == GGML_TYPE_F16); GGML_TENSOR_BINARY_OP_LOCALS - const int64_t ne_dst = ggml_nelements(dst); SYCL_CHECK(ggml_sycl_set_device(ctx.device)); queue_ptr main_stream = ctx.stream();; @@ -3380,6 +3409,7 @@ catch (sycl::exception const &exc) { inline bool ggml_sycl_supports_mmq(enum ggml_type type) { // TODO: accuracy issues in MMQ + GGML_UNUSED(type); return false; } @@ -3447,8 +3477,15 @@ static void ggml_sycl_mul_mat(ggml_backend_sycl_context & ctx, const ggml_tensor use_dequantize_mul_mat_vec = use_dequantize_mul_mat_vec && !use_mul_mat_vec_q; if (!split && src0->type == GGML_TYPE_F16 && ggml_is_permuted(src0) && ggml_is_permuted(src1) && src1->ne[1] == 1) { - // KQ single-batch - ggml_sycl_mul_mat_vec_p021(ctx, src0, src1, dst); + // TODO: Refactor and cleanup of mul mat dispatching. + if (src0->ne[3] == 1 && src1->ne[3] == 1) { + // KQ single-batch + // mmv p021 was specific for these dimensions + ggml_sycl_mul_mat_vec_p021(ctx, src0, src1, dst); + } else { + // The kernel from the if path is faster for that specific case, but does not support all mul mats. + ggml_sycl_mul_mat_batched_sycl(ctx, src0, src1, dst); + } } else if (!split && src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && src1->ne[1] == 1) { // KQV single-batch ggml_sycl_mul_mat_vec_nc(ctx, src0, src1, dst); @@ -3532,9 +3569,10 @@ __dpct_inline__ static void k_copy_dst_from_contiguous( } } -static void ggml_sycl_mul_mat_id(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, +static void ggml_sycl_mul_mat_id(ggml_backend_sycl_context & ctx, ggml_tensor *dst) try { + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; GGML_ASSERT(!ggml_backend_buffer_is_sycl_split(src0->buffer) && "mul_mat_id does not support split buffers"); const ggml_tensor *ids = dst->src[2]; @@ -3700,12 +3738,12 @@ catch (sycl::exception const &exc) { std::exit(1); } -static void ggml_sycl_scale(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_scale); +static void ggml_sycl_scale(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_scale); } -static void ggml_sycl_clamp(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_clamp); +static void ggml_sycl_clamp(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_clamp); } static void ggml_sycl_cpy(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, const ggml_tensor *src1, @@ -3743,12 +3781,11 @@ static void ggml_sycl_cpy(ggml_backend_sycl_context & ctx, const ggml_tensor *sr } else if (src0->type == GGML_TYPE_I32 && src1->type == GGML_TYPE_I32) { ggml_cpy_i32_i32_sycl (src0_ddc, src1_ddc, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, main_stream); } else { - fprintf(stderr, "%s: unsupported type combination (%s to %s)\n", __func__, + GGML_LOG_ERROR("%s: unsupported type combination (%s to %s)\n", __func__, ggml_type_name(src0->type), ggml_type_name(src1->type)); GGML_ABORT("fatal error"); } - - (void) dst; + GGML_UNUSED(dst); } catch (sycl::exception const &exc) { std::cerr << exc.what() << "Exception caught at file:" << __FILE__ @@ -3756,61 +3793,57 @@ catch (sycl::exception const &exc) { std::exit(1); } -static void ggml_sycl_dup(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { +static void ggml_sycl_dup(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { // TODO: why do we pass dst as src1 here? - ggml_sycl_cpy(ctx, src0, dst, nullptr); - (void) src1; + ggml_sycl_cpy(ctx, dst->src[0], dst, nullptr); } -static void ggml_sycl_diag_mask_inf(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_diag_mask_inf); +static void ggml_sycl_diag_mask_inf(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_diag_mask_inf); } -static void ggml_sycl_soft_max(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_soft_max); +static void ggml_sycl_soft_max(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_soft_max); } -static void ggml_sycl_rope(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_contiguous(src0)); // TODO: this restriction is temporary until non-cont support is implemented - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_rope); +static void ggml_sycl_rope(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + GGML_ASSERT(ggml_is_contiguous(dst->src[0])); // TODO: this restriction is temporary until non-cont support is implemented + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_rope); } -static void ggml_sycl_pool2d(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_pool2d); +static void ggml_sycl_pool2d(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_pool2d); } -static void ggml_sycl_im2col(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_im2col); +static void ggml_sycl_im2col(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_im2col); } -static void ggml_sycl_sum(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_contiguous(src0)); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sum); +static void ggml_sycl_sum(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + GGML_ASSERT(ggml_is_contiguous(dst->src[0])); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sum); } -static void ggml_sycl_sum_rows(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_contiguous(src0)); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_sum_rows); +static void ggml_sycl_sum_rows(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + GGML_ASSERT(ggml_is_contiguous(dst->src[0])); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_sum_rows); } -static void ggml_sycl_argsort(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_contiguous(src0)); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_argsort); +static void ggml_sycl_argsort(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + GGML_ASSERT(ggml_is_contiguous(dst->src[0])); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_argsort); } -static void ggml_sycl_argmax(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - GGML_ASSERT(ggml_is_contiguous(src0)); - ggml_sycl_op_flatten(ctx, src0, src1, dst, ggml_sycl_op_argmax); +static void ggml_sycl_argmax(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + GGML_ASSERT(ggml_is_contiguous(dst->src[0])); + ggml_sycl_op_flatten(ctx, dst->src[0], dst->src[1], dst, ggml_sycl_op_argmax); } -static void ggml_sycl_nop(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { - (void) src0; - (void) src1; - (void) dst; -} void ggml_sycl_set_main_device(const int main_device) try { - if (dpct::get_current_device_id() == main_device) return; + if (dpct::get_current_device_id() == static_cast (main_device)) { + return; + } check_allow_gpu_index(main_device); dpct::select_device(main_device); @@ -3818,7 +3851,7 @@ void ggml_sycl_set_main_device(const int main_device) try { dpct::device_info prop; SYCL_CHECK(CHECK_TRY_ERROR(dpct::get_device_info( prop, dpct::dev_mgr::instance().get_device(main_device)))); - fprintf(stderr, "Using device %d (%s) as main device\n", + GGML_LOG_INFO("Using device %d (%s) as main device\n", main_device, prop.get_name()); } } @@ -3828,191 +3861,192 @@ catch (sycl::exception const &exc) { std::exit(1); } -bool ggml_sycl_compute_forward(ggml_backend_sycl_context & ctx, struct ggml_tensor * tensor) { +bool ggml_sycl_compute_forward(ggml_backend_sycl_context & ctx, struct ggml_tensor * dst) { if (!g_sycl_loaded) return false; - ggml_sycl_func_t func; + if (dst->src[0] != nullptr && ggml_backend_buffer_is_sycl_split(dst->src[0]->buffer)) { + ggml_sycl_set_peer_access(dst->src[1]->ne[1], ctx.device); + } - switch (tensor->op) { + switch (dst->op) { case GGML_OP_ARGMAX: - func = ggml_sycl_argmax; + ggml_sycl_argmax(ctx, dst); break; case GGML_OP_CONV_TRANSPOSE_1D: - func = ggml_sycl_op_conv_transpose_1d; + ggml_sycl_op_conv_transpose_1d(ctx, dst); break; case GGML_OP_REPEAT: - func = ggml_sycl_repeat; + ggml_sycl_repeat(ctx, dst); break; case GGML_OP_GET_ROWS: - func = ggml_sycl_get_rows; + ggml_sycl_get_rows(ctx, dst); break; case GGML_OP_DUP: - func = ggml_sycl_dup; + ggml_sycl_dup(ctx, dst); break; case GGML_OP_ADD: case GGML_OP_ADD1: // TODO: more efficient implementation - func = ggml_sycl_add; + ggml_sycl_add(ctx, dst); break; case GGML_OP_SUB: - func = ggml_sycl_sub; + ggml_sycl_sub(ctx, dst); break; case GGML_OP_ACC: - func = ggml_sycl_acc; + ggml_sycl_acc(ctx, dst); break; case GGML_OP_MUL: - func = ggml_sycl_mul; + ggml_sycl_mul(ctx, dst); break; case GGML_OP_LOG: - func = ggml_sycl_log; + ggml_sycl_log(ctx, dst); break; case GGML_OP_DIV: - func = ggml_sycl_div; + ggml_sycl_div(ctx, dst); break; case GGML_OP_UNARY: - switch (ggml_get_unary_op(tensor)) { + switch (ggml_get_unary_op(dst)) { case GGML_UNARY_OP_NEG: - func = ggml_sycl_neg; + ggml_sycl_neg(ctx, dst); break; case GGML_UNARY_OP_STEP: - func = ggml_sycl_step; + ggml_sycl_step(ctx, dst); break; case GGML_UNARY_OP_GELU: - func = ggml_sycl_gelu; + ggml_sycl_gelu(ctx, dst); break; case GGML_UNARY_OP_SILU: - func = ggml_sycl_silu; + ggml_sycl_silu(ctx, dst); break; case GGML_UNARY_OP_GELU_QUICK: - func = ggml_sycl_gelu_quick; + ggml_sycl_gelu_quick(ctx, dst); break; case GGML_UNARY_OP_TANH: - func = ggml_sycl_tanh; + ggml_sycl_tanh(ctx, dst); break; case GGML_UNARY_OP_RELU: - func = ggml_sycl_relu; + ggml_sycl_relu(ctx, dst); break; case GGML_UNARY_OP_SIGMOID: - func = ggml_sycl_sigmoid; + ggml_sycl_sigmoid(ctx, dst); break; case GGML_UNARY_OP_HARDSIGMOID: - func = ggml_sycl_hardsigmoid; + ggml_sycl_hardsigmoid(ctx, dst); break; case GGML_UNARY_OP_HARDSWISH: - func = ggml_sycl_hardswish; + ggml_sycl_hardswish(ctx, dst); break; case GGML_UNARY_OP_EXP: - func = ggml_sycl_exp; + ggml_sycl_exp(ctx, dst); break; default: return false; } break; case GGML_OP_NORM: - func = ggml_sycl_norm; + ggml_sycl_norm(ctx, dst); break; case GGML_OP_GROUP_NORM: - func = ggml_sycl_group_norm; + ggml_sycl_group_norm(ctx, dst); break; case GGML_OP_CONCAT: - func = ggml_sycl_op_concat; + ggml_sycl_op_concat(ctx, dst); break; case GGML_OP_UPSCALE: - func = ggml_sycl_upscale; + ggml_sycl_upscale(ctx, dst); break; case GGML_OP_PAD: - func = ggml_sycl_pad; + ggml_sycl_pad(ctx, dst); break; case GGML_OP_LEAKY_RELU: - func = ggml_sycl_leaky_relu; + ggml_sycl_leaky_relu(ctx, dst); break; case GGML_OP_RMS_NORM: - func = ggml_sycl_rms_norm; + ggml_sycl_rms_norm(ctx, dst); break; case GGML_OP_MUL_MAT: - if (tensor->src[0]->ne[3] != tensor->src[1]->ne[3]) { + if (dst->src[0]->ne[3] != dst->src[1]->ne[3]) { return false; } - func = ggml_sycl_mul_mat; + /* ggml_sycl_mul_mat_id is dependent on ggml_sycl_mul_mat */ + ggml_sycl_mul_mat(ctx, dst->src[0], dst->src[1], dst); break; case GGML_OP_MUL_MAT_ID: - if (tensor->src[0]->ne[3] != tensor->src[1]->ne[3]) { + if (dst->src[0]->ne[3] != dst->src[1]->ne[3]) { return false; } - func = ggml_sycl_mul_mat_id; + ggml_sycl_mul_mat_id(ctx, dst); break; case GGML_OP_OUT_PROD: - func = ggml_sycl_op_out_prod; + ggml_sycl_op_out_prod(ctx, dst); break; case GGML_OP_SCALE: - func = ggml_sycl_scale; + ggml_sycl_scale(ctx, dst); break; case GGML_OP_SQR: - func = ggml_sycl_sqr; + ggml_sycl_sqr(ctx, dst); break; case GGML_OP_SQRT: - func = ggml_sycl_sqrt; + ggml_sycl_sqrt(ctx, dst); break; case GGML_OP_SIN: - func = ggml_sycl_sin; + ggml_sycl_sin(ctx, dst); break; case GGML_OP_COS: - func = ggml_sycl_cos; + ggml_sycl_cos(ctx, dst); break; case GGML_OP_CLAMP: - func = ggml_sycl_clamp; + ggml_sycl_clamp(ctx, dst); break; case GGML_OP_CPY: - func = ggml_sycl_cpy; + ggml_sycl_cpy(ctx, dst->src[0], dst->src[1], dst); break; case GGML_OP_CONT: - func = ggml_sycl_dup; + ggml_sycl_dup(ctx, dst); break; case GGML_OP_NONE: case GGML_OP_RESHAPE: case GGML_OP_VIEW: case GGML_OP_PERMUTE: case GGML_OP_TRANSPOSE: - func = ggml_sycl_nop; + GGML_SYCL_DEBUG("%s: Tensor NO-OP\n", __func__); break; case GGML_OP_DIAG_MASK_INF: - func = ggml_sycl_diag_mask_inf; + ggml_sycl_diag_mask_inf(ctx, dst); break; case GGML_OP_SOFT_MAX: - func = ggml_sycl_soft_max; + ggml_sycl_soft_max(ctx, dst); break; case GGML_OP_ROPE: - func = ggml_sycl_rope; + ggml_sycl_rope(ctx, dst); break; case GGML_OP_IM2COL: - func = ggml_sycl_im2col; + ggml_sycl_im2col(ctx, dst); break; case GGML_OP_POOL_2D: - func = ggml_sycl_pool2d; + ggml_sycl_pool2d(ctx, dst); break; case GGML_OP_SUM: - func = ggml_sycl_sum; + ggml_sycl_sum(ctx, dst); break; case GGML_OP_SUM_ROWS: - func = ggml_sycl_sum_rows; + ggml_sycl_sum_rows(ctx, dst); break; case GGML_OP_ARGSORT: - func = ggml_sycl_argsort; + ggml_sycl_argsort(ctx, dst); break; case GGML_OP_TIMESTEP_EMBEDDING: - func = ggml_sycl_op_timestep_embedding; + ggml_sycl_op_timestep_embedding(ctx, dst); break; case GGML_OP_RWKV_WKV6: - func = ggml_sycl_op_rwkv_wkv6; + ggml_sycl_op_rwkv_wkv6(ctx, dst); + break; + case GGML_OP_GATED_LINEAR_ATTN: + ggml_sycl_op_gated_linear_attn(ctx, dst); break; default: return false; } - if (tensor->src[0] != nullptr && ggml_backend_buffer_is_sycl_split(tensor->src[0]->buffer)) { - ggml_sycl_set_peer_access(tensor->src[1]->ne[1], ctx.device); - } - - func(ctx, tensor->src[0], tensor->src[1], tensor); return true; } @@ -4165,7 +4199,7 @@ static ggml_status ggml_backend_sycl_graph_compute(ggml_backend_t backend, ggml_ #endif bool ok = ggml_sycl_compute_forward(*sycl_ctx, node); if (!ok) { - fprintf(stderr, "%s: error: op not supported %s (%s)\n", __func__, node->name, ggml_op_name(node->op)); + GGML_LOG_ERROR("%s: error: op not supported %s (%s)\n", __func__, node->name, ggml_op_name(node->op)); } GGML_ASSERT(ok); } @@ -4178,6 +4212,7 @@ try { ggml_backend_sycl_context *sycl_ctx = (ggml_backend_sycl_context *)backend->context; + sycl::event *sycl_event = static_cast(event->context); const queue_ptr &stream = sycl_ctx->stream(sycl_ctx->device, 0); @@ -4192,7 +4227,7 @@ catch (sycl::exception const &exc) } static void ggml_backend_sycl_event_wait(ggml_backend_t backend, ggml_backend_event_t event) try { - ggml_backend_sycl_context* sycl_ctx = static_cast(backend->context); + sycl::event* sycl_event = static_cast(event->context); if (ggml_backend_is_sycl(backend)) { @@ -4451,7 +4486,16 @@ static bool ggml_backend_sycl_device_supports_op(ggml_backend_dev_t dev, const g case GGML_OP_SOFT_MAX: return true; case GGML_OP_ROPE: - return ggml_is_contiguous(op->src[0]); + { + const int mode = ((const int32_t *) op->op_params)[2]; + if (mode & GGML_ROPE_TYPE_MROPE) { + return false; + } + if (mode & GGML_ROPE_TYPE_VISION) { + return false; + } + return ggml_is_contiguous(op->src[0]); + } case GGML_OP_IM2COL: // TODO: add support for the new F32 operations return op->src[0]->type == GGML_TYPE_F16; @@ -4466,6 +4510,7 @@ static bool ggml_backend_sycl_device_supports_op(ggml_backend_dev_t dev, const g case GGML_OP_LEAKY_RELU: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_RWKV_WKV6: + case GGML_OP_GATED_LINEAR_ATTN: return true; default: return false; @@ -4486,7 +4531,7 @@ static bool ggml_backend_sycl_device_supports_buft(ggml_backend_dev_t dev, ggml_ static int64_t get_op_batch_size(const ggml_tensor * op) { switch (op->op) { case GGML_OP_GET_ROWS: - return op->ne[1]; // this will increse the speed of prefill in test + return 0; case GGML_OP_MUL_MAT: return op->ne[1]; case GGML_OP_MUL_MAT_ID: @@ -4592,21 +4637,21 @@ static ggml_backend_dev_t ggml_backend_sycl_reg_get_device(ggml_backend_reg_t re static void *ggml_backend_sycl_reg_get_proc_address(ggml_backend_reg_t reg, const char *name) { GGML_UNUSED(reg); - // TODO: update to the current function signature - //if (strcmp(name, "ggml_backend_split_buffer_type") == 0) { - // return (void *)ggml_backend_sycl_split_buffer_type; - //} + if (strcmp(name, "ggml_backend_split_buffer_type") == 0) { + return (void *)ggml_backend_sycl_split_buffer_type; + } // SYCL doesn't support registering host memory, left here for reference // "ggml_backend_register_host_buffer" // "ggml_backend_unregister_host_buffer" + GGML_UNUSED(name); return nullptr; } static const ggml_backend_reg_i ggml_backend_sycl_reg_interface = { /* .get_name = */ ggml_backend_sycl_reg_get_name, /* .get_device_count = */ ggml_backend_sycl_reg_get_device_count, - /* .get_device_get = */ ggml_backend_sycl_reg_get_device, + /* .get_device = */ ggml_backend_sycl_reg_get_device, /* .get_proc_address = */ ggml_backend_sycl_reg_get_proc_address, }; @@ -4637,16 +4682,17 @@ ggml_backend_reg_t ggml_backend_sycl_reg() { dev_ctx->description = prop.get_name(); ggml_backend_dev_t dev = new ggml_backend_device { - /* .interface = */ ggml_backend_sycl_device_interface, - /* .reg = */ ®, - /* .context = */ dev_ctx + /* .iface = */ ggml_backend_sycl_device_interface, + /* .reg = */ ®, + /* .context = */ dev_ctx }; ctx->devices.push_back(dev); } reg = ggml_backend_reg { - /* .interface = */ ggml_backend_sycl_reg_interface, - /* .context = */ ctx + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_sycl_reg_interface, + /* .context = */ ctx }; } @@ -4664,7 +4710,7 @@ ggml_backend_t ggml_backend_sycl_init(int device) { ggml_backend_sycl_context * ctx = new ggml_backend_sycl_context(device); if (ctx == nullptr) { - fprintf(stderr, "%s: error: failed to allocate context\n", __func__); + GGML_LOG_ERROR("%s: error: failed to allocate context\n", __func__); return nullptr; }; @@ -4678,3 +4724,4 @@ ggml_backend_t ggml_backend_sycl_init(int device) { return sycl_backend; } +GGML_BACKEND_DL_IMPL(ggml_backend_sycl_reg) diff --git a/ggml/src/ggml-sycl/gla.cpp b/ggml/src/ggml-sycl/gla.cpp new file mode 100644 index 000000000..eedb47486 --- /dev/null +++ b/ggml/src/ggml-sycl/gla.cpp @@ -0,0 +1,105 @@ +#include + +#include "common.hpp" + +template +static void gated_linear_attn_f32_kernel(const dpct::queue_ptr stream, u_int B, u_int T, u_int C, u_int H, float scale, + const float * k, const float * v, const float * r, const float * td, + const float * s, float * dst) { + const u_int head_size = HEAD_SIZE; + const u_int state_size = C * head_size; + const u_int n_seq_tokens = T / B; + sycl::range<1> block_dims((C / H)); + sycl::range<1> grid_dims((B * H)); + stream->submit([&](sycl::handler & cgh) { + /* local memory accessors*/ + auto _k = sycl::local_accessor(sycl::range<1>(head_size), cgh); + auto _r = sycl::local_accessor(sycl::range<1>(head_size), cgh); + auto _td = sycl::local_accessor(sycl::range<1>(head_size), cgh); + + cgh.parallel_for(sycl::nd_range<1>(grid_dims * block_dims, block_dims), [=](sycl::nd_item<1> item) { + u_int tid = item.get_local_id(0); + u_int bid = item.get_group(0); + + u_int batch_i = bid / H; + u_int head_i = bid % H; + + float state[head_size]; + +#pragma unroll + for (u_int i = 0; i < head_size; i++) { + state[i] = s[batch_i * state_size + head_i * head_size * head_size + i * head_size + tid]; + } + + for (u_int t = batch_i * n_seq_tokens * C + head_i * head_size + tid; + t < (batch_i + 1) * n_seq_tokens * C + head_i * head_size + tid; t += C) { + + item.barrier(sycl::access::fence_space::local_space); //sync threads + _k[tid] = k[t]; + _r[tid] = r[t]; + _td[tid] = td[t]; + item.barrier(sycl::access::fence_space::local_space); //sync threads + + const float _v = v[t]; + float y = 0; + + for (u_int j = 0; j < head_size; j += 4) { + const sycl::float4 & k = (sycl::float4 &) (_k[j]); + const sycl::float4 & r = (sycl::float4 &) (_r[j]); + const sycl::float4 & td = (sycl::float4 &) (_td[j]); + sycl::float4 & s = (sycl::float4 &) (state[j]); + sycl::float4 kv; + + kv.x() = k.x() * _v; + kv.y() = k.y() * _v; + kv.z() = k.z() * _v; + kv.w() = k.w() * _v; + + s.x() = s.x() * td.x() + kv.x(); + s.y() = s.y() * td.y() + kv.y(); + s.z() = s.z() * td.z() + kv.z(); + s.w() = s.w() * td.w() + kv.w(); + + y += r.x() * s.x(); + y += r.y() * s.y(); + y += r.z() * s.z(); + y += r.w() * s.w(); + } + dst[t] = y * scale; + } +#pragma unroll + for (u_int i = 0; i < head_size; i++) { + dst[T * C + batch_i * state_size + head_i * head_size * head_size + i * head_size + tid] = state[i]; + } + }); + }); +} + +void ggml_sycl_op_gated_linear_attn(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + const float * k_d = static_cast(dst->src[0]->data); + const float * v_d = static_cast(dst->src[1]->data); + const float * r_d = static_cast(dst->src[2]->data); + const float * td_d = static_cast(dst->src[3]->data); + const float * s_d = static_cast(dst->src[4]->data); + + const int64_t B = dst->src[4]->ne[1]; + const int64_t T = dst->src[0]->ne[2]; + const int64_t C = dst->ne[0]; + const int64_t H = dst->src[0]->ne[1]; + + dpct::queue_ptr stream = ctx.stream(); + GGML_ASSERT(dst->src[4]->type == GGML_TYPE_F32); + GGML_ASSERT(C % H == 0); + GGML_ASSERT(C / H == 64 || C / H == 128); + + float scale; + memcpy(&scale, dst->op_params, sizeof(float)); + + float * dst_d = (float *) dst->data; + + if (C / H == 64) { + gated_linear_attn_f32_kernel<64>(stream, B, T, C, H, scale, k_d, v_d, r_d, td_d, s_d, dst_d); + } else { + gated_linear_attn_f32_kernel<128>(stream, B, T, C, H, scale, k_d, v_d, r_d, td_d, s_d, dst_d); + } +} diff --git a/ggml/src/ggml-sycl/gla.hpp b/ggml/src/ggml-sycl/gla.hpp new file mode 100644 index 000000000..607cf3a7f --- /dev/null +++ b/ggml/src/ggml-sycl/gla.hpp @@ -0,0 +1,8 @@ +#ifndef GGML_SYCL_GLA_HPP +#define GGML_SYCL_GLA_HPP + +#include "common.hpp" + +void ggml_sycl_op_gated_linear_attn(ggml_backend_sycl_context & ctx, ggml_tensor * dst); + +#endif // GGML_SYCL_GLA_HPP diff --git a/ggml/src/ggml-sycl/im2col.cpp b/ggml/src/ggml-sycl/im2col.cpp index 6a0a0fcd0..6146a99ed 100644 --- a/ggml/src/ggml-sycl/im2col.cpp +++ b/ggml/src/ggml-sycl/im2col.cpp @@ -120,6 +120,7 @@ void ggml_sycl_op_im2col( im2col_sycl(src1_dd, (float *)dst_dd, IW, IH, OW, OH, KW, KH, IC, batch, batch_offset, delta_offset, s0, s1, p0, p1, d0, d1, main_stream); } - (void) src0; - (void) src0_dd; + GGML_UNUSED(src0); + GGML_UNUSED(src0_dd); + GGML_UNUSED(ctx); } diff --git a/ggml/src/ggml-sycl/mmq.cpp b/ggml/src/ggml-sycl/mmq.cpp index e952533d3..8ea82c940 100644 --- a/ggml/src/ggml-sycl/mmq.cpp +++ b/ggml/src/ggml-sycl/mmq.cpp @@ -813,7 +813,7 @@ load_tiles_q4_K(const void *__restrict__ vx, int *__restrict__ x_ql, x_ql[i * (WARP_SIZE + 1) + k] = get_int_from_uint8_aligned(bxi->qs, kqsx); } - const int blocks_per_tile_x_row = WARP_SIZE / QI4_K; // == 1 if QK_K == 256 + constexpr int blocks_per_tile_x_row = QI4_K > WARP_SIZE ? 1 : WARP_SIZE / QI4_K; // == 1 if QK_K == 256 const int kbxd = k % blocks_per_tile_x_row; // == 0 if QK_K == 256 #pragma unroll @@ -961,7 +961,7 @@ load_tiles_q5_K(const void *__restrict__ vx, int *__restrict__ x_ql, x_ql[i * (2*WARP_SIZE + 1) + kq1] = ql1 | qh1; } - const int blocks_per_tile_x_row = WARP_SIZE / QI5_K; // == 1 if QK_K == 256 + constexpr int blocks_per_tile_x_row = QI5_K > WARP_SIZE ? 1 : WARP_SIZE / QI5_K; // == 1 if QK_K == 256 const int kbxd = k % blocks_per_tile_x_row; // == 0 if QK_K == 256 #pragma unroll @@ -1109,7 +1109,7 @@ load_tiles_q6_K(const void *__restrict__ vx, int *__restrict__ x_ql, dpct::sub_sat()); } - const int blocks_per_tile_x_row = WARP_SIZE / QI6_K; // == 1 if QK_K == 256 + constexpr int blocks_per_tile_x_row = QI6_K > WARP_SIZE ? 1 : WARP_SIZE / QI6_K; // == 1 if QK_K == 256 const int kbxd = k % blocks_per_tile_x_row; // == 0 if QK_K == 256 float * x_dmf = (float *) x_dm; @@ -3020,9 +3020,9 @@ void ggml_sycl_op_mul_mat_q( break; } - (void) src1; - (void) dst; - (void) src1_ddf_i; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_ddf_i); } catch (sycl::exception const &exc) { std::cerr << exc.what() << "Exception caught at file:" << __FILE__ diff --git a/ggml/src/ggml-sycl/mmvq.cpp b/ggml/src/ggml-sycl/mmvq.cpp index 7b10cf688..221f65c21 100644 --- a/ggml/src/ggml-sycl/mmvq.cpp +++ b/ggml/src/ggml-sycl/mmvq.cpp @@ -753,11 +753,7 @@ static void mul_mat_vec_iq2_xs_q8_1_sycl(const void *vx, const void *vy, const sycl::range<3> block_nums(1, 1, block_num_y); const sycl::range<3> block_dims(1, GGML_SYCL_MMV_Y, QK_WARP_SIZE); { - - stream->submit([&](sycl::handler &cgh) { - auto iq2xs_grid_ptr_ct1 = &iq2xs_grid[0]; - auto ksigns64_ptr_ct1 = &ksigns64[0]; - + stream->submit([&](sycl::handler & cgh) { cgh.parallel_for( sycl::nd_range<3>(block_nums * block_dims, block_dims), [=](sycl::nd_item<3> item_ct1) @@ -780,9 +776,6 @@ static void mul_mat_vec_iq2_s_q8_1_sycl(const void *vx, const void *vy, { stream->submit([&](sycl::handler &cgh) { - auto iq2xs_grid_ptr_ct1 = &iq2xs_grid[0]; - auto ksigns64_ptr_ct1 = &ksigns64[0]; - cgh.parallel_for( sycl::nd_range<3>(block_nums * block_dims, block_dims), [=](sycl::nd_item<3> item_ct1) @@ -805,9 +798,6 @@ static void mul_mat_vec_iq3_xxs_q8_1_sycl(const void *vx, const void *vy, { stream->submit([&](sycl::handler &cgh) { - auto iq3xxs_grid_ptr_ct1 = &iq3xxs_grid[0]; - auto ksigns64_ptr_ct1 = &ksigns64[0]; - cgh.parallel_for( sycl::nd_range<3>(block_nums * block_dims, block_dims), [=](sycl::nd_item<3> item_ct1) @@ -830,8 +820,6 @@ static void mul_mat_vec_iq3_s_q8_1_sycl(const void *vx, const void *vy, { stream->submit([&](sycl::handler &cgh) { - auto iq3s_grid_ptr_ct1 = &iq3s_grid[0]; - cgh.parallel_for( sycl::nd_range<3>(block_nums * block_dims, block_dims), [=](sycl::nd_item<3> item_ct1) @@ -854,9 +842,6 @@ static void mul_mat_vec_iq1_s_q8_1_sycl(const void *vx, const void *vy, { stream->submit([&](sycl::handler &cgh) { - auto iq1s_grid_ptr_ct1 = &iq1s_grid_gpu[0]; - auto ksigns64_ptr_ct1 = &ksigns64[0]; - cgh.parallel_for( sycl::nd_range<3>(block_nums * block_dims, block_dims), [=](sycl::nd_item<3> item_ct1) @@ -954,7 +939,7 @@ void ggml_sycl_op_mul_mat_vec_q( const size_t q8_1_bs = QK8_1; // the main device has a larger memory buffer to hold the results from all GPUs // nrows_dst == nrows of the matrix that the kernel writes into - const int64_t nrows_dst = id == ctx.device ? ne00 : row_diff; + for (int i = 0; i < src1_ncols; i++) { const size_t src1_ddq_i_offset = i * src1_padded_col_size * q8_1_ts / q8_1_bs; @@ -1023,7 +1008,8 @@ void ggml_sycl_op_mul_mat_vec_q( break; } } - (void) src1; - (void) dst; - (void) src1_ddf_i; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_ddf_i); + GGML_UNUSED(ctx); } diff --git a/ggml/src/ggml-sycl/norm.cpp b/ggml/src/ggml-sycl/norm.cpp index b3159b9d1..9cf2be155 100644 --- a/ggml/src/ggml-sycl/norm.cpp +++ b/ggml/src/ggml-sycl/norm.cpp @@ -8,7 +8,6 @@ static void norm_f32(const float* x, float* dst, const int ncols, const float ep const int nthreads = item_ct1.get_local_range(2); const int nwarps = nthreads / WARP_SIZE; - assert(nwarps % WARP_SIZE == 0); sycl::float2 mean_var = sycl::float2(0.f, 0.f); for (int col = tid; col < ncols; col += block_size) { @@ -32,7 +31,7 @@ static void norm_f32(const float* x, float* dst, const int ncols, const float ep */ item_ct1.barrier(sycl::access::fence_space::local_space); mean_var = 0.f; - int nreduce = nwarps / WARP_SIZE; + size_t nreduce = nwarps / WARP_SIZE; for (size_t i = 0; i < nreduce; i += 1) { mean_var += s_sum[lane_id + i * WARP_SIZE]; @@ -55,9 +54,8 @@ static void group_norm_f32(const float* x, float* dst, const int group_size, con int end = start + group_size; const int nthreads = item_ct1.get_local_range(2); const int nwarps = nthreads / WARP_SIZE; - assert(nwarps % WARP_SIZE == 0); start += item_ct1.get_local_id(2); - int nreduce = nwarps / WARP_SIZE; + size_t nreduce = nwarps / WARP_SIZE; if (end >= ne_elements) { end = ne_elements; @@ -144,7 +142,6 @@ static void rms_norm_f32(const float* x, float* dst, const int ncols, const floa const int tid = item_ct1.get_local_id(2); const int nthreads = item_ct1.get_local_range(2); const int nwarps = nthreads / WARP_SIZE; - assert(nwarps % WARP_SIZE == 0); float tmp = 0.0f; // partial sum for thread in warp for (int col = tid; col < ncols; col += block_size) { @@ -166,7 +163,7 @@ static void rms_norm_f32(const float* x, float* dst, const int ncols, const floa converged control flow. You may need to adjust the code. */ item_ct1.barrier(sycl::access::fence_space::local_space); - int nreduce = nwarps / WARP_SIZE; + size_t nreduce = nwarps / WARP_SIZE; tmp = 0.f; for (size_t i = 0; i < nreduce; i += 1) { @@ -202,6 +199,7 @@ static void norm_f32_sycl(const float* x, float* dst, const int ncols, } else { const int work_group_size = ggml_sycl_info().max_work_group_sizes[device]; + assert(work_group_size % (WARP_SIZE * WARP_SIZE) == 0); const sycl::range<3> block_dims(1, 1, work_group_size); /* DPCT1049:17: The work-group size passed to the SYCL kernel may exceed @@ -244,6 +242,7 @@ static void group_norm_f32_sycl(const float* x, float* dst, } else { const int work_group_size = ggml_sycl_info().max_work_group_sizes[device]; + assert(work_group_size % (WARP_SIZE * WARP_SIZE) == 0); const sycl::range<3> block_dims(1, 1, work_group_size); /* DPCT1049:18: The work-group size passed to the SYCL kernel may exceed @@ -290,6 +289,7 @@ static void rms_norm_f32_sycl(const float* x, float* dst, const int ncols, } else { const int work_group_size = ggml_sycl_info().max_work_group_sizes[device]; + assert(work_group_size % (WARP_SIZE * WARP_SIZE) == 0); const sycl::range<3> block_dims(1, 1, work_group_size); /* DPCT1049:19: The work-group size passed to the SYCL kernel may exceed @@ -352,6 +352,7 @@ void ggml_sycl_op_group_norm(ggml_backend_sycl_context& ctx, const ggml_tensor* (void)src1; (void)dst; (void)src1_dd; + GGML_UNUSED(ctx); } void ggml_sycl_op_rms_norm(ggml_backend_sycl_context& ctx, const ggml_tensor* src0, diff --git a/ggml/src/ggml-sycl/outprod.cpp b/ggml/src/ggml-sycl/outprod.cpp index c2779df0e..8e8347ff4 100644 --- a/ggml/src/ggml-sycl/outprod.cpp +++ b/ggml/src/ggml-sycl/outprod.cpp @@ -1,10 +1,11 @@ #include +#include #include "outprod.hpp" -void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, const ggml_tensor* src0, - const ggml_tensor* src1, ggml_tensor* dst) { - +void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, ggml_tensor* dst) { + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; GGML_ASSERT(src0->type == GGML_TYPE_F32); GGML_ASSERT(src1->type == GGML_TYPE_F32); @@ -39,14 +40,14 @@ void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, const ggml_tensor* sr try { // Perform matrix multiplication using oneMKL GEMM - oneapi::mkl::blas::gemm(*stream, - oneapi::mkl::transpose::nontrans, src1_op, - ne0, ne1, ne01, - alpha, - src0_d, ne00, - src1_d, ldb, - beta, - dst_d, ne0); +#ifdef GGML_SYCL_NVIDIA + oneapi::mkl::blas::column_major::gemm(oneapi::mkl::backend_selector{ *stream }, + oneapi::mkl::transpose::nontrans, src1_op, ne0, ne1, ne01, alpha, src0_d, + ne00, src1_d, ldb, beta, dst_d, ne0); +#else + oneapi::mkl::blas::column_major::gemm(*stream, oneapi::mkl::transpose::nontrans, src1_op, ne0, ne1, ne01, alpha, + src0_d, ne00, src1_d, ldb, beta, dst_d, ne0); +#endif } catch (sycl::exception const& exc) { std::cerr << exc.what() << std::endl; diff --git a/ggml/src/ggml-sycl/outprod.hpp b/ggml/src/ggml-sycl/outprod.hpp index 9c042738a..f50413d3f 100644 --- a/ggml/src/ggml-sycl/outprod.hpp +++ b/ggml/src/ggml-sycl/outprod.hpp @@ -3,8 +3,7 @@ #include "common.hpp" -void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, const ggml_tensor* src0, - const ggml_tensor* src1, ggml_tensor* dst); +void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, ggml_tensor* dst); #endif // GGML_SYCL_OUTPROD_HPP diff --git a/ggml/src/ggml-sycl/rope.cpp b/ggml/src/ggml-sycl/rope.cpp index 1f06f78fa..1244b231a 100644 --- a/ggml/src/ggml-sycl/rope.cpp +++ b/ggml/src/ggml-sycl/rope.cpp @@ -269,7 +269,8 @@ void ggml_sycl_op_rope( } } - (void) src1; - (void) dst; - (void) src1_dd; + GGML_UNUSED(src1); + GGML_UNUSED(dst); + GGML_UNUSED(src1_dd); + GGML_UNUSED(ctx); } diff --git a/ggml/src/ggml-sycl/softmax.cpp b/ggml/src/ggml-sycl/softmax.cpp index 17a542e49..a9b3fce0d 100644 --- a/ggml/src/ggml-sycl/softmax.cpp +++ b/ggml/src/ggml-sycl/softmax.cpp @@ -16,7 +16,7 @@ static void soft_max_f32(const float * x, const float * mask, float * dst, const const int lane_id = item_ct1.get_local_id(2) % WARP_SIZE; const int nthreads = block_size; const int nwarps = nthreads / WARP_SIZE; - int nreduce = nwarps / WARP_SIZE; + size_t nreduce = nwarps / WARP_SIZE; float slope = 1.0f; // ALiBi @@ -53,8 +53,9 @@ static void soft_max_f32(const float * x, const float * mask, float * dst, const if (block_size > WARP_SIZE) { if (warp_id == 0) { buf[lane_id] = -INFINITY; - for (size_t i = 1; i < nreduce; i += 1) + for (size_t i = 1; i < nreduce; i += 1) { buf[lane_id + i * WARP_SIZE] = -INFINITY; + } } item_ct1.barrier(sycl::access::fence_space::local_space); @@ -63,8 +64,7 @@ static void soft_max_f32(const float * x, const float * mask, float * dst, const } item_ct1.barrier(sycl::access::fence_space::local_space); max_val = buf[lane_id]; - for (size_t i = 1; i < nreduce; i += 1) - { + for (size_t i = 1; i < nreduce; i += 1) { max_val = std::max(max_val, buf[lane_id + i * WARP_SIZE]); } max_val = warp_reduce_max(max_val, item_ct1); @@ -89,8 +89,9 @@ static void soft_max_f32(const float * x, const float * mask, float * dst, const item_ct1.barrier(sycl::access::fence_space::local_space); if (warp_id == 0) { buf[lane_id] = 0.f; - for (size_t i = 1; i < nreduce; i += 1) + for (size_t i = 1; i < nreduce; i += 1) { buf[lane_id + i * WARP_SIZE] = 0.f; + } } item_ct1.barrier(sycl::access::fence_space::local_space); @@ -100,8 +101,7 @@ static void soft_max_f32(const float * x, const float * mask, float * dst, const item_ct1.barrier(sycl::access::fence_space::local_space); tmp = buf[lane_id]; - for (size_t i = 1; i < nreduce; i += 1) - { + for (size_t i = 1; i < nreduce; i += 1) { tmp += buf[lane_id + i * WARP_SIZE]; } tmp = warp_reduce_sum(tmp, item_ct1); diff --git a/ggml/src/ggml-sycl/tsembd.cpp b/ggml/src/ggml-sycl/tsembd.cpp index d5c227cd1..b877d18c1 100644 --- a/ggml/src/ggml-sycl/tsembd.cpp +++ b/ggml/src/ggml-sycl/tsembd.cpp @@ -55,8 +55,9 @@ static void timestep_embedding_f32_sycl( }); } -void ggml_sycl_op_timestep_embedding(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor * dst) { +void ggml_sycl_op_timestep_embedding(ggml_backend_sycl_context & ctx, ggml_tensor * dst) { + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; const float * src0_d = (const float *)src0->data; float * dst_d = (float *)dst->data; dpct::queue_ptr stream = ctx.stream(); @@ -68,4 +69,5 @@ void ggml_sycl_op_timestep_embedding(ggml_backend_sycl_context & ctx, const ggml const int max_period = dst->op_params[1]; timestep_embedding_f32_sycl(src0_d, dst_d, src0->ne[0], dst->nb[1], dim, max_period, stream); + GGML_UNUSED(src1); } diff --git a/ggml/src/ggml-sycl/tsembd.hpp b/ggml/src/ggml-sycl/tsembd.hpp index ff854c337..4c18748bb 100644 --- a/ggml/src/ggml-sycl/tsembd.hpp +++ b/ggml/src/ggml-sycl/tsembd.hpp @@ -15,7 +15,6 @@ #include "common.hpp" -void ggml_sycl_op_timestep_embedding(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor * dst); +void ggml_sycl_op_timestep_embedding(ggml_backend_sycl_context & ctx, ggml_tensor * dst); #endif // GGML_SYCL_TSEMBD_HPP diff --git a/ggml/src/ggml-sycl/vecdotq.hpp b/ggml/src/ggml-sycl/vecdotq.hpp index d2dccade2..c5942008a 100644 --- a/ggml/src/ggml-sycl/vecdotq.hpp +++ b/ggml/src/ggml-sycl/vecdotq.hpp @@ -968,8 +968,8 @@ vec_dot_iq3_xxs_q8_1(const void *__restrict__ vbq, grid1[0] ^ signs[0], signs[0], std::minus<>()); const int grid_h = dpct::vectorized_binary( grid2[0] ^ signs[1], signs[1], std::minus<>()); - sumi = dpct::dp4a(grid_l, *((int *)q8 + 0), sumi); - sumi = dpct::dp4a(grid_h, *((int *)q8 + 1), sumi); + sumi = dpct::dp4a(grid_l, *((const int *)q8 + 0), sumi); + sumi = dpct::dp4a(grid_h, *((const int *)q8 + 1), sumi); q8 += 8; aux32 >>= 7; } @@ -1009,8 +1009,8 @@ vec_dot_iq3_s_q8_1(const void *__restrict__ vbq, grid1[0] ^ signs0, signs0, std::minus<>()); const int grid_h = dpct::vectorized_binary( grid2[0] ^ signs1, signs1, std::minus<>()); - sumi = dpct::dp4a(grid_l, *((int *)q8 + 0), sumi); - sumi = dpct::dp4a(grid_h, *((int *)q8 + 1), sumi); + sumi = dpct::dp4a(grid_l, *((const int *)q8 + 0), sumi); + sumi = dpct::dp4a(grid_h, *((const int *)q8 + 1), sumi); q8 += 8; } const float d = diff --git a/ggml/src/ggml-sycl/wkv6.cpp b/ggml/src/ggml-sycl/wkv6.cpp index 4c737f4bf..b54c20964 100644 --- a/ggml/src/ggml-sycl/wkv6.cpp +++ b/ggml/src/ggml-sycl/wkv6.cpp @@ -59,7 +59,7 @@ static void rwkv_wkv_f32_kernel( float y = 0; // Process in chunks of 4 for better vectorization - sycl::float4 k4, r4, tf4, td4, s4, kv4; + sycl::float4 k4, r4, tf4, td4, s4; #pragma unroll for (int j = 0; j < head_size; j += 4) { // Load data in vec4 chunks @@ -95,8 +95,10 @@ static void rwkv_wkv_f32_kernel( } } -void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context& ctx, const ggml_tensor* src0, - const ggml_tensor* src1, ggml_tensor* dst) { +void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context& ctx, ggml_tensor* dst) { + + const ggml_tensor *src0 = dst->src[0]; + const ggml_tensor *src1 = dst->src[1]; const float* k_d = (const float*)dst->src[0]->data; const float* v_d = (const float*)dst->src[1]->data; @@ -107,9 +109,9 @@ void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context& ctx, const ggml_tensor* s float* dst_d = (float*)dst->data; const int64_t B = dst->src[5]->ne[1]; - const int64_t T = dst->src[0]->ne[3]; + const int64_t T = dst->src[0]->ne[2]; const int64_t C = dst->ne[0]; - const int64_t H = dst->src[0]->ne[2]; + const int64_t H = dst->src[0]->ne[1]; GGML_ASSERT(dst->src[5]->type == GGML_TYPE_F32); GGML_ASSERT(C % H == 0); @@ -131,8 +133,11 @@ void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context& ctx, const ggml_tensor* s [=](sycl::nd_item<3> item_ct1) { rwkv_wkv_f32_kernel( B, T, C, H, k_d, v_d, r_d, tf_d, td_d, s_d, dst_d, - item_ct1, shared_mem_acc.get_pointer() + item_ct1, (float*)shared_mem_acc.get_multi_ptr().get() ); }); }); + + GGML_UNUSED(src0); + GGML_UNUSED(src1); } diff --git a/ggml/src/ggml-sycl/wkv6.hpp b/ggml/src/ggml-sycl/wkv6.hpp index ddfa3377b..8c596a997 100644 --- a/ggml/src/ggml-sycl/wkv6.hpp +++ b/ggml/src/ggml-sycl/wkv6.hpp @@ -3,8 +3,7 @@ #include "common.hpp" -void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context & ctx, const ggml_tensor *src0, - const ggml_tensor *src1, ggml_tensor * dst); +void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context & ctx, ggml_tensor * dst); #endif // GGML_SYCL_WKV6_HPP diff --git a/ggml/src/ggml-threading.cpp b/ggml/src/ggml-threading.cpp new file mode 100644 index 000000000..25a19eedb --- /dev/null +++ b/ggml/src/ggml-threading.cpp @@ -0,0 +1,12 @@ +#include "ggml-threading.h" +#include + +std::mutex ggml_critical_section_mutex; + +void ggml_critical_section_start() { + ggml_critical_section_mutex.lock(); +} + +void ggml_critical_section_end(void) { + ggml_critical_section_mutex.unlock(); +} diff --git a/ggml/src/ggml-threading.h b/ggml/src/ggml-threading.h new file mode 100644 index 000000000..dec2c8840 --- /dev/null +++ b/ggml/src/ggml-threading.h @@ -0,0 +1,14 @@ +#pragma once + +#include "ggml.h" + +#ifdef __cplusplus +extern "C" { +#endif + +GGML_API void ggml_critical_section_start(void); +GGML_API void ggml_critical_section_end(void); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-vulkan/CMakeLists.txt b/ggml/src/ggml-vulkan/CMakeLists.txt new file mode 100644 index 000000000..d970f7e20 --- /dev/null +++ b/ggml/src/ggml-vulkan/CMakeLists.txt @@ -0,0 +1,162 @@ +cmake_minimum_required(VERSION 3.19) +cmake_policy(SET CMP0114 NEW) + +find_package(Vulkan COMPONENTS glslc REQUIRED) + +function(detect_host_compiler) + if (CMAKE_HOST_SYSTEM_NAME STREQUAL "Windows") + find_program(HOST_C_COMPILER NAMES cl gcc clang NO_CMAKE_FIND_ROOT_PATH) + find_program(HOST_CXX_COMPILER NAMES cl g++ clang++ NO_CMAKE_FIND_ROOT_PATH) + else() + find_program(HOST_C_COMPILER NAMES gcc clang NO_CMAKE_FIND_ROOT_PATH) + find_program(HOST_CXX_COMPILER NAMES g++ clang++ NO_CMAKE_FIND_ROOT_PATH) + endif() + set(HOST_C_COMPILER "${HOST_C_COMPILER}" PARENT_SCOPE) + set(HOST_CXX_COMPILER "${HOST_CXX_COMPILER}" PARENT_SCOPE) +endfunction() + +if (Vulkan_FOUND) + message(STATUS "Vulkan found") + + ggml_add_backend_library(ggml-vulkan + ggml-vulkan.cpp + ../../include/ggml-vulkan.h + ) + + # Compile a test shader to determine whether GL_KHR_cooperative_matrix is supported. + # If it's not, there will be an error to stderr. + # If it's supported, set a define to indicate that we should compile those shaders + execute_process(COMMAND ${Vulkan_GLSLC_EXECUTABLE} -o - -fshader-stage=compute --target-env=vulkan1.3 "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_coopmat_support.comp" + OUTPUT_VARIABLE glslc_output + ERROR_VARIABLE glslc_error) + + if (${glslc_error} MATCHES ".*extension not supported: GL_KHR_cooperative_matrix.*") + message(STATUS "GL_KHR_cooperative_matrix not supported by glslc") + else() + message(STATUS "GL_KHR_cooperative_matrix supported by glslc") + add_compile_definitions(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) + endif() + + # Compile a test shader to determine whether GL_NV_cooperative_matrix2 is supported. + # If it's not, there will be an error to stderr. + # If it's supported, set a define to indicate that we should compile those shaders + execute_process(COMMAND ${Vulkan_GLSLC_EXECUTABLE} -o - -fshader-stage=compute --target-env=vulkan1.3 "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_coopmat2_support.comp" + OUTPUT_VARIABLE glslc_output + ERROR_VARIABLE glslc_error) + + if (${glslc_error} MATCHES ".*extension not supported: GL_NV_cooperative_matrix2.*") + message(STATUS "GL_NV_cooperative_matrix2 not supported by glslc") + else() + message(STATUS "GL_NV_cooperative_matrix2 supported by glslc") + add_compile_definitions(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + endif() + + target_link_libraries(ggml-vulkan PRIVATE Vulkan::Vulkan) + target_include_directories(ggml-vulkan PRIVATE ${CMAKE_CURRENT_BINARY_DIR}) + + # Workaround to the "can't dereference invalidated vector iterator" bug in clang-cl debug build + # Posssibly relevant: https://stackoverflow.com/questions/74748276/visual-studio-no-displays-the-correct-length-of-stdvector + if (MSVC AND CMAKE_CXX_COMPILER_ID STREQUAL "Clang") + add_compile_definitions(_ITERATOR_DEBUG_LEVEL=0) + endif() + + if (GGML_VULKAN_CHECK_RESULTS) + add_compile_definitions(GGML_VULKAN_CHECK_RESULTS) + endif() + + if (GGML_VULKAN_DEBUG) + add_compile_definitions(GGML_VULKAN_DEBUG) + endif() + + if (GGML_VULKAN_MEMORY_DEBUG) + add_compile_definitions(GGML_VULKAN_MEMORY_DEBUG) + endif() + + if (GGML_VULKAN_SHADER_DEBUG_INFO) + add_compile_definitions(GGML_VULKAN_SHADER_DEBUG_INFO) + endif() + + if (GGML_VULKAN_PERF) + add_compile_definitions(GGML_VULKAN_PERF) + endif() + + if (GGML_VULKAN_VALIDATE) + add_compile_definitions(GGML_VULKAN_VALIDATE) + endif() + + if (GGML_VULKAN_RUN_TESTS) + add_compile_definitions(GGML_VULKAN_RUN_TESTS) + endif() + + if (NOT CMAKE_CROSSCOMPILING) + add_subdirectory(vulkan-shaders) + if (MSVC) + foreach(CONFIG ${CMAKE_CONFIGURATION_TYPES}) + string(TOUPPER ${CONFIG} CONFIG) + set_target_properties(vulkan-shaders-gen PROPERTIES + RUNTIME_OUTPUT_DIRECTORY_${CONFIG} ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}) + endforeach() + endif() + else() + if (GGML_VULKAN_SHADERS_GEN_TOOLCHAIN) + set(HOST_CMAKE_TOOLCHAIN_FILE ${GGML_VULKAN_SHADERS_GEN_TOOLCHAIN}) + else() + detect_host_compiler() + if (NOT HOST_C_COMPILER OR NOT HOST_CXX_COMPILER) + message(FATAL_ERROR "Host compiler not found") + else() + message(STATUS "Host compiler: ${HOST_C_COMPILER} ${HOST_CXX_COMPILER}") + endif() + configure_file(${CMAKE_CURRENT_SOURCE_DIR}/cmake/host-toolchain.cmake.in ${CMAKE_BINARY_DIR}/host-toolchain.cmake @ONLY) + set(HOST_CMAKE_TOOLCHAIN_FILE ${CMAKE_BINARY_DIR}/host-toolchain.cmake) + endif() + message(STATUS "vulkan-shaders-gen toolchain file: ${HOST_CMAKE_TOOLCHAIN_FILE}") + + include(ExternalProject) + # Native build through ExternalProject_Add + ExternalProject_Add( + vulkan-shaders-gen + SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders + CMAKE_ARGS -DCMAKE_TOOLCHAIN_FILE=${HOST_CMAKE_TOOLCHAIN_FILE} + -DCMAKE_INSTALL_PREFIX=${CMAKE_BINARY_DIR} + BUILD_COMMAND ${CMAKE_COMMAND} --build . + INSTALL_COMMAND ${CMAKE_COMMAND} --install . + INSTALL_DIR ${CMAKE_BINARY_DIR} + ) + ExternalProject_Add_StepTargets(vulkan-shaders-gen build install) + endif() + set (_ggml_vk_host_suffix $,.exe,>) + set (_ggml_vk_genshaders_cmd ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/vulkan-shaders-gen${_ggml_vk_host_suffix}) + set (_ggml_vk_header ${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.hpp) + set (_ggml_vk_source ${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.cpp) + set (_ggml_vk_input_dir ${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders) + set (_ggml_vk_output_dir ${CMAKE_CURRENT_BINARY_DIR}/vulkan-shaders.spv) + + file(GLOB _ggml_vk_shader_deps "${_ggml_vk_input_dir}/*.comp") + set (_ggml_vk_shader_deps ${_ggml_vk_shader_deps} vulkan-shaders-gen) + + if (CMAKE_CROSSCOMPILING) + set(_ggml_vk_shader_deps ${_ggml_vk_shader_deps} vulkan-shaders-gen-build vulkan-shaders-gen-install) + endif() + + add_custom_command( + OUTPUT ${_ggml_vk_header} + ${_ggml_vk_source} + + COMMAND ${_ggml_vk_genshaders_cmd} + --glslc ${Vulkan_GLSLC_EXECUTABLE} + --input-dir ${_ggml_vk_input_dir} + --output-dir ${_ggml_vk_output_dir} + --target-hpp ${_ggml_vk_header} + --target-cpp ${_ggml_vk_source} + --no-clean + + DEPENDS ${_ggml_vk_shader_deps} + COMMENT "Generate vulkan shaders" + ) + + target_sources(ggml-vulkan PRIVATE ${_ggml_vk_source} ${_ggml_vk_header}) + +else() + message(WARNING "Vulkan not found") +endif() diff --git a/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in b/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in new file mode 100644 index 000000000..b6af747a5 --- /dev/null +++ b/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in @@ -0,0 +1,15 @@ +set(CMAKE_BUILD_TYPE Release) +set(CMAKE_C_FLAGS -O2) +set(CMAKE_CXX_FLAGS -O2) +set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER) +set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY NEVER) +set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE NEVER) +set(CMAKE_C_COMPILER @HOST_C_COMPILER@) +set(CMAKE_CXX_COMPILER @HOST_CXX_COMPILER@) +set(CMAKE_RUNTIME_OUTPUT_DIRECTORY @CMAKE_RUNTIME_OUTPUT_DIRECTORY@) + +if("@CMAKE_C_COMPILER_ID@" STREQUAL "MSVC") + foreach(CONFIG IN ITEMS DEBUG RELEASE MINSIZEREL RELWITHDEBINFO) + set(CMAKE_RUNTIME_OUTPUT_DIRECTORY_${CONFIG} ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}) + endforeach() +endif() diff --git a/ggml/src/ggml-vulkan.cpp b/ggml/src/ggml-vulkan/ggml-vulkan.cpp similarity index 70% rename from ggml/src/ggml-vulkan.cpp rename to ggml/src/ggml-vulkan/ggml-vulkan.cpp index a8e78c4db..649146d7b 100644 --- a/ggml/src/ggml-vulkan.cpp +++ b/ggml/src/ggml-vulkan/ggml-vulkan.cpp @@ -1,7 +1,8 @@ #include "ggml-vulkan.h" #include -#if defined(GGML_VULKAN_RUN_TESTS) || defined(GGML_VULKAN_PERF) +#if defined(GGML_VULKAN_RUN_TESTS) || defined(GGML_VULKAN_PERF) || defined(GGML_VULKAN_CHECK_RESULTS) #include +#include "ggml-cpu.h" #endif #include @@ -43,12 +44,6 @@ #define MAX_VK_BUFFERS 256 -#ifndef K_QUANTS_PER_ITERATION -#define K_QUANTS_PER_ITERATION 1 -#else -static_assert(K_QUANTS_PER_ITERATION == 1 || K_QUANTS_PER_ITERATION == 2, "K_QUANTS_PER_ITERATION must be 1 or 2"); -#endif - #define VK_CHECK(err, msg) \ do { \ vk::Result err_ = (err); \ @@ -106,6 +101,15 @@ struct vk_matmul_pipeline_struct { typedef std::shared_ptr vk_matmul_pipeline; +struct vk_matmul_pipeline2 { + vk_matmul_pipeline2() { + f16acc = std::make_shared(); + f32acc = std::make_shared(); + } + vk_matmul_pipeline f32acc; + vk_matmul_pipeline f16acc; +}; + struct vk_device_struct; typedef std::shared_ptr vk_device; typedef std::weak_ptr vk_device_ref; @@ -141,6 +145,8 @@ class vk_perf_logger; #endif static void ggml_vk_destroy_buffer(vk_buffer& buf); +static constexpr uint32_t mul_mat_vec_max_cols = 8; + struct vk_device_struct { std::mutex mutex; @@ -149,33 +155,57 @@ struct vk_device_struct { std::string name; uint64_t max_memory_allocation_size; bool fp16; + bool pipeline_robustness; vk::Device device; uint32_t vendor_id; vk_queue compute_queue; vk_queue transfer_queue; bool single_queue; uint32_t subgroup_size; + uint32_t shader_core_count; bool uma; + bool float_controls_rte_fp16; + + bool subgroup_size_control; + uint32_t subgroup_min_size; + uint32_t subgroup_max_size; + bool subgroup_require_full_support; + + bool coopmat_support; + bool coopmat_acc_f32_support; + bool coopmat_acc_f16_support; + uint32_t coopmat_m; + uint32_t coopmat_n; + uint32_t coopmat_k; + bool coopmat2; size_t idx; + bool mul_mat_l; + bool mul_mat_m; + bool mul_mat_s; + bool mul_mat_id_l; + bool mul_mat_id_m; + bool mul_mat_id_s; + vk_matmul_pipeline pipeline_matmul_f32; vk_matmul_pipeline pipeline_matmul_f32_f16; - vk_matmul_pipeline pipeline_matmul_f16; - vk_matmul_pipeline pipeline_matmul_f16_f32; + vk_matmul_pipeline2 pipeline_matmul_f16; + vk_matmul_pipeline2 pipeline_matmul_f16_f32; vk_pipeline pipeline_matmul_split_k_reduce; - vk_matmul_pipeline pipeline_dequant_mul_mat_mat[GGML_TYPE_COUNT]; + vk_matmul_pipeline2 pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_COUNT]; + vk_matmul_pipeline2 pipeline_dequant_mul_mat_mat[GGML_TYPE_COUNT]; vk_matmul_pipeline pipeline_matmul_id_f32; - vk_matmul_pipeline pipeline_matmul_id_f16; - vk_matmul_pipeline pipeline_matmul_id_f16_f32; + vk_matmul_pipeline2 pipeline_matmul_id_f16; + vk_matmul_pipeline2 pipeline_matmul_id_f16_f32; - vk_matmul_pipeline pipeline_dequant_mul_mat_mat_id[GGML_TYPE_COUNT]; + vk_matmul_pipeline2 pipeline_dequant_mul_mat_mat_id[GGML_TYPE_COUNT]; vk_pipeline pipeline_dequant[GGML_TYPE_COUNT]; - vk_pipeline pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_COUNT]; - vk_pipeline pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_COUNT]; + vk_pipeline pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_COUNT][mul_mat_vec_max_cols]; + vk_pipeline pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_COUNT][mul_mat_vec_max_cols]; vk_pipeline pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_COUNT]; vk_pipeline pipeline_mul_mat_vec_p021_f16_f32; @@ -183,9 +213,10 @@ struct vk_device_struct { vk_pipeline pipeline_get_rows[GGML_TYPE_COUNT]; vk_pipeline pipeline_get_rows_f32[GGML_TYPE_COUNT]; vk_pipeline pipeline_acc_f32; - vk_pipeline pipeline_add_f32, pipeline_add_f16_f32_f16; - vk_pipeline pipeline_mul_f32; - vk_pipeline pipeline_div_f32; + vk_pipeline pipeline_add_f32, pipeline_add_f32_norepeat; + vk_pipeline pipeline_add_f16_f32_f16, pipeline_add_f16_f32_f16_norepeat; + vk_pipeline pipeline_mul_f32, pipeline_mul_f32_norepeat; + vk_pipeline pipeline_div_f32, pipeline_div_f32_norepeat; vk_pipeline pipeline_concat_f32, pipeline_concat_f16, pipeline_concat_i32; vk_pipeline pipeline_upscale_f32; vk_pipeline pipeline_scale_f32; @@ -196,6 +227,7 @@ struct vk_device_struct { vk_pipeline pipeline_pad_f32; vk_pipeline pipeline_repeat_f32; vk_pipeline pipeline_cpy_f32_f32, pipeline_cpy_f32_f16, pipeline_cpy_f16_f16; + vk_pipeline pipeline_contig_cpy_f32_f32, pipeline_contig_cpy_f32_f16, pipeline_contig_cpy_f16_f16; vk_pipeline pipeline_norm_f32; vk_pipeline pipeline_group_norm_f32; vk_pipeline pipeline_rms_norm_f32; @@ -207,6 +239,7 @@ struct vk_device_struct { vk_pipeline pipeline_tanh_f32; vk_pipeline pipeline_diag_mask_inf_f32; vk_pipeline pipeline_soft_max_f32, pipeline_soft_max_f32_f16; + vk_pipeline pipeline_soft_max_f32_wg512, pipeline_soft_max_f32_f16_wg512; vk_pipeline pipeline_rope_norm_f32, pipeline_rope_norm_f16; vk_pipeline pipeline_rope_neox_f32, pipeline_rope_neox_f16; vk_pipeline pipeline_argsort_f32; @@ -214,6 +247,15 @@ struct vk_device_struct { vk_pipeline pipeline_im2col_f32, pipeline_im2col_f32_f16; vk_pipeline pipeline_timestep_embedding_f32; vk_pipeline pipeline_pool2d_f32; + vk_pipeline pipeline_rwkv_wkv6_f32; + + // [2][2][2] is for {f16acc,f32acc}x{large,small_rows}x{unaligned, aligned} + vk_pipeline pipeline_flash_attn_f32_f16_D64[GGML_TYPE_COUNT][2][2][2]; + vk_pipeline pipeline_flash_attn_f32_f16_D80[GGML_TYPE_COUNT][2][2][2]; + vk_pipeline pipeline_flash_attn_f32_f16_D96[GGML_TYPE_COUNT][2][2][2]; + vk_pipeline pipeline_flash_attn_f32_f16_D112[GGML_TYPE_COUNT][2][2][2]; + vk_pipeline pipeline_flash_attn_f32_f16_D128[GGML_TYPE_COUNT][2][2][2]; + vk_pipeline pipeline_flash_attn_f32_f16_D256[GGML_TYPE_COUNT][2][2][2]; std::unordered_map pipelines; std::unordered_map pipeline_descriptor_set_requirements; @@ -326,6 +368,40 @@ struct vk_mat_vec_id_push_constants { uint32_t nei0; uint32_t ne11; }; +struct vk_flash_attn_push_constants { + uint32_t N; + uint32_t KV; + + uint32_t ne1; + uint32_t ne2; + uint32_t ne3; + + uint32_t neq2; + uint32_t neq3; + uint32_t nek2; + uint32_t nek3; + uint32_t nev2; + uint32_t nev3; + uint32_t nem1; + + uint32_t nb02; + uint32_t nb03; + uint32_t nb12; + uint32_t nb13; + uint32_t nb22; + uint32_t nb23; + uint32_t nb31; + + float scale; + float max_bias; + float logit_softcap; + + uint32_t mask; + uint32_t n_head_log2; + float m0; + float m1; +}; + struct vk_op_push_constants { uint32_t KX; uint32_t KY; @@ -337,16 +413,55 @@ struct vk_op_unary_push_constants { uint32_t ne; uint32_t ne00; uint32_t ne01; uint32_t ne02; uint32_t ne03; uint32_t nb00; uint32_t nb01; uint32_t nb02; uint32_t nb03; uint32_t ne10; uint32_t ne11; uint32_t ne12; uint32_t ne13; uint32_t nb10; uint32_t nb11; uint32_t nb12; uint32_t nb13; - uint32_t d_offset; + uint32_t misalign_offsets; float param1; float param2; + uint32_t ne0_012mp; uint32_t ne0_012L; + uint32_t ne0_01mp; uint32_t ne0_01L; + uint32_t ne0_0mp; uint32_t ne0_0L; + uint32_t ne1_012mp; uint32_t ne1_012L; + uint32_t ne1_01mp; uint32_t ne1_01L; + uint32_t ne1_0mp; uint32_t ne1_0L; }; +static_assert(sizeof(vk_op_unary_push_constants) <= 128, "sizeof(vk_op_unary_push_constants) must be <= 128"); + +// See https://gmplib.org/~tege/divcnst-pldi94.pdf figure 4.1. +// Precompute mp (m' in the paper) and L such that division +// can be computed using a multiply (high 32b of 64b result) +// and a shift: +// +// n/d = (mulhi(n, mp) + n) >> L; +static void init_fastdiv_values(uint32_t d, uint32_t &mp, uint32_t &L) +{ + // compute L = ceil(log2(d)); + L = 0; + while (L < 32 && (uint32_t{1} << L) < d) { + L++; + } + + mp = (uint32_t)((uint64_t{1} << 32) * ((uint64_t{1} << L) - d) / d + 1); +} + +template void init_pushconst_fastdiv(T &p) { + GGML_UNUSED(p); + static_assert(!std::is_const::value, "unexpected type"); +} + +template <> void init_pushconst_fastdiv(vk_op_unary_push_constants &p) { + // Compute magic values to divide by these six numbers. + init_fastdiv_values(p.ne02*p.ne01*p.ne00, p.ne0_012mp, p.ne0_012L); + init_fastdiv_values(p.ne01*p.ne00, p.ne0_01mp, p.ne0_01L); + init_fastdiv_values(p.ne00, p.ne0_0mp, p.ne0_0L); + init_fastdiv_values(p.ne12*p.ne11*p.ne10, p.ne1_012mp, p.ne1_012L); + init_fastdiv_values(p.ne11*p.ne10, p.ne1_01mp, p.ne1_01L); + init_fastdiv_values(p.ne10, p.ne1_0mp, p.ne1_0L); +} struct vk_op_binary_push_constants { uint32_t ne; uint32_t ne00; uint32_t ne01; uint32_t ne02; uint32_t ne03; uint32_t nb00; uint32_t nb01; uint32_t nb02; uint32_t nb03; uint32_t ne10; uint32_t ne11; uint32_t ne12; uint32_t ne13; uint32_t nb10; uint32_t nb11; uint32_t nb12; uint32_t nb13; uint32_t ne20; uint32_t ne21; uint32_t ne22; uint32_t ne23; uint32_t nb20; uint32_t nb21; uint32_t nb22; uint32_t nb23; - uint32_t d_offset; + uint32_t misalign_offsets; float param1; float param2; int32_t param3; }; @@ -377,6 +492,7 @@ struct vk_op_soft_max_push_constants { float m0; float m1; uint32_t n_head_log2; + uint32_t nrows_x; }; struct vk_op_argsort_push_constants { @@ -415,6 +531,13 @@ struct vk_op_pool2d_push_constants { int32_t p0; int32_t p1; }; +struct vk_op_rwkv_wkv6_push_constants { + uint32_t B; + uint32_t T; + uint32_t C; + uint32_t H; +}; + // Allow pre-recording command buffers struct vk_staging_memcpy { vk_staging_memcpy(void * _dst, const void * _src, size_t _n) : dst(_dst), src(_src), n(_n) {} @@ -425,7 +548,7 @@ struct vk_staging_memcpy { }; struct vk_op_upscale_push_constants { - uint32_t ne; uint32_t d_offset; + uint32_t ne; uint32_t a_offset; uint32_t d_offset; uint32_t nb00; uint32_t nb01; uint32_t nb02; uint32_t nb03; uint32_t ne10; uint32_t ne11; uint32_t ne12; uint32_t ne13; float sf0; float sf1; float sf2; float sf3; @@ -641,8 +764,12 @@ static uint32_t compile_count = 0; static std::mutex compile_count_mutex; static std::condition_variable compile_count_cond; -static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipeline, const std::string name, size_t spv_size, const void* spv_data, const std::string entrypoint, uint32_t parameter_count, uint32_t push_constant_size, std::array wg_denoms, std::vector specialization_constants, uint32_t align) { - VK_LOG_DEBUG("ggml_vk_create_pipeline(" << device->name << ", " << name << ", " << entrypoint << ", " << parameter_count << ", " << push_constant_size << ", (" << wg_denoms[0] << "," << wg_denoms[1] << "," << wg_denoms[2] << "), specialization_constants, " << align << ")"); +static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipeline, const std::string name, size_t spv_size, const void* spv_data, const std::string entrypoint, + uint32_t parameter_count, uint32_t push_constant_size, std::array wg_denoms, std::vector specialization_constants, + uint32_t align, bool disable_robustness, bool require_full_subgroups, uint32_t required_subgroup_size) { + VK_LOG_DEBUG("ggml_vk_create_pipeline(" << device->name << ", " << name << ", " << entrypoint << ", " << parameter_count << ", " << push_constant_size << + ", (" << wg_denoms[0] << "," << wg_denoms[1] << "," << wg_denoms[2] << "), specialization_constants, " << align << + ", " << disable_robustness << ", " << require_full_subgroups << ", " << required_subgroup_size << ")"); GGML_ASSERT(parameter_count > 0); GGML_ASSERT(wg_denoms[0] > 0 && wg_denoms[1] > 0 && wg_denoms[2] > 0); // NOLINT @@ -701,16 +828,39 @@ static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipelin specialization_constants.data() ); + vk::PipelineShaderStageCreateFlags pipeline_shader_stage_create_flags{}; + + if (device->subgroup_require_full_support && require_full_subgroups) { + pipeline_shader_stage_create_flags |= vk::PipelineShaderStageCreateFlagBits::eRequireFullSubgroupsEXT; + } + vk::PipelineShaderStageCreateInfo pipeline_shader_create_info( - vk::PipelineShaderStageCreateFlags(), + pipeline_shader_stage_create_flags, vk::ShaderStageFlagBits::eCompute, pipeline->shader_module, entrypoint.c_str(), &specialization_info); + + vk::PipelineShaderStageRequiredSubgroupSizeCreateInfoEXT pipeline_shader_stage_required_subgroup_size_create_info; + pipeline_shader_stage_required_subgroup_size_create_info.requiredSubgroupSize = required_subgroup_size; + if (device->subgroup_size_control && required_subgroup_size > 0) { + GGML_ASSERT(device->subgroup_min_size <= required_subgroup_size && required_subgroup_size <= device->subgroup_max_size); + pipeline_shader_create_info.setPNext(&pipeline_shader_stage_required_subgroup_size_create_info); + } + vk::ComputePipelineCreateInfo compute_pipeline_create_info( - vk::PipelineCreateFlags(), + vk::PipelineCreateFlags{}, pipeline_shader_create_info, pipeline->layout); + + vk::PipelineRobustnessCreateInfoEXT rci; + + if (device->pipeline_robustness && disable_robustness) { + rci.storageBuffers = vk::PipelineRobustnessBufferBehaviorEXT::eDisabled; + rci.uniformBuffers = vk::PipelineRobustnessBufferBehaviorEXT::eDisabled; + compute_pipeline_create_info.setPNext(&rci); + } + pipeline->pipeline = device->device.createComputePipeline(VK_NULL_HANDLE, compute_pipeline_create_info).value; { @@ -722,6 +872,12 @@ static void ggml_vk_create_pipeline_func(vk_device& device, vk_pipeline& pipelin std::lock_guard guard(compile_count_mutex); assert(compile_count > 0); compile_count--; + + // "Progress bar" for shader compiles + static uint32_t total_compile_count = 0; + if ((total_compile_count++ % 10) == 0) { + std::cerr << "."; + } } compile_count_cond.notify_all(); } @@ -1197,59 +1353,186 @@ static void ggml_vk_wait_events(vk_context& ctx, std::vector&& events ); } +// number of rows/cols for flash attention shader +static constexpr uint32_t flash_attention_num_small_rows = 32; +static std::array fa_rows_cols(uint32_t D, uint32_t clamp, ggml_type type, bool small_rows) { + GGML_UNUSED(clamp); + + // small rows, large cols + if (small_rows) { + return {flash_attention_num_small_rows, 128}; + } + // small cols to reduce register count + if (ggml_is_quantized(type) || D == 256) { + return {64, 32}; + } + return {64, 64}; +}; + +static bool ggml_vk_matmul_shmem_support(const vk_device& device, const std::vector& warptile, bool mul_mat_id) { + // Needs to be kept up to date on shader changes + const uint32_t bank_conflict_offset = device->coopmat_support ? 8 : 1; + const uint32_t type_size = device->fp16 ? sizeof(ggml_fp16_t) : sizeof(float); + const uint32_t warps = warptile[0] / warptile[10]; + + const uint32_t load_bufs = (warptile[1] + warptile[2]) * (warptile[3] + bank_conflict_offset) * type_size; + const uint32_t mmid_row_ids = mul_mat_id ? 3072 * sizeof(uint32_t) : 0; + const uint32_t coopmat_stage = device->coopmat_support ? warptile[7] * warptile[8] / warps * sizeof(float) : 0; + + return (load_bufs + mmid_row_ids + coopmat_stage) <= device->properties.limits.maxComputeSharedMemorySize; +} + static void ggml_vk_load_shaders(vk_device& device) { VK_LOG_DEBUG("ggml_vk_load_shaders(" << device->name << ")"); + std::cerr << "ggml_vulkan: Compiling shaders"; + + // some shaders have a minimum subgroup size + const uint32_t subgroup_size_16 = std::max(device->subgroup_size, 16u); + const uint32_t subgroup_size_32 = std::max(device->subgroup_size, 32u); + // mulmat - std::initializer_list warptile_l = { 128, 128, 128, 16, device->subgroup_size * 2, 64, 2, 4, 4, device->subgroup_size }; - std::initializer_list warptile_m = { 128, 64, 64, 16, device->subgroup_size, 32, 2, 4, 2, device->subgroup_size }; - std::initializer_list warptile_s = { std::max(device->subgroup_size, 16u), 32, 32, 16, 32, 32, 2, 2, 2, device->subgroup_size }; + std::vector l_warptile, m_warptile, s_warptile, + l_warptile_mmq, m_warptile_mmq, s_warptile_mmq, + l_warptile_mmq_k, m_warptile_mmq_k, s_warptile_mmq_k, + l_warptile_mmqid, m_warptile_mmqid, s_warptile_mmqid; + std::array l_wg_denoms, m_wg_denoms, s_wg_denoms, + l_mmq_wg_denoms, m_mmq_wg_denoms, s_mmq_wg_denoms, + l_mmq_wg_denoms_k, m_mmq_wg_denoms_k, s_mmq_wg_denoms_k, + l_mmqid_wg_denoms, m_mmqid_wg_denoms, s_mmqid_wg_denoms; - std::initializer_list warptile_mmq_l = { 128, 128, 128, 32, device->subgroup_size * 2, 64, 2, 4, 4, device->subgroup_size }; - std::initializer_list warptile_mmq_m = { 128, 64, 64, 32, device->subgroup_size, 32, 2, 4, 2, device->subgroup_size }; - std::initializer_list warptile_mmq_s = { std::max(device->subgroup_size, 16u), 32, 32, 32, 32, 32, 2, 2, 2, device->subgroup_size }; + uint32_t l_align, m_align, s_align; + if (device->coopmat2) { + // spec constants and tile sizes for non-quant matmul/matmul_id + l_warptile = { 256, 128, 256, 64 }; + m_warptile = { 256, 128, 128, 64 }; + s_warptile = { 128, 64, 64, 64 }; + l_wg_denoms = {128, 256, 1 }; + m_wg_denoms = {128, 128, 1 }; + s_wg_denoms = { 64, 64, 1 }; - std::array l_wg_denoms = {128, 128, 1 }; - std::array m_wg_denoms = { 64, 64, 1 }; - std::array s_wg_denoms = { 32, 32, 1 }; + // spec constants and tile sizes for quant matmul (non-Qi_K) + l_warptile_mmq = { 256, 128, 256, 64 }; + m_warptile_mmq = { 256, 128, 128, 64 }; + s_warptile_mmq = { 256, 128, 128, 64 }; + l_mmq_wg_denoms = { 128, 256, 1 }; + m_mmq_wg_denoms = { 128, 128, 1 }; + s_mmq_wg_denoms = { 128, 128, 1 }; - uint32_t l_align = 128; - uint32_t m_align = 64; - uint32_t s_align = 32; + // spec constants and tile sizes for quant matmul (Qi_K) + l_warptile_mmq_k = { 256, 128, 512, 16 }; + m_warptile_mmq_k = { 256, 128, 256, 16 }; + s_warptile_mmq_k = { 256, 32, 128, 64 }; + l_mmq_wg_denoms_k = { 128, 512, 1 }; + m_mmq_wg_denoms_k = { 128, 256, 1 }; + s_mmq_wg_denoms_k = { 32, 128, 1 }; + + // spec constants and tile sizes for quant matmul_id + l_warptile_mmqid = { 256, 128, 128, 16 }; + m_warptile_mmqid = { 256, 128, 64, 16 }; + s_warptile_mmqid = { 256, 64, 64, 16 }; + l_mmqid_wg_denoms = { 128, 128, 1 }; + m_mmqid_wg_denoms = { 128, 64, 1 }; + s_mmqid_wg_denoms = { 64, 64, 1 }; + + l_align = 128; + m_align = 64; + s_align = 32; + } else { + // Matrix cores require different warp group sizes + const uint32_t tm_l = device->coopmat_support ? device->coopmat_m : 4; + const uint32_t tm_m = device->coopmat_support ? device->coopmat_m : 4; + const uint32_t tm_s = device->coopmat_support ? device->coopmat_m : 2; + const uint32_t tn_l = device->coopmat_support ? device->coopmat_n : 4; + const uint32_t tn_m = device->coopmat_support ? device->coopmat_n : 2; + const uint32_t tn_s = device->coopmat_support ? device->coopmat_n : 2; + const uint32_t tk_l = device->coopmat_support ? device->coopmat_k : 1; + const uint32_t tk_m = device->coopmat_support ? device->coopmat_k : 1; + const uint32_t tk_s = device->coopmat_support ? device->coopmat_k : 1; + + l_warptile = { 128, 128, 128, 16, device->subgroup_size * 2, 64, 2, tm_l, tn_l, tk_l, device->subgroup_size }; + m_warptile = { 128, 64, 64, 16, device->subgroup_size, 32, 2, tm_m, tn_m, tk_m, device->subgroup_size }; + s_warptile = { subgroup_size_16, 32, 32, 16, 32, 32, 2, tm_s, tn_s, tk_s, device->subgroup_size }; + + l_warptile_mmq = { 128, 128, 128, 32, device->subgroup_size * 2, 64, 2, tm_l, tn_l, tk_l, device->subgroup_size }; + m_warptile_mmq = { 128, 64, 64, 32, device->subgroup_size, 32, 2, tm_m, tn_m, tk_m, device->subgroup_size }; + s_warptile_mmq = { subgroup_size_32, 32, 32, 32, 32, 32, 2, tm_s, tn_s, tk_s, device->subgroup_size }; + + l_mmq_wg_denoms = l_wg_denoms = {128, 128, 1 }; + m_mmq_wg_denoms = m_wg_denoms = { 64, 64, 1 }; + s_mmq_wg_denoms = s_wg_denoms = { 32, 32, 1 }; + l_align = 128; + m_align = 64; + s_align = 32; + + // Fallback to smaller sizes if there's not enough shared memory. Given the current shaders + // and tile sizes, this should handle 16KB, 32KB, and 48KB+. + // This logic doesn't explicitly account for the 12KB row_ids in the mul_mat_mat_id shaders. + // But the numbers happen to work out for 32KB shared memory size that when using the medium + // size there's enough room for everything, and we assert for this. + uint32_t shmem_needed = (l_warptile[1] + l_warptile[2]) * (l_warptile[3] + 1) * sizeof(float); + if (shmem_needed > device->properties.limits.maxComputeSharedMemorySize) { + l_warptile = m_warptile; + l_wg_denoms = m_wg_denoms; + shmem_needed = (l_warptile[1] + l_warptile[2]) * (l_warptile[3] + 1) * sizeof(float); + GGML_ASSERT(shmem_needed <= device->properties.limits.maxComputeSharedMemorySize); + } + if (device->properties.limits.maxComputeSharedMemorySize >= 32768) { + // assert mul_mat_mat_id shaders will fit. + GGML_ASSERT(shmem_needed + 3072*4 <= device->properties.limits.maxComputeSharedMemorySize); + } + + shmem_needed = (l_warptile_mmq[1] + l_warptile_mmq[2]) * (l_warptile_mmq[3] + 1) * sizeof(float); + if (shmem_needed > device->properties.limits.maxComputeSharedMemorySize) { + if (device->properties.limits.maxComputeSharedMemorySize == 32768) { + l_warptile_mmq = m_warptile_mmq; + l_mmq_wg_denoms = m_mmq_wg_denoms; + } else { + l_warptile_mmq = s_warptile_mmq; + l_mmq_wg_denoms = s_mmq_wg_denoms; + } + shmem_needed = (l_warptile_mmq[1] + l_warptile_mmq[2]) * (l_warptile_mmq[3] + 1) * sizeof(float); + GGML_ASSERT(shmem_needed <= device->properties.limits.maxComputeSharedMemorySize); + } + if (device->properties.limits.maxComputeSharedMemorySize >= 32768) { + // assert mul_mat_mat_id shaders will fit. + GGML_ASSERT(shmem_needed + 3072*4 <= device->properties.limits.maxComputeSharedMemorySize); + } + // Disable medium and large matrix multiplication if not enough shared memory is available + // Check mmq warptiles as the largest configuration + // Throw an error if not enough for any matrix multiplication is available + if (!ggml_vk_matmul_shmem_support(device, s_warptile_mmq, false)) { + std::cerr << "ggml_vulkan: Error: Shared memory size too small for matrix multiplication." << std::endl; + throw std::runtime_error("Shared memory size too small for matrix multiplication."); + } else if (!ggml_vk_matmul_shmem_support(device, m_warptile_mmq, false)) { + device->mul_mat_m = false; + device->mul_mat_l = false; + } else if (!ggml_vk_matmul_shmem_support(device, l_warptile_mmq, false)) { + device->mul_mat_l = false; + } + + // Disable mul_mat_id if not enough shared memory is available + if (!ggml_vk_matmul_shmem_support(device, s_warptile_mmq, true)) { + device->mul_mat_id_s = false; + device->mul_mat_id_m = false; + device->mul_mat_id_l = false; + } else if (!ggml_vk_matmul_shmem_support(device, m_warptile_mmq, true)) { + device->mul_mat_id_m = false; + device->mul_mat_id_l = false; + } else if (!ggml_vk_matmul_shmem_support(device, l_warptile_mmq, true)) { + device->mul_mat_id_l = false; + } + } device->pipeline_matmul_f32 = std::make_shared(); device->pipeline_matmul_f32_f16 = std::make_shared(); - device->pipeline_matmul_f16_f32 = std::make_shared(); - device->pipeline_matmul_f16 = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL] = std::make_shared(); device->pipeline_matmul_id_f32 = std::make_shared(); - device->pipeline_matmul_id_f16_f32 = std::make_shared(); - device->pipeline_matmul_id_f16 = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K] = std::make_shared(); - device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL] = std::make_shared(); std::vector> compiles; - auto const &ggml_vk_create_pipeline = [&](vk_device& device, vk_pipeline& pipeline, const std::string &name, size_t spv_size, const void* spv_data, const std::string &entrypoint, uint32_t parameter_count, uint32_t push_constant_size, std::array wg_denoms, std::vector&& specialization_constants, uint32_t align) { + auto const &ggml_vk_create_pipeline = [&](vk_device& device, vk_pipeline& pipeline, const std::string &name, size_t spv_size, const void* spv_data, const std::string &entrypoint, + uint32_t parameter_count, uint32_t push_constant_size, std::array wg_denoms, const std::vector& specialization_constants, + uint32_t align, bool disable_robustness = false, bool require_full_subgroups = false, uint32_t required_subgroup_size = 0) { { // wait until fewer than N compiles are in progress uint32_t N = std::max(1u, std::thread::hardware_concurrency()); @@ -1259,459 +1542,378 @@ static void ggml_vk_load_shaders(vk_device& device) { } compile_count++; } - compiles.push_back(std::async(ggml_vk_create_pipeline_func, std::ref(device), std::ref(pipeline), name, spv_size, spv_data, entrypoint, parameter_count, push_constant_size, wg_denoms, specialization_constants, align)); + compiles.push_back(std::async(ggml_vk_create_pipeline_func, std::ref(device), std::ref(pipeline), name, spv_size, spv_data, entrypoint, + parameter_count, push_constant_size, wg_denoms, specialization_constants, align, disable_robustness, require_full_subgroups, required_subgroup_size)); }; +#if defined(VK_NV_cooperative_matrix2) && defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + if (device->coopmat2) { + + auto const &fa_wg_denoms = [&](uint32_t D, uint32_t clamp, ggml_type type, bool small_rows) -> std::array { + return {fa_rows_cols(D, clamp, type, small_rows)[0], 1, 1}; + }; + + auto const &fa_spec_constants = [&](uint32_t D, uint32_t clamp, ggml_type type, bool small_rows) -> std::vector { + // For large number of rows, 128 invocations seems to work best. + // For small number of rows (e.g. N==1), 256 works better. But matrix granularity for 256 is 32, so we + // can't use 256 for D==80. + uint32_t wg_size = (small_rows && (D % 32) == 0) ? 256 : 128; + auto rows_cols = fa_rows_cols(D, clamp, type, small_rows); + return {wg_size, rows_cols[0], rows_cols[1], (D), clamp}; + }; + +#define CREATE_FA2(TYPE, NAMELC, D) \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][0][0][0], "flash_attn_f32_f16_D" #D "_f16acc" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_len, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,1,TYPE,false), fa_spec_constants(D,1,TYPE,false), 1); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][0][0][1], "flash_attn_f32_f16_D" #D "_aligned_f16acc" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_len, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,0,TYPE,false), fa_spec_constants(D,0,TYPE,false), fa_rows_cols(D,0,TYPE,false)[1]); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][1][0][0], "flash_attn_f32_f16_D" #D "_f32acc" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _cm2_len, flash_attn_f32_f16_ ## NAMELC ## _cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,1,TYPE,false), fa_spec_constants(D,1,TYPE,false), 1); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][1][0][1], "flash_attn_f32_f16_D" #D "_aligned_f32acc" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _cm2_len, flash_attn_f32_f16_ ## NAMELC ## _cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,0,TYPE,false), fa_spec_constants(D,0,TYPE,false), fa_rows_cols(D,0,TYPE,false)[1]); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][0][1][0], "flash_attn_f32_f16_D" #D "_f16acc_smallrows" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_len, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,1,TYPE,true), fa_spec_constants(D,1,TYPE,true), 1); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][0][1][1], "flash_attn_f32_f16_D" #D "_aligned_f16acc_smallrows" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_len, flash_attn_f32_f16_ ## NAMELC ## _f16acc_cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,0,TYPE,true), fa_spec_constants(D,0,TYPE,true), fa_rows_cols(D,0,TYPE,true)[1]); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][1][1][0], "flash_attn_f32_f16_D" #D "_f32acc_smallrows" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _cm2_len, flash_attn_f32_f16_ ## NAMELC ## _cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,1,TYPE,true), fa_spec_constants(D,1,TYPE,true), 1); \ + ggml_vk_create_pipeline(device, device->pipeline_flash_attn_f32_f16_D ## D[TYPE][1][1][1], "flash_attn_f32_f16_D" #D "_aligned_f32acc_smallrows" #NAMELC, flash_attn_f32_f16_ ## NAMELC ## _cm2_len, flash_attn_f32_f16_ ## NAMELC ## _cm2_data, "main", 5, sizeof(vk_flash_attn_push_constants), fa_wg_denoms(D,0,TYPE,true), fa_spec_constants(D,0,TYPE,true), fa_rows_cols(D,0,TYPE,true)[1]); \ + +#define CREATE_FA(TYPE, NAMELC) \ + CREATE_FA2(TYPE, NAMELC, 64) \ + CREATE_FA2(TYPE, NAMELC, 80) \ + CREATE_FA2(TYPE, NAMELC, 96) \ + CREATE_FA2(TYPE, NAMELC, 112) \ + CREATE_FA2(TYPE, NAMELC, 128) \ + CREATE_FA2(TYPE, NAMELC, 256) + + CREATE_FA(GGML_TYPE_F16, f16) + CREATE_FA(GGML_TYPE_Q4_0, q4_0) + CREATE_FA(GGML_TYPE_Q4_1, q4_1) + CREATE_FA(GGML_TYPE_Q5_0, q5_0) + CREATE_FA(GGML_TYPE_Q5_1, q5_1) + CREATE_FA(GGML_TYPE_Q8_0, q8_0) + // K dequants currently disabled because D dimension is rounded up to 256 and runs inefficiently + //CREATE_FA(GGML_TYPE_Q2_K, q2_k) + //CREATE_FA(GGML_TYPE_Q3_K, q3_k) + //CREATE_FA(GGML_TYPE_Q4_K, q4_k) + //CREATE_FA(GGML_TYPE_Q5_K, q5_k) + //CREATE_FA(GGML_TYPE_Q6_K, q6_k) + CREATE_FA(GGML_TYPE_IQ4_NL, iq4_nl) +#undef CREATE_FA + + // Create 6 variants, {s,m,l}x{unaligned,aligned} +#define CREATE_MM(PIPELINE_NAME, NAMELC, F16ACC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->l, #NAMELC #F16ACC "_l", NAMELC ## F16ACC ## _cm2_len, NAMELC ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, 1); \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->m, #NAMELC #F16ACC "_m", NAMELC ## F16ACC ## _cm2_len, NAMELC ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, 1); \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->s, #NAMELC #F16ACC "_s", NAMELC ## F16ACC ## _cm2_len, NAMELC ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, 1); \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_l, #NAMELC #F16ACC "_aligned_l", NAMELC ## _aligned ## F16ACC ## _cm2_len, NAMELC ## _aligned ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, l_align); \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_m, #NAMELC #F16ACC "_aligned_m", NAMELC ## _aligned ## F16ACC ## _cm2_len, NAMELC ## _aligned ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, m_align); \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_s, #NAMELC #F16ACC "_aligned_s", NAMELC ## _aligned ## F16ACC ## _cm2_len, NAMELC ## _aligned ## F16ACC ## _cm2_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, s_align); \ + + // Create 2 variants, {f16,f32} accumulator +#define CREATE_MM2(PIPELINE_NAME, NAMELC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT) \ + CREATE_MM(PIPELINE_NAME . f16acc, NAMELC, _f16acc, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT) \ + CREATE_MM(PIPELINE_NAME . f32acc, NAMELC, , WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT) \ + + CREATE_MM(pipeline_matmul_f32, matmul_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_matmul_f32_f16, matmul_f32_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 3) + + CREATE_MM2(pipeline_matmul_f16, matmul_f16, wg_denoms, warptile, vk_mat_mat_push_constants, 3) + CREATE_MM2(pipeline_matmul_f16_f32, matmul_f16_f32, wg_denoms, warptile, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q4_0].f16acc, matmul_q4_0_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q4_1].f16acc, matmul_q4_1_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q5_0].f16acc, matmul_q5_0_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q5_1].f16acc, matmul_q5_1_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q8_0].f16acc, matmul_q8_0_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q2_K].f16acc, matmul_q2_k_f16, _f16acc, mmq_wg_denoms_k, warptile_mmq_k, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q3_K].f16acc, matmul_q3_k_f16, _f16acc, mmq_wg_denoms_k, warptile_mmq_k, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q4_K].f16acc, matmul_q4_k_f16, _f16acc, mmq_wg_denoms_k, warptile_mmq_k, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q5_K].f16acc, matmul_q5_k_f16, _f16acc, mmq_wg_denoms_k, warptile_mmq_k, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_Q6_K].f16acc, matmul_q6_k_f16, _f16acc, mmq_wg_denoms_k, warptile_mmq_k, vk_mat_mat_push_constants, 3) + CREATE_MM(pipeline_dequant_mul_mat_mat_f16[GGML_TYPE_IQ4_NL].f16acc, matmul_iq4_nl_f16, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3) + + CREATE_MM(pipeline_matmul_id_f32, matmul_id_f32_f32, , wg_denoms, warptile, vk_mat_mat_id_push_constants, 4) + CREATE_MM2(pipeline_matmul_id_f16, matmul_id_f16, wg_denoms, warptile, vk_mat_mat_id_push_constants, 4) + CREATE_MM2(pipeline_matmul_id_f16_f32, matmul_id_f16_f32, wg_denoms, warptile, vk_mat_mat_id_push_constants, 4) + + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0].f16acc, matmul_id_q4_0_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1].f16acc, matmul_id_q4_1_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0].f16acc, matmul_id_q5_0_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1].f16acc, matmul_id_q5_1_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0].f16acc, matmul_id_q8_0_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K].f16acc, matmul_id_q2_k_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K].f16acc, matmul_id_q3_k_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K].f16acc, matmul_id_q4_k_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K].f16acc, matmul_id_q5_k_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K].f16acc, matmul_id_q6_k_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL].f16acc, matmul_id_iq4_nl_f32, , mmqid_wg_denoms, warptile_mmqid, vk_mat_mat_id_push_constants, 4) +#undef CREATE_MM +#undef CREATE_MM2 + } else +#endif // defined(VK_NV_cooperative_matrix2) && defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) +#if defined(VK_KHR_cooperative_matrix) && defined(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) + if (device->coopmat_support) { + // Create 6 variants, {s,m,l}x{unaligned,aligned} +#define CREATE_MM(PIPELINE_NAME, NAMELC, F16ACC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->l, #NAMELC #F16ACC "_l", NAMELC ## F16ACC ## _coopmat_len, NAMELC ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, 1, false, true); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->m, #NAMELC #F16ACC "_m", NAMELC ## F16ACC ## _coopmat_len, NAMELC ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, 1, false, true); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->s, #NAMELC #F16ACC "_s", NAMELC ## F16ACC ## _coopmat_len, NAMELC ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, 1, false, true); \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_l, #NAMELC #F16ACC "_aligned_l", NAMELC ## _aligned ## F16ACC ## _coopmat_len, NAMELC ## _aligned ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, l_align, false, true); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_m, #NAMELC #F16ACC "_aligned_m", NAMELC ## _aligned ## F16ACC ## _coopmat_len, NAMELC ## _aligned ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, m_align, false, true); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_s, #NAMELC #F16ACC "_aligned_s", NAMELC ## _aligned ## F16ACC ## _coopmat_len, NAMELC ## _aligned ## F16ACC ## _coopmat_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, s_align, false, true); \ + + // Create 2 variants, {f16,f32} accumulator +#define CREATE_MM2(PIPELINE_NAME, NAMELC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + if (device->coopmat_acc_f16_support) { \ + CREATE_MM(PIPELINE_NAME . f16acc, NAMELC, _f16acc, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + } \ + if (device->coopmat_acc_f32_support) { \ + CREATE_MM(PIPELINE_NAME . f32acc, NAMELC, , WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + } \ + + CREATE_MM(pipeline_matmul_f32, matmul_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_matmul_f32_f16, matmul_f32_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM2(pipeline_matmul_f16, matmul_f16, wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM2(pipeline_matmul_f16_f32, matmul_f16_f32, wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + + if (device->coopmat_acc_f16_support) { + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0].f16acc, matmul_q4_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1].f16acc, matmul_q4_1_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0].f16acc, matmul_q5_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1].f16acc, matmul_q5_1_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0].f16acc, matmul_q8_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K].f16acc, matmul_q2_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K].f16acc, matmul_q3_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K].f16acc, matmul_q4_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K].f16acc, matmul_q5_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K].f16acc, matmul_q6_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL].f16acc, matmul_iq4_nl_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + } else { + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0].f16acc, matmul_q4_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1].f16acc, matmul_q4_1_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0].f16acc, matmul_q5_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1].f16acc, matmul_q5_1_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0].f16acc, matmul_q8_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K].f16acc, matmul_q2_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K].f16acc, matmul_q3_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K].f16acc, matmul_q4_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K].f16acc, matmul_q5_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K].f16acc, matmul_q6_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL].f16acc, matmul_iq4_nl_f32, , wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + } + + // If there's not enough shared memory for row_ids and the result tile, don't create these pipelines. + if (device->mul_mat_id_s || device->mul_mat_id_m || device->mul_mat_id_l) { + CREATE_MM(pipeline_matmul_id_f32, matmul_id_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM2(pipeline_matmul_id_f16, matmul_id_f16, wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM2(pipeline_matmul_id_f16_f32, matmul_id_f16_f32, wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + + if (device->coopmat_acc_f16_support) { + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0].f16acc, matmul_id_q4_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1].f16acc, matmul_id_q4_1_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0].f16acc, matmul_id_q5_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1].f16acc, matmul_id_q5_1_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0].f16acc, matmul_id_q8_0_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K].f16acc, matmul_id_q2_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K].f16acc, matmul_id_q3_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K].f16acc, matmul_id_q4_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K].f16acc, matmul_id_q5_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K].f16acc, matmul_id_q6_k_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL].f16acc, matmul_id_iq4_nl_f32, _f16acc, wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + } else { + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0].f16acc, matmul_id_q4_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1].f16acc, matmul_id_q4_1_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0].f16acc, matmul_id_q5_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1].f16acc, matmul_id_q5_1_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0].f16acc, matmul_id_q8_0_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K].f16acc, matmul_id_q2_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K].f16acc, matmul_id_q3_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K].f16acc, matmul_id_q4_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K].f16acc, matmul_id_q5_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K].f16acc, matmul_id_q6_k_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL].f16acc, matmul_id_iq4_nl_f32, , wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + } + } +#undef CREATE_MM2 +#undef CREATE_MM + } else +#endif // defined(VK_KHR_cooperative_matrix) && defined(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) if (device->fp16) { - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->l, "matmul_f32_l", matmul_f32_f32_len, matmul_f32_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->m, "matmul_f32_m", matmul_f32_f32_len, matmul_f32_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->s, "matmul_f32_s", matmul_f32_f32_len, matmul_f32_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_l, "matmul_f32_aligned_l", matmul_f32_f32_aligned_len, matmul_f32_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_m, "matmul_f32_aligned_m", matmul_f32_f32_aligned_len, matmul_f32_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_s, "matmul_f32_aligned_s", matmul_f32_f32_aligned_len, matmul_f32_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + // Create 6 variants, {s,m,l}x{unaligned,aligned} +#define CREATE_MM(PIPELINE_NAME, NAMELC, F16ACC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->l, #NAMELC #F16ACC "_l", NAMELC ## F16ACC ## _len, NAMELC ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->m, #NAMELC #F16ACC "_m", NAMELC ## F16ACC ## _len, NAMELC ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->s, #NAMELC #F16ACC "_s", NAMELC ## F16ACC ## _len, NAMELC ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_l, #NAMELC #F16ACC "_aligned_l", NAMELC ## _aligned ## F16ACC ## _len, NAMELC ## _aligned ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, l_align); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_m, #NAMELC #F16ACC "_aligned_m", NAMELC ## _aligned ## F16ACC ## _len, NAMELC ## _aligned ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, m_align); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_s, #NAMELC #F16ACC "_aligned_s", NAMELC ## _aligned ## F16ACC ## _len, NAMELC ## _aligned ## F16ACC ## _data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, s_align); \ - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->l, "matmul_f32_f16_l", matmul_f32_f16_len, matmul_f32_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->m, "matmul_f32_f16_m", matmul_f32_f16_len, matmul_f32_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->s, "matmul_f32_f16_s", matmul_f32_f16_len, matmul_f32_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_l, "matmul_f32_f16_aligned_l", matmul_f32_f16_aligned_len, matmul_f32_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_m, "matmul_f32_f16_aligned_m", matmul_f32_f16_aligned_len, matmul_f32_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_s, "matmul_f32_f16_aligned_s", matmul_f32_f16_aligned_len, matmul_f32_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + // Create 2 variants, {f16,f32} accumulator +#define CREATE_MM2(PIPELINE_NAME, NAMELC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + CREATE_MM(PIPELINE_NAME . f16acc, NAMELC, _f16acc, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + CREATE_MM(PIPELINE_NAME . f32acc, NAMELC, , WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->l, "matmul_f16_l", matmul_f16_len, matmul_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->m, "matmul_f16_m", matmul_f16_len, matmul_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->s, "matmul_f16_s", matmul_f16_len, matmul_f16_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_l, "matmul_f16_aligned_l", matmul_f16_aligned_len, matmul_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_m, "matmul_f16_aligned_m", matmul_f16_aligned_len, matmul_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_s, "matmul_f16_aligned_s", matmul_f16_aligned_len, matmul_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + CREATE_MM(pipeline_matmul_f32, matmul_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_matmul_f32_f16, matmul_f32_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM2(pipeline_matmul_f16, matmul_f16, wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM2(pipeline_matmul_f16_f32, matmul_f16_f32, wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->l, "matmul_f16_f32_l", matmul_f16_f32_len, matmul_f16_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->m, "matmul_f16_f32_m", matmul_f16_f32_len, matmul_f16_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->s, "matmul_f16_f32_s", matmul_f16_f32_len, matmul_f16_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_l, "matmul_f16_f32_aligned_l", matmul_f16_f32_aligned_len, matmul_f16_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_m, "matmul_f16_f32_aligned_m", matmul_f16_f32_aligned_len, matmul_f16_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_s, "matmul_f16_f32_aligned_s", matmul_f16_f32_aligned_len, matmul_f16_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0].f16acc, matmul_q4_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1].f16acc, matmul_q4_1_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0].f16acc, matmul_q5_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1].f16acc, matmul_q5_1_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0].f16acc, matmul_q8_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->l, "matmul_q4_0_f32_l", matmul_q4_0_f32_len, matmul_q4_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->m, "matmul_q4_0_f32_m", matmul_q4_0_f32_len, matmul_q4_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->s, "matmul_q4_0_f32_s", matmul_q4_0_f32_len, matmul_q4_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_l, "matmul_q4_0_f32_aligned_l", matmul_q4_0_f32_aligned_len, matmul_q4_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_m, "matmul_q4_0_f32_aligned_m", matmul_q4_0_f32_aligned_len, matmul_q4_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_s, "matmul_q4_0_f32_aligned_s", matmul_q4_0_f32_aligned_len, matmul_q4_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K].f16acc, matmul_q2_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K].f16acc, matmul_q3_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K].f16acc, matmul_q4_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K].f16acc, matmul_q5_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K].f16acc, matmul_q6_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL].f16acc, matmul_iq4_nl_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->l, "matmul_q4_1_f32_l", matmul_q4_1_f32_len, matmul_q4_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->m, "matmul_q4_1_f32_m", matmul_q4_1_f32_len, matmul_q4_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->s, "matmul_q4_1_f32_s", matmul_q4_1_f32_len, matmul_q4_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_l, "matmul_q4_1_f32_aligned_l", matmul_q4_1_f32_aligned_len, matmul_q4_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_m, "matmul_q4_1_f32_aligned_m", matmul_q4_1_f32_aligned_len, matmul_q4_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_s, "matmul_q4_1_f32_aligned_s", matmul_q4_1_f32_aligned_len, matmul_q4_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + // If there's not enough shared memory for row_ids and the result tile, don't create these pipelines. + if (device->mul_mat_id_s || device->mul_mat_id_m || device->mul_mat_id_l) { + CREATE_MM(pipeline_matmul_id_f32, matmul_id_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM2(pipeline_matmul_id_f16, matmul_id_f16, wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM2(pipeline_matmul_id_f16_f32, matmul_id_f16_f32, wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->l, "matmul_q5_0_f32_l", matmul_q5_0_f32_len, matmul_q5_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->m, "matmul_q5_0_f32_m", matmul_q5_0_f32_len, matmul_q5_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->s, "matmul_q5_0_f32_s", matmul_q5_0_f32_len, matmul_q5_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_l, "matmul_q5_0_f32_aligned_l", matmul_q5_0_f32_aligned_len, matmul_q5_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_m, "matmul_q5_0_f32_aligned_m", matmul_q5_0_f32_aligned_len, matmul_q5_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_s, "matmul_q5_0_f32_aligned_s", matmul_q5_0_f32_aligned_len, matmul_q5_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0].f16acc, matmul_id_q4_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1].f16acc, matmul_id_q4_1_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0].f16acc, matmul_id_q5_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1].f16acc, matmul_id_q5_1_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0].f16acc, matmul_id_q8_0_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->l, "matmul_q5_1_f32_l", matmul_q5_1_f32_len, matmul_q5_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->m, "matmul_q5_1_f32_m", matmul_q5_1_f32_len, matmul_q5_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->s, "matmul_q5_1_f32_s", matmul_q5_1_f32_len, matmul_q5_1_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_l, "matmul_q5_1_f32_aligned_l", matmul_q5_1_f32_aligned_len, matmul_q5_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_m, "matmul_q5_1_f32_aligned_m", matmul_q5_1_f32_aligned_len, matmul_q5_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_s, "matmul_q5_1_f32_aligned_s", matmul_q5_1_f32_aligned_len, matmul_q5_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->l, "matmul_q8_0_f32_l", matmul_q8_0_f32_len, matmul_q8_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->m, "matmul_q8_0_f32_m", matmul_q8_0_f32_len, matmul_q8_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->s, "matmul_q8_0_f32_s", matmul_q8_0_f32_len, matmul_q8_0_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_l, "matmul_q8_0_f32_aligned_l", matmul_q8_0_f32_aligned_len, matmul_q8_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_m, "matmul_q8_0_f32_aligned_m", matmul_q8_0_f32_aligned_len, matmul_q8_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_s, "matmul_q8_0_f32_aligned_s", matmul_q8_0_f32_aligned_len, matmul_q8_0_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->l, "matmul_q2_k_f32_l", matmul_q2_k_f32_len, matmul_q2_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->m, "matmul_q2_k_f32_m", matmul_q2_k_f32_len, matmul_q2_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->s, "matmul_q2_k_f32_s", matmul_q2_k_f32_len, matmul_q2_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_l, "matmul_q2_k_f32_aligned_l", matmul_q2_k_f32_aligned_len, matmul_q2_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_m, "matmul_q2_k_f32_aligned_m", matmul_q2_k_f32_aligned_len, matmul_q2_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_s, "matmul_q2_k_f32_aligned_s", matmul_q2_k_f32_aligned_len, matmul_q2_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->l, "matmul_q3_k_f32_l", matmul_q3_k_f32_len, matmul_q3_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->m, "matmul_q3_k_f32_m", matmul_q3_k_f32_len, matmul_q3_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->s, "matmul_q3_k_f32_s", matmul_q3_k_f32_len, matmul_q3_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_l, "matmul_q3_k_f32_aligned_l", matmul_q3_k_f32_aligned_len, matmul_q3_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_m, "matmul_q3_k_f32_aligned_m", matmul_q3_k_f32_aligned_len, matmul_q3_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_s, "matmul_q3_k_f32_aligned_s", matmul_q3_k_f32_aligned_len, matmul_q3_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->l, "matmul_q4_k_f32_l", matmul_q4_k_f32_len, matmul_q4_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->m, "matmul_q4_k_f32_m", matmul_q4_k_f32_len, matmul_q4_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->s, "matmul_q4_k_f32_s", matmul_q4_k_f32_len, matmul_q4_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_l, "matmul_q4_k_f32_aligned_l", matmul_q4_k_f32_aligned_len, matmul_q4_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_m, "matmul_q4_k_f32_aligned_m", matmul_q4_k_f32_aligned_len, matmul_q4_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_s, "matmul_q4_k_f32_aligned_s", matmul_q4_k_f32_aligned_len, matmul_q4_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->l, "matmul_q5_k_f32_l", matmul_q5_k_f32_len, matmul_q5_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->m, "matmul_q5_k_f32_m", matmul_q5_k_f32_len, matmul_q5_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->s, "matmul_q5_k_f32_s", matmul_q5_k_f32_len, matmul_q5_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_l, "matmul_q5_k_f32_aligned_l", matmul_q5_k_f32_aligned_len, matmul_q5_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_m, "matmul_q5_k_f32_aligned_m", matmul_q5_k_f32_aligned_len, matmul_q5_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_s, "matmul_q5_k_f32_aligned_s", matmul_q5_k_f32_aligned_len, matmul_q5_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->l, "matmul_q6_k_f32_l", matmul_q6_k_f32_len, matmul_q6_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->m, "matmul_q6_k_f32_m", matmul_q6_k_f32_len, matmul_q6_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->s, "matmul_q6_k_f32_s", matmul_q6_k_f32_len, matmul_q6_k_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_l, "matmul_q6_k_f32_aligned_l", matmul_q6_k_f32_aligned_len, matmul_q6_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_m, "matmul_q6_k_f32_aligned_m", matmul_q6_k_f32_aligned_len, matmul_q6_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_s, "matmul_q6_k_f32_aligned_s", matmul_q6_k_f32_aligned_len, matmul_q6_k_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->l, "matmul_iq4_nl_f32_l", matmul_iq4_nl_f32_len, matmul_iq4_nl_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->m, "matmul_iq4_nl_f32_m", matmul_iq4_nl_f32_len, matmul_iq4_nl_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->s, "matmul_iq4_nl_f32_s", matmul_iq4_nl_f32_len, matmul_iq4_nl_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_l, "matmul_iq4_nl_f32_aligned_l", matmul_iq4_nl_f32_aligned_len, matmul_iq4_nl_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_m, "matmul_iq4_nl_f32_aligned_m", matmul_iq4_nl_f32_aligned_len, matmul_iq4_nl_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_s, "matmul_iq4_nl_f32_aligned_s", matmul_iq4_nl_f32_aligned_len, matmul_iq4_nl_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->l, "matmul_id_f32_l", matmul_id_f32_f32_len, matmul_id_f32_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->m, "matmul_id_f32_m", matmul_id_f32_f32_len, matmul_id_f32_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->s, "matmul_id_f32_s", matmul_id_f32_f32_len, matmul_id_f32_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_l, "matmul_id_f32_aligned_l", matmul_id_f32_f32_aligned_len, matmul_id_f32_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_m, "matmul_id_f32_aligned_m", matmul_id_f32_f32_aligned_len, matmul_id_f32_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_s, "matmul_id_f32_aligned_s", matmul_id_f32_f32_aligned_len, matmul_id_f32_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->l, "matmul_id_f16_l", matmul_id_f16_len, matmul_id_f16_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->m, "matmul_id_f16_m", matmul_id_f16_len, matmul_id_f16_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->s, "matmul_id_f16_s", matmul_id_f16_len, matmul_id_f16_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_l, "matmul_id_f16_aligned_l", matmul_id_f16_aligned_len, matmul_id_f16_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_m, "matmul_id_f16_aligned_m", matmul_id_f16_aligned_len, matmul_id_f16_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_s, "matmul_id_f16_aligned_s", matmul_id_f16_aligned_len, matmul_id_f16_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->l, "matmul_id_f16_f32_l", matmul_id_f16_f32_len, matmul_id_f16_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->m, "matmul_id_f16_f32_m", matmul_id_f16_f32_len, matmul_id_f16_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->s, "matmul_id_f16_f32_s", matmul_id_f16_f32_len, matmul_id_f16_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_l, "matmul_id_f16_f32_aligned_l", matmul_id_f16_f32_aligned_len, matmul_id_f16_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_m, "matmul_id_f16_f32_aligned_m", matmul_id_f16_f32_aligned_len, matmul_id_f16_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_s, "matmul_id_f16_f32_aligned_s", matmul_id_f16_f32_aligned_len, matmul_id_f16_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->l, "matmul_id_q4_0_f32_l", matmul_id_q4_0_f32_len, matmul_id_q4_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->m, "matmul_id_q4_0_f32_m", matmul_id_q4_0_f32_len, matmul_id_q4_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->s, "matmul_id_q4_0_f32_s", matmul_id_q4_0_f32_len, matmul_id_q4_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_l, "matmul_id_q4_0_f32_aligned_l", matmul_id_q4_0_f32_aligned_len, matmul_id_q4_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_m, "matmul_id_q4_0_f32_aligned_m", matmul_id_q4_0_f32_aligned_len, matmul_id_q4_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_s, "matmul_id_q4_0_f32_aligned_s", matmul_id_q4_0_f32_aligned_len, matmul_id_q4_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->l, "matmul_id_q4_1_f32_l", matmul_id_q4_1_f32_len, matmul_id_q4_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->m, "matmul_id_q4_1_f32_m", matmul_id_q4_1_f32_len, matmul_id_q4_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->s, "matmul_id_q4_1_f32_s", matmul_id_q4_1_f32_len, matmul_id_q4_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_l, "matmul_id_q4_1_f32_aligned_l", matmul_id_q4_1_f32_aligned_len, matmul_id_q4_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_m, "matmul_id_q4_1_f32_aligned_m", matmul_id_q4_1_f32_aligned_len, matmul_id_q4_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_s, "matmul_id_q4_1_f32_aligned_s", matmul_id_q4_1_f32_aligned_len, matmul_id_q4_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->l, "matmul_id_q5_0_f32_l", matmul_id_q5_0_f32_len, matmul_id_q5_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->m, "matmul_id_q5_0_f32_m", matmul_id_q5_0_f32_len, matmul_id_q5_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->s, "matmul_id_q5_0_f32_s", matmul_id_q5_0_f32_len, matmul_id_q5_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_l, "matmul_id_q5_0_f32_aligned_l", matmul_id_q5_0_f32_aligned_len, matmul_id_q5_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_m, "matmul_id_q5_0_f32_aligned_m", matmul_id_q5_0_f32_aligned_len, matmul_id_q5_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_s, "matmul_id_q5_0_f32_aligned_s", matmul_id_q5_0_f32_aligned_len, matmul_id_q5_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->l, "matmul_id_q5_1_f32_l", matmul_id_q5_1_f32_len, matmul_id_q5_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->m, "matmul_id_q5_1_f32_m", matmul_id_q5_1_f32_len, matmul_id_q5_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->s, "matmul_id_q5_1_f32_s", matmul_id_q5_1_f32_len, matmul_id_q5_1_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_l, "matmul_id_q5_1_f32_aligned_l", matmul_id_q5_1_f32_aligned_len, matmul_id_q5_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_m, "matmul_id_q5_1_f32_aligned_m", matmul_id_q5_1_f32_aligned_len, matmul_id_q5_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_s, "matmul_id_q5_1_f32_aligned_s", matmul_id_q5_1_f32_aligned_len, matmul_id_q5_1_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->l, "matmul_id_q8_0_f32_l", matmul_id_q8_0_f32_len, matmul_id_q8_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->m, "matmul_id_q8_0_f32_m", matmul_id_q8_0_f32_len, matmul_id_q8_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->s, "matmul_id_q8_0_f32_s", matmul_id_q8_0_f32_len, matmul_id_q8_0_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_l, "matmul_id_q8_0_f32_aligned_l", matmul_id_q8_0_f32_aligned_len, matmul_id_q8_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_m, "matmul_id_q8_0_f32_aligned_m", matmul_id_q8_0_f32_aligned_len, matmul_id_q8_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_s, "matmul_id_q8_0_f32_aligned_s", matmul_id_q8_0_f32_aligned_len, matmul_id_q8_0_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->l, "matmul_id_q2_k_f32_l", matmul_id_q2_k_f32_len, matmul_id_q2_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->m, "matmul_id_q2_k_f32_m", matmul_id_q2_k_f32_len, matmul_id_q2_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->s, "matmul_id_q2_k_f32_s", matmul_id_q2_k_f32_len, matmul_id_q2_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_l, "matmul_id_q2_k_f32_aligned_l", matmul_id_q2_k_f32_aligned_len, matmul_id_q2_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_m, "matmul_id_q2_k_f32_aligned_m", matmul_id_q2_k_f32_aligned_len, matmul_id_q2_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_s, "matmul_id_q2_k_f32_aligned_s", matmul_id_q2_k_f32_aligned_len, matmul_id_q2_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->l, "matmul_id_q3_k_f32_l", matmul_id_q3_k_f32_len, matmul_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->m, "matmul_id_q3_k_f32_m", matmul_id_q3_k_f32_len, matmul_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->s, "matmul_id_q3_k_f32_s", matmul_id_q3_k_f32_len, matmul_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_l, "matmul_id_q3_k_f32_aligned_l", matmul_id_q3_k_f32_aligned_len, matmul_id_q3_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_m, "matmul_id_q3_k_f32_aligned_m", matmul_id_q3_k_f32_aligned_len, matmul_id_q3_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_s, "matmul_id_q3_k_f32_aligned_s", matmul_id_q3_k_f32_aligned_len, matmul_id_q3_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->l, "matmul_id_q4_k_f32_l", matmul_id_q4_k_f32_len, matmul_id_q4_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->m, "matmul_id_q4_k_f32_m", matmul_id_q4_k_f32_len, matmul_id_q4_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->s, "matmul_id_q4_k_f32_s", matmul_id_q4_k_f32_len, matmul_id_q4_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_l, "matmul_id_q4_k_f32_aligned_l", matmul_id_q4_k_f32_aligned_len, matmul_id_q4_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_m, "matmul_id_q4_k_f32_aligned_m", matmul_id_q4_k_f32_aligned_len, matmul_id_q4_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_s, "matmul_id_q4_k_f32_aligned_s", matmul_id_q4_k_f32_aligned_len, matmul_id_q4_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->l, "matmul_id_q5_k_f32_l", matmul_id_q5_k_f32_len, matmul_id_q5_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->m, "matmul_id_q5_k_f32_m", matmul_id_q5_k_f32_len, matmul_id_q5_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->s, "matmul_id_q5_k_f32_s", matmul_id_q5_k_f32_len, matmul_id_q5_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_l, "matmul_id_q5_k_f32_aligned_l", matmul_id_q5_k_f32_aligned_len, matmul_id_q5_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_m, "matmul_id_q5_k_f32_aligned_m", matmul_id_q5_k_f32_aligned_len, matmul_id_q5_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_s, "matmul_id_q5_k_f32_aligned_s", matmul_id_q5_k_f32_aligned_len, matmul_id_q5_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->l, "matmul_id_q6_k_f32_l", matmul_id_q6_k_f32_len, matmul_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->m, "matmul_id_q6_k_f32_m", matmul_id_q6_k_f32_len, matmul_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->s, "matmul_id_q6_k_f32_s", matmul_id_q6_k_f32_len, matmul_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_l, "matmul_id_q6_k_f32_aligned_l", matmul_id_q6_k_f32_aligned_len, matmul_id_q6_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_m, "matmul_id_q6_k_f32_aligned_m", matmul_id_q6_k_f32_aligned_len, matmul_id_q6_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_s, "matmul_id_q6_k_f32_aligned_s", matmul_id_q6_k_f32_aligned_len, matmul_id_q6_k_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->l, "matmul_id_iq4_nl_f32_l", matmul_id_iq4_nl_f32_len, matmul_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->m, "matmul_id_iq4_nl_f32_m", matmul_id_iq4_nl_f32_len, matmul_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->s, "matmul_id_iq4_nl_f32_s", matmul_id_iq4_nl_f32_len, matmul_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_l, "matmul_id_iq4_nl_f32_aligned_l", matmul_id_iq4_nl_f32_aligned_len, matmul_id_iq4_nl_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_m, "matmul_id_iq4_nl_f32_aligned_m", matmul_id_iq4_nl_f32_aligned_len, matmul_id_iq4_nl_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_s, "matmul_id_iq4_nl_f32_aligned_s", matmul_id_iq4_nl_f32_aligned_len, matmul_id_iq4_nl_f32_aligned_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K].f16acc, matmul_id_q2_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K].f16acc, matmul_id_q3_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K].f16acc, matmul_id_q4_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K].f16acc, matmul_id_q5_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K].f16acc, matmul_id_q6_k_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL].f16acc, matmul_id_iq4_nl_f32, _f16acc, mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + } +#undef CREATE_MM2 +#undef CREATE_MM } else { - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->l, "matmul_f32_l", matmul_f32_f32_fp32_len, matmul_f32_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->m, "matmul_f32_m", matmul_f32_f32_fp32_len, matmul_f32_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->s, "matmul_f32_s", matmul_f32_f32_fp32_len, matmul_f32_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_l, "matmul_f32_aligned_l", matmul_f32_f32_aligned_fp32_len, matmul_f32_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_m, "matmul_f32_aligned_m", matmul_f32_f32_aligned_fp32_len, matmul_f32_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->a_s, "matmul_f32_aligned_s", matmul_f32_f32_aligned_fp32_len, matmul_f32_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + // Create 6 variants, {s,m,l}x{unaligned,aligned} +#define CREATE_MM(PIPELINE_NAME, NAMELC, F16ACC, WG_DENOMS, WARPTILE, PUSHCONST, PARAMCOUNT, ID) \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->l, #NAMELC #F16ACC "_l", NAMELC ## F16ACC ## _fp32_len, NAMELC ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->m, #NAMELC #F16ACC "_m", NAMELC ## F16ACC ## _fp32_len, NAMELC ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->s, #NAMELC #F16ACC "_s", NAMELC ## F16ACC ## _fp32_len, NAMELC ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, 1); \ + if (device->mul_mat ## ID ## _l) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_l, #NAMELC #F16ACC "_aligned_l", NAMELC ## _aligned ## F16ACC ## _fp32_len, NAMELC ## _aligned ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), l_ ## WG_DENOMS, l_ ## WARPTILE, l_align); \ + if (device->mul_mat ## ID ## _m) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_m, #NAMELC #F16ACC "_aligned_m", NAMELC ## _aligned ## F16ACC ## _fp32_len, NAMELC ## _aligned ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), m_ ## WG_DENOMS, m_ ## WARPTILE, m_align); \ + if (device->mul_mat ## ID ## _s) \ + ggml_vk_create_pipeline(device, device-> PIPELINE_NAME ->a_s, #NAMELC #F16ACC "_aligned_s", NAMELC ## _aligned ## F16ACC ## _fp32_len, NAMELC ## _aligned ## F16ACC ## _fp32_data, "main", PARAMCOUNT, sizeof(PUSHCONST), s_ ## WG_DENOMS, s_ ## WARPTILE, s_align); \ - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->l, "matmul_f32_f16_l", matmul_f32_f16_fp32_len, matmul_f32_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->m, "matmul_f32_f16_m", matmul_f32_f16_fp32_len, matmul_f32_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->s, "matmul_f32_f16_s", matmul_f32_f16_fp32_len, matmul_f32_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_l, "matmul_f32_f16_aligned_l", matmul_f32_f16_aligned_fp32_len, matmul_f32_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_m, "matmul_f32_f16_aligned_m", matmul_f32_f16_aligned_fp32_len, matmul_f32_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f32_f16->a_s, "matmul_f32_f16_aligned_s", matmul_f32_f16_aligned_fp32_len, matmul_f32_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + CREATE_MM(pipeline_matmul_f32, matmul_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_matmul_f32_f16, matmul_f32_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_matmul_f16.f32acc, matmul_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_matmul_f16_f32.f32acc, matmul_f16_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->l, "matmul_f16_l", matmul_f16_fp32_len, matmul_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->m, "matmul_f16_m", matmul_f16_fp32_len, matmul_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->s, "matmul_f16_s", matmul_f16_fp32_len, matmul_f16_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_l, "matmul_f16_aligned_l", matmul_f16_aligned_fp32_len, matmul_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_m, "matmul_f16_aligned_m", matmul_f16_aligned_fp32_len, matmul_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_s, "matmul_f16_aligned_s", matmul_f16_aligned_fp32_len, matmul_f16_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0].f32acc, matmul_q4_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1].f32acc, matmul_q4_1_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0].f32acc, matmul_q5_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1].f32acc, matmul_q5_1_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0].f32acc, matmul_q8_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->l, "matmul_f16_f32_l", matmul_f16_f32_fp32_len, matmul_f16_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->m, "matmul_f16_f32_m", matmul_f16_f32_fp32_len, matmul_f16_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->s, "matmul_f16_f32_s", matmul_f16_f32_fp32_len, matmul_f16_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_l, "matmul_f16_f32_aligned_l", matmul_f16_f32_aligned_fp32_len, matmul_f16_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_m, "matmul_f16_f32_aligned_m", matmul_f16_f32_aligned_fp32_len, matmul_f16_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_f16_f32->a_s, "matmul_f16_f32_aligned_s", matmul_f16_f32_aligned_fp32_len, matmul_f16_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K].f32acc, matmul_q2_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K].f32acc, matmul_q3_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K].f32acc, matmul_q4_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K].f32acc, matmul_q5_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K].f32acc, matmul_q6_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); + CREATE_MM(pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL].f32acc, matmul_iq4_nl_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_push_constants, 3, ); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->l, "matmul_q4_0_f32_l", matmul_q4_0_f32_fp32_len, matmul_q4_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->m, "matmul_q4_0_f32_m", matmul_q4_0_f32_fp32_len, matmul_q4_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->s, "matmul_q4_0_f32_s", matmul_q4_0_f32_fp32_len, matmul_q4_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_l, "matmul_q4_0_f32_aligned_l", matmul_q4_0_f32_aligned_fp32_len, matmul_q4_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_m, "matmul_q4_0_f32_aligned_m", matmul_q4_0_f32_aligned_fp32_len, matmul_q4_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_0]->a_s, "matmul_q4_0_f32_aligned_s", matmul_q4_0_f32_aligned_fp32_len, matmul_q4_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + // If there's not enough shared memory for row_ids and the result tile, don't create these pipelines. + if (device->mul_mat_id_s || device->mul_mat_id_m || device->mul_mat_id_l) { + CREATE_MM(pipeline_matmul_id_f32, matmul_id_f32_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM(pipeline_matmul_id_f16.f32acc, matmul_id_f16, , wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); + CREATE_MM(pipeline_matmul_id_f16_f32.f32acc, matmul_id_f16_f32, , wg_denoms, warptile, vk_mat_mat_push_constants, 4, _id); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->l, "matmul_q4_1_f32_l", matmul_q4_1_f32_fp32_len, matmul_q4_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->m, "matmul_q4_1_f32_m", matmul_q4_1_f32_fp32_len, matmul_q4_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->s, "matmul_q4_1_f32_s", matmul_q4_1_f32_fp32_len, matmul_q4_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_l, "matmul_q4_1_f32_aligned_l", matmul_q4_1_f32_aligned_fp32_len, matmul_q4_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_m, "matmul_q4_1_f32_aligned_m", matmul_q4_1_f32_aligned_fp32_len, matmul_q4_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_s, "matmul_q4_1_f32_aligned_s", matmul_q4_1_f32_aligned_fp32_len, matmul_q4_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0].f32acc, matmul_id_q4_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1].f32acc, matmul_id_q4_1_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0].f32acc, matmul_id_q5_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1].f32acc, matmul_id_q5_1_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0].f32acc, matmul_id_q8_0_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->l, "matmul_q5_0_f32_l", matmul_q5_0_f32_fp32_len, matmul_q5_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->m, "matmul_q5_0_f32_m", matmul_q5_0_f32_fp32_len, matmul_q5_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->s, "matmul_q5_0_f32_s", matmul_q5_0_f32_fp32_len, matmul_q5_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_l, "matmul_q5_0_f32_aligned_l", matmul_q5_0_f32_aligned_fp32_len, matmul_q5_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_m, "matmul_q5_0_f32_aligned_m", matmul_q5_0_f32_aligned_fp32_len, matmul_q5_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_0]->a_s, "matmul_q5_0_f32_aligned_s", matmul_q5_0_f32_aligned_fp32_len, matmul_q5_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->l, "matmul_q5_1_f32_l", matmul_q5_1_f32_fp32_len, matmul_q5_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->m, "matmul_q5_1_f32_m", matmul_q5_1_f32_fp32_len, matmul_q5_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->s, "matmul_q5_1_f32_s", matmul_q5_1_f32_fp32_len, matmul_q5_1_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_l, "matmul_q5_1_f32_aligned_l", matmul_q5_1_f32_aligned_fp32_len, matmul_q5_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_m, "matmul_q5_1_f32_aligned_m", matmul_q5_1_f32_aligned_fp32_len, matmul_q5_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_1]->a_s, "matmul_q5_1_f32_aligned_s", matmul_q5_1_f32_aligned_fp32_len, matmul_q5_1_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->l, "matmul_q8_0_f32_l", matmul_q8_0_f32_fp32_len, matmul_q8_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->m, "matmul_q8_0_f32_m", matmul_q8_0_f32_fp32_len, matmul_q8_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->s, "matmul_q8_0_f32_s", matmul_q8_0_f32_fp32_len, matmul_q8_0_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_l, "matmul_q8_0_f32_aligned_l", matmul_q8_0_f32_aligned_fp32_len, matmul_q8_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_m, "matmul_q8_0_f32_aligned_m", matmul_q8_0_f32_aligned_fp32_len, matmul_q8_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q8_0]->a_s, "matmul_q8_0_f32_aligned_s", matmul_q8_0_f32_aligned_fp32_len, matmul_q8_0_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->l, "matmul_q2_k_f32_l", matmul_q2_k_f32_fp32_len, matmul_q2_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->m, "matmul_q2_k_f32_m", matmul_q2_k_f32_fp32_len, matmul_q2_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->s, "matmul_q2_k_f32_s", matmul_q2_k_f32_fp32_len, matmul_q2_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_l, "matmul_q2_k_f32_aligned_l", matmul_q2_k_f32_aligned_fp32_len, matmul_q2_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_m, "matmul_q2_k_f32_aligned_m", matmul_q2_k_f32_aligned_fp32_len, matmul_q2_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q2_K]->a_s, "matmul_q2_k_f32_aligned_s", matmul_q2_k_f32_aligned_fp32_len, matmul_q2_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->l, "matmul_q3_k_f32_l", matmul_q3_k_f32_fp32_len, matmul_q3_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->m, "matmul_q3_k_f32_m", matmul_q3_k_f32_fp32_len, matmul_q3_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->s, "matmul_q3_k_f32_s", matmul_q3_k_f32_fp32_len, matmul_q3_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_l, "matmul_q3_k_f32_aligned_l", matmul_q3_k_f32_aligned_fp32_len, matmul_q3_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_m, "matmul_q3_k_f32_aligned_m", matmul_q3_k_f32_aligned_fp32_len, matmul_q3_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q3_K]->a_s, "matmul_q3_k_f32_aligned_s", matmul_q3_k_f32_aligned_fp32_len, matmul_q3_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->l, "matmul_q4_k_f32_l", matmul_q4_k_f32_fp32_len, matmul_q4_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->m, "matmul_q4_k_f32_m", matmul_q4_k_f32_fp32_len, matmul_q4_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->s, "matmul_q4_k_f32_s", matmul_q4_k_f32_fp32_len, matmul_q4_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_l, "matmul_q4_k_f32_aligned_l", matmul_q4_k_f32_aligned_fp32_len, matmul_q4_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_m, "matmul_q4_k_f32_aligned_m", matmul_q4_k_f32_aligned_fp32_len, matmul_q4_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_K]->a_s, "matmul_q4_k_f32_aligned_s", matmul_q4_k_f32_aligned_fp32_len, matmul_q4_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->l, "matmul_q5_k_f32_l", matmul_q5_k_f32_fp32_len, matmul_q5_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->m, "matmul_q5_k_f32_m", matmul_q5_k_f32_fp32_len, matmul_q5_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->s, "matmul_q5_k_f32_s", matmul_q5_k_f32_fp32_len, matmul_q5_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_l, "matmul_q5_k_f32_aligned_l", matmul_q5_k_f32_aligned_fp32_len, matmul_q5_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_m, "matmul_q5_k_f32_aligned_m", matmul_q5_k_f32_aligned_fp32_len, matmul_q5_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q5_K]->a_s, "matmul_q5_k_f32_aligned_s", matmul_q5_k_f32_aligned_fp32_len, matmul_q5_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->l, "matmul_q6_k_f32_l", matmul_q6_k_f32_fp32_len, matmul_q6_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->m, "matmul_q6_k_f32_m", matmul_q6_k_f32_fp32_len, matmul_q6_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->s, "matmul_q6_k_f32_s", matmul_q6_k_f32_fp32_len, matmul_q6_k_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_l, "matmul_q6_k_f32_aligned_l", matmul_q6_k_f32_aligned_fp32_len, matmul_q6_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_m, "matmul_q6_k_f32_aligned_m", matmul_q6_k_f32_aligned_fp32_len, matmul_q6_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q6_K]->a_s, "matmul_q6_k_f32_aligned_s", matmul_q6_k_f32_aligned_fp32_len, matmul_q6_k_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->l, "matmul_iq4_nl_f32_l", matmul_iq4_nl_f32_fp32_len, matmul_iq4_nl_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->m, "matmul_iq4_nl_f32_m", matmul_iq4_nl_f32_fp32_len, matmul_iq4_nl_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->s, "matmul_iq4_nl_f32_s", matmul_iq4_nl_f32_fp32_len, matmul_iq4_nl_f32_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_l, "matmul_iq4_nl_f32_aligned_l", matmul_iq4_nl_f32_aligned_fp32_len, matmul_iq4_nl_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_m, "matmul_iq4_nl_f32_aligned_m", matmul_iq4_nl_f32_aligned_fp32_len, matmul_iq4_nl_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_IQ4_NL]->a_s, "matmul_iq4_nl_f32_aligned_s", matmul_iq4_nl_f32_aligned_fp32_len, matmul_iq4_nl_f32_aligned_fp32_data, "main", 3, sizeof(vk_mat_mat_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->l, "matmul_id_f32_l", matmul_id_f32_f32_fp32_len, matmul_id_f32_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->m, "matmul_id_f32_m", matmul_id_f32_f32_fp32_len, matmul_id_f32_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->s, "matmul_id_f32_s", matmul_id_f32_f32_fp32_len, matmul_id_f32_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_l, "matmul_id_f32_aligned_l", matmul_id_f32_f32_aligned_fp32_len, matmul_id_f32_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_m, "matmul_id_f32_aligned_m", matmul_id_f32_f32_aligned_fp32_len, matmul_id_f32_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f32->a_s, "matmul_id_f32_aligned_s", matmul_id_f32_f32_aligned_fp32_len, matmul_id_f32_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->l, "matmul_id_f16_l", matmul_id_f16_fp32_len, matmul_id_f16_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->m, "matmul_id_f16_m", matmul_id_f16_fp32_len, matmul_id_f16_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->s, "matmul_id_f16_s", matmul_id_f16_fp32_len, matmul_id_f16_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_l, "matmul_id_f16_aligned_l", matmul_id_f16_aligned_fp32_len, matmul_id_f16_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_m, "matmul_id_f16_aligned_m", matmul_id_f16_aligned_fp32_len, matmul_id_f16_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16->a_s, "matmul_id_f16_aligned_s", matmul_id_f16_aligned_fp32_len, matmul_id_f16_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->l, "matmul_id_f16_f32_l", matmul_id_f16_f32_fp32_len, matmul_id_f16_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->m, "matmul_id_f16_f32_m", matmul_id_f16_f32_fp32_len, matmul_id_f16_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->s, "matmul_id_f16_f32_s", matmul_id_f16_f32_fp32_len, matmul_id_f16_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_l, "matmul_id_f16_f32_aligned_l", matmul_id_f16_f32_aligned_fp32_len, matmul_id_f16_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_m, "matmul_id_f16_f32_aligned_m", matmul_id_f16_f32_aligned_fp32_len, matmul_id_f16_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_matmul_id_f16_f32->a_s, "matmul_id_f16_f32_aligned_s", matmul_id_f16_f32_aligned_fp32_len, matmul_id_f16_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->l, "matmul_id_q4_0_f32_l", matmul_id_q4_0_f32_fp32_len, matmul_id_q4_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->m, "matmul_id_q4_0_f32_m", matmul_id_q4_0_f32_fp32_len, matmul_id_q4_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->s, "matmul_id_q4_0_f32_s", matmul_id_q4_0_f32_fp32_len, matmul_id_q4_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_l, "matmul_id_q4_0_f32_aligned_l", matmul_id_q4_0_f32_aligned_fp32_len, matmul_id_q4_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_m, "matmul_id_q4_0_f32_aligned_m", matmul_id_q4_0_f32_aligned_fp32_len, matmul_id_q4_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->a_s, "matmul_id_q4_0_f32_aligned_s", matmul_id_q4_0_f32_aligned_fp32_len, matmul_id_q4_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->l, "matmul_id_q4_1_f32_l", matmul_id_q4_1_f32_fp32_len, matmul_id_q4_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->m, "matmul_id_q4_1_f32_m", matmul_id_q4_1_f32_fp32_len, matmul_id_q4_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->s, "matmul_id_q4_1_f32_s", matmul_id_q4_1_f32_fp32_len, matmul_id_q4_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_l, "matmul_id_q4_1_f32_aligned_l", matmul_id_q4_1_f32_aligned_fp32_len, matmul_id_q4_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_m, "matmul_id_q4_1_f32_aligned_m", matmul_id_q4_1_f32_aligned_fp32_len, matmul_id_q4_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_1]->a_s, "matmul_id_q4_1_f32_aligned_s", matmul_id_q4_1_f32_aligned_fp32_len, matmul_id_q4_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->l, "matmul_id_q5_0_f32_l", matmul_id_q5_0_f32_fp32_len, matmul_id_q5_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->m, "matmul_id_q5_0_f32_m", matmul_id_q5_0_f32_fp32_len, matmul_id_q5_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->s, "matmul_id_q5_0_f32_s", matmul_id_q5_0_f32_fp32_len, matmul_id_q5_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_l, "matmul_id_q5_0_f32_aligned_l", matmul_id_q5_0_f32_aligned_fp32_len, matmul_id_q5_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_m, "matmul_id_q5_0_f32_aligned_m", matmul_id_q5_0_f32_aligned_fp32_len, matmul_id_q5_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_0]->a_s, "matmul_id_q5_0_f32_aligned_s", matmul_id_q5_0_f32_aligned_fp32_len, matmul_id_q5_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->l, "matmul_id_q5_1_f32_l", matmul_id_q5_1_f32_fp32_len, matmul_id_q5_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->m, "matmul_id_q5_1_f32_m", matmul_id_q5_1_f32_fp32_len, matmul_id_q5_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->s, "matmul_id_q5_1_f32_s", matmul_id_q5_1_f32_fp32_len, matmul_id_q5_1_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_l, "matmul_id_q5_1_f32_aligned_l", matmul_id_q5_1_f32_aligned_fp32_len, matmul_id_q5_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_m, "matmul_id_q5_1_f32_aligned_m", matmul_id_q5_1_f32_aligned_fp32_len, matmul_id_q5_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_1]->a_s, "matmul_id_q5_1_f32_aligned_s", matmul_id_q5_1_f32_aligned_fp32_len, matmul_id_q5_1_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->l, "matmul_id_q8_0_f32_l", matmul_id_q8_0_f32_fp32_len, matmul_id_q8_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->m, "matmul_id_q8_0_f32_m", matmul_id_q8_0_f32_fp32_len, matmul_id_q8_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->s, "matmul_id_q8_0_f32_s", matmul_id_q8_0_f32_fp32_len, matmul_id_q8_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_l, "matmul_id_q8_0_f32_aligned_l", matmul_id_q8_0_f32_aligned_fp32_len, matmul_id_q8_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_m, "matmul_id_q8_0_f32_aligned_m", matmul_id_q8_0_f32_aligned_fp32_len, matmul_id_q8_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q8_0]->a_s, "matmul_id_q8_0_f32_aligned_s", matmul_id_q8_0_f32_aligned_fp32_len, matmul_id_q8_0_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->l, "matmul_id_q2_k_f32_l", matmul_id_q2_k_f32_fp32_len, matmul_id_q2_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->m, "matmul_id_q2_k_f32_m", matmul_id_q2_k_f32_fp32_len, matmul_id_q2_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->s, "matmul_id_q2_k_f32_s", matmul_id_q2_k_f32_fp32_len, matmul_id_q2_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_l, "matmul_id_q2_k_f32_aligned_l", matmul_id_q2_k_f32_aligned_fp32_len, matmul_id_q2_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_m, "matmul_id_q2_k_f32_aligned_m", matmul_id_q2_k_f32_aligned_fp32_len, matmul_id_q2_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K]->a_s, "matmul_id_q2_k_f32_aligned_s", matmul_id_q2_k_f32_aligned_fp32_len, matmul_id_q2_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->l, "matmul_id_q3_k_f32_l", matmul_id_q3_k_f32_fp32_len, matmul_id_q3_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->m, "matmul_id_q3_k_f32_m", matmul_id_q3_k_f32_fp32_len, matmul_id_q3_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->s, "matmul_id_q3_k_f32_s", matmul_id_q3_k_f32_fp32_len, matmul_id_q3_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_l, "matmul_id_q3_k_f32_aligned_l", matmul_id_q3_k_f32_aligned_fp32_len, matmul_id_q3_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_m, "matmul_id_q3_k_f32_aligned_m", matmul_id_q3_k_f32_aligned_fp32_len, matmul_id_q3_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K]->a_s, "matmul_id_q3_k_f32_aligned_s", matmul_id_q3_k_f32_aligned_fp32_len, matmul_id_q3_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->l, "matmul_id_q4_k_f32_l", matmul_id_q4_k_f32_fp32_len, matmul_id_q4_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->m, "matmul_id_q4_k_f32_m", matmul_id_q4_k_f32_fp32_len, matmul_id_q4_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->s, "matmul_id_q4_k_f32_s", matmul_id_q4_k_f32_fp32_len, matmul_id_q4_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_l, "matmul_id_q4_k_f32_aligned_l", matmul_id_q4_k_f32_aligned_fp32_len, matmul_id_q4_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_m, "matmul_id_q4_k_f32_aligned_m", matmul_id_q4_k_f32_aligned_fp32_len, matmul_id_q4_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K]->a_s, "matmul_id_q4_k_f32_aligned_s", matmul_id_q4_k_f32_aligned_fp32_len, matmul_id_q4_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->l, "matmul_id_q5_k_f32_l", matmul_id_q5_k_f32_fp32_len, matmul_id_q5_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->m, "matmul_id_q5_k_f32_m", matmul_id_q5_k_f32_fp32_len, matmul_id_q5_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->s, "matmul_id_q5_k_f32_s", matmul_id_q5_k_f32_fp32_len, matmul_id_q5_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_l, "matmul_id_q5_k_f32_aligned_l", matmul_id_q5_k_f32_aligned_fp32_len, matmul_id_q5_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_m, "matmul_id_q5_k_f32_aligned_m", matmul_id_q5_k_f32_aligned_fp32_len, matmul_id_q5_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_s, "matmul_id_q5_k_f32_aligned_s", matmul_id_q5_k_f32_aligned_fp32_len, matmul_id_q5_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->l, "matmul_id_q6_k_f32_l", matmul_id_q6_k_f32_fp32_len, matmul_id_q6_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->m, "matmul_id_q6_k_f32_m", matmul_id_q6_k_f32_fp32_len, matmul_id_q6_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->s, "matmul_id_q6_k_f32_s", matmul_id_q6_k_f32_fp32_len, matmul_id_q6_k_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_l, "matmul_id_q6_k_f32_aligned_l", matmul_id_q6_k_f32_aligned_fp32_len, matmul_id_q6_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_m, "matmul_id_q6_k_f32_aligned_m", matmul_id_q6_k_f32_aligned_fp32_len, matmul_id_q6_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K]->a_s, "matmul_id_q6_k_f32_aligned_s", matmul_id_q6_k_f32_aligned_fp32_len, matmul_id_q6_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->l, "matmul_id_iq4_nl_f32_l", matmul_id_iq4_nl_f32_fp32_len, matmul_id_iq4_nl_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->m, "matmul_id_iq4_nl_f32_m", matmul_id_iq4_nl_f32_fp32_len, matmul_id_iq4_nl_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->s, "matmul_id_iq4_nl_f32_s", matmul_id_iq4_nl_f32_fp32_len, matmul_id_iq4_nl_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_l, "matmul_id_iq4_nl_f32_aligned_l", matmul_id_iq4_nl_f32_aligned_fp32_len, matmul_id_iq4_nl_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_m, "matmul_id_iq4_nl_f32_aligned_m", matmul_id_iq4_nl_f32_aligned_fp32_len, matmul_id_iq4_nl_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), m_wg_denoms, warptile_mmq_m, m_align); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL]->a_s, "matmul_id_iq4_nl_f32_aligned_s", matmul_id_iq4_nl_f32_aligned_fp32_len, matmul_id_iq4_nl_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), s_wg_denoms, warptile_mmq_s, s_align); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q2_K].f32acc, matmul_id_q2_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q3_K].f32acc, matmul_id_q3_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_K].f32acc, matmul_id_q4_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K].f32acc, matmul_id_q5_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q6_K].f32acc, matmul_id_q6_k_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + CREATE_MM(pipeline_dequant_mul_mat_mat_id[GGML_TYPE_IQ4_NL].f32acc, matmul_id_iq4_nl_f32, , mmq_wg_denoms, warptile_mmq, vk_mat_mat_id_push_constants, 4, _id); + } +#undef CREATE_MM } // mul mat vec - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_F32 ], "mul_mat_vec_f32_f32_f32", mul_mat_vec_f32_f32_f32_len, mul_mat_vec_f32_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_F16 ], "mul_mat_vec_f16_f32_f32", mul_mat_vec_f16_f32_f32_len, mul_mat_vec_f16_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_0], "mul_mat_vec_q4_0_f32_f32", mul_mat_vec_q4_0_f32_f32_len, mul_mat_vec_q4_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_1], "mul_mat_vec_q4_1_f32_f32", mul_mat_vec_q4_1_f32_f32_len, mul_mat_vec_q4_1_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_0], "mul_mat_vec_q5_0_f32_f32", mul_mat_vec_q5_0_f32_f32_len, mul_mat_vec_q5_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_1], "mul_mat_vec_q5_1_f32_f32", mul_mat_vec_q5_1_f32_f32_len, mul_mat_vec_q5_1_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q8_0], "mul_mat_vec_q8_0_f32_f32", mul_mat_vec_q8_0_f32_f32_len, mul_mat_vec_q8_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q2_K], "mul_mat_vec_q2_k_f32_f32", mul_mat_vec_q2_k_f32_f32_len, mul_mat_vec_q2_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f32_f32", mul_mat_vec_q3_k_f32_f32_len, mul_mat_vec_q3_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_K], "mul_mat_vec_q4_k_f32_f32", mul_mat_vec_q4_k_f32_f32_len, mul_mat_vec_q4_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_K], "mul_mat_vec_q5_k_f32_f32", mul_mat_vec_q5_k_f32_f32_len, mul_mat_vec_q5_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q6_K], "mul_mat_vec_q6_k_f32_f32", mul_mat_vec_q6_k_f32_f32_len, mul_mat_vec_q6_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f32_f32", mul_mat_vec_iq4_nl_f32_f32_len, mul_mat_vec_iq4_nl_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_F32 ], "mul_mat_vec_f32_f16_f32", mul_mat_vec_f32_f16_f32_len, mul_mat_vec_f32_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_F16 ], "mul_mat_vec_f16_f16_f32", mul_mat_vec_f16_f16_f32_len, mul_mat_vec_f16_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_0], "mul_mat_vec_q4_0_f16_f32", mul_mat_vec_q4_0_f16_f32_len, mul_mat_vec_q4_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_1], "mul_mat_vec_q4_1_f16_f32", mul_mat_vec_q4_1_f16_f32_len, mul_mat_vec_q4_1_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_0], "mul_mat_vec_q5_0_f16_f32", mul_mat_vec_q5_0_f16_f32_len, mul_mat_vec_q5_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_1], "mul_mat_vec_q5_1_f16_f32", mul_mat_vec_q5_1_f16_f32_len, mul_mat_vec_q5_1_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q8_0], "mul_mat_vec_q8_0_f16_f32", mul_mat_vec_q8_0_f16_f32_len, mul_mat_vec_q8_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q2_K], "mul_mat_vec_q2_k_f16_f32", mul_mat_vec_q2_k_f16_f32_len, mul_mat_vec_q2_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f16_f32", mul_mat_vec_q3_k_f16_f32_len, mul_mat_vec_q3_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_K], "mul_mat_vec_q4_k_f16_f32", mul_mat_vec_q4_k_f16_f32_len, mul_mat_vec_q4_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_K], "mul_mat_vec_q5_k_f16_f32", mul_mat_vec_q5_k_f16_f32_len, mul_mat_vec_q5_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q6_K], "mul_mat_vec_q6_k_f16_f32", mul_mat_vec_q6_k_f16_f32_len, mul_mat_vec_q6_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f16_f32", mul_mat_vec_iq4_nl_f16_f32_len, mul_mat_vec_iq4_nl_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); + // the number of rows computed per shader depends on GPU model and quant + uint32_t rm_stdq = 1; + uint32_t rm_kq = 2; + if (device->vendor_id == VK_VENDOR_ID_AMD) { + if (device->subgroup_min_size == 64 && device->subgroup_max_size == 64) { // GCN + rm_stdq = 2; + rm_kq = 4; + } + } else if (device->vendor_id == VK_VENDOR_ID_INTEL) + rm_stdq = 2; - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_F32 ], "mul_mat_vec_id_f32_f32", mul_mat_vec_id_f32_f32_len, mul_mat_vec_id_f32_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_F16 ], "mul_mat_vec_id_f16_f32", mul_mat_vec_id_f16_f32_len, mul_mat_vec_id_f16_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_0], "mul_mat_vec_id_q4_0_f32", mul_mat_vec_id_q4_0_f32_len, mul_mat_vec_id_q4_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_1], "mul_mat_vec_id_q4_1_f32", mul_mat_vec_id_q4_1_f32_len, mul_mat_vec_id_q4_1_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_0], "mul_mat_vec_id_q5_0_f32", mul_mat_vec_id_q5_0_f32_len, mul_mat_vec_id_q5_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_1], "mul_mat_vec_id_q5_1_f32", mul_mat_vec_id_q5_1_f32_len, mul_mat_vec_id_q5_1_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q8_0], "mul_mat_vec_id_q8_0_f32", mul_mat_vec_id_q8_0_f32_len, mul_mat_vec_id_q8_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q2_K], "mul_mat_vec_id_q2_k_f32", mul_mat_vec_id_q2_k_f32_len, mul_mat_vec_id_q2_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q3_K], "mul_mat_vec_id_q3_k_f32", mul_mat_vec_id_q3_k_f32_len, mul_mat_vec_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_K], "mul_mat_vec_id_q4_k_f32", mul_mat_vec_id_q4_k_f32_len, mul_mat_vec_id_q4_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_K], "mul_mat_vec_id_q5_k_f32", mul_mat_vec_id_q5_k_f32_len, mul_mat_vec_id_q5_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q6_K], "mul_mat_vec_id_q6_k_f32", mul_mat_vec_id_q6_k_f32_len, mul_mat_vec_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_id_iq4_nl_f32", mul_mat_vec_id_iq4_nl_f32_len, mul_mat_vec_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); + for (uint32_t i = 0; i < mul_mat_vec_max_cols; ++i) { + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_F32 ][i], "mul_mat_vec_f32_f32_f32_"+std::to_string(i+1), mul_mat_vec_f32_f32_f32_len, mul_mat_vec_f32_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2, 1, 1}, {device->subgroup_size, 2, i+1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_F16 ][i], "mul_mat_vec_f16_f32_f32_"+std::to_string(i+1), mul_mat_vec_f16_f32_f32_len, mul_mat_vec_f16_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2, 1, 1}, {device->subgroup_size, 2, i+1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_0][i], "mul_mat_vec_q4_0_f32_f32_"+std::to_string(i+1), mul_mat_vec_q4_0_f32_f32_len, mul_mat_vec_q4_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_1][i], "mul_mat_vec_q4_1_f32_f32_"+std::to_string(i+1), mul_mat_vec_q4_1_f32_f32_len, mul_mat_vec_q4_1_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_0][i], "mul_mat_vec_q5_0_f32_f32_"+std::to_string(i+1), mul_mat_vec_q5_0_f32_f32_len, mul_mat_vec_q5_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_1][i], "mul_mat_vec_q5_1_f32_f32_"+std::to_string(i+1), mul_mat_vec_q5_1_f32_f32_len, mul_mat_vec_q5_1_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q8_0][i], "mul_mat_vec_q8_0_f32_f32_"+std::to_string(i+1), mul_mat_vec_q8_0_f32_f32_len, mul_mat_vec_q8_0_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1*rm_stdq, 1, 1}, {device->subgroup_size, 1*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q2_K][i], "mul_mat_vec_q2_k_f32_f32_"+std::to_string(i+1), mul_mat_vec_q2_k_f32_f32_len, mul_mat_vec_q2_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q3_K][i], "mul_mat_vec_q3_k_f32_f32_"+std::to_string(i+1), mul_mat_vec_q3_k_f32_f32_len, mul_mat_vec_q3_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q4_K][i], "mul_mat_vec_q4_k_f32_f32_"+std::to_string(i+1), mul_mat_vec_q4_k_f32_f32_len, mul_mat_vec_q4_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q5_K][i], "mul_mat_vec_q5_k_f32_f32_"+std::to_string(i+1), mul_mat_vec_q5_k_f32_f32_len, mul_mat_vec_q5_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q6_K][i], "mul_mat_vec_q6_k_f32_f32_"+std::to_string(i+1), mul_mat_vec_q6_k_f32_f32_len, mul_mat_vec_q6_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_IQ4_NL][i], "mul_mat_vec_iq4_nl_f32_f32_"+std::to_string(i+1), mul_mat_vec_iq4_nl_f32_f32_len, mul_mat_vec_iq4_nl_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {subgroup_size_16, 2*rm_stdq, i+1}, 1, true); + + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_F32 ][i], "mul_mat_vec_f32_f16_f32_"+std::to_string(i+1), mul_mat_vec_f32_f16_f32_len, mul_mat_vec_f32_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2, 1, 1}, {device->subgroup_size, 2, i+1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_F16 ][i], "mul_mat_vec_f16_f16_f32_"+std::to_string(i+1), mul_mat_vec_f16_f16_f32_len, mul_mat_vec_f16_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2, 1, 1}, {device->subgroup_size, 2, i+1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_0][i], "mul_mat_vec_q4_0_f16_f32_"+std::to_string(i+1), mul_mat_vec_q4_0_f16_f32_len, mul_mat_vec_q4_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_1][i], "mul_mat_vec_q4_1_f16_f32_"+std::to_string(i+1), mul_mat_vec_q4_1_f16_f32_len, mul_mat_vec_q4_1_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_0][i], "mul_mat_vec_q5_0_f16_f32_"+std::to_string(i+1), mul_mat_vec_q5_0_f16_f32_len, mul_mat_vec_q5_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_1][i], "mul_mat_vec_q5_1_f16_f32_"+std::to_string(i+1), mul_mat_vec_q5_1_f16_f32_len, mul_mat_vec_q5_1_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q8_0][i], "mul_mat_vec_q8_0_f16_f32_"+std::to_string(i+1), mul_mat_vec_q8_0_f16_f32_len, mul_mat_vec_q8_0_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1*rm_stdq, 1, 1}, {device->subgroup_size, 1*rm_stdq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q2_K][i], "mul_mat_vec_q2_k_f16_f32_"+std::to_string(i+1), mul_mat_vec_q2_k_f16_f32_len, mul_mat_vec_q2_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q3_K][i], "mul_mat_vec_q3_k_f16_f32_"+std::to_string(i+1), mul_mat_vec_q3_k_f16_f32_len, mul_mat_vec_q3_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q4_K][i], "mul_mat_vec_q4_k_f16_f32_"+std::to_string(i+1), mul_mat_vec_q4_k_f16_f32_len, mul_mat_vec_q4_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q5_K][i], "mul_mat_vec_q5_k_f16_f32_"+std::to_string(i+1), mul_mat_vec_q5_k_f16_f32_len, mul_mat_vec_q5_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q6_K][i], "mul_mat_vec_q6_k_f16_f32_"+std::to_string(i+1), mul_mat_vec_q6_k_f16_f32_len, mul_mat_vec_q6_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq, i+1}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_IQ4_NL][i], "mul_mat_vec_iq4_nl_f16_f32_"+std::to_string(i+1), mul_mat_vec_iq4_nl_f16_f32_len, mul_mat_vec_iq4_nl_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {2*rm_stdq, 1, 1}, {subgroup_size_16, 2*rm_stdq, i+1}, 1, true); + } + + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_F32 ], "mul_mat_vec_id_f32_f32", mul_mat_vec_id_f32_f32_len, mul_mat_vec_id_f32_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2, 1, 1}, {device->subgroup_size, 2}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_F16 ], "mul_mat_vec_id_f16_f32", mul_mat_vec_id_f16_f32_len, mul_mat_vec_id_f16_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2, 1, 1}, {device->subgroup_size, 2}, 1); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_0], "mul_mat_vec_id_q4_0_f32", mul_mat_vec_id_q4_0_f32_len, mul_mat_vec_id_q4_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_1], "mul_mat_vec_id_q4_1_f32", mul_mat_vec_id_q4_1_f32_len, mul_mat_vec_id_q4_1_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_0], "mul_mat_vec_id_q5_0_f32", mul_mat_vec_id_q5_0_f32_len, mul_mat_vec_id_q5_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_1], "mul_mat_vec_id_q5_1_f32", mul_mat_vec_id_q5_1_f32_len, mul_mat_vec_id_q5_1_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2*rm_stdq, 1, 1}, {device->subgroup_size, 2*rm_stdq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q8_0], "mul_mat_vec_id_q8_0_f32", mul_mat_vec_id_q8_0_f32_len, mul_mat_vec_id_q8_0_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1*rm_stdq, 1, 1}, {device->subgroup_size, 1*rm_stdq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q2_K], "mul_mat_vec_id_q2_k_f32", mul_mat_vec_id_q2_k_f32_len, mul_mat_vec_id_q2_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q3_K], "mul_mat_vec_id_q3_k_f32", mul_mat_vec_id_q3_k_f32_len, mul_mat_vec_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q4_K], "mul_mat_vec_id_q4_k_f32", mul_mat_vec_id_q4_k_f32_len, mul_mat_vec_id_q4_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q5_K], "mul_mat_vec_id_q5_k_f32", mul_mat_vec_id_q5_k_f32_len, mul_mat_vec_id_q5_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q6_K], "mul_mat_vec_id_q6_k_f32", mul_mat_vec_id_q6_k_f32_len, mul_mat_vec_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {rm_kq, 1, 1}, {subgroup_size_16, rm_kq}, 1, true); + ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_id_iq4_nl_f32", mul_mat_vec_id_iq4_nl_f32_len, mul_mat_vec_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {2*rm_stdq, 1, 1}, {subgroup_size_16, 2*rm_stdq}, 1, true); // dequant shaders ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_F32 ], "f32_to_f16", dequant_f32_len, dequant_f32_data, "main", 2, 5 * sizeof(uint32_t), {256 * 16, 1, 1}, {}, 1); @@ -1746,7 +1948,7 @@ static void ggml_vk_load_shaders(vk_device& device) { ggml_vk_create_pipeline(device, device->pipeline_get_rows_f32[GGML_TYPE_Q8_0], "get_rows_q8_0_f32", get_rows_q8_0_f32_len, get_rows_q8_0_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {1024, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_get_rows_f32[GGML_TYPE_IQ4_NL], "get_rows_iq4_nl_f32", get_rows_iq4_nl_f32_len, get_rows_iq4_nl_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {1024, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_matmul_split_k_reduce, "split_k_reduce", split_k_reduce_len, split_k_reduce_data, "main", 2, 2 * sizeof(uint32_t), {256, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_matmul_split_k_reduce, "split_k_reduce", split_k_reduce_len, split_k_reduce_data, "main", 2, 2 * sizeof(uint32_t), {256 * 4, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_mul_mat_vec_p021_f16_f32, "mul_mat_vec_p021_f16_f32", mul_mat_vec_p021_f16_f32_len, mul_mat_vec_p021_f16_f32_data, "main", 3, 6 * sizeof(uint32_t), {1, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_mul_mat_vec_nc_f16_f32, "mul_mat_vec_nc_f16_f32", mul_mat_vec_nc_f16_f32_len, mul_mat_vec_nc_f16_f32_data, "main", 3, 7 * sizeof(uint32_t), {1, 1, 1}, {}, 1); @@ -1759,13 +1961,21 @@ static void ggml_vk_load_shaders(vk_device& device) { ggml_vk_create_pipeline(device, device->pipeline_cpy_f32_f16, "cpy_f32_f16", cpy_f32_f16_len, cpy_f32_f16_data, "main", 2, sizeof(vk_op_unary_push_constants), {512, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_cpy_f16_f16, "cpy_f16_f16", cpy_f16_f16_len, cpy_f16_f16_data, "main", 2, sizeof(vk_op_unary_push_constants), {512, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_add_f32, "add_f32", add_f32_len, add_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_add_f16_f32_f16, "add_f16_f32_f16", add_f16_f32_f16_len, add_f16_f32_f16_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_contig_cpy_f32_f32, "contig_cpy_f32_f32", contig_cpy_f32_f32_len, contig_cpy_f32_f32_data, "main", 2, sizeof(vk_op_unary_push_constants), {512, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_contig_cpy_f32_f16, "contig_cpy_f32_f16", contig_cpy_f32_f16_len, contig_cpy_f32_f16_data, "main", 2, sizeof(vk_op_unary_push_constants), {512, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_contig_cpy_f16_f16, "contig_cpy_f16_f16", contig_cpy_f16_f16_len, contig_cpy_f16_f16_data, "main", 2, sizeof(vk_op_unary_push_constants), {512, 1, 1}, {}, 1); + + ggml_vk_create_pipeline(device, device->pipeline_add_f32, "add_f32", add_f32_len, add_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {0}, 1); + ggml_vk_create_pipeline(device, device->pipeline_add_f32_norepeat, "add_f32_norepeat", add_f32_len, add_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_add_f16_f32_f16, "add_f16_f32_f16", add_f16_f32_f16_len, add_f16_f32_f16_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {0}, 1); + ggml_vk_create_pipeline(device, device->pipeline_add_f16_f32_f16_norepeat, "add_f16_f32_f16_norepeat", add_f16_f32_f16_len, add_f16_f32_f16_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {1}, 1); ggml_vk_create_pipeline(device, device->pipeline_acc_f32, "acc_f32", acc_f32_len, acc_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_mul_f32, "mul_f32", mul_f32_len, mul_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_div_f32, "div_f32", div_f32_len, div_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_mul_f32, "mul_f32", mul_f32_len, mul_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {0}, 1); + ggml_vk_create_pipeline(device, device->pipeline_mul_f32_norepeat, "mul_f32_norepeat", mul_f32_len, mul_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {1}, 1); + ggml_vk_create_pipeline(device, device->pipeline_div_f32, "div_f32", div_f32_len, div_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {0}, 1); + ggml_vk_create_pipeline(device, device->pipeline_div_f32_norepeat, "div_f32_norepeat", div_f32_len, div_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {1}, 1); ggml_vk_create_pipeline(device, device->pipeline_concat_f32, "concat_f32", concat_f32_len, concat_f32_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_concat_f16, "concat_f16", concat_f16_len, concat_f16_data, "main", 3, sizeof(vk_op_binary_push_constants), {512, 1, 1}, {}, 1); @@ -1794,31 +2004,47 @@ static void ggml_vk_load_shaders(vk_device& device) { ggml_vk_create_pipeline(device, device->pipeline_diag_mask_inf_f32, "diag_mask_inf_f32", diag_mask_inf_f32_len, diag_mask_inf_f32_data, "main", 2, sizeof(vk_op_diag_mask_push_constants), {512, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32, "soft_max_f32", soft_max_f32_len, soft_max_f32_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32_f16, "soft_max_f32_f16", soft_max_f32_f16_len, soft_max_f32_f16_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32, "soft_max_f32", soft_max_f32_len, soft_max_f32_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); + ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32_wg512, "soft_max_f32_wg512", soft_max_f32_len, soft_max_f32_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, { 512 }, 1); + ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32_f16, "soft_max_f32_f16", soft_max_f32_f16_len, soft_max_f32_f16_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); + ggml_vk_create_pipeline(device, device->pipeline_soft_max_f32_f16_wg512, "soft_max_f32_f16_wg512", soft_max_f32_f16_len, soft_max_f32_f16_data, "main", 3, sizeof(vk_op_soft_max_push_constants), {1, 1, 1}, { 512 }, 1); ggml_vk_create_pipeline(device, device->pipeline_rope_norm_f32, "rope_norm_f32", rope_norm_f32_len, rope_norm_f32_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_rope_norm_f16, "rope_norm_f16", rope_norm_f16_len, rope_norm_f16_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_rope_neox_f32, "rope_neox_f32", rope_neox_f32_len, rope_neox_f32_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_rope_neox_f16, "rope_neox_f16", rope_neox_f16_len, rope_neox_f16_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); + + if (device->float_controls_rte_fp16) { + ggml_vk_create_pipeline(device, device->pipeline_rope_norm_f16, "rope_norm_f16", rope_norm_f16_rte_len, rope_norm_f16_rte_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_rope_neox_f16, "rope_neox_f16", rope_neox_f16_rte_len, rope_neox_f16_rte_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); + } else { + ggml_vk_create_pipeline(device, device->pipeline_rope_norm_f16, "rope_norm_f16", rope_norm_f16_len, rope_norm_f16_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_rope_neox_f16, "rope_neox_f16", rope_neox_f16_len, rope_neox_f16_data, "main", 4, sizeof(vk_op_rope_push_constants), {1, 512, 1}, {}, 1); + } ggml_vk_create_pipeline(device, device->pipeline_argsort_f32, "argsort_f32", argsort_f32_len, argsort_f32_data, "main", 2, sizeof(vk_op_argsort_push_constants), {1024, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_sum_rows_f32, "sum_rows_f32", sum_rows_f32_len, sum_rows_f32_data, "main", 2, sizeof(vk_op_push_constants), {1, 1, 1}, { device->subgroup_size }, 1); - ggml_vk_create_pipeline(device, device->pipeline_im2col_f32, "im2col_f32", im2col_f32_len, im2col_f32_data, "main", 2, sizeof(vk_op_im2col_push_constants), {256, 1, 1}, {}, 1); - ggml_vk_create_pipeline(device, device->pipeline_im2col_f32_f16, "im2col_f32_f16", im2col_f32_f16_len, im2col_f32_f16_data, "main", 2, sizeof(vk_op_im2col_push_constants), {256, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_im2col_f32, "im2col_f32", im2col_f32_len, im2col_f32_data, "main", 2, sizeof(vk_op_im2col_push_constants), {512, 1, 1}, { device->subgroup_size }, 1, true); + if (device->float_controls_rte_fp16) { + ggml_vk_create_pipeline(device, device->pipeline_im2col_f32_f16, "im2col_f32_f16", im2col_f32_f16_rte_len, im2col_f32_f16_rte_data, "main", 2, sizeof(vk_op_im2col_push_constants), {512, 1, 1}, { device->subgroup_size }, 1, true); + } else { + ggml_vk_create_pipeline(device, device->pipeline_im2col_f32_f16, "im2col_f32_f16", im2col_f32_f16_len, im2col_f32_f16_data, "main", 2, sizeof(vk_op_im2col_push_constants), {512, 1, 1}, { device->subgroup_size }, 1, true); + } ggml_vk_create_pipeline(device, device->pipeline_timestep_embedding_f32, "timestep_embedding_f32", timestep_embedding_f32_len, timestep_embedding_f32_data, "main", 2, sizeof(vk_op_timestep_embedding_push_constants), {256, 1, 1}, {}, 1); ggml_vk_create_pipeline(device, device->pipeline_pool2d_f32, "pool2d_f32", pool2d_f32_len, pool2d_f32_data, "main", 2, sizeof(vk_op_pool2d_push_constants), {512, 1, 1}, {}, 1); + ggml_vk_create_pipeline(device, device->pipeline_rwkv_wkv6_f32, "rwkv_wkv6_f32", rwkv_wkv6_f32_len, rwkv_wkv6_f32_data, "main", 7, sizeof(vk_op_rwkv_wkv6_push_constants), {1, 1, 1}, {device->subgroup_size}, 1); + for (auto &c : compiles) { c.wait(); } + std::cerr << "Done!" << std::endl; } +static bool ggml_vk_khr_cooperative_matrix_support(const vk::PhysicalDeviceProperties& props, const vk::PhysicalDeviceDriverProperties& driver_props); + static vk_device ggml_vk_get_device(size_t idx) { VK_LOG_DEBUG("ggml_vk_get_device(" << idx << ")"); @@ -1846,12 +2072,40 @@ static vk_device ggml_vk_get_device(size_t idx) { device->physical_device = physical_devices[dev_num]; const std::vector ext_props = device->physical_device.enumerateDeviceExtensionProperties(); + bool fp16_storage = false; + bool fp16_compute = false; bool maintenance4_support = false; + bool sm_builtins = false; + bool amd_shader_core_properties2 = false; + bool pipeline_robustness = false; + bool coopmat2_support = false; + device->coopmat_support = false; // Check if maintenance4 is supported for (const auto& properties : ext_props) { if (strcmp("VK_KHR_maintenance4", properties.extensionName) == 0) { maintenance4_support = true; + } else if (strcmp("VK_KHR_16bit_storage", properties.extensionName) == 0) { + fp16_storage = true; + } else if (strcmp("VK_KHR_shader_float16_int8", properties.extensionName) == 0) { + fp16_compute = true; + } else if (strcmp("VK_NV_shader_sm_builtins", properties.extensionName) == 0) { + sm_builtins = true; + } else if (strcmp("VK_AMD_shader_core_properties2", properties.extensionName) == 0) { + amd_shader_core_properties2 = true; + } else if (strcmp("VK_EXT_pipeline_robustness", properties.extensionName) == 0) { + pipeline_robustness = true; + } else if (strcmp("VK_EXT_subgroup_size_control", properties.extensionName) == 0) { + device->subgroup_size_control = true; + } else if (strcmp("VK_KHR_cooperative_matrix", properties.extensionName) == 0 && + !getenv("GGML_VK_DISABLE_COOPMAT")) { + device->coopmat_support = true; + device->coopmat_m = 0; + device->coopmat_n = 0; + device->coopmat_k = 0; + } else if (strcmp("VK_NV_cooperative_matrix2", properties.extensionName) == 0 && + !getenv("GGML_VK_DISABLE_COOPMAT2")) { + coopmat2_support = true; } } @@ -1859,18 +2113,51 @@ static vk_device ggml_vk_get_device(size_t idx) { vk::PhysicalDeviceMaintenance3Properties props3; vk::PhysicalDeviceMaintenance4Properties props4; vk::PhysicalDeviceSubgroupProperties subgroup_props; + vk::PhysicalDeviceDriverProperties driver_props; + vk::PhysicalDeviceShaderSMBuiltinsPropertiesNV sm_props; + vk::PhysicalDeviceShaderCoreProperties2AMD amd_shader_core_properties2_props; + vk::PhysicalDeviceVulkan12Properties vk12_props; + vk::PhysicalDeviceSubgroupSizeControlPropertiesEXT subgroup_size_control_props; + props2.pNext = &props3; props3.pNext = &subgroup_props; + subgroup_props.pNext = &driver_props; + driver_props.pNext = &vk12_props; + + VkBaseOutStructure * last_struct = (VkBaseOutStructure *)&vk12_props; + if (maintenance4_support) { - subgroup_props.pNext = &props4; + last_struct->pNext = (VkBaseOutStructure *)&props4; + last_struct = (VkBaseOutStructure *)&props4; } + if (sm_builtins) { + last_struct->pNext = (VkBaseOutStructure *)&sm_props; + last_struct = (VkBaseOutStructure *)&sm_props; + } + if (amd_shader_core_properties2) { + last_struct->pNext = (VkBaseOutStructure *)&amd_shader_core_properties2_props; + last_struct = (VkBaseOutStructure *)&amd_shader_core_properties2_props; + } + if (device->subgroup_size_control) { + last_struct->pNext = (VkBaseOutStructure *)&subgroup_size_control_props; + last_struct = (VkBaseOutStructure *)&subgroup_size_control_props; + } + +#if defined(VK_NV_cooperative_matrix2) + vk::PhysicalDeviceCooperativeMatrix2PropertiesNV coopmat2_props; + if (coopmat2_support) { + last_struct->pNext = (VkBaseOutStructure *)&coopmat2_props; + last_struct = (VkBaseOutStructure *)&coopmat2_props; + } +#endif + device->physical_device.getProperties2(&props2); device->properties = props2.properties; const char* GGML_VK_FORCE_MAX_ALLOCATION_SIZE = getenv("GGML_VK_FORCE_MAX_ALLOCATION_SIZE"); if (GGML_VK_FORCE_MAX_ALLOCATION_SIZE != nullptr) { - device->max_memory_allocation_size = std::stoi(GGML_VK_FORCE_MAX_ALLOCATION_SIZE); + device->max_memory_allocation_size = std::stoul(GGML_VK_FORCE_MAX_ALLOCATION_SIZE); } else if (maintenance4_support) { device->max_memory_allocation_size = std::min(props3.maxMemoryAllocationSize, props4.maxBufferSize); } else { @@ -1880,23 +2167,23 @@ static vk_device ggml_vk_get_device(size_t idx) { device->vendor_id = device->properties.vendorID; device->subgroup_size = subgroup_props.subgroupSize; device->uma = device->properties.deviceType == vk::PhysicalDeviceType::eIntegratedGpu; - - bool fp16_storage = false; - bool fp16_compute = false; - - for (const auto& properties : ext_props) { - if (strcmp("VK_KHR_16bit_storage", properties.extensionName) == 0) { - fp16_storage = true; - } else if (strcmp("VK_KHR_shader_float16_int8", properties.extensionName) == 0) { - fp16_compute = true; - } + if (sm_builtins) { + device->shader_core_count = sm_props.shaderSMCount; + } else if (amd_shader_core_properties2) { + device->shader_core_count = amd_shader_core_properties2_props.activeComputeUnitCount; + } else { + device->shader_core_count = 0; } + device->float_controls_rte_fp16 = vk12_props.shaderRoundingModeRTEFloat16; - const char* GGML_VK_DISABLE_F16 = getenv("GGML_VK_DISABLE_F16"); - const bool force_disable_f16 = GGML_VK_DISABLE_F16 != nullptr; + const bool force_disable_f16 = getenv("GGML_VK_DISABLE_F16") != nullptr; device->fp16 = !force_disable_f16 && fp16_storage && fp16_compute; + if (!ggml_vk_khr_cooperative_matrix_support(device->properties, driver_props)) { + device->coopmat_support = false; + } + std::vector queue_family_props = device->physical_device.getQueueFamilyProperties(); // Try to find a non-graphics compute queue and transfer-focused queues @@ -1934,10 +2221,153 @@ static vk_device ggml_vk_get_device(size_t idx) { vk12_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_2_FEATURES; vk11_features.pNext = &vk12_features; + last_struct = (VkBaseOutStructure *)&vk12_features; + + VkPhysicalDevicePipelineRobustnessFeaturesEXT pl_robustness_features; + pl_robustness_features.pNext = nullptr; + pl_robustness_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PIPELINE_ROBUSTNESS_FEATURES_EXT; + pl_robustness_features.pipelineRobustness = VK_FALSE; + + if (pipeline_robustness) { + last_struct->pNext = (VkBaseOutStructure *)&pl_robustness_features; + last_struct = (VkBaseOutStructure *)&pl_robustness_features; + device_extensions.push_back("VK_EXT_pipeline_robustness"); + } + + VkPhysicalDeviceSubgroupSizeControlFeaturesEXT subgroup_size_control_features; + subgroup_size_control_features.pNext = nullptr; + subgroup_size_control_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_SIZE_CONTROL_FEATURES_EXT; + subgroup_size_control_features.computeFullSubgroups = false; + subgroup_size_control_features.subgroupSizeControl = false; + + if (device->subgroup_size_control) { + last_struct->pNext = (VkBaseOutStructure *)&subgroup_size_control_features; + last_struct = (VkBaseOutStructure *)&subgroup_size_control_features; + } + +#if defined(VK_KHR_cooperative_matrix) + VkPhysicalDeviceCooperativeMatrixFeaturesKHR coopmat_features; + coopmat_features.pNext = nullptr; + coopmat_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_COOPERATIVE_MATRIX_FEATURES_KHR; + coopmat_features.cooperativeMatrix = VK_FALSE; + + if (device->coopmat_support) { + last_struct->pNext = (VkBaseOutStructure *)&coopmat_features; + last_struct = (VkBaseOutStructure *)&coopmat_features; + } +#endif + +#if defined(VK_NV_cooperative_matrix2) + VkPhysicalDeviceCooperativeMatrix2FeaturesNV coopmat2_features {}; + coopmat2_features.pNext = nullptr; + coopmat2_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_COOPERATIVE_MATRIX_2_FEATURES_NV; + if (coopmat2_support) { + last_struct->pNext = (VkBaseOutStructure *)&coopmat2_features; + last_struct = (VkBaseOutStructure *)&coopmat2_features; + device_extensions.push_back("VK_NV_cooperative_matrix2"); + } +#endif + vkGetPhysicalDeviceFeatures2(device->physical_device, &device_features2); device->fp16 = device->fp16 && vk12_features.shaderFloat16; + device->pipeline_robustness = pl_robustness_features.pipelineRobustness; + + if (device->subgroup_size_control) { + device->subgroup_min_size = subgroup_size_control_props.minSubgroupSize; + device->subgroup_max_size = subgroup_size_control_props.maxSubgroupSize; + device_extensions.push_back("VK_EXT_subgroup_size_control"); + } + + device->subgroup_size_control = device->subgroup_size_control && + (subgroup_size_control_props.requiredSubgroupSizeStages & vk::ShaderStageFlagBits::eCompute) && + subgroup_size_control_features.subgroupSizeControl; + + if (device->subgroup_size_control) { + device->subgroup_require_full_support = subgroup_size_control_features.computeFullSubgroups; + } + +#if defined(VK_KHR_cooperative_matrix) + device->coopmat_support = device->coopmat_support && coopmat_features.cooperativeMatrix; +#endif + + if (coopmat2_support) { +#if defined(VK_NV_cooperative_matrix2) && defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + if (coopmat2_features.cooperativeMatrixWorkgroupScope && + coopmat2_features.cooperativeMatrixFlexibleDimensions && + coopmat2_features.cooperativeMatrixReductions && + coopmat2_features.cooperativeMatrixConversions && + coopmat2_features.cooperativeMatrixPerElementOperations && + coopmat2_features.cooperativeMatrixTensorAddressing && + coopmat2_features.cooperativeMatrixBlockLoads && + vk12_features.bufferDeviceAddress) { + + std::vector flexible_dimensions; + uint32_t count = 0; + + PFN_vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV + _vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV = + (PFN_vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV) + vk_instance.instance.getProcAddr("vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV"); + + _vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV(device->physical_device, &count, nullptr); + + VkCooperativeMatrixFlexibleDimensionsPropertiesNV empty_prop {}; + empty_prop.sType = VK_STRUCTURE_TYPE_COOPERATIVE_MATRIX_FLEXIBLE_DIMENSIONS_PROPERTIES_NV; + flexible_dimensions.resize(count, empty_prop); + + _vkGetPhysicalDeviceCooperativeMatrixFlexibleDimensionsPropertiesNV(device->physical_device, &count, flexible_dimensions.data()); + + bool found_fp16_128 = false, + found_fp16_256 = false, + found_fp32_128 = false, + found_fp32_256 = false; + // need to support fp16*fp16 with fp16/fp32 accumulator, for workgroupsize 128 + // with 32x16x16 and 256 with 32x32x16. + for (auto &prop : flexible_dimensions) { + if (prop.saturatingAccumulation == VK_FALSE && + prop.scope == VK_SCOPE_WORKGROUP_KHR && + prop.AType == VK_COMPONENT_TYPE_FLOAT16_KHR && + prop.BType == VK_COMPONENT_TYPE_FLOAT16_KHR) { + + if (prop.workgroupInvocations == 128 && + prop.MGranularity <= 32 && + prop.NGranularity <= 16 && + prop.KGranularity <= 16) { + if (prop.CType == VK_COMPONENT_TYPE_FLOAT16_KHR && + prop.ResultType == VK_COMPONENT_TYPE_FLOAT16_KHR) { + found_fp16_128 = true; + } + if (prop.CType == VK_COMPONENT_TYPE_FLOAT32_KHR && + prop.ResultType == VK_COMPONENT_TYPE_FLOAT32_KHR) { + found_fp32_128 = true; + } + } + if (prop.workgroupInvocations == 256 && + prop.MGranularity <= 32 && + prop.NGranularity <= 32 && + prop.KGranularity <= 16) { + if (prop.CType == VK_COMPONENT_TYPE_FLOAT16_KHR && + prop.ResultType == VK_COMPONENT_TYPE_FLOAT16_KHR) { + found_fp16_256 = true; + } + if (prop.CType == VK_COMPONENT_TYPE_FLOAT32_KHR && + prop.ResultType == VK_COMPONENT_TYPE_FLOAT32_KHR) { + found_fp32_256 = true; + } + } + } + } + if (found_fp16_128 && found_fp16_256 && + found_fp32_128 && found_fp32_256 && + coopmat2_props.cooperativeMatrixFlexibleDimensionsMaxDimension >= 512) { + device->coopmat2 = true; + } + } +#endif + } + if (!vk11_features.storageBuffer16BitAccess) { std::cerr << "ggml_vulkan: device " << GGML_VK_NAME << idx << " does not support 16-bit storage." << std::endl; throw std::runtime_error("Unsupported device"); @@ -1952,6 +2382,75 @@ static vk_device ggml_vk_get_device(size_t idx) { if (device->fp16) { device_extensions.push_back("VK_KHR_shader_float16_int8"); } + +#if defined(VK_KHR_cooperative_matrix) + if (device->coopmat_support) { + // Query supported shapes + std::vector cm_props; + + PFN_vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR pfn_vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR = + (PFN_vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR)vkGetInstanceProcAddr(vk_instance.instance, "vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR"); + + uint32_t cm_props_num; + + pfn_vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR(device->physical_device, &cm_props_num, nullptr); + + cm_props.resize(cm_props_num); + + for (auto& prop : cm_props) { + prop.sType = VK_STRUCTURE_TYPE_COOPERATIVE_MATRIX_PROPERTIES_KHR; + } + + pfn_vkGetPhysicalDeviceCooperativeMatrixPropertiesKHR(device->physical_device, &cm_props_num, cm_props.data()); + + VK_LOG_DEBUG("ggml_vulkan: Cooperative Matrix Shapes: " << cm_props.size()); + + for (auto& prop : cm_props) { + VK_LOG_DEBUG("ggml_vulkan: M: " << prop.MSize << " N: " << prop.NSize << " K: " << prop.KSize << " A: " << vk::to_string((vk::ComponentTypeKHR)prop.AType) << " B: " << vk::to_string((vk::ComponentTypeKHR)prop.BType) << " C: " << vk::to_string((vk::ComponentTypeKHR)prop.CType) << " Result: " << vk::to_string((vk::ComponentTypeKHR)prop.ResultType) << " saturatingAccumulation: " << prop.saturatingAccumulation << " scope: " << vk::to_string((vk::ScopeKHR)prop.scope)); + + if ((vk::ComponentTypeKHR)prop.AType == vk::ComponentTypeKHR::eFloat16 && + (vk::ComponentTypeKHR)prop.BType == vk::ComponentTypeKHR::eFloat16 && + (vk::ScopeKHR)prop.scope == vk::ScopeKHR::eSubgroup + ) { + if ((vk::ComponentTypeKHR)prop.CType == vk::ComponentTypeKHR::eFloat32 && + (vk::ComponentTypeKHR)prop.ResultType == vk::ComponentTypeKHR::eFloat32) { + // coopmat sizes not set yet + if (device->coopmat_m == 0) { + device->coopmat_acc_f32_support = true; + device->coopmat_m = prop.MSize; + device->coopmat_n = prop.NSize; + device->coopmat_k = prop.KSize; + } else if (device->coopmat_m == prop.MSize && device->coopmat_n == prop.NSize && device->coopmat_k == prop.KSize) { + // Only enable if shape is identical + device->coopmat_acc_f32_support = true; + } + } else if ((vk::ComponentTypeKHR)prop.CType == vk::ComponentTypeKHR::eFloat16 && + (vk::ComponentTypeKHR)prop.ResultType == vk::ComponentTypeKHR::eFloat16) { + // coopmat sizes not set yet + if (device->coopmat_m == 0) { + device->coopmat_acc_f16_support = true; + device->coopmat_m = prop.MSize; + device->coopmat_n = prop.NSize; + device->coopmat_k = prop.KSize; + } else if (device->coopmat_m == prop.MSize && device->coopmat_n == prop.NSize && device->coopmat_k == prop.KSize) { + // Only enable if shape is identical + device->coopmat_acc_f16_support = true; + } + } + } + } + + if (device->coopmat_m == 0 || !device->coopmat_acc_f32_support) { + // No suitable matmul mode found + GGML_LOG_DEBUG("ggml_vulkan: WARNING: No suitable matrix core mode found. Disabling matrix cores.\n"); + device->coopmat_support = false; + } + } + + if (device->coopmat_support) { + device_extensions.push_back("VK_KHR_cooperative_matrix"); + } +#endif device->name = GGML_VK_NAME + std::to_string(idx); device_create_info = { @@ -1967,6 +2466,37 @@ static vk_device ggml_vk_get_device(size_t idx) { ggml_vk_create_queue(device, device->compute_queue, compute_queue_family_index, 0, { vk::PipelineStageFlagBits::eComputeShader | vk::PipelineStageFlagBits::eTransfer }, false); // Shaders + // Disable matmul tile sizes early if performance low or not supported + switch (device->vendor_id) { +#ifndef GGML_VULKAN_RUN_TESTS + case VK_VENDOR_ID_AMD: + case VK_VENDOR_ID_INTEL: + device->mul_mat_l = false; + device->mul_mat_m = true; + device->mul_mat_s = true; + device->mul_mat_id_l = false; + device->mul_mat_id_m = true; + device->mul_mat_id_s = true; + break; + case VK_VENDOR_ID_APPLE: + device->mul_mat_l = false; + device->mul_mat_m = true; + device->mul_mat_s = false; + device->mul_mat_id_l = false; + device->mul_mat_id_m = true; + device->mul_mat_id_s = false; + break; +#endif + default: + device->mul_mat_l = true; + device->mul_mat_m = true; + device->mul_mat_s = true; + device->mul_mat_id_l = true; + device->mul_mat_id_m = true; + device->mul_mat_id_s = true; + break; + } + ggml_vk_load_shaders(device); if (!device->single_queue) { @@ -1993,7 +2523,6 @@ static vk_device ggml_vk_get_device(size_t idx) { return vk_instance.devices[idx]; } - static void ggml_vk_print_gpu_info(size_t idx) { GGML_ASSERT(idx < vk_instance.device_indices.size()); size_t dev_num = vk_instance.device_indices[idx]; @@ -2024,15 +2553,31 @@ static void ggml_vk_print_gpu_info(size_t idx) { bool fp16_storage = false; bool fp16_compute = false; + bool coopmat_support = false; + bool coopmat2_support = false; for (auto properties : ext_props) { if (strcmp("VK_KHR_16bit_storage", properties.extensionName) == 0) { fp16_storage = true; } else if (strcmp("VK_KHR_shader_float16_int8", properties.extensionName) == 0) { fp16_compute = true; +#if defined(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) + } else if (strcmp("VK_KHR_cooperative_matrix", properties.extensionName) == 0 && + !getenv("GGML_VK_DISABLE_COOPMAT")) { + coopmat_support = true; +#endif +#if defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + } else if (strcmp("VK_NV_cooperative_matrix2", properties.extensionName) == 0 && + !getenv("GGML_VK_DISABLE_COOPMAT2")) { + coopmat2_support = true; +#endif } } + if (!ggml_vk_khr_cooperative_matrix_support(props2.properties, driver_props)) { + coopmat_support = false; + } + const char* GGML_VK_DISABLE_F16 = getenv("GGML_VK_DISABLE_F16"); bool force_disable_f16 = GGML_VK_DISABLE_F16 != nullptr; @@ -2055,15 +2600,35 @@ static void ggml_vk_print_gpu_info(size_t idx) { vk12_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VULKAN_1_2_FEATURES; vk11_features.pNext = &vk12_features; + // Pointer to the last chain element + VkBaseOutStructure * last_struct = (VkBaseOutStructure *)&vk12_features; + +#if defined(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) + VkPhysicalDeviceCooperativeMatrixFeaturesKHR coopmat_features; + coopmat_features.pNext = nullptr; + coopmat_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_COOPERATIVE_MATRIX_FEATURES_KHR; + coopmat_features.cooperativeMatrix = VK_FALSE; + + if (coopmat_support) { + last_struct->pNext = (VkBaseOutStructure *)&coopmat_features; + last_struct = (VkBaseOutStructure *)&coopmat_features; + } + vkGetPhysicalDeviceFeatures2(physical_device, &device_features2); fp16 = fp16 && vk12_features.shaderFloat16; + coopmat_support = coopmat_support && coopmat_features.cooperativeMatrix; +#endif + + std::string matrix_cores = coopmat2_support ? "NV_coopmat2" : coopmat_support ? "KHR_coopmat" : "none"; + std::string device_name = props2.properties.deviceName.data(); - std::cerr << GGML_VK_NAME << idx << ": " << device_name << " (" << driver_props.driverName << ") | uma: " << uma << " | fp16: " << fp16 << " | warp size: " << subgroup_size << std::endl; + GGML_LOG_DEBUG("ggml_vulkan: %zu = %s (%s) | uma: %d | fp16: %d | warp size: %zu | matrix cores: %s\n", + idx, device_name.c_str(), driver_props.driverName.data(), uma, fp16, subgroup_size, matrix_cores.c_str()); if (props2.properties.deviceType == vk::PhysicalDeviceType::eCpu) { - std::cerr << "ggml_vulkan: Warning: Device type is CPU. This is probably not the device you want." << std::endl; + GGML_LOG_DEBUG("ggml_vulkan: Warning: Device type is CPU. This is probably not the device you want.\n"); } } @@ -2118,8 +2683,7 @@ void ggml_vk_instance_init() { }; validation_features.setPNext(nullptr); instance_create_info.setPNext(&validation_features); - - std::cerr << "ggml_vulkan: Validation layers enabled" << std::endl; + GGML_LOG_DEBUG("ggml_vulkan: Validation layers enabled\n"); } vk_instance.instance = vk::createInstance(instance_create_info); @@ -2233,8 +2797,7 @@ void ggml_vk_instance_init() { vk_instance.device_indices.push_back(0); } } - - std::cerr << "ggml_vulkan: Found " << vk_instance.device_indices.size() << " Vulkan devices:" << std::endl; + GGML_LOG_DEBUG("ggml_vulkan: Found %zu Vulkan devices:\n", vk_instance.device_indices.size()); for (size_t i = 0; i < vk_instance.device_indices.size(); i++) { ggml_vk_print_gpu_info(i); @@ -2290,7 +2853,7 @@ static vk_pipeline ggml_vk_get_to_fp16(ggml_backend_vk_context * ctx, ggml_type return ctx->device->pipeline_dequant[type]; } -static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_pipeline(ggml_backend_vk_context * ctx, ggml_type src0_type, ggml_type src1_type) { +static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_pipeline(ggml_backend_vk_context * ctx, ggml_type src0_type, ggml_type src1_type, ggml_prec prec) { VK_LOG_DEBUG("ggml_vk_get_mul_mat_mat_pipeline(" << ggml_type_name(src0_type) << ", " << ggml_type_name(src1_type) << ")"); if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_F32) { return ctx->device->pipeline_matmul_f32; @@ -2298,14 +2861,23 @@ static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_pipeline(ggml_backend_vk_conte if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_F16) { return ctx->device->pipeline_matmul_f32_f16; } - if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { - return ctx->device->pipeline_matmul_f16_f32; - } - if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { - return ctx->device->pipeline_matmul_f16; + if (prec == GGML_PREC_DEFAULT && ctx->device->fp16 && !(ctx->device->coopmat_support && !ctx->device->coopmat_acc_f16_support)) { + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { + return ctx->device->pipeline_matmul_f16_f32.f16acc; + } + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { + return ctx->device->pipeline_matmul_f16.f16acc; + } + } else { + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { + return ctx->device->pipeline_matmul_f16_f32.f32acc; + } + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { + return ctx->device->pipeline_matmul_f16.f32acc; + } } - if (src1_type != GGML_TYPE_F32) { + if (src1_type != GGML_TYPE_F32 && !ctx->device->coopmat2) { return nullptr; } @@ -2326,12 +2898,17 @@ static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_pipeline(ggml_backend_vk_conte return nullptr; } - return ctx->device->pipeline_dequant_mul_mat_mat[src0_type]; + if (ctx->device->coopmat2) { + assert(src1_type == GGML_TYPE_F16); + return ctx->device->pipeline_dequant_mul_mat_mat_f16[src0_type].f16acc; + } + return ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[src0_type].f16acc : ctx->device->pipeline_dequant_mul_mat_mat[src0_type].f32acc; } -static vk_pipeline ggml_vk_get_dequantize_mul_mat_vec(ggml_backend_vk_context * ctx, ggml_type a_type, ggml_type b_type) { +static vk_pipeline ggml_vk_get_dequantize_mul_mat_vec(ggml_backend_vk_context * ctx, ggml_type a_type, ggml_type b_type, uint32_t num_cols) { VK_LOG_DEBUG("ggml_vk_get_dequantize_mul_mat_vec()"); GGML_ASSERT(b_type == GGML_TYPE_F32 || b_type == GGML_TYPE_F16); + GGML_ASSERT(num_cols >= 1 && num_cols <= mul_mat_vec_max_cols); switch (a_type) { case GGML_TYPE_F32: @@ -2352,19 +2929,28 @@ static vk_pipeline ggml_vk_get_dequantize_mul_mat_vec(ggml_backend_vk_context * return nullptr; } - return b_type == GGML_TYPE_F32 ? ctx->device->pipeline_dequant_mul_mat_vec_f32_f32[a_type] : ctx->device->pipeline_dequant_mul_mat_vec_f16_f32[a_type]; + return b_type == GGML_TYPE_F32 ? ctx->device->pipeline_dequant_mul_mat_vec_f32_f32[a_type][num_cols-1] : ctx->device->pipeline_dequant_mul_mat_vec_f16_f32[a_type][num_cols-1]; } -static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_id_pipeline(ggml_backend_vk_context * ctx, ggml_type src0_type, ggml_type src1_type) { +static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_id_pipeline(ggml_backend_vk_context * ctx, ggml_type src0_type, ggml_type src1_type, ggml_prec prec) { VK_LOG_DEBUG("ggml_vk_get_mul_mat_mat_id_pipeline()"); if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_F32) { return ctx->device->pipeline_matmul_id_f32; } - if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { - return ctx->device->pipeline_matmul_id_f16_f32; - } - if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { - return ctx->device->pipeline_matmul_id_f16; + if (prec == GGML_PREC_DEFAULT && ctx->device->fp16 && !(ctx->device->coopmat_support && !ctx->device->coopmat_acc_f16_support)) { + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { + return ctx->device->pipeline_matmul_id_f16_f32.f16acc; + } + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { + return ctx->device->pipeline_matmul_id_f16.f16acc; + } + } else { + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F32) { + return ctx->device->pipeline_matmul_id_f16_f32.f32acc; + } + if (src0_type == GGML_TYPE_F16 && src1_type == GGML_TYPE_F16) { + return ctx->device->pipeline_matmul_id_f16.f32acc; + } } GGML_ASSERT(src1_type == GGML_TYPE_F32); @@ -2386,7 +2972,7 @@ static vk_matmul_pipeline ggml_vk_get_mul_mat_mat_id_pipeline(ggml_backend_vk_co return nullptr; } - return ctx->device->pipeline_dequant_mul_mat_mat_id[src0_type]; + return ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat_id[src0_type].f16acc : ctx->device->pipeline_dequant_mul_mat_mat_id[src0_type].f32acc; } static vk_pipeline ggml_vk_get_dequantize_mul_mat_vec_id(ggml_backend_vk_context * ctx, ggml_type a_type, ggml_type b_type) { @@ -2638,8 +3224,8 @@ static void ggml_vk_buffer_write_nc_async(ggml_backend_vk_context * ctx, vk_cont GGML_ABORT("fatal error"); } // Check if src is pinned memory - vk_buffer buf; - size_t buf_offset; + vk_buffer buf = nullptr; + size_t buf_offset = 0; ggml_vk_host_get(ctx->device, tensor->data, buf, buf_offset); const uint64_t ne0 = tensor->ne[0]; @@ -2702,7 +3288,7 @@ static void ggml_vk_buffer_write_nc_async(ggml_backend_vk_context * ctx, vk_cont VkBufferCopy buf_copy{ 0, offset, copy_size }; ggml_vk_sync_buffers(subctx); - vkCmdCopyBuffer(subctx->s->buffer, staging->buffer, dst->buffer, 1, &buf_copy); + vkCmdCopyBuffer(subctx->s->buffer, (VkBuffer)staging->buffer, (VkBuffer)dst->buffer, 1, &buf_copy); for (uint64_t i3 = 0; i3 < ne3; i3++) { for (uint64_t i2 = 0; i2 < ne2; i2++) { @@ -2735,7 +3321,7 @@ static void ggml_vk_buffer_write_2d_async(vk_context subctx, vk_buffer& dst, siz } // Check if src is pinned memory vk_buffer buf = nullptr; - size_t buf_offset; + size_t buf_offset = 0; ggml_vk_host_get(dst->device, src, buf, buf_offset); if (buf != nullptr) { @@ -2777,7 +3363,7 @@ static void ggml_vk_buffer_write_2d_async(vk_context subctx, vk_buffer& dst, siz copy_size}; ggml_vk_sync_buffers(subctx); - vkCmdCopyBuffer(subctx->s->buffer, staging_buffer->buffer, dst->buffer, 1, &buf_copy); + vkCmdCopyBuffer(subctx->s->buffer, (VkBuffer)staging_buffer->buffer, (VkBuffer)dst->buffer, 1, &buf_copy); if (width == spitch) { deferred_memcpy((uint8_t *)staging_buffer->ptr, src, width * height, &subctx->in_memcpys); @@ -2833,7 +3419,7 @@ static void ggml_vk_buffer_read_2d_async(vk_context subctx, vk_buffer& src, size // Check if dst is pinned memory vk_buffer buf = nullptr; - size_t buf_offset; + size_t buf_offset = 0; ggml_vk_host_get(src->device, dst, buf, buf_offset); std::vector slices(1); @@ -2913,7 +3499,7 @@ static void ggml_vk_buffer_copy_async(vk_context& ctx, vk_buffer& dst, size_t ds VkBufferCopy bc{ src_offset, dst_offset, size }; - vkCmdCopyBuffer(ctx->s->buffer, src->buffer, dst->buffer, 1, &bc); + vkCmdCopyBuffer(ctx->s->buffer, (VkBuffer)src->buffer, (VkBuffer)dst->buffer, 1, &bc); } static void ggml_vk_buffer_copy(vk_buffer& dst, size_t dst_offset, vk_buffer& src, size_t src_offset, size_t size) { @@ -2955,55 +3541,44 @@ static void ggml_vk_buffer_memset(vk_buffer& dst, size_t offset, uint32_t c, siz dst->device->device.resetFences({ dst->device->fence }); } -static uint32_t ggml_vk_guess_split_k(int m, int n, int k) { +static uint32_t ggml_vk_guess_split_k(ggml_backend_vk_context * ctx, int m, int n, int k, const vk_pipeline& pipeline) { VK_LOG_DEBUG("ggml_vk_guess_split_k(" << m << ", " << n << ", " << k << ")"); - // if (k > 128 && (m < 128 || n < 128) && m > 2 && n > 2) { - // return 4; - // } - return 1; - - GGML_UNUSED(m); GGML_UNUSED(n); GGML_UNUSED(k); -} - -static vk_pipeline ggml_vk_guess_matmul_pipeline_amd(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, int m, int n, bool aligned) { - if (m <= 32 || n <= 32) { - return aligned ? mmp->a_s : mmp->s; + uint32_t split_k = 1; + if (ctx->device->shader_core_count != 0 && m >= (int)pipeline->wg_denoms[0] && n >= (int)pipeline->wg_denoms[1]) { + // If k is 'large' and the SMs will fill less than halfway, use split_k. + uint32_t m_tiles = CEIL_DIV(m, pipeline->wg_denoms[0]); + uint32_t n_tiles = CEIL_DIV(n, pipeline->wg_denoms[1]); + if (k >= 2048 && m_tiles * n_tiles < ctx->device->shader_core_count / 2) { + split_k = ctx->device->shader_core_count / (m_tiles * n_tiles); + // Clamp to 2 or 4 + split_k = std::min(split_k, 4u); + if (split_k == 3) { + split_k = 2; + } + } } - return aligned ? mmp->a_m : mmp->m; - GGML_UNUSED(ctx); -} - -static vk_pipeline ggml_vk_guess_matmul_pipeline_apple(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, bool aligned) { - return aligned ? mmp->a_m : mmp->m; - - GGML_UNUSED(ctx); -} - -static vk_pipeline ggml_vk_guess_matmul_pipeline_intel(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, bool aligned) { - return aligned ? mmp->a_s : mmp->s; - - GGML_UNUSED(ctx); + return split_k; } static vk_pipeline ggml_vk_guess_matmul_pipeline(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, int m, int n, bool aligned) { VK_LOG_DEBUG("ggml_vk_guess_matmul_pipeline(" << m << ", " << n << ", " << aligned << ")"); - switch (ctx->device->vendor_id) { - case VK_VENDOR_ID_AMD: - return ggml_vk_guess_matmul_pipeline_amd(ctx, mmp, m, n, aligned); - case VK_VENDOR_ID_APPLE: - return ggml_vk_guess_matmul_pipeline_apple(ctx, mmp, aligned); - case VK_VENDOR_ID_INTEL: - return ggml_vk_guess_matmul_pipeline_intel(ctx, mmp, aligned); - default: - break; - } - if (m <= 32 || n <= 32) { + if (ctx->device->coopmat2) { + if ((ctx->device->mul_mat_l && (m % mmp->l->wg_denoms[0]) == 0 && (n % mmp->l->wg_denoms[1]) == 0) || (!ctx->device->mul_mat_m && !ctx->device->mul_mat_s)) { + return aligned ? mmp->a_l : mmp->l; + } + if ((ctx->device->mul_mat_m && (m % mmp->m->wg_denoms[0]) == 0 && (n % mmp->m->wg_denoms[1]) == 0) || !ctx->device->mul_mat_s) { + return aligned ? mmp->a_m : mmp->m; + } return aligned ? mmp->a_s : mmp->s; } - if (m <= 64 || n <= 64) { + + if ((ctx->device->mul_mat_s && (m <= 32 || n <= 32)) || (!ctx->device->mul_mat_m && !ctx->device->mul_mat_l)) { + return aligned ? mmp->a_s : mmp->s; + } + if ((ctx->device->mul_mat_m && (m <= 64 || n <= 64)) || !ctx->device->mul_mat_l) { return aligned ? mmp->a_m : mmp->m; } return aligned ? mmp->a_l : mmp->l; @@ -3038,6 +3613,33 @@ static void ggml_vk_matmul( ggml_vk_dispatch_pipeline(ctx, subctx, ctx->device->pipeline_matmul_split_k_reduce, { split_k_buffer, d }, pc2.size() * sizeof(uint32_t), pc2.data(), { m * n * batch, 1, 1 }); } +static vk_pipeline ggml_vk_guess_matmul_id_pipeline(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, int m, int n, bool aligned) { + VK_LOG_DEBUG("ggml_vk_guess_matmul_pipeline(" << m << ", " << n << ", " << aligned << ")"); + + if (ctx->device->coopmat2) { + if ((ctx->device->mul_mat_id_l && (m % mmp->l->wg_denoms[0]) == 0 && (n % mmp->l->wg_denoms[1]) == 0) || (!ctx->device->mul_mat_id_m && !ctx->device->mul_mat_id_s)) { + return aligned ? mmp->a_l : mmp->l; + } + if ((ctx->device->mul_mat_id_m && (m % mmp->m->wg_denoms[0]) == 0 && (n % mmp->m->wg_denoms[1]) == 0) || !ctx->device->mul_mat_id_s) { + return aligned ? mmp->a_m : mmp->m; + } + return aligned ? mmp->a_s : mmp->s; + } + + if ((ctx->device->mul_mat_id_s && (m <= 32 || n <= 32)) || (!ctx->device->mul_mat_id_m && !ctx->device->mul_mat_id_l)) { + return aligned ? mmp->a_s : mmp->s; + } + if ((ctx->device->mul_mat_id_m && (m <= 64 || n <= 64)) || !ctx->device->mul_mat_id_l) { + return aligned ? mmp->a_m : mmp->m; + } + return aligned ? mmp->a_l : mmp->l; +} + +static uint32_t ggml_vk_guess_matmul_id_pipeline_align(ggml_backend_vk_context * ctx, vk_matmul_pipeline& mmp, int m, int n) { + VK_LOG_DEBUG("ggml_vk_guess_matmul_pipeline_align(" << m << ", " << n << ")"); + return ggml_vk_guess_matmul_id_pipeline(ctx, mmp, m, n, true)->align; +} + static void ggml_vk_matmul_id( ggml_backend_vk_context * ctx, vk_context& subctx, vk_pipeline& pipeline, vk_subbuffer&& a, vk_subbuffer&& b, vk_subbuffer&& d, vk_subbuffer&& ids, @@ -3061,18 +3663,34 @@ static bool ggml_vk_dim01_contiguous(const ggml_tensor * tensor) { tensor->nb[3] == tensor->nb[2]*tensor->ne[2]; } -static vk_pipeline ggml_vk_get_cpy_pipeline(ggml_backend_vk_context * ctx, ggml_type from, ggml_type to) { - if (from == GGML_TYPE_F32 && to == GGML_TYPE_F32) { - return ctx->device->pipeline_cpy_f32_f32; +static vk_pipeline ggml_vk_get_cpy_pipeline(ggml_backend_vk_context * ctx, const ggml_tensor * src, const ggml_tensor * dst, ggml_type to) { + + // Choose "contiguous copy" shader if src/dst are contiguous + bool contig = ggml_is_contiguous(src) && (!dst || ggml_is_contiguous(dst)); + + if (src->type == GGML_TYPE_F32 && to == GGML_TYPE_F32) { + if (contig) { + return ctx->device->pipeline_contig_cpy_f32_f32; + } else { + return ctx->device->pipeline_cpy_f32_f32; + } } - if (from == GGML_TYPE_F32 && to == GGML_TYPE_F16) { - return ctx->device->pipeline_cpy_f32_f16; + if (src->type == GGML_TYPE_F32 && to == GGML_TYPE_F16) { + if (contig) { + return ctx->device->pipeline_contig_cpy_f32_f16; + } else { + return ctx->device->pipeline_cpy_f32_f16; + } } - if (from == GGML_TYPE_F16 && to == GGML_TYPE_F16) { - return ctx->device->pipeline_cpy_f16_f16; + if (src->type == GGML_TYPE_F16 && to == GGML_TYPE_F16) { + if (contig) { + return ctx->device->pipeline_contig_cpy_f16_f16; + } else { + return ctx->device->pipeline_cpy_f16_f16; + } } - std::cerr << "Missing CPY op for types: " << ggml_type_name(from) << " " << ggml_type_name(to) << std::endl; + std::cerr << "Missing CPY op for types: " << ggml_type_name(src->type) << " " << ggml_type_name(to) << std::endl; GGML_ABORT("fatal error"); } @@ -3082,16 +3700,27 @@ static void ggml_vk_cpy_to_contiguous(ggml_backend_vk_context * ctx, vk_context& const int tensor_type_size = ggml_type_size(tensor->type); const uint32_t ne = ggml_nelements(tensor); + std::array elements; - const vk_op_unary_push_constants pc = { + if (ne > 262144) { + elements = { 512, 512, CEIL_DIV(ne, 262144) }; + } else if (ne > 512) { + elements = { 512, CEIL_DIV(ne, 512), 1 }; + } else { + elements = { ne, 1, 1 }; + } + + vk_op_unary_push_constants pc = { (uint32_t)ne, (uint32_t)tensor->ne[0], (uint32_t)tensor->ne[1], (uint32_t)tensor->ne[2], (uint32_t)tensor->ne[3], (uint32_t)tensor->nb[0] / tensor_type_size, (uint32_t)tensor->nb[1] / tensor_type_size, (uint32_t)tensor->nb[2] / tensor_type_size, (uint32_t)tensor->nb[3] / tensor_type_size, (uint32_t)tensor->ne[0], (uint32_t)tensor->ne[1], (uint32_t)tensor->ne[2], (uint32_t)tensor->ne[3], 1 , (uint32_t)tensor->ne[0] , (uint32_t)(tensor->ne[0] * tensor->ne[1]) , (uint32_t)(tensor->ne[0] * tensor->ne[1] * tensor->ne[2]), 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }; + init_pushconst_fastdiv(pc); ggml_vk_sync_buffers(subctx); - ggml_vk_dispatch_pipeline(ctx, subctx, pipeline, { in, out }, sizeof(vk_op_unary_push_constants), &pc, { ne, 1, 1 }); + ggml_vk_dispatch_pipeline(ctx, subctx, pipeline, { in, out }, sizeof(vk_op_unary_push_constants), &pc, elements); } static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, bool dryrun = false) { @@ -3122,9 +3751,9 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub ggml_backend_vk_buffer_context * src0_buf_ctx = (ggml_backend_vk_buffer_context *)src0->buffer->context; ggml_backend_vk_buffer_context * src1_buf_ctx = (ggml_backend_vk_buffer_context *)src1->buffer->context; - vk_buffer d_Qx; + vk_buffer d_Qx = nullptr; size_t qx_buf_offset = 0; - vk_buffer d_Qy; + vk_buffer d_Qy = nullptr; size_t qy_buf_offset = 0; bool src0_uma = false; @@ -3138,18 +3767,20 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub } const bool x_non_contig = !ggml_vk_dim01_contiguous(src0); - const bool y_non_contig = !ggml_vk_dim01_contiguous(src1); + // Reformat and convert to fp16 if src1 is non-contiguous, or for coopmat2 for better perf + const bool y_non_contig = (ctx->device->coopmat2 && src1->type == GGML_TYPE_F32) || + !ggml_vk_dim01_contiguous(src1); const bool y_f32_kernel = src1->type == GGML_TYPE_F32 && !y_non_contig; - vk_matmul_pipeline mmp = ggml_vk_get_mul_mat_mat_pipeline(ctx, src0->type, y_non_contig ? GGML_TYPE_F16 : src1->type); + vk_matmul_pipeline mmp = ggml_vk_get_mul_mat_mat_pipeline(ctx, src0->type, y_non_contig ? GGML_TYPE_F16 : src1->type, (ggml_prec)dst->op_params[0]); const bool qx_needs_dequant = mmp == nullptr || x_non_contig; const bool qy_needs_dequant = (src1->type != GGML_TYPE_F16 && !y_f32_kernel) || y_non_contig; - if (mmp == nullptr) { + if (qx_needs_dequant) { // Fall back to dequant + f16 mulmat - mmp = ggml_vk_get_mul_mat_mat_pipeline(ctx, GGML_TYPE_F16, y_f32_kernel ? GGML_TYPE_F32 : GGML_TYPE_F16); + mmp = ggml_vk_get_mul_mat_mat_pipeline(ctx, GGML_TYPE_F16, y_f32_kernel ? GGML_TYPE_F32 : GGML_TYPE_F16, (ggml_prec)dst->op_params[0]); } // Not implemented @@ -3162,10 +3793,10 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub const uint32_t kpad = ggml_vk_align_size(ne10, ggml_vk_guess_matmul_pipeline_align(ctx, mmp, ne01, ne11)); const bool aligned = ne10 == kpad && ne01 > 8 && ne11 > 8; - const uint32_t split_k = ggml_vk_guess_split_k(ne01, ne11, ne10); - vk_pipeline pipeline = ggml_vk_guess_matmul_pipeline(ctx, mmp, ne01, ne11, aligned); + const uint32_t split_k = ggml_vk_guess_split_k(ctx, ne01, ne11, ne10, pipeline); + const uint64_t qx_sz = ggml_type_size(src0->type) * x_ne / ggml_blck_size(src0->type); const uint64_t qy_sz = ggml_type_size(src1->type) * y_ne / ggml_blck_size(src1->type); const uint64_t x_sz = !qx_needs_dequant ? qx_sz : sizeof(ggml_fp16_t) * x_ne; @@ -3176,12 +3807,12 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub vk_pipeline to_fp16_vk_1 = nullptr; if (x_non_contig) { - to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0->type, GGML_TYPE_F16); + to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0, nullptr, GGML_TYPE_F16); } else { to_fp16_vk_0 = ggml_vk_get_to_fp16(ctx, src0->type); } if (y_non_contig) { - to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1->type, GGML_TYPE_F16); + to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1, nullptr, GGML_TYPE_F16); } else { to_fp16_vk_1 = ggml_vk_get_to_fp16(ctx, src1->type); } @@ -3191,7 +3822,7 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub if (dryrun) { const uint64_t x_sz_upd = x_sz * ne02 * ne03; const uint64_t y_sz_upd = y_sz * ne12 * ne13; - const uint64_t split_k_size = split_k > 1 ? d_sz * ne12 * ne13 * 4 : 0; + const uint64_t split_k_size = split_k > 1 ? d_sz * ne12 * ne13 * split_k : 0; if ( (qx_needs_dequant && x_sz_upd > ctx->device->max_memory_allocation_size) || (qy_needs_dequant && y_sz_upd > ctx->device->max_memory_allocation_size) || @@ -3308,8 +3939,6 @@ static void ggml_vk_mul_mat_vec_q_f16(ggml_backend_vk_context * ctx, vk_context& const uint64_t ne12 = src1->ne[2]; const uint64_t ne13 = src1->ne[3]; - GGML_ASSERT(ne11 == 1); - const uint64_t ne20 = dst->ne[0]; const uint64_t ne21 = dst->ne[1]; const uint64_t ne22 = dst->ne[2]; @@ -3318,13 +3947,18 @@ static void ggml_vk_mul_mat_vec_q_f16(ggml_backend_vk_context * ctx, vk_context& const uint64_t r2 = ne12 / ne02; const uint64_t r3 = ne13 / ne03; + // batch_n indicates that we need to compute a few vector results, and this assumes + // ne12 and ne13 are 1. It overloads the batch_strides to hold the row strides. + GGML_ASSERT(ne11 == 1 || ne12 * ne13 == 1); + bool batch_n = ne11 > 1; + ggml_backend_vk_buffer_context * dst_buf_ctx = (ggml_backend_vk_buffer_context *)dst->buffer->context; ggml_backend_vk_buffer_context * src0_buf_ctx = (ggml_backend_vk_buffer_context *)src0->buffer->context; ggml_backend_vk_buffer_context * src1_buf_ctx = (ggml_backend_vk_buffer_context *)src1->buffer->context; - vk_buffer d_Qx; + vk_buffer d_Qx = nullptr; size_t qx_buf_offset = 0; - vk_buffer d_Qy; + vk_buffer d_Qy = nullptr; size_t qy_buf_offset = 0; bool src0_uma = false; @@ -3361,14 +3995,14 @@ static void ggml_vk_mul_mat_vec_q_f16(ggml_backend_vk_context * ctx, vk_context& vk_pipeline to_fp16_vk_0 = nullptr; vk_pipeline to_fp16_vk_1 = nullptr; if (x_non_contig) { - to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0->type, src0->type); + to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0, nullptr, src0->type); } if (y_non_contig) { - to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1->type, src1->type); + to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1, nullptr, src1->type); } else { to_fp16_vk_1 = ggml_vk_get_to_fp16(ctx, src1->type); } - vk_pipeline dmmv = ggml_vk_get_dequantize_mul_mat_vec(ctx, src0->type, src1->type); + vk_pipeline dmmv = ggml_vk_get_dequantize_mul_mat_vec(ctx, src0->type, src1->type, ne11); GGML_ASSERT(!qx_needs_dequant || to_fp16_vk_0 != nullptr); // NOLINT GGML_ASSERT(!qy_needs_dequant || to_fp16_vk_1 != nullptr); // NOLINT GGML_ASSERT(dmmv != nullptr); @@ -3440,8 +4074,10 @@ static void ggml_vk_mul_mat_vec_q_f16(ggml_backend_vk_context * ctx, vk_context& ggml_vk_cpy_to_contiguous(ctx, subctx, to_fp16_vk_1, src1, { d_Qy, qy_buf_offset, VK_WHOLE_SIZE }, { d_Y, 0, VK_WHOLE_SIZE }); } - uint32_t stride_batch_x = ne00*ne01; - uint32_t stride_batch_y = ne10*ne11; + // For batch_n, the A matrix is the same for each batch, and B/D use the row stride as the batch stride + uint32_t stride_batch_x = batch_n ? 0 : ne00*ne01; + uint32_t stride_batch_y = batch_n ? ne10 : (ne10*ne11); + uint32_t stride_batch_d = batch_n ? ne20 : (ne20*ne21); if (!ggml_vk_dim01_contiguous(src0) && !qx_needs_dequant) { stride_batch_x = src0->nb[0] / ggml_type_size(src0->type); @@ -3458,13 +4094,13 @@ static void ggml_vk_mul_mat_vec_q_f16(ggml_backend_vk_context * ctx, vk_context& if (ne01 > max_groups_x) { groups_z = 64; - groups_x /= groups_z; + groups_x = CEIL_DIV(groups_x, groups_z); } // compute const vk_mat_vec_push_constants pc = { (uint32_t)ne00, (uint32_t)ne10, (uint32_t)ne10, (uint32_t)ne01, - stride_batch_x, stride_batch_y, (uint32_t)(ne20*ne21), + stride_batch_x, stride_batch_y, stride_batch_d, (uint32_t)ne02, (uint32_t)ne12, (uint32_t)r2, (uint32_t)r3, }; ggml_vk_sync_buffers(subctx); @@ -3500,7 +4136,7 @@ static void ggml_vk_mul_mat_vec_p021_f16_f32(ggml_backend_vk_context * ctx, vk_c ggml_backend_vk_buffer_context * src0_buf_ctx = (ggml_backend_vk_buffer_context *)src0->buffer->context; ggml_backend_vk_buffer_context * src1_buf_ctx = (ggml_backend_vk_buffer_context *)src1->buffer->context; - vk_buffer d_Qy; + vk_buffer d_Qy = nullptr; size_t qy_buf_offset = 0; bool src1_uma = false; @@ -3630,11 +4266,24 @@ static void ggml_vk_mul_mat_vec_nc_f16_f32(ggml_backend_vk_context * ctx, vk_con static void ggml_vk_mul_mat(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, bool dryrun = false) { VK_LOG_DEBUG("ggml_vk_mul_mat(" << src0 << ", " << src1 << ", " << dst << ")"); - if (src0->type == GGML_TYPE_F16 && ggml_is_permuted(src0) && ggml_is_permuted(src1) && dst->ne[1] == 1) { + if (src0->type == GGML_TYPE_F16 && ggml_is_permuted(src0) && ggml_is_permuted(src1) && dst->ne[1] == 1 && + // detect 0213 permutation, and batch size of 1 + src0->nb[0] <= src0->nb[2] && + src0->nb[2] <= src0->nb[1] && + src0->nb[1] <= src0->nb[3] && + src1->nb[0] <= src1->nb[2] && + src1->nb[2] <= src1->nb[1] && + src1->nb[1] <= src1->nb[3] && + src0->ne[3] == 1 && + src1->ne[3] == 1) { ggml_vk_mul_mat_vec_p021_f16_f32(ctx, subctx, src0, src1, dst, dryrun); - } else if (src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && dst->ne[1] == 1) { + } else if (src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && dst->ne[1] == 1 && + !ggml_is_permuted(src0) && !ggml_is_permuted(src1)) { ggml_vk_mul_mat_vec_nc_f16_f32(ctx, subctx, src0, src1, dst, dryrun); - } else if (dst->ne[1] == 1 && (src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16 || ggml_is_quantized(src0->type))) { + // mul_mat_vec supports batching ne12*ne13 when ne11==1, or treating ne11 as the batch size (up to four) + // when ne12 and ne13 are one. + } else if ((dst->ne[1] == 1 || (dst->ne[1] <= mul_mat_vec_max_cols && src1->ne[2] * src1->ne[3] == 1)) && + (src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16 || ggml_is_quantized(src0->type))) { ggml_vk_mul_mat_vec_q_f16(ctx, subctx, src0, src1, dst, dryrun); } else { ggml_vk_mul_mat_q_f16(ctx, subctx, src0, src1, dst, dryrun); @@ -3678,11 +4327,11 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context& ggml_backend_vk_buffer_context * src1_buf_ctx = (ggml_backend_vk_buffer_context *)src1->buffer->context; ggml_backend_vk_buffer_context * ids_buf_ctx = (ggml_backend_vk_buffer_context *)ids->buffer->context; - vk_buffer d_Qx; + vk_buffer d_Qx = nullptr; size_t qx_buf_offset = 0; - vk_buffer d_Qy; + vk_buffer d_Qy = nullptr; size_t qy_buf_offset = 0; - vk_buffer d_ids; + vk_buffer d_ids = nullptr; size_t ids_buf_offset = 0; bool src0_uma = false; @@ -3703,12 +4352,12 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context& const bool y_f32_kernel = src1->type == GGML_TYPE_F32 && !y_non_contig; - vk_matmul_pipeline mmp = ggml_vk_get_mul_mat_mat_id_pipeline(ctx, src0->type, y_non_contig ? GGML_TYPE_F16 : src1->type); + vk_matmul_pipeline mmp = ggml_vk_get_mul_mat_mat_id_pipeline(ctx, src0->type, y_non_contig ? GGML_TYPE_F16 : src1->type, (ggml_prec)dst->op_params[0]); const bool qx_needs_dequant = mmp == nullptr || x_non_contig; const bool qy_needs_dequant = (src1->type != GGML_TYPE_F16 && !y_f32_kernel) || y_non_contig; - if (mmp == nullptr) { + if (qx_needs_dequant) { GGML_ABORT("fatal error"); } @@ -3719,10 +4368,10 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context& const uint64_t y_ne = ne11 * ne10; const uint64_t d_ne = ne21 * ne20; - const uint32_t kpad = ggml_vk_align_size(ne10, ggml_vk_guess_matmul_pipeline_align(ctx, mmp, ne01, nei1)); + const uint32_t kpad = ggml_vk_align_size(ne10, ggml_vk_guess_matmul_id_pipeline_align(ctx, mmp, ne01, nei1)); const bool aligned = ne10 == kpad && ne01 > 8 && nei1 > 8; - vk_pipeline pipeline = ggml_vk_guess_matmul_pipeline(ctx, mmp, ne01, nei1, aligned); + vk_pipeline pipeline = ggml_vk_guess_matmul_id_pipeline(ctx, mmp, ne01, nei1, aligned); const uint64_t qx_sz = ggml_type_size(src0->type) * x_ne / ggml_blck_size(src0->type); const uint64_t qy_sz = ggml_type_size(src1->type) * y_ne / ggml_blck_size(src1->type); @@ -3735,12 +4384,12 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context& vk_pipeline to_fp16_vk_1 = nullptr; if (x_non_contig) { - to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0->type, GGML_TYPE_F16); + to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0, nullptr, GGML_TYPE_F16); } else { to_fp16_vk_0 = ggml_vk_get_to_fp16(ctx, src0->type); } if (y_non_contig) { - to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1->type, GGML_TYPE_F16); + to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1, nullptr, GGML_TYPE_F16); } else { to_fp16_vk_1 = ggml_vk_get_to_fp16(ctx, src1->type); } @@ -3883,11 +4532,11 @@ static void ggml_vk_mul_mat_vec_id_q_f16(ggml_backend_vk_context * ctx, vk_conte ggml_backend_vk_buffer_context * src1_buf_ctx = (ggml_backend_vk_buffer_context *)src1->buffer->context; ggml_backend_vk_buffer_context * ids_buf_ctx = (ggml_backend_vk_buffer_context *)ids->buffer->context; - vk_buffer d_Qx; + vk_buffer d_Qx = nullptr; size_t qx_buf_offset = 0; - vk_buffer d_Qy; + vk_buffer d_Qy = nullptr; size_t qy_buf_offset = 0; - vk_buffer d_ids; + vk_buffer d_ids = nullptr; size_t ids_buf_offset = 0; bool src0_uma = false; @@ -3928,10 +4577,10 @@ static void ggml_vk_mul_mat_vec_id_q_f16(ggml_backend_vk_context * ctx, vk_conte vk_pipeline to_fp16_vk_0 = nullptr; vk_pipeline to_fp16_vk_1 = nullptr; if (x_non_contig) { - to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0->type, src0->type); + to_fp16_vk_0 = ggml_vk_get_cpy_pipeline(ctx, src0, nullptr, src0->type); } if (y_non_contig) { - to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1->type, src1->type); + to_fp16_vk_1 = ggml_vk_get_cpy_pipeline(ctx, src1, nullptr, src1->type); } else { to_fp16_vk_1 = ggml_vk_get_to_fp16(ctx, src1->type); } @@ -4025,7 +4674,7 @@ static void ggml_vk_mul_mat_vec_id_q_f16(ggml_backend_vk_context * ctx, vk_conte if (ne01 > max_groups_x) { groups_z = 64; - groups_x /= groups_z; + groups_x = CEIL_DIV(groups_x, groups_z); } // compute @@ -4050,6 +4699,167 @@ static void ggml_vk_mul_mat_id(ggml_backend_vk_context * ctx, vk_context& subctx } } +static void ggml_vk_flash_attn(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * q, const ggml_tensor * k, const ggml_tensor * v, const ggml_tensor * mask, ggml_tensor * dst, bool dryrun = false) { + VK_LOG_DEBUG("ggml_vk_flash_attn((" << q << ", name=" << q->name << ", type=" << q->type << ", ne0=" << q->ne[0] << ", ne1=" << q->ne[1] << ", ne2=" << q->ne[2] << ", ne3=" << q->ne[3] << ", nb0=" << q->nb[0] << ", nb1=" << q->nb[1] << ", nb2=" << q->nb[2] << ", nb3=" << q->nb[3]; + std::cerr << "), (" << k << ", name=" << k->name << ", type=" << k->type << ", ne0=" << k->ne[0] << ", ne1=" << k->ne[1] << ", ne2=" << k->ne[2] << ", ne3=" << k->ne[3] << ", nb0=" << k->nb[0] << ", nb1=" << k->nb[1] << ", nb2=" << k->nb[2] << ", nb3=" << k->nb[3]; + std::cerr << "), (" << v << ", name=" << v->name << ", type=" << v->type << ", ne0=" << v->ne[0] << ", ne1=" << v->ne[1] << ", ne2=" << v->ne[2] << ", ne3=" << v->ne[3] << ", nb0=" << v->nb[0] << ", nb1=" << v->nb[1] << ", nb2=" << v->nb[2] << ", nb3=" << v->nb[3]; + std::cerr << "), (" << dst << ", name=" << dst->name << ", type=" << dst->type << ", ne0=" << dst->ne[0] << ", ne1=" << dst->ne[1] << ", ne2=" << dst->ne[2] << ", ne3=" << dst->ne[3] << ", nb0=" << dst->nb[0] << ", nb1=" << dst->nb[1] << ", nb2=" << dst->nb[2] << ", nb3=" << dst->nb[3]; + std::cerr << "), " << (dryrun ? "dryrun" : "") << ")"); + + GGML_TENSOR_LOCALS(int64_t, neq, q, ne) + GGML_TENSOR_LOCALS(size_t, nbq, q, nb) + GGML_TENSOR_LOCALS(int64_t, nek, k, ne) + GGML_TENSOR_LOCALS(size_t, nbk, k, nb) + GGML_TENSOR_LOCALS(int64_t, nev, v, ne) + GGML_TENSOR_LOCALS(size_t, nbv, v, nb) + GGML_TENSOR_LOCALS(int64_t, ne, dst, ne) + GGML_TENSOR_LOCALS(size_t, nb, dst, nb) + + const uint32_t nem1 = mask ? mask->ne[1] : 0; + const uint32_t nbm1 = mask ? mask->nb[1] : 0; + + const uint32_t D = neq0; + const uint32_t N = neq1; + const uint32_t KV = nek1; + + GGML_ASSERT(ne0 == D); + GGML_ASSERT(ne2 == N); + + // input tensor rows must be contiguous + GGML_ASSERT(nbq0 == ggml_type_size(q->type)); + GGML_ASSERT(nbk0 == ggml_type_size(k->type)); + GGML_ASSERT(nbv0 == ggml_type_size(v->type)); + + GGML_ASSERT(neq0 == D); + GGML_ASSERT(nek0 == D); + GGML_ASSERT(nev0 == D); + + GGML_ASSERT(neq1 == N); + GGML_ASSERT(nev0 == D); + + GGML_ASSERT(nev1 == nek1); + + // dst cannot be transposed or permuted + GGML_ASSERT(nb0 == sizeof(float)); + GGML_ASSERT(nb0 <= nb1); + GGML_ASSERT(nb1 <= nb2); + GGML_ASSERT(nb2 <= nb3); + + assert(dst->type == GGML_TYPE_F32); + assert(q->type == GGML_TYPE_F32); + assert(k->type == v->type); + + vk_pipeline *pipelines; + // XXX TODO other backends may be changing accumulator precision to default to f32 soon + bool f32acc = dst->op_params[3] == GGML_PREC_F32; + bool small_rows = N <= flash_attention_num_small_rows; + switch (D) { + case 64: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D64[k->type][f32acc][small_rows][0]; break; + case 80: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D80[k->type][f32acc][small_rows][0]; break; + case 96: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D96[k->type][f32acc][small_rows][0]; break; + case 112: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D112[k->type][f32acc][small_rows][0]; break; + case 128: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D128[k->type][f32acc][small_rows][0]; break; + case 256: pipelines = &ctx->device->pipeline_flash_attn_f32_f16_D256[k->type][f32acc][small_rows][0]; break; + default: + assert(!"unsupported D value"); + return; + } + assert(pipelines); + + bool aligned = (KV % pipelines[1]->align) == 0; + vk_pipeline pipeline = pipelines[aligned]; + assert(pipeline); + + if (dryrun) { + // Request descriptor sets + ggml_pipeline_request_descriptor_sets(ctx->device, pipeline, 1); + return; + } + + float scale = 1.0f; + float max_bias = 0.0f; + float logit_softcap = 0.0f; + + memcpy(&scale, (const float *) dst->op_params + 0, sizeof(float)); + memcpy(&max_bias, (const float *) dst->op_params + 1, sizeof(float)); + memcpy(&logit_softcap, (const float *) dst->op_params + 2, sizeof(float)); + + if (logit_softcap != 0) { + scale /= logit_softcap; + } + + const uint32_t n_head_kv = neq2; + const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head_kv)); + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + ggml_vk_sync_buffers(subctx); + + vk_buffer d_Q = nullptr, d_K = nullptr, d_V = nullptr, d_D = nullptr, d_M = nullptr; + size_t q_buf_offset = 0, k_buf_offset = 0, v_buf_offset = 0, d_buf_offset = 0, m_buf_offset = 0; + + bool Q_uma = false, K_uma = false, V_uma = false, D_uma = false, M_uma = false; + + if (ctx->device->uma) { + ggml_vk_host_get(ctx->device, q->data, d_Q, q_buf_offset); + ggml_vk_host_get(ctx->device, k->data, d_K, q_buf_offset); + ggml_vk_host_get(ctx->device, v->data, d_V, q_buf_offset); + ggml_vk_host_get(ctx->device, dst->data, d_D, q_buf_offset); + Q_uma = d_Q != nullptr; + K_uma = d_K != nullptr; + V_uma = d_V != nullptr; + D_uma = d_D != nullptr; + if (mask) { + ggml_vk_host_get(ctx->device, mask->data, d_M, q_buf_offset); + M_uma = d_M != nullptr; + } + } + + + ggml_backend_vk_buffer_context * d_buf_ctx = (ggml_backend_vk_buffer_context *)dst->buffer->context; + ggml_backend_vk_buffer_context * q_buf_ctx = (ggml_backend_vk_buffer_context *)q->buffer->context; + ggml_backend_vk_buffer_context * k_buf_ctx = (ggml_backend_vk_buffer_context *)k->buffer->context; + ggml_backend_vk_buffer_context * v_buf_ctx = (ggml_backend_vk_buffer_context *)v->buffer->context; + + if (!Q_uma) { + d_Q = q_buf_ctx->dev_buffer; + q_buf_offset = vk_tensor_offset(q) + q->view_offs; + } + if (!K_uma) { + d_K = k_buf_ctx->dev_buffer; + k_buf_offset = vk_tensor_offset(k) + k->view_offs; + } + if (!V_uma) { + d_V = v_buf_ctx->dev_buffer; + v_buf_offset = vk_tensor_offset(v) + v->view_offs; + } + if (!D_uma) { + d_D = d_buf_ctx->dev_buffer; + d_buf_offset = vk_tensor_offset(dst) + dst->view_offs; + } + + if (!M_uma) { + d_M = d_Q; + m_buf_offset = q_buf_offset; + if (mask) { + ggml_backend_vk_buffer_context * m_buf_ctx = (ggml_backend_vk_buffer_context*)mask->buffer->context; + d_M = m_buf_ctx->dev_buffer; + m_buf_offset = vk_tensor_offset(mask) + mask->view_offs; + } + } + + const vk_flash_attn_push_constants pc = { N, KV, (uint32_t)ne1, (uint32_t)ne2, (uint32_t)ne3, (uint32_t)neq2, (uint32_t)neq3, (uint32_t)nek2, (uint32_t)nek3, (uint32_t)nev2, (uint32_t)nev3, nem1, (uint32_t)nbq2, (uint32_t)nbq3, (uint32_t)nbk2, (uint32_t)nbk3, (uint32_t)nbv2, (uint32_t)nbv3, nbm1, scale, max_bias, logit_softcap, mask != nullptr, n_head_log2, m0, m1 }; + ggml_vk_dispatch_pipeline(ctx, subctx, pipeline, + { + vk_subbuffer{d_Q, q_buf_offset, VK_WHOLE_SIZE}, + vk_subbuffer{d_K, k_buf_offset, VK_WHOLE_SIZE}, + vk_subbuffer{d_V, v_buf_offset, VK_WHOLE_SIZE}, + vk_subbuffer{d_M, m_buf_offset, VK_WHOLE_SIZE}, + vk_subbuffer{d_D, d_buf_offset, VK_WHOLE_SIZE}, + }, + sizeof(vk_flash_attn_push_constants), &pc, { (uint32_t)neq1, (uint32_t)neq2, (uint32_t)neq3 }); +} + static vk_pipeline ggml_vk_op_get_pipeline(ggml_backend_vk_context * ctx, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst, ggml_op op) { switch (op) { case GGML_OP_GET_ROWS: @@ -4068,20 +4878,20 @@ static vk_pipeline ggml_vk_op_get_pipeline(ggml_backend_vk_context * ctx, const return nullptr; case GGML_OP_ADD: if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { - return ctx->device->pipeline_add_f32; + return ggml_are_same_shape(src0, src1) ? ctx->device->pipeline_add_f32_norepeat : ctx->device->pipeline_add_f32; } if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F16) { - return ctx->device->pipeline_add_f16_f32_f16; + return ggml_are_same_shape(src0, src1) ? ctx->device->pipeline_add_f16_f32_f16_norepeat : ctx->device->pipeline_add_f16_f32_f16; } return nullptr; case GGML_OP_MUL: if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { - return ctx->device->pipeline_mul_f32; + return ggml_are_same_shape(src0, src1) ? ctx->device->pipeline_mul_f32_norepeat : ctx->device->pipeline_mul_f32; } return nullptr; case GGML_OP_DIV: if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { - return ctx->device->pipeline_div_f32; + return ggml_are_same_shape(src0, src1) ? ctx->device->pipeline_div_f32_norepeat : ctx->device->pipeline_div_f32; } return nullptr; case GGML_OP_CONCAT: @@ -4138,7 +4948,7 @@ static vk_pipeline ggml_vk_op_get_pipeline(ggml_backend_vk_context * ctx, const case GGML_OP_CPY: case GGML_OP_CONT: case GGML_OP_DUP: - return ggml_vk_get_cpy_pipeline(ctx, src0->type, dst->type); + return ggml_vk_get_cpy_pipeline(ctx, src0, dst, dst->type); case GGML_OP_NORM: if (src0->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { return ctx->device->pipeline_norm_f32; @@ -4194,10 +5004,10 @@ static vk_pipeline ggml_vk_op_get_pipeline(ggml_backend_vk_context * ctx, const GGML_ASSERT(!src1 || src1->type == GGML_TYPE_F32 || src1->type == GGML_TYPE_F16); if (src0->type == GGML_TYPE_F32 && (src1 == nullptr || src1->type == GGML_TYPE_F32) && dst->type == GGML_TYPE_F32) { - return ctx->device->pipeline_soft_max_f32; + return src0->ne[0] > 1024 ? ctx->device->pipeline_soft_max_f32_wg512 : ctx->device->pipeline_soft_max_f32; } if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_F16 && dst->type == GGML_TYPE_F32) { - return ctx->device->pipeline_soft_max_f32_f16; + return src0->ne[0] > 1024 ? ctx->device->pipeline_soft_max_f32_f16_wg512 : ctx->device->pipeline_soft_max_f32_f16; } return nullptr; case GGML_OP_ROPE: @@ -4250,6 +5060,11 @@ static vk_pipeline ggml_vk_op_get_pipeline(ggml_backend_vk_context * ctx, const return ctx->device->pipeline_pool2d_f32; } return nullptr; + case GGML_OP_RWKV_WKV6: + if (src0->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { + return ctx->device->pipeline_rwkv_wkv6_f32; + } + return nullptr; case GGML_OP_LEAKY_RELU: if (src0->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32) { return ctx->device->pipeline_leaky_relu_f32; @@ -4271,7 +5086,6 @@ static bool ggml_vk_op_supports_incontiguous(ggml_op op) { case GGML_OP_DIV: case GGML_OP_CONCAT: case GGML_OP_UPSCALE: - case GGML_OP_SCALE: case GGML_OP_SQR: case GGML_OP_SIN: case GGML_OP_COS: @@ -4284,8 +5098,59 @@ static bool ggml_vk_op_supports_incontiguous(ggml_op op) { } } +static uint32_t get_misalign_bytes(ggml_backend_vk_context * ctx, const ggml_tensor * t) +{ + return ((vk_tensor_offset(t) + t->view_offs) & (ctx->device->properties.limits.minStorageBufferOffsetAlignment - 1));; +} + +template void init_pushconst_tensor_offsets(ggml_backend_vk_context * ctx, T &p, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst) { + GGML_UNUSED(p); + GGML_UNUSED(src0); + GGML_UNUSED(src1); + GGML_UNUSED(src2); + GGML_UNUSED(dst); + static_assert(!std::is_const::value, "unexpected type"); + GGML_ASSERT(!src0 || get_misalign_bytes(ctx, src0) == 0); + GGML_ASSERT(!src1 || get_misalign_bytes(ctx, src1) == 0); + GGML_ASSERT(!src2 || get_misalign_bytes(ctx, src2) == 0); + GGML_ASSERT(!dst || get_misalign_bytes(ctx, dst) == 0); +} + +template <> void init_pushconst_tensor_offsets(ggml_backend_vk_context * ctx, vk_op_unary_push_constants &p, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst) { + const uint32_t a_offset = get_misalign_bytes(ctx, src0) / ggml_type_size(src0->type); + const uint32_t d_offset = get_misalign_bytes(ctx, dst) / ggml_type_size(dst->type); + + p.misalign_offsets = (a_offset << 16) | d_offset; + + GGML_UNUSED(src1); + GGML_UNUSED(src2); +} + +template <> void init_pushconst_tensor_offsets(ggml_backend_vk_context * ctx, vk_op_binary_push_constants &p, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst) { + const uint32_t a_offset = get_misalign_bytes(ctx, src0) / ggml_type_size(src0->type); + const uint32_t b_offset = get_misalign_bytes(ctx, src1) / ggml_type_size(src1->type); + const uint32_t d_offset = get_misalign_bytes(ctx, dst) / ggml_type_size(dst->type); + + GGML_ASSERT(dst->op != GGML_OP_GET_ROWS || (a_offset == 0 && b_offset == 0 && d_offset == 0)); + + p.misalign_offsets = (a_offset << 16) | (b_offset << 8) | d_offset; + + GGML_UNUSED(src2); +} + +template <> void init_pushconst_tensor_offsets(ggml_backend_vk_context * ctx, vk_op_upscale_push_constants &p, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst) { + const uint32_t a_offset = get_misalign_bytes(ctx, src0) / ggml_type_size(src0->type); + const uint32_t d_offset = get_misalign_bytes(ctx, dst) / ggml_type_size(dst->type); + + p.a_offset = a_offset; + p.d_offset = d_offset; + + GGML_UNUSED(src1); + GGML_UNUSED(src2); +} + template -static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst, ggml_op op, const PC&& pc, bool dryrun = false) { +static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, const ggml_tensor * src2, ggml_tensor * dst, ggml_op op, PC&& pc, bool dryrun = false) { VK_LOG_DEBUG("ggml_vk_op_f32((" << src0 << ", name=" << src0->name << ", type=" << src0->type << ", ne0=" << src0->ne[0] << ", ne1=" << src0->ne[1] << ", ne2=" << src0->ne[2] << ", ne3=" << src0->ne[3] << ", nb0=" << src0->nb[0] << ", nb1=" << src0->nb[1] << ", nb2=" << src0->nb[2] << ", nb3=" << src0->nb[3]; if (src1 != nullptr) { std::cerr << "), (" << src1 << ", name=" << src1->name << ", type=" << src1->type << ", ne0=" << src1->ne[0] << ", ne1=" << src1->ne[1] << ", ne2=" << src1->ne[2] << ", ne3=" << src1->ne[3] << ", nb0=" << src1->nb[0] << ", nb1=" << src1->nb[1] << ", nb2=" << src1->nb[2] << ", nb3=" << src1->nb[3]; @@ -4325,6 +5190,8 @@ static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, co const uint64_t ned3 = dst->ne[3]; const uint64_t ned = ned0 * ned1; + init_pushconst_fastdiv(pc); + vk_pipeline pipeline = ggml_vk_op_get_pipeline(ctx, src0, src1, src2, dst, op); if (pipeline == nullptr) { @@ -4385,8 +5252,7 @@ static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, co } GGML_ASSERT(d_D != nullptr); - uint64_t d_buf_offset = ((vk_tensor_offset(dst) + dst->view_offs) / ctx->device->properties.limits.minStorageBufferOffsetAlignment) * ctx->device->properties.limits.minStorageBufferOffsetAlignment; - GGML_ASSERT(d_buf_offset == vk_tensor_offset(dst) || op == GGML_OP_CPY); // NOLINT + uint64_t d_buf_offset = vk_tensor_offset(dst) + dst->view_offs; if(!src0_uma) { d_X = src0_buf_ctx->dev_buffer; x_buf_offset = vk_tensor_offset(src0) + src0->view_offs; @@ -4402,6 +5268,12 @@ static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, co z_buf_offset = vk_tensor_offset(src2) + src2->view_offs; GGML_ASSERT(d_Z != nullptr); } + // Compute misalignment offset for descriptors and store it in in push constants, then align the descriptor offsets. + init_pushconst_tensor_offsets(ctx, pc, src0, src1, src2, dst); + x_buf_offset &= ~(ctx->device->properties.limits.minStorageBufferOffsetAlignment - 1); + y_buf_offset &= ~(ctx->device->properties.limits.minStorageBufferOffsetAlignment - 1); + z_buf_offset &= ~(ctx->device->properties.limits.minStorageBufferOffsetAlignment - 1); + d_buf_offset &= ~(ctx->device->properties.limits.minStorageBufferOffsetAlignment - 1); if (op_supports_incontiguous) { x_sz = ggml_nbytes(src0); @@ -4470,7 +5342,7 @@ static void ggml_vk_op_f32(ggml_backend_vk_context * ctx, vk_context& subctx, co const uint32_t OH = is_2D ? dst->ne[2] : 1; const uint32_t OW = dst->ne[1]; - const uint32_t batch = src1->ne[3]; + const uint32_t batch = src1->ne[is_2D ? 3 : 2]; elements = { OW * KW * KH, OH, batch * IC }; } break; @@ -4589,7 +5461,6 @@ static void ggml_vk_acc(ggml_backend_vk_context * ctx, vk_context& subctx, const const uint32_t src0_type_size = ggml_type_size(src0->type); const uint32_t src1_type_size = ggml_type_size(src1->type); const uint32_t dst_type_size = ggml_type_size(dst->type); - const uint32_t d_offset = ((vk_tensor_offset(dst) + dst->view_offs) % ctx->device->properties.limits.minStorageBufferOffsetAlignment) / dst_type_size; int nb1 = dst->op_params[0] / 4; // 4 bytes of float32 int nb2 = dst->op_params[1] / 4; // 4 bytes of float32 @@ -4601,7 +5472,7 @@ static void ggml_vk_acc(ggml_backend_vk_context * ctx, vk_context& subctx, const (uint32_t)src0->ne[0], (uint32_t)src0->ne[1], (uint32_t)src0->ne[2],(uint32_t)src0->ne[3], (uint32_t)src0->nb[0] / src0_type_size, (uint32_t)nb1, (uint32_t)nb2, (uint32_t)src0->nb[3] / src0_type_size, (uint32_t)src1->ne[0], (uint32_t)src1->ne[1], (uint32_t)src1->ne[2],(uint32_t)src1->ne[3], (uint32_t)src1->nb[0] / src1_type_size, (uint32_t)src1->nb[1] / src1_type_size, (uint32_t)src1->nb[2] / src1_type_size, (uint32_t)src1->nb[3] / src1_type_size, (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2],(uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t)nb1, (uint32_t)nb2, (uint32_t) dst->nb[3] / dst_type_size, - d_offset, + 0, 0.0f, 0.0f, offset, }, dryrun); } @@ -4651,6 +5522,134 @@ static void ggml_vk_div(ggml_backend_vk_context * ctx, vk_context& subctx, const }, dryrun); } +static void ggml_vk_op_f32_rwkv6(ggml_backend_vk_context * ctx, vk_context& subctx, ggml_tensor * dst, const vk_op_rwkv_wkv6_push_constants&& pc, bool dryrun = false) { + const ggml_tensor * k = dst->src[0]; + const ggml_tensor * v = dst->src[1]; + const ggml_tensor * r = dst->src[2]; + const ggml_tensor * tf = dst->src[3]; + const ggml_tensor * td = dst->src[4]; + const ggml_tensor * state = dst->src[5]; + + GGML_ASSERT(!ggml_is_quantized(k->type)); + GGML_ASSERT(!ggml_is_quantized(v->type)); + GGML_ASSERT(!ggml_is_quantized(r->type)); + GGML_ASSERT(!ggml_is_quantized(tf->type)); + GGML_ASSERT(!ggml_is_quantized(td->type)); + GGML_ASSERT(!ggml_is_quantized(state->type)); + GGML_ASSERT(dst->buffer != nullptr); + + vk_pipeline pipeline = ggml_vk_op_get_pipeline(ctx, k, v, r, dst, GGML_OP_RWKV_WKV6); + GGML_ASSERT(pipeline != nullptr); + + if (dryrun) { + ggml_pipeline_request_descriptor_sets(ctx->device, pipeline, 1); + return; + } + + ggml_backend_vk_buffer_context * dst_buf_ctx = (ggml_backend_vk_buffer_context *)dst->buffer->context; + ggml_backend_vk_buffer_context * k_buf_ctx = (ggml_backend_vk_buffer_context *)k->buffer->context; + ggml_backend_vk_buffer_context * v_buf_ctx = (ggml_backend_vk_buffer_context *)v->buffer->context; + ggml_backend_vk_buffer_context * r_buf_ctx = (ggml_backend_vk_buffer_context *)r->buffer->context; + ggml_backend_vk_buffer_context * tf_buf_ctx = (ggml_backend_vk_buffer_context *)tf->buffer->context; + ggml_backend_vk_buffer_context * td_buf_ctx = (ggml_backend_vk_buffer_context *)td->buffer->context; + ggml_backend_vk_buffer_context * state_buf_ctx = (ggml_backend_vk_buffer_context *)state->buffer->context; + + ggml_vk_sync_buffers(subctx); + + vk_buffer d_D = nullptr, d_K = nullptr, d_V = nullptr, d_R = nullptr, d_TF = nullptr, d_TD = nullptr, d_State = nullptr; + size_t k_offset = 0, v_offset = 0, r_offset = 0, tf_offset = 0, td_offset = 0, state_offset = 0, dst_offset = 0; + bool K_uma = false, V_uma = false, R_uma = false, TF_uma = false, TD_uma = false, STATE_uma = false, DST_uma = false; + + if (ctx->device->uma) { + ggml_vk_host_get(ctx->device, k->data, d_K, k_offset); + ggml_vk_host_get(ctx->device, v->data, d_V, v_offset); + ggml_vk_host_get(ctx->device, r->data, d_R, r_offset); + ggml_vk_host_get(ctx->device, tf->data, d_TF, tf_offset); + ggml_vk_host_get(ctx->device, td->data, d_TD, td_offset); + ggml_vk_host_get(ctx->device, state->data, d_State, state_offset); + ggml_vk_host_get(ctx->device, dst->data, d_D, dst_offset); + + K_uma = d_K != nullptr; + V_uma = d_V != nullptr; + R_uma = d_R != nullptr; + TF_uma = d_TF != nullptr; + TD_uma = d_TD != nullptr; + STATE_uma = d_State != nullptr; + DST_uma = d_D != nullptr; + } + + if (!K_uma) { + d_K = k_buf_ctx->dev_buffer; + k_offset = vk_tensor_offset(k) + k->view_offs; + } + if (!V_uma) { + d_V = v_buf_ctx->dev_buffer; + v_offset = vk_tensor_offset(v) + v->view_offs; + } + if (!R_uma) { + d_R = r_buf_ctx->dev_buffer; + r_offset = vk_tensor_offset(r) + r->view_offs; + } + if (!TF_uma) { + d_TF = tf_buf_ctx->dev_buffer; + tf_offset = vk_tensor_offset(tf) + tf->view_offs; + } + if (!TD_uma) { + d_TD = td_buf_ctx->dev_buffer; + td_offset = vk_tensor_offset(td) + td->view_offs; + } + if (!STATE_uma) { + d_State = state_buf_ctx->dev_buffer; + state_offset = vk_tensor_offset(state) + state->view_offs; + } + if (!DST_uma) { + d_D = dst_buf_ctx->dev_buffer; + dst_offset = vk_tensor_offset(dst) + dst->view_offs; + } + + const uint64_t k_size = ggml_nbytes(k); + const uint64_t v_size = ggml_nbytes(v); + const uint64_t r_size = ggml_nbytes(r); + const uint64_t tf_size = ggml_nbytes(tf); + const uint64_t td_size = ggml_nbytes(td); + const uint64_t state_size = ggml_nbytes(state); + const uint64_t dst_size = ggml_nbytes(dst); + + std::array elements = { + (uint32_t)(pc.B * pc.H), + 1, + 1 + }; + + ggml_vk_dispatch_pipeline(ctx, subctx, pipeline, { + vk_subbuffer{ d_K, k_offset, k_size }, + vk_subbuffer{ d_V, v_offset, v_size }, + vk_subbuffer{ d_R, r_offset, r_size }, + vk_subbuffer{ d_TF, tf_offset, tf_size }, + vk_subbuffer{ d_TD, td_offset, td_size }, + vk_subbuffer{ d_State, state_offset, state_size }, + vk_subbuffer{ d_D, dst_offset, dst_size } + }, sizeof(vk_op_rwkv_wkv6_push_constants), &pc, elements); +} + +static void ggml_vk_rwkv_wkv6(ggml_backend_vk_context * ctx, vk_context& subctx, ggml_tensor * dst, bool dryrun = false) { + const size_t seq_length = dst->src[0]->ne[2]; + const size_t n_embed = dst->ne[0]; + const size_t n_heads = dst->src[0]->ne[1]; + const size_t n_seqs = dst->src[5]->ne[1]; + + ggml_vk_op_f32_rwkv6( + ctx, subctx, dst, + { + (uint32_t)n_seqs, + (uint32_t)seq_length, + (uint32_t)n_embed, + (uint32_t)n_heads, + }, + dryrun + ); +} + static void ggml_vk_concat(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, bool dryrun = false) { int * op_params = (int *)dst->op_params; @@ -4677,7 +5676,7 @@ static void ggml_vk_upscale(ggml_backend_vk_context * ctx, vk_context& subctx, c const float sf3 = (float)dst->ne[3] / src0->ne[3]; ggml_vk_op_f32(ctx, subctx, src0, nullptr, nullptr, dst, GGML_OP_UPSCALE, { - (uint32_t)ggml_nelements(dst), 0, + (uint32_t)ggml_nelements(dst), 0, 0, (uint32_t)src0->nb[0] / src0_type_size, (uint32_t)src0->nb[1] / src0_type_size, (uint32_t)src0->nb[2] / src0_type_size, (uint32_t)src0->nb[3] / src0_type_size, (uint32_t)dst->ne[0], (uint32_t)dst->ne[1], (uint32_t)dst->ne[2],(uint32_t)dst->ne[3], sf0, sf1, sf2, sf3, @@ -4694,7 +5693,8 @@ static void ggml_vk_scale(ggml_backend_vk_context * ctx, vk_context& subctx, con (uint32_t)src0->ne[0], (uint32_t)src0->ne[1], (uint32_t)src0->ne[2], (uint32_t)src0->ne[3], (uint32_t)src0->nb[0] / src0_type_size, (uint32_t)src0->nb[1] / src0_type_size, (uint32_t)src0->nb[2] / src0_type_size, (uint32_t)src0->nb[3] / src0_type_size, (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, - op_params[0], 0.0f + op_params[0], 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4708,6 +5708,7 @@ static void ggml_vk_sqr(ggml_backend_vk_context * ctx, vk_context& subctx, const (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4721,6 +5722,7 @@ static void ggml_vk_sin(ggml_backend_vk_context * ctx, vk_context& subctx, const (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4734,6 +5736,7 @@ static void ggml_vk_cos(ggml_backend_vk_context * ctx, vk_context& subctx, const (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4748,6 +5751,7 @@ static void ggml_vk_clamp(ggml_backend_vk_context * ctx, vk_context& subctx, con (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, op_params[0], op_params[1], + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4761,6 +5765,7 @@ static void ggml_vk_pad(ggml_backend_vk_context * ctx, vk_context& subctx, const (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4774,20 +5779,21 @@ static void ggml_vk_repeat(ggml_backend_vk_context * ctx, vk_context& subctx, co (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } static void ggml_vk_cpy(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, ggml_tensor * dst, bool dryrun = false) { const uint32_t src0_type_size = ggml_type_size(src0->type); const uint32_t dst_type_size = ggml_type_size(dst->type); - const uint32_t d_offset = ((vk_tensor_offset(dst) + dst->view_offs) % ctx->device->properties.limits.minStorageBufferOffsetAlignment) / dst_type_size; ggml_vk_op_f32(ctx, subctx, src0, nullptr, nullptr, dst, GGML_OP_CPY, { (uint32_t)ggml_nelements(src0), (uint32_t)src0->ne[0], (uint32_t)src0->ne[1], (uint32_t)src0->ne[2], (uint32_t)src0->ne[3], (uint32_t)src0->nb[0] / src0_type_size, (uint32_t)src0->nb[1] / src0_type_size, (uint32_t)src0->nb[2] / src0_type_size, (uint32_t)src0->nb[3] / src0_type_size, (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], (uint32_t) dst->nb[0] / dst_type_size, (uint32_t) dst->nb[1] / dst_type_size, (uint32_t) dst->nb[2] / dst_type_size, (uint32_t) dst->nb[3] / dst_type_size, - d_offset, + 0, 0.0f, 0.0f, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, dryrun); } @@ -4844,6 +5850,7 @@ static void ggml_vk_soft_max(ggml_backend_vk_context * ctx, vk_context& subctx, scale, max_bias, m0, m1, n_head_log2, + nrows_x, }, dryrun); } @@ -4915,7 +5922,7 @@ static void ggml_vk_im2col(ggml_backend_vk_context * ctx, vk_context& subctx, co const uint32_t OW = dst->ne[1]; const uint32_t offset_delta = src1->nb[is_2D ? 2 : 1] / 4; // nb is byte offset, src is type float32 - const uint32_t batch_offset = src1->nb[3] / 4; // nb is byte offset, src is type float32 + const uint32_t batch_offset = src1->nb[is_2D ? 3 : 2] / 4; // nb is byte offset, src is type float32 const uint32_t pelements = OW * KW * KH; @@ -5022,10 +6029,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->a_s; shname = "F32_F16_ALIGNED_S"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->a_s; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->a_s; shname = "F16_F32_ALIGNED_S"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->a_s; + p = ctx->device->pipeline_matmul_f16.f32acc->a_s; shname = "F16_ALIGNED_S"; } else { GGML_ABORT("fatal error"); @@ -5038,10 +6045,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->a_m; shname = "F32_F16_ALIGNED_M"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->a_m; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->a_m; shname = "F16_F32_ALIGNED_M"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->a_m; + p = ctx->device->pipeline_matmul_f16.f32acc->a_m; shname = "F16_ALIGNED_M"; } else { GGML_ABORT("fatal error"); @@ -5054,10 +6061,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->a_l; shname = "F32_F16_ALIGNED_L"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->a_l; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->a_l; shname = "F16_F32_ALIGNED_L"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->a_l; + p = ctx->device->pipeline_matmul_f16.f32acc->a_l; shname = "F16_ALIGNED_L"; } else { GGML_ABORT("fatal error"); @@ -5077,10 +6084,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->s; shname = "F32_F16_S"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->s; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->s; shname = "F16_F32_S"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->s; + p = ctx->device->pipeline_matmul_f16.f32acc->s; shname = "F16_S"; } } else if (shader_size == 1) { @@ -5091,10 +6098,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->m; shname = "F32_F16_M"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->m; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->m; shname = "F16_F32_M"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->m; + p = ctx->device->pipeline_matmul_f16.f32acc->m; shname = "F16_M"; } } else if (shader_size == 2) { @@ -5105,10 +6112,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t p = ctx->device->pipeline_matmul_f32_f16->l; shname = "F32_F16_L"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16_f32->l; + p = ctx->device->pipeline_matmul_f16_f32.f32acc->l; shname = "F16_F32_L"; } else if (std::is_same() && std::is_same()) { - p = ctx->device->pipeline_matmul_f16->l; + p = ctx->device->pipeline_matmul_f16.f32acc->l; shname = "F16_L"; } } @@ -5140,19 +6147,27 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t for (size_t i = 0; i < x_ne; i++) { if (std::is_same()) { x[i] = (rand() / (float)RAND_MAX) * 2.0f - 1.0f; + // x[i] = 1.0f; + // x[i] = i + 1; + // x[i] = (i % k == i / k) ? 1.0f : 0.0f; } else if (std::is_same()) { x[i] = ggml_fp32_to_fp16((rand() / (float)RAND_MAX) * 2.0f - 1.0f); + // x[i] = ggml_fp32_to_fp16(1.0f); + // x[i] = ggml_fp32_to_fp16(i + 1); + // x[i] = ggml_fp32_to_fp16((i % k == i / k) ? 1.0f : 0.0f); } else { GGML_ABORT("fatal error"); } } for (size_t i = 0; i < y_ne; i++) { if (std::is_same()) { - // y[i] = (rand() / (float)RAND_MAX) * 2.0f - 1.0f; - y[i] = (i % k == i / k) ? 1.0f : 0.0f; + y[i] = (rand() / (float)RAND_MAX) * 2.0f - 1.0f; + // y[i] = (i % k == i / k) ? 1.0f : 0.0f; + // y[i] = i + 1; } else if (std::is_same()) { - // y[i] = ggml_fp32_to_fp16((rand() / (float)RAND_MAX) * 2.0f - 1.0f); - y[i] = ggml_fp32_to_fp16((i % k == i / k) ? 1.0f : 0.0f); + y[i] = ggml_fp32_to_fp16((rand() / (float)RAND_MAX) * 2.0f - 1.0f); + // y[i] = ggml_fp32_to_fp16((i % k == i / k) ? 1.0f : 0.0f); + // y[i] = ggml_fp32_to_fp16(i + 1); } else { GGML_ABORT("fatal error"); } @@ -5162,16 +6177,16 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t ggml_vk_buffer_write(d_Y, 0, y, sizeof(Y_TYPE) * k * n * batch); vk_context subctx = ggml_vk_create_context(ctx, ctx->device->compute_queue); + ggml_vk_ctx_begin(ctx->device, subctx); for (size_t i = 0; i < num_it; i++) { - ggml_vk_ctx_begin(ctx->device, subctx); ggml_vk_matmul( ctx, subctx, p, ggml_vk_subbuffer(d_X), ggml_vk_subbuffer(d_Y), ggml_vk_subbuffer(d_D), ggml_vk_subbuffer(ctx->prealloc_split_k), m, n, k, k, k, m, k*m, k*n, m*n, split_k, batch, batch, batch, 1, 1 ); - ggml_vk_ctx_end(subctx); } + ggml_vk_ctx_end(subctx); auto begin = std::chrono::high_resolution_clock::now(); ggml_vk_submit(subctx, ctx->fence); @@ -5236,7 +6251,7 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t double err = std::fabs(d[i] - d_chk[i]); avg_err += err; - if (err > 0.05f && first_err_n == -1) { + if ((err > 0.05f || std::isnan(err)) && first_err_n == -1) { first_err_b = i / (m * n); first_err_n = (i % (m * n)) / m; first_err_m = (i % (m * n)) % m; @@ -5249,12 +6264,10 @@ static void ggml_vk_test_matmul(ggml_backend_vk_context * ctx, size_t m, size_t std::cerr << "TEST " << shname << " m=" << m << " n=" << n << " k=" << k << " batch=" << batch << " split_k=" << split_k << " matmul " << time / num_it << "ms " << tflops << " TFLOPS avg_err=" << avg_err << std::endl; - if (avg_err > 0.1) { + if (avg_err > 0.1 || std::isnan(avg_err)) { std::cerr << "m = " << first_err_m << " n = " << first_err_n << " b = " << first_err_b << std::endl; std::cerr << "Actual result: " << std::endl << std::endl; ggml_vk_print_matrix_area(d, GGML_TYPE_F32, m, n, first_err_m, first_err_n, first_err_b); - std::cerr << std::endl; - ggml_vk_print_matrix_area(d, GGML_TYPE_F32, m, n, first_err_m, first_err_n + 15, first_err_b); std::cerr << "Expected result: " << std::endl << std::endl; ggml_vk_print_matrix_area(d_chk, GGML_TYPE_F32, m, n, first_err_m, first_err_n, first_err_b); @@ -5437,13 +6450,13 @@ static void ggml_vk_test_dequant_matmul(ggml_backend_vk_context * ctx, size_t m, vk_pipeline p; std::string shname; if (shader_size == 0) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->a_s; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->a_s : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->a_s; shname = std::string(ggml_type_name(quant)) + "_ALIGNED_S"; } else if (shader_size == 1) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->a_m; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->a_m : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->a_m; shname = std::string(ggml_type_name(quant)) + "_ALIGNED_M"; } else if (shader_size == 2) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->a_l; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->a_l : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->a_l; shname = std::string(ggml_type_name(quant)) + "_ALIGNED_L"; } else { GGML_ASSERT(0); @@ -5453,13 +6466,13 @@ static void ggml_vk_test_dequant_matmul(ggml_backend_vk_context * ctx, size_t m, if (k != kpad) { if (shader_size == 0) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->s; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->s : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->s; shname = std::string(ggml_type_name(quant)) + "_S"; } else if (shader_size == 1) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->m; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->m : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->m; shname = std::string(ggml_type_name(quant)) + "_M"; } else if (shader_size == 2) { - p = ctx->device->pipeline_dequant_mul_mat_mat[quant]->l; + p = ctx->device->fp16 ? ctx->device->pipeline_dequant_mul_mat_mat[quant].f16acc->l : ctx->device->pipeline_dequant_mul_mat_mat[quant].f32acc->l; shname = std::string(ggml_type_name(quant)) + "_L"; } else { GGML_ASSERT(0); @@ -5509,16 +6522,16 @@ static void ggml_vk_test_dequant_matmul(ggml_backend_vk_context * ctx, size_t m, ggml_vk_buffer_write(y_buf, 0, y, y_sz); vk_context subctx = ggml_vk_create_context(ctx, ctx->device->compute_queue); + ggml_vk_ctx_begin(ctx->device, subctx); for (size_t i = 0; i < num_it; i++) { - ggml_vk_ctx_begin(ctx->device, subctx); ggml_vk_matmul( ctx, subctx, p, ggml_vk_subbuffer(qx_buf), ggml_vk_subbuffer(y_buf), ggml_vk_subbuffer(d_buf), ggml_vk_subbuffer(ctx->prealloc_split_k), m, n, k, k, k, m, k*m, k*n, m*n, split_k, batch, batch, batch, 1, 1 ); - ggml_vk_ctx_end(subctx); } + ggml_vk_ctx_end(subctx); auto begin = std::chrono::high_resolution_clock::now(); @@ -5618,105 +6631,13 @@ static void ggml_vk_test_dequant_matmul(ggml_backend_vk_context * ctx, size_t m, static void ggml_vk_preallocate_buffers(ggml_backend_vk_context * ctx) { #if defined(GGML_VULKAN_RUN_TESTS) - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_F32); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q4_0); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q4_1); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q5_0); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q5_1); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q8_0); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q2_K); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q3_K); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q4_K); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q5_K); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_Q6_K); - ggml_vk_test_dequant(ctx, 7680, GGML_TYPE_IQ4_NL); - - ggml_vk_test_matmul(ctx, 512, 512, 100, 32, 100, 1, 2); - - ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 1, 0); - ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 1, 1); - ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 1, 2); - // ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 4, 0); - // ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 4, 1); - // ggml_vk_test_matmul(ctx, 128, 512, 512, 2, 100, 4, 2); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q4_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q4_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q4_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q4_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q4_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q4_0); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q4_1); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q4_1); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q4_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q4_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q4_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q4_1); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q5_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q5_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q5_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q5_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q5_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q5_0); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q5_1); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q5_1); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q5_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q5_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q5_1); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q5_1); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q8_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q8_0); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q8_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q8_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q8_0); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q8_0); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q2_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q2_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q2_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q2_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q2_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q2_K); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q3_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q3_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q3_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q3_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q3_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q3_K); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q4_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q4_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q4_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q4_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q4_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q4_K); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q5_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q5_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q5_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q5_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q5_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q5_K); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_Q6_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_Q6_K); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_Q6_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 0, GGML_TYPE_Q6_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 1, GGML_TYPE_Q6_K); - // ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 4, 2, GGML_TYPE_Q6_K); - - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 0, GGML_TYPE_IQ4_NL); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 1, GGML_TYPE_IQ4_NL); - ggml_vk_test_dequant_matmul(ctx, 128, 512, 512, 2, 100, 1, 2, GGML_TYPE_IQ4_NL); - - std::cerr << std::endl; - const std::vector vals { + 512, 512, 128, + 128, 512, 512, + 4096, 512, 4096, + 11008, 512, 4096, + 4096, 512, 11008, + 32000, 512, 4096, 8, 8, 8, 100, 46, 576, 623, 111, 128, @@ -5729,25 +6650,52 @@ static void ggml_vk_preallocate_buffers(ggml_backend_vk_context * ctx) { 49, 49, 128, 128, 49, 49, 4096, 49, 4096, - 11008, 49, 4096, - 4096, 49, 11008, - 32000, 49, 4096, - 512, 512, 128, - 128, 512, 512, - 4096, 512, 4096, - 11008, 512, 4096, - 4096, 512, 11008, - 32000, 512, 4096, }; - const size_t num_it = 1; + const size_t num_it = 100; + for (size_t i = 0; i < vals.size(); i += 3) { ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 0); ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 1); ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 2); - // ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 0); - // ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 1); - // ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 2); - std::cerr << std::endl; + std::cerr << '\n'; + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 0); + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 1); + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 2); + std::cerr << '\n'; + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 0); + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 1); + ggml_vk_test_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 2); + std::cerr << '\n' << std::endl; + + if (vals[i + 2] % 32 == 0) { + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 0, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 1, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 2, GGML_TYPE_Q4_0); + std::cerr << '\n'; + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 0, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 1, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 2, GGML_TYPE_Q4_0); + std::cerr << '\n'; + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 0, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 1, GGML_TYPE_Q4_0); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 2, GGML_TYPE_Q4_0); + std::cerr << '\n' << std::endl; + } + + if (vals[i + 2] % 256 == 0) { + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 0, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 1, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 1, 2, GGML_TYPE_Q4_K); + std::cerr << '\n'; + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 0, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 1, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 2, 2, GGML_TYPE_Q4_K); + std::cerr << '\n'; + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 0, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 1, GGML_TYPE_Q4_K); + ggml_vk_test_dequant_matmul(ctx, vals[i], vals[i + 1], vals[i + 2], 2, num_it, 4, 2, GGML_TYPE_Q4_K); + std::cerr << '\n' << std::endl; + } } GGML_ABORT("fatal error"); @@ -5794,6 +6742,7 @@ static bool ggml_vk_build_graph(ggml_backend_vk_context * ctx, ggml_tensor * nod const ggml_tensor * src0 = node->src[0]; const ggml_tensor * src1 = node->src[1]; const ggml_tensor * src2 = node->src[2]; + const ggml_tensor * src3 = node->src[3]; switch (node->op) { // Return on empty ops to avoid generating a compute_ctx and setting exit_tensor @@ -5845,7 +6794,9 @@ static bool ggml_vk_build_graph(ggml_backend_vk_context * ctx, ggml_tensor * nod case GGML_OP_IM2COL: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_POOL_2D: + case GGML_OP_RWKV_WKV6: case GGML_OP_LEAKY_RELU: + case GGML_OP_FLASH_ATTN_EXT: break; default: std::cerr << "ggml_vulkan: Error: Missing op: " << ggml_op_name(node->op) << std::endl; @@ -5863,6 +6814,48 @@ static bool ggml_vk_build_graph(ggml_backend_vk_context * ctx, ggml_tensor * nod } else { compute_ctx = ctx->compute_ctx.lock(); } + } else { + switch (node->op) { + case GGML_OP_REPEAT: + case GGML_OP_ACC: + case GGML_OP_GET_ROWS: + case GGML_OP_ADD: + case GGML_OP_MUL: + case GGML_OP_DIV: + case GGML_OP_CONCAT: + case GGML_OP_UPSCALE: + case GGML_OP_SCALE: + case GGML_OP_SQR: + case GGML_OP_SIN: + case GGML_OP_COS: + case GGML_OP_CLAMP: + case GGML_OP_PAD: + case GGML_OP_CPY: + case GGML_OP_CONT: + case GGML_OP_DUP: + case GGML_OP_NORM: + case GGML_OP_GROUP_NORM: + case GGML_OP_RMS_NORM: + case GGML_OP_UNARY: + case GGML_OP_DIAG_MASK_INF: + case GGML_OP_SOFT_MAX: + case GGML_OP_ROPE: + case GGML_OP_ARGSORT: + case GGML_OP_SUM_ROWS: + case GGML_OP_IM2COL: + case GGML_OP_TIMESTEP_EMBEDDING: + case GGML_OP_POOL_2D: + case GGML_OP_LEAKY_RELU: + { + // These operations all go through ggml_vk_op_f32, so short-circuit and + // do the only thing needed for the dryrun. + vk_pipeline pipeline = ggml_vk_op_get_pipeline(ctx, src0, src1, src2, node, node->op); + ggml_pipeline_request_descriptor_sets(ctx->device, pipeline, 1); + return false; + } + default: + break; + } } switch (node->op) { @@ -5996,6 +6989,16 @@ static bool ggml_vk_build_graph(ggml_backend_vk_context * ctx, ggml_tensor * nod case GGML_OP_MUL_MAT_ID: ggml_vk_mul_mat_id(ctx, compute_ctx, src0, src1, src2, node, dryrun); + break; + + case GGML_OP_FLASH_ATTN_EXT: + ggml_vk_flash_attn(ctx, compute_ctx, src0, src1, src2, src3, node, dryrun); + + break; + + case GGML_OP_RWKV_WKV6: + ggml_vk_rwkv_wkv6(ctx, compute_ctx, node, dryrun); + break; default: return false; @@ -6076,6 +7079,7 @@ static bool ggml_vk_compute_forward(ggml_backend_vk_context * ctx, ggml_tensor * case GGML_OP_IM2COL: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_POOL_2D: + case GGML_OP_RWKV_WKV6: case GGML_OP_LEAKY_RELU: case GGML_OP_REPEAT: buf = tensor->buffer; @@ -6096,6 +7100,7 @@ static bool ggml_vk_compute_forward(ggml_backend_vk_context * ctx, ggml_tensor * break; case GGML_OP_MUL_MAT: case GGML_OP_MUL_MAT_ID: + case GGML_OP_FLASH_ATTN_EXT: buf = tensor->buffer; break; @@ -6592,16 +7597,17 @@ static ggml_status ggml_backend_vk_graph_compute(ggml_backend_t backend, ggml_cg bool first_node_in_batch = true; // true if next node will be first node in a batch int submit_node_idx = 0; // index to first node in a batch - // submit work every submit_count node to overlap CPU cmdbuffer generation with GPU execution - constexpr int submit_count = 100; + // Submit work every nodes_per_submit nodes to overlap CPU cmdbuffer generation with GPU execution. + // Start with a smaller count to get work submitted right away, and increase it after each submit. + int nodes_per_submit = 20; int submitted_nodes = 0; + int submit_count = 0; for (int i = 0; i < cgraph->n_nodes; i++) { if (first_node_in_batch) { submit_node_idx = i; } - bool submit = (submitted_nodes >= submit_count) || (i == last_node); - + bool submit = (submitted_nodes >= nodes_per_submit) || (i == last_node); bool enqueued = ggml_vk_build_graph(ctx, cgraph->nodes[i], i, cgraph->nodes[submit_node_idx], submit_node_idx, false, i == last_node, submit); @@ -6618,6 +7624,15 @@ static ggml_status ggml_backend_vk_graph_compute(ggml_backend_t backend, ggml_cg if (submit) { first_node_in_batch = true; submitted_nodes = 0; + switch (submit_count) { + case 0: + nodes_per_submit = 50; + break; + default: + nodes_per_submit = 100; + break; + } + submit_count++; } } @@ -6774,6 +7789,12 @@ static bool ggml_backend_vk_device_supports_op(ggml_backend_dev_t dev, const ggm case GGML_OP_MUL_MAT: case GGML_OP_MUL_MAT_ID: { + ggml_backend_vk_device_context * ctx = (ggml_backend_vk_device_context *)dev->context; + const vk_device& device = ggml_vk_get_device(ctx->device); + if (op->op == GGML_OP_MUL_MAT_ID && !device->mul_mat_id_s && !device->mul_mat_id_m && !device->mul_mat_id_l) { + // If there's not enough shared memory for row_ids and the result tile, fallback to CPU + return false; + } switch (op->src[0]->type) { case GGML_TYPE_F32: case GGML_TYPE_F16: @@ -6804,8 +7825,64 @@ static bool ggml_backend_vk_device_supports_op(ggml_backend_dev_t dev, const ggm if (a->ne[3] != b->ne[3]) { return false; } + if (!(ggml_vk_dim01_contiguous(op->src[0]) || op->src[0]->type == GGML_TYPE_F32 || op->src[0]->type == GGML_TYPE_F16) || + !(ggml_vk_dim01_contiguous(op->src[1]) || op->src[1]->type == GGML_TYPE_F32 || op->src[1]->type == GGML_TYPE_F16)) { + return false; + } + return true; } break; + case GGML_OP_FLASH_ATTN_EXT: + { + ggml_backend_vk_device_context * ctx = (ggml_backend_vk_device_context *)dev->context; + if (!ggml_vk_get_device(ctx->device)->coopmat2) { + return false; + } + switch (op->src[0]->ne[0]) { + case 64: + case 80: + case 96: + case 112: + case 128: + case 256: + break; + default: + return false; + } + if (op->src[0]->type != GGML_TYPE_F32) { + return false; + } + if (op->type != GGML_TYPE_F32) { + return false; + } + if (op->src[3] && op->src[3]->type != GGML_TYPE_F16) { + return false; + } + // It's straightforward to support different K/V dequant, but would + // significantly increase the number of pipelines + if (op->src[1]->type != op->src[2]->type) { + return false; + } + switch (op->src[1]->type) { + case GGML_TYPE_F16: + case GGML_TYPE_Q4_0: + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q5_0: + case GGML_TYPE_Q5_1: + case GGML_TYPE_Q8_0: + // K dequants currently disabled because D dimension is rounded up to 256 and runs inefficiently + //case GGML_TYPE_Q2_K: + //case GGML_TYPE_Q3_K: + //case GGML_TYPE_Q4_K: + //case GGML_TYPE_Q5_K: + //case GGML_TYPE_Q6_K: + case GGML_TYPE_IQ4_NL: + break; + default: + return false; + } + return true; + } case GGML_OP_GET_ROWS: { switch (op->src[0]->type) { @@ -6842,7 +7919,16 @@ static bool ggml_backend_vk_device_supports_op(ggml_backend_dev_t dev, const ggm case GGML_OP_REPEAT: return ggml_type_size(op->type) == sizeof(float) && ggml_type_size(op->src[0]->type) == sizeof(float); case GGML_OP_ROPE: - return ggml_is_contiguous(op->src[0]); + { + const int mode = ((const int32_t *) op->op_params)[2]; + if (mode & GGML_ROPE_TYPE_MROPE) { + return false; + } + if (mode & GGML_ROPE_TYPE_VISION) { + return false; + } + return ggml_is_contiguous(op->src[0]); + } case GGML_OP_NONE: case GGML_OP_RESHAPE: case GGML_OP_VIEW: @@ -6870,6 +7956,7 @@ static bool ggml_backend_vk_device_supports_op(ggml_backend_dev_t dev, const ggm case GGML_OP_IM2COL: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_POOL_2D: + case GGML_OP_RWKV_WKV6: case GGML_OP_LEAKY_RELU: return true; default: @@ -6966,8 +8053,9 @@ static const struct ggml_backend_reg_i ggml_backend_vk_reg_i = { ggml_backend_reg_t ggml_backend_vk_reg() { static ggml_backend_reg reg = { - /* .iface = */ ggml_backend_vk_reg_i, - /* .context = */ nullptr, + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_vk_reg_i, + /* .context = */ nullptr, }; return ® @@ -7009,6 +8097,25 @@ static bool ggml_vk_instance_portability_enumeration_ext_available(const std::ve UNUSED(instance_extensions); } +static bool ggml_vk_khr_cooperative_matrix_support(const vk::PhysicalDeviceProperties& props, const vk::PhysicalDeviceDriverProperties& driver_props) { + switch (props.vendorID) { + case VK_VENDOR_ID_INTEL: + // Intel drivers don't support coopmat properly yet + return false; + case VK_VENDOR_ID_AMD: + if (driver_props.driverID == vk::DriverId::eAmdProprietary || driver_props.driverID == vk::DriverId::eAmdOpenSource) { + // Workaround for AMD proprietary driver reporting support on all GPUs + const std::string name = props.deviceName; + return name.rfind("AMD Radeon RX 7", 0) == 0 || name.rfind("AMD Radeon(TM) RX 7", 0) == 0 || // RDNA 3 consumer GPUs + name.rfind("AMD Radeon PRO W7", 0) == 0 || name.rfind("AMD Radeon(TM) PRO W7", 0) == 0 || // RDNA 3 workstation GPUs + name.rfind("AMD Radeon 7", 0) == 0 || name.rfind("AMD Radeon(TM) 7", 0) == 0; // RDNA 3 APUs + } + return true; + default: + return true; + } +} + // checks #ifdef GGML_VULKAN_CHECK_RESULTS @@ -7119,6 +8226,7 @@ static void ggml_vk_check_results_0(ggml_tensor * tensor) { ggml_tensor * src0 = tensor->src[0]; ggml_tensor * src1 = tensor->src[1]; ggml_tensor * src2 = tensor->src[2]; + ggml_tensor * src3 = tensor->src[3]; struct ggml_init_params iparams = { /*.mem_size =*/ 2ul*1024ul*1024ul*1024ul, @@ -7131,15 +8239,18 @@ static void ggml_vk_check_results_0(ggml_tensor * tensor) { struct ggml_tensor * src0_clone = nullptr; struct ggml_tensor * src1_clone = nullptr; struct ggml_tensor * src2_clone = nullptr; + struct ggml_tensor * src3_clone = nullptr; struct ggml_tensor * tensor_clone = nullptr; size_t src0_size; size_t src1_size; size_t src2_size; + size_t src3_size; void * src0_buffer = nullptr; void * src1_buffer = nullptr; void * src2_buffer = nullptr; + void * src3_buffer = nullptr; if (src0 != nullptr) { src0_clone = ggml_dup_tensor(ggml_ctx, src0); @@ -7267,8 +8378,53 @@ static void ggml_vk_check_results_0(ggml_tensor * tensor) { ggml_vk_print_tensor(src2, "src2"); } } + if (src3 != nullptr) { + src3_clone = ggml_dup_tensor(ggml_ctx, src3); - if (tensor->op == GGML_OP_MUL_MAT) { + src3_size = ggml_nbytes(src3); + + src3_buffer = malloc(src3_size); + src3_clone->data = src3_buffer; + if (ggml_backend_buffer_is_host(src3->buffer)) { + memcpy(src3_clone->data, src3->data, src3_size); + memcpy(src3_clone->nb, src3->nb, sizeof(size_t) * GGML_MAX_DIMS); + } else if (ggml_backend_buffer_is_vk(src3->buffer)) { + ggml_backend_vk_buffer_context * buf_ctx = (ggml_backend_vk_buffer_context *)src3->buffer->context; + vk_buffer& buffer_gpu = buf_ctx->dev_buffer; + uint64_t offset = vk_tensor_offset(src3) + src3->view_offs; + if (!ggml_is_contiguous(src3) && ggml_vk_dim01_contiguous(src3)) { + for (int i3 = 0; i3 < src3->ne[3]; i3++) { + for (int i2 = 0; i2 < src3->ne[2]; i2++) { + const int idx = i3*src3->ne[2] + i2; + ggml_vk_buffer_read(buffer_gpu, offset + idx * src3->nb[2], ((char *)src3_clone->data + idx * src3_clone->nb[2]), src3->ne[1] * src3->nb[1]); + } + } + + src3_clone->nb[0] = src3->nb[0]; + src3_clone->nb[1] = src3->nb[1]; + for (int i = 2; i < GGML_MAX_DIMS; i++) { + src3_clone->nb[i] = src3_clone->nb[i - 1]*src3_clone->ne[i - 1]; + } + } else { + if (offset + src3_size >= buffer_gpu->size) { + src3_size = buffer_gpu->size - offset; + } + ggml_vk_buffer_read(buffer_gpu, offset, src3_clone->data, src3_size); + memcpy(src3_clone->nb, src3->nb, sizeof(size_t) * GGML_MAX_DIMS); + } + } else { + GGML_ABORT("fatal error"); + } + + if (vk_output_tensor > 0 && vk_output_tensor == check_counter) { + ggml_vk_print_tensor(src3, "src3"); + } + } + + if (tensor->op == GGML_OP_FLASH_ATTN_EXT) { + const float *params = (const float *)tensor->op_params; + tensor_clone = ggml_flash_attn_ext(ggml_ctx, src0_clone, src1_clone, src2_clone, src3_clone, params[0], params[1], params[2]); + } else if (tensor->op == GGML_OP_MUL_MAT) { tensor_clone = ggml_mul_mat(ggml_ctx, src0_clone, src1_clone); } else if (tensor->op == GGML_OP_MUL_MAT_ID) { tensor_clone = ggml_mul_mat_id(ggml_ctx, src0_clone, src1_clone, src2_clone); @@ -7384,7 +8540,7 @@ static void ggml_vk_check_results_0(ggml_tensor * tensor) { const int32_t max_period = tensor->op_params[1]; tensor_clone = ggml_timestep_embedding(ggml_ctx, src0_clone, dim, max_period); } else if (tensor->op == GGML_OP_POOL_2D) { - enum ggml_op_pool op = static_cast(dst->op_params[0]); + enum ggml_op_pool op = static_cast(tensor->op_params[0]); const int32_t k0 = tensor->op_params[1]; const int32_t k1 = tensor->op_params[2]; const int32_t s0 = tensor->op_params[3]; @@ -7396,7 +8552,11 @@ static void ggml_vk_check_results_0(ggml_tensor * tensor) { } else if (tensor->op == GGML_OP_LEAKY_RELU) { const float * op_params = (const float *)tensor->op_params; tensor_clone = ggml_leaky_relu(ggml_ctx, src0_clone, op_params[0], false); - } else { + } else if (tensor->op == GGML_OP_RWKV_WKV6) { + tensor_clone = ggml_rwkv_wkv6(ggml_ctx, tensor->src[0], tensor->src[1], tensor->src[2], tensor->src[3], + tensor->src[4], tensor->src[5]); + } + else { std::cerr << "Missing vk_check_results OP: " << ggml_op_name(tensor->op) << std::endl; GGML_ABORT("fatal error"); } @@ -7593,3 +8753,5 @@ static void ggml_vk_check_results_1(ggml_tensor * tensor) { VK_LOG_DEBUG("END ggml_vk_check_results_1(" << tensor->name << ")"); } #endif + +GGML_BACKEND_DL_IMPL(ggml_backend_vk_reg) diff --git a/ggml/src/vulkan-shaders/CMakeLists.txt b/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt similarity index 56% rename from ggml/src/vulkan-shaders/CMakeLists.txt rename to ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt index 10075db33..074031087 100644 --- a/ggml/src/vulkan-shaders/CMakeLists.txt +++ b/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt @@ -1,7 +1,11 @@ find_package (Threads REQUIRED) +find_program(GLSLC_EXECUTABLE glslc) +if(NOT GLSLC_EXECUTABLE) + message(FATAL_ERROR "glslc not found.") +endif() set(TARGET vulkan-shaders-gen) add_executable(${TARGET} vulkan-shaders-gen.cpp) install(TARGETS ${TARGET} RUNTIME) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) target_link_libraries(vulkan-shaders-gen PUBLIC Threads::Threads) diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp b/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp new file mode 100644 index 000000000..d896f1ef0 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp @@ -0,0 +1,29 @@ +#version 450 + +#include "types.comp" +#include "generic_binary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = gl_GlobalInvocationID.x; + if (idx >= p.ne) { + return; + } + + const uint offset = p.param3; + const uint src1_i = idx - offset; + const uint oz = src1_i / p.nb02; + const uint oy = (src1_i - (oz * p.nb02)) / p.nb01; + const uint ox = src1_i % p.nb01; + + uint i00, i01, i02, i03; + get_indices(idx, i00, i01, i02, i03); + + if (ox < p.ne10 && oy < p.ne11 && oz < p.ne12) { + data_d[get_doffset() + dst_idx(i00, i01, i02, i03)] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + src0_idx(i00, i01, i02, i03)]) + FLOAT_TYPE(data_b[get_boffset() + ox + oy * p.ne10 + oz * p.ne10 * p.ne11])); + } else { + data_d[get_doffset() + dst_idx(i00, i01, i02, i03)] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + src0_idx(i00, i01, i02, i03)])); + } +} + diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/add.comp b/ggml/src/ggml-vulkan/vulkan-shaders/add.comp new file mode 100644 index 000000000..2b4085c4f --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/add.comp @@ -0,0 +1,29 @@ +#version 450 + +#extension GL_EXT_shader_16bit_storage : require + +#include "types.comp" +#include "generic_binary_head.comp" + +const uint num_threads = 256; + +layout(local_size_x = num_threads, local_size_y = 1, local_size_z = 1) in; + +void main() { + uint idx = get_idx(); + + // num_threads * num_iter must equal 512, to match the wg_denoms and get_idx calculation + const uint num_iter = 2; + + [[unroll]] for (uint i = 0; i < num_iter; ++i) { + if (idx >= p.ne) { + continue; + } + uint i00, i01, i02, i03; + get_indices(idx, i00, i01, i02, i03); + + data_d[get_doffset() + dst_idx(i00, i01, i02, i03)] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + src0_idx(i00, i01, i02, i03)]) + FLOAT_TYPE(data_b[get_boffset() + src1_idx(i00, i01, i02, i03)])); + + idx += num_threads; + } +} diff --git a/ggml/src/vulkan-shaders/argsort.comp b/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp similarity index 100% rename from ggml/src/vulkan-shaders/argsort.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp b/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp new file mode 100644 index 000000000..1e5cb8dae --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp @@ -0,0 +1,17 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = get_idx(); + + if (idx >= p.ne) { + return; + } + + const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]); + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(val < p.param1 ? p.param1 : (val > p.param2 ? p.param2 : val)); +} diff --git a/ggml/src/vulkan-shaders/concat.comp b/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp similarity index 76% rename from ggml/src/vulkan-shaders/concat.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/concat.comp index c23b6eb1b..9ee2f1fae 100644 --- a/ggml/src/vulkan-shaders/concat.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp @@ -3,6 +3,8 @@ #include "types.comp" #include "generic_binary_head.comp" +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + void main() { const uint idx = gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x; const int dim = p.param3; @@ -28,12 +30,12 @@ void main() { const bool is_src0 = i0 < p.ne00 && i1 < p.ne01 && i2 < p.ne02 && i3 < p.ne03; #ifndef OPTIMIZATION_ERROR_WORKAROUND - data_d[p.d_offset + dst_idx] = D_TYPE(is_src0 ? data_a[src0_idx] : data_b[src1_idx]); + data_d[get_doffset() + dst_idx] = D_TYPE(is_src0 ? data_a[get_aoffset() + src0_idx] : data_b[get_boffset() + src1_idx]); #else if (is_src0) { - data_d[p.d_offset + dst_idx] = data_a[src0_idx]; + data_d[get_doffset() + dst_idx] = data_a[get_aoffset() + src0_idx]; } else { - data_d[p.d_offset + dst_idx] = data_b[src1_idx]; + data_d[get_doffset() + dst_idx] = data_b[get_boffset() + src1_idx]; } #endif } diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp b/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp new file mode 100644 index 000000000..dd828c232 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp @@ -0,0 +1,42 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +#extension GL_EXT_control_flow_attributes : require + +const uint num_threads = 128; + +layout(local_size_x = num_threads, local_size_y = 1, local_size_z = 1) in; + +void main() { + uint idx = get_idx(); + + // num_threads * num_iter must equal 512, to match the wg_denoms and get_idx calculation + const uint num_iter = 4; + + // fast path for when all four iterations are in-bounds + if (idx + (num_iter-1)*num_threads < p.ne) { + [[unroll]] for (uint i = 0; i < num_iter; ++i) { +#ifndef OPTIMIZATION_ERROR_WORKAROUND + data_d[get_doffset() + idx] = D_TYPE(data_a[get_aoffset() + idx]); +#else + data_d[get_doffset() + idx] = data_a[get_aoffset() + idx]; +#endif + idx += num_threads; + } + } else { + [[unroll]] for (uint i = 0; i < num_iter; ++i) { + if (idx >= p.ne) { + continue; + } + +#ifndef OPTIMIZATION_ERROR_WORKAROUND + data_d[get_doffset() + idx] = D_TYPE(data_a[get_aoffset() + idx]); +#else + data_d[get_doffset() + idx] = data_a[get_aoffset() + idx]; +#endif + idx += num_threads; + } + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp b/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp new file mode 100644 index 000000000..29c906494 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp @@ -0,0 +1,20 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = get_idx(); + + if (idx >= p.ne) { + return; + } + +#ifndef OPTIMIZATION_ERROR_WORKAROUND + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(data_a[get_aoffset() + src0_idx(idx)]); +#else + data_d[get_doffset() + dst_idx(idx)] = data_a[get_aoffset() + src0_idx(idx)]; +#endif +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp b/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp new file mode 100644 index 000000000..0b8d02f58 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp @@ -0,0 +1,17 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = get_idx(); + + if (idx >= p.ne) { + return; + } + + const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]); + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(cos(val)); +} diff --git a/ggml/src/vulkan-shaders/dequant_f32.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_f32.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp new file mode 100644 index 000000000..91bb8f8db --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp @@ -0,0 +1,118 @@ +#if !defined(DATA_A_F32) && !defined(DATA_A_F16) +#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require +#endif + +#include "types.comp" + +#if defined(A_TYPE_PACKED16) +layout (binding = 0) readonly buffer A_PACKED16 {A_TYPE_PACKED16 data_a_packed16[];}; +#endif +#if defined(A_TYPE_PACKED32) +layout (binding = 0) readonly buffer A_PACKED32 {A_TYPE_PACKED32 data_a_packed32[];}; +#endif + +#if defined(DATA_A_F32) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + return vec2(data_a[a_offset + ib], data_a[a_offset + ib + 1]); +} +#endif + +#if defined(DATA_A_F16) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + return vec2(data_a[a_offset + ib], data_a[a_offset + ib + 1]); +} +#endif + +#if defined(DATA_A_Q4_0) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a[a_offset + ib].qs[iqs]); + return (vec2(vui & 0xF, vui >> 4) - 8.0f); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a_packed16[a_offset + ib].qs[iqs/2]); + return (vec4(vui & 0xF, (vui >> 4) & 0xF, (vui >> 8) & 0xF, vui >> 12) - 8.0f); +} +#endif + +#if defined(DATA_A_Q4_1) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a[a_offset + ib].qs[iqs]); + return vec2(vui & 0xF, vui >> 4); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a_packed16[a_offset + ib].qs[iqs/2]); + return vec4(vui & 0xF, (vui >> 4) & 0xF, (vui >> 8) & 0xF, vui >> 12); +} +#endif + +#if defined(DATA_A_Q5_0) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + const uint uint_qh = uint(data_a[a_offset + ib].qh[1]) << 16 | data_a[a_offset + ib].qh[0]; + const ivec2 qh = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); + const uint vui = uint(data_a[a_offset + ib].qs[iqs]); + return (vec2((vui & 0xF) | qh.x, (vui >> 4) | qh.y) - 16.0f); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + const uint uint_qh = uint(data_a_packed16[a_offset + ib].qh[1]) << 16 | data_a_packed16[a_offset + ib].qh[0]; + const ivec2 qh0 = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); + const ivec2 qh1 = ivec2(((uint_qh >> (iqs + 1)) << 4) & 0x10, (uint_qh >> (iqs + 13)) & 0x10); + const uint vui = uint(data_a_packed16[a_offset + ib].qs[iqs/2]); + return (vec4((vui & 0xF) | qh0.x, ((vui >> 4) & 0xF) | qh0.y, ((vui >> 8) & 0xF) | qh1.x, (vui >> 12) | qh1.y) - 16.0f); +} +#endif + +#if defined(DATA_A_Q5_1) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + const uint uint_qh = data_a[a_offset + ib].qh; + const ivec2 qh = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); + const uint vui = uint(data_a[a_offset + ib].qs[iqs]); + return vec2((vui & 0xF) | qh.x, (vui >> 4) | qh.y); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + const uint uint_qh = data_a_packed16[a_offset + ib].qh; + const ivec2 qh0 = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); + const ivec2 qh1 = ivec2(((uint_qh >> (iqs + 1)) << 4) & 0x10, (uint_qh >> (iqs + 13)) & 0x10); + const uint vui = uint(data_a_packed16[a_offset + ib].qs[iqs/2]); + return vec4((vui & 0xF) | qh0.x, ((vui >> 4) & 0xF) | qh0.y, ((vui >> 8) & 0xF) | qh1.x, (vui >> 12) | qh1.y); +} +#endif + +#if defined(DATA_A_Q8_0) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + return vec2(int(data_a[a_offset + ib].qs[iqs]), int(data_a[a_offset + ib].qs[iqs + 1])); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + uint32_t v0 = data_a_packed16[a_offset + ib].qs[iqs/2]; + uint32_t v1 = data_a_packed16[a_offset + ib].qs[iqs/2 + 1]; + return vec4(int8_t(v0 & 0xFF), int8_t(v0 >> 8), int8_t(v1 & 0xFF), int8_t(v1 >> 8)); +} +#endif + +#if defined(DATA_A_IQ4_NL) +vec2 dequantize(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a[a_offset + ib].qs[iqs]); + return vec2(kvalues_iq4nl[vui & 0xF], kvalues_iq4nl[vui >> 4]); +} +vec4 dequantize4(uint ib, uint iqs, uint a_offset) { + const uint vui = uint(data_a_packed16[a_offset + ib].qs[iqs/2]); + return vec4(kvalues_iq4nl[vui & 0xF], kvalues_iq4nl[(vui >> 4) & 0xF], kvalues_iq4nl[(vui >> 8) & 0xF], kvalues_iq4nl[vui >> 12]); +} +#endif + +#if defined(DATA_A_F32) || defined(DATA_A_F16) +vec2 get_dm(uint ib, uint a_offset) { + return vec2(0, 0); +} +#endif + +#if defined(DATA_A_Q4_0) || defined(DATA_A_Q5_0) || defined(DATA_A_Q8_0) || defined(DATA_A_IQ4_NL) +vec2 get_dm(uint ib, uint a_offset) { + return vec2(float(data_a[a_offset + ib].d), 0); +} +#endif + +#if defined(DATA_A_Q4_1) || defined(DATA_A_Q5_1) +vec2 get_dm(uint ib, uint a_offset) { + return vec2(float(data_a[a_offset + ib].d), float(data_a[a_offset + ib].m)); +} +#endif diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp new file mode 100644 index 000000000..94b78598e --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp @@ -0,0 +1,325 @@ + +#include "types.comp" + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ4_0 { + block_q4_0_packed16 block; +}; + +float16_t dequantFuncQ4_0(const in decodeBufQ4_0 bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const uint idx = coordInBlock[1]; + const uint shift = (idx & 0x10) >> 2; + uint32_t qs = uint32_t(bl.block.qs[(idx & 0xE) >> 1]); + qs >>= shift; + qs &= 0x0F0F; + qs = unpack8(qs)[idx & 1]; + float16_t ret = (float16_t(qs) - float16_t(8)) * d; + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 4) buffer decodeBufQ4_1 { + block_q4_1 block; +}; + +float16_t dequantFuncQ4_1(const in decodeBufQ4_1 bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const float16_t m = bl.block.m; + const uint idx = coordInBlock[1]; + const uint iqs = idx & 0xF; + const uint shift = (idx & 0x10) >> 2; + uint32_t qs = bl.block.qs[iqs]; + qs >>= shift; + qs &= 0xF; + float16_t ret = float16_t(qs) * d + m; + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ5_0 { + block_q5_0 block; +}; + +float16_t dequantFuncQ5_0(const in decodeBufQ5_0 bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const uint idx = coordInBlock[1]; + const uint iqs = idx & 0xF; + + const uint uint_qh = uint(bl.block.qh[1]) << 16 | bl.block.qh[0]; + const uint qh = ((uint_qh >> idx) << 4) & 0x10; + + const uint shift = (idx & 0x10) >> 2; + uint32_t qs = bl.block.qs[iqs]; + qs >>= shift; + qs &= 0xF; + + float16_t ret = (float16_t(qs | qh) - float16_t(16)) * d; + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 8) buffer decodeBufQ5_1 { + block_q5_1 block; +}; + +float16_t dequantFuncQ5_1(const in decodeBufQ5_1 bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const float16_t m = bl.block.m; + const uint idx = coordInBlock[1]; + const uint iqs = idx & 0xF; + + const uint uint_qh = bl.block.qh; + const uint qh = ((uint_qh >> idx) << 4) & 0x10; + + const uint shift = (idx & 0x10) >> 2; + uint32_t qs = bl.block.qs[iqs]; + qs >>= shift; + qs &= 0xF; + + float16_t ret = float16_t(qs | qh) * d + m; + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ8_0 { + block_q8_0_packed16 block; +}; + +float16_t dequantFuncQ8_0(const in decodeBufQ8_0 bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const uint idx = coordInBlock[1]; + const uint iqs = idx; + + // Load 16b and select the byte for this element + int32_t qs = unpack8(int32_t(bl.block.qs[(iqs & 0x1E) >> 1]))[iqs & 1]; + float16_t ret = float16_t(qs) * d; + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 4) buffer decodeBufQ2_K { + block_q2_K block; +}; + +float16_t dequantFuncQ2_K(const in decodeBufQ2_K bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const f16vec2 d = bl.block.d; + const uint idx = coordInBlock[1]; + const uint iqs = idx; + + const uint qsi = (iqs / 128) * 32 + (iqs % 32); // 0..31 + const uint scalesi = iqs / 16; // 0..15 + const uint qsshift = ((iqs % 128) / 32) * 2; // 0,2,4,6 + + uint32_t qs = bl.block.qs[qsi]; + const uint scales = bl.block.scales[scalesi]; + float16_t ret = d.x * float16_t(scales & 0xF) * float16_t((qs >> qsshift) & 3) - d.y * float16_t(scales >> 4); + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ3_K { + block_q3_K block; +}; + +float16_t dequantFuncQ3_K(const in decodeBufQ3_K bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const uint idx = coordInBlock[1]; + const uint iqs = idx; + + const uint n = iqs / 128; // 0,1 + const uint qsi = n * 32 + (iqs % 32); // 0..63 + const uint hmi = (iqs % 32); // 0..31 + const uint j = (iqs % 128) / 8; // 0..15 + const uint is = iqs / 16; // 0..15 + const uint halfsplit = ((iqs % 128) / 32); // 0,1,2,3 + const uint qsshift = halfsplit * 2; // 0,2,4,6 + const uint m = 1 << (4 * n + halfsplit); // 1,2,4,8,16,32,64,128 + + uint32_t scaleidx0 = (is < 8) ? is : (is-8); + uint32_t scaleidx0shift = (is < 8) ? 0 : 4; + uint32_t scaleidx1 = is + 8 - (is/4)*4; + uint32_t scaleidx1shift = (is/4)*2; + + const int8_t us = int8_t(((bl.block.scales[scaleidx0] >> scaleidx0shift) & 0xF) | (((bl.block.scales[scaleidx1] >> scaleidx1shift) & 3) << 4)); + + const float16_t dl = bl.block.d * float16_t(us - 32); + + float16_t ret = dl * float16_t(int8_t((bl.block.qs[qsi ] >> qsshift) & 3) - (((bl.block.hmask[hmi ] & m) != 0) ? 0 : 4)); + + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 16) buffer decodeBufQ4_K { + block_q4_K block; +}; + +layout(buffer_reference, std430, buffer_reference_align = 16) buffer decodeBufQ4_K_packed16 { + block_q4_K_packed16 block; +}; + +float16_t dequantFuncQ4_K(const in decodeBufQ4_K bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + decodeBufQ4_K_packed16 bl16 = decodeBufQ4_K_packed16(bl); + const uint idx = coordInBlock[1]; + + const uint b = (idx & 0x20) >> 5; // 0,1 + const uint is = (idx & 0xE0) >> 5; // 0..7 + + const f16vec2 loadd = bl.block.d; + + uint32_t sc; + uint32_t mbyte; + + uint32_t scidx0 = (is < 4) ? is : (is + 4); + uint32_t scidx1 = (is < 4) ? is : (is - 4); + uint32_t scidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint32_t scidxshift1 = (is < 4) ? 0 : 2; + uint32_t mbidx0 = is + 4; + uint32_t mbidx1 = (is < 4) ? is + 4 : is; + uint32_t mbidxmask0 = (is < 4) ? 0xF : 0xF0; + uint32_t mbidxshift0 = (is < 4) ? 0 : 4; + uint32_t mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint32_t mbidxshift1 = (is < 4) ? 0 : 2; + + sc = uint8_t((bl.block.scales[scidx0] & 0xF) | ((bl.block.scales[scidx1] & scidxmask1) >> scidxshift1)); + mbyte = uint8_t(((bl.block.scales[mbidx0] & mbidxmask0) >> mbidxshift0) | ((bl.block.scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const float16_t d = loadd.x * float16_t(sc); + const float16_t m = loadd.y * float16_t(mbyte); + + uint qs = uint32_t(bl16.block.qs[((idx & 0xC0) >> 2) + ((idx & 0x1E) >> 1)]); + qs = (qs >> (b * 4)) & 0x0F0F; + qs = unpack8(qs)[idx & 1]; + + float16_t ret = d * float16_t(qs) - m; + + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 16) buffer decodeBufQ5_K { + block_q5_K block; +}; + +layout(buffer_reference, std430, buffer_reference_align = 16) buffer decodeBufQ5_K_packed16 { + block_q5_K_packed16 block; +}; + +float16_t dequantFuncQ5_K(const in decodeBufQ5_K bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + decodeBufQ5_K_packed16 bl16 = decodeBufQ5_K_packed16(bl); + const uint idx = coordInBlock[1]; + + const uint b = (idx & 0x20) >> 5; // 0,1 + const uint is = (idx & 0xE0) >> 5; // 0..7 + + const uint32_t hm = 0x0101 << is; + + const f16vec2 loadd = bl.block.d; + + uint32_t sc; + uint32_t mbyte; + + uint32_t scidx0 = (is < 4) ? is : (is + 4); + uint32_t scidx1 = (is < 4) ? is : (is - 4); + uint32_t scidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint32_t scidxshift1 = (is < 4) ? 0 : 2; + uint32_t mbidx0 = is + 4; + uint32_t mbidx1 = (is < 4) ? is + 4 : is; + uint32_t mbidxmask0 = (is < 4) ? 0xF : 0xF0; + uint32_t mbidxshift0 = (is < 4) ? 0 : 4; + uint32_t mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint32_t mbidxshift1 = (is < 4) ? 0 : 2; + + sc = uint8_t((bl.block.scales[scidx0] & 0xF) | ((bl.block.scales[scidx1] & scidxmask1) >> scidxshift1)); + mbyte = uint8_t(((bl.block.scales[mbidx0] & mbidxmask0) >> mbidxshift0) | ((bl.block.scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const float16_t d = loadd.x * float16_t(sc); + const float16_t m = loadd.y * float16_t(mbyte); + + uint qh = uint32_t(bl16.block.qh[(idx & 0x1E) >> 1]); + qh = qh & hm; + qh = unpack8(qh)[idx & 1]; + + uint qs = uint32_t(bl16.block.qs[((idx & 0xC0) >> 2) + ((idx & 0x1E) >> 1)]); + qs = (qs >> (b * 4)) & 0x0F0F; + qs = unpack8(qs)[idx & 1]; + + float16_t ret = d * (float16_t(qs) + (qh != 0 ? float16_t(16) : float16_t(0))) - m; + + return ret; +} + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ6_K { + block_q6_K block; +}; + +layout(buffer_reference, std430, buffer_reference_align = 16) buffer decodeBufQ6_K_packed16 { + block_q6_K_packed16 block; +}; + +float16_t dequantFuncQ6_K(const in decodeBufQ6_K bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + decodeBufQ6_K_packed16 bl16 = decodeBufQ6_K_packed16(bl); + const uint idx = coordInBlock[1]; + + const uint b = (idx & 0x40) >> 6; // 0,1 + const uint qhshift = (idx & 0x60) >> 4; // 0,2,4,6 + const uint is = (idx & 0xF0) >> 4; // 0..15 + + const float16_t dscale = bl.block.d * float16_t(bl.block.scales[is]); + + uint ql = uint32_t(bl16.block.ql[((idx & 0x80) >> 2) + ((idx & 0x3E) >> 1)]); + ql = (ql >> (b * 4)) & 0x0F0F; + + uint qh = uint32_t(bl16.block.qh[((idx & 0x80) >> 3) + ((idx & 0x1E) >> 1)]); + qh = ((qh >> qhshift) & 0x0303) << 4; + + int q = unpack8(ql | qh)[idx & 1]; + + float16_t ret = dscale * float16_t(q - 32); + + return ret; +} + +#if defined(DATA_A_IQ4_NL) +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufIQ4_NL { + block_iq4_nl block; +}; + +float16_t dequantFuncIQ4_NL(const in decodeBufIQ4_NL bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const float16_t d = bl.block.d; + const uint idx = coordInBlock[1]; + const uint iqs = idx & 0xF; + const uint shift = (idx & 0x10) >> 2; + uint32_t qs = bl.block.qs[iqs]; + qs >>= shift; + qs &= 0xF; + float16_t ret = float16_t(kvalues_iq4nl[qs]) * d; + return ret; +} +#endif + +#if defined(DATA_A_Q4_0) +#define dequantFuncA dequantFuncQ4_0 +#elif defined(DATA_A_Q4_1) +#define dequantFuncA dequantFuncQ4_1 +#elif defined(DATA_A_Q5_0) +#define dequantFuncA dequantFuncQ5_0 +#elif defined(DATA_A_Q5_1) +#define dequantFuncA dequantFuncQ5_1 +#elif defined(DATA_A_Q8_0) +#define dequantFuncA dequantFuncQ8_0 +#elif defined(DATA_A_Q2_K) +#define dequantFuncA dequantFuncQ2_K +#elif defined(DATA_A_Q3_K) +#define dequantFuncA dequantFuncQ3_K +#elif defined(DATA_A_Q4_K) +#define dequantFuncA dequantFuncQ4_K +#elif defined(DATA_A_Q5_K) +#define dequantFuncA dequantFuncQ5_K +#elif defined(DATA_A_Q6_K) +#define dequantFuncA dequantFuncQ6_K +#elif defined(DATA_A_IQ4_NL) +#define dequantFuncA dequantFuncIQ4_NL +#endif diff --git a/ggml/src/vulkan-shaders/dequant_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_head.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp diff --git a/ggml/src/vulkan-shaders/dequant_iq4_nl.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp similarity index 97% rename from ggml/src/vulkan-shaders/dequant_iq4_nl.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp index 34ef3da30..8de14fc03 100644 --- a/ggml/src/vulkan-shaders/dequant_iq4_nl.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp @@ -10,6 +10,8 @@ layout (binding = 1) writeonly buffer D {D_TYPE data_b[];}; void main() { const uint i = gl_WorkGroupID.x * 4 + gl_LocalInvocationID.x / 64; + init_iq4nl_shmem(); + const uint tid = gl_LocalInvocationID.x % 64; const uint il = tid/32; const uint ir = tid%32; diff --git a/ggml/src/vulkan-shaders/dequant_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q2_k.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp diff --git a/ggml/src/vulkan-shaders/dequant_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q3_k.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp diff --git a/ggml/src/vulkan-shaders/dequant_q4_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q4_0.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp diff --git a/ggml/src/vulkan-shaders/dequant_q4_1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q4_1.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp new file mode 100644 index 000000000..987f113a3 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp @@ -0,0 +1,68 @@ +#version 450 + +#include "dequant_head.comp" + +layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; + +layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; +layout (binding = 1) writeonly buffer D {D_TYPE data_b[];}; + +void main() { + [[unroll]] for (uint wgy = 0; wgy < 256; wgy++) { + const uint ib = gl_WorkGroupID.x * 256 + wgy; + if (ib >= p.M * p.K / QUANT_K) { + return; + } + + const uint tid = gl_LocalInvocationID.x; + const uint il = tid / 8; + const uint ir = tid % 8; + const uint is = 2 * il; + const uint n = 4; + + const FLOAT_TYPE dall = FLOAT_TYPE(data_a[ib].d.x); + const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[ib].d.y); + + const uint y_idx = ib * QUANT_K + 64 * il + n * ir; + const uint qs_idx = 32*il + n * ir; + + uint scidx0 = (is < 4) ? is : (is + 4); + uint scidx1 = (is < 4) ? is : (is - 4); + uint scidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint scidxshift1 = (is < 4) ? 0 : 2; + uint mbidx0 = is + 4; + uint mbidx1 = (is < 4) ? is + 4 : is; + uint mbidxmask0 = (is < 4) ? 0xF : 0xF0; + uint mbidxshift0 = (is < 4) ? 0 : 4; + uint mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint mbidxshift1 = (is < 4) ? 0 : 2; + + uint8_t sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + uint8_t mbyte = uint8_t((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0 | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const FLOAT_TYPE d1 = dall * sc; + const FLOAT_TYPE m1 = dmin * mbyte; + + scidx0 = (is < 4) ? is + 1 : (is + 5); + scidx1 = (is < 4) ? is + 1 : (is - 3); + scidxmask1 = (is < 4) ? 0x30 : 0xC0; + scidxshift1 = (is < 4) ? 0 : 2; + mbidx0 = is + 5; + mbidx1 = (is < 4) ? is + 5 : is + 1; + mbidxmask0 = (is < 4) ? 0xF : 0xF0; + mbidxshift0 = (is < 4) ? 0 : 4; + mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + mbidxshift1 = (is < 4) ? 0 : 2; + + sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + mbyte = uint8_t((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0 | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const FLOAT_TYPE d2 = dall * sc; + const FLOAT_TYPE m2 = dmin * mbyte; + + [[unroll]] for (uint l = 0; l < n; ++l) { + data_b[y_idx + l ] = D_TYPE(d1 * FLOAT_TYPE(data_a[ib].qs[qs_idx + l] & 0xF) - m1); + data_b[y_idx + l + 32] = D_TYPE(d2 * FLOAT_TYPE(data_a[ib].qs[qs_idx + l] >> 4) - m2); + } + } +} diff --git a/ggml/src/vulkan-shaders/dequant_q5_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q5_0.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp diff --git a/ggml/src/vulkan-shaders/dequant_q5_1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q5_1.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp new file mode 100644 index 000000000..6db5403b6 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp @@ -0,0 +1,70 @@ +#version 450 + +#include "dequant_head.comp" + +layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in; + +layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; +layout (binding = 1) writeonly buffer D {D_TYPE data_b[];}; + +void main() { + [[unroll]] for (uint wgy = 0; wgy < 256; wgy++) { + const uint ib = gl_WorkGroupID.x * 256 + wgy; + if (ib >= p.M * p.K / QUANT_K) { + return; + } + + const uint tid = gl_LocalInvocationID.x; + const uint il = tid / 16; + const uint ir = tid % 16; + const uint is = 2 * il; + + const FLOAT_TYPE dall = FLOAT_TYPE(data_a[ib].d.x); + const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[ib].d.y); + + const uint y_idx = ib * QUANT_K + 64 * il + 2 * ir; + const uint qs_idx = 32*il + 2 * ir; + const uint qh_idx = 2 * ir; + + uint scidx0 = (is < 4) ? is : (is + 4); + uint scidx1 = (is < 4) ? is : (is - 4); + uint scidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint scidxshift1 = (is < 4) ? 0 : 2; + uint mbidx0 = is + 4; + uint mbidx1 = (is < 4) ? is + 4 : is; + uint mbidxmask0 = (is < 4) ? 0xF : 0xF0; + uint mbidxshift0 = (is < 4) ? 0 : 4; + uint mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + uint mbidxshift1 = (is < 4) ? 0 : 2; + + uint8_t sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + uint8_t mbyte = uint8_t((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0 | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const FLOAT_TYPE d1 = dall * sc; + const FLOAT_TYPE m1 = dmin * mbyte; + + scidx0 = (is < 4) ? is + 1 : (is + 5); + scidx1 = (is < 4) ? is + 1 : (is - 3); + scidxmask1 = (is < 4) ? 0x30 : 0xC0; + scidxshift1 = (is < 4) ? 0 : 2; + mbidx0 = is + 5; + mbidx1 = (is < 4) ? is + 5 : is + 1; + mbidxmask0 = (is < 4) ? 0xF : 0xF0; + mbidxshift0 = (is < 4) ? 0 : 4; + mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + mbidxshift1 = (is < 4) ? 0 : 2; + + sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + mbyte = uint8_t((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0 | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + + const FLOAT_TYPE d2 = dall * sc; + const FLOAT_TYPE m2 = dmin * mbyte; + + const uint8_t hm1 = uint8_t(1 << (2 * il )); + const uint8_t hm2 = uint8_t(1 << (2 * il + 1)); + data_b[y_idx ] = D_TYPE(d1 * FLOAT_TYPE((data_a[ib].qs[qs_idx ] & 0xF) + (((data_a[ib].qh[qh_idx ] & hm1) != 0) ? 16 : 0)) - m1); + data_b[y_idx + 1] = D_TYPE(d1 * FLOAT_TYPE((data_a[ib].qs[qs_idx + 1] & 0xF) + (((data_a[ib].qh[qh_idx + 1] & hm1) != 0) ? 16 : 0)) - m1); + data_b[y_idx + 32] = D_TYPE(d2 * FLOAT_TYPE((data_a[ib].qs[qs_idx ] >> 4) + (((data_a[ib].qh[qh_idx ] & hm2) != 0) ? 16 : 0)) - m2); + data_b[y_idx + 33] = D_TYPE(d2 * FLOAT_TYPE((data_a[ib].qs[qs_idx + 1] >> 4) + (((data_a[ib].qh[qh_idx + 1] & hm2) != 0) ? 16 : 0)) - m2); + } +} diff --git a/ggml/src/vulkan-shaders/dequant_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q6_k.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp diff --git a/ggml/src/vulkan-shaders/dequant_q8_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp similarity index 100% rename from ggml/src/vulkan-shaders/dequant_q8_0.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp diff --git a/ggml/src/vulkan-shaders/diag_mask_inf.comp b/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp similarity index 100% rename from ggml/src/vulkan-shaders/diag_mask_inf.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/div.comp b/ggml/src/ggml-vulkan/vulkan-shaders/div.comp new file mode 100644 index 000000000..9fb69c6c1 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/div.comp @@ -0,0 +1,27 @@ +#version 450 + +#include "types.comp" +#include "generic_binary_head.comp" + +const uint num_threads = 256; + +layout(local_size_x = num_threads, local_size_y = 1, local_size_z = 1) in; + +void main() { + uint idx = get_idx(); + + // num_threads * num_iter must equal 512, to match the wg_denoms and get_idx calculation + const uint num_iter = 2; + + [[unroll]] for (uint i = 0; i < num_iter; ++i) { + if (idx >= p.ne) { + continue; + } + uint i00, i01, i02, i03; + get_indices(idx, i00, i01, i02, i03); + + data_d[get_doffset() + dst_idx(i00, i01, i02, i03)] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + src0_idx(i00, i01, i02, i03)]) / FLOAT_TYPE(data_b[get_boffset() + src1_idx(i00, i01, i02, i03)])); + + idx += num_threads; + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp new file mode 100644 index 000000000..c5be8131b --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp @@ -0,0 +1,289 @@ +#version 450 + +#extension GL_EXT_control_flow_attributes : enable +#extension GL_EXT_shader_16bit_storage : require + +#extension GL_EXT_shader_explicit_arithmetic_types_float16 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require + +#extension GL_KHR_memory_scope_semantics : enable +#extension GL_KHR_cooperative_matrix : enable +#extension GL_NV_cooperative_matrix2 : enable +#extension GL_EXT_buffer_reference : enable +#extension GL_KHR_shader_subgroup_ballot : enable +#extension GL_KHR_shader_subgroup_vote : enable +#extension GL_EXT_null_initializer : enable + +#include "types.comp" +#include "dequant_funcs_cm2.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +layout (constant_id = 1) const uint32_t Br = 32; +layout (constant_id = 2) const uint32_t Bc = 32; +layout (constant_id = 3) const uint32_t D = 32; +layout (constant_id = 4) const uint32_t Clamp = gl_CooperativeMatrixClampModeConstantNV; + +layout (push_constant) uniform parameter { + uint32_t N; + uint32_t KV; + + uint32_t ne1; + uint32_t ne2; + uint32_t ne3; + + uint32_t neq2; + uint32_t neq3; + uint32_t nek2; + uint32_t nek3; + uint32_t nev2; + uint32_t nev3; + uint32_t nem1; + + uint32_t nb02; + uint32_t nb03; + uint32_t nb12; + uint32_t nb13; + uint32_t nb22; + uint32_t nb23; + uint32_t nb31; + + float scale; + float max_bias; + float logit_softcap; + + uint32_t mask; + uint32_t n_head_log2; + float m0; + float m1; +} p; + +layout (binding = 0) readonly buffer Q {uint8_t data_q[];}; +layout (binding = 1) readonly buffer K {uint8_t data_k[];}; +layout (binding = 2) readonly buffer V {uint8_t data_v[];}; +layout (binding = 3) readonly buffer M {uint8_t data_m[];}; +layout (binding = 4) writeonly buffer O {D_TYPE data_o[];}; + +#define CEIL_DIV(a, b) (((a) + (b) - 1) / (b)) + +ACC_TYPE maxReduce(const in ACC_TYPE x, const in ACC_TYPE y) { + return max(x, y); +} + +ACC_TYPE smearReduce(const in ACC_TYPE x, const in ACC_TYPE y) { + return x; +} + +// Replace matrix elements >= numRows or numCols with 'replace' +ACC_TYPE replacePadding(const in uint32_t row, const in uint32_t col, const in ACC_TYPE elem, const in ACC_TYPE replace, const in uint32_t numRows, const in uint32_t numCols) { + if (row >= numRows || col >= numCols) { + return replace; + } + return elem; +} + +ACC_TYPE Exp(const in uint32_t row, const in uint32_t col, const in ACC_TYPE elem) +{ + return exp(elem); +} + +ACC_TYPE Max(const in uint32_t row, const in uint32_t col, const in ACC_TYPE elem0, const in ACC_TYPE elem1) +{ + return max(elem0, elem1); +} + +#if defined(BLOCK_SIZE) +#define DECODEFUNC , DEQUANTFUNC +#else +#define DECODEFUNC +#endif + +void main() { +#if defined(DATA_A_IQ4_NL) + init_iq4nl_shmem(); +#endif + + const uint32_t N = p.N; + const uint32_t KV = p.KV; + + const uint32_t Tr = CEIL_DIV(N, Br); + const uint32_t Tc = CEIL_DIV(KV, Bc); + + const uint32_t i = gl_WorkGroupID.x; + + const uint32_t iq2 = gl_WorkGroupID.y; + const uint32_t iq3 = gl_WorkGroupID.z; + + // broadcast factors + const uint32_t rk2 = p.neq2/p.nek2; + const uint32_t rk3 = p.neq3/p.nek3; + + const uint32_t rv2 = p.neq2/p.nev2; + const uint32_t rv3 = p.neq3/p.nev3; + + // k indices + const uint32_t ik3 = iq3 / rk3; + const uint32_t ik2 = iq2 / rk2; + + // v indices + const uint32_t iv3 = iq3 / rv3; + const uint32_t iv2 = iq2 / rv2; + + tensorLayoutNV<2, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutQ = createTensorLayoutNV(2, gl_CooperativeMatrixClampModeConstantNV); + tensorLayoutNV<2, Clamp> tensorLayoutK = createTensorLayoutNV(2, Clamp); + tensorLayoutNV<2, Clamp> tensorLayoutV = createTensorLayoutNV(2, Clamp); + + tensorViewNV<2, false, 1, 0> tensorViewTranspose = createTensorViewNV(2, false, 1, 0); + +#if defined(BLOCK_SIZE) + tensorLayoutK = setTensorLayoutBlockSizeNV(tensorLayoutK, 1, BLOCK_SIZE); + tensorLayoutV = setTensorLayoutBlockSizeNV(tensorLayoutV, 1, BLOCK_SIZE); +#endif + + tensorLayoutQ = setTensorLayoutDimensionNV(tensorLayoutQ, N, D); + tensorLayoutK = setTensorLayoutDimensionNV(tensorLayoutK, KV, D); + tensorLayoutV = setTensorLayoutDimensionNV(tensorLayoutV, KV, D); + + coopmat Q; + coopmat Qf16; + + uint32_t q_offset = iq2*p.nb02+iq3*p.nb03; + coopMatLoadTensorNV(Q, data_q, q_offset, sliceTensorLayoutNV(tensorLayoutQ, i * Br, Br, 0, D)); + + Qf16 = coopmat(Q); + Qf16 *= float16_t(p.scale); + + coopmat O = coopmat(0); + + coopmat L, M; + + L = coopmat(0); + M = coopmat(-1.0/0.0); + + ACC_TYPE slope = ACC_TYPE(1.0); + + // ALiBi + if (p.max_bias > 0.0f) { + const uint32_t h = iq2; + + const ACC_TYPE base = ACC_TYPE(h < p.n_head_log2 ? p.m0 : p.m1); + const int exph = int(h < p.n_head_log2 ? h + 1 : 2*(h - p.n_head_log2) + 1); + + slope = pow(base, ACC_TYPE(exph)); + } + + [[dont_unroll]] + for (uint32_t j = 0; j < Tc; ++j) { + + coopmat S = coopmat(0); + + coopmat K_T; + + uint32_t k_offset = ik2*p.nb12 + ik3*p.nb13; + coopMatLoadTensorNV(K_T, data_k, k_offset, sliceTensorLayoutNV(tensorLayoutK, j * Bc, Bc, 0, D), tensorViewTranspose DECODEFUNC); + S = coopMatMulAdd(Qf16, K_T, S); + + if (p.logit_softcap != 0.0f) { + [[unroll]] + for (int k = 0; k < S.length(); ++k) { + S[k] = ACC_TYPE(p.logit_softcap)*tanh(S[k]); + } + } + + if (p.mask != 0) { + tensorLayoutNV<2, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutM = createTensorLayoutNV(2, gl_CooperativeMatrixClampModeConstantNV); + tensorLayoutM = setTensorLayoutDimensionNV(tensorLayoutM, p.nem1, KV); + + coopmat mv; + + coopMatLoadTensorNV(mv, data_m, 0, sliceTensorLayoutNV(tensorLayoutM, i * Br, Br, j * Bc, Bc)); + + S += slope*coopmat(mv); + } + + // Clear padding elements to -inf, so they don't contribute to rowmax + if (Clamp != 0 && + ((j + 1) * Bc > KV || + (i + 1) * Br > N)) { + + uint R = ((i + 1) * Br > N) ? (N % Br) : Br; + uint C = ((j + 1) * Bc > KV) ? (KV % Bc) : Bc; + + coopMatPerElementNV(S, S, replacePadding, ACC_TYPE(-1.0/0.0), R, C); + } + + coopmat rowmax, P, rowsum, eM; + + coopMatReduceNV(rowmax, S, gl_CooperativeMatrixReduceRowNV, maxReduce); + + coopmat Mold = M; + + // M = max(rowmax, Mold) + // P = e^(S - M) + // eM = e^(Mold - M) + coopMatPerElementNV(M, rowmax, Max, Mold); + coopMatPerElementNV(P, S - M, Exp); + coopMatPerElementNV(eM, Mold - M, Exp); + + // Clear padding elements to 0, so they don't contribute to rowsum + if (Clamp != 0 && + ((j + 1) * Bc > KV || + (i + 1) * Br > N)) { + + uint R = ((i + 1) * Br > N) ? (N % Br) : Br; + uint C = ((j + 1) * Bc > KV) ? (KV % Bc) : Bc; + + coopMatPerElementNV(P, P, replacePadding, ACC_TYPE(0.0), R, C); + } + + coopmat P_A = coopmat(P); + + // compute rowsum by multiplying by matrix of all ones. + coopmat One = coopmat(1.0); + + rowsum = coopmat(0.0); + rowsum = coopMatMulAdd(P_A, One, rowsum); + + coopmat V; + uint32_t v_offset = iv2*p.nb22 + iv3*p.nb23; + coopMatLoadTensorNV(V, data_v, v_offset, sliceTensorLayoutNV(tensorLayoutV, j * Bc, Bc, 0, D) DECODEFUNC); + + L = eM*L + rowsum; + + // This is the "diagonal" matrix in the paper, but since we do componentwise + // multiply rather than matrix multiply it has the diagonal element smeared + // across the row + coopmat eMdiag; + + // resize eM by using smear/reduce + coopMatReduceNV(eMdiag, eM, gl_CooperativeMatrixReduceRowNV, smearReduce); + + O = eMdiag * O; + + O = coopMatMulAdd(P_A, V, O); + } + + coopmat Ldiag; + + // resize L by using smear/reduce + coopMatReduceNV(Ldiag, L, gl_CooperativeMatrixReduceRowNV, smearReduce); + + [[unroll]] + for (int k = 0; k < Ldiag.length(); ++k) { + Ldiag[k] = ACC_TYPE(1.0) / Ldiag[k]; + } + + O = Ldiag*O; + + tensorLayoutNV<3, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutD = createTensorLayoutNV(3, gl_CooperativeMatrixClampModeConstantNV); + tensorLayoutD = setTensorLayoutDimensionNV(tensorLayoutD, p.ne2, p.ne1, D); + + // permute dimensions + tensorViewNV<3, false, 1, 0, 2> tensorViewPermute = createTensorViewNV(3, false, 1, 0, 2); + uint32_t o_offset = iq3*p.ne2*p.ne1; + + coopmat O_D = coopmat(O); + coopMatStoreTensorNV(O_D, data_o, o_offset, sliceTensorLayoutNV(tensorLayoutD, i * Br, Br, iq2, 1, 0, D), tensorViewPermute); +} diff --git a/ggml/src/vulkan-shaders/gelu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp similarity index 100% rename from ggml/src/vulkan-shaders/gelu.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp diff --git a/ggml/src/vulkan-shaders/gelu_quick.comp b/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp similarity index 100% rename from ggml/src/vulkan-shaders/gelu_quick.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp new file mode 100644 index 000000000..062e2a4cd --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp @@ -0,0 +1,64 @@ +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_control_flow_attributes : require + +layout (push_constant) uniform parameter +{ + uint ne; + uint ne00; uint ne01; uint ne02; uint ne03; uint nb00; uint nb01; uint nb02; uint nb03; + uint ne10; uint ne11; uint ne12; uint ne13; uint nb10; uint nb11; uint nb12; uint nb13; + uint ne20; uint ne21; uint ne22; uint ne23; uint nb20; uint nb21; uint nb22; uint nb23; + uint misalign_offsets; + float param1; float param2; int param3; +} p; + +layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; +layout (binding = 1) readonly buffer B {B_TYPE data_b[];}; +layout (binding = 2) writeonly buffer D {D_TYPE data_d[];}; + +// true if src0/src1 are the same shape and the indices can be reused without additional modulus +layout(constant_id = 0) const bool norepeat = false; + +uint get_idx() { + return gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x; +} + +uint get_aoffset() { return p.misalign_offsets >> 16; } +uint get_boffset() { return (p.misalign_offsets >> 8) & 0xFF; } +uint get_doffset() { return p.misalign_offsets & 0xFF; } + +// mod and div are expensive and coordinates/dimensions are often power of 2 or equal to 1 +uint fastmod(uint a, uint b) { + if ((b & (b-1)) == 0) { + return a & (b-1); + } + return a % b; +} + +uint fastdiv(uint a, uint b) { + return (a < b) ? 0 : (a / b); +} + +void get_indices(uint idx, out uint i00, out uint i01, out uint i02, out uint i03) { + i03 = fastdiv(idx, (p.ne02*p.ne01*p.ne00)); + const uint i03_offset = i03 * p.ne02*p.ne01*p.ne00; + i02 = fastdiv((idx - i03_offset), (p.ne01*p.ne00)); + const uint i02_offset = i02*p.ne01*p.ne00; + i01 = (idx - i03_offset - i02_offset) / p.ne00; + i00 = idx - i03_offset - i02_offset - i01*p.ne00; +} + +uint src0_idx(uint i00, uint i01, uint i02, uint i03) { + return i03*p.nb03 + i02*p.nb02 + i01*p.nb01 + i00*p.nb00; +} + +uint src1_idx(uint i00, uint i01, uint i02, uint i03) { + if (norepeat) { + return i03*p.nb13 + i02*p.nb12 + i01*p.nb11 + i00*p.nb10; + } else { + return fastmod(i03, p.ne13)*p.nb13 + fastmod(i02, p.ne12)*p.nb12 + fastmod(i01, p.ne11)*p.nb11 + fastmod(i00, p.ne10)*p.nb10; + } +} + +uint dst_idx(uint i00, uint i01, uint i02, uint i03) { + return i03*p.nb23 + i02*p.nb22 + i01*p.nb21 + i00*p.nb20; +} diff --git a/ggml/src/vulkan-shaders/generic_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp similarity index 100% rename from ggml/src/vulkan-shaders/generic_head.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp diff --git a/ggml/src/vulkan-shaders/generic_unary_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp similarity index 51% rename from ggml/src/vulkan-shaders/generic_unary_head.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp index eacdefc7d..68d1bc9f1 100644 --- a/ggml/src/vulkan-shaders/generic_unary_head.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp @@ -1,15 +1,21 @@ #extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_control_flow_attributes : require layout (push_constant) uniform parameter { uint ne; uint ne00; uint ne01; uint ne02; uint ne03; uint nb00; uint nb01; uint nb02; uint nb03; uint ne10; uint ne11; uint ne12; uint ne13; uint nb10; uint nb11; uint nb12; uint nb13; - uint d_offset; + uint misalign_offsets; float param1; float param2; -} p; -layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + uint ne0_012mp; uint ne0_012L; + uint ne0_01mp; uint ne0_01L; + uint ne0_0mp; uint ne0_0L; + uint ne1_012mp; uint ne1_012L; + uint ne1_01mp; uint ne1_01L; + uint ne1_0mp; uint ne1_0L; +} p; layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; layout (binding = 1) writeonly buffer D {D_TYPE data_d[];}; @@ -18,22 +24,33 @@ uint get_idx() { return gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x; } +uint get_aoffset() { return p.misalign_offsets >> 16; } +uint get_doffset() { return p.misalign_offsets & 0xFFFF; } + +// see init_fastdiv_values in ggml-vulkan.cpp +uint fastdiv(uint n, uint mp, uint L) { + uint msbs, lsbs; + // msbs = mulhi(n, mp) + umulExtended(n, mp, msbs, lsbs); + return (msbs + n) >> L; +} + uint src0_idx(uint idx) { - const uint i03 = idx / (p.ne02*p.ne01*p.ne00); + const uint i03 = fastdiv(idx, p.ne0_012mp, p.ne0_012L); const uint i03_offset = i03 * p.ne02*p.ne01*p.ne00; - const uint i02 = (idx - i03_offset) / (p.ne01*p.ne00); + const uint i02 = fastdiv(idx - i03_offset, p.ne0_01mp, p.ne0_01L); const uint i02_offset = i02*p.ne01*p.ne00; - const uint i01 = (idx - i03_offset - i02_offset) / p.ne00; + const uint i01 = fastdiv(idx - i03_offset - i02_offset, p.ne0_0mp, p.ne0_0L); const uint i00 = idx - i03_offset - i02_offset - i01*p.ne00; return i03*p.nb03 + i02*p.nb02 + i01*p.nb01 + i00*p.nb00; } uint dst_idx(uint idx) { - const uint i13 = idx / (p.ne12*p.ne11*p.ne10); + const uint i13 = fastdiv(idx, p.ne1_012mp, p.ne1_012L); const uint i13_offset = i13 * p.ne12*p.ne11*p.ne10; - const uint i12 = (idx - i13_offset) / (p.ne11*p.ne10); + const uint i12 = fastdiv(idx - i13_offset, p.ne1_01mp, p.ne1_01L); const uint i12_offset = i12*p.ne11*p.ne10; - const uint i11 = (idx - i13_offset - i12_offset) / p.ne10; + const uint i11 = fastdiv(idx - i13_offset - i12_offset, p.ne1_0mp, p.ne1_0L); const uint i10 = idx - i13_offset - i12_offset - i11*p.ne10; return i13*p.nb13 + i12*p.nb12 + i11*p.nb11 + i10*p.nb10; } diff --git a/ggml/src/vulkan-shaders/get_rows.comp b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp similarity index 61% rename from ggml/src/vulkan-shaders/get_rows.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp index e9ff22efa..e877ed779 100644 --- a/ggml/src/vulkan-shaders/get_rows.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp @@ -3,6 +3,8 @@ #include "types.comp" #include "generic_binary_head.comp" +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + void main() { const uint i00 = gl_GlobalInvocationID.x; const uint i10 = gl_GlobalInvocationID.y; @@ -13,10 +15,10 @@ void main() { return; } - const uint i01 = data_b[i10*p.nb10 + i11*p.nb11 + i12*p.nb12]; + const uint i01 = data_b[get_boffset() + i10*p.nb10 + i11*p.nb11 + i12*p.nb12]; - const uint a_offset = i01*p.nb01 + i11*p.nb02 + i12*p.nb03; - const uint d_offset = i10*p.nb21 + i11*p.nb22 + i12*p.nb23; + const uint a_offset = get_aoffset() + i01*p.nb01 + i11*p.nb02 + i12*p.nb03; + const uint d_offset = get_doffset() + i10*p.nb21 + i11*p.nb22 + i12*p.nb23; #ifndef OPTIMIZATION_ERROR_WORKAROUND data_d[d_offset + i00] = D_TYPE(data_a[a_offset + i00]); diff --git a/ggml/src/vulkan-shaders/get_rows_quant.comp b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp similarity index 83% rename from ggml/src/vulkan-shaders/get_rows_quant.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp index 53a9a96f2..1426fde65 100644 --- a/ggml/src/vulkan-shaders/get_rows_quant.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp @@ -4,12 +4,18 @@ #include "generic_binary_head.comp" #include "dequant_funcs.comp" +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + void main() { const uint i00 = (gl_GlobalInvocationID.x)*2; const uint i10 = gl_GlobalInvocationID.y; const uint i11 = (gl_GlobalInvocationID.z)/p.ne12; const uint i12 = (gl_GlobalInvocationID.z)%p.ne12; +#if defined(DATA_A_IQ4_NL) + init_iq4nl_shmem(); +#endif + if (i00 >= p.ne00) { return; } @@ -25,6 +31,8 @@ void main() { const uint y_offset = QUANT_R == 1 ? 1 : QUANT_K/2; vec2 v = dequantize(ib, iqs, 0); + const vec2 dm = get_dm(ib, 0); + v = v * dm.x + dm.y; data_d[d_offset + iybs + iqs ] = D_TYPE(v.x); data_d[d_offset + iybs + iqs + y_offset] = D_TYPE(v.y); diff --git a/ggml/src/vulkan-shaders/group_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp similarity index 96% rename from ggml/src/vulkan-shaders/group_norm.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp index 5ad9b28da..b6a0d5645 100644 --- a/ggml/src/vulkan-shaders/group_norm.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp @@ -19,7 +19,7 @@ void main() { const uint tid = gl_LocalInvocationID.x; const uint start = gl_WorkGroupID.x * group_size + tid; - const uint end = start + group_size; + const uint end = (gl_WorkGroupID.x + 1) * group_size; tmp[tid] = 0.0f; diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp b/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp new file mode 100644 index 000000000..122b1e93f --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp @@ -0,0 +1,87 @@ +#version 450 + +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_spirv_intrinsics: enable +#extension GL_EXT_control_flow_attributes : require + +#if RTE16 +spirv_execution_mode(capabilities = [4467], 4462, 16); // RoundingModeRTE, 16 bits +#endif + +layout (push_constant) uniform parameter +{ + uint batch_offset; uint offset_delta; + uint IC; + uint IW; uint IH; + uint OW; uint OH; + uint KW; uint KH; + uint pelements; + uint CHW; + int s0; int s1; + int p0; int p1; + int d0; int d1; +} p; + +#include "types.comp" + +layout(constant_id = 0) const uint BLOCK_SIZE = 32; + +const uint NUM_ITER = 512 / BLOCK_SIZE; + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +layout (binding = 0) readonly buffer X {A_TYPE data_a[];}; +layout (binding = 1) writeonly buffer D {D_TYPE data_d[];}; + +void main() { + const uint gidx = gl_GlobalInvocationID.x; + + const uint oh = gl_GlobalInvocationID.y; + const uint batch = gl_GlobalInvocationID.z / p.IC; + const uint ic = gl_GlobalInvocationID.z % p.IC; + + A_TYPE values[NUM_ITER]; + uint offset_dst[NUM_ITER]; + [[unroll]] for (uint idx = 0; idx < NUM_ITER; ++idx) { + values[idx] = A_TYPE(0); + } + + [[unroll]] for (uint idx = 0; idx < NUM_ITER; ++idx) { + + const uint i = gidx * NUM_ITER + idx; + + const uint ksize = p.OW * (p.KH > 1 ? p.KW : 1); + const uint kx = i / ksize; + const uint kd = kx * ksize; + const uint ky = (i - kd) / p.OW; + const uint ix = i % p.OW; + + const uint iiw = ix * p.s0 + kx * p.d0 - p.p0; + const uint iih = oh * p.s1 + ky * p.d1 - p.p1; + + offset_dst[idx] = + ((batch * p.OH + oh) * p.OW + ix) * p.CHW + + (ic * (p.KW * p.KH) + ky * p.KW + kx); + + if (i >= p.pelements) { + continue; + } + + if (iih < p.IH && iiw < p.IW) { + const uint offset_src = ic * p.offset_delta + batch * p.batch_offset; + values[idx] = data_a[offset_src + iih * p.IW + iiw]; + } + } + + [[unroll]] for (uint idx = 0; idx < NUM_ITER; ++idx) { + + const uint i = gidx * NUM_ITER + idx; + + if (i >= p.pelements) { + continue; + } + + data_d[offset_dst[idx]] = D_TYPE(values[idx]); + } + +} diff --git a/ggml/src/vulkan-shaders/leaky_relu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp similarity index 100% rename from ggml/src/vulkan-shaders/leaky_relu.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp new file mode 100644 index 000000000..43de19df8 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp @@ -0,0 +1,27 @@ +#version 450 + +#include "types.comp" +#include "generic_binary_head.comp" + +const uint num_threads = 256; + +layout(local_size_x = num_threads, local_size_y = 1, local_size_z = 1) in; + +void main() { + uint idx = get_idx(); + + // num_threads * num_iter must equal 512, to match the wg_denoms and get_idx calculation + const uint num_iter = 2; + + [[unroll]] for (uint i = 0; i < num_iter; ++i) { + if (idx >= p.ne) { + continue; + } + uint i00, i01, i02, i03; + get_indices(idx, i00, i01, i02, i03); + + data_d[get_doffset() + dst_idx(i00, i01, i02, i03)] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + src0_idx(i00, i01, i02, i03)]) * FLOAT_TYPE(data_b[get_boffset() + src1_idx(i00, i01, i02, i03)])); + + idx += num_threads; + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_split_k_reduce.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_split_k_reduce.comp new file mode 100644 index 000000000..4c64fd47a --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_split_k_reduce.comp @@ -0,0 +1,48 @@ +#version 450 + +#extension GL_EXT_control_flow_attributes : enable + +layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in; + +layout (binding = 0) readonly buffer A {float data_a[];}; +layout (binding = 0) readonly buffer A4 {vec4 data_a4[];}; +layout (binding = 1) writeonly buffer D {float data_d[];}; +layout (binding = 1) writeonly buffer D4 {vec4 data_d4[];}; + +layout (push_constant) uniform parameter { + uint ne; + uint k_num; +} p; + +void main() { + // Each invocation handles four consecutive components + const uint idx = gl_GlobalInvocationID.x * 4; + + if (idx >= p.ne) { + return; + } + + // Check if all four components are in bounds and aligned, + // then use vector loads + if (idx + 3 < p.ne && (p.ne % 4) == 0) { + vec4 result = vec4(0.0f); + + [[unroll]] for (uint i = 0; i < p.k_num; i++) { + result += data_a4[(i * p.ne + idx) / 4]; + } + + data_d4[idx / 4] = result; + } else { + [[unroll]] for (uint j = 0; j < 4; ++j) { + if (idx + j < p.ne) { + float result = 0.0f; + + [[unroll]] for (uint i = 0; i < p.k_num; i++) { + result += data_a[i * p.ne + idx + j]; + } + + data_d[idx + j] = result; + } + } + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp new file mode 100644 index 000000000..53902858d --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp @@ -0,0 +1,149 @@ +#version 450 + +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +#if !defined(DATA_A_F32) && !defined(DATA_A_F16) +#define K_PER_ITER 8 +#else +#define K_PER_ITER 2 +#endif + + +uint a_offset, b_offset, d_offset, y_offset; + +void iter(inout FLOAT_TYPE temp[NUM_COLS][NUM_ROWS], const uint first_row, const uint num_rows, const uint tid, const uint i, bool lastiter) +{ + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + const uint col = i*BLOCK_SIZE + K_PER_ITER*tid; + const uint iqs = (col%QUANT_K)/QUANT_R; // quant index + const uint iybs = col - col%QUANT_K; // y block start index + +#if K_PER_ITER == 8 +#if QUANT_R == 2 + const vec4 bv02 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + iybs + iqs) / 4]); + const vec4 bv13 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + iybs + iqs + y_offset) / 4]); + const vec4 bv0 = vec4(bv02.x, bv13.x, bv02.y, bv13.y); + const vec4 bv1 = vec4(bv02.z, bv13.z, bv02.w, bv13.w); +#else + const vec4 bv0 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + iybs + iqs) / 4]); + const vec4 bv1 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + iybs + iqs) / 4 + 1]); +#endif +#else + // Check if the second of the pair of elements is OOB, and don't fetch B or + // accumulate it. We still fetch a pair of elements for A, which is fine for + // quantized formats since they'll be within the same block. We should + // probably skip fetching the second element for F16/F32, but as of now we + // still do. + const bool OOB = lastiter && (iybs + iqs + y_offset >= p.ncols); + + FLOAT_TYPE b0 = 0, b1 = 0; + b0 = FLOAT_TYPE(data_b[j*p.batch_stride_b + b_offset + iybs + iqs]); + if (!OOB) { + b1 = FLOAT_TYPE(data_b[j*p.batch_stride_b + b_offset + iybs + iqs + y_offset]); + } +#endif + uint ibi = first_row*p.ncols; + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib = (ibi + col)/QUANT_K; // block index + ibi += p.ncols; + +#if K_PER_ITER == 8 + vec4 v = dequantize4(ib, iqs, a_offset); + vec4 v2 = dequantize4(ib, iqs+(4/QUANT_R), a_offset); + + const vec2 dm = get_dm(ib, a_offset); + if (dm.y != 0) { // quant has min component + v = v * dm.x + dm.y; + v2 = v2 * dm.x + dm.y; + } + + // matrix multiplication + FLOAT_TYPE rowtmp = dot(bv0, v); + rowtmp += dot(bv1, v2); + + if (dm.y == 0) + rowtmp *= dm.x; + + temp[j][n] += rowtmp; +#else + const vec2 v = dequantize(ib, iqs, a_offset); + + // matrix multiplication + temp[j][n] = fma(FLOAT_TYPE(v.x), b0, temp[j][n]); + if (!OOB) { + temp[j][n] = fma(FLOAT_TYPE(v.y), b1, temp[j][n]); + } +#endif + } + } +} + +void compute_outputs(const uint32_t first_row, const uint32_t num_rows) { + const uint tid = gl_LocalInvocationID.x; + + get_offsets(a_offset, b_offset, d_offset); + a_offset /= QUANT_K; + + y_offset = QUANT_R == 1 ? 1 : QUANT_K/2; + + FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + uint num_iters = p.ncols / (K_PER_ITER * BLOCK_SIZE); + if (num_iters * K_PER_ITER * BLOCK_SIZE + K_PER_ITER*tid < p.ncols) { + num_iters++; + } + int unroll_count = 4; + uint unrolled_iters = num_iters & ~(unroll_count - 1); + + uint i = 0; + while (i < unrolled_iters) { + // Manually partially unroll the loop + [[unroll]] for (uint k = 0; k < unroll_count; ++k) { + iter(temp, first_row, num_rows, tid, i*K_PER_ITER, false); + i++; + } + } + unroll_count = 2; + unrolled_iters = num_iters & ~(unroll_count - 1); + while (i < unrolled_iters) { + // Manually partially unroll the loop + [[unroll]] for (uint k = 0; k < unroll_count; ++k) { + iter(temp, first_row, num_rows, tid, i*K_PER_ITER, false); + i++; + } + } + while (i < num_iters) { + iter(temp, first_row, num_rows, tid, i*K_PER_ITER, true); + i++; + } + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + +#if defined(DATA_A_IQ4_NL) + init_iq4nl_shmem(); +#endif + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp new file mode 100644 index 000000000..903753c7e --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp @@ -0,0 +1,118 @@ +#extension GL_EXT_control_flow_attributes : enable +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_shader_8bit_storage : require + +#ifdef MUL_MAT_ID +#define EXPERT_COUNT 8 +#endif + +#include "types.comp" + +layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; +layout (binding = 1) readonly buffer B {B_TYPE data_b[];}; +layout (binding = 1) readonly buffer BV2 {B_TYPE_VEC2 data_b_v2[];}; +layout (binding = 1) readonly buffer BV4 {B_TYPE_VEC4 data_b_v4[];}; + +layout (binding = 2) writeonly buffer D {D_TYPE data_d[];}; +#ifdef MUL_MAT_ID +layout (binding = 3) readonly buffer IDS {int data_ids[];}; +#endif + +#include "dequant_funcs.comp" + +layout (push_constant) uniform parameter +{ + uint ncols; + uint stride_a; + uint stride_b; + uint stride_d; + + uint batch_stride_a; + uint batch_stride_b; + uint batch_stride_d; + +#ifdef MUL_MAT_ID + uint nei0; + uint ne11; +#else + uint ne02; + uint ne12; + uint broadcast2; + uint broadcast3; +#endif +} p; + +void get_offsets(out uint a_offset, out uint b_offset, out uint d_offset) { +#ifdef MUL_MAT_ID + const uint expert_idx = gl_GlobalInvocationID.y; +#else + const uint batch_idx = gl_GlobalInvocationID.y; +#endif + +#ifndef MUL_MAT_ID + uint batch_idx_a = 0; + if (batch_idx != 0) { + const uint i13 = batch_idx / p.ne12; + const uint i12 = batch_idx % p.ne12; + + const uint i03 = i13 / p.broadcast3; + const uint i02 = i12 / p.broadcast2; + + batch_idx_a = i03 * p.ne02 + i02; + } +#else + const uint expert_id = data_ids[expert_idx]; +#endif + + a_offset = +#ifdef MUL_MAT_ID + expert_id * p.batch_stride_a; +#else + batch_idx_a * p.batch_stride_a; +#endif + b_offset = +#ifdef MUL_MAT_ID + (expert_idx % p.ne11) * p.stride_b; +#else + batch_idx * p.batch_stride_b; +#endif + d_offset = +#ifdef MUL_MAT_ID + expert_idx * p.stride_d; +#else + batch_idx * p.batch_stride_d; +#endif +} + +layout (constant_id = 0) const uint BLOCK_SIZE = 32; +layout (constant_id = 1) const uint NUM_ROWS = 1; +layout (constant_id = 2) const uint NUM_COLS = 1; + +shared FLOAT_TYPE tmpsh[NUM_COLS][NUM_ROWS][BLOCK_SIZE]; + +void reduce_result(const in FLOAT_TYPE temp[NUM_COLS][NUM_ROWS], const in uint32_t d_offset, const in uint32_t first_row, const in uint32_t num_rows, const in uint32_t tid) { + // sum up partial sums and write back result + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + tmpsh[j][n][tid] = temp[j][n]; + } + } + barrier(); + [[unroll]] for (uint s = BLOCK_SIZE/2; s > 0; s >>= 1) { + if (tid < s) { + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + tmpsh[j][n][tid] += tmpsh[j][n][tid + s]; + } + } + } + barrier(); + } + if (tid == 0) { + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + data_d[j*p.batch_stride_d + d_offset + first_row + n] = D_TYPE(tmpsh[j][n][0]); + } + } + } +} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_nc.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_nc.comp similarity index 100% rename from ggml/src/vulkan-shaders/mul_mat_vec_nc.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_nc.comp diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_p021.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_p021.comp similarity index 100% rename from ggml/src/vulkan-shaders/mul_mat_vec_p021.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_p021.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp new file mode 100644 index 000000000..8cdc640e8 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp @@ -0,0 +1,129 @@ +#version 450 +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +shared FLOAT_TYPE sccache1[BLOCK_SIZE/16][16]; +shared FLOAT_TYPE sccache2[BLOCK_SIZE/16][16]; + +FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + +void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint v_im, const uint ix, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) { + const uint y_idx = i * QUANT_K + y_offset; + + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row; + + barrier(); + if (!all_threads) { // when we don't have enough blocks to use all threads + if (i < num_blocks_per_row) { + const uint32_t scale = uint32_t(data_a[ib0 + i].scales[itid]); + sccache1[ix][itid] = FLOAT_TYPE(scale & 0xF); + sccache2[ix][itid] = FLOAT_TYPE((scale >> 4) & 0xF); + } + barrier(); + + if (i >= num_blocks_per_row) + continue; + } else { + const uint32_t scale = uint32_t(data_a[ib0 + i].scales[itid]); + sccache1[ix][itid] = FLOAT_TYPE(scale & 0xF); + sccache2[ix][itid] = FLOAT_TYPE((scale >> 4) & 0xF); + barrier(); + } + + const uint32_t qs_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16); + const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303)); + const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303)); + const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303)); + const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303)); + + vec2 d = vec2(data_a[ib0 + i].d); + const FLOAT_TYPE dall = FLOAT_TYPE(d.x); + const FLOAT_TYPE dmin = FLOAT_TYPE(d.y); + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + vec2 b0 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 0]); + vec2 b16 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8]); + vec2 b32 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16]); + vec2 b48 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24]); + vec2 b64 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32]); + vec2 b80 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40]); + vec2 b96 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48]); + vec2 b112 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56]); + + FLOAT_TYPE sum1 = FLOAT_TYPE(0.0); + FLOAT_TYPE sum2 = FLOAT_TYPE(0.0); + [[unroll]] for (int l = 0; l < 2; ++l) { + sum1 = fma(FLOAT_TYPE(b0[l]), sccache1[ix][ 8*v_im] * qs_u32_0[l ], + fma(FLOAT_TYPE(b16[l]), sccache1[ix][1 + 8*v_im] * qs_u32_0[l+2], + fma(FLOAT_TYPE(b32[l]), sccache1[ix][2 + 8*v_im] * qs_u32_2[l ], + fma(FLOAT_TYPE(b48[l]), sccache1[ix][3 + 8*v_im] * qs_u32_2[l+2], + fma(FLOAT_TYPE(b64[l]), sccache1[ix][4 + 8*v_im] * qs_u32_4[l ], + fma(FLOAT_TYPE(b80[l]), sccache1[ix][5 + 8*v_im] * qs_u32_4[l+2], + fma(FLOAT_TYPE(b96[l]), sccache1[ix][6 + 8*v_im] * qs_u32_6[l ], + fma(FLOAT_TYPE(b112[l]), sccache1[ix][7 + 8*v_im] * qs_u32_6[l+2], sum1)))))))); + sum2 = fma(FLOAT_TYPE(b0[l]), sccache2[ix][ 8*v_im], + fma(FLOAT_TYPE(b16[l]), sccache2[ix][1 + 8*v_im], + fma(FLOAT_TYPE(b32[l]), sccache2[ix][2 + 8*v_im], + fma(FLOAT_TYPE(b48[l]), sccache2[ix][3 + 8*v_im], + fma(FLOAT_TYPE(b64[l]), sccache2[ix][4 + 8*v_im], + fma(FLOAT_TYPE(b80[l]), sccache2[ix][5 + 8*v_im], + fma(FLOAT_TYPE(b96[l]), sccache2[ix][6 + 8*v_im], + fma(FLOAT_TYPE(b112[l]), sccache2[ix][7 + 8*v_im], sum2)))))))); + } + temp[j][n] = fma(dall, sum1, fma(-dmin, sum2, temp[j][n])); + } + } +} + +void compute_outputs(const uint32_t first_row, const uint32_t num_rows) { + uint a_offset, b_offset, d_offset; + get_offsets(a_offset, b_offset, d_offset); + + const uint num_blocks_per_row = p.ncols / QUANT_K; + + // 16 threads are used to process each block + const uint it_size = gl_WorkGroupSize.x/16; + const uint tid = gl_LocalInvocationID.x; + const uint itid = tid%16; // 0...15 + const uint ix = tid/16; + + const uint v_im = itid/8; // 0 or 1. 0 computes 0..., 1 computes 128... + const uint v_in = itid - 8*v_im; // 0...7 + + const uint l0 = 2*v_in; // 0...15 + const uint q_offset = 32*v_im + l0; + const uint y_offset = 128*v_im + l0; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + const uint nbr_par_th = num_blocks_per_row%it_size; + const uint nbr_all_th = num_blocks_per_row - nbr_par_th; + uint i0 = 0; + [[unroll]] for (; i0 < nbr_all_th; i0 += it_size) + calc_superblock(a_offset, b_offset, itid, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true); + calc_superblock(a_offset, b_offset, itid, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false); + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp new file mode 100644 index 000000000..3116fad16 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp @@ -0,0 +1,132 @@ +#version 450 +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8]; + +FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + +void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, const uint itid8, const uint v_im, const uint v_im4, const uint v_in, const uint32_t hm_m[4], const uint q_offset, const uint y_offset, const uint s_shift, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) { + const uint y_idx = i * QUANT_K + y_offset; + + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row; + + if (!all_threads) { // when we don't have enough blocks to use all threads + barrier(); + if (i < num_blocks_per_row) + sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32); + barrier(); + + if (i >= num_blocks_per_row) + continue; + } + + const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16)); + const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> ( v_im4)) << 2)); + const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2)); + const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2)); + const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2)); + + // 0, 1, 16, 17 + uint32_t qs_u32 = uint32_t(data_a[ib0 + i].qs[q_offset]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 1]) << 8); + qs_u32 |= (uint32_t(data_a[ib0 + i].qs[q_offset + 16]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 17]) << 8)) << 16; + const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303)); + const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303)); + const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303)); + const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303)); + + if (all_threads) { + barrier(); + sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32); + barrier(); + } + + const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d); + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + vec2 b0 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 0]); + vec2 b16 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8]); + vec2 b32 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16]); + vec2 b48 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24]); + vec2 b64 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32]); + vec2 b80 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40]); + vec2 b96 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48]); + vec2 b112 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56]); + + FLOAT_TYPE sum = FLOAT_TYPE(0.0); + [[unroll]] for (int l = 0; l < 2; ++l) { + sum = fma(FLOAT_TYPE( b0[l]) * sccache[ix][v_im][0], qs_u32_0[l ] - hmk_0[l ], + fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - hmk_0[l+2], + fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l ] - hmk_1[l ], + fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - hmk_1[l+2], + fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l ] - hmk_2[l ], + fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - hmk_2[l+2], + fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l ] - hmk_3[l ], + fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - hmk_3[l+2], sum)))))))); + } + temp[j][n] = fma(d, sum, temp[j][n]); + } + } +} + +void compute_outputs(const uint32_t first_row, const uint32_t num_rows) { + uint a_offset, b_offset, d_offset; + get_offsets(a_offset, b_offset, d_offset); + + const uint num_blocks_per_row = p.ncols / QUANT_K; + + // 16 threads are used to process each block + const uint it_size = gl_WorkGroupSize.x/16; + const uint tid = gl_LocalInvocationID.x; + const uint itid = tid%16; // 0...15 + const uint ix = tid/16; + const uint itid8 = itid%8; + + const uint v_im = itid/8; // 0 or 1. 0 computes 0..., 1 computes 128... + const uint v_im4 = v_im*4; + const uint v_in = itid - 8*v_im; // 0...7 + + const uint32_t m = 0x01010101 << (4 * v_im); + uint32_t hm_m[4]; + [[unroll]] for (uint j = 0; j < 4; ++j) + hm_m[j] = m << j; + + const uint l0 = 2*v_in; // 0...15 + const uint q_offset = 32*v_im + l0; + const uint y_offset = 128*v_im + l0; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + const uint s_shift = v_im4 + 2*(itid8/4); + + const uint nbr_par_th = num_blocks_per_row%it_size; + const uint nbr_all_th = num_blocks_per_row - nbr_par_th; + uint i0 = 0; + [[unroll]] for (; i0 < nbr_all_th; i0 += it_size) + calc_superblock(a_offset, b_offset, ix, itid8, v_im, v_im4, v_in, hm_m, q_offset, y_offset, s_shift, i0 + ix, num_blocks_per_row, first_row, num_rows, true); + calc_superblock(a_offset, b_offset, ix, itid8, v_im, v_im4, v_in, hm_m, q_offset, y_offset, s_shift, i0 + ix, num_blocks_per_row, first_row, num_rows, false); + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp new file mode 100644 index 000000000..f9cde0648 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp @@ -0,0 +1,136 @@ +#version 450 + +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + +void calc_superblock(const uint a_offset, const uint b_offset, const uint v_im, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows) { + const uint y1_idx = i * QUANT_K + y_offset; + const uint y2_idx = y1_idx + 128; + + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row; + vec2 d = vec2(data_a[ib0 + i].d); + const FLOAT_TYPE dall = FLOAT_TYPE(d.x); + const FLOAT_TYPE dmin = FLOAT_TYPE(d.y); + + const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im ]; + const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2]; + const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4]; + + const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32; + const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2; + const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F)); + const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h)); + + const FLOAT_TYPE sc0 = scale_0_4_l_f.x; + const FLOAT_TYPE sc1 = scale_0_4_l_f.y; + const FLOAT_TYPE sc2 = scale_0_4_l_f.z; + const FLOAT_TYPE sc3 = scale_0_4_l_f.w; + const FLOAT_TYPE sc4 = scale8_f.x; + const FLOAT_TYPE sc5 = scale8_f.y; + const FLOAT_TYPE sc6 = scale8_f.z; + const FLOAT_TYPE sc7 = scale8_f.w; + + const uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4]; + const uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16]; + + const uint32_t qs0_u32_lo4 = qs0_u32 & 0x0F0F0F0F; + const uint32_t qs0_u32_hi4 = (qs0_u32 >> 4) & 0x0F0F0F0F; + const uint32_t qs64_u32_lo4 = qs64_u32 & 0x0F0F0F0F; + const uint32_t qs64_u32_hi4 = (qs64_u32 >> 4) & 0x0F0F0F0F; + + const vec4 qs0_lo4 = vec4(unpack8(qs0_u32_lo4)); + const vec4 qs64_lo4 = vec4(unpack8(qs64_u32_lo4)); + const vec4 qs0_hi4 = vec4(unpack8(qs0_u32_hi4)); + const vec4 qs64_hi4 = vec4(unpack8(qs64_u32_hi4)); + + const FLOAT_TYPE q4_0 = qs0_lo4.x; + const FLOAT_TYPE q4_1 = qs0_lo4.y; + const FLOAT_TYPE q4_2 = qs0_lo4.z; + const FLOAT_TYPE q4_3 = qs0_lo4.w; + const FLOAT_TYPE q4_4 = qs0_hi4.x; + const FLOAT_TYPE q4_5 = qs0_hi4.y; + const FLOAT_TYPE q4_6 = qs0_hi4.z; + const FLOAT_TYPE q4_7 = qs0_hi4.w; + const FLOAT_TYPE q4_8 = qs64_lo4.x; + const FLOAT_TYPE q4_9 = qs64_lo4.y; + const FLOAT_TYPE q4_10 = qs64_lo4.z; + const FLOAT_TYPE q4_11 = qs64_lo4.w; + const FLOAT_TYPE q4_12 = qs64_hi4.x; + const FLOAT_TYPE q4_13 = qs64_hi4.y; + const FLOAT_TYPE q4_14 = qs64_hi4.z; + const FLOAT_TYPE q4_15 = qs64_hi4.w; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + vec4 by10 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4 ]); + vec4 by132 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4 + 8]); + vec4 by20 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4 ]); + vec4 by232 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4 + 8]); + + const FLOAT_TYPE sx = fma(FLOAT_TYPE(by10.x), q4_0, fma(FLOAT_TYPE(by10.y), q4_1, fma(FLOAT_TYPE(by10.z), q4_2, FLOAT_TYPE(by10.w) * q4_3))); + const FLOAT_TYPE sy = fma(FLOAT_TYPE(by132.x), q4_4, fma(FLOAT_TYPE(by132.y), q4_5, fma(FLOAT_TYPE(by132.z), q4_6, FLOAT_TYPE(by132.w) * q4_7))); + const FLOAT_TYPE sz = fma(FLOAT_TYPE(by20.x), q4_8, fma(FLOAT_TYPE(by20.y), q4_9, fma(FLOAT_TYPE(by20.z), q4_10, FLOAT_TYPE(by20.w) * q4_11))); + const FLOAT_TYPE sw = fma(FLOAT_TYPE(by232.x), q4_12, fma(FLOAT_TYPE(by232.y), q4_13, fma(FLOAT_TYPE(by232.z), q4_14, FLOAT_TYPE(by232.w) * q4_15))); + const FLOAT_TYPE smin = + fma(FLOAT_TYPE(by10.x), sc2, fma(FLOAT_TYPE(by132.x), sc3, fma(FLOAT_TYPE(by20.x), sc6, fma(FLOAT_TYPE(by232.x), sc7, + fma(FLOAT_TYPE(by10.y), sc2, fma(FLOAT_TYPE(by132.y), sc3, fma(FLOAT_TYPE(by20.y), sc6, fma(FLOAT_TYPE(by232.y), sc7, + fma(FLOAT_TYPE(by10.z), sc2, fma(FLOAT_TYPE(by132.z), sc3, fma(FLOAT_TYPE(by20.z), sc6, fma(FLOAT_TYPE(by232.z), sc7, + fma(FLOAT_TYPE(by10.w), sc2, fma(FLOAT_TYPE(by132.w), sc3, fma(FLOAT_TYPE(by20.w), sc6, FLOAT_TYPE(by232.w) * sc7))))))))))))))); + temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n])); + } + } +} + +void compute_outputs(const uint32_t first_row, const uint32_t num_rows) { + uint a_offset, b_offset, d_offset; + get_offsets(a_offset, b_offset, d_offset); + + const uint num_blocks_per_row = p.ncols / QUANT_K; + + // 16 threads are used to process each block + const uint it_size = gl_WorkGroupSize.x/16; + const uint tid = gl_LocalInvocationID.x; + const uint itid = tid%16; // 0...15 + const uint ix = tid/16; + + const uint il = itid/4; // 0...3 + const uint ir = itid - 4*il; // 0...3 + const uint n = 4; + + const uint v_im = il / 2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 + const uint v_in = il % 2; + + const uint l0 = n * (2 * ir + v_in); // 0...15 + const uint q_offset = 32*v_im + l0; + const uint y_offset = 64*v_im + l0; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) + calc_superblock(a_offset, b_offset, v_im, q_offset, y_offset, i, num_blocks_per_row, first_row, num_rows); + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp new file mode 100644 index 000000000..6c84ef3cd --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp @@ -0,0 +1,167 @@ +#version 450 + +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + +void calc_superblock(const uint a_offset, const uint b_offset, const uint v_im, const uint l0, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows) { + const uint y1_idx = i * QUANT_K + y_offset; + const uint y2_idx = y1_idx + 128; + + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row; + vec2 d = vec2(data_a[ib0 + i].d); + const FLOAT_TYPE dall = FLOAT_TYPE(d.x); + const FLOAT_TYPE dmin = FLOAT_TYPE(d.y); + + const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im ]; + const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2]; + const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4]; + + const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32; + const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2; + const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F)); + const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h)); + + const FLOAT_TYPE sc0 = scale_0_4_l_f.x; + const FLOAT_TYPE sc1 = scale_0_4_l_f.y; + const FLOAT_TYPE sc2 = scale_0_4_l_f.z; + const FLOAT_TYPE sc3 = scale_0_4_l_f.w; + const FLOAT_TYPE sc4 = scale8_f.x; + const FLOAT_TYPE sc5 = scale8_f.y; + const FLOAT_TYPE sc6 = scale8_f.z; + const FLOAT_TYPE sc7 = scale8_f.w; + + const uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16); + const uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16); + + uint32_t qs0_16_u32_lo4 = qs0_16_u32 & 0x0F0F0F0F; + uint32_t qs0_16_u32_hi4 = (qs0_16_u32 >> 4) & 0x0F0F0F0F; + uint32_t qs64_80_u32_lo4 = qs64_80_u32 & 0x0F0F0F0F; + uint32_t qs64_80_u32_hi4 = (qs64_80_u32 >> 4) & 0x0F0F0F0F; + + const uint32_t qh = pack32(u16vec2(data_a_packed16[ib0 + i].qh[l0 / 2], data_a_packed16[ib0 + i].qh[l0 / 2 + 8])); + + const uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4; + const uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3; + const uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010); + const uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1; + + qs0_16_u32_lo4 += qs0_16_lo4_offset16; + qs0_16_u32_hi4 += qs0_16_hi4_offset16; + qs64_80_u32_lo4 += qs64_80_lo4_offset16; + qs64_80_u32_hi4 += qs64_80_hi4_offset16; + + const vec4 qs0_16_lo4 = vec4(unpack8(qs0_16_u32_lo4)); + const vec4 qs64_80_lo4 = vec4(unpack8(qs64_80_u32_lo4)); + const vec4 qs0_16_hi4 = vec4(unpack8(qs0_16_u32_hi4)); + const vec4 qs64_80_hi4 = vec4(unpack8(qs64_80_u32_hi4)); + + const FLOAT_TYPE q4_0 = qs0_16_lo4.x; + const FLOAT_TYPE q4_1 = qs0_16_lo4.y; + const FLOAT_TYPE q4_2 = qs0_16_lo4.z; + const FLOAT_TYPE q4_3 = qs0_16_lo4.w; + const FLOAT_TYPE q4_4 = qs0_16_hi4.x; + const FLOAT_TYPE q4_5 = qs0_16_hi4.y; + const FLOAT_TYPE q4_6 = qs0_16_hi4.z; + const FLOAT_TYPE q4_7 = qs0_16_hi4.w; + const FLOAT_TYPE q4_8 = qs64_80_lo4.x; + const FLOAT_TYPE q4_9 = qs64_80_lo4.y; + const FLOAT_TYPE q4_10 = qs64_80_lo4.z; + const FLOAT_TYPE q4_11 = qs64_80_lo4.w; + const FLOAT_TYPE q4_12 = qs64_80_hi4.x; + const FLOAT_TYPE q4_13 = qs64_80_hi4.y; + const FLOAT_TYPE q4_14 = qs64_80_hi4.z; + const FLOAT_TYPE q4_15 = qs64_80_hi4.w; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + vec2 by10 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 ]); + vec2 by116 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 8]); + vec2 by132 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 16]); + vec2 by148 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 24]); + vec2 by20 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 ]); + vec2 by216 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 8]); + vec2 by232 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 16]); + vec2 by248 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 24]); + + const FLOAT_TYPE sx = + fma(FLOAT_TYPE(by10.x), q4_0, + fma(FLOAT_TYPE(by10.y), q4_1, + fma(FLOAT_TYPE(by116.x), q4_2, + FLOAT_TYPE(by116.y) * q4_3))); + const FLOAT_TYPE sy = + fma(FLOAT_TYPE(by132.x), q4_4, + fma(FLOAT_TYPE(by132.y), q4_5, + fma(FLOAT_TYPE(by148.x), q4_6, + FLOAT_TYPE(by148.y) * q4_7))); + const FLOAT_TYPE sz = + fma(FLOAT_TYPE(by20.x), q4_8, + fma(FLOAT_TYPE(by20.y), q4_9, + fma(FLOAT_TYPE(by216.x), q4_10, + FLOAT_TYPE(by216.y) * q4_11))); + const FLOAT_TYPE sw = + fma(FLOAT_TYPE(by232.x), q4_12, + fma(FLOAT_TYPE(by232.y), q4_13, + fma(FLOAT_TYPE(by248.x), q4_14, + FLOAT_TYPE(by248.y) * q4_15))); + const FLOAT_TYPE smin = + fma(FLOAT_TYPE(by10.x) + FLOAT_TYPE(by10.y) + FLOAT_TYPE(by116.x) + FLOAT_TYPE(by116.y), sc2, + fma(FLOAT_TYPE(by132.x) + FLOAT_TYPE(by132.y) + FLOAT_TYPE(by148.x) + FLOAT_TYPE(by148.y), sc3, + fma(FLOAT_TYPE(by20.x) + FLOAT_TYPE(by20.y) + FLOAT_TYPE(by216.x) + FLOAT_TYPE(by216.y), sc6, + (FLOAT_TYPE(by232.x) + FLOAT_TYPE(by232.y) + FLOAT_TYPE(by248.x) + FLOAT_TYPE(by248.y)) * sc7))); + temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n])); + } + } +} + +void compute_outputs(const uint32_t first_row, const uint32_t num_rows) { + uint a_offset, b_offset, d_offset; + get_offsets(a_offset, b_offset, d_offset); + + const uint num_blocks_per_row = p.ncols / QUANT_K; + + // 16 threads are used to process each block + const uint it_size = gl_WorkGroupSize.x/16; + const uint tid = gl_LocalInvocationID.x; + const uint itid = tid%16; // 0...15 + const uint ix = tid/16; + + const uint il = itid/4; // 0...3 + const uint ir = itid - 4*il; // 0...3 + + const uint v_im = il / 2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 + const uint v_in = il % 2; + + const uint l0 = 4*ir + 2*v_in; // 0...15 + const uint q_offset = 32*v_im + l0; + const uint y_offset = 64*v_im + l0; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) + calc_superblock(a_offset, b_offset, v_im, l0, q_offset, y_offset, i, num_blocks_per_row, first_row, num_rows); + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp new file mode 100644 index 000000000..f05f96b5e --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp @@ -0,0 +1,130 @@ +#version 450 + +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require + +#include "mul_mat_vec_base.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +shared FLOAT_TYPE sccache[BLOCK_SIZE/16][16]; + +FLOAT_TYPE temp[NUM_COLS][NUM_ROWS]; + +void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint ix, const uint ql_offset, const uint qh_offset, const uint s_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) { + const uint y_idx = i * QUANT_K + y_offset; + + [[unroll]] for (uint n = 0; n < num_rows; ++n) { + const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row; + + if (!all_threads) { // when we don't have enough blocks to use all threads + barrier(); + if (i < num_blocks_per_row) + sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]); + barrier(); + + if (i >= num_blocks_per_row) + continue; + } + + const uint32_t ql0_u32 = uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 1]) << 16); + const uint32_t ql32_u32 = uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 16]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 17]) << 16); + + const uint32_t ql0_u32_lo4 = ql0_u32 & 0x0F0F0F0F; + const uint32_t ql0_u32_hi4 = (ql0_u32 >> 4) & 0x0F0F0F0F; + const uint32_t ql32_u32_lo4 = ql32_u32 & 0x0F0F0F0F; + const uint32_t ql32_u32_hi4 = (ql32_u32 >> 4) & 0x0F0F0F0F; + + const uint32_t qh_u32 = uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2 + 1]) << 16); + const uint32_t qh0_u32 = (qh_u32 & 0x03030303) << 4; + const uint32_t qh2_u32 = (qh_u32 & 0x0C0C0C0C) << 2; + const uint32_t qh4_u32 = (qh_u32 & 0x30303030); + const uint32_t qh6_u32 = (qh_u32 & 0xC0C0C0C0) >> 2; + + const uint32_t q0_u32 = ql0_u32_lo4 | qh0_u32; + const uint32_t q1_u32 = ql32_u32_lo4 | qh2_u32; + const uint32_t q2_u32 = ql0_u32_hi4 | qh4_u32; + const uint32_t q3_u32 = ql32_u32_hi4 | qh6_u32; + + const vec4 q0 = vec4(unpack8(q0_u32)) - 32; + const vec4 q1 = vec4(unpack8(q1_u32)) - 32; + const vec4 q2 = vec4(unpack8(q2_u32)) - 32; + const vec4 q3 = vec4(unpack8(q3_u32)) - 32; + + if (all_threads) { + barrier(); + sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]); + barrier(); + } + + const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d); + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + vec4 by0 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 ]); + vec4 by32 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 8]); + vec4 by64 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 16]); + vec4 by96 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 24]); + + FLOAT_TYPE sum[4] = {0, 0, 0, 0}; + [[unroll]] for (uint l = 0; l < 4; ++l) { + sum[0] = fma(FLOAT_TYPE(by0[l]), q0[l], sum[0]); + sum[1] = fma(FLOAT_TYPE(by32[l]), q1[l], sum[1]); + sum[2] = fma(FLOAT_TYPE(by64[l]), q2[l], sum[2]); + sum[3] = fma(FLOAT_TYPE(by96[l]), q3[l], sum[3]); + } + temp[j][n] = fma(fma(sum[0], sccache[ix][s_offset], fma(sum[1], sccache[ix][s_offset + 2], fma(sum[2], sccache[ix][s_offset + 4], sum[3] * sccache[ix][s_offset + 6]))), d, temp[j][n]); + } + } +} + +void compute_outputs(const uint first_row, const uint num_rows) { + uint a_offset, b_offset, d_offset; + get_offsets(a_offset, b_offset, d_offset); + + const uint num_blocks_per_row = p.ncols / QUANT_K; + + // 16 threads are used to process each block + const uint it_size = gl_WorkGroupSize.x/16; + const uint tid = gl_LocalInvocationID.x; + const uint itid = tid%16; // 0...15 + const uint ix = tid/16; + + const uint v_im = itid/8; // 0 or 1. 0 computes 0..., 1 computes 128... + const uint v_in = itid - 8*v_im; // 0...7 + + const uint l0 = 4 * v_in; // 0, 4, 8, ..., 28 + const uint is = v_in / 4; + + const uint ql_offset = 64*v_im + l0; + const uint qh_offset = 32*v_im + l0; + const uint s_offset = 8*v_im + is; + const uint y_offset = 128*v_im + l0; + + [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) { + [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) { + temp[j][i] = FLOAT_TYPE(0); + } + } + + const uint nbr_par_th = num_blocks_per_row%it_size; + const uint nbr_all_th = num_blocks_per_row - nbr_par_th; + uint i0 = 0; + [[unroll]] for (; i0 < nbr_all_th; i0 += it_size) + calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true); + calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false); + + reduce_result(temp, d_offset, first_row, num_rows, tid); +} + +void main() { + const uint first_row = NUM_ROWS * (gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z); + + // do NUM_ROWS at a time, unless there aren't enough remaining rows + if (first_row + NUM_ROWS <= p.stride_d) { + compute_outputs(first_row, NUM_ROWS); + } else { + if (first_row >= p.stride_d) { + return; + } + compute_outputs(first_row, p.stride_d - first_row); + } +} diff --git a/ggml/src/vulkan-shaders/mul_mm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp similarity index 66% rename from ggml/src/vulkan-shaders/mul_mm.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp index fffdd1818..48122cbef 100644 --- a/ggml/src/vulkan-shaders/mul_mm.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp @@ -7,6 +7,12 @@ #extension GL_EXT_shader_explicit_arithmetic_types_float16 : require #endif +#ifdef COOPMAT +#extension GL_KHR_cooperative_matrix : enable +#extension GL_KHR_memory_scope_semantics : enable +#extension GL_KHR_shader_subgroup_basic : enable +#endif + #ifdef MUL_MAT_ID #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require #endif @@ -57,6 +63,7 @@ layout (push_constant) uniform parameter #endif } p; +layout (constant_id = 0) const uint BLOCK_SIZE = 64; layout (constant_id = 1) const uint BM = 64; layout (constant_id = 2) const uint BN = 64; layout (constant_id = 3) const uint BK = 16; // Assumed to be 32 if working with a quant @@ -65,16 +72,33 @@ layout (constant_id = 5) const uint WN = 32; layout (constant_id = 6) const uint WMITER = 2; layout (constant_id = 7) const uint TM = 4; layout (constant_id = 8) const uint TN = 2; -layout (constant_id = 9) const uint WARP = 32; +layout (constant_id = 9) const uint TK = 1; // Only needed for coopmat +layout (constant_id = 10) const uint WARP = 32; -shared FLOAT_TYPE buf_a[BM * (BK+1)]; -shared FLOAT_TYPE buf_b[BN * (BK+1)]; +#ifdef COOPMAT +#define SHMEM_STRIDE (BK + 8) +#else +#define SHMEM_STRIDE (BK + 1) +#endif + +shared FLOAT_TYPE buf_a[BM * SHMEM_STRIDE]; +shared FLOAT_TYPE buf_b[BN * SHMEM_STRIDE]; #ifdef MUL_MAT_ID shared u16vec2 row_ids[3072]; +#endif // MUL_MAT_ID + +#define NUM_WARPS (BLOCK_SIZE / WARP) + +#ifdef COOPMAT +shared ACC_TYPE coopmat_stage[TM * TN * NUM_WARPS]; #endif void main() { +#if defined(DATA_A_IQ4_NL) + init_iq4nl_shmem(); +#endif + #ifdef MUL_MAT_ID const uint expert_idx = gl_GlobalInvocationID.z; #else @@ -94,17 +118,32 @@ void main() { const uint ik = gl_WorkGroupID.x / blocks_m; const uint ic = gl_WorkGroupID.y; - const uint warp_i = gl_LocalInvocationID.x / WARP; - const uint warp_r = warp_i % (BM / WM); - const uint warp_c = warp_i / (BM / WM); - const uint WNITER = (WM * WN) / (WARP * TM * TN * WMITER); const uint WSUBM = WM / WMITER; const uint WSUBN = WN / WNITER; +#ifdef COOPMAT + const uint warp_i = gl_SubgroupID; + + const uint tiw = gl_SubgroupInvocationID; + + const uint cms_per_row = WM / TM; + const uint cms_per_col = WN / TN; + + const uint storestride = WARP / TM; + const uint store_r = tiw % TM; + const uint store_c = tiw / TM; +#else + const uint warp_i = gl_LocalInvocationID.x / WARP; + const uint tiw = gl_LocalInvocationID.x % WARP; + const uint tiwr = tiw % (WSUBM / TM); const uint tiwc = tiw / (WSUBM / TM); +#endif + + const uint warp_r = warp_i % (BM / WM); + const uint warp_c = warp_i / (BM / WM); const uint loadr_a = gl_LocalInvocationID.x % (BK / LOAD_VEC_A); const uint loadc_a = gl_LocalInvocationID.x / (BK / LOAD_VEC_A); @@ -152,21 +191,31 @@ void main() { uint pos_b = (batch_idx * p.batch_stride_b + ic * BN * p.stride_b + start_k) / LOAD_VEC_B; #endif - float sums[WMITER * TM * WNITER * TN]; +#ifdef COOPMAT + coopmat cache_a; + coopmat cache_b; + coopmat sums[cms_per_row * cms_per_col]; + + [[unroll]] for (uint i = 0; i < cms_per_row * cms_per_col; i++) { + sums[i] = coopmat(0.0f); + } +#else + ACC_TYPE sums[WMITER * TM * WNITER * TN]; FLOAT_TYPE cache_a[WMITER * TM]; FLOAT_TYPE cache_b[WNITER * TN]; [[unroll]] for (uint i = 0; i < WMITER*TM*WNITER*TN; i++) { - sums[i] = 0.0f; + sums[i] = ACC_TYPE(0.0f); } +#endif - [[unroll]] for (uint block = start_k; block < end_k; block += BK) { + for (uint block = start_k; block < end_k; block += BK) { [[unroll]] for (uint l = 0; l < BM; l += loadstride_a) { #if defined(DATA_A_F32) || defined(DATA_A_F16) #if LOAD_VEC_A == 8 const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; buf_a[buf_idx ] = FLOAT_TYPE(data_a[idx][0].x); buf_a[buf_idx + 1] = FLOAT_TYPE(data_a[idx][0].y); buf_a[buf_idx + 2] = FLOAT_TYPE(data_a[idx][0].z); @@ -177,21 +226,21 @@ void main() { buf_a[buf_idx + 7] = FLOAT_TYPE(data_a[idx][1].w); #elif LOAD_VEC_A == 4 const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; buf_a[buf_idx ] = FLOAT_TYPE(data_a[idx].x); buf_a[buf_idx + 1] = FLOAT_TYPE(data_a[idx].y); buf_a[buf_idx + 2] = FLOAT_TYPE(data_a[idx].z); buf_a[buf_idx + 3] = FLOAT_TYPE(data_a[idx].w); #else if (ir * BM + loadc_a + l < p.M && block + loadr_a < end_k) { - buf_a[(loadc_a + l) * (BK+1) + loadr_a] = FLOAT_TYPE(data_a[pos_a + (loadc_a + l) * p.stride_a + loadr_a]); + buf_a[(loadc_a + l) * SHMEM_STRIDE + loadr_a] = FLOAT_TYPE(data_a[pos_a + (loadc_a + l) * p.stride_a + loadr_a]); } else { - buf_a[(loadc_a + l) * (BK+1) + loadr_a] = FLOAT_TYPE(0.0f); + buf_a[(loadc_a + l) * SHMEM_STRIDE + loadr_a] = FLOAT_TYPE(0.0f); } #endif #elif defined(DATA_A_Q4_0) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a; const uint ib = idx / 16; const uint iqs = idx & 0xF; @@ -204,7 +253,7 @@ void main() { buf_a[buf_idx + 16] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q4_1) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a; const uint ib = idx / 16; const uint iqs = idx & 0xF; @@ -218,7 +267,7 @@ void main() { buf_a[buf_idx + 16] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q5_0) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a; const uint ib = idx / 16; const uint iqs = idx & 0xF; @@ -233,7 +282,7 @@ void main() { buf_a[buf_idx + 16] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q5_1) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a; const uint ib = idx / 16; const uint iqs = idx & 0xF; @@ -249,7 +298,7 @@ void main() { buf_a[buf_idx + 16] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q8_0) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 16; const uint iqs = (idx & 0xF) * 2; @@ -261,7 +310,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q2_K) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 128; // 2 values per idx const uint iqs = idx % 128; // 0..127 @@ -280,7 +329,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(v.y); #elif defined(DATA_A_Q3_K) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 128; // 2 values per idx const uint iqs = idx % 128; // 0..127 @@ -304,7 +353,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(dl * float(int8_t((data_a[ib].qs[qsi + 1] >> qsshift) & 3) - (((data_a[ib].hmask[hmi + 1] & m) != 0) ? 0 : 4))); #elif defined(DATA_A_Q4_K) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 128; // 2 values per idx const uint iqs = idx % 128; // 0..127 @@ -316,15 +365,20 @@ void main() { const vec2 loadd = vec2(data_a[ib].d); - uint8_t sc; - uint8_t mbyte; - if (is < 4) { - sc = uint8_t(data_a[ib].scales[is ] & 63); - mbyte = uint8_t(data_a[ib].scales[is + 4] & 63); - } else { - sc = uint8_t((data_a[ib].scales[is + 4] & 0xF) | ((data_a[ib].scales[is - 4] >> 6) << 4)); - mbyte = uint8_t((data_a[ib].scales[is + 4] >> 4) | ((data_a[ib].scales[is ] >> 6) << 4)); - } + const uint scidx0 = (is < 4) ? is : (is + 4); + const uint scidx1 = (is < 4) ? is : (is - 4); + const uint scidxmask1 = (is < 4) ? 0x30 : 0xC0; + const uint scidxshift1 = (is < 4) ? 0 : 2; + const uint mbidx0 = is + 4; + const uint mbidx1 = (is < 4) ? is + 4 : is; + const uint mbidxmask0 = (is < 4) ? 0xF : 0xF0; + const uint mbidxshift0 = (is < 4) ? 0 : 4; + const uint mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + const uint mbidxshift1 = (is < 4) ? 0 : 2; + + const uint8_t sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + const uint8_t mbyte = uint8_t((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0 | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + const float d = loadd.x * sc; const float m = -loadd.y * mbyte; @@ -332,7 +386,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(fma(d, float((data_a[ib].qs[qsi + 1] >> (b * 4)) & 0xF), m)); #elif defined(DATA_A_Q5_K) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 128; // 2 values per idx const uint iqs = idx % 128; // 0..127 @@ -347,15 +401,20 @@ void main() { const vec2 loadd = vec2(data_a[ib].d); - uint8_t sc; - uint8_t mbyte; - if (is < 4) { - sc = uint8_t(data_a[ib].scales[is ] & 63); - mbyte = uint8_t(data_a[ib].scales[is + 4] & 63); - } else { - sc = uint8_t((data_a[ib].scales[is + 4] & 0xF) | ((data_a[ib].scales[is - 4] >> 6) << 4)); - mbyte = uint8_t((data_a[ib].scales[is + 4] >> 4) | ((data_a[ib].scales[is ] >> 6) << 4)); - } + const uint scidx0 = (is < 4) ? is : (is + 4); + const uint scidx1 = (is < 4) ? is : (is - 4); + const uint scidxmask1 = (is < 4) ? 0x30 : 0xC0; + const uint scidxshift1 = (is < 4) ? 0 : 2; + const uint mbidx0 = is + 4; + const uint mbidx1 = (is < 4) ? is + 4 : is; + const uint mbidxmask0 = (is < 4) ? 0xF : 0xF0; + const uint mbidxshift0 = (is < 4) ? 0 : 4; + const uint mbidxmask1 = (is < 4) ? 0x30 : 0xC0; + const uint mbidxshift1 = (is < 4) ? 0 : 2; + + const uint8_t sc = uint8_t((data_a[ib].scales[scidx0] & 0xF) | ((data_a[ib].scales[scidx1] & scidxmask1) >> scidxshift1)); + const uint8_t mbyte = uint8_t(((data_a[ib].scales[mbidx0] & mbidxmask0) >> mbidxshift0) | ((data_a[ib].scales[mbidx1] & mbidxmask1) >> mbidxshift1)); + const float d = loadd.x * sc; const float m = -loadd.y * mbyte; @@ -363,7 +422,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(fma(d, float((data_a[ib].qs[qsi + 1] >> (b * 4)) & 0xF) + float((data_a[ib].qh[qhi + 1] & hm) != 0 ? 16 : 0), m)); #elif defined(DATA_A_Q6_K) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a * LOAD_VEC_A; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a * LOAD_VEC_A; const uint ib = idx / 128; // 2 values per idx const uint iqs = idx % 128; // 0..127 @@ -382,7 +441,7 @@ void main() { buf_a[buf_idx + 1] = FLOAT_TYPE(dscale * float(int8_t(((data_a[ib].ql[qsi + 1] >> (b * 4)) & 0xF) | (((data_a[ib].qh[qhi + 1] >> qhshift) & 3) << 4)) - 32)); #elif defined(DATA_A_IQ4_NL) const uint idx = pos_a + (loadc_a + l) * p.stride_a / LOAD_VEC_A + loadr_a; - const uint buf_idx = (loadc_a + l) * (BK+1) + loadr_a; + const uint buf_idx = (loadc_a + l) * SHMEM_STRIDE + loadr_a; const uint ib = idx / 16; const uint iqs = idx & 0xF; @@ -403,7 +462,7 @@ void main() { #else const uint idx = pos_b + (loadc_b + l) * p.stride_b / LOAD_VEC_B + loadr_b; #endif - const uint buf_idx = (loadc_b + l) * (BK+1) + loadr_b * LOAD_VEC_B; + const uint buf_idx = (loadc_b + l) * SHMEM_STRIDE + loadr_b * LOAD_VEC_B; buf_b[buf_idx + 0] = FLOAT_TYPE(data_b[idx][0].x); buf_b[buf_idx + 1] = FLOAT_TYPE(data_b[idx][0].y); buf_b[buf_idx + 2] = FLOAT_TYPE(data_b[idx][0].z); @@ -419,24 +478,24 @@ void main() { #else const uint idx = pos_b + (loadc_b + l) * p.stride_b / LOAD_VEC_B + loadr_b; #endif - const uint buf_idx = (loadc_b + l) * (BK+1) + loadr_b * LOAD_VEC_B; + const uint buf_idx = (loadc_b + l) * SHMEM_STRIDE + loadr_b * LOAD_VEC_B; buf_b[buf_idx + 0] = FLOAT_TYPE(data_b[idx].x); buf_b[buf_idx + 1] = FLOAT_TYPE(data_b[idx].y); buf_b[buf_idx + 2] = FLOAT_TYPE(data_b[idx].z); buf_b[buf_idx + 3] = FLOAT_TYPE(data_b[idx].w); #elif !MUL_MAT_ID if (ic * BN + loadc_b + l < p.N && block + loadr_b < end_k) { - buf_b[(loadc_b + l) * (BK+1) + loadr_b] = FLOAT_TYPE(data_b[pos_b + (loadc_b + l) * p.stride_b + loadr_b]); + buf_b[(loadc_b + l) * SHMEM_STRIDE + loadr_b] = FLOAT_TYPE(data_b[pos_b + (loadc_b + l) * p.stride_b + loadr_b]); } else { - buf_b[(loadc_b + l) * (BK+1) + loadr_b] = FLOAT_TYPE(0.0f); + buf_b[(loadc_b + l) * SHMEM_STRIDE + loadr_b] = FLOAT_TYPE(0.0f); } #else const uint row_i = ic * BN + loadc_b + l; if (row_i < _ne1) { const u16vec2 row_idx = row_ids[row_i]; - buf_b[(loadc_b + l) * (BK+1) + loadr_b] = FLOAT_TYPE(data_b[pos_b + row_idx.y * p.batch_stride_b + (row_idx.x % p.ne11) * p.stride_b + loadr_b]); + buf_b[(loadc_b + l) * SHMEM_STRIDE + loadr_b] = FLOAT_TYPE(data_b[pos_b + row_idx.y * p.batch_stride_b + (row_idx.x % p.ne11) * p.stride_b + loadr_b]); } else { - buf_b[(loadc_b + l) * (BK+1) + loadr_b] = FLOAT_TYPE(0.0f); + buf_b[(loadc_b + l) * SHMEM_STRIDE + loadr_b] = FLOAT_TYPE(0.0f); } #endif } @@ -446,16 +505,30 @@ void main() { pos_a += BK / LOAD_VEC_A; pos_b += BK / LOAD_VEC_B; - for (uint i = 0; i < BK; i++) { +#ifdef COOPMAT + [[unroll]] for (uint i = 0; i < BK; i += TK) { + [[unroll]] for (uint cm_row = 0; cm_row < cms_per_row; cm_row++) { + // Load from shared into cache + coopMatLoad(cache_a, buf_a, (warp_r * WM + cm_row * TM) * SHMEM_STRIDE + i, SHMEM_STRIDE, gl_CooperativeMatrixLayoutRowMajor); + + [[unroll]] for (uint cm_col = 0; cm_col < cms_per_col; cm_col++) { + coopMatLoad(cache_b, buf_b, (warp_c * WN + cm_col * TN) * SHMEM_STRIDE + i, SHMEM_STRIDE, gl_CooperativeMatrixLayoutColumnMajor); + + sums[cm_col * cms_per_row + cm_row] = coopMatMulAdd(cache_a, cache_b, sums[cm_col * cms_per_row + cm_row]); + } + } + } +#else + [[unroll]] for (uint i = 0; i < BK; i++) { // Load from shared into cache [[unroll]] for (uint wsir = 0; wsir < WMITER; wsir++) { [[unroll]] for (uint j = 0; j < TM; j++) { - cache_a[wsir * TM + j] = buf_a[(warp_r * WM + wsir * WSUBM + tiwr * TM + j) * (BK+1) + i]; + cache_a[wsir * TM + j] = buf_a[(warp_r * WM + wsir * WSUBM + tiwr * TM + j) * SHMEM_STRIDE + i]; } } [[unroll]] for (uint wsic = 0; wsic < WNITER; wsic++) { [[unroll]] for (uint j = 0; j < TN; j++) { - cache_b[wsic * TN + j] = buf_b[(warp_c * WN + wsic * WSUBN + tiwc * TN + j) * (BK+1) + i]; + cache_b[wsic * TN + j] = buf_b[(warp_c * WN + wsic * WSUBN + tiwc * TN + j) * SHMEM_STRIDE + i]; } } @@ -464,12 +537,13 @@ void main() { [[unroll]] for (uint cc = 0; cc < TN; cc++) { [[unroll]] for (uint cr = 0; cr < TM; cr++) { const uint sums_idx = (wsic * TN + cc) * (WMITER * TM) + wsir * TM + cr; - sums[sums_idx] = fma(float(cache_a[wsir * TM + cr]), float(cache_b[wsic * TN + cc]), sums[sums_idx]); + sums[sums_idx] = fma(ACC_TYPE(cache_a[wsir * TM + cr]), ACC_TYPE(cache_b[wsic * TN + cc]), sums[sums_idx]); } } } } } +#endif barrier(); } @@ -481,6 +555,54 @@ void main() { const uint offsets = batch_idx * p.batch_stride_d + ik * p.batch_stride_d * gl_NumWorkGroups.z; #endif +#ifdef COOPMAT +#ifdef MUL_MAT_ID + [[unroll]] for (uint cm_row = 0; cm_row < cms_per_row; cm_row++) { + [[unroll]] for (uint cm_col = 0; cm_col < cms_per_col; cm_col++) { + coopMatStore(sums[cm_col * cms_per_row + cm_row], coopmat_stage, warp_i * TM * TN, TM, gl_CooperativeMatrixLayoutColumnMajor); + + [[unroll]] for (uint col = 0; col < BN; col += storestride) { + const uint row_i = dc + cm_col * TN + col + store_c; + if (row_i >= _ne1) break; + + const u16vec2 row_idx = row_ids[row_i]; + + data_d[row_idx.y * p.batch_stride_d + row_idx.x * p.stride_d + dr + cm_row * TM + store_r] = D_TYPE(coopmat_stage[warp_i * TM * TN + (col + store_c) * TM + store_r]); + } + } + } +#else + const bool is_aligned = p.stride_d % 4 == 0; // Assumption: D_TYPE == float + + [[unroll]] for (uint cm_row = 0; cm_row < cms_per_row; cm_row++) { + [[unroll]] for (uint cm_col = 0; cm_col < cms_per_col; cm_col++) { + const bool is_in_bounds = dr + (cm_row + 1) * TM <= p.M && dc + (cm_col + 1) * TN <= p.N; + + if (is_aligned && is_in_bounds) { + // Full coopMat is within bounds and stride_d is aligned with 16B + coopmat cm_dtype = coopmat(sums[cm_col * cms_per_row + cm_row]); + coopMatStore(cm_dtype, data_d, offsets + (dc + cm_col * TN) * p.stride_d + dr + cm_row * TM, p.stride_d, gl_CooperativeMatrixLayoutColumnMajor); + } else if (is_in_bounds) { + // Full coopMat is within bounds, but stride_d is not aligned + coopMatStore(sums[cm_col * cms_per_row + cm_row], coopmat_stage, warp_i * TM * TN, TM, gl_CooperativeMatrixLayoutColumnMajor); + + [[unroll]] for (uint col = 0; col < TN; col += storestride) { + data_d[offsets + (dc + cm_col * TN + col + store_c) * p.stride_d + dr + cm_row * TM + store_r] = D_TYPE(coopmat_stage[warp_i * TM * TN + (col + store_c) * TM + store_r]); + } + } else if (dr + cm_row * TM < p.M && dc + cm_col * TN < p.N) { + // Partial coopMat is within bounds + coopMatStore(sums[cm_col * cms_per_row + cm_row], coopmat_stage, warp_i * TM * TN, TM, gl_CooperativeMatrixLayoutColumnMajor); + + [[unroll]] for (uint col = 0; col < TN; col += storestride) { + if (dr + cm_row * TM + store_r < p.M && dc + cm_col * TN + col + store_c < p.N) { + data_d[offsets + (dc + cm_col * TN + col + store_c) * p.stride_d + dr + cm_row * TM + store_r] = D_TYPE(coopmat_stage[warp_i * TM * TN + (col + store_c) * TM + store_r]); + } + } + } + } + } +#endif // MUL_MAT_ID +#else [[unroll]] for (uint wsic = 0; wsic < WNITER; wsic++) { [[unroll]] for (uint wsir = 0; wsir < WMITER; wsir++) { @@ -492,7 +614,7 @@ void main() { if (row_i >= _ne1) break; const u16vec2 row_idx = row_ids[row_i]; -#endif +#endif // MUL_MAT_ID [[unroll]] for (uint cr = 0; cr < TM; cr++) { #ifdef MUL_MAT_ID data_d[row_idx.y * p.batch_stride_d + row_idx.x * p.stride_d + dr_warp + cr] = D_TYPE(sums[(wsic * TN + cc) * (WMITER * TM) + wsir * TM + cr]); @@ -500,9 +622,10 @@ void main() { if (dr_warp + cr < p.M && dc_warp + cc < p.N) { data_d[offsets + (dc_warp + cc) * p.stride_d + dr_warp + cr] = D_TYPE(sums[(wsic * TN + cc) * (WMITER * TM) + wsir * TM + cr]); } -#endif +#endif // MUL_MAT_ID } } } } +#endif // COOPMAT } diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp new file mode 100644 index 000000000..cbfa5dce1 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp @@ -0,0 +1,328 @@ +#version 450 + +#extension GL_EXT_control_flow_attributes : enable +#extension GL_EXT_shader_16bit_storage : require + +#extension GL_EXT_shader_explicit_arithmetic_types_float16 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require + +#extension GL_KHR_memory_scope_semantics : enable +#extension GL_KHR_cooperative_matrix : enable +#extension GL_NV_cooperative_matrix2 : enable +#extension GL_EXT_buffer_reference : enable +#extension GL_KHR_shader_subgroup_ballot : enable +#extension GL_KHR_shader_subgroup_vote : enable + +#include "types.comp" + +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +layout (constant_id = 1) const uint BM = 64; +layout (constant_id = 2) const uint BN = 64; +layout (constant_id = 3) const uint BK = 16; // Assumed to be 32 if working with a quant + +layout (push_constant) uniform parameter +{ + uint M; + uint N; + uint K; + uint stride_a; + uint stride_b; + uint stride_d; + + uint batch_stride_a; + uint batch_stride_b; + uint batch_stride_d; + +#ifdef MUL_MAT_ID + uint nei0; + uint nei1; + uint nbi1; + uint ne11; +#else + uint k_split; + uint ne02; + uint ne12; + uint broadcast2; + uint broadcast3; +#endif +} p; + + +layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; +layout (binding = 1) readonly buffer B {B_TYPE data_b[];}; +layout (binding = 2) writeonly buffer D {D_TYPE data_d[];}; + +#if QUANT_K > 1 +#define DECODEFUNCA , dequantFuncA +#define MAT_A_TYPE float16_t + +#include "dequant_funcs_cm2.comp" + +#else +#define DECODEFUNCA +#define MAT_A_TYPE A_TYPE +#endif + +#define MAT_B_TYPE B_TYPE + +#ifdef MUL_MAT_ID +layout (binding = 3) readonly buffer IDS {int data_ids[];}; + +shared u16vec4 row_ids[3072]; + +layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufB { + B_TYPE b[]; +}; + +uint _ne1; +shared uint _ne1_sh; + +B_TYPE decodeFuncB(const in decodeBufB bl, const in uint blockCoords[2], const in uint coordInBlock[2]) +{ + const uint row_i = blockCoords[0]; + + if (row_i >= _ne1) { + return B_TYPE(0.0); + } + + const u16vec4 row_idx = row_ids[row_i]; + B_TYPE ret = data_b[row_idx.y * p.batch_stride_b + row_idx.x * p.stride_b + blockCoords[1]]; + + return ret; +} + +D_TYPE perElemOpD(const in uint32_t r, const in uint32_t c, const in D_TYPE elem, const in uint32_t ir, const in uint32_t ic) +{ + uint dr = ir * BM + r; + uint dc = ic * BN + c; + + if (dr < p.M && dc < _ne1) { + uint row_i = dc; + const u16vec4 row_idx = row_ids[row_i]; + data_d[row_idx.y * p.batch_stride_d + row_idx.z * p.stride_d + dr] = elem; + } + return elem; +} + +#endif + +void main() { +#if defined(DATA_A_IQ4_NL) + init_iq4nl_shmem(); +#endif + +#ifdef MUL_MAT_ID + const uint expert_idx = gl_GlobalInvocationID.z; +#else + const uint batch_idx = gl_GlobalInvocationID.z; + + const uint i13 = batch_idx / p.ne12; + const uint i12 = batch_idx % p.ne12; + + const uint i03 = i13 / p.broadcast3; + const uint i02 = i12 / p.broadcast2; + + const uint batch_idx_a = i03 * p.ne02 + i02; +#endif + + const uint blocks_m = (p.M + BM - 1) / BM; + const uint ir = gl_WorkGroupID.x % blocks_m; + const uint ik = gl_WorkGroupID.x / blocks_m; + const uint ic = gl_WorkGroupID.y; + +#ifdef MUL_MAT_ID + // Spread the search across all elements in the first subgroup + if (gl_SubgroupID == 0) { + _ne1 = 0; + uint num_elements = p.nei1 * p.nei0; + + for (uint i = gl_SubgroupInvocationID; subgroupAny(i < num_elements); i += gl_SubgroupSize) { + bool in_range = i < num_elements; + uint ii0 = i % p.nei0; + uint ii1 = i / p.nei0; + uint id = in_range ? data_ids[ii1*p.nbi1 + ii0] : 0; + uvec4 ballot = subgroupBallot(in_range && id == expert_idx); + uint idx = subgroupBallotExclusiveBitCount(ballot); + if (in_range && id == expert_idx) { + row_ids[_ne1 + idx] = u16vec4(ii0 % p.ne11, ii1, ii0, 0); + } + _ne1 += subgroupBallotBitCount(ballot); + } + _ne1_sh = _ne1; + } + + barrier(); + + _ne1 = _ne1_sh; + + // Workgroup has no work + if (ic * BN >= _ne1) return; +#endif + +#ifdef MUL_MAT_ID + uint start_k = 0; + const uint end_k = p.K; +#else + uint start_k = ik * p.k_split; + const uint end_k = min(p.K, (ik + 1) * p.k_split); +#endif + + coopmat sum; + sum = coopmat(0.0); + +#ifdef MUL_MAT_ID + uint pos_a = (expert_idx * p.batch_stride_a) / QUANT_K; + uint pos_b = 0; +#else + uint pos_a = (batch_idx_a * p.batch_stride_a) / QUANT_K; + uint pos_b = batch_idx * p.batch_stride_b; +#endif + + uint stride_a = p.stride_a / QUANT_K; + uint stride_b = p.stride_b; + + // Hint to the compiler that values are aligned (want 16B alignment). + // Quants are always block-aligned, no alignment needed. +#if ALIGNED +#if QUANT_K == 1 + stride_a &= ~7; +#endif + stride_b &= ~7; +#endif + + // Create layouts for both clamped and unclamped accesses + tensorLayoutNV<2> tensorLayoutA = createTensorLayoutNV(2); + tensorLayoutNV<2, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutAClamp = createTensorLayoutNV(2, gl_CooperativeMatrixClampModeConstantNV); + tensorLayoutNV<2> tensorLayoutB = createTensorLayoutNV(2); + tensorLayoutNV<2, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutBClamp = createTensorLayoutNV(2, gl_CooperativeMatrixClampModeConstantNV); + tensorLayoutNV<2, gl_CooperativeMatrixClampModeConstantNV> tensorLayoutD = createTensorLayoutNV(2, gl_CooperativeMatrixClampModeConstantNV); + +#if QUANT_K > 1 + tensorLayoutA = setTensorLayoutBlockSizeNV(tensorLayoutA, 1, QUANT_K); + tensorLayoutAClamp = setTensorLayoutBlockSizeNV(tensorLayoutAClamp, 1, QUANT_K); +#endif + + // Use end_k rather than p.K as the dimension because that's what + // we need to bound check against when using split_k + tensorLayoutA = setTensorLayoutDimensionNV(tensorLayoutA, p.M, end_k); + tensorLayoutB = setTensorLayoutDimensionNV(tensorLayoutB, p.N, end_k); + tensorLayoutD = setTensorLayoutDimensionNV(tensorLayoutD, p.N, p.M); + tensorLayoutAClamp = setTensorLayoutDimensionNV(tensorLayoutAClamp, p.M, end_k); + tensorLayoutBClamp = setTensorLayoutDimensionNV(tensorLayoutBClamp, p.N, end_k); + + tensorViewNV<2, false, 1, 0> tensorViewTranspose = createTensorViewNV(2, false, 1, 0); + +#if !defined(MUL_MAT_ID) + // Detect a fast path where all loads are entirely in bounds and no clamping is required + if ((ir + 1) * BM <= p.M && (ic + 1) * BN <= p.N && (start_k % BK) == 0 && (end_k % BK) == 0 && +#if QUANT_K == 1 + (stride_a % 8) == 0 && +#endif + (stride_b % 8) == 0 && (start_k % 8) == 0) { + // Hint to the compiler that values are aligned (want 16B alignment) + start_k &= ~7; + stride_b &= ~7; +#if QUANT_K == 1 + stride_a &= ~7; +#endif + + tensorLayoutA = setTensorLayoutStrideNV(tensorLayoutA, stride_a, 1); + tensorLayoutB = setTensorLayoutStrideNV(tensorLayoutB, stride_b, 1); + + uint k_iters = (end_k - start_k + BK - 1) / BK; + + for (uint block_k = start_k, i = 0; i < k_iters; block_k += BK, ++i) { + + coopmat mat_a; + coopmat mat_b; + + coopMatLoadTensorNV(mat_a, data_a, pos_a, sliceTensorLayoutNV(tensorLayoutA, ir * BM, BM, block_k, BK) DECODEFUNCA); + coopmat mat_a_ft = coopmat(mat_a); + + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutB, ic * BN, BN, block_k, BK), tensorViewTranspose); + coopmat mat_b_ft = coopmat(mat_b); + + sum = coopMatMulAdd(mat_a_ft, mat_b_ft, sum); + } + } else +#endif // !defined(MUL_MAT_ID) + { + tensorLayoutA = setTensorLayoutStrideNV(tensorLayoutA, stride_a, 1); + + tensorLayoutAClamp = setTensorLayoutStrideNV(tensorLayoutAClamp, stride_a, 1); + + tensorLayoutB = setTensorLayoutStrideNV(tensorLayoutB, stride_b, 1); + + tensorLayoutBClamp = setTensorLayoutStrideNV(tensorLayoutBClamp, stride_b, 1); + + [[dont_unroll]] + for (uint block_k = start_k; block_k < end_k; block_k += BK) { + + coopmat mat_a; + coopmat mat_b; + coopmat mat_a_ft; + coopmat mat_b_ft; + + // Clamping is expensive, so detect different code paths for each combination + // of A and B needing clamping. + bool unclampedA = (ir + 1) * BM <= p.M && block_k + BK <= end_k && (block_k % 8) == 0; +#ifdef MUL_MAT_ID + bool unclampedB = true; +#else + bool unclampedB = (ic + 1) * BN <= p.N && block_k + BK <= end_k && (block_k % 8) == 0; +#endif + if (unclampedA && unclampedB) { + coopMatLoadTensorNV(mat_a, data_a, pos_a, sliceTensorLayoutNV(tensorLayoutA, ir * BM, BM, (block_k & ~7), BK) DECODEFUNCA); +#ifdef MUL_MAT_ID + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutB, ic * BN, BN, block_k, BK), tensorViewTranspose, decodeFuncB); +#else + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutB, ic * BN, BN, (block_k & ~7), BK), tensorViewTranspose); +#endif + mat_a_ft = coopmat(mat_a); + mat_b_ft = coopmat(mat_b); + sum = coopMatMulAdd(mat_a_ft, mat_b_ft, sum); + } else if (unclampedA && !unclampedB) { + coopMatLoadTensorNV(mat_a, data_a, pos_a, sliceTensorLayoutNV(tensorLayoutA, ir * BM, BM, (block_k & ~7), BK) DECODEFUNCA); + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutBClamp, ic * BN, BN, block_k, BK), tensorViewTranspose); + + mat_a_ft = coopmat(mat_a); + mat_b_ft = coopmat(mat_b); + sum = coopMatMulAdd(mat_a_ft, mat_b_ft, sum); + } else if (!unclampedA && unclampedB) { + coopMatLoadTensorNV(mat_a, data_a, pos_a, sliceTensorLayoutNV(tensorLayoutAClamp, ir * BM, BM, block_k, BK) DECODEFUNCA); +#ifdef MUL_MAT_ID + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutB, ic * BN, BN, block_k, BK), tensorViewTranspose, decodeFuncB); +#else + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutB, ic * BN, BN, (block_k & ~7), BK), tensorViewTranspose); +#endif + mat_a_ft = coopmat(mat_a); + mat_b_ft = coopmat(mat_b); + sum = coopMatMulAdd(mat_a_ft, mat_b_ft, sum); + } else if (!unclampedA && !unclampedB) { + coopMatLoadTensorNV(mat_a, data_a, pos_a, sliceTensorLayoutNV(tensorLayoutAClamp, ir * BM, BM, block_k, BK) DECODEFUNCA); + coopMatLoadTensorNV(mat_b, data_b, pos_b, sliceTensorLayoutNV(tensorLayoutBClamp, ic * BN, BN, block_k, BK), tensorViewTranspose); + + mat_a_ft = coopmat(mat_a); + mat_b_ft = coopmat(mat_b); + sum = coopMatMulAdd(mat_a_ft, mat_b_ft, sum); + } + } + } + + // Convert from ACC_TYPE to D_TYPE + coopmat mat_d; + mat_d = coopmat(sum); + +#ifdef MUL_MAT_ID + // Call callback to store each element, remapping row through shared memory + coopMatPerElementNV(mat_d, mat_d, perElemOpD, ir, ic); +#else + tensorLayoutD = setTensorLayoutStrideNV(tensorLayoutD, p.stride_d, 1); + + uint pos_d = batch_idx * p.batch_stride_d + ik * p.batch_stride_d * gl_NumWorkGroups.z; + coopMatStoreTensorNV(mat_d, data_d, pos_d, sliceTensorLayoutNV(tensorLayoutD, ic * BN, BN, ir * BM, BM), tensorViewTranspose); +#endif +} diff --git a/ggml/src/vulkan-shaders/norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp similarity index 100% rename from ggml/src/vulkan-shaders/norm.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/norm.comp diff --git a/ggml/src/vulkan-shaders/pad.comp b/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp similarity index 83% rename from ggml/src/vulkan-shaders/pad.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/pad.comp index a465cd52b..450b67fc5 100644 --- a/ggml/src/vulkan-shaders/pad.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp @@ -3,6 +3,8 @@ #include "types.comp" #include "generic_unary_head.comp" +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + void main() { const uint idx = gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x; @@ -22,5 +24,5 @@ void main() { const bool is_src0 = i0 < p.ne00 && i1 < p.ne01 && i2 < p.ne02 && i3 < p.ne03; - data_d[p.d_offset + dst_idx] = D_TYPE(is_src0 ? data_a[src0_idx] : 0.0f); + data_d[get_doffset() + dst_idx] = D_TYPE(is_src0 ? data_a[get_aoffset() + src0_idx] : 0.0f); } diff --git a/ggml/src/vulkan-shaders/pool2d.comp b/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp similarity index 100% rename from ggml/src/vulkan-shaders/pool2d.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp diff --git a/ggml/src/vulkan-shaders/relu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp similarity index 100% rename from ggml/src/vulkan-shaders/relu.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/relu.comp diff --git a/ggml/src/vulkan-shaders/repeat.comp b/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp similarity index 79% rename from ggml/src/vulkan-shaders/repeat.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp index a86af87e7..1568b141d 100644 --- a/ggml/src/vulkan-shaders/repeat.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp @@ -3,6 +3,8 @@ #include "types.comp" #include "generic_unary_head.comp" +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + uint src0_idx_mod(uint idx) { const uint i13 = idx / (p.ne12*p.ne11*p.ne10); const uint i13_offset = i13 * p.ne12*p.ne11*p.ne10; @@ -20,5 +22,5 @@ void main() { return; } - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(data_a[src0_idx_mod(idx)]); + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(data_a[get_aoffset() + src0_idx_mod(idx)]); } diff --git a/ggml/src/vulkan-shaders/rms_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp similarity index 100% rename from ggml/src/vulkan-shaders/rms_norm.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp diff --git a/ggml/src/vulkan-shaders/rope_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp similarity index 91% rename from ggml/src/vulkan-shaders/rope_head.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp index ea8954226..574b51ca5 100644 --- a/ggml/src/vulkan-shaders/rope_head.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp @@ -1,6 +1,11 @@ #include "types.comp" #extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_spirv_intrinsics: enable + +#if RTE16 +spirv_execution_mode(capabilities = [4467], 4462, 16); // RoundingModeRTE, 16 bits +#endif layout(local_size_x = 1, local_size_y = 256, local_size_z = 1) in; diff --git a/ggml/src/vulkan-shaders/rope_neox.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp similarity index 100% rename from ggml/src/vulkan-shaders/rope_neox.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp diff --git a/ggml/src/vulkan-shaders/rope_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp similarity index 100% rename from ggml/src/vulkan-shaders/rope_norm.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp b/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp new file mode 100644 index 000000000..4663428de --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp @@ -0,0 +1,24 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +const uint num_threads = 128; + +layout(local_size_x = num_threads, local_size_y = 1, local_size_z = 1) in; + +void main() { + uint idx = get_idx(); + + // num_threads * num_iter must equal 512, to match the wg_denoms and get_idx calculation + const uint num_iter = 4; + + [[unroll]] for (uint i = 0; i < num_iter; ++i) { + if (idx >= p.ne) { + continue; + } + + data_d[get_doffset() + idx] = D_TYPE(FLOAT_TYPE(data_a[get_aoffset() + idx]) * FLOAT_TYPE(p.param1)); + idx += num_threads; + } +} diff --git a/ggml/src/vulkan-shaders/silu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp similarity index 100% rename from ggml/src/vulkan-shaders/silu.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/silu.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp new file mode 100644 index 000000000..d7c15a169 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp @@ -0,0 +1,17 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = get_idx(); + + if (idx >= p.ne) { + return; + } + + const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]); + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(sin(val)); +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp new file mode 100644 index 000000000..51fc2dc7e --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp @@ -0,0 +1,173 @@ +#version 450 + +#extension GL_EXT_control_flow_attributes : enable + +layout (push_constant) uniform parameter +{ + uint KX; + uint KY; + float scale; + float max_bias; + float m0; + float m1; + uint n_head_log2; + uint nrows_x; +} p; + +#include "types.comp" + +layout(constant_id = 0) const uint BLOCK_SIZE = 32; +layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; + +layout (binding = 0) readonly buffer X {A_TYPE data_a[];}; +layout (binding = 1) readonly buffer Y {B_TYPE data_b[];}; +layout (binding = 2) buffer D {D_TYPE data_d[];}; + +shared FLOAT_TYPE vals[BLOCK_SIZE]; + +// num_iters is the number of BLOCK_SIZE loop iterations we need to iterate +// over all the columns. The main function tries to pass a constant here, +// as if it were a template function, to allow unrolling. +void soft_max(uint num_iters) { + const uint tid = gl_LocalInvocationID.x; + const uint rowx = gl_WorkGroupID.z * 262144 + gl_WorkGroupID.y * 512 + gl_WorkGroupID.x; + const uint rowy = (p.KY > 0) ? (rowx % p.KY) : 0; + + if (rowx >= p.nrows_x) { + return; + } + + float slope = 1.0f; + + // ALiBi + if (p.max_bias > 0.0f) { + const uint h = rowx/p.KY; // head index + + const float base = h < p.n_head_log2 ? p.m0 : p.m1; + const uint exp = h < p.n_head_log2 ? h + 1 : 2*(h - p.n_head_log2) + 1; + + slope = pow(base, exp); + } + + // Find max + FLOAT_TYPE max_val = uintBitsToFloat(0xFF800000); + + // Cache values while we compute the max, so we don't need to read them + // again when we're ready to compute exp(x-max). + const uint DATA_CACHE_SIZE = 16; + FLOAT_TYPE data_cache[DATA_CACHE_SIZE]; + + [[unroll]] for (uint col0 = 0, idx = 0; idx < num_iters; col0 += BLOCK_SIZE, ++idx) { + const uint col = col0 + tid; + + FLOAT_TYPE a = FLOAT_TYPE(0); + if (col < p.KX) { + a = data_a[rowx * p.KX + col]; + } + + FLOAT_TYPE b = FLOAT_TYPE(0); + if (p.KY > 0 && col < p.KX) { + b = data_b[rowy * p.KX + col]; + } + + FLOAT_TYPE v = a * p.scale + slope * b; + + if (col < p.KX) { + max_val = max(max_val, v); + } + + if (idx < DATA_CACHE_SIZE) { + data_cache[idx] = v; + } + } + + // reduce across the workgroup + vals[tid] = max_val; + barrier(); + [[unroll]] for (uint s = BLOCK_SIZE / 2; s > 0; s >>= 1) { + if (tid < s) { + vals[tid] = max(vals[tid], vals[tid + s]); + } + barrier(); + } + + max_val = vals[0]; + barrier(); + + FLOAT_TYPE sum = FLOAT_TYPE(0.0f); + + // Compute sum{exp(x - max)} + [[unroll]] for (uint col0 = 0, idx = 0; idx < num_iters; col0 += BLOCK_SIZE, ++idx) { + const uint col = col0 + tid; + + if (col >= p.KX) { + break; + } + + // compute exp(a*scale+b*slope), add it to sum, and cache the new value + // in data_cache if possible. + const uint i = rowx * p.KX + col; + FLOAT_TYPE val; + if (idx < DATA_CACHE_SIZE) { + val = exp(data_cache[idx] - max_val); + } else { + val = exp(FLOAT_TYPE(data_a[i]) * p.scale + (p.KY > 0 ? slope * FLOAT_TYPE(data_b[rowy * p.KX + col]) : FLOAT_TYPE(0.0f)) - max_val); + } + sum += val; + if (idx < DATA_CACHE_SIZE) { + data_cache[idx] = val; + } else { + data_d[i] = D_TYPE(val); + } + } + + // reduce across the workgroup + vals[tid] = sum; + barrier(); + [[unroll]] for (uint s = BLOCK_SIZE / 2; s > 0; s >>= 1) { + if (tid < s) { + vals[tid] += vals[tid + s]; + } + barrier(); + } + sum = vals[0]; + + FLOAT_TYPE rcpdivisor = 1.0/sum; + + [[unroll]] for (uint col0 = 0, idx = 0; idx < num_iters; col0 += BLOCK_SIZE, ++idx) { + const uint col = col0 + tid; + + if (col >= p.KX) { + continue; + } + + if (idx < DATA_CACHE_SIZE) { + data_d[rowx*p.KX + col] = D_TYPE(data_cache[idx] * rcpdivisor); + } else { + data_d[rowx*p.KX + col] *= D_TYPE(rcpdivisor); + } + } +} + +void main() { + // instantiate the soft_max function for several different + // dimensions, to allow loop unrolling + uint num_blocks = (p.KX + BLOCK_SIZE - 1) / BLOCK_SIZE; + if (num_blocks > 32) { + soft_max(num_blocks); + } else if (num_blocks > 16) { + soft_max(32); + } else if (num_blocks > 8) { + soft_max(16); + } else if (num_blocks > 4) { + soft_max(8); + } else if (num_blocks == 4) { + soft_max(4); + } else if (num_blocks == 3) { + soft_max(3); + } else if (num_blocks == 2) { + soft_max(2); + } else if (num_blocks == 1) { + soft_max(1); + } +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/square.comp b/ggml/src/ggml-vulkan/vulkan-shaders/square.comp new file mode 100644 index 000000000..ef43598ba --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/square.comp @@ -0,0 +1,17 @@ +#version 450 + +#include "types.comp" +#include "generic_unary_head.comp" + +layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; + +void main() { + const uint idx = get_idx(); + + if (idx >= p.ne) { + return; + } + + const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]); + data_d[get_doffset() + dst_idx(idx)] = D_TYPE(val * val); +} diff --git a/ggml/src/vulkan-shaders/sum_rows.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp similarity index 100% rename from ggml/src/vulkan-shaders/sum_rows.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp diff --git a/ggml/src/vulkan-shaders/tanh.comp b/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp similarity index 88% rename from ggml/src/vulkan-shaders/tanh.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp index 74630dc7f..495f966bd 100644 --- a/ggml/src/vulkan-shaders/tanh.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp @@ -16,6 +16,5 @@ void main() { if (i >= p.KX) { return; } - - data_d[i] = D_TYPE(tanh(data_a[i])); + data_d[i] = D_TYPE(1. - 2. / (exp(2.*data_a[i]) + 1.)); } diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp new file mode 100644 index 000000000..28eb24e11 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp @@ -0,0 +1,7 @@ +#version 460 + +#extension GL_NV_cooperative_matrix2 : require + +void main() +{ +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp new file mode 100644 index 000000000..8c5dd1bd1 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp @@ -0,0 +1,7 @@ +#version 460 + +#extension GL_KHR_cooperative_matrix : require + +void main() +{ +} diff --git a/ggml/src/vulkan-shaders/timestep_embedding.comp b/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp similarity index 100% rename from ggml/src/vulkan-shaders/timestep_embedding.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/types.comp b/ggml/src/ggml-vulkan/vulkan-shaders/types.comp new file mode 100644 index 000000000..f12e61bbe --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/types.comp @@ -0,0 +1,326 @@ + +#if !defined(GGML_TYPES_COMP) +#define GGML_TYPES_COMP + +#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require +#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require +#extension GL_EXT_shader_16bit_storage : require + +#if defined(DATA_A_F32) +#define QUANT_K 1 +#define QUANT_R 1 + +#if !defined(LOAD_VEC_A) || LOAD_VEC_A == 1 +#define A_TYPE float +#elif LOAD_VEC_A == 4 +#define A_TYPE vec4 +#elif LOAD_VEC_A == 8 +#define A_TYPE mat2x4 +#endif +#endif + +#if defined(DATA_A_F16) +#define QUANT_K 1 +#define QUANT_R 1 + +#if !defined(LOAD_VEC_A) || LOAD_VEC_A == 1 +#define A_TYPE float16_t +#elif LOAD_VEC_A == 4 +#define A_TYPE f16vec4 +#elif LOAD_VEC_A == 8 +#define A_TYPE f16mat2x4 +#endif +#endif + +#define QUANT_K_Q4_0 32 +#define QUANT_R_Q4_0 2 + +struct block_q4_0 +{ + float16_t d; + uint8_t qs[16]; +}; +struct block_q4_0_packed16 +{ + float16_t d; + uint16_t qs[16/2]; +}; + +#if defined(DATA_A_Q4_0) +#define QUANT_K QUANT_K_Q4_0 +#define QUANT_R QUANT_R_Q4_0 +#define A_TYPE block_q4_0 +#define A_TYPE_PACKED16 block_q4_0_packed16 +#endif + +#define QUANT_K_Q4_1 32 +#define QUANT_R_Q4_1 2 + +struct block_q4_1 +{ + float16_t d; + float16_t m; + uint8_t qs[16]; +}; + +struct block_q4_1_packed16 +{ + float16_t d; + float16_t m; + uint16_t qs[16/2]; +}; + +#if defined(DATA_A_Q4_1) +#define QUANT_K QUANT_K_Q4_1 +#define QUANT_R QUANT_R_Q4_1 +#define A_TYPE block_q4_1 +#define A_TYPE_PACKED16 block_q4_1_packed16 +#endif + +#define QUANT_K_Q5_0 32 +#define QUANT_R_Q5_0 2 + +struct block_q5_0 +{ + float16_t d; + uint16_t qh[2]; + uint8_t qs[16]; +}; + +struct block_q5_0_packed16 +{ + float16_t d; + uint16_t qh[2]; + uint16_t qs[16/2]; +}; + +#if defined(DATA_A_Q5_0) +#define QUANT_K QUANT_K_Q5_0 +#define QUANT_R QUANT_R_Q5_0 +#define A_TYPE block_q5_0 +#define A_TYPE_PACKED16 block_q5_0_packed16 +#endif + +#define QUANT_K_Q5_1 32 +#define QUANT_R_Q5_1 2 + +struct block_q5_1 +{ + float16_t d; + float16_t m; + uint qh; + uint8_t qs[16]; +}; + +struct block_q5_1_packed16 +{ + float16_t d; + float16_t m; + uint qh; + uint16_t qs[16/2]; +}; + +#if defined(DATA_A_Q5_1) +#define QUANT_K QUANT_K_Q5_1 +#define QUANT_R QUANT_R_Q5_1 +#define A_TYPE block_q5_1 +#define A_TYPE_PACKED16 block_q5_1_packed16 +#endif + +#define QUANT_K_Q8_0 32 +#define QUANT_R_Q8_0 1 + +struct block_q8_0 +{ + float16_t d; + int8_t qs[32]; +}; +struct block_q8_0_packed16 +{ + float16_t d; + uint16_t qs[32/2]; +}; + +#if defined(DATA_A_Q8_0) +#define QUANT_K QUANT_K_Q8_0 +#define QUANT_R QUANT_R_Q8_0 +#define A_TYPE block_q8_0 +#define A_TYPE_PACKED16 block_q8_0_packed16 +#endif + +// K-quants +#define QUANT_K_Q2_K 256 + +struct block_q2_K +{ + uint8_t scales[QUANT_K_Q2_K/16]; + uint8_t qs[QUANT_K_Q2_K/4]; + f16vec2 d; +}; + +struct block_q2_K_packed16 +{ + uint16_t scales[QUANT_K_Q2_K/16/2]; + uint16_t qs[QUANT_K_Q2_K/4/2]; + f16vec2 d; +}; + +struct block_q2_K_packed32 +{ + uint32_t scales[QUANT_K_Q2_K/16/4]; + uint32_t qs[QUANT_K_Q2_K/4/4]; + f16vec2 d; +}; + +#if defined(DATA_A_Q2_K) +#define QUANT_K QUANT_K_Q2_K +#define A_TYPE block_q2_K +#define A_TYPE_PACKED16 block_q2_K_packed16 +#define A_TYPE_PACKED32 block_q2_K_packed32 +#endif + +#define QUANT_K_Q3_K 256 + +struct block_q3_K +{ + uint8_t hmask[QUANT_K_Q3_K/8]; + uint8_t qs[QUANT_K_Q3_K/4]; + uint8_t scales[12]; + float16_t d; +}; + +struct block_q3_K_packed16 +{ + uint16_t hmask[QUANT_K_Q3_K/8/2]; + uint16_t qs[QUANT_K_Q3_K/4/2]; + uint16_t scales[12/2]; + float16_t d; +}; + +#if defined(DATA_A_Q3_K) +#define QUANT_K QUANT_K_Q3_K +#define A_TYPE block_q3_K +#define A_TYPE_PACKED16 block_q3_K_packed16 +#endif + +#define QUANT_K_Q4_K 256 + +struct block_q4_K +{ + f16vec2 d; + uint8_t scales[3*QUANT_K_Q4_K/64]; + uint8_t qs[QUANT_K_Q4_K/2]; +}; + +struct block_q4_K_packed16 +{ + f16vec2 d; + uint16_t scales[3*QUANT_K_Q4_K/64/2]; + uint16_t qs[QUANT_K_Q4_K/2/2]; +}; + +struct block_q4_K_packed32 +{ + f16vec2 d; + uint32_t scales[3*QUANT_K_Q4_K/64/4]; + uint32_t qs[QUANT_K_Q4_K/2/4]; +}; + +#if defined(DATA_A_Q4_K) +#define QUANT_K QUANT_K_Q4_K +#define A_TYPE block_q4_K +#define A_TYPE_PACKED16 block_q4_K_packed16 +#define A_TYPE_PACKED32 block_q4_K_packed32 +#endif + +#define QUANT_K_Q5_K 256 + +struct block_q5_K +{ + f16vec2 d; + uint8_t scales[12]; + uint8_t qh[QUANT_K_Q5_K/8]; + uint8_t qs[QUANT_K_Q5_K/2]; +}; + +struct block_q5_K_packed16 +{ + f16vec2 d; + uint16_t scales[12/2]; + uint16_t qh[QUANT_K_Q5_K/8/2]; + uint16_t qs[QUANT_K_Q5_K/2/2]; +}; + +#if defined(DATA_A_Q5_K) +#define QUANT_K QUANT_K_Q5_K +#define A_TYPE block_q5_K +#define A_TYPE_PACKED16 block_q5_K_packed16 +#endif + +#define QUANT_K_Q6_K 256 + +struct block_q6_K +{ + uint8_t ql[QUANT_K_Q6_K/2]; + uint8_t qh[QUANT_K_Q6_K/4]; + int8_t scales[QUANT_K_Q6_K/16]; + float16_t d; +}; + +struct block_q6_K_packed16 +{ + uint16_t ql[QUANT_K_Q6_K/2/2]; + uint16_t qh[QUANT_K_Q6_K/4/2]; + int8_t scales[QUANT_K_Q6_K/16]; + float16_t d; +}; + +#if defined(DATA_A_Q6_K) +#define QUANT_K QUANT_K_Q6_K +#define A_TYPE block_q6_K +#define A_TYPE_PACKED16 block_q6_K_packed16 +#endif + +// IQuants + +#define QUANT_K_IQ4_NL 32 +#define QUANT_R_IQ4_NL 2 + +struct block_iq4_nl +{ + float16_t d; + uint8_t qs[QUANT_K_IQ4_NL/2]; +}; + +struct block_iq4_nl_packed16 +{ + float16_t d; + uint16_t qs[QUANT_K_IQ4_NL/2/2]; +}; + +#if defined(DATA_A_IQ4_NL) + +const int8_t kvalues_iq4nl_const[16] = { + int8_t(-127), int8_t(-104), int8_t(-83), int8_t(-65), int8_t(-49), int8_t(-35), int8_t(-22), int8_t(-10), + int8_t(1), int8_t(13), int8_t(25), int8_t(38), int8_t(53), int8_t(69), int8_t(89), int8_t(113) +}; + +shared FLOAT_TYPE kvalues_iq4nl[16]; + +void init_iq4nl_shmem() +{ + // copy the table into shared memory and sync + if (gl_LocalInvocationIndex.x < 16) { + kvalues_iq4nl[gl_LocalInvocationIndex.x] = FLOAT_TYPE(kvalues_iq4nl_const[gl_LocalInvocationIndex.x]); + } + barrier(); +} + +#define QUANT_K QUANT_K_IQ4_NL +#define QUANT_R QUANT_R_IQ4_NL +#define A_TYPE block_iq4_nl +#define A_TYPE_PACKED16 block_iq4_nl_packed16 +#endif + +#endif // !defined(GGML_TYPES_COMP) diff --git a/ggml/src/vulkan-shaders/upscale.comp b/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp similarity index 85% rename from ggml/src/vulkan-shaders/upscale.comp rename to ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp index 511a086ea..6f607380d 100644 --- a/ggml/src/vulkan-shaders/upscale.comp +++ b/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp @@ -2,7 +2,7 @@ layout (push_constant) uniform parameter { - uint ne; uint d_offset; + uint ne; uint a_offset; uint d_offset; uint nb00; uint nb01; uint nb02; uint nb03; uint ne10; uint ne11; uint ne12; uint ne13; float sf0; float sf1; float sf2; float sf3; @@ -32,5 +32,5 @@ void main() { const uint i02 = uint(i12 / p.sf2); const uint i03 = uint(i13 / p.sf3); - data_d[p.d_offset + idx] = D_TYPE(data_a[i03 * p.nb03 + i02 * p.nb02 + i01 * p.nb01 + i00 * p.nb00]); + data_d[p.d_offset + idx] = D_TYPE(data_a[p.a_offset + i03 * p.nb03 + i02 * p.nb02 + i01 * p.nb01 + i00 * p.nb00]); } diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp new file mode 100644 index 000000000..243839917 --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp @@ -0,0 +1,594 @@ + + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef _WIN32 + #include + #include // For _mkdir on Windows + #include // For std::replace on w64devkit +#else + #include + #include + #include +#endif + +#define ASYNCIO_CONCURRENCY 64 + +std::mutex lock; +std::vector> shader_fnames; + +std::string GLSLC = "glslc"; +std::string input_dir = "vulkan-shaders"; +std::string output_dir = "/tmp"; +std::string target_hpp = "ggml-vulkan-shaders.hpp"; +std::string target_cpp = "ggml-vulkan-shaders.cpp"; +bool no_clean = false; + +const std::vector type_names = { + "f32", + "f16", + "q4_0", + "q4_1", + "q5_0", + "q5_1", + "q8_0", + "q2_k", + "q3_k", + "q4_k", + "q5_k", + "q6_k", + "iq4_nl" +}; + +namespace { +void execute_command(const std::string& command, std::string& stdout_str, std::string& stderr_str) { +#ifdef _WIN32 + HANDLE stdout_read, stdout_write; + HANDLE stderr_read, stderr_write; + SECURITY_ATTRIBUTES sa = { sizeof(SECURITY_ATTRIBUTES), NULL, TRUE }; + + if (!CreatePipe(&stdout_read, &stdout_write, &sa, 0) || + !SetHandleInformation(stdout_read, HANDLE_FLAG_INHERIT, 0)) { + throw std::runtime_error("Failed to create stdout pipe"); + } + + if (!CreatePipe(&stderr_read, &stderr_write, &sa, 0) || + !SetHandleInformation(stderr_read, HANDLE_FLAG_INHERIT, 0)) { + throw std::runtime_error("Failed to create stderr pipe"); + } + + PROCESS_INFORMATION pi; + STARTUPINFOA si = {}; + si.cb = sizeof(STARTUPINFOA); + si.dwFlags = STARTF_USESTDHANDLES; + si.hStdOutput = stdout_write; + si.hStdError = stderr_write; + + std::vector cmd(command.begin(), command.end()); + cmd.push_back('\0'); + + if (!CreateProcessA(NULL, cmd.data(), NULL, NULL, TRUE, 0, NULL, NULL, &si, &pi)) { + throw std::runtime_error("Failed to create process"); + } + + CloseHandle(stdout_write); + CloseHandle(stderr_write); + + std::array buffer; + DWORD bytes_read; + + while (ReadFile(stdout_read, buffer.data(), (DWORD)buffer.size(), &bytes_read, NULL) && bytes_read > 0) { + stdout_str.append(buffer.data(), bytes_read); + } + + while (ReadFile(stderr_read, buffer.data(), (DWORD)buffer.size(), &bytes_read, NULL) && bytes_read > 0) { + stderr_str.append(buffer.data(), bytes_read); + } + + CloseHandle(stdout_read); + CloseHandle(stderr_read); + WaitForSingleObject(pi.hProcess, INFINITE); + CloseHandle(pi.hProcess); + CloseHandle(pi.hThread); +#else +int stdout_pipe[2]; + int stderr_pipe[2]; + + if (pipe(stdout_pipe) != 0 || pipe(stderr_pipe) != 0) { + throw std::runtime_error("Failed to create pipes"); + } + + pid_t pid = fork(); + if (pid < 0) { + throw std::runtime_error("Failed to fork process"); + } + + if (pid == 0) { + close(stdout_pipe[0]); + close(stderr_pipe[0]); + dup2(stdout_pipe[1], STDOUT_FILENO); + dup2(stderr_pipe[1], STDERR_FILENO); + close(stdout_pipe[1]); + close(stderr_pipe[1]); + execl("/bin/sh", "sh", "-c", command.c_str(), (char*) nullptr); + _exit(EXIT_FAILURE); + } else { + close(stdout_pipe[1]); + close(stderr_pipe[1]); + + std::array buffer; + ssize_t bytes_read; + + while ((bytes_read = read(stdout_pipe[0], buffer.data(), buffer.size())) > 0) { + stdout_str.append(buffer.data(), bytes_read); + } + + while ((bytes_read = read(stderr_pipe[0], buffer.data(), buffer.size())) > 0) { + stderr_str.append(buffer.data(), bytes_read); + } + + close(stdout_pipe[0]); + close(stderr_pipe[0]); + waitpid(pid, nullptr, 0); + } +#endif +} + +bool directory_exists(const std::string& path) { + struct stat info; + if (stat(path.c_str(), &info) != 0) { + return false; // Path doesn't exist or can't be accessed + } + return (info.st_mode & S_IFDIR) != 0; // Check if it is a directory +} + +bool create_directory(const std::string& path) { +#ifdef _WIN32 + return _mkdir(path.c_str()) == 0 || errno == EEXIST; // EEXIST means the directory already exists +#else + return mkdir(path.c_str(), 0755) == 0 || errno == EEXIST; // 0755 is the directory permissions +#endif +} + +std::string to_uppercase(const std::string& input) { + std::string result = input; + for (char& c : result) { + c = std::toupper(c); + } + return result; +} + +bool string_ends_with(const std::string& str, const std::string& suffix) { + if (suffix.size() > str.size()) { + return false; + } + return std::equal(suffix.rbegin(), suffix.rend(), str.rbegin()); +} + +static const char path_separator = '/'; + +std::string join_paths(const std::string& path1, const std::string& path2) { + return path1 + path_separator + path2; +} + +std::string basename(const std::string &path) { + return path.substr(path.find_last_of("/\\") + 1); +} + +// variables to track number of compiles in progress +static uint32_t compile_count = 0; +static std::mutex compile_count_mutex; +static std::condition_variable compile_count_cond; + +void string_to_spv_func(const std::string& _name, const std::string& in_fname, const std::map& defines, bool fp16 = true, bool coopmat = false, bool coopmat2 = false, bool f16acc = false) { + std::string name = _name + (f16acc ? "_f16acc" : "") + (coopmat ? "_coopmat" : "") + (coopmat2 ? "_cm2" : (fp16 ? "" : "_fp32")); + std::string out_fname = join_paths(output_dir, name + ".spv"); + std::string in_path = join_paths(input_dir, in_fname); + + std::string target_env = (name.find("_cm2") != std::string::npos) ? "--target-env=vulkan1.3" : "--target-env=vulkan1.2"; + + // disable spirv-opt for coopmat shaders for https://github.com/ggerganov/llama.cpp/issues/10734 + std::string opt_level = coopmat ? "" : "-O"; + + #ifdef _WIN32 + std::vector cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, "\"" + in_path + "\"", "-o", "\"" + out_fname + "\""}; + #else + std::vector cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, in_path, "-o", out_fname}; + #endif + + #ifdef GGML_VULKAN_SHADER_DEBUG_INFO + cmd.push_back("-g"); + #endif + + for (const auto& define : defines) { + cmd.push_back("-D" + define.first + "=" + define.second); + } + + std::string command; + for (const auto& part : cmd) { + command += part + " "; + } + + std::string stdout_str, stderr_str; + try { + // std::cout << "Executing command: "; + // for (const auto& part : cmd) { + // std::cout << part << " "; + // } + // std::cout << std::endl; + + execute_command(command, stdout_str, stderr_str); + if (!stderr_str.empty()) { + std::cerr << "cannot compile " << name << "\n\n" << command << "\n\n" << stderr_str << std::endl; + return; + } + + std::lock_guard guard(lock); + shader_fnames.push_back(std::make_pair(name, out_fname)); + } catch (const std::exception& e) { + std::cerr << "Error executing command for " << name << ": " << e.what() << std::endl; + } + { + std::lock_guard guard(compile_count_mutex); + assert(compile_count > 0); + compile_count--; + } + compile_count_cond.notify_all(); +} + +std::map merge_maps(const std::map& a, const std::map& b) { + std::map result = a; + result.insert(b.begin(), b.end()); + return result; +} + +static std::vector> compiles; +void string_to_spv(const std::string& _name, const std::string& in_fname, const std::map& defines, bool fp16 = true, bool coopmat = false, bool coopmat2 = false, bool f16acc = false) { + { + // wait until fewer than N compiles are in progress. + // 16 is an arbitrary limit, the goal is to avoid "failed to create pipe" errors. + uint32_t N = 16; + std::unique_lock guard(compile_count_mutex); + while (compile_count >= N) { + compile_count_cond.wait(guard); + } + compile_count++; + } + compiles.push_back(std::async(string_to_spv_func, _name, in_fname, defines, fp16, coopmat, coopmat2, f16acc)); +} + +void matmul_shaders(bool fp16, bool matmul_id, bool coopmat, bool coopmat2, bool f16acc) { + std::string load_vec = coopmat2 ? "1" : fp16 ? "8" : "4"; + std::string aligned_b_type_f32 = coopmat2 ? "float" : fp16 ? "mat2x4" : "vec4"; + std::string aligned_b_type_f16 = coopmat2 ? "float16_t" : fp16 ? "f16mat2x4" : "f16vec4"; + + std::map base_dict = {{"FLOAT_TYPE", (coopmat2 || fp16) ? "float16_t" : "float"}}; + std::string shader_name = "matmul"; + + if (matmul_id) { + base_dict["MUL_MAT_ID"] = "1"; + shader_name = "matmul_id"; + } + + if (fp16) { + base_dict["FLOAT16"] = "1"; + } + + base_dict["ACC_TYPE"] = f16acc ? "float16_t" : "float"; + + if (coopmat) { + base_dict["COOPMAT"] = "1"; + } + + base_dict["ACC_TYPE"] = f16acc ? "float16_t" : "float"; + + std::string source_name = coopmat2 ? "mul_mm_cm2.comp" : "mul_mm.comp"; + + // Shaders with f16 B_TYPE + string_to_spv(shader_name + "_f32_f16", source_name, merge_maps(base_dict, {{"DATA_A_F32", "1"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}, }), fp16, coopmat, coopmat2, f16acc); + string_to_spv(shader_name + "_f32_f16_aligned", source_name, merge_maps(base_dict, {{"DATA_A_F32", "1"}, {"LOAD_VEC_A", load_vec}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f16}, {"D_TYPE", "float"}, {"ALIGNED", "1"}}), fp16, coopmat, coopmat2, f16acc); + + string_to_spv(shader_name + "_f16_aligned", source_name, merge_maps(base_dict, {{"DATA_A_F16", "1"}, {"LOAD_VEC_A", load_vec}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f16}, {"D_TYPE", "float"}, {"ALIGNED", "1"}}), fp16, coopmat, coopmat2, f16acc); + string_to_spv(shader_name + "_f16", source_name, merge_maps(base_dict, {{"DATA_A_F16", "1"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}}), fp16, coopmat, coopmat2, f16acc); + + for (const auto& tname : type_names) { + std::string data_a_key = "DATA_A_" + to_uppercase(tname); + // For unaligned, load one at a time for f32/f16, or two at a time for quants + std::string load_vec_a_unaligned = (coopmat2 || tname == "f32" || tname == "f16") ? "1" : "2"; + // For aligned matmul loads + std::string load_vec_a = (coopmat2 || tname == "f32" || tname == "f16") ? load_vec : "2"; + + string_to_spv(shader_name + "_" + tname + "_f32", source_name, merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a_unaligned}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"B_IS_FLOAT", "1"}}), fp16, coopmat, coopmat2, f16acc); + string_to_spv(shader_name + "_" + tname + "_f32_aligned", source_name, merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f32}, {"D_TYPE", "float"}, {"B_IS_FLOAT", "1"}, {"ALIGNED", "1"}}), fp16, coopmat, coopmat2, f16acc); + + if (tname != "f16" && tname != "f32") { + string_to_spv(shader_name + "_" + tname + "_f16", source_name, merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a_unaligned}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}, {"B_IS_FLOAT", "1"}}), fp16, coopmat, coopmat2, f16acc); + string_to_spv(shader_name + "_" + tname + "_f16_aligned", source_name, merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f16}, {"D_TYPE", "float"}, {"B_IS_FLOAT", "1"}, {"ALIGNED", "1"}}), fp16, coopmat, coopmat2, f16acc); + } + } +} + +void process_shaders() { + std::cout << "ggml_vulkan: Generating and compiling shaders to SPIR-V" << std::endl; + std::map base_dict = {{"FLOAT_TYPE", "float"}}; + + // matmul + for (const auto& matmul_id : {false, true}) { + // No coopmats + // fp32 + matmul_shaders(false, matmul_id, false, false, false); + + // fp16, fp32acc and fp16acc + matmul_shaders(true, matmul_id, false, false, false); + matmul_shaders(true, matmul_id, false, false, true); + +#if defined(GGML_VULKAN_COOPMAT_GLSLC_SUPPORT) + // Coopmat, fp32acc and fp16acc + matmul_shaders(true, matmul_id, true, false, false); + matmul_shaders(true, matmul_id, true, false, true); +#endif + +#if defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + // Coopmat2, fp32acc and fp16acc + matmul_shaders(true, matmul_id, false, true, false); + matmul_shaders(true, matmul_id, false, true, true); +#endif + } + +#if defined(GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT) + // flash attention + for (const auto& f16acc : {false, true}) { + std::string acctype = f16acc ? "float16_t" : "float"; + + for (const auto& tname : type_names) { + if (tname == "f32") { + continue; + } + + if (tname == "f16") { + string_to_spv("flash_attn_f32_f16_" + tname, "flash_attn_cm2.comp", + merge_maps(base_dict, {{"Q_TYPE", "float"}, {"D_TYPE", "float"}, {"ACC_TYPE", acctype}}), true, false, true, f16acc); + } else { + std::string data_a_key = "DATA_A_" + to_uppercase(tname); + string_to_spv("flash_attn_f32_f16_" + tname, "flash_attn_cm2.comp", + merge_maps(base_dict, {{data_a_key, "1"}, {"Q_TYPE", "float"}, {"D_TYPE", "float"}, {"ACC_TYPE", acctype}, {"DEQUANTFUNC", "dequantFunc"+to_uppercase(tname) }, {"BLOCK_SIZE", "QUANT_K_"+to_uppercase(tname) }}), true, false, true, f16acc); + } + } + } +#endif + + for (const auto& tname : type_names) { + // mul mat vec + std::string data_a_key = "DATA_A_" + to_uppercase(tname); + std::string shader = (string_ends_with(tname, "_k")) ? "mul_mat_vec_" + tname + ".comp" : "mul_mat_vec.comp"; + + string_to_spv("mul_mat_vec_" + tname + "_f32_f32", shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "float"}, {"B_TYPE_VEC2", "vec2"}, {"B_TYPE_VEC4", "vec4"}, {"D_TYPE", "float"}})); + string_to_spv("mul_mat_vec_" + tname + "_f16_f32", shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "float16_t"}, {"B_TYPE_VEC2", "f16vec2"}, {"B_TYPE_VEC4", "f16vec4"}, {"D_TYPE", "float"}})); + + string_to_spv("mul_mat_vec_id_" + tname + "_f32", shader, merge_maps(base_dict, {{"MUL_MAT_ID", "1"}, {data_a_key, "1"}, {"B_TYPE", "float"}, {"B_TYPE_VEC2", "vec2"}, {"B_TYPE_VEC4", "vec4"}, {"D_TYPE", "float"}})); + + // Dequant shaders + if (tname != "f16") { + string_to_spv("dequant_" + tname, "dequant_" + tname + ".comp", merge_maps(base_dict, {{data_a_key, "1"}, {"D_TYPE", "float16_t"}})); + } + + if (!string_ends_with(tname, "_k")) { + shader = (tname == "f32" || tname == "f16") ? "get_rows.comp" : "get_rows_quant.comp"; + + if (tname == "f16") { + string_to_spv("get_rows_" + tname, shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}})); + } else { + string_to_spv("get_rows_" + tname, shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float16_t"}})); + } + string_to_spv("get_rows_" + tname + "_f32", shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float"}})); + } + } + + string_to_spv("mul_mat_vec_p021_f16_f32", "mul_mat_vec_p021.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("mul_mat_vec_nc_f16_f32", "mul_mat_vec_nc.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); + + // Norms + string_to_spv("norm_f32", "norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + string_to_spv("group_norm_f32", "group_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + string_to_spv("rms_norm_f32", "rms_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + + string_to_spv("cpy_f32_f32", "copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("cpy_f32_f16", "copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}}); + string_to_spv("cpy_f16_f16", "copy.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); + string_to_spv("contig_cpy_f32_f32", "contig_copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("contig_cpy_f32_f16", "contig_copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}}); + string_to_spv("contig_cpy_f16_f16", "contig_copy.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); + + string_to_spv("add_f32", "add.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + string_to_spv("add_f16_f32_f16", "add.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float16_t"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("acc_f32", "acc.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("split_k_reduce", "mul_mat_split_k_reduce.comp", {}); + + string_to_spv("mul_f32", "mul.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("div_f32", "div.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("repeat_f32", "repeat.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + + string_to_spv("scale_f32", "scale.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("sqr_f32", "square.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("sin_f32", "sin.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("cos_f32", "cos.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("clamp_f32", "clamp.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); + + string_to_spv("pad_f32", "pad.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + + string_to_spv("concat_f32", "concat.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("concat_f16", "concat.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); + string_to_spv("concat_i32", "concat.comp", {{"A_TYPE", "int"}, {"B_TYPE", "int"}, {"D_TYPE", "int"}}); + + string_to_spv("upscale_f32", "upscale.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); + + string_to_spv("gelu_f32", "gelu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("gelu_quick_f32", "gelu_quick.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("silu_f32", "silu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("relu_f32", "relu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("leaky_relu_f32", "leaky_relu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("tanh_f32", "tanh.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + + string_to_spv("diag_mask_inf_f32", "diag_mask_inf.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + + string_to_spv("soft_max_f32", "soft_max.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}})); + string_to_spv("soft_max_f32_f16", "soft_max.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}})); + + string_to_spv("rope_norm_f32", "rope_norm.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("rope_norm_f16", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}}); + string_to_spv("rope_norm_f16_rte", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}}); + + string_to_spv("rope_neox_f32", "rope_neox.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); + string_to_spv("rope_neox_f16", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}}); + string_to_spv("rope_neox_f16_rte", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}}); + + string_to_spv("argsort_f32", "argsort.comp", {{"A_TYPE", "float"}}); + + string_to_spv("sum_rows_f32", "sum_rows.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + + string_to_spv("im2col_f32", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + string_to_spv("im2col_f32_f16", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}})); + string_to_spv("im2col_f32_f16_rte", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}, {"RTE16", "1"}})); + + string_to_spv("timestep_embedding_f32", "timestep_embedding.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + + string_to_spv("pool2d_f32", "pool2d.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); + + string_to_spv("rwkv_wkv6_f32", "wkv6.comp", merge_maps(base_dict, {{"A_TYPE", "float"}})); + + for (auto &c : compiles) { + c.wait(); + } +} + +void write_output_files() { + FILE* hdr = fopen(target_hpp.c_str(), "w"); + FILE* src = fopen(target_cpp.c_str(), "w"); + + fprintf(hdr, "#include \n\n"); + fprintf(src, "#include \"%s\"\n\n", basename(target_hpp).c_str()); + + for (const auto& pair : shader_fnames) { + const std::string& name = pair.first; + #ifdef _WIN32 + std::string path = pair.second; + std::replace(path.begin(), path.end(), '/', '\\' ); + #else + const std::string& path = pair.second; + #endif + + FILE* spv = fopen(path.c_str(), "rb"); + if (!spv) { + std::cerr << "Error opening SPIR-V file: " << path << " (" << strerror(errno) << ")\n"; + continue; + } + + fseek(spv, 0, SEEK_END); + size_t size = ftell(spv); + fseek(spv, 0, SEEK_SET); + + std::vector data(size); + size_t read_size = fread(data.data(), 1, size, spv); + fclose(spv); + if (read_size != size) { + std::cerr << "Error reading SPIR-V file: " << path << " (" << strerror(errno) << ")\n"; + continue; + } + + fprintf(hdr, "extern unsigned char %s_data[%zu];\n", name.c_str(), size); + fprintf(hdr, "const uint64_t %s_len = %zu;\n\n", name.c_str(), size); + + fprintf(src, "unsigned char %s_data[%zu] = {\n", name.c_str(), size); + for (size_t i = 0; i < size; ++i) { + fprintf(src, "0x%02x,", data[i]); + if ((i + 1) % 12 == 0) fprintf(src, "\n"); + } + fprintf(src, "\n};\n\n"); + + if (!no_clean) { + std::remove(path.c_str()); + } + } + + fclose(hdr); + fclose(src); +} +} + +int main(int argc, char** argv) { + std::map args; + for (int i = 1; i < argc; ++i) { + std::string arg = argv[i]; + if (arg.rfind("--", 0) == 0) { + if (i + 1 < argc && argv[i + 1][0] != '-') { + args[arg] = argv[i + 1]; + ++i; + } else { + args[arg] = ""; + } + } + } + + if (args.find("--glslc") != args.end()) { + GLSLC = args["--glslc"]; // Path to glslc + } + if (args.find("--input-dir") != args.end()) { + input_dir = args["--input-dir"]; // Directory containing shader sources + } + if (args.find("--output-dir") != args.end()) { + output_dir = args["--output-dir"]; // Directory for containing SPIR-V output + } + if (args.find("--target-hpp") != args.end()) { + target_hpp = args["--target-hpp"]; // Path to generated header file + } + if (args.find("--target-cpp") != args.end()) { + target_cpp = args["--target-cpp"]; // Path to generated cpp file + } + if (args.find("--no-clean") != args.end()) { + no_clean = true; // Keep temporary SPIR-V files in output-dir after build + } + + if (!directory_exists(input_dir)) { + std::cerr << "\"" << input_dir << "\" must be a valid directory containing shader sources" << std::endl; + return EXIT_FAILURE; + } + + if (!directory_exists(output_dir)) { + if (!create_directory(output_dir)) { + std::cerr << "Error creating output directory: " << output_dir << "\n"; + return EXIT_FAILURE; + } + } + + process_shaders(); + + write_output_files(); + + return EXIT_SUCCESS; +} diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/wkv6.comp b/ggml/src/ggml-vulkan/vulkan-shaders/wkv6.comp new file mode 100644 index 000000000..35cc6c45f --- /dev/null +++ b/ggml/src/ggml-vulkan/vulkan-shaders/wkv6.comp @@ -0,0 +1,87 @@ +#version 450 + +#extension GL_EXT_control_flow_attributes : require + +#define BLOCK_SIZE 64 +layout(local_size_x = BLOCK_SIZE, local_size_y = 1, local_size_z = 1) in; + +layout(push_constant) uniform Parameters { + uint B; + uint T; + uint C; + uint H; +}; + +layout(binding = 0) readonly buffer KBuf { A_TYPE k[]; }; +layout(binding = 1) readonly buffer VBuf { A_TYPE v[]; }; +layout(binding = 2) readonly buffer RBuf { A_TYPE r[]; }; +layout(binding = 3) readonly buffer TimeFBuf { A_TYPE tf[]; }; +layout(binding = 4) readonly buffer TimeDBuf { A_TYPE td[]; }; +layout(binding = 5) readonly buffer StateBuf { A_TYPE state_in[]; }; +layout(binding = 6) buffer DstBuf { A_TYPE dst[]; }; + +shared A_TYPE _k[BLOCK_SIZE], _r[BLOCK_SIZE], _tf[BLOCK_SIZE], _td[BLOCK_SIZE]; + +void main() { + const uint head_size = BLOCK_SIZE; + const uint batch_id = gl_WorkGroupID.x / H; + const uint head_id = gl_WorkGroupID.x % H; + const uint tid = gl_LocalInvocationID.x; + + const uint state_size = C * head_size; + const uint n_seq_tokens = T / B; + + if (batch_id >= B || head_id >= H) { + return; + } + + A_TYPE state[BLOCK_SIZE]; + [[unroll]] for (uint i = 0; i < head_size; i++) { + state[i] = state_in[batch_id * state_size + head_id * head_size * head_size + + i * head_size + tid]; + } + + barrier(); + _tf[tid] = tf[head_id * head_size + tid]; + barrier(); + + const uint start_t = batch_id * n_seq_tokens * C + head_id * head_size + tid; + const uint end_t = (batch_id + 1) * n_seq_tokens * C + head_id * head_size + tid; + + for (uint t = start_t; t < end_t; t += C) { + barrier(); + _k[tid] = k[t]; + _r[tid] = r[t]; + _td[tid] = td[t]; + barrier(); + + const A_TYPE v_val = v[t]; + A_TYPE y = 0.0; + + [[unroll]] for (uint j = 0; j < head_size; j += 4) { + vec4 k_vec = vec4(_k[j], _k[j+1], _k[j+2], _k[j+3]); + vec4 r_vec = vec4(_r[j], _r[j+1], _r[j+2], _r[j+3]); + vec4 tf_vec = vec4(_tf[j], _tf[j+1], _tf[j+2], _tf[j+3]); + vec4 td_vec = vec4(_td[j], _td[j+1], _td[j+2], _td[j+3]); + vec4 s_vec = vec4(state[j], state[j+1], state[j+2], state[j+3]); + + vec4 kv = k_vec * v_val; + + vec4 temp = tf_vec * kv + s_vec; + y += dot(r_vec, temp); + + s_vec = s_vec * td_vec + kv; + state[j] = s_vec.x; + state[j+1] = s_vec.y; + state[j+2] = s_vec.z; + state[j+3] = s_vec.w; + } + + dst[t] = y; + } + + [[unroll]] for (uint i = 0; i < head_size; i++) { + dst[T * C + batch_id * state_size + head_id * head_size * head_size + + i * head_size + tid] = state[i]; + } +} diff --git a/ggml/src/ggml.c b/ggml/src/ggml.c index bc034015f..ecfb84a80 100644 --- a/ggml/src/ggml.c +++ b/ggml/src/ggml.c @@ -3,10 +3,15 @@ #include "ggml-backend.h" #include "ggml-impl.h" -#include "ggml-cpu-impl.h" -#include "ggml-quants.h" +#include "ggml-threading.h" #include "ggml.h" -#include "ggml-aarch64.h" + +// FIXME: required here for quantization functions +#include "ggml-quants.h" + +#ifdef GGML_USE_CPU_HBM +#include +#endif #if defined(_MSC_VER) || defined(__MINGW32__) #include // using malloc.h with MSC/MINGW @@ -47,6 +52,17 @@ #define UNUSED GGML_UNUSED +#if defined(_MSC_VER) +#define m512bh(p) p +#define m512i(p) p +#else +#define m512bh(p) (__m512bh)(p) +#define m512i(p) (__m512i)(p) +#endif + +// precomputed f32 table for f16 (256 KB) (ggml-impl.h) +float ggml_table_f32_f16[1 << 16]; + #if (defined(__linux__) || defined(__APPLE__) || defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__)) && \ (!defined(TARGET_OS_TV) && !defined(TARGET_OS_WATCH)) #include @@ -363,7 +379,7 @@ void ggml_fp16_to_fp32_row(const ggml_fp16_t * x, float * y, int64_t n) { void ggml_fp32_to_fp16_row(const float * x, ggml_fp16_t * y, int64_t n) { int64_t i = 0; #if defined(__F16C__) - if (ggml_cpu_has_f16c()) { + //if (ggml_cpu_has_f16c()) { for (; i + 7 < n; i += 8) { __m256 x_vec = _mm256_loadu_ps(x + i); __m128i y_vec = _mm256_cvtps_ph(x_vec, _MM_FROUND_TO_NEAREST_INT); @@ -374,7 +390,7 @@ void ggml_fp32_to_fp16_row(const float * x, ggml_fp16_t * y, int64_t n) { __m128i y_vec = _mm_cvtps_ph(x_vec, _MM_FROUND_TO_NEAREST_INT); _mm_storel_epi64((__m128i *)(y + i), y_vec); } - } + //} #endif for (; i < n; i++) { y[i] = GGML_FP32_TO_FP16(x[i]); @@ -384,7 +400,7 @@ void ggml_fp32_to_fp16_row(const float * x, ggml_fp16_t * y, int64_t n) { void ggml_bf16_to_fp32_row(const ggml_bf16_t * x, float * y, int64_t n) { int64_t i = 0; #if defined(__AVX512F__) - if (ggml_cpu_has_avx512()) { + //if (ggml_cpu_has_avx512()) { for (; i + 16 <= n; i += 16) { _mm512_storeu_ps(y + i, _mm512_castsi512_ps( @@ -394,10 +410,10 @@ void ggml_bf16_to_fp32_row(const ggml_bf16_t * x, float * y, int64_t n) { (const __m256i *)(x + i))), 16))); } - } + //} #endif #if defined(__AVX2__) - if (ggml_cpu_has_avx2()) { + //if (ggml_cpu_has_avx2()) { for (; i + 8 <= n; i += 8) { _mm256_storeu_ps(y + i, _mm256_castsi256_ps( @@ -407,7 +423,7 @@ void ggml_bf16_to_fp32_row(const ggml_bf16_t * x, float * y, int64_t n) { (const __m128i *)(x + i))), 16))); } - } + //} #endif for (; i < n; i++) { y[i] = GGML_BF16_TO_FP32(x[i]); @@ -588,7 +604,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(ggml_fp16_t), .is_quantized = false, .to_float = (ggml_to_float_t) ggml_fp16_to_fp32_row, - .from_float = (ggml_from_float_t) ggml_fp32_to_fp16_row, .from_float_ref = (ggml_from_float_t) ggml_fp32_to_fp16_row, }, [GGML_TYPE_Q4_0] = { @@ -597,7 +612,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q4_0), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q4_0, - .from_float = quantize_row_q4_0, .from_float_ref = (ggml_from_float_t) quantize_row_q4_0_ref, }, [GGML_TYPE_Q4_1] = { @@ -606,7 +620,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q4_1), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q4_1, - .from_float = quantize_row_q4_1, .from_float_ref = (ggml_from_float_t) quantize_row_q4_1_ref, }, [4] = { // GGML_TYPE_Q4_2 @@ -614,18 +627,12 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .blck_size = 0, .type_size = 0, .is_quantized = false, - .to_float = NULL, - .from_float = NULL, - .from_float_ref = NULL, }, [5] = { // GGML_TYPE_Q4_3 .type_name = "DEPRECATED", .blck_size = 0, .type_size = 0, .is_quantized = false, - .to_float = NULL, - .from_float = NULL, - .from_float_ref = NULL, }, [GGML_TYPE_Q5_0] = { .type_name = "q5_0", @@ -633,7 +640,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q5_0), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q5_0, - .from_float = quantize_row_q5_0, .from_float_ref = (ggml_from_float_t) quantize_row_q5_0_ref, }, [GGML_TYPE_Q5_1] = { @@ -642,7 +648,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q5_1), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q5_1, - .from_float = quantize_row_q5_1, .from_float_ref = (ggml_from_float_t) quantize_row_q5_1_ref, }, [GGML_TYPE_Q8_0] = { @@ -651,7 +656,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q8_0), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q8_0, - .from_float = quantize_row_q8_0, .from_float_ref = (ggml_from_float_t) quantize_row_q8_0_ref, }, [GGML_TYPE_Q8_1] = { @@ -659,7 +663,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .blck_size = QK8_1, .type_size = sizeof(block_q8_1), .is_quantized = true, - .from_float = quantize_row_q8_1, .from_float_ref = (ggml_from_float_t) quantize_row_q8_1_ref, }, [GGML_TYPE_Q2_K] = { @@ -668,7 +671,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q2_K), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q2_K, - .from_float = quantize_row_q2_K, .from_float_ref = (ggml_from_float_t) quantize_row_q2_K_ref, }, [GGML_TYPE_Q3_K] = { @@ -677,7 +679,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q3_K), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q3_K, - .from_float = quantize_row_q3_K, .from_float_ref = (ggml_from_float_t) quantize_row_q3_K_ref, }, [GGML_TYPE_Q4_K] = { @@ -686,7 +687,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q4_K), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q4_K, - .from_float = quantize_row_q4_K, .from_float_ref = (ggml_from_float_t) quantize_row_q4_K_ref, }, [GGML_TYPE_Q5_K] = { @@ -695,7 +695,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q5_K), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q5_K, - .from_float = quantize_row_q5_K, .from_float_ref = (ggml_from_float_t) quantize_row_q5_K_ref, }, [GGML_TYPE_Q6_K] = { @@ -704,7 +703,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_q6_K), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_q6_K, - .from_float = quantize_row_q6_K, .from_float_ref = (ggml_from_float_t) quantize_row_q6_K_ref, }, [GGML_TYPE_IQ2_XXS] = { @@ -713,7 +711,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq2_xxs), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq2_xxs, - .from_float = NULL, .from_float_ref = NULL, }, [GGML_TYPE_IQ2_XS] = { @@ -722,7 +719,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq2_xs), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq2_xs, - .from_float = NULL, .from_float_ref = NULL, }, [GGML_TYPE_IQ3_XXS] = { @@ -731,7 +727,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq3_xxs), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq3_xxs, - .from_float = quantize_row_iq3_xxs, .from_float_ref = (ggml_from_float_t)quantize_row_iq3_xxs_ref, }, [GGML_TYPE_IQ3_S] = { @@ -740,7 +735,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq3_s), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq3_s, - .from_float = quantize_row_iq3_s, .from_float_ref = (ggml_from_float_t)quantize_row_iq3_s_ref, }, [GGML_TYPE_IQ2_S] = { @@ -749,7 +743,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq2_s), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq2_s, - .from_float = quantize_row_iq2_s, .from_float_ref = (ggml_from_float_t)quantize_row_iq2_s_ref, }, [GGML_TYPE_IQ1_S] = { @@ -758,7 +751,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq1_s), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq1_s, - .from_float = NULL, .from_float_ref = NULL, }, [GGML_TYPE_IQ1_M] = { @@ -767,7 +759,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq1_m), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq1_m, - .from_float = NULL, .from_float_ref = NULL, }, [GGML_TYPE_IQ4_NL] = { @@ -776,7 +767,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq4_nl), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq4_nl, - .from_float = quantize_row_iq4_nl, .from_float_ref = (ggml_from_float_t)quantize_row_iq4_nl_ref, }, [GGML_TYPE_IQ4_XS] = { @@ -785,7 +775,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_iq4_xs), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_iq4_xs, - .from_float = quantize_row_iq4_xs, .from_float_ref = (ggml_from_float_t)quantize_row_iq4_xs_ref, }, [GGML_TYPE_Q8_K] = { @@ -793,7 +782,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .blck_size = QK_K, .type_size = sizeof(block_q8_K), .is_quantized = true, - .from_float = quantize_row_q8_K, }, [GGML_TYPE_BF16] = { .type_name = "bf16", @@ -801,38 +789,25 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(ggml_bf16_t), .is_quantized = false, .to_float = (ggml_to_float_t) ggml_bf16_to_fp32_row, - .from_float = (ggml_from_float_t) ggml_fp32_to_bf16_row, .from_float_ref = (ggml_from_float_t) ggml_fp32_to_bf16_row_ref, }, - [GGML_TYPE_Q4_0_4_4] = { - .type_name = "q4_0_4x4", - .blck_size = QK4_0, - .blck_size_interleave = 4, - .type_size = sizeof(block_q4_0), - .is_quantized = true, - .to_float = NULL, - .from_float = NULL, - .from_float_ref = NULL, + [31] = { // GGML_TYPE_Q4_0_4_4 + .type_name = "TYPE_Q4_0_4_4 REMOVED, use Q4_0 with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, }, - [GGML_TYPE_Q4_0_4_8] = { - .type_name = "q4_0_4x8", - .blck_size = QK4_0, - .blck_size_interleave = 8, - .type_size = sizeof(block_q4_0), - .is_quantized = true, - .to_float = NULL, - .from_float = NULL, - .from_float_ref = NULL, + [32] = { // GGML_TYPE_Q4_0_4_8 + .type_name = "TYPE_Q4_0_4_8 REMOVED, use Q4_0 with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, }, - [GGML_TYPE_Q4_0_8_8] = { - .type_name = "q4_0_8x8", - .blck_size = QK4_0, - .blck_size_interleave = 8, - .type_size = sizeof(block_q4_0), - .is_quantized = true, - .to_float = NULL, - .from_float = NULL, - .from_float_ref = NULL, + [33] = { // GGML_TYPE_Q4_0_8_8 + .type_name = "TYPE_Q4_0_8_8 REMOVED, use Q4_0 with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, }, [GGML_TYPE_TQ1_0] = { .type_name = "tq1_0", @@ -840,7 +815,6 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_tq1_0), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_tq1_0, - .from_float = quantize_row_tq1_0, .from_float_ref = (ggml_from_float_t) quantize_row_tq1_0_ref, }, [GGML_TYPE_TQ2_0] = { @@ -849,9 +823,26 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = { .type_size = sizeof(block_tq2_0), .is_quantized = true, .to_float = (ggml_to_float_t) dequantize_row_tq2_0, - .from_float = quantize_row_tq2_0, .from_float_ref = (ggml_from_float_t) quantize_row_tq2_0_ref, }, + [36] = { // GGML_TYPE_IQ4_NL_4_4 + .type_name = "TYPE_IQ4_NL_4_4 REMOVED, use IQ4_NL with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, + }, + [37] = { // GGML_TYPE_IQ4_NL_4_8 + .type_name = "TYPE_IQ4_NL_4_8 REMOVED, use IQ4_NL with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, + }, + [38] = { // GGML_TYPE_IQ4_NL_8_8 + .type_name = "TYPE_IQ4_NL_8_8 REMOVED, use IQ4_NL with runtime repacking", + .blck_size = 0, + .type_size = 0, + .is_quantized = false, + }, }; const struct ggml_type_traits * ggml_get_type_traits(enum ggml_type type) { @@ -962,6 +953,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = { "POOL_2D_BACK", "UPSCALE", "PAD", + "PAD_REFLECT_1D", "ARANGE", "TIMESTEP_EMBEDDING", "ARGSORT", @@ -976,6 +968,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = { "GET_REL_POS", "ADD_REL_POS", "RWKV_WKV6", + "GATED_LINEAR_ATTN", "UNARY", @@ -995,7 +988,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = { "OPT_STEP_ADAMW", }; -static_assert(GGML_OP_COUNT == 81, "GGML_OP_COUNT != 81"); +static_assert(GGML_OP_COUNT == 83, "GGML_OP_COUNT != 83"); static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = { "none", @@ -1057,6 +1050,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = { "pool_2d_back(x)", "upscale(x)", "pad(x)", + "pad_reflect_1d(x)", "arange(start, stop, step)", "timestep_embedding(timesteps, dim, max_period)", "argsort(x)", @@ -1071,6 +1065,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = { "get_rel_pos(x)", "add_rel_pos(x)", "rwkv_wkv6(k, v, r, tf, td, s)", + "gated_linear_attn(k, v, q, gate, s)", "unary(x)", @@ -1090,7 +1085,7 @@ static const char * GGML_OP_SYMBOL[GGML_OP_COUNT] = { "adamw(x)", }; -static_assert(GGML_OP_COUNT == 81, "GGML_OP_COUNT != 81"); +static_assert(GGML_OP_COUNT == 83, "GGML_OP_COUNT != 83"); static_assert(GGML_OP_POOL_COUNT == 2, "GGML_OP_POOL_COUNT != 2"); @@ -1280,9 +1275,6 @@ enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype) { case GGML_FTYPE_MOSTLY_IQ4_XS: wtype = GGML_TYPE_IQ4_XS; break; case GGML_FTYPE_MOSTLY_IQ3_S: wtype = GGML_TYPE_IQ3_S; break; case GGML_FTYPE_MOSTLY_IQ2_S: wtype = GGML_TYPE_IQ2_S; break; - case GGML_FTYPE_MOSTLY_Q4_0_4_4: wtype = GGML_TYPE_Q4_0_4_4; break; - case GGML_FTYPE_MOSTLY_Q4_0_4_8: wtype = GGML_TYPE_Q4_0_4_8; break; - case GGML_FTYPE_MOSTLY_Q4_0_8_8: wtype = GGML_TYPE_Q4_0_8_8; break; case GGML_FTYPE_UNKNOWN: wtype = GGML_TYPE_COUNT; break; case GGML_FTYPE_MOSTLY_Q4_1_SOME_F16: wtype = GGML_TYPE_COUNT; break; } @@ -1598,35 +1590,23 @@ static struct ggml_tensor * ggml_new_tensor_impl( struct ggml_tensor * const result = (struct ggml_tensor *)((char *)ctx->mem_buffer + obj_new->offs); -#ifdef __clang__ - // temporary until ggml_tensor::backend is removed - #pragma clang diagnostic push - #pragma clang diagnostic ignored "-Wdeprecated-declarations" -#endif - *result = (struct ggml_tensor) { /*.type =*/ type, - /*.backend =*/ GGML_BACKEND_TYPE_CPU, /*.buffer =*/ NULL, /*.ne =*/ { 1, 1, 1, 1 }, /*.nb =*/ { 0, 0, 0, 0 }, /*.op =*/ GGML_OP_NONE, /*.op_params =*/ { 0 }, /*.flags =*/ 0, - /*.grad =*/ NULL, /*.src =*/ { NULL }, /*.view_src =*/ view_src, /*.view_offs =*/ view_offs, /*.data =*/ obj_alloc_size > 0 ? (void *)(result + 1) : data, /*.name =*/ { 0 }, /*.extra =*/ NULL, - ///*.padding =*/ { 0 }, + /*.padding =*/ { 0 }, }; -#ifdef __clang__ - #pragma clang diagnostic pop -#endif - // TODO: this should not be needed as long as we don't rely on aligned SIMD loads //GGML_ASSERT_ALIGNED(result->data); @@ -2277,6 +2257,7 @@ struct ggml_tensor * ggml_argmax( struct ggml_context * ctx, struct ggml_tensor * a) { GGML_ASSERT(ggml_is_matrix(a)); + GGML_ASSERT(a->ne[0] <= INT32_MAX); struct ggml_tensor * result = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, a->ne[1]); @@ -3527,15 +3508,18 @@ static struct ggml_tensor * ggml_rope_impl( GGML_ASSERT(c->ne[0] >= n_dims / 2); } + int sections[4] = {0, 0, 0, 0}; + struct ggml_tensor * result = inplace ? ggml_view_tensor(ctx, a) : ggml_dup_tensor(ctx, a); - int32_t params[11] = { /*n_past*/ 0, n_dims, mode, /*n_ctx*/ 0, n_ctx_orig }; + int32_t params[15] = { /*n_past*/ 0, n_dims, mode, /*n_ctx*/ 0, n_ctx_orig }; memcpy(params + 5, &freq_base, sizeof(float)); memcpy(params + 6, &freq_scale, sizeof(float)); memcpy(params + 7, &ext_factor, sizeof(float)); memcpy(params + 8, &attn_factor, sizeof(float)); memcpy(params + 9, &beta_fast, sizeof(float)); memcpy(params + 10, &beta_slow, sizeof(float)); + memcpy(params + 11, §ions, sizeof(int)*4); ggml_set_op_params(result, params, sizeof(params)); result->op = GGML_OP_ROPE; @@ -3557,6 +3541,53 @@ struct ggml_tensor * ggml_rope( ); } +struct ggml_tensor * ggml_rope_multi( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + struct ggml_tensor * c, + int n_dims, + int sections[4], + int mode, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow) { + // Multimodal Rotary Position Embedding + GGML_ASSERT((mode & 1) == 0 && "mode & 1 == 1 is no longer supported"); + + GGML_ASSERT(ggml_is_vector(b)); + GGML_ASSERT(b->type == GGML_TYPE_I32); + GGML_ASSERT(a->ne[2] * 4 == b->ne[0]); // mrope expecting 4 position ids per token + + if (c) { + GGML_ASSERT(c->type == GGML_TYPE_F32); + GGML_ASSERT(c->ne[0] >= n_dims / 2); + } + + struct ggml_tensor * result = ggml_dup_tensor(ctx, a); + + int32_t params[11 + 4] = { /*n_past*/ 0, n_dims, mode, /*n_ctx*/ 0, n_ctx_orig }; + memcpy(params + 5, &freq_base, sizeof(float)); + memcpy(params + 6, &freq_scale, sizeof(float)); + memcpy(params + 7, &ext_factor, sizeof(float)); + memcpy(params + 8, &attn_factor, sizeof(float)); + memcpy(params + 9, &beta_fast, sizeof(float)); + memcpy(params + 10, &beta_slow, sizeof(float)); + memcpy(¶ms[11], sections, sizeof(int)*4); + ggml_set_op_params(result, params, sizeof(params)); + + result->op = GGML_OP_ROPE; + result->src[0] = a; + result->src[1] = b; + result->src[2] = c; + + return result; +} + struct ggml_tensor * ggml_rope_inplace( struct ggml_context * ctx, struct ggml_tensor * a, @@ -3646,9 +3677,25 @@ struct ggml_tensor * ggml_rope_custom_inplace( ); } +// Apparently solving `n_rot = 2pi * x * base^((2 * max_pos_emb) / n_dims)` for x, we get +// `corr_dim(n_rot) = n_dims * log(max_pos_emb / (n_rot * 2pi)) / (2 * log(base))` +static float ggml_rope_yarn_corr_dim(int n_dims, int n_ctx_orig, float n_rot, float base) { + return n_dims * logf(n_ctx_orig / (n_rot * 2 * (float)M_PI)) / (2 * logf(base)); +} + +void ggml_rope_yarn_corr_dims( + int n_dims, int n_ctx_orig, float freq_base, float beta_fast, float beta_slow, float dims[2] +) { + // start and end correction dims + float start = floorf(ggml_rope_yarn_corr_dim(n_dims, n_ctx_orig, beta_fast, freq_base)); + float end = ceilf(ggml_rope_yarn_corr_dim(n_dims, n_ctx_orig, beta_slow, freq_base)); + dims[0] = MAX(0, start); + dims[1] = MIN(n_dims - 1, end); +} + // ggml_rope_back -struct ggml_tensor * ggml_rope_back( +struct ggml_tensor * ggml_rope_ext_back( struct ggml_context * ctx, struct ggml_tensor * a, struct ggml_tensor * b, @@ -3662,29 +3709,32 @@ struct ggml_tensor * ggml_rope_back( float attn_factor, float beta_fast, float beta_slow) { - GGML_ASSERT(ggml_is_vector(b)); - GGML_ASSERT(b->type == GGML_TYPE_I32); - GGML_ASSERT(a->ne[2] == b->ne[0]); - - struct ggml_tensor * result = ggml_dup_tensor(ctx, a); - - int32_t params[11] = { /*n_past*/ 0, n_dims, mode, /*n_ctx*/ 0, n_ctx_orig }; - memcpy(params + 5, &freq_base, sizeof(float)); - memcpy(params + 6, &freq_scale, sizeof(float)); - memcpy(params + 7, &ext_factor, sizeof(float)); - memcpy(params + 8, &attn_factor, sizeof(float)); - memcpy(params + 9, &beta_fast, sizeof(float)); - memcpy(params + 10, &beta_slow, sizeof(float)); - ggml_set_op_params(result, params, sizeof(params)); - - result->op = GGML_OP_ROPE_BACK; - result->src[0] = a; - result->src[1] = b; - result->src[2] = c; - + struct ggml_tensor * result = ggml_rope_ext( + ctx, a, b, c, n_dims, mode, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow); + result->op = GGML_OP_ROPE_BACK; return result; } +struct ggml_tensor * ggml_rope_multi_back( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + struct ggml_tensor * c, + int n_dims, + int sections[4], + int mode, + int n_ctx_orig, + float freq_base, + float freq_scale, + float ext_factor, + float attn_factor, + float beta_fast, + float beta_slow) { + struct ggml_tensor * result = ggml_rope_multi( + ctx, a, b, c, n_dims, sections, mode, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow); + result->op = GGML_OP_ROPE_BACK; + return result; +} // ggml_clamp struct ggml_tensor * ggml_clamp( @@ -3704,104 +3754,10 @@ struct ggml_tensor * ggml_clamp( return result; } -// ggml_conv_1d - static int64_t ggml_calc_conv_output_size(int64_t ins, int64_t ks, int s, int p, int d) { return (ins + 2 * p - d * (ks - 1) - 1) / s + 1; } -GGML_API struct ggml_tensor * ggml_conv_1d( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - int s0, - int p0, - int d0) { - struct ggml_tensor * im2col = ggml_im2col(ctx, a, b, s0, 0, p0, 0, d0, 0, false, GGML_TYPE_F16); // [N, OL, IC * K] - - struct ggml_tensor * result = - ggml_mul_mat(ctx, - ggml_reshape_2d(ctx, im2col, im2col->ne[0], (im2col->ne[2] * im2col->ne[1])), // [N, OL, IC * K] => [N*OL, IC * K] - ggml_reshape_2d(ctx, a, (a->ne[0] * a->ne[1]), a->ne[2])); // [OC，IC, K] => [OC, IC * K] - - result = ggml_reshape_3d(ctx, result, im2col->ne[1], a->ne[2], im2col->ne[2]); // [N, OC, OL] - - return result; -} - -// ggml_conv_1d_ph - -struct ggml_tensor* ggml_conv_1d_ph( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - int s, - int d) { - return ggml_conv_1d(ctx, a, b, s, a->ne[0] / 2, d); -} - -// ggml_conv_transpose_1d - -static int64_t ggml_calc_conv_transpose_1d_output_size(int64_t ins, int64_t ks, int s, int p, int d) { - return (ins - 1) * s - 2 * p + d * (ks - 1) + 1; -} - -GGML_API struct ggml_tensor * ggml_conv_transpose_1d( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - int s0, - int p0, - int d0) { - GGML_ASSERT(ggml_is_matrix(b)); - GGML_ASSERT(a->ne[2] == b->ne[1]); - GGML_ASSERT(a->ne[3] == 1); - - GGML_ASSERT(p0 == 0); - GGML_ASSERT(d0 == 1); - - const int64_t ne[4] = { - ggml_calc_conv_transpose_1d_output_size(b->ne[0], a->ne[0], s0, 0 /*p0*/, 1 /*d0*/), - a->ne[1], b->ne[2], 1, - }; - struct ggml_tensor * result = ggml_new_tensor(ctx, GGML_TYPE_F32, 4, ne); - - int32_t params[] = { s0, p0, d0 }; - ggml_set_op_params(result, params, sizeof(params)); - - result->op = GGML_OP_CONV_TRANSPOSE_1D; - result->src[0] = a; - result->src[1] = b; - - return result; -} - -// ggml_conv_depthwise - -struct ggml_tensor * ggml_conv_depthwise_2d( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - int s0, - int s1, - int p0, - int p1, - int d0, - int d1) { - struct ggml_tensor * new_a = ggml_reshape_4d(ctx, a, a->ne[0], a->ne[1], 1, a->ne[2] * a->ne[3]); - struct ggml_tensor * im2col = ggml_im2col(ctx, new_a, - ggml_reshape_4d(ctx, b, b->ne[0], b->ne[1], 1, b->ne[2] * b->ne[3]), - s0, s1, p0, p1, d0, d1, true, GGML_TYPE_F16); // [N * IC, OH, OW, KH * KW] - struct ggml_tensor * new_b = ggml_reshape_4d(ctx, im2col, im2col->ne[0], im2col->ne[2] * im2col->ne[1], b->ne[2], b->ne[3]); // [N * IC, OH, OW, KH * KW] => [N, IC, OH * OW, KH * KW] - - new_a = ggml_reshape_4d(ctx, new_a, (new_a->ne[0] * new_a->ne[1]), new_a->ne[2], new_a->ne[3], 1); // [OC，1, KH, KW] => [1, OC, 1, KH * KW] - struct ggml_tensor * result = ggml_mul_mat(ctx, new_a, new_b); - result = ggml_reshape_4d(ctx, result, im2col->ne[1], im2col->ne[2], b->ne[2], b->ne[3]); // [N, OC, OH, OW] - - return result; -} -// ggml_conv_2d - // im2col: [N, IC, IH, IW] => [N, OH, OW, IC*KH*KW] // a: [OC，IC, KH, KW] // b: [N, IC, IH, IW] @@ -3818,10 +3774,11 @@ struct ggml_tensor * ggml_im2col( int d1, bool is_2D, enum ggml_type dst_type) { - if(is_2D) { + if (is_2D) { GGML_ASSERT(a->ne[2] == b->ne[2]); } else { - GGML_ASSERT(a->ne[1] == b->ne[1]); + //GGML_ASSERT(b->ne[1] % a->ne[1] == 0); + GGML_ASSERT(b->ne[1] == a->ne[1]); GGML_ASSERT(b->ne[3] == 1); } @@ -3872,6 +3829,108 @@ struct ggml_tensor * ggml_im2col_back( return result; } +// ggml_conv_1d + +struct ggml_tensor * ggml_conv_1d( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s0, + int p0, + int d0) { + struct ggml_tensor * im2col = ggml_im2col(ctx, a, b, s0, 0, p0, 0, d0, 0, false, GGML_TYPE_F16); // [N, OL, IC * K] + + struct ggml_tensor * result = + ggml_mul_mat(ctx, + ggml_reshape_2d(ctx, im2col, im2col->ne[0], (im2col->ne[2] * im2col->ne[1])), // [N, OL, IC * K] => [N*OL, IC * K] + ggml_reshape_2d(ctx, a, (a->ne[0] * a->ne[1]), a->ne[2])); // [OC，IC, K] => [OC, IC * K] + + result = ggml_reshape_3d(ctx, result, im2col->ne[1], a->ne[2], im2col->ne[2]); // [N, OC, OL] + + return result; +} + +// ggml_conv_1d_ph + +struct ggml_tensor* ggml_conv_1d_ph( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s, + int d) { + return ggml_conv_1d(ctx, a, b, s, a->ne[0] / 2, d); +} + +// ggml_conv_1d_dw + +struct ggml_tensor * ggml_conv_1d_dw( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s0, + int p0, + int d0) { + struct ggml_tensor * new_a = ggml_reshape_4d(ctx, a, a->ne[0], 1, a->ne[1], a->ne[2]); + struct ggml_tensor * new_b = ggml_reshape_4d(ctx, b, b->ne[0], 1, b->ne[1], b->ne[2]); + + struct ggml_tensor * im2col = ggml_im2col(ctx, new_a, new_b, s0, 0, p0, 0, d0, 0, false, GGML_TYPE_F16); + + struct ggml_tensor * result = ggml_mul_mat(ctx, im2col, a); + + result = ggml_reshape_3d(ctx, result, b->ne[0], b->ne[1], 1); + + return result; +} + +// ggml_conv_1d_dw_ph + +struct ggml_tensor * ggml_conv_1d_dw_ph( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s0, + int d0) { + return ggml_conv_1d_dw(ctx, a, b, s0, a->ne[0] / 2, d0); +} + +// ggml_conv_transpose_1d + +static int64_t ggml_calc_conv_transpose_1d_output_size(int64_t ins, int64_t ks, int s, int p, int d) { + return (ins - 1) * s - 2 * p + d * (ks - 1) + 1; +} + +GGML_API struct ggml_tensor * ggml_conv_transpose_1d( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s0, + int p0, + int d0) { + GGML_ASSERT(ggml_is_matrix(b)); + GGML_ASSERT(a->ne[2] == b->ne[1]); + GGML_ASSERT(a->ne[3] == 1); + + GGML_ASSERT(p0 == 0); + GGML_ASSERT(d0 == 1); + + const int64_t ne[4] = { + ggml_calc_conv_transpose_1d_output_size(b->ne[0], a->ne[0], s0, 0 /*p0*/, 1 /*d0*/), + a->ne[1], b->ne[2], 1, + }; + struct ggml_tensor * result = ggml_new_tensor(ctx, GGML_TYPE_F32, 4, ne); + + int32_t params[] = { s0, p0, d0 }; + ggml_set_op_params(result, params, sizeof(params)); + + result->op = GGML_OP_CONV_TRANSPOSE_1D; + result->src[0] = a; + result->src[1] = b; + + return result; +} + +// ggml_conv_2d + // a: [OC，IC, KH, KW] // b: [N, IC, IH, IW] // result: [N, OC, OH, OW] @@ -3917,6 +3976,31 @@ struct ggml_tensor * ggml_conv_2d_s1_ph( return ggml_conv_2d(ctx, a, b, 1, 1, a->ne[0] / 2, a->ne[1] / 2, 1, 1); } +// ggml_conv_2d_dw + +struct ggml_tensor * ggml_conv_2d_dw( + struct ggml_context * ctx, + struct ggml_tensor * a, + struct ggml_tensor * b, + int s0, + int s1, + int p0, + int p1, + int d0, + int d1) { + struct ggml_tensor * new_a = ggml_reshape_4d(ctx, a, a->ne[0], a->ne[1], 1, a->ne[2] * a->ne[3]); + struct ggml_tensor * im2col = ggml_im2col(ctx, new_a, + ggml_reshape_4d(ctx, b, b->ne[0], b->ne[1], 1, b->ne[2] * b->ne[3]), + s0, s1, p0, p1, d0, d1, true, GGML_TYPE_F16); // [N * IC, OH, OW, KH * KW] + struct ggml_tensor * new_b = ggml_reshape_4d(ctx, im2col, im2col->ne[0], im2col->ne[2] * im2col->ne[1], b->ne[2], b->ne[3]); // [N * IC, OH, OW, KH * KW] => [N, IC, OH * OW, KH * KW] + + new_a = ggml_reshape_4d(ctx, new_a, (new_a->ne[0] * new_a->ne[1]), new_a->ne[2], new_a->ne[3], 1); // [OC，1, KH, KW] => [1, OC, 1, KH * KW] + struct ggml_tensor * result = ggml_mul_mat(ctx, new_a, new_b); + result = ggml_reshape_4d(ctx, result, im2col->ne[1], im2col->ne[2], b->ne[2], b->ne[3]); // [N, OC, OH, OW] + + return result; +} + // ggml_conv_transpose_2d_p0 static int64_t ggml_calc_conv_transpose_output_size(int64_t ins, int64_t ks, int s, int p) { @@ -4093,6 +4177,37 @@ struct ggml_tensor * ggml_pad( return result; } +// ggml_pad_reflect_1d + +struct ggml_tensor * ggml_pad_reflect_1d( + struct ggml_context * ctx, + struct ggml_tensor * a, + int p0, + int p1) { + GGML_ASSERT(p0 >= 0); + GGML_ASSERT(p1 >= 0); + + GGML_ASSERT(p0 < a->ne[0]); // padding length on each size must be less than the + GGML_ASSERT(p1 < a->ne[0]); // existing length of the dimension being padded + + GGML_ASSERT(ggml_is_contiguous(a)); + GGML_ASSERT(a->type == GGML_TYPE_F32); + + struct ggml_tensor * result = ggml_new_tensor_4d(ctx, a->type, + a->ne[0] + p0 + p1, + a->ne[1], + a->ne[2], + a->ne[3]); + + int32_t params[] = { p0, p1 }; + ggml_set_op_params(result, params, sizeof(params)); + + result->op = GGML_OP_PAD_REFLECT_1D; + result->src[0] = a; + + return result; +} + // ggml_arange struct ggml_tensor * ggml_arange( @@ -4144,6 +4259,7 @@ struct ggml_tensor * ggml_argsort( struct ggml_context * ctx, struct ggml_tensor * a, enum ggml_sort_order order) { + GGML_ASSERT(a->ne[0] <= INT32_MAX); struct ggml_tensor * result = ggml_new_tensor(ctx, GGML_TYPE_I32, GGML_MAX_DIMS, a->ne); ggml_set_op_params_i32(result, 0, (int32_t) order); @@ -4199,8 +4315,6 @@ struct ggml_tensor * ggml_flash_attn_ext( GGML_ASSERT(mask); } - bool is_node = false; - // permute(0, 2, 1, 3) int64_t ne[4] = { q->ne[0], q->ne[2], q->ne[1], q->ne[3] }; struct ggml_tensor * result = ggml_new_tensor(ctx, GGML_TYPE_F32, 4, ne); @@ -4208,8 +4322,7 @@ struct ggml_tensor * ggml_flash_attn_ext( float params[] = { scale, max_bias, logit_softcap }; ggml_set_op_params(result, params, sizeof(params)); - result->op = GGML_OP_FLASH_ATTN_EXT; - result->grad = is_node ? ggml_dup_tensor(ctx, result) : NULL; + result->op = GGML_OP_FLASH_ATTN_EXT; result->src[0] = q; result->src[1] = k; result->src[2] = v; @@ -4228,6 +4341,15 @@ void ggml_flash_attn_ext_set_prec( ggml_set_op_params_i32(a, 3, prec_i32); // scale is on first pos, max_bias on second } +enum ggml_prec ggml_flash_attn_ext_get_prec( + const struct ggml_tensor * a) { + GGML_ASSERT(a->op == GGML_OP_FLASH_ATTN_EXT); + + const int32_t prec_i32 = ggml_get_op_params_i32(a, 3); + + return (enum ggml_prec) prec_i32; +} + // ggml_flash_attn_back struct ggml_tensor * ggml_flash_attn_back( @@ -4268,14 +4390,6 @@ struct ggml_tensor * ggml_flash_attn_back( GGML_ASSERT(ne2 % kvne2 == 0); - bool is_node = false; - - if (q->grad || k->grad || v->grad) { - // when using this operation (in backwards pass) these grads are set. - // we don't want to create (big) grad of our result, so is_node is false. - is_node = false; - } - // store gradients of q, k and v as continuous tensors concatenated in result. // note: v and gradv are actually transposed, i.e. v->ne[0] != D. const int64_t elem_q = ggml_nelements(q); @@ -4298,8 +4412,7 @@ struct ggml_tensor * ggml_flash_attn_back( int32_t masked_i = masked ? 1 : 0; ggml_set_op_params(result, &masked_i, sizeof(masked_i)); - result->op = GGML_OP_FLASH_ATTN_BACK; - result->grad = is_node ? ggml_dup_tensor(ctx, result) : NULL; + result->op = GGML_OP_FLASH_ATTN_BACK; result->src[0] = q; result->src[1] = k; result->src[2] = v; @@ -4521,15 +4634,13 @@ struct ggml_tensor * ggml_rwkv_wkv6( GGML_ASSERT(ggml_is_contiguous(state)); const int64_t S = k->ne[0]; - const int64_t H = k->ne[2]; - const int64_t n_tokens = k->ne[3]; + const int64_t H = k->ne[1]; + const int64_t n_tokens = k->ne[2]; const int64_t n_seqs = state->ne[1]; { - GGML_ASSERT(k->ne[1] == 1); - GGML_ASSERT(v->ne[0] == 1 && v->ne[1] == S && v->ne[2] == H && v->ne[3] == n_tokens); - GGML_ASSERT(r->ne[0] == 1 && r->ne[1] == S && r->ne[2] == H && r->ne[3] == n_tokens); - // TODO: RWKV v4 and v5 - GGML_ASSERT(td->ne[0] == 1 && td->ne[1] == S && td->ne[2] == H && td->ne[3] == n_tokens); + GGML_ASSERT(v->ne[0] == S && v->ne[1] == H && v->ne[2] == n_tokens); + GGML_ASSERT(r->ne[0] == S && r->ne[1] == H && r->ne[2] == n_tokens); + GGML_ASSERT(td->ne[0] == S && td->ne[1] == H && td->ne[2] == n_tokens); GGML_ASSERT(ggml_nelements(state) == S * S * H * n_seqs); } @@ -4548,6 +4659,49 @@ struct ggml_tensor * ggml_rwkv_wkv6( return result; } +// ggml_gated_linear_attn + +struct ggml_tensor * ggml_gated_linear_attn( + struct ggml_context * ctx, + struct ggml_tensor * k, + struct ggml_tensor * v, + struct ggml_tensor * q, + struct ggml_tensor * g, + struct ggml_tensor * state, + float scale) { + GGML_ASSERT(ggml_is_contiguous(k)); + GGML_ASSERT(ggml_is_contiguous(v)); + GGML_ASSERT(ggml_is_contiguous(q)); + GGML_ASSERT(ggml_is_contiguous(g)); + GGML_ASSERT(ggml_is_contiguous(state)); + + const int64_t S = k->ne[0]; + const int64_t H = k->ne[1]; + const int64_t n_tokens = k->ne[2]; + const int64_t n_seqs = state->ne[1]; + { + GGML_ASSERT(v->ne[0] == S && v->ne[1] == H && v->ne[2] == n_tokens); + GGML_ASSERT(q->ne[0] == S && q->ne[1] == H && q->ne[2] == n_tokens); + GGML_ASSERT(g->ne[0] == S && g->ne[1] == H && g->ne[2] == n_tokens); + GGML_ASSERT(ggml_nelements(state) == S * S * H * n_seqs); + } + + // concat output and new_state + const int64_t ne[4] = { S * H, n_tokens + S * n_seqs, 1, 1 }; + struct ggml_tensor * result = ggml_new_tensor(ctx, GGML_TYPE_F32, 4, ne); + + ggml_set_op_params_f32(result, 0, scale); + + result->op = GGML_OP_GATED_LINEAR_ATTN; + result->src[0] = k; + result->src[1] = v; + result->src[2] = q; + result->src[3] = g; + result->src[4] = state; + + return result; +} + // ggml_unary static struct ggml_tensor * ggml_unary_impl( @@ -4941,34 +5095,24 @@ struct ggml_tensor * ggml_opt_step_adamw( struct ggml_context * ctx, struct ggml_tensor * a, struct ggml_tensor * grad, - float alpha, - float beta1, - float beta2, - float eps, - float wd) { + struct ggml_tensor * m, + struct ggml_tensor * v, + struct ggml_tensor * adamw_params) { GGML_ASSERT(a->flags & GGML_TENSOR_FLAG_PARAM); GGML_ASSERT(ggml_are_same_shape(a, grad)); - GGML_ASSERT(alpha > 0.0f); - GGML_ASSERT(beta1 >= 0.0f && beta1 <= 1.0f); - GGML_ASSERT(beta2 >= 0.0f && beta2 <= 1.0f); - GGML_ASSERT(eps >= 0.0f); - GGML_ASSERT(wd >= 0.0f && wd <= 1.0f); + GGML_ASSERT(ggml_are_same_shape(a, m)); + GGML_ASSERT(ggml_are_same_shape(a, v)); + GGML_ASSERT(adamw_params->type == GGML_TYPE_F32); + GGML_ASSERT(ggml_nelements(adamw_params) == 7); struct ggml_tensor * result = ggml_view_tensor(ctx, a); - const int64_t iter = 1; - memcpy(&result->op_params[0], &iter, sizeof(int64_t)); - ggml_set_op_params_f32(result, 2, alpha); - ggml_set_op_params_f32(result, 3, beta1); - ggml_set_op_params_f32(result, 4, beta2); - ggml_set_op_params_f32(result, 5, eps); - ggml_set_op_params_f32(result, 6, wd); - result->op = GGML_OP_OPT_STEP_ADAMW; result->src[0] = a; result->src[1] = grad; - result->src[2] = ggml_dup_tensor(ctx, grad); - result->src[3] = ggml_dup_tensor(ctx, grad); + result->src[2] = m; + result->src[3] = v; + result->src[4] = adamw_params; return result; } @@ -5037,1112 +5181,531 @@ static void ggml_hash_map_free(struct hash_map * map) { GGML_FREE(map); } -// gradient checkpointing +// utility functions to change gradients +// isrc is the index of tensor in cgraph->visited_has_set.keys +// the corresponding gradient (accumulators) are also at position isrc +// if tensor has a gradient accumulator, modify that accumulator in-place +// else if there is no gradient for tensor, set the corresponding value +// else, just add/subtract/etc. the gradients -static struct ggml_tensor * ggml_recompute_graph_node( +static void ggml_add_or_set( struct ggml_context * ctx, - struct ggml_cgraph * graph, - struct hash_map * replacements, - struct ggml_tensor * node) { - - if (node == NULL) { - return NULL; + struct ggml_cgraph * cgraph, + size_t isrc, + struct ggml_tensor * tensor) { + struct ggml_tensor * src = cgraph->visited_hash_set.keys[isrc]; + GGML_ASSERT(src); + if (cgraph->grads[isrc]) { + cgraph->grads[isrc] = ggml_add_impl(ctx, cgraph->grads[isrc], tensor, /*inplace =*/ cgraph->grad_accs[isrc]); + } else { + cgraph->grads[isrc] = tensor; } - - if (node->flags & GGML_TENSOR_FLAG_PARAM) { - return node; - } - - if (!ggml_hash_contains(&graph->visited_hash_set, node)) { - return node; - } - - int count_children = 0; - for (int k = 0; k < GGML_MAX_SRC; ++k) { - if (node->src[k]) { - ++count_children; - } - } - - if (count_children == 0) { - return node; - } - - size_t i = ggml_hash_find(&replacements->set, node); - GGML_ASSERT(i != GGML_HASHSET_FULL); // assert that not full - if (replacements->set.keys[i] == node) { - return replacements->vals[i]; - } - - struct ggml_tensor * clone = ggml_new_tensor(ctx, node->type, GGML_MAX_DIMS, node->ne); - - // insert clone into replacements - GGML_ASSERT(replacements->set.keys[i] == NULL); // assert that we don't overwrite - replacements->set.keys[i] = node; - replacements->vals[i] = clone; - - clone->op = node->op; - clone->grad = node->grad; - clone->flags = node->flags; - clone->extra = node->extra; - for (int k = 0; k < GGML_MAX_DIMS; ++k) { - clone->nb[k] = node->nb[k]; - } - for (int k = 0; k < GGML_MAX_SRC; ++k) { - clone->src[k] = ggml_recompute_graph_node(ctx, graph, replacements, node->src[k]); - } - if (node->view_src != NULL) { - clone->data = (node->view_src->data == NULL) - ? NULL // view_src not yet allocated - : (char *) node->view_src->data // view_src already allocated - + node->view_offs; - clone->view_src = node->view_src; - clone->view_offs = node->view_offs; - } - - GGML_ASSERT(sizeof(node->op_params) == sizeof(int32_t) * (GGML_MAX_OP_PARAMS / sizeof(int32_t))); - GGML_ASSERT(sizeof(node->name) == GGML_MAX_NAME); - memcpy(clone->op_params, node->op_params, sizeof(node->op_params)); - ggml_format_name(clone, "%s (clone)", ggml_get_name(node)); - - return clone; + ggml_format_name(cgraph->grads[isrc], "grad for %s", src->name); + ggml_build_forward_expand(cgraph, cgraph->grads[isrc]); } -void ggml_build_backward_gradient_checkpointing( - struct ggml_context * ctx, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - struct ggml_cgraph * gb_tmp, - struct ggml_tensor * * checkpoints, - int n_checkpoints) { - ggml_graph_cpy(gf, gb_tmp); - ggml_build_backward_expand(ctx, gf, gb_tmp, false); +static void ggml_acc_or_set( + struct ggml_context * ctx, + struct ggml_cgraph * cgraph, + size_t isrc, + struct ggml_tensor * tensor, + const size_t nb1, + const size_t nb2, + const size_t nb3, + const size_t offset) { + struct ggml_tensor * src = cgraph->visited_hash_set.keys[isrc]; + GGML_ASSERT(src); + if (cgraph->grads[isrc]) { + cgraph->grads[isrc] = ggml_acc_impl(ctx, cgraph->grads[isrc], tensor, nb1, nb2, nb3, offset, cgraph->grad_accs[isrc]); + } else { + struct ggml_tensor * a_zero = ggml_scale(ctx, src, 0.0f); // FIXME this is going to produce NaN if a contains inf/NaN + cgraph->grads[isrc] = ggml_acc_impl(ctx, a_zero, tensor, nb1, nb2, nb3, offset, false); + } + ggml_format_name(cgraph->grads[isrc], "grad for %s", cgraph->visited_hash_set.keys[isrc]->name); + ggml_build_forward_expand(cgraph, cgraph->grads[isrc]); +} - if (n_checkpoints <= 0) { - ggml_graph_cpy(gb_tmp, gb); +static void ggml_add1_or_set( + struct ggml_context * ctx, + struct ggml_cgraph * cgraph, + size_t isrc, + struct ggml_tensor * tensor) { + struct ggml_tensor * src = cgraph->visited_hash_set.keys[isrc]; + GGML_ASSERT(src); + if (cgraph->grads[isrc]) { + cgraph->grads[isrc] = ggml_add1_impl(ctx, cgraph->grads[isrc], tensor, cgraph->grad_accs[isrc]); + } else { + cgraph->grads[isrc] = ggml_repeat(ctx, tensor, src); + } + ggml_format_name(cgraph->grads[isrc], "grad for %s", src->name); + ggml_build_forward_expand(cgraph, cgraph->grads[isrc]); +} + +static void ggml_sub_or_set( + struct ggml_context * ctx, + struct ggml_cgraph * cgraph, + size_t isrc, + struct ggml_tensor * tensor) { + struct ggml_tensor * src = cgraph->visited_hash_set.keys[isrc]; + GGML_ASSERT(src); + if (cgraph->grads[isrc]) { + cgraph->grads[isrc] = ggml_sub_impl(ctx, cgraph->grads[isrc], tensor, cgraph->grad_accs[isrc]); + } else { + cgraph->grads[isrc] = ggml_neg(ctx, tensor); + } + ggml_format_name(cgraph->grads[isrc], "grad for %s", src->name); + ggml_build_forward_expand(cgraph, cgraph->grads[isrc]); +} + +static void ggml_compute_backward( + struct ggml_context * ctx, struct ggml_cgraph * cgraph, int i, bool * grads_needed) { + struct ggml_tensor * tensor = cgraph->nodes[i]; + struct ggml_tensor * grad = ggml_graph_get_grad(cgraph, tensor); + + if (!grad) { return; } - struct hash_map * replacements = ggml_new_hash_map(gf->n_nodes + gf->n_leafs + n_checkpoints); - - // insert checkpoints in replacements - for (int i = 0; i < n_checkpoints; ++i) { - size_t k = ggml_hash_find(&replacements->set, checkpoints[i]); - GGML_ASSERT(k != GGML_HASHSET_FULL); // assert that not full - GGML_ASSERT(replacements->set.keys[k] == NULL); // assert that we don't overwrite - replacements->set.keys[k] = checkpoints[i]; - replacements->vals[k] = checkpoints[i]; - } - - ggml_graph_cpy(gf, gb); - // rewrite gb_tmp->nodes[gf->n_nodes:gb_tmp->n_nodes], - // replacing references to gb_tmp->nodes[0:gf->n_nodes] ( == gf->nodes[0:gf->n_nodes]), - // by recomputing them from checkpoints - for (int i = gf->n_nodes; in_nodes; ++i) { - struct ggml_tensor * node = gb_tmp->nodes[i]; - for (int k = 0; k < GGML_MAX_SRC; ++k) { - // insert new tensors recomputing src, reusing already made replacements, - // remember replacements: remember new tensors with mapping from corresponding gf nodes - // recurse for input tensors, - // unless (i.e. terminating when) input tensors are replacements (like checkpoints) - node->src[k] = ggml_recompute_graph_node(ctx, gf, replacements, node->src[k]); - } - // insert rewritten backward node with replacements made into resulting backward graph gb - ggml_build_forward_expand(gb, node); - } - - ggml_hash_map_free(replacements); -} - -// utility functions to change gradients -// if a is in acc_table, modify gradients in-place and mark result as gradient accumulator -// else if a is in zero_table, replace a -// else, just add/subtract/etc. the gradients - -static struct ggml_tensor * ggml_add_or_set( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - struct ggml_hash_set * zero_table, - struct ggml_hash_set * acc_table) { - if (ggml_hash_contains(acc_table, a)) { - struct ggml_tensor * ret = ggml_add_impl(ctx, a, b, true); - const size_t insert_result = ggml_hash_insert(acc_table, ret); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - return ret; - } - if (ggml_hash_contains(zero_table, a)) { - return b; - } - return ggml_add_impl(ctx, a, b, false); -} - -static struct ggml_tensor * ggml_acc_or_set( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - const size_t nb1, - const size_t nb2, - const size_t nb3, - const size_t offset, - struct ggml_hash_set * zero_table, - struct ggml_hash_set * acc_table) { - if (ggml_hash_contains(acc_table, a)) { - struct ggml_tensor * ret = ggml_acc_impl(ctx, a, b, nb1, nb2, nb3, offset, true); - const size_t insert_result = ggml_hash_insert(acc_table, ret); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - return ret; - } - if (ggml_hash_contains(zero_table, a)) { - struct ggml_tensor * a_zero = ggml_scale(ctx, a, 0.0f); // FIXME this is going to produce NaN if a contains inf/NaN - return ggml_acc_impl(ctx, a_zero, b, nb1, nb2, nb3, offset, false); - } - return ggml_acc_impl(ctx, a, b, nb1, nb2, nb3, offset, false); -} - -static struct ggml_tensor * ggml_add1_or_set( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - struct ggml_hash_set * zero_table, - struct ggml_hash_set * acc_table) { - if (ggml_hash_contains(acc_table, a)) { - struct ggml_tensor * ret = ggml_add1_impl(ctx, a, b, true); - const size_t insert_result = ggml_hash_insert(acc_table, ret); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - return ret; - } - if (ggml_hash_contains(zero_table, a)) { - return ggml_repeat(ctx, b, a); - } - return ggml_add1_impl(ctx, a, b, false); -} - -static struct ggml_tensor * ggml_sub_or_set( - struct ggml_context * ctx, - struct ggml_tensor * a, - struct ggml_tensor * b, - struct ggml_hash_set * zero_table, - struct ggml_hash_set * acc_table) { - if (ggml_hash_contains(acc_table, a)) { - struct ggml_tensor * ret = ggml_sub_impl(ctx, a, b, true); - const size_t insert_result = ggml_hash_insert(acc_table, ret); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - return ret; - } - if (ggml_hash_contains(zero_table, a)) { - return ggml_neg(ctx, b); - } - return ggml_sub_impl(ctx, a, b, false); -} - -static void ggml_compute_backward(struct ggml_context * ctx, struct ggml_tensor * tensor, struct ggml_hash_set * zero_table, struct ggml_hash_set * acc_table) { struct ggml_tensor * src0 = tensor->src[0]; struct ggml_tensor * src1 = tensor->src[1]; struct ggml_tensor * src2 = tensor->src[2]; + struct ggml_hash_set * hash_set = &cgraph->visited_hash_set; + const size_t isrc0 = src0 ? ggml_hash_find(hash_set, src0) : (size_t) -1; + const size_t isrc1 = src1 ? ggml_hash_find(hash_set, src1) : (size_t) -1; + const size_t isrc2 = src2 ? ggml_hash_find(hash_set, src2) : (size_t) -1; + const bool src0_needs_grads = src0 && isrc0 != GGML_HASHSET_FULL && ggml_bitset_get(hash_set->used, isrc0) && grads_needed[isrc0]; + const bool src1_needs_grads = src1 && isrc1 != GGML_HASHSET_FULL && ggml_bitset_get(hash_set->used, isrc1) && grads_needed[isrc1]; + const bool src2_needs_grads = src2 && isrc2 != GGML_HASHSET_FULL && ggml_bitset_get(hash_set->used, isrc2) && grads_needed[isrc2]; switch (tensor->op) { - case GGML_OP_DUP: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - } break; - case GGML_OP_ADD: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - if (src1->grad) { - if (ggml_are_same_shape(src0, src1)) { - src1->grad = ggml_add_or_set(ctx, src1->grad, tensor->grad, zero_table, acc_table); - } else { - src1->grad = ggml_add_or_set(ctx, src1->grad, ggml_repeat_back(ctx, tensor->grad, src1), zero_table, acc_table); - } - } - } break; - case GGML_OP_ADD1: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - if (src1->grad) { - src1->grad = ggml_add_or_set(ctx, - src1->grad, - ggml_mean(ctx, tensor->grad), // TODO: should probably be sum instead of mean - zero_table, acc_table); - } - } break; - case GGML_OP_ACC: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - if (src1->grad) { - const size_t nb1 = ((int32_t *) tensor->op_params)[0]; - const size_t nb2 = ((int32_t *) tensor->op_params)[1]; - const size_t nb3 = ((int32_t *) tensor->op_params)[2]; - const size_t offset = ((int32_t *) tensor->op_params)[3]; - - struct ggml_tensor * tensor_grad_view = ggml_view_4d(ctx, - tensor->grad, - src1->grad->ne[0], - src1->grad->ne[1], - src1->grad->ne[2], - src1->grad->ne[3], - nb1, nb2, nb3, offset); - - src1->grad = - ggml_add_or_set(ctx, - src1->grad, - ggml_reshape(ctx, - ggml_cont(ctx, tensor_grad_view), - src1->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_SUB: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - if (src1->grad) { - src1->grad = ggml_sub_or_set(ctx, src1->grad, tensor->grad, zero_table, acc_table); - } - } break; - case GGML_OP_MUL: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_mul(ctx, src1, tensor->grad), - zero_table, acc_table); - } - if (src1->grad) { - src1->grad = - ggml_add_or_set(ctx, - src1->grad, - ggml_mul(ctx, src0, tensor->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_DIV: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_div(ctx, tensor->grad, src1), - zero_table, acc_table); - } - if (src1->grad) { - src1->grad = - ggml_sub_or_set(ctx, - src1->grad, - ggml_mul(ctx, - tensor->grad, - ggml_div(ctx, tensor, src1)), - zero_table, acc_table); - } - } break; - case GGML_OP_SQR: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_scale(ctx, - ggml_mul(ctx, src0, tensor->grad), - 2.0f), - zero_table, acc_table); - } - } break; - case GGML_OP_SQRT: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_scale(ctx, - ggml_div(ctx, - tensor->grad, - tensor), - 0.5f), - zero_table, acc_table); - } - } break; - case GGML_OP_LOG: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_div(ctx, - tensor->grad, - src0), - zero_table, acc_table); - } - } break; - case GGML_OP_SIN: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_mul(ctx, - tensor->grad, - ggml_cos(ctx, src0)), - zero_table, acc_table); - } - } break; - case GGML_OP_COS: - { - if (src0->grad) { - src0->grad = - ggml_sub_or_set(ctx, - src0->grad, - ggml_mul(ctx, - tensor->grad, - ggml_sin(ctx, src0)), - zero_table, acc_table); - } - } break; - case GGML_OP_SUM: - { - if (src0->grad) { - src0->grad = - ggml_add1_or_set(ctx, - src0->grad, - tensor->grad, - zero_table, acc_table); - } - } break; - case GGML_OP_SUM_ROWS: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_repeat(ctx, - tensor->grad, - src0->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_MEAN: - case GGML_OP_ARGMAX: - case GGML_OP_COUNT_EQUAL: - { - GGML_ABORT("fatal error"); // TODO: implement + case GGML_OP_DUP: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_REPEAT: - { - // necessary for llama - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_repeat_back(ctx, tensor->grad, src0->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_REPEAT_BACK: - { - if (src0->grad) { - // TODO: test this - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_repeat(ctx, tensor->grad, src0->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_CONCAT: - { - GGML_ABORT("fatal error"); // TODO: implement + } break; + case GGML_OP_ADD: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_SILU_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented + if (src1_needs_grads) { + struct ggml_tensor * tmp = grad; + if (!ggml_are_same_shape(src0, src1)) { + tmp = ggml_repeat_back(ctx, tmp, src1); + } + ggml_add_or_set(ctx, cgraph, isrc1, tmp); } - case GGML_OP_NORM: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_ADD1: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_RMS_NORM: - { - // necessary for llama - if (src0->grad) { - float eps; - memcpy(&eps, tensor->op_params, sizeof(float)); - - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_rms_norm_back(ctx, src0, tensor->grad, eps), - zero_table, acc_table); - } - } break; - case GGML_OP_RMS_NORM_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented + if (src1_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc1, ggml_mean(ctx, grad)); // TODO: should probably be sum instead of mean } - case GGML_OP_GROUP_NORM: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_ACC: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_MUL_MAT: - { - // https://cs231n.github.io/optimization-2/#staged - // # forward pass - // s0 = np.random.randn(5, 10) - // s1 = np.random.randn(10, 3) - // t = s0.dot(s1) + if (src1_needs_grads) { + const size_t nb1 = ((int32_t *) tensor->op_params)[0]; + const size_t nb2 = ((int32_t *) tensor->op_params)[1]; + const size_t nb3 = ((int32_t *) tensor->op_params)[2]; + const size_t offset = ((int32_t *) tensor->op_params)[3]; - // # now suppose we had the gradient on t from above in the circuit - // dt = np.random.randn(*t.shape) # same shape as t - // ds0 = dt.dot(s1.T) #.T gives the transpose of the matrix - // ds1 = t.T.dot(dt) + struct ggml_tensor * tensor_grad_view = ggml_view_4d(ctx, + grad, src1->ne[0], src1->ne[1], src1->ne[2], src1->ne[3], + nb1, nb2, nb3, offset); - // tensor.shape [m,p,qq,rr] - // src0.shape [n,m,q1,r1] - // src1.shape [n,p,qq,rr] - - // necessary for llama - if (src0->grad) { - struct ggml_tensor * s1_tg = - ggml_out_prod(ctx, // [n,m,qq,rr] - src1, // [n,p,qq,rr] - tensor->grad); // [m,p,qq,rr] - const int64_t qq = s1_tg->ne[2]; - const int64_t rr = s1_tg->ne[3]; - const int64_t q1 = src0->ne[2]; - const int64_t r1 = src0->ne[3]; - const bool ne2_broadcasted = qq > q1; - const bool ne3_broadcasted = rr > r1; - if (ne2_broadcasted || ne3_broadcasted) { - // sum broadcast repetitions of s1_tg into shape of src0 - s1_tg = ggml_repeat_back(ctx, s1_tg, src0); - } - src0->grad = - ggml_add_or_set(ctx, - src0->grad, // [n,m,q1,r1] - s1_tg, // [n,m,q1,r1] - zero_table, acc_table); - } - if (src1->grad) { - src1->grad = - ggml_add_or_set(ctx, - src1->grad, // [n,p,qq,rr] - // ggml_mul_mat(ctx, // [n,p,qq,rr] - // ggml_cont(ctx, // [m,n,q1,r1] - // ggml_transpose(ctx, src0)), // [m,n,q1,r1] - // tensor->grad), // [m,p,qq,rr] - - // // when src0 is bigger than tensor->grad (this is mostly the case in llama), - // // avoid transpose of src0, rather transpose smaller tensor->grad - // // and then use ggml_out_prod - ggml_out_prod(ctx, // [n,p,qq,rr] - src0, // [n,m,q1,r1] - ggml_transpose(ctx, // [p,m,qq,rr] - tensor->grad)), // [m,p,qq,rr] - zero_table, acc_table); - } - } break; - case GGML_OP_MUL_MAT_ID: - { - GGML_ABORT("fatal error"); // TODO: not implemented + ggml_add_or_set(ctx, cgraph, isrc1, ggml_reshape(ctx, ggml_cont(ctx, tensor_grad_view), src1)); } - case GGML_OP_OUT_PROD: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_SUB: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_SCALE: - { - // necessary for llama - if (src0->grad) { - float s; - memcpy(&s, tensor->op_params, sizeof(float)); - - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_scale_impl(ctx, tensor->grad, s, false), - zero_table, acc_table); - } - } break; - case GGML_OP_SET: - { - const size_t nb1 = ((int32_t *) tensor->op_params)[0]; - const size_t nb2 = ((int32_t *) tensor->op_params)[1]; - const size_t nb3 = ((int32_t *) tensor->op_params)[2]; - const size_t offset = ((int32_t *) tensor->op_params)[3]; - - struct ggml_tensor * tensor_grad_view = NULL; - - if (src0->grad || src1->grad) { - GGML_ASSERT(src0->type == tensor->type); - GGML_ASSERT(tensor->grad->type == tensor->type); - GGML_ASSERT(!src1->grad || src1->grad->type == tensor->grad->type); - - tensor_grad_view = ggml_view_4d(ctx, - tensor->grad, src1->ne[0], src1->ne[1], src1->ne[2], src1->ne[3], - nb1, nb2, nb3, offset); - } - - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_acc_impl(ctx, - tensor->grad, - ggml_neg(ctx, tensor_grad_view), - nb1, nb2, nb3, offset, false), - zero_table, acc_table); - } - - if (src1->grad) { - src1->grad = - ggml_add_or_set(ctx, - src1->grad, - ggml_reshape(ctx, - ggml_cont(ctx, tensor_grad_view), - src1->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_CPY: - { - // necessary for llama - // cpy overwrites value of src1 by src0 and returns view(src1) - // the overwriting is mathematically equivalent to: - // tensor = src0 * 1 + src1 * 0 - if (src0->grad) { - // dsrc0 = dtensor * 1 - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - if (src1->grad) { - // dsrc1 = dtensor * 0 -> noop - } - } break; - case GGML_OP_CONT: - { - // same as cpy - if (src0->grad) { - GGML_ASSERT(ggml_is_contiguous(src0->grad)); - GGML_ASSERT(ggml_is_contiguous(tensor->grad)); - src0->grad = ggml_add_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - } break; - case GGML_OP_RESHAPE: - { - // necessary for llama - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, src0->grad, - ggml_reshape(ctx, - ggml_is_contiguous(tensor->grad) - ? tensor->grad - : ggml_cont(ctx, tensor->grad), - src0->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_VIEW: - { - // necessary for llama - if (src0->grad) { - size_t offset; - - memcpy(&offset, tensor->op_params, sizeof(offset)); - - size_t nb1 = tensor->nb[1]; - size_t nb2 = tensor->nb[2]; - size_t nb3 = tensor->nb[3]; - - if (src0->type != src0->grad->type) { - // gradient is typically F32, but src0 could be other type - size_t ng = ggml_element_size(src0->grad); - size_t n0 = ggml_element_size(src0); - GGML_ASSERT(offset % n0 == 0); - GGML_ASSERT(nb1 % n0 == 0); - GGML_ASSERT(nb2 % n0 == 0); - GGML_ASSERT(nb3 % n0 == 0); - offset = (offset / n0) * ng; - nb1 = (nb1 / n0) * ng; - nb2 = (nb2 / n0) * ng; - nb3 = (nb3 / n0) * ng; - } - - src0->grad = ggml_acc_or_set(ctx, src0->grad, tensor->grad, nb1, nb2, nb3, offset, zero_table, acc_table); - } - } break; - case GGML_OP_PERMUTE: - { - // necessary for llama - if (src0->grad) { - int32_t * axes = (int32_t *) tensor->op_params; - int axis0 = axes[0] & 0x3; - int axis1 = axes[1] & 0x3; - int axis2 = axes[2] & 0x3; - int axis3 = axes[3] & 0x3; - int axes_backward[4] = {0,0,0,0}; - axes_backward[axis0] = 0; - axes_backward[axis1] = 1; - axes_backward[axis2] = 2; - axes_backward[axis3] = 3; - src0->grad = - ggml_add_or_set(ctx, src0->grad, - ggml_permute(ctx, - tensor->grad, - axes_backward[0], - axes_backward[1], - axes_backward[2], - axes_backward[3]), - zero_table, acc_table); - } - } break; - case GGML_OP_TRANSPOSE: - { - // necessary for llama - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, src0->grad, - ggml_transpose(ctx, tensor->grad), - zero_table, acc_table); - } - } break; - case GGML_OP_GET_ROWS: - { - // necessary for llama (only for tokenizer) - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, src0->grad, - // last ggml_get_rows_back argument src0->grad is only - // necessary to setup correct output shape - ggml_get_rows_back(ctx, tensor->grad, src1, src0->grad), - zero_table, acc_table); - } - if (src1->grad) { - // noop - } - } break; - case GGML_OP_GET_ROWS_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented + if (src1_needs_grads) { + ggml_sub_or_set(ctx, cgraph, isrc1, grad); } - case GGML_OP_DIAG: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_MUL: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, src1, grad)); } - case GGML_OP_DIAG_MASK_INF: - { - // necessary for llama - if (src0->grad) { - const int n_past = ((int32_t *) tensor->op_params)[0]; - src0->grad = - ggml_add_or_set(ctx, src0->grad, - /* ggml_diag_mask_inf_impl() shouldn't be here */ - /* ref: https://github.com/ggerganov/llama.cpp/pull/4203#discussion_r1412377992 */ - ggml_diag_mask_zero_impl(ctx, tensor->grad, n_past, false), - zero_table, acc_table); + if (src1_needs_grads) { + struct ggml_tensor * tmp = ggml_mul(ctx, src0, grad); + if (!ggml_are_same_shape(src0, src1)) { + tmp = ggml_repeat_back(ctx, tmp, src1); } - } break; - case GGML_OP_DIAG_MASK_ZERO: - { - // necessary for llama - if (src0->grad) { - const int n_past = ((int32_t *) tensor->op_params)[0]; - src0->grad = - ggml_add_or_set(ctx, src0->grad, - ggml_diag_mask_zero_impl(ctx, tensor->grad, n_past, false), - zero_table, acc_table); - } - } break; - case GGML_OP_SOFT_MAX: - { - // necessary for llama - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, src0->grad, - ggml_soft_max_back(ctx, tensor->grad, tensor), - zero_table, acc_table); - } - GGML_ASSERT((!src1 || !src1->grad) && "backward pass for softmax mask not implemented"); - } break; - case GGML_OP_SOFT_MAX_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented + ggml_add_or_set(ctx, cgraph, isrc1, tmp); } - case GGML_OP_ROPE: - { - // necessary for llama - if (src0->grad) { - //const int n_past = ((int32_t *) tensor->op_params)[0]; - const int n_dims = ((int32_t *) tensor->op_params)[1]; - const int mode = ((int32_t *) tensor->op_params)[2]; - //const int n_ctx = ((int32_t *) tensor->op_params)[3]; - const int n_ctx_orig = ((int32_t *) tensor->op_params)[4]; - float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + } break; + case GGML_OP_DIV: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_div(ctx, grad, src1)); + } + if (src1_needs_grads) { + ggml_sub_or_set(ctx, cgraph, isrc1, ggml_mul(ctx, grad, ggml_div(ctx, tensor, src1))); + } + } break; + case GGML_OP_SQR: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_scale(ctx, ggml_mul(ctx, src0, grad), 2.0f)); + } + } break; + case GGML_OP_SQRT: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_scale(ctx, ggml_div(ctx, grad, tensor), 0.5f)); + } + } break; + case GGML_OP_LOG: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_div(ctx, grad, src0)); + } + } break; + case GGML_OP_SIN: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, grad, ggml_cos(ctx, src0))); + } + } break; + case GGML_OP_COS: { + if (src0_needs_grads) { + ggml_sub_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, grad, ggml_sin(ctx, src0))); + } + } break; + case GGML_OP_SUM: { + if (src0_needs_grads) { + ggml_add1_or_set(ctx, cgraph, isrc0, grad); + } + } break; + case GGML_OP_SUM_ROWS: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_repeat(ctx, grad, src0)); + } + } break; + case GGML_OP_MEAN: { + if (src0_needs_grads) { + ggml_add1_or_set(ctx, cgraph, isrc0, ggml_scale_impl(ctx, grad, 1.0f/src0->ne[0], false)); + } + } break; + case GGML_OP_REPEAT: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_repeat_back(ctx, grad, src0)); + } + } break; + case GGML_OP_REPEAT_BACK: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_repeat(ctx, grad, src0)); + } + } break; + case GGML_OP_RMS_NORM: { + if (src0_needs_grads) { + float eps; + memcpy(&eps, tensor->op_params, sizeof(float)); + ggml_add_or_set(ctx, cgraph, isrc0, ggml_rms_norm_back(ctx, src0, grad, eps)); + } + } break; + case GGML_OP_MUL_MAT: { + // https://cs231n.github.io/optimization-2/#staged + // # forward pass + // s0 = np.random.randn(5, 10) + // s1 = np.random.randn(10, 3) + // t = s0.dot(s1) - memcpy(&freq_base, (int32_t *) tensor->op_params + 5, sizeof(float)); - memcpy(&freq_scale, (int32_t *) tensor->op_params + 6, sizeof(float)); - memcpy(&ext_factor, (int32_t *) tensor->op_params + 7, sizeof(float)); - memcpy(&attn_factor, (int32_t *) tensor->op_params + 8, sizeof(float)); - memcpy(&beta_fast, (int32_t *) tensor->op_params + 9, sizeof(float)); - memcpy(&beta_slow, (int32_t *) tensor->op_params + 10, sizeof(float)); + // # now suppose we had the gradient on t from above in the circuit + // dt = np.random.randn(*t.shape) # same shape as t + // ds0 = dt.dot(s1.T) #.T gives the transpose of the matrix + // ds1 = t.T.dot(dt) - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_rope_back(ctx, - tensor->grad, - src1, - src2, - n_dims, - mode, - n_ctx_orig, - freq_base, - freq_scale, - ext_factor, - attn_factor, - beta_fast, - beta_slow), - zero_table, acc_table); + // tensor.shape [m,p,qq,rr] + // src0.shape [n,m,q1,r1] + // src1.shape [n,p,qq,rr] + + if (src0_needs_grads) { + struct ggml_tensor * s1_tg = + ggml_out_prod(ctx, // [n,m,qq,rr] + src1, // [n,p,qq,rr] + grad); // [m,p,qq,rr] + const int64_t qq = s1_tg->ne[2]; + const int64_t rr = s1_tg->ne[3]; + const int64_t q1 = src0->ne[2]; + const int64_t r1 = src0->ne[3]; + const bool ne2_broadcasted = qq > q1; + const bool ne3_broadcasted = rr > r1; + if (ne2_broadcasted || ne3_broadcasted) { + // sum broadcast repetitions of s1_tg into shape of src0 + s1_tg = ggml_repeat_back(ctx, s1_tg, src0); } - GGML_ASSERT((!src2 || !src2->grad) && "gradients for freq factors not implemented"); - } break; - case GGML_OP_ROPE_BACK: - { - if (src0->grad) { - //const int n_past = ((int32_t *) tensor->op_params)[0]; - const int n_dims = ((int32_t *) tensor->op_params)[1]; - const int mode = ((int32_t *) tensor->op_params)[2]; - //const int n_ctx = ((int32_t *) tensor->op_params)[3]; - const int n_ctx_orig = ((int32_t *) tensor->op_params)[4]; - float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + ggml_add_or_set(ctx, cgraph, isrc0, s1_tg /*= [n,m,q1,r1]*/); + } + if (src1_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc1, + // ggml_mul_mat(ctx, // [n,p,qq,rr] + // ggml_cont(ctx, // [m,n,q1,r1] + // ggml_transpose(ctx, src0)), // [m,n,q1,r1] + // grad), // [m,p,qq,rr] - memcpy(&freq_base, (int32_t *) tensor->op_params + 5, sizeof(float)); - memcpy(&freq_scale, (int32_t *) tensor->op_params + 6, sizeof(float)); - memcpy(&ext_factor, (int32_t *) tensor->op_params + 7, sizeof(float)); - memcpy(&attn_factor, (int32_t *) tensor->op_params + 8, sizeof(float)); - memcpy(&beta_fast, (int32_t *) tensor->op_params + 9, sizeof(float)); - memcpy(&beta_slow, (int32_t *) tensor->op_params + 10, sizeof(float)); + // when src0 is bigger than tensor->grad (this is mostly the case in llama), + // avoid transpose of src0, rather transpose smaller tensor->grad + // and then use ggml_out_prod + ggml_out_prod(ctx, // [n,p,qq,rr] + src0, // [n,m,q1,r1] + ggml_transpose(ctx, // [p,m,qq,rr] + grad))); // [m,p,qq,rr] + } + } break; + case GGML_OP_SCALE: { + if (src0_needs_grads) { + float s; + memcpy(&s, tensor->op_params, sizeof(float)); + ggml_add_or_set(ctx, cgraph, isrc0, ggml_scale_impl(ctx, grad, s, false)); + } + } break; + case GGML_OP_SET: { + const size_t nb1 = ((const int32_t *) tensor->op_params)[0]; + const size_t nb2 = ((const int32_t *) tensor->op_params)[1]; + const size_t nb3 = ((const int32_t *) tensor->op_params)[2]; + const size_t offset = ((const int32_t *) tensor->op_params)[3]; - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_rope_impl(ctx, - tensor->grad, - src1, - src2, - n_dims, - mode, - n_ctx_orig, - freq_base, - freq_scale, - ext_factor, - attn_factor, - beta_fast, - beta_slow, - false), - zero_table, acc_table); - } - } break; - case GGML_OP_CLAMP: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_OP_CONV_TRANSPOSE_1D: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_OP_IM2COL: - { - if (src1->grad) { - const int32_t s0 = ggml_get_op_params_i32(tensor, 0); - const int32_t s1 = ggml_get_op_params_i32(tensor, 1); - const int32_t p0 = ggml_get_op_params_i32(tensor, 2); - const int32_t p1 = ggml_get_op_params_i32(tensor, 3); - const int32_t d0 = ggml_get_op_params_i32(tensor, 4); - const int32_t d1 = ggml_get_op_params_i32(tensor, 5); - const bool is_2D = ggml_get_op_params_i32(tensor, 6) == 1; + struct ggml_tensor * tensor_grad_view = NULL; - src1->grad = ggml_add_or_set(ctx, - src1->grad, - ggml_im2col_back(ctx, src0, tensor->grad, src1->ne, s0, s1, p0, p1, d0, d1, is_2D), - zero_table, acc_table); - } - } break; - case GGML_OP_IM2COL_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_OP_CONV_TRANSPOSE_2D: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_OP_POOL_1D: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_OP_POOL_2D: - { - if (src0->grad) { - const enum ggml_op_pool op = ggml_get_op_params_i32(tensor, 0); - const int32_t k0 = ggml_get_op_params_i32(tensor, 1); - const int32_t k1 = ggml_get_op_params_i32(tensor, 2); - const int32_t s0 = ggml_get_op_params_i32(tensor, 3); - const int32_t s1 = ggml_get_op_params_i32(tensor, 4); - const int32_t p0 = ggml_get_op_params_i32(tensor, 5); - const int32_t p1 = ggml_get_op_params_i32(tensor, 6); + if (src0_needs_grads || src1_needs_grads) { + GGML_ASSERT(src0->type == tensor->type); + GGML_ASSERT(!cgraph->grads[isrc0] || cgraph->grads[isrc0]->type == grad->type); + GGML_ASSERT(!cgraph->grads[isrc1] || !src1_needs_grads || cgraph->grads[isrc1]->type == grad->type); - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_pool_2d_back(ctx, tensor->grad, src0, op, k0, k1, s0, s1, p0, p1), - zero_table, acc_table); - } - } break; - case GGML_OP_POOL_2D_BACK: - { - GGML_ABORT("fatal error"); // TODO: not implemented + tensor_grad_view = ggml_view_4d(ctx, + grad, src1->ne[0], src1->ne[1], src1->ne[2], src1->ne[3], + nb1, nb2, nb3, offset); } - case GGML_OP_UPSCALE: - { - GGML_ABORT("fatal error"); // TODO: not implemented + + if (src0_needs_grads) { + struct ggml_tensor * tmp = ggml_neg(ctx, tensor_grad_view); + ggml_add_or_set(ctx, cgraph, isrc0, ggml_acc_impl(ctx, grad, tmp, nb1, nb2, nb3, offset, false)); } - case GGML_OP_PAD: - { - GGML_ABORT("fatal error"); // TODO: not implemented + + if (src1_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc1, ggml_reshape(ctx, ggml_cont(ctx, tensor_grad_view), src1)); } - case GGML_OP_ARANGE: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_CPY: { + // cpy overwrites value of src1 by src0 and returns view(src1) + // the overwriting is mathematically equivalent to: + // tensor = src0 * 1 + src1 * 0 + if (src0_needs_grads) { + // dsrc0 = dtensor * 1 + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_TIMESTEP_EMBEDDING: - { - GGML_ABORT("fatal error"); // TODO: not implemented + if (src1_needs_grads) { + // dsrc1 = dtensor * 0 -> noop } - case GGML_OP_ARGSORT: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_CONT: { + // same as cpy + if (src0_needs_grads) { + GGML_ASSERT(!cgraph->grads[isrc0] || ggml_is_contiguous(cgraph->grads[isrc0])); + GGML_ASSERT(ggml_is_contiguous(grad)); + ggml_add_or_set(ctx, cgraph, isrc0, grad); } - case GGML_OP_LEAKY_RELU: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_RESHAPE: { + if (src0_needs_grads) { + struct ggml_tensor * grad_cont = ggml_is_contiguous(grad) ? grad : ggml_cont(ctx, grad); + ggml_add_or_set(ctx, cgraph, isrc0, ggml_reshape(ctx, grad_cont, src0)); } - case GGML_OP_FLASH_ATTN_EXT: - { - GGML_ABORT("FA backward pass not adapted after rework"); - struct ggml_tensor * flash_grad = NULL; - if (src0->grad || src1->grad || tensor->src[2]->grad) { - int32_t t = ggml_get_op_params_i32(tensor, 0); - GGML_ASSERT(t == 0 || t == 1); - bool masked = t != 0; - flash_grad = - ggml_flash_attn_back(ctx, - src0, - src1, - tensor->src[2], - tensor->grad, - masked); + } break; + case GGML_OP_VIEW: { + if (src0_needs_grads) { + size_t offset; + + memcpy(&offset, tensor->op_params, sizeof(offset)); + + size_t nb1 = tensor->nb[1]; + size_t nb2 = tensor->nb[2]; + size_t nb3 = tensor->nb[3]; + + if (cgraph->grads[isrc0] && src0->type != cgraph->grads[isrc0]->type) { + // gradient is typically F32, but src0 could be other type + size_t ng = ggml_element_size(cgraph->grads[isrc0]); + size_t n0 = ggml_element_size(src0); + GGML_ASSERT(offset % n0 == 0); + GGML_ASSERT(nb1 % n0 == 0); + GGML_ASSERT(nb2 % n0 == 0); + GGML_ASSERT(nb3 % n0 == 0); + offset = (offset / n0) * ng; + nb1 = (nb1 / n0) * ng; + nb2 = (nb2 / n0) * ng; + nb3 = (nb3 / n0) * ng; } - const int64_t elem_q = ggml_nelements(src0); - const int64_t elem_k = ggml_nelements(src1); - const int64_t elem_v = ggml_nelements(src2); - - enum ggml_type result_type = flash_grad->type; - GGML_ASSERT(ggml_blck_size(result_type) == 1); - const size_t tsize = ggml_type_size(result_type); - - const size_t offs_q = 0; - const size_t offs_k = offs_q + GGML_PAD(elem_q * tsize, GGML_MEM_ALIGN); - const size_t offs_v = offs_k + GGML_PAD(elem_k * tsize, GGML_MEM_ALIGN); - - if (src0->grad) { - struct ggml_tensor * view_q = ggml_view_1d(ctx, flash_grad, elem_q, offs_q); - struct ggml_tensor * grad_q = ggml_reshape(ctx, view_q, src0); - src0->grad = ggml_add_or_set(ctx, - src0->grad, - grad_q, - zero_table, acc_table); - } - if (src1->grad) { - struct ggml_tensor * view_k = ggml_view_1d(ctx, flash_grad, elem_k, offs_k); - struct ggml_tensor * grad_k = ggml_reshape(ctx, view_k, src1); - src1->grad = ggml_add_or_set(ctx, - src1->grad, - grad_k, - zero_table, acc_table); - } - if (src2->grad) { - struct ggml_tensor * view_v = ggml_view_1d(ctx, flash_grad, elem_v, offs_v); - struct ggml_tensor * grad_v = ggml_reshape(ctx, view_v, src2); - src2->grad = ggml_add_or_set(ctx, - src2->grad, - grad_v, - zero_table, acc_table); - } - } break; - case GGML_OP_FLASH_ATTN_BACK: - { - GGML_ABORT("fatal error"); // not supported + ggml_acc_or_set(ctx, cgraph, isrc0, grad, nb1, nb2, nb3, offset); } - case GGML_OP_SSM_CONV: - case GGML_OP_SSM_SCAN: - { - GGML_ABORT("fatal error"); // TODO: not implemented + } break; + case GGML_OP_PERMUTE: { + if (src0_needs_grads) { + const int32_t * axes = (const int32_t *) tensor->op_params; + const int axis0 = axes[0] & 0x3; + const int axis1 = axes[1] & 0x3; + const int axis2 = axes[2] & 0x3; + const int axis3 = axes[3] & 0x3; + int axb[4] = {0,0,0,0}; // axes backward + axb[axis0] = 0; + axb[axis1] = 1; + axb[axis2] = 2; + axb[axis3] = 3; + ggml_add_or_set(ctx, cgraph, isrc0, ggml_permute(ctx, grad, axb[0], axb[1], axb[2], axb[3])); } + } break; + case GGML_OP_TRANSPOSE: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_transpose(ctx, grad)); + } + } break; + case GGML_OP_GET_ROWS: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_get_rows_back(ctx, grad, src1, src0)); + } + if (src1_needs_grads) { + // noop + } + } break; + case GGML_OP_DIAG_MASK_INF: { + if (src0_needs_grads) { + /* ggml_diag_mask_inf_impl() shouldn't be here */ + /* ref: https://github.com/ggerganov/llama.cpp/pull/4203#discussion_r1412377992 */ + const int n_past = ((const int32_t *) tensor->op_params)[0]; + ggml_add_or_set(ctx, cgraph, isrc0, ggml_diag_mask_zero_impl(ctx, grad, n_past, false)); + } + } break; + case GGML_OP_DIAG_MASK_ZERO: { + if (src0_needs_grads) { + const int n_past = ((const int32_t *) tensor->op_params)[0]; + ggml_add_or_set(ctx, cgraph, isrc0, ggml_diag_mask_zero_impl(ctx, grad, n_past, false)); + } + } break; + case GGML_OP_SOFT_MAX: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_soft_max_back(ctx, grad, tensor)); + } + GGML_ASSERT((!src1 || !src1_needs_grads) && "backward pass for softmax mask not implemented"); + } break; + case GGML_OP_ROPE: { + if (src0_needs_grads) { + //const int n_past = ((int32_t *) tensor->op_params)[0]; + const int n_dims = ((const int32_t *) tensor->op_params)[1]; + const int mode = ((const int32_t *) tensor->op_params)[2]; + //const int n_ctx = ((int32_t *) tensor->op_params)[3]; + const int n_ctx_orig = ((const int32_t *) tensor->op_params)[4]; + float freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow; + int sections[4] = {0, 0, 0, 0}; + + memcpy(&freq_base, (const float *) tensor->op_params + 5, sizeof(float)); + memcpy(&freq_scale, (const float *) tensor->op_params + 6, sizeof(float)); + memcpy(&ext_factor, (const float *) tensor->op_params + 7, sizeof(float)); + memcpy(&attn_factor, (const float *) tensor->op_params + 8, sizeof(float)); + memcpy(&beta_fast, (const float *) tensor->op_params + 9, sizeof(float)); + memcpy(&beta_slow, (const float *) tensor->op_params + 10, sizeof(float)); + memcpy(§ions, tensor->op_params + 11, sizeof(sections)); + + struct ggml_tensor * rope_back = grad->ne[2] == src1->ne[0] ? + ggml_rope_ext_back(ctx, grad, src1, src2, n_dims, + mode, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow) : + ggml_rope_multi_back(ctx, grad, src1, src2, n_dims, sections, + mode, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow); + ggml_add_or_set(ctx, cgraph, isrc0, rope_back); + } + GGML_ASSERT((!src2 || !src2_needs_grads) && "gradients for freq factors not implemented"); + } break; + case GGML_OP_IM2COL: { + if (src1_needs_grads) { + const int32_t s0 = ggml_get_op_params_i32(tensor, 0); + const int32_t s1 = ggml_get_op_params_i32(tensor, 1); + const int32_t p0 = ggml_get_op_params_i32(tensor, 2); + const int32_t p1 = ggml_get_op_params_i32(tensor, 3); + const int32_t d0 = ggml_get_op_params_i32(tensor, 4); + const int32_t d1 = ggml_get_op_params_i32(tensor, 5); + const bool is_2D = ggml_get_op_params_i32(tensor, 6) == 1; + + ggml_add_or_set(ctx, cgraph, isrc1, ggml_im2col_back(ctx, src0, grad, src1->ne, s0, s1, p0, p1, d0, d1, is_2D)); + } + } break; + case GGML_OP_POOL_2D: { + if (src0_needs_grads) { + const enum ggml_op_pool op = ggml_get_op_params_i32(tensor, 0); + const int32_t k0 = ggml_get_op_params_i32(tensor, 1); + const int32_t k1 = ggml_get_op_params_i32(tensor, 2); + const int32_t s0 = ggml_get_op_params_i32(tensor, 3); + const int32_t s1 = ggml_get_op_params_i32(tensor, 4); + const int32_t p0 = ggml_get_op_params_i32(tensor, 5); + const int32_t p1 = ggml_get_op_params_i32(tensor, 6); + + ggml_add_or_set(ctx, cgraph, isrc0, ggml_pool_2d_back(ctx, grad, src0, op, k0, k1, s0, s1, p0, p1)); + } + } break; case GGML_OP_WIN_PART: case GGML_OP_WIN_UNPART: - case GGML_OP_UNARY: - { - switch (ggml_get_unary_op(tensor)) { - case GGML_UNARY_OP_ABS: - { - if (src0->grad) { - src0->grad = - ggml_add_or_set(ctx, - src0->grad, - ggml_mul(ctx, - ggml_sgn(ctx, src0), - tensor->grad), - zero_table, acc_table); - } - } break; - case GGML_UNARY_OP_SGN: - { - if (src0->grad) { - // noop - } - } break; - case GGML_UNARY_OP_NEG: - { - if (src0->grad) { - src0->grad = ggml_sub_or_set(ctx, src0->grad, tensor->grad, zero_table, acc_table); - } - } break; - case GGML_UNARY_OP_STEP: - { - if (src0->grad) { - // noop - } - } break; - case GGML_UNARY_OP_TANH: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_UNARY_OP_ELU: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_UNARY_OP_RELU: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_mul(ctx, - ggml_step(ctx, src0), - tensor->grad), - zero_table, acc_table); - } - } break; - case GGML_UNARY_OP_SIGMOID: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_UNARY_OP_GELU: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_UNARY_OP_GELU_QUICK: - { - GGML_ABORT("fatal error"); // TODO: not implemented - } - case GGML_UNARY_OP_SILU: - { - // necessary for llama - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_silu_back(ctx, src0, tensor->grad), - zero_table, acc_table); - } - } break; - case GGML_UNARY_OP_EXP: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_mul(ctx, tensor, tensor->grad), - zero_table, acc_table); - } - } break; - default: - GGML_ABORT("fatal error"); - } - } break; - case GGML_OP_GET_REL_POS: - case GGML_OP_ADD_REL_POS: - case GGML_OP_RWKV_WKV6: - case GGML_OP_MAP_UNARY: - case GGML_OP_MAP_BINARY: - case GGML_OP_MAP_CUSTOM1_F32: - case GGML_OP_MAP_CUSTOM2_F32: - case GGML_OP_MAP_CUSTOM3_F32: - case GGML_OP_MAP_CUSTOM1: - case GGML_OP_MAP_CUSTOM2: - case GGML_OP_MAP_CUSTOM3: - { - GGML_ABORT("fatal error"); // not supported + case GGML_OP_UNARY: { + switch (ggml_get_unary_op(tensor)) { + case GGML_UNARY_OP_ABS: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, ggml_sgn(ctx, src0), grad)); + } + } break; + case GGML_UNARY_OP_SGN: { + // noop + } break; + case GGML_UNARY_OP_NEG: { + if (src0_needs_grads) { + ggml_sub_or_set(ctx, cgraph, isrc0, grad); + } + } break; + case GGML_UNARY_OP_STEP: { + // noop + } break; + case GGML_UNARY_OP_RELU: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, ggml_step(ctx, src0), grad)); + } + } break; + case GGML_UNARY_OP_SILU: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_silu_back(ctx, src0, grad)); + } + } break; + case GGML_UNARY_OP_EXP: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_mul(ctx, tensor, grad)); + } + } break; + default: { + fprintf(stderr, "%s: unsupported unary op for backward pass: %s\n", + __func__, ggml_unary_op_name(ggml_get_unary_op(tensor))); + GGML_ABORT("fatal error"); + } //break; } - case GGML_OP_CROSS_ENTROPY_LOSS: - { - if (src0->grad) { - src0->grad = ggml_add_or_set(ctx, - src0->grad, - ggml_cross_entropy_loss_back(ctx, - src0, - src1, - tensor->grad), - zero_table, acc_table); - } - GGML_ASSERT(!src1->grad && "backward pass for labels not implemented"); - } break; - case GGML_OP_CROSS_ENTROPY_LOSS_BACK: - { - GGML_ABORT("fatal error"); // not supported + } break; + case GGML_OP_CROSS_ENTROPY_LOSS: { + if (src0_needs_grads) { + ggml_add_or_set(ctx, cgraph, isrc0, ggml_cross_entropy_loss_back(ctx, src0, src1, grad)); } - case GGML_OP_OPT_STEP_ADAMW: - { - GGML_ABORT("fatal error"); // not supported - } - case GGML_OP_NONE: - { - // nop - } break; + GGML_ASSERT(!src1_needs_grads && "backward pass for labels not implemented"); + } break; + case GGML_OP_NONE: { + // noop + } break; case GGML_OP_COUNT: - { - GGML_ABORT("fatal error"); - } + default: { + fprintf(stderr, "%s: unsupported ggml op for backward pass: %s\n", __func__, ggml_op_name(tensor->op)); + GGML_ABORT("fatal error"); + } //break; } - for (int i = 0; i < GGML_MAX_SRC; ++i) { - if (tensor->src[i] && tensor->src[i]->grad) { - GGML_ASSERT(ggml_are_same_shape(tensor->src[i], tensor->src[i]->grad)); - } - } + GGML_ASSERT(!src0_needs_grads || ggml_are_same_shape(src0, cgraph->grads[isrc0])); + GGML_ASSERT(!src1_needs_grads || ggml_are_same_shape(src1, cgraph->grads[isrc1])); + GGML_ASSERT(!src2_needs_grads || ggml_are_same_shape(src2, cgraph->grads[isrc2])); } static void ggml_visit_parents(struct ggml_cgraph * cgraph, struct ggml_tensor * node) { - if (node->grad == NULL) { - // this usually happens when we generate intermediate nodes from constants in the backward pass - // it can also happen during forward pass, if the user performs computations with constants - if (node->op != GGML_OP_NONE) { - //GGML_PRINT_DEBUG("%s: warning: node %p has no grad, but op %d\n", __func__, (void *) node, node->op); - } - } - // check if already visited if (ggml_hash_insert(&cgraph->visited_hash_set, node) == GGML_HASHSET_ALREADY_EXISTS) { return; @@ -6203,18 +5766,41 @@ void ggml_build_forward_expand(struct ggml_cgraph * cgraph, struct ggml_tensor * ggml_build_forward_impl(cgraph, tensor, true); } -void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * gf, struct ggml_cgraph * gb, bool accumulate) { - GGML_ASSERT(gf->n_nodes > 0); - GGML_ASSERT(gf->grads); +void ggml_build_backward_expand( + struct ggml_context * ctx_static, + struct ggml_context * ctx_compute, + struct ggml_cgraph * cgraph, + bool accumulate) { + GGML_ASSERT(cgraph->n_nodes > 0); + GGML_ASSERT(cgraph->grads); + GGML_ASSERT(cgraph->grad_accs); - for (int i = 0; i < gf->n_nodes; ++i) { - struct ggml_tensor * node = gf->nodes[i]; + const int n_nodes_f = cgraph->n_nodes; + + memset(cgraph->grads, 0, cgraph->visited_hash_set.size*sizeof(struct ggml_tensor *)); + memset(cgraph->grad_accs, 0, cgraph->visited_hash_set.size*sizeof(struct ggml_tensor *)); + bool * grads_needed = calloc(cgraph->visited_hash_set.size, sizeof(bool)); + + { + bool any_params = false; + bool any_loss = false; + for (int i = 0; i < n_nodes_f; ++i) { + struct ggml_tensor * node = cgraph->nodes[i]; + any_params = any_params || (node->flags & GGML_TENSOR_FLAG_PARAM); + any_loss = any_loss || (node->flags & GGML_TENSOR_FLAG_LOSS); + } + GGML_ASSERT(any_params && "no trainable parameters found, did you forget to call ggml_set_param?"); + GGML_ASSERT(any_loss && "no training loss found, did you forget to call ggml_set_loss?"); + } + + for (int i = 0; i < n_nodes_f; ++i) { + struct ggml_tensor * node = cgraph->nodes[i]; if (node->type == GGML_TYPE_I32) { continue; } - bool needs_grad = node->flags & GGML_TENSOR_FLAG_PARAM; + bool node_needs_grad = (node->flags & GGML_TENSOR_FLAG_PARAM) || (node->flags & GGML_TENSOR_FLAG_LOSS); bool ignore_src[GGML_MAX_SRC] = {false}; switch (node->op) { // gradients in node->src[0] for one reason or another have no effect on output gradients @@ -6231,7 +5817,7 @@ void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * } break; // gradients in node->src[1] for one reason or another have no effect on output gradients - case GGML_OP_CPY: // gradients in CPY target are irrelevant + case GGML_OP_CPY: // gradients in CPY target are irrelevant case GGML_OP_GET_ROWS: // row indices not differentiable case GGML_OP_GET_ROWS_BACK: // same as for GET_ROWS case GGML_OP_ROPE: // positions not differentiable @@ -6242,14 +5828,14 @@ void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * break; } for (int j = 0; j < GGML_MAX_SRC; ++j) { - if (!node->src[j] || !node->src[j]->grad || ignore_src[j]) { + if (!node->src[j] || ignore_src[j] || !grads_needed[ggml_hash_find(&cgraph->visited_hash_set, node->src[j])]) { continue; } GGML_ASSERT(node->src[j]->type == GGML_TYPE_F32 || node->src[j]->type == GGML_TYPE_F16); - needs_grad = true; + node_needs_grad = true; break; } - if (!needs_grad) { + if (!node_needs_grad) { continue; } @@ -6257,73 +5843,24 @@ void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * GGML_ASSERT(!node->view_src || node->op == GGML_OP_CPY || node->op == GGML_OP_VIEW || node->op == GGML_OP_RESHAPE || node->op == GGML_OP_PERMUTE || node->op == GGML_OP_TRANSPOSE); - // create a new tensor with the same type and shape as the node and set it as grad - node->grad = ggml_dup_tensor(ctx, node); - } - - // keep tables of original gradients for replacement/accumulation logic - struct ggml_hash_set zero_table = ggml_hash_set_new(gf->size); - struct ggml_hash_set acc_table = ggml_hash_set_new(gf->size); - for (int i = 0; i < gf->n_nodes; i++) { - struct ggml_tensor * node = gf->nodes[i]; - - if (node->grad) { - { - const size_t insert_result = ggml_hash_insert(&zero_table, node->grad); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - } - - // only gradients of trainable parameters should be accumulated - if (accumulate && (node->flags & GGML_TENSOR_FLAG_PARAM)) { - const size_t insert_result = ggml_hash_insert(&acc_table, node->grad); - GGML_ASSERT(insert_result != GGML_HASHSET_FULL); - GGML_ASSERT(insert_result != GGML_HASHSET_ALREADY_EXISTS); - } + const size_t igrad = ggml_hash_find(&cgraph->visited_hash_set, node); + GGML_ASSERT(igrad != GGML_HASHSET_FULL); + GGML_ASSERT(ggml_bitset_get(cgraph->visited_hash_set.used, igrad)); + if ((accumulate && (node->flags & GGML_TENSOR_FLAG_PARAM)) || (node->flags & GGML_TENSOR_FLAG_LOSS)) { + cgraph->grad_accs[igrad] = ggml_dup_tensor(ctx_static, node); + cgraph->grads[igrad] = cgraph->grad_accs[igrad]; + ggml_format_name(cgraph->grad_accs[igrad], "grad acc for %s", node->name); } + grads_needed[igrad] = true; } - for (int i = gf->n_nodes - 1; i >= 0; i--) { - struct ggml_tensor * node = gf->nodes[i]; - + for (int i = n_nodes_f - 1; i >= 0; --i) { // inplace operations to add gradients are not created by ggml_compute_backward except for gradient accumulation // use allocator to automatically make inplace operations - if (node->grad) { - ggml_compute_backward(ctx, node, &zero_table, &acc_table); - } + ggml_compute_backward(ctx_compute, cgraph, i, grads_needed); } - for (int i = 0; i < gf->n_nodes; i++) { - struct ggml_tensor * node = gf->nodes[i]; - - if (node->flags & GGML_TENSOR_FLAG_PARAM) { - GGML_PRINT_DEBUG("%s: found root node %p\n", __func__, (void *) node); - ggml_build_forward_expand(gb, node->grad); - } - } - - ggml_hash_set_free(&zero_table); - ggml_hash_set_free(&acc_table); -} - -void ggml_build_opt_adamw( - struct ggml_context * ctx, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - float alpha, - float beta1, - float beta2, - float eps, - float wd) { - for (int i = 0; i < gf->n_nodes; i++) { - struct ggml_tensor * node = gf->nodes[i]; - - if (node->flags & GGML_TENSOR_FLAG_PARAM) { - GGML_PRINT_DEBUG("%s: found root node %p\n", __func__, (void *) node); - struct ggml_tensor * opt_step = ggml_opt_step_adamw(ctx, node, node->grad, alpha, beta1, beta2, eps, wd); - ggml_build_forward_expand(gb, opt_step); - } - } + free(grads_needed); } static void * incr_ptr_aligned(void ** p, size_t size, size_t align) { @@ -6341,7 +5878,8 @@ static size_t ggml_graph_nbytes(size_t size, bool grads) { incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); // leafs incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); // hash keys if (grads) { - incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); // grads + incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); // grads + incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); // grad_accs } incr_ptr_aligned(&p, ggml_bitset_size(hash_size) * sizeof(ggml_bitset_t), sizeof(ggml_bitset_t)); @@ -6367,10 +5905,12 @@ struct ggml_cgraph * ggml_new_graph_custom(struct ggml_context * ctx, size_t siz void * p = cgraph + 1; - struct ggml_tensor ** nodes_ptr = incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); - struct ggml_tensor ** leafs_ptr = incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); - struct ggml_tensor ** hash_keys_ptr = incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); - struct ggml_tensor ** grads_ptr = grads ? incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)) : NULL; + struct ggml_tensor ** nodes_ptr = incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); + struct ggml_tensor ** leafs_ptr = incr_ptr_aligned(&p, size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); + struct ggml_tensor ** hash_keys_ptr = incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)); + struct ggml_tensor ** grads_ptr = grads ? incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)) : NULL; + struct ggml_tensor ** grad_accs_ptr = grads ? incr_ptr_aligned(&p, hash_size * sizeof(struct ggml_tensor *), sizeof(struct ggml_tensor *)) : NULL; + ggml_bitset_t * hash_used = incr_ptr_aligned(&p, ggml_bitset_size(hash_size) * sizeof(ggml_bitset_t), sizeof(ggml_bitset_t)); // check that we allocated the correct amount of memory @@ -6382,12 +5922,17 @@ struct ggml_cgraph * ggml_new_graph_custom(struct ggml_context * ctx, size_t siz /*.n_leafs =*/ 0, /*.nodes =*/ nodes_ptr, /*.grads =*/ grads_ptr, + /*.grad_accs =*/ grad_accs_ptr, /*.leafs =*/ leafs_ptr, /*.hash_table =*/ { hash_size, hash_used, hash_keys_ptr }, /*.order =*/ GGML_CGRAPH_EVAL_ORDER_LEFT_TO_RIGHT, }; ggml_hash_set_reset(&cgraph->visited_hash_set); + if (grads) { + memset(cgraph->grads, 0, hash_size*sizeof(struct ggml_tensor *)); + memset(cgraph->grad_accs, 0, hash_size*sizeof(struct ggml_tensor *)); + } return cgraph; } @@ -6398,14 +5943,15 @@ struct ggml_cgraph * ggml_new_graph(struct ggml_context * ctx) { struct ggml_cgraph ggml_graph_view(struct ggml_cgraph * cgraph0, int i0, int i1) { struct ggml_cgraph cgraph = { - /*.size =*/ 0, - /*.n_nodes =*/ i1 - i0, - /*.n_leafs =*/ 0, - /*.nodes =*/ cgraph0->nodes + i0, - /*.grads =*/ cgraph0->grads ? cgraph0->grads + i0 : NULL, - /*.leafs =*/ NULL, - /*.hash_table =*/ { 0, NULL, NULL }, - /*.order =*/ cgraph0->order, + /*.size =*/ 0, + /*.n_nodes =*/ i1 - i0, + /*.n_leafs =*/ 0, + /*.nodes =*/ cgraph0->nodes + i0, + /*.grads =*/ NULL, // gradients would need visited_hash_set + /*.grad_accs =*/ NULL, + /*.leafs =*/ NULL, + /*.visited_hash_set =*/ { 0, NULL, NULL }, + /*.order =*/ cgraph0->order, }; return cgraph; @@ -6428,19 +5974,33 @@ void ggml_graph_cpy(struct ggml_cgraph * src, struct ggml_cgraph * dst) { dst->nodes[i] = src->nodes[i]; } - if (src->grads) { - GGML_ASSERT(dst->grads != NULL); - for (int i = 0; i < src->n_nodes; ++i) { - dst->grads[i] = src->grads[i]; - } - } - for (size_t i = 0; i < src->visited_hash_set.size; ++i) { // copy all hashset keys (tensors) that are in use if (ggml_bitset_get(src->visited_hash_set.used, i)) { ggml_hash_insert(&dst->visited_hash_set, src->visited_hash_set.keys[i]); } } + + if (dst->grads) { + memset(dst->grads, 0, dst->visited_hash_set.size*sizeof(struct ggml_tensor *)); + memset(dst->grad_accs, 0, dst->visited_hash_set.size*sizeof(struct ggml_tensor *)); + } + if (src->grads) { + GGML_ASSERT(dst->grads != NULL); + GGML_ASSERT(dst->grad_accs != NULL); + for (int i = 0; i < src->n_nodes; ++i) { + const size_t igrad_src = ggml_hash_find(&src->visited_hash_set, src->nodes[i]); + const size_t igrad_dst = ggml_hash_find(&dst->visited_hash_set, dst->nodes[i]); + + GGML_ASSERT(igrad_src != GGML_HASHSET_FULL); + GGML_ASSERT(ggml_bitset_get(src->visited_hash_set.used, igrad_src)); + GGML_ASSERT(igrad_dst != GGML_HASHSET_FULL); + GGML_ASSERT(ggml_bitset_get(dst->visited_hash_set.used, igrad_dst)); + + dst->grads[igrad_dst] = src->grads[igrad_src]; + dst->grad_accs[igrad_dst] = src->grad_accs[igrad_src]; + } + } } struct ggml_cgraph * ggml_graph_dup(struct ggml_context * ctx, struct ggml_cgraph * cgraph) { @@ -6466,29 +6026,32 @@ void ggml_graph_reset(struct ggml_cgraph * cgraph) { GGML_ASSERT(cgraph->grads != NULL); for (int i = 0; i < cgraph->n_nodes; i++) { - struct ggml_tensor * node = cgraph->nodes[i]; + struct ggml_tensor * node = cgraph->nodes[i]; + struct ggml_tensor * grad_acc = ggml_graph_get_grad_acc(cgraph, node); - // initial gradients of loss should be 1, 0 otherwise - if (node->grad) { - if (node->flags & GGML_TENSOR_FLAG_LOSS) { - GGML_ASSERT(node->grad->buffer); - GGML_ASSERT(node->type == GGML_TYPE_F32); - GGML_ASSERT(ggml_is_scalar(node)); - - const float onef = 1.0f; - ggml_backend_tensor_set(node->grad, &onef, 0, ggml_nbytes(node->grad)); - } else { - ggml_set_zero(node->grad); - } - } - - GGML_ASSERT(node); if (node->op == GGML_OP_OPT_STEP_ADAMW) { - // set iteration to 1 and clear momenta - ggml_set_op_params_i32(node, 0, 1); + // clear momenta ggml_set_zero(node->src[2]); ggml_set_zero(node->src[3]); } + + // initial gradients of loss should be 1, 0 otherwise + if (grad_acc) { + if (node->flags & GGML_TENSOR_FLAG_LOSS) { + GGML_ASSERT(grad_acc->type == GGML_TYPE_F32); + GGML_ASSERT(ggml_is_scalar(grad_acc)); + + const float onef = 1.0f; + if (grad_acc->buffer) { + ggml_backend_tensor_set(grad_acc, &onef, 0, sizeof(float)); + } else { + GGML_ASSERT(grad_acc->data); + *((float *) grad_acc->data) = onef; + } + } else { + ggml_set_zero(grad_acc); + } + } } } @@ -6526,7 +6089,7 @@ void ggml_graph_add_node(struct ggml_cgraph * cgraph, struct ggml_tensor * tenso cgraph->n_nodes++; } -struct ggml_tensor * ggml_graph_get_tensor(struct ggml_cgraph * cgraph, const char * name) { +struct ggml_tensor * ggml_graph_get_tensor(const struct ggml_cgraph * cgraph, const char * name) { for (int i = 0; i < cgraph->n_leafs; i++) { struct ggml_tensor * leaf = cgraph->leafs[i]; @@ -6546,6 +6109,16 @@ struct ggml_tensor * ggml_graph_get_tensor(struct ggml_cgraph * cgraph, const ch return NULL; } +struct ggml_tensor * ggml_graph_get_grad(const struct ggml_cgraph * cgraph, const struct ggml_tensor * node) { + const size_t igrad = ggml_hash_find(&cgraph->visited_hash_set, node); + return igrad != GGML_HASHSET_FULL && ggml_bitset_get(cgraph->visited_hash_set.used, igrad) && cgraph->grads ? cgraph->grads[igrad] : NULL; +} + +struct ggml_tensor * ggml_graph_get_grad_acc(const struct ggml_cgraph * cgraph, const struct ggml_tensor * node) { + const size_t igrad = ggml_hash_find(&cgraph->visited_hash_set, node); + return igrad != GGML_HASHSET_FULL && ggml_bitset_get(cgraph->visited_hash_set.used, igrad) && cgraph->grad_accs ? cgraph->grad_accs[igrad] : NULL; +} + void ggml_graph_print(const struct ggml_cgraph * cgraph) { GGML_LOG_INFO("=== GRAPH ===\n"); @@ -6556,7 +6129,8 @@ void ggml_graph_print(const struct ggml_cgraph * cgraph) { GGML_LOG_INFO(" - %3d: [ %5" PRId64 ", %5" PRId64 ", %5" PRId64 "] %16s %s\n", i, node->ne[0], node->ne[1], node->ne[2], - ggml_op_name(node->op), (node->flags & GGML_TENSOR_FLAG_PARAM) ? "x" : node->grad ? "g" : " "); + ggml_op_name(node->op), (node->flags & GGML_TENSOR_FLAG_PARAM) ? "x" : + ggml_graph_get_grad(cgraph, node) ? "g" : " "); } GGML_LOG_INFO("n_leafs = %d\n", cgraph->n_leafs); @@ -6591,8 +6165,9 @@ static bool ggml_graph_find(const struct ggml_cgraph * cgraph, const struct ggml static struct ggml_tensor * ggml_graph_get_parent(const struct ggml_cgraph * cgraph, const struct ggml_tensor * node) { for (int i = 0; i < cgraph->n_nodes; i++) { struct ggml_tensor * parent = cgraph->nodes[i]; + struct ggml_tensor * grad = ggml_graph_get_grad(cgraph, parent); - if (parent->grad == node) { + if (grad == node) { return parent; } } @@ -6632,6 +6207,7 @@ void ggml_graph_dump_dot(const struct ggml_cgraph * gb, const struct ggml_cgraph for (int i = 0; i < gb->n_nodes; i++) { struct ggml_tensor * node = gb->nodes[i]; + struct ggml_tensor * grad = ggml_graph_get_grad(gb, node); if (ggml_graph_get_parent(gb, node) != NULL) { continue; @@ -6639,7 +6215,7 @@ void ggml_graph_dump_dot(const struct ggml_cgraph * gb, const struct ggml_cgraph if (node->flags & GGML_TENSOR_FLAG_PARAM) { snprintf(color, sizeof(color), "yellow"); - } else if (node->grad) { + } else if (grad) { if (ggml_graph_find(gf, node)) { snprintf(color, sizeof(color), "green"); } else { @@ -6666,8 +6242,8 @@ void ggml_graph_dump_dot(const struct ggml_cgraph * gb, const struct ggml_cgraph fprintf(fp, "%d [%" PRId64 ", %" PRId64 ", %" PRId64 "] | %s", i, node->ne[0], node->ne[1], node->ne[2], ggml_op_symbol(node->op)); } - if (node->grad) { - fprintf(fp, " | %s\"; ]\n", ggml_op_symbol(node->grad->op)); + if (grad) { + fprintf(fp, " | %s\"; ]\n", ggml_op_symbol(grad->op)); } else { fprintf(fp, "\"; ]\n"); } @@ -6852,9 +6428,6 @@ size_t ggml_quantize_chunk( case GGML_TYPE_IQ1_M: result = quantize_iq1_m (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ4_NL: result = quantize_iq4_nl (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_IQ4_XS: result = quantize_iq4_xs (src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; - case GGML_TYPE_Q4_0_4_4: result = quantize_q4_0_4x4(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; - case GGML_TYPE_Q4_0_4_8: result = quantize_q4_0_4x8(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; - case GGML_TYPE_Q4_0_8_8: result = quantize_q4_0_8x8(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break; case GGML_TYPE_F16: { size_t elemsize = sizeof(ggml_fp16_t); @@ -6884,1495 +6457,30 @@ size_t ggml_quantize_chunk( //////////////////////////////////////////////////////////////////////////////// -struct gguf_str { - uint64_t n; // GGUFv2 - char * data; -}; - -static const size_t GGUF_TYPE_SIZE[GGUF_TYPE_COUNT] = { - [GGUF_TYPE_UINT8] = sizeof(uint8_t), - [GGUF_TYPE_INT8] = sizeof(int8_t), - [GGUF_TYPE_UINT16] = sizeof(uint16_t), - [GGUF_TYPE_INT16] = sizeof(int16_t), - [GGUF_TYPE_UINT32] = sizeof(uint32_t), - [GGUF_TYPE_INT32] = sizeof(int32_t), - [GGUF_TYPE_FLOAT32] = sizeof(float), - [GGUF_TYPE_BOOL] = sizeof(bool), - [GGUF_TYPE_STRING] = sizeof(struct gguf_str), - [GGUF_TYPE_UINT64] = sizeof(uint64_t), - [GGUF_TYPE_INT64] = sizeof(int64_t), - [GGUF_TYPE_FLOAT64] = sizeof(double), - [GGUF_TYPE_ARRAY] = 0, // undefined -}; -static_assert(GGUF_TYPE_COUNT == 13, "GGUF_TYPE_COUNT != 13"); - -static const char * GGUF_TYPE_NAME[GGUF_TYPE_COUNT] = { - [GGUF_TYPE_UINT8] = "u8", - [GGUF_TYPE_INT8] = "i8", - [GGUF_TYPE_UINT16] = "u16", - [GGUF_TYPE_INT16] = "i16", - [GGUF_TYPE_UINT32] = "u32", - [GGUF_TYPE_INT32] = "i32", - [GGUF_TYPE_FLOAT32] = "f32", - [GGUF_TYPE_BOOL] = "bool", - [GGUF_TYPE_STRING] = "str", - [GGUF_TYPE_ARRAY] = "arr", - [GGUF_TYPE_UINT64] = "u64", - [GGUF_TYPE_INT64] = "i64", - [GGUF_TYPE_FLOAT64] = "f64", -}; -static_assert(GGUF_TYPE_COUNT == 13, "GGUF_TYPE_COUNT != 13"); - -union gguf_value { - uint8_t uint8; - int8_t int8; - uint16_t uint16; - int16_t int16; - uint32_t uint32; - int32_t int32; - float float32; - uint64_t uint64; - int64_t int64; - double float64; - bool bool_; - - struct gguf_str str; - - struct { - enum gguf_type type; - - uint64_t n; // GGUFv2 - void * data; - } arr; -}; - -struct gguf_kv { - struct gguf_str key; - - enum gguf_type type; - union gguf_value value; -}; - -struct gguf_header { - char magic[4]; - - uint32_t version; - uint64_t n_tensors; // GGUFv2 - uint64_t n_kv; // GGUFv2 -}; - -struct gguf_tensor_info { - struct gguf_str name; - - uint32_t n_dims; - uint64_t ne[GGML_MAX_DIMS]; - - enum ggml_type type; - - uint64_t offset; // offset from start of `data`, must be a multiple of `ALIGNMENT` - - // for writing API - const void * data; - size_t size; -}; - -struct gguf_context { - struct gguf_header header; - - struct gguf_kv * kv; - struct gguf_tensor_info * infos; - - size_t alignment; - size_t offset; // offset of `data` from beginning of file - size_t size; // size of `data` in bytes - - //uint8_t * padding; - void * data; -}; - -static size_t gguf_type_size(enum gguf_type type) { - GGML_ASSERT(0 <= type && type < GGUF_TYPE_COUNT); - return GGUF_TYPE_SIZE[type]; -} - -static bool gguf_tensor_info_sanitize(struct gguf_tensor_info * info) { - if (info->n_dims > GGML_MAX_DIMS) { - fprintf(stderr, "%s: invalid number of dimensions (%" PRIu32 ")\n", __func__, info->n_dims); - return false; - } - - if (info->type < 0 || info->type >= GGML_TYPE_COUNT) { - fprintf(stderr, "%s: invalid type (%d)\n", __func__, info->type); - return false; - } - - if (strlen(info->name.data) >= GGML_MAX_NAME) { - fprintf(stderr, "%s: tensor '%s' name is too long\n", __func__, info->name.data); - return false; - } - - for (uint32_t i = 0; i < info->n_dims; ++i) { - if (info->ne[i] <= 0) { - fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[i]); - return false; - } - } - - // prevent overflow for total number of elements - if (INT64_MAX/info->ne[1] <= info->ne[0]) { - fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[1]); - return false; - } - - if (INT64_MAX/info->ne[2] <= info->ne[0]*info->ne[1]) { - fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[2]); - return false; - } - - if (INT64_MAX/info->ne[3] <= info->ne[0]*info->ne[1]*info->ne[2]) { - fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[3]); - return false; - } - - return true; -} - -static bool gguf_fread_el(FILE * file, void * dst, size_t size, size_t * offset) { - const size_t n = fread(dst, 1, size, file); - *offset += n; - return n == size; -} - -static bool gguf_fread_str(FILE * file, struct gguf_str * p, size_t * offset) { - p->n = 0; - p->data = NULL; - - bool ok = true; - - ok = ok && gguf_fread_el(file, &p->n, sizeof(p->n), offset); - - // early exit if string length is invalid, prevents from integer overflow - if (p->n == SIZE_MAX) { - fprintf(stderr, "%s: invalid string length (%" PRIu64 ")\n", __func__, p->n); - return false; - } - - p->data = calloc(p->n + 1, 1); - if (!p->data) { - fprintf(stderr, "%s: failed to allocate memory for string of length %" PRIu64 "\n", __func__, p->n); - return false; - } - - ok = ok && gguf_fread_el(file, p->data, p->n, offset); - - return ok; -} - -static void gguf_free_kv(struct gguf_kv * kv) { - if (kv->key.data) { - GGML_FREE(kv->key.data); - } - - if (kv->type == GGUF_TYPE_STRING) { - if (kv->value.str.data) { - GGML_FREE(kv->value.str.data); - } - } - - if (kv->type == GGUF_TYPE_ARRAY) { - if (kv->value.arr.data) { - if (kv->value.arr.type == GGUF_TYPE_STRING) { - for (uint64_t j = 0; j < kv->value.arr.n; ++j) { - struct gguf_str * str = &((struct gguf_str *) kv->value.arr.data)[j]; - if (str->data) { - GGML_FREE(str->data); - } - } - } - GGML_FREE(kv->value.arr.data); - } - } -} - -struct gguf_context * gguf_init_empty(void) { - struct gguf_context * ctx = calloc(1, sizeof(struct gguf_context)); - if (!ctx) { - fprintf(stderr, "%s: failed to allocate memory for context\n", __func__); - return NULL; - } - - memcpy(ctx->header.magic, GGUF_MAGIC, sizeof(ctx->header.magic)); - ctx->header.version = GGUF_VERSION; - ctx->header.n_tensors = 0; - ctx->header.n_kv = 0; - - ctx->kv = NULL; - ctx->infos = NULL; - - ctx->alignment = GGUF_DEFAULT_ALIGNMENT; - ctx->offset = 0; - ctx->size = 0; - - ctx->data = NULL; - - return ctx; -} - -struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_params params) { - FILE * file = ggml_fopen(fname, "rb"); - if (!file) { - fprintf(stderr, "%s: failed to open '%s': '%s'\n", __func__, fname, strerror(errno)); - return NULL; - } - - // offset from start of file - size_t offset = 0; - - char magic[4]; - - // check the magic before making allocations - { - gguf_fread_el(file, &magic, sizeof(magic), &offset); - - for (uint32_t i = 0; i < sizeof(magic); i++) { - if (magic[i] != GGUF_MAGIC[i]) { - fprintf(stderr, "%s: invalid magic characters '%c%c%c%c'\n", __func__, magic[0], magic[1], magic[2], magic[3]); - fclose(file); - return NULL; - } - } - } - - bool ok = true; - - struct gguf_context * ctx = calloc(1, sizeof(struct gguf_context)); - if (!ctx) { - fprintf(stderr, "%s: failed to allocate memory for context\n", __func__); - fclose(file); - return NULL; - } - - // read the header - { - strncpy(ctx->header.magic, magic, 4); - - ctx->kv = NULL; - ctx->infos = NULL; - ctx->data = NULL; - - ok = ok && gguf_fread_el(file, &ctx->header.version, sizeof(ctx->header.version), &offset); - ok = ok && gguf_fread_el(file, &ctx->header.n_tensors, sizeof(ctx->header.n_tensors), &offset); - ok = ok && gguf_fread_el(file, &ctx->header.n_kv, sizeof(ctx->header.n_kv), &offset); - - if (ctx->header.version == 1) { - fprintf(stderr, "%s: GGUFv1 is no longer supported. please use a more up-to-date version\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - // sanity-checks to prevent from integer/buffer overflows - - ok = ok && (ctx->header.n_tensors < (SIZE_MAX/2)/sizeof(struct gguf_tensor_info)); - ok = ok && (ctx->header.n_tensors < (SIZE_MAX/2)/ggml_tensor_overhead()); - ok = ok && (ctx->header.n_kv < (SIZE_MAX/2)/sizeof(struct gguf_kv)); - - if (!ok) { - fprintf(stderr, "%s: failed to read header\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - } - - // read the kv pairs - { - const uint64_t n_kv = ctx->header.n_kv; - - ctx->kv = calloc(n_kv, sizeof(struct gguf_kv)); - if (!ctx->kv) { - fprintf(stderr, "%s: failed to allocate memory for kv pairs\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - for (uint64_t i = 0; i < n_kv; ++i) { - struct gguf_kv * kv = &ctx->kv[i]; - - //fprintf(stderr, "%s: reading kv %d\n", __func__, i); - - ok = ok && gguf_fread_str(file, &kv->key, &offset); - ok = ok && gguf_fread_el (file, &kv->type, sizeof(kv->type), &offset); - - //fprintf(stderr, "%s: reading kv with key %s\n", __func__, kv->key.data); - - switch (kv->type) { - case GGUF_TYPE_UINT8: ok = ok && gguf_fread_el (file, &kv->value.uint8, sizeof(kv->value.uint8), &offset); break; - case GGUF_TYPE_INT8: ok = ok && gguf_fread_el (file, &kv->value.int8, sizeof(kv->value.int8), &offset); break; - case GGUF_TYPE_UINT16: ok = ok && gguf_fread_el (file, &kv->value.uint16, sizeof(kv->value.uint16), &offset); break; - case GGUF_TYPE_INT16: ok = ok && gguf_fread_el (file, &kv->value.int16, sizeof(kv->value.int16), &offset); break; - case GGUF_TYPE_UINT32: ok = ok && gguf_fread_el (file, &kv->value.uint32, sizeof(kv->value.uint32), &offset); break; - case GGUF_TYPE_INT32: ok = ok && gguf_fread_el (file, &kv->value.int32, sizeof(kv->value.int32), &offset); break; - case GGUF_TYPE_FLOAT32: ok = ok && gguf_fread_el (file, &kv->value.float32, sizeof(kv->value.float32), &offset); break; - case GGUF_TYPE_UINT64: ok = ok && gguf_fread_el (file, &kv->value.uint64, sizeof(kv->value.uint64), &offset); break; - case GGUF_TYPE_INT64: ok = ok && gguf_fread_el (file, &kv->value.int64, sizeof(kv->value.int64), &offset); break; - case GGUF_TYPE_FLOAT64: ok = ok && gguf_fread_el (file, &kv->value.float64, sizeof(kv->value.float64), &offset); break; - case GGUF_TYPE_BOOL: ok = ok && gguf_fread_el (file, &kv->value.bool_, sizeof(kv->value.bool_), &offset); break; - case GGUF_TYPE_STRING: ok = ok && gguf_fread_str(file, &kv->value.str, &offset); break; - case GGUF_TYPE_ARRAY: - { - ok = ok && gguf_fread_el(file, &kv->value.arr.type, sizeof(kv->value.arr.type), &offset); - ok = ok && gguf_fread_el(file, &kv->value.arr.n, sizeof(kv->value.arr.n), &offset); - - switch (kv->value.arr.type) { - case GGUF_TYPE_UINT8: - case GGUF_TYPE_INT8: - case GGUF_TYPE_UINT16: - case GGUF_TYPE_INT16: - case GGUF_TYPE_UINT32: - case GGUF_TYPE_INT32: - case GGUF_TYPE_FLOAT32: - case GGUF_TYPE_UINT64: - case GGUF_TYPE_INT64: - case GGUF_TYPE_FLOAT64: - case GGUF_TYPE_BOOL: - { - // prevent from integer overflow in the malloc below - if (kv->value.arr.n >= SIZE_MAX/gguf_type_size(kv->value.arr.type)) { - fprintf(stderr, "%s: array size is too large (%" PRIu64 ")\n", __func__, kv->value.arr.n); - fclose(file); - gguf_free(ctx); - return NULL; - } - - kv->value.arr.data = calloc(kv->value.arr.n, gguf_type_size(kv->value.arr.type)); - if (!kv->value.arr.data) { - fprintf(stderr, "%s: failed to allocate memory for array\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - ok = ok && gguf_fread_el(file, kv->value.arr.data, kv->value.arr.n * gguf_type_size(kv->value.arr.type), &offset); - } break; - case GGUF_TYPE_STRING: - { - // prevent from integer overflow in the malloc below - if (kv->value.arr.n >= SIZE_MAX/sizeof(struct gguf_str)) { - fprintf(stderr, "%s: array size is too large (%" PRIu64 ")\n", __func__, kv->value.arr.n); - fclose(file); - gguf_free(ctx); - return NULL; - } - - kv->value.arr.data = calloc(kv->value.arr.n, sizeof(struct gguf_str)); - if (!kv->value.arr.data) { - fprintf(stderr, "%s: failed to allocate memory for array\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - for (uint64_t j = 0; j < kv->value.arr.n; ++j) { - ok = ok && gguf_fread_str(file, &((struct gguf_str *) kv->value.arr.data)[j], &offset); - } - } break; - case GGUF_TYPE_ARRAY: - default: - { - fprintf(stderr, "%s: invalid array type %d\n", __func__, kv->value.arr.type); - ok = false; - } break; - } - } break; - default: - { - fprintf(stderr, "%s: invalid type %d\n", __func__, kv->type); - ok = false; - } break; - } - - if (!ok) { - break; - } - } - - if (!ok) { - fprintf(stderr, "%s: failed to read key-value pairs\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - } - - // read the tensor infos - if (ctx->header.n_tensors > 0) { - ctx->infos = calloc(ctx->header.n_tensors, sizeof(struct gguf_tensor_info)); - if (!ctx->infos) { - fprintf(stderr, "%s: failed to allocate memory for tensor infos\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - for (uint64_t i = 0; i < ctx->header.n_tensors; ++i) { - struct gguf_tensor_info * info = &ctx->infos[i]; - - for (int j = 0; j < GGML_MAX_DIMS; ++j) { - info->ne[j] = 1; - } - - ok = ok && gguf_fread_str(file, &info->name, &offset); - ok = ok && gguf_fread_el (file, &info->n_dims, sizeof(info->n_dims), &offset); - - ok = ok && (info->n_dims <= GGML_MAX_DIMS); - - for (uint32_t j = 0; j < info->n_dims; ++j) { - ok = ok && gguf_fread_el(file, &info->ne[j], sizeof(info->ne[j]), &offset); - } - - ok = ok && gguf_fread_el (file, &info->type, sizeof(info->type), &offset); - ok = ok && gguf_fread_el (file, &info->offset, sizeof(info->offset), &offset); - - ok = ok && gguf_tensor_info_sanitize(info); - - // make sure there is no duplicated tensor names - for (uint64_t j = 0; j < i && ok; ++j) { - if (strcmp(info->name.data, ctx->infos[j].name.data) == 0) { - fprintf(stderr, "%s: duplicated tensor name %s\n", __func__, info->name.data); - ok = false; - } - } - - if (!ok) { - fprintf(stderr, "%s: failed to read tensor info\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - } - } - - ctx->alignment = GGUF_DEFAULT_ALIGNMENT; - - int alignment_idx = gguf_find_key(ctx, "general.alignment"); - if (alignment_idx != -1) { - ctx->alignment = gguf_get_val_u32(ctx, alignment_idx); - } - - // we require the data section to be aligned, so take into account any padding - { - const size_t offset_pad = offset % ctx->alignment; - - if (offset_pad != 0) { - offset += ctx->alignment - offset_pad; - fseek(file, offset, SEEK_SET); - } - } - - // store the current file offset - this is where the data section starts - ctx->offset = offset; - - // compute the total size of the data section, taking into account the alignment - { - ctx->size = 0; - for (uint64_t i = 0; i < ctx->header.n_tensors; ++i) { - struct gguf_tensor_info * info = &ctx->infos[i]; - - const int64_t ne = - (int64_t) info->ne[0] * - (int64_t) info->ne[1] * - (int64_t) info->ne[2] * - (int64_t) info->ne[3]; - - if (ggml_blck_size(info->type) == 0 || ne % ggml_blck_size(info->type) != 0) { - fprintf(stderr, "%s: tensor '%s' of type %d (%s) number of elements (%" PRId64 ") is not a multiple of block size (%" PRId64 ")\n", - __func__, info->name.data, (int) info->type, ggml_type_name(info->type), ne, ggml_blck_size(info->type)); - fclose(file); - gguf_free(ctx); - return NULL; - } - - const size_t size_cur = ggml_row_size(info->type, ne); - - ctx->size += GGML_PAD(size_cur, ctx->alignment); - } - } - - // load the tensor data only if requested - if (params.ctx != NULL) { - // if the provided gguf_context is no_alloc, then we create "empty" tensors and do not read the binary blob - // otherwise, we load the binary blob into the created ggml_context as well, and point the "data" members of - // the ggml_tensor structs to the appropriate locations in the binary blob - - // compute the exact size needed for the new ggml_context - const size_t mem_size = - params.no_alloc ? - (ctx->header.n_tensors )*ggml_tensor_overhead() : - (ctx->header.n_tensors + 1)*ggml_tensor_overhead() + ctx->size; - - struct ggml_init_params pdata = { - .mem_size = mem_size, - .mem_buffer = NULL, - .no_alloc = params.no_alloc, - }; - - *params.ctx = ggml_init(pdata); - if (*params.ctx == NULL) { - fprintf(stderr, "%s: failed to initialize context\n", __func__); - fclose(file); - gguf_free(ctx); - return NULL; - } - - struct ggml_context * ctx_data = *params.ctx; - - struct ggml_tensor * data = NULL; - - if (!params.no_alloc) { - data = ggml_new_tensor_1d(ctx_data, GGML_TYPE_I8, ctx->size); - - ok = ok && data != NULL; - - // read the binary blob with the tensor data - ok = ok && gguf_fread_el(file, data->data, ctx->size, &offset); - - if (!ok) { - fprintf(stderr, "%s: failed to read tensor data\n", __func__); - fclose(file); - ggml_free(ctx_data); - gguf_free(ctx); - return NULL; - } - - ctx->data = data->data; - } - - ggml_set_no_alloc(ctx_data, true); - - // create the tensors - for (uint64_t i = 0; i < ctx->header.n_tensors; ++i) { - const int64_t ne[GGML_MAX_DIMS] = { - ctx->infos[i].ne[0], - ctx->infos[i].ne[1], - ctx->infos[i].ne[2], - ctx->infos[i].ne[3], - }; - - struct ggml_tensor * cur = ggml_new_tensor(ctx_data, ctx->infos[i].type, ctx->infos[i].n_dims, ne); - - ok = ok && cur != NULL; - - if (!ok) { - break; - } - - ggml_set_name(cur, ctx->infos[i].name.data); - - // point the data member to the appropriate location in the binary blob using the tensor infos - if (!params.no_alloc) { - //cur->data = (char *) data->data + ctx->infos[i].offset - ctx->offset; // offset from start of file - cur->data = (char *) data->data + ctx->infos[i].offset; // offset from data - } - } - - if (!ok) { - fprintf(stderr, "%s: failed to read the tensor data\n", __func__); - fclose(file); - ggml_free(ctx_data); - gguf_free(ctx); - return NULL; - } - - ggml_set_no_alloc(ctx_data, params.no_alloc); - } - - fclose(file); - - return ctx; -} - -void gguf_free(struct gguf_context * ctx) { - if (ctx == NULL) { - return; - } - - if (ctx->kv) { - // free string memory - not great.. - for (uint64_t i = 0; i < ctx->header.n_kv; ++i) { - gguf_free_kv(&ctx->kv[i]); - } - - GGML_FREE(ctx->kv); - } - - if (ctx->infos) { - for (uint64_t i = 0; i < ctx->header.n_tensors; ++i) { - struct gguf_tensor_info * info = &ctx->infos[i]; - - if (info->name.data) { - GGML_FREE(info->name.data); - } - } - - GGML_FREE(ctx->infos); - } - - GGML_FREE(ctx); -} - -const char * gguf_type_name(enum gguf_type type) { - return GGUF_TYPE_NAME[type]; -} - -int gguf_get_version(const struct gguf_context * ctx) { - return ctx->header.version; -} - -size_t gguf_get_alignment(const struct gguf_context * ctx) { - return ctx->alignment; -} - -size_t gguf_get_data_offset(const struct gguf_context * ctx) { - return ctx->offset; -} - -void * gguf_get_data(const struct gguf_context * ctx) { - return ctx->data; -} - -int gguf_get_n_kv(const struct gguf_context * ctx) { - return ctx->header.n_kv; -} - -int gguf_find_key(const struct gguf_context * ctx, const char * key) { - // return -1 if key not found - int keyfound = -1; - - const int n_kv = gguf_get_n_kv(ctx); - - for (int i = 0; i < n_kv; ++i) { - if (strcmp(key, gguf_get_key(ctx, i)) == 0) { - keyfound = i; - break; - } - } - - return keyfound; -} - -const char * gguf_get_key(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - return ctx->kv[key_id].key.data; -} - -enum gguf_type gguf_get_kv_type(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - return ctx->kv[key_id].type; -} - -enum gguf_type gguf_get_arr_type(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY); - return ctx->kv[key_id].value.arr.type; -} - -const void * gguf_get_arr_data(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY); - return ctx->kv[key_id].value.arr.data; -} - -const char * gguf_get_arr_str(const struct gguf_context * ctx, int key_id, int i) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY); - struct gguf_kv * kv = &ctx->kv[key_id]; - struct gguf_str * str = &((struct gguf_str *) kv->value.arr.data)[i]; - return str->data; -} - -int gguf_get_arr_n(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY); - return ctx->kv[key_id].value.arr.n; -} - -uint8_t gguf_get_val_u8(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_UINT8); - return ctx->kv[key_id].value.uint8; -} - -int8_t gguf_get_val_i8(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_INT8); - return ctx->kv[key_id].value.int8; -} - -uint16_t gguf_get_val_u16(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_UINT16); - return ctx->kv[key_id].value.uint16; -} - -int16_t gguf_get_val_i16(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_INT16); - return ctx->kv[key_id].value.int16; -} - -uint32_t gguf_get_val_u32(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_UINT32); - return ctx->kv[key_id].value.uint32; -} - -int32_t gguf_get_val_i32(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_INT32); - return ctx->kv[key_id].value.int32; -} - -float gguf_get_val_f32(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_FLOAT32); - return ctx->kv[key_id].value.float32; -} - -uint64_t gguf_get_val_u64(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_UINT64); - return ctx->kv[key_id].value.uint64; -} - -int64_t gguf_get_val_i64(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_INT64); - return ctx->kv[key_id].value.int64; -} - -double gguf_get_val_f64(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_FLOAT64); - return ctx->kv[key_id].value.float64; -} - -bool gguf_get_val_bool(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_BOOL); - return ctx->kv[key_id].value.bool_; -} - -const char * gguf_get_val_str(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_STRING); - return ctx->kv[key_id].value.str.data; -} - -const void * gguf_get_val_data(const struct gguf_context * ctx, int key_id) { - GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); - GGML_ASSERT(ctx->kv[key_id].type != GGUF_TYPE_ARRAY); - GGML_ASSERT(ctx->kv[key_id].type != GGUF_TYPE_STRING); - return &ctx->kv[key_id].value; -} - -int gguf_get_n_tensors(const struct gguf_context * ctx) { - return ctx->header.n_tensors; -} - -int gguf_find_tensor(const struct gguf_context * ctx, const char * name) { - // return -1 if tensor not found - int tensorfound = -1; - - const int n_tensors = gguf_get_n_tensors(ctx); - - for (int i = 0; i < n_tensors; ++i) { - if (strcmp(name, gguf_get_tensor_name(ctx, i)) == 0) { - tensorfound = i; - break; - } - } - - return tensorfound; -} - -size_t gguf_get_tensor_offset(const struct gguf_context * ctx, int i) { - return ctx->infos[i].offset; -} - -char * gguf_get_tensor_name(const struct gguf_context * ctx, int i) { - return ctx->infos[i].name.data; -} - -enum ggml_type gguf_get_tensor_type(const struct gguf_context * ctx, int i) { - return ctx->infos[i].type; -} - -// returns the index -static int gguf_get_or_add_key(struct gguf_context * ctx, const char * key) { - const int idx = gguf_find_key(ctx, key); - if (idx >= 0) { - return idx; - } - - const int n_kv = gguf_get_n_kv(ctx); - - ctx->kv = realloc(ctx->kv, (n_kv + 1) * sizeof(struct gguf_kv)); - ctx->kv[n_kv].key.n = strlen(key); - ctx->kv[n_kv].key.data = strdup(key); - ctx->header.n_kv++; - - return n_kv; -} - -void gguf_remove_key(struct gguf_context * ctx, const char * key) { - const int idx = gguf_find_key(ctx, key); - if (idx >= 0) { - const int n_kv = gguf_get_n_kv(ctx); - gguf_free_kv(&ctx->kv[idx]); - for (int i = idx; i < n_kv-1; ++i) { - ctx->kv[i] = ctx->kv[i+1]; - } - ctx->kv = realloc(ctx->kv, (n_kv - 1) * sizeof(struct gguf_kv)); - ctx->header.n_kv--; - } -} - -void gguf_set_val_u8(struct gguf_context * ctx, const char * key, uint8_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_UINT8; - ctx->kv[idx].value.uint8 = val; -} - -void gguf_set_val_i8(struct gguf_context * ctx, const char * key, int8_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_INT8; - ctx->kv[idx].value.int8 = val; -} - -void gguf_set_val_u16(struct gguf_context * ctx, const char * key, uint16_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_UINT16; - ctx->kv[idx].value.uint16 = val; -} - -void gguf_set_val_i16(struct gguf_context * ctx, const char * key, int16_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_INT16; - ctx->kv[idx].value.int16 = val; -} - -void gguf_set_val_u32(struct gguf_context * ctx, const char * key, uint32_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_UINT32; - ctx->kv[idx].value.uint32 = val; -} - -void gguf_set_val_i32(struct gguf_context * ctx, const char * key, int32_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_INT32; - ctx->kv[idx].value.int32 = val; -} - -void gguf_set_val_f32(struct gguf_context * ctx, const char * key, float val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_FLOAT32; - ctx->kv[idx].value.float32 = val; -} - -void gguf_set_val_u64(struct gguf_context * ctx, const char * key, uint64_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_UINT64; - ctx->kv[idx].value.uint64 = val; -} - -void gguf_set_val_i64(struct gguf_context * ctx, const char * key, int64_t val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_INT64; - ctx->kv[idx].value.int64 = val; -} - -void gguf_set_val_f64(struct gguf_context * ctx, const char * key, double val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_FLOAT64; - ctx->kv[idx].value.float64 = val; -} - -void gguf_set_val_bool(struct gguf_context * ctx, const char * key, bool val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_BOOL; - ctx->kv[idx].value.bool_ = val; -} - -void gguf_set_val_str(struct gguf_context * ctx, const char * key, const char * val) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_STRING; - ctx->kv[idx].value.str.n = strlen(val); - ctx->kv[idx].value.str.data = strdup(val); -} - -void gguf_set_arr_data(struct gguf_context * ctx, const char * key, enum gguf_type type, const void * data, int n) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_ARRAY; - ctx->kv[idx].value.arr.type = type; - ctx->kv[idx].value.arr.n = n; - ctx->kv[idx].value.arr.data = GGML_CALLOC(n, gguf_type_size(type)); - memcpy(ctx->kv[idx].value.arr.data, data, n*gguf_type_size(type)); -} - -void gguf_set_arr_str(struct gguf_context * ctx, const char * key, const char ** data, int n) { - const int idx = gguf_get_or_add_key(ctx, key); - - ctx->kv[idx].type = GGUF_TYPE_ARRAY; - ctx->kv[idx].value.arr.type = GGUF_TYPE_STRING; - ctx->kv[idx].value.arr.n = n; - ctx->kv[idx].value.arr.data = GGML_CALLOC(n, sizeof(struct gguf_str)); - for (int i = 0; i < n; i++) { - struct gguf_str * str = &((struct gguf_str *)ctx->kv[idx].value.arr.data)[i]; - str->n = strlen(data[i]); - str->data = strdup(data[i]); - } -} - -// set or add KV pairs from another context -void gguf_set_kv(struct gguf_context * ctx, struct gguf_context * src) { - for (uint32_t i = 0; i < src->header.n_kv; i++) { - switch (src->kv[i].type) { - case GGUF_TYPE_UINT8: gguf_set_val_u8 (ctx, src->kv[i].key.data, src->kv[i].value.uint8); break; - case GGUF_TYPE_INT8: gguf_set_val_i8 (ctx, src->kv[i].key.data, src->kv[i].value.int8); break; - case GGUF_TYPE_UINT16: gguf_set_val_u16 (ctx, src->kv[i].key.data, src->kv[i].value.uint16); break; - case GGUF_TYPE_INT16: gguf_set_val_i16 (ctx, src->kv[i].key.data, src->kv[i].value.int16); break; - case GGUF_TYPE_UINT32: gguf_set_val_u32 (ctx, src->kv[i].key.data, src->kv[i].value.uint32); break; - case GGUF_TYPE_INT32: gguf_set_val_i32 (ctx, src->kv[i].key.data, src->kv[i].value.int32); break; - case GGUF_TYPE_FLOAT32: gguf_set_val_f32 (ctx, src->kv[i].key.data, src->kv[i].value.float32); break; - case GGUF_TYPE_UINT64: gguf_set_val_u64 (ctx, src->kv[i].key.data, src->kv[i].value.uint64); break; - case GGUF_TYPE_INT64: gguf_set_val_i64 (ctx, src->kv[i].key.data, src->kv[i].value.int64); break; - case GGUF_TYPE_FLOAT64: gguf_set_val_f64 (ctx, src->kv[i].key.data, src->kv[i].value.float64); break; - case GGUF_TYPE_BOOL: gguf_set_val_bool(ctx, src->kv[i].key.data, src->kv[i].value.bool_); break; - case GGUF_TYPE_STRING: gguf_set_val_str (ctx, src->kv[i].key.data, src->kv[i].value.str.data); break; - case GGUF_TYPE_ARRAY: - { - if (src->kv[i].value.arr.type == GGUF_TYPE_STRING) { - const char ** data = GGML_CALLOC(src->kv[i].value.arr.n, sizeof(char *)); - for (uint32_t j = 0; j < src->kv[i].value.arr.n; j++) { - data[j] = ((struct gguf_str *)src->kv[i].value.arr.data)[j].data; - } - gguf_set_arr_str(ctx, src->kv[i].key.data, data, src->kv[i].value.arr.n); - GGML_FREE((void *)data); - } else if (src->kv[i].value.arr.type == GGUF_TYPE_ARRAY) { - GGML_ABORT("nested arrays not supported"); - } else { - gguf_set_arr_data(ctx, src->kv[i].key.data, src->kv[i].value.arr.type, src->kv[i].value.arr.data, src->kv[i].value.arr.n); - } - } break; - default: GGML_ABORT("invalid type"); - } - } -} - -void gguf_add_tensor( - struct gguf_context * ctx, - const struct ggml_tensor * tensor) { - GGML_ASSERT(tensor); - if (gguf_find_tensor(ctx, tensor->name) != -1) { - GGML_ABORT("duplicated tensor name"); - } - - const int idx = ctx->header.n_tensors; - ctx->infos = realloc(ctx->infos, (idx + 1)*sizeof(struct gguf_tensor_info)); - - ctx->infos[idx].name.n = strlen(tensor->name); - ctx->infos[idx].name.data = strdup(tensor->name); - - for (int i = 0; i < GGML_MAX_DIMS; ++i) { - ctx->infos[idx].ne[i] = 1; - } - - ctx->infos[idx].n_dims = ggml_n_dims(tensor); - for (uint32_t i = 0; i < ctx->infos[idx].n_dims; i++) { - ctx->infos[idx].ne[i] = tensor->ne[i]; - } - - ctx->infos[idx].type = tensor->type; - ctx->infos[idx].offset = 0; - ctx->infos[idx].data = tensor->data; - ctx->infos[idx].size = ggml_nbytes(tensor); - - if (ctx->header.n_tensors > 0) { - ctx->infos[idx].offset = ctx->infos[idx - 1].offset + GGML_PAD(ctx->infos[idx - 1].size, ctx->alignment); - } - - ctx->header.n_tensors++; -} - -void gguf_set_tensor_type(struct gguf_context * ctx, const char * name, enum ggml_type type) { - const int idx = gguf_find_tensor(ctx, name); - if (idx < 0) { - GGML_ABORT("tensor not found"); - } - - ctx->infos[idx].type = type; -} - -void gguf_set_tensor_data(struct gguf_context * ctx, const char * name, const void * data, size_t size) { - const int idx = gguf_find_tensor(ctx, name); - if (idx < 0) { - GGML_ABORT("tensor not found"); - } - - ctx->infos[idx].data = data; - ctx->infos[idx].size = size; - - // update offsets - for (uint32_t i = idx + 1; i < ctx->header.n_tensors; ++i) { - ctx->infos[i].offset = ctx->infos[i - 1].offset + GGML_PAD(ctx->infos[i - 1].size, ctx->alignment); - } -} - -//static void gguf_fwrite_str(FILE * file, const struct gguf_str * val) { -// fwrite(&val->n, sizeof(val->n), 1, file); -// fwrite(val->data, sizeof(char), val->n, file); -//} -// -//static void gguf_fwrite_el(FILE * file, const void * val, size_t size) { -// fwrite(val, sizeof(char), size, file); -//} - -struct gguf_buf { - void * data; - size_t size; - size_t offset; -}; - -static struct gguf_buf gguf_buf_init(size_t size) { - struct gguf_buf buf = { - /*buf.data =*/ size == 0 ? NULL : GGML_CALLOC(1, size), - /*buf.size =*/ size, - /*buf.offset =*/ 0, - }; - - return buf; -} - -static void gguf_buf_free(struct gguf_buf buf) { - if (buf.data) { - GGML_FREE(buf.data); - } -} - -static void gguf_buf_grow(struct gguf_buf * buf, size_t size) { - if (buf->offset + size > buf->size) { - buf->size = 1.5*(buf->offset + size); - if (buf->data) { - buf->data = realloc(buf->data, buf->size); - } - } -} - -static void gguf_bwrite_str(struct gguf_buf * buf, const struct gguf_str * val) { - gguf_buf_grow(buf, sizeof(val->n) + val->n); - - if (buf->data) { - memcpy((char *) buf->data + buf->offset, &val->n, sizeof(val->n)); - } - buf->offset += sizeof(val->n); - - if (buf->data) { - memcpy((char *) buf->data + buf->offset, val->data, val->n); - } - buf->offset += val->n; -} - -static void gguf_bwrite_el(struct gguf_buf * buf, const void * val, size_t el_size) { - gguf_buf_grow(buf, el_size); - - if (buf->data) { - memcpy((char *) buf->data + buf->offset, val, el_size); - } - buf->offset += el_size; -} - -static void gguf_write_to_buf(const struct gguf_context * ctx, struct gguf_buf * buf, bool only_meta) { - // write header - gguf_bwrite_el(buf, &ctx->header.magic, sizeof(ctx->header.magic)); - gguf_bwrite_el(buf, &ctx->header.version, sizeof(ctx->header.version)); - gguf_bwrite_el(buf, &ctx->header.n_tensors, sizeof(ctx->header.n_tensors)); - gguf_bwrite_el(buf, &ctx->header.n_kv, sizeof(ctx->header.n_kv)); - - // write key-value pairs - for (uint32_t i = 0; i < ctx->header.n_kv; ++i) { - struct gguf_kv * kv = &ctx->kv[i]; - - gguf_bwrite_str(buf, &kv->key); - gguf_bwrite_el (buf, &kv->type, sizeof(kv->type)); - - switch (kv->type) { - case GGUF_TYPE_UINT8: gguf_bwrite_el( buf, &kv->value.uint8, sizeof(kv->value.uint8) ); break; - case GGUF_TYPE_INT8: gguf_bwrite_el (buf, &kv->value.int8, sizeof(kv->value.int8) ); break; - case GGUF_TYPE_UINT16: gguf_bwrite_el (buf, &kv->value.uint16, sizeof(kv->value.uint16) ); break; - case GGUF_TYPE_INT16: gguf_bwrite_el (buf, &kv->value.int16, sizeof(kv->value.int16) ); break; - case GGUF_TYPE_UINT32: gguf_bwrite_el (buf, &kv->value.uint32, sizeof(kv->value.uint32) ); break; - case GGUF_TYPE_INT32: gguf_bwrite_el (buf, &kv->value.int32, sizeof(kv->value.int32) ); break; - case GGUF_TYPE_FLOAT32: gguf_bwrite_el (buf, &kv->value.float32, sizeof(kv->value.float32)); break; - case GGUF_TYPE_UINT64: gguf_bwrite_el (buf, &kv->value.uint64, sizeof(kv->value.uint64) ); break; - case GGUF_TYPE_INT64: gguf_bwrite_el (buf, &kv->value.int64, sizeof(kv->value.int64) ); break; - case GGUF_TYPE_FLOAT64: gguf_bwrite_el (buf, &kv->value.float64, sizeof(kv->value.float64)); break; - case GGUF_TYPE_BOOL: gguf_bwrite_el (buf, &kv->value.bool_, sizeof(kv->value.bool_) ); break; - case GGUF_TYPE_STRING: gguf_bwrite_str(buf, &kv->value.str ); break; - case GGUF_TYPE_ARRAY: - { - gguf_bwrite_el(buf, &kv->value.arr.type, sizeof(kv->value.arr.type)); - gguf_bwrite_el(buf, &kv->value.arr.n, sizeof(kv->value.arr.n) ); - - switch (kv->value.arr.type) { - case GGUF_TYPE_UINT8: - case GGUF_TYPE_INT8: - case GGUF_TYPE_UINT16: - case GGUF_TYPE_INT16: - case GGUF_TYPE_UINT32: - case GGUF_TYPE_INT32: - case GGUF_TYPE_FLOAT32: - case GGUF_TYPE_UINT64: - case GGUF_TYPE_INT64: - case GGUF_TYPE_FLOAT64: - case GGUF_TYPE_BOOL: - { - gguf_bwrite_el(buf, kv->value.arr.data, kv->value.arr.n * gguf_type_size(kv->value.arr.type)); - } break; - case GGUF_TYPE_STRING: - { - for (uint32_t j = 0; j < kv->value.arr.n; ++j) { - gguf_bwrite_str(buf, &((struct gguf_str *) kv->value.arr.data)[j]); - } - } break; - case GGUF_TYPE_ARRAY: - default: GGML_ABORT("invalid type"); - } - } break; - default: GGML_ABORT("invalid type"); - } - } - - // write tensor infos - for (uint32_t i = 0; i < ctx->header.n_tensors; ++i) { - struct gguf_tensor_info * info = &ctx->infos[i]; - - gguf_bwrite_str(buf, &info->name); - gguf_bwrite_el (buf, &info->n_dims, sizeof(info->n_dims)); - for (uint32_t j = 0; j < info->n_dims; ++j) { - gguf_bwrite_el(buf, &info->ne[j], sizeof(info->ne[j])); - } - gguf_bwrite_el(buf, &info->type, sizeof(info->type)); - gguf_bwrite_el(buf, &info->offset, sizeof(info->offset)); - } - - // we require the data section to be aligned, so take into account any padding - { - const size_t offset = buf->offset; - const size_t offset_pad = GGML_PAD(offset, ctx->alignment); - - if (offset_pad != offset) { - uint8_t pad = 0; - for (size_t i = 0; i < offset_pad - offset; ++i) { - gguf_bwrite_el(buf, &pad, sizeof(pad)); - } - } - } - - if (only_meta) { - return; - } - - size_t offset = 0; - - // write tensor data - for (uint32_t i = 0; i < ctx->header.n_tensors; ++i) { - struct gguf_tensor_info * info = &ctx->infos[i]; - - const size_t size = info->size; - const size_t size_pad = GGML_PAD(size, ctx->alignment); - - gguf_bwrite_el(buf, info->data, size); - - if (size_pad != size) { - uint8_t pad = 0; - for (size_t j = 0; j < size_pad - size; ++j) { - gguf_bwrite_el(buf, &pad, sizeof(pad)); - } - } - - GGML_ASSERT(offset == info->offset); - - offset += size_pad; - } -} - -void gguf_write_to_file(const struct gguf_context * ctx, const char * fname, bool only_meta) { - FILE * file = ggml_fopen(fname, "wb"); - if (!file) { - GGML_ABORT("failed to open file for writing"); - } - - struct gguf_buf buf = gguf_buf_init(16*1024); - - gguf_write_to_buf(ctx, &buf, only_meta); - - fwrite(buf.data, 1, buf.offset, file); - - gguf_buf_free(buf); - - fclose(file); -} - -size_t gguf_get_meta_size(const struct gguf_context * ctx) { - // no allocs - only compute size - struct gguf_buf buf = gguf_buf_init(0); - - gguf_write_to_buf(ctx, &buf, true); - - return buf.offset; -} - -void gguf_get_meta_data(const struct gguf_context * ctx, void * data) { - struct gguf_buf buf = gguf_buf_init(16*1024); - - gguf_write_to_buf(ctx, &buf, true); - - memcpy(data, buf.data, buf.offset); - - gguf_buf_free(buf); -} - -//////////////////////////////////////////////////////////////////////////////// - -int ggml_cpu_has_avx(void) { -#if defined(__AVX__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx_vnni(void) { -#if defined(__AVXVNNI__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx2(void) { -#if defined(__AVX2__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx512(void) { -#if defined(__AVX512F__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx512_vbmi(void) { -#if defined(__AVX512VBMI__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx512_vnni(void) { -#if defined(__AVX512VNNI__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_avx512_bf16(void) { -#if defined(__AVX512BF16__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_amx_int8(void) { -#if defined(__AMX_INT8__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_fma(void) { -#if defined(__FMA__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_arm_fma(void) { -#if defined(__ARM_FEATURE_FMA) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_riscv_v(void) { -#if defined(__riscv_v_intrinsic) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_metal(void) { -#if defined(GGML_USE_METAL) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_f16c(void) { -#if defined(__F16C__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_fp16_va(void) { -#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_wasm_simd(void) { -#if defined(__wasm_simd128__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_blas(void) { -#if defined(GGML_USE_BLAS) || defined(GGML_USE_CUDA) || defined(GGML_USE_VULKAN) || defined(GGML_USE_SYCL) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_cuda(void) { -#if defined(GGML_USE_CUDA) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_vulkan(void) { -#if defined(GGML_USE_VULKAN) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_kompute(void) { -#if defined(GGML_USE_KOMPUTE) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_sycl(void) { -#if defined(GGML_USE_SYCL) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_rpc(void) { -#if defined(GGML_USE_RPC) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_cann(void) { -#if defined(GGML_USE_CANN) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_llamafile(void) { -#if defined(GGML_USE_LLAMAFILE) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_gpublas(void) { - return ggml_cpu_has_cuda() || ggml_cpu_has_vulkan() || ggml_cpu_has_kompute() || ggml_cpu_has_sycl(); -} - -int ggml_cpu_has_sse3(void) { -#if defined(__SSE3__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_ssse3(void) { -#if defined(__SSSE3__) - return 1; -#else - return 0; -#endif -} - -int ggml_cpu_has_vsx(void) { -#if defined(__POWER9_VECTOR__) - return 1; -#else - return 0; -#endif -} - void ggml_log_set(ggml_log_callback log_callback, void * user_data) { g_logger_state.log_callback = log_callback ? log_callback : ggml_log_callback_default; g_logger_state.log_callback_user_data = user_data; } -//////////////////////////////////////////////////////////////////////////////// + +void ggml_threadpool_params_init(struct ggml_threadpool_params * p, int n_threads) { + p->n_threads = n_threads; + p->prio = 0; // default priority (usually means normal or inherited) + p->poll = 50; // hybrid-polling enabled + p->strict_cpu = false; // no strict placement (all threads share same cpumask) + p->paused = false; // threads are ready to go + memset(p->cpumask, 0, GGML_MAX_N_THREADS); // all-zero means use the default affinity (usually inherited) +} + +struct ggml_threadpool_params ggml_threadpool_params_default(int n_threads) { + struct ggml_threadpool_params p; + ggml_threadpool_params_init(&p, n_threads); + return p; +} + +bool ggml_threadpool_params_match(const struct ggml_threadpool_params * p0, const struct ggml_threadpool_params * p1) { + if (p0->n_threads != p1->n_threads ) return false; + if (p0->prio != p1->prio ) return false; + if (p0->poll != p1->poll ) return false; + if (p0->strict_cpu != p1->strict_cpu ) return false; + return memcmp(p0->cpumask, p1->cpumask, GGML_MAX_N_THREADS) == 0; +} diff --git a/ggml/src/gguf.cpp b/ggml/src/gguf.cpp new file mode 100644 index 000000000..655ed600a --- /dev/null +++ b/ggml/src/gguf.cpp @@ -0,0 +1,1325 @@ +#include "ggml.h" +#include "ggml-backend.h" +#include "ggml-impl.h" +#include "gguf.h" + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +template +struct type_to_gguf_type; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_UINT8; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_INT8; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_UINT16; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_INT16; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_UINT32; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_INT32; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_FLOAT32; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_BOOL; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_STRING; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_UINT64; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_INT64; +}; + +template <> +struct type_to_gguf_type { + static constexpr enum gguf_type value = GGUF_TYPE_FLOAT64; +}; + +static const std::map GGUF_TYPE_SIZE = { + {GGUF_TYPE_UINT8, sizeof(uint8_t)}, + {GGUF_TYPE_INT8, sizeof(int8_t)}, + {GGUF_TYPE_UINT16, sizeof(uint16_t)}, + {GGUF_TYPE_INT16, sizeof(int16_t)}, + {GGUF_TYPE_UINT32, sizeof(uint32_t)}, + {GGUF_TYPE_INT32, sizeof(int32_t)}, + {GGUF_TYPE_FLOAT32, sizeof(float)}, + {GGUF_TYPE_BOOL, sizeof(int8_t)}, + {GGUF_TYPE_STRING, 0}, // undefined + {GGUF_TYPE_ARRAY, 0}, // undefined + {GGUF_TYPE_UINT64, sizeof(uint64_t)}, + {GGUF_TYPE_INT64, sizeof(int64_t)}, + {GGUF_TYPE_FLOAT64, sizeof(double)}, +}; +static_assert(GGUF_TYPE_COUNT == 13, "GGUF_TYPE_COUNT != 13"); + +static const std::map GGUF_TYPE_NAME = { + {GGUF_TYPE_UINT8, "u8"}, + {GGUF_TYPE_INT8, "i8"}, + {GGUF_TYPE_UINT16, "u16"}, + {GGUF_TYPE_INT16, "i16"}, + {GGUF_TYPE_UINT32, "u32"}, + {GGUF_TYPE_INT32, "i32"}, + {GGUF_TYPE_FLOAT32, "f32"}, + {GGUF_TYPE_BOOL, "bool"}, + {GGUF_TYPE_STRING, "str"}, + {GGUF_TYPE_ARRAY, "arr"}, + {GGUF_TYPE_UINT64, "u64"}, + {GGUF_TYPE_INT64, "i64"}, + {GGUF_TYPE_FLOAT64, "f64"}, +}; +static_assert(GGUF_TYPE_COUNT == 13, "GGUF_TYPE_COUNT != 13"); + +size_t gguf_type_size(enum gguf_type type) { + auto it = GGUF_TYPE_SIZE.find(type); + return it == GGUF_TYPE_SIZE.end() ? 0 : it->second; +} + +struct gguf_kv { + std::string key; + + bool is_array; + enum gguf_type type; + + std::vector data; + std::vector data_string; + + template + gguf_kv(const std::string & key, const T value) + : key(key), is_array(false), type(type_to_gguf_type::value) { + GGML_ASSERT(!key.empty()); + data.resize(sizeof(T)); + memcpy(data.data(), &value, sizeof(T)); + } + + template + gguf_kv(const std::string & key, const std::vector & value) + : key(key), is_array(true), type(type_to_gguf_type::value) { + GGML_ASSERT(!key.empty()); + data.resize(value.size()*sizeof(T)); + for (size_t i = 0; i < value.size(); ++i) { + const T tmp = value[i]; + memcpy(data.data() + i*sizeof(T), &tmp, sizeof(T)); + } + } + + gguf_kv(const std::string & key, const std::string & value) + : key(key), is_array(false), type(GGUF_TYPE_STRING) { + GGML_ASSERT(!key.empty()); + data_string.push_back(value); + } + + gguf_kv(const std::string & key, const std::vector & value) + : key(key), is_array(true), type(GGUF_TYPE_STRING) { + GGML_ASSERT(!key.empty()); + data_string = value; + } + + const std::string & get_key() const { + return key; + } + + const enum gguf_type & get_type() const { + return type; + } + + size_t get_ne() const { + if (type == GGUF_TYPE_STRING) { + const size_t ne = data_string.size(); + GGML_ASSERT(is_array || ne == 1); + return ne; + } + const size_t type_size = gguf_type_size(type); + GGML_ASSERT(data.size() % type_size == 0); + const size_t ne = data.size() / type_size; + GGML_ASSERT(is_array || ne == 1); + return ne; + } + + template + const T & get_val(const size_t i = 0) const { + GGML_ASSERT(type_to_gguf_type::value == type); + if constexpr (std::is_same::value) { + GGML_ASSERT(data_string.size() >= i+1); + return data_string[i]; + } + const size_t type_size = gguf_type_size(type); + GGML_ASSERT(data.size() % type_size == 0); + GGML_ASSERT(data.size() >= (i+1)*type_size); + return reinterpret_cast(data.data())[i]; + } + + void cast(const enum gguf_type new_type) { + const size_t new_type_size = gguf_type_size(new_type); + GGML_ASSERT(data.size() % new_type_size == 0); + type = new_type; + } +}; + +struct gguf_tensor_info { + struct ggml_tensor t; // for holding the equivalent info + uint64_t offset; // offset from start of `data`, must be a multiple of `ALIGNMENT` +}; + +struct gguf_context { + uint32_t version = GGUF_VERSION; + + std::vector kv; + std::vector info; + + size_t alignment = GGUF_DEFAULT_ALIGNMENT; + size_t offset = 0; // offset of `data` from beginning of file + size_t size = 0; // size of `data` in bytes + + void * data = nullptr; +}; + +struct gguf_reader { + FILE * file; + + gguf_reader(FILE * file) : file(file) {} + + template + bool read(T & dst) const { + return fread(&dst, 1, sizeof(dst), file) == sizeof(dst); + } + + template + bool read(std::vector & dst, const size_t n) const { + dst.resize(n); + for (size_t i = 0; i < dst.size(); ++i) { + if constexpr (std::is_same::value) { + bool tmp; + if (!read(tmp)) { + return false; + } + dst[i] = tmp; + } else { + if (!read(dst[i])) { + return false; + } + } + } + return true; + } + + bool read(bool & dst) const { + int8_t tmp = -1; + if (!read(tmp)) { + return false; + } + dst = tmp != 0; + return true; + } + + bool read(enum ggml_type & dst) const { + int32_t tmp = -1; + if (!read(tmp)) { + return false; + } + dst = ggml_type(tmp); + return true; + } + + bool read(enum gguf_type & dst) const { + int32_t tmp = -1; + if (!read(tmp)) { + return false; + } + dst = gguf_type(tmp); + return true; + } + + bool read(std::string & dst) const { + uint64_t size = -1; + if (!read(size)) { + return false; + } + dst.resize(size); + return fread(dst.data(), 1, dst.length(), file) == dst.length(); + } + + bool read(void * dst, const size_t size) const { + return fread(dst, 1, size, file) == size; + } +}; + +struct gguf_context * gguf_init_empty(void) { + return new gguf_context; +} + +template +bool gguf_read_emplace_helper(const struct gguf_reader & gr, std::vector & kv, const std::string & key, const bool is_array, const size_t n) { + if (is_array) { + std::vector value; + try { + if (!gr.read(value, n)) { + return false; + } + } catch (std::length_error &) { + fprintf(stderr, "%s: encountered length_error while reading value for key '%s'\n", __func__, key.c_str()); + return false; + } catch (std::bad_alloc &) { + fprintf(stderr, "%s: encountered bad_alloc error while reading value for key '%s'\n", __func__, key.c_str()); + return false; + } + kv.emplace_back(key, value); + } else { + T value; + if (!gr.read(value)) { + return false; + } + kv.emplace_back(key, value); + } + return true; +} + +struct gguf_context * gguf_init_from_file_impl(FILE * file, struct gguf_init_params params) { + const struct gguf_reader gr(file); + struct gguf_context * ctx = new gguf_context; + + bool ok = true; + + // file magic + { + std::vector magic; + ok = ok && gr.read(magic, 4); + + if (!ok) { + fprintf(stderr, "%s: failed to read magic\n", __func__); + gguf_free(ctx); + return nullptr; + } + + for (uint32_t i = 0; i < magic.size(); i++) { + if (magic[i] != GGUF_MAGIC[i]) { + fprintf(stderr, "%s: invalid magic characters: '%c%c%c%c', expected 'GGUF'\n", __func__, magic[0], magic[1], magic[2], magic[3]); + gguf_free(ctx); + return nullptr; + } + } + } + + // header + int64_t n_kv = 0; + int64_t n_tensors = 0; + + if (ok && gr.read(ctx->version)) { + if (ctx->version == 1) { + fprintf(stderr, "%s: GGUFv1 is no longer supported, please use a more up-to-date version\n", __func__); + ok = false; + } + if (ctx->version > GGUF_VERSION) { + fprintf(stderr, "%s: this GGUF file is version %" PRIu32 " but this software only supports up to version %d\n", + __func__, ctx->version, GGUF_VERSION); + ok = false; + } + } else { + ok = false; + } + + if (ok && gr.read(n_tensors)) { + static_assert(sizeof(size_t) <= 8 && sizeof(gguf_tensor_info) >= 2, "int64_t insufficient for indexing"); + if (n_tensors < 0 || n_tensors > int64_t(SIZE_MAX/sizeof(gguf_tensor_info))) { + fprintf(stderr, "%s: number of tensors is %" PRIi64 " but must be in [0, %zu]\n", + __func__, n_tensors, SIZE_MAX/sizeof(gguf_tensor_info)); + ok = false; + } + } else { + ok = false; + } + + if (ok && gr.read(n_kv)) { + static_assert(sizeof(size_t) <= 8 && sizeof(gguf_tensor_info) >= 2, "int64_t insufficient for indexing"); + if (n_kv < 0 || n_kv > int64_t(SIZE_MAX/sizeof(gguf_kv))) { + fprintf(stderr, "%s: number of key value pairs is %" PRIi64 " but must be in [0, %zu]\n", + __func__, n_kv, SIZE_MAX/sizeof(gguf_kv)); + ok = false; + } + } else { + ok = false; + } + + if (!ok) { + fprintf(stderr, "%s: failed to read header\n", __func__); + gguf_free(ctx); + return nullptr; + } + + // KV pairs + { + for (int64_t i = 0; ok && i < n_kv; ++i) { + std::string key; + gguf_type type = gguf_type(-1); + bool is_array = false; + uint64_t n = 1; + + try { + ok = ok && gr.read(key); + } catch (std::length_error &) { + fprintf(stderr, "%s: encountered length_error while reading key %" PRIi64 "\n", __func__, i); + ok = false; + } catch (std::bad_alloc &) { + fprintf(stderr, "%s: encountered bad_alloc error while reading key %" PRIi64 "\n", __func__, i); + ok = false; + } + for (size_t j = 0; ok && j < ctx->kv.size(); ++j) { + if (key == ctx->kv[j].key) { + fprintf(stderr, "%s: duplicate key '%s' for tensors %zu and %" PRIi64 " \n", __func__, key.c_str(), j, i); + ok = false; + } + } + if (!ok) { + break; + } + + ok = ok && gr.read(type); + if (type == GGUF_TYPE_ARRAY) { + is_array = true; + ok = ok && gr.read(type); + ok = ok && gr.read(n); + } + if (!ok) { + break; + } + + switch (type) { + case GGUF_TYPE_UINT8: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_INT8: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_UINT16: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_INT16: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_UINT32: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_INT32: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_FLOAT32: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_BOOL: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_STRING: ok = ok && gguf_read_emplace_helper(gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_UINT64: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_INT64: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_FLOAT64: ok = ok && gguf_read_emplace_helper (gr, ctx->kv, key, is_array, n); break; + case GGUF_TYPE_ARRAY: + default: + { + fprintf(stderr, "%s: key '%s' has invalid GGUF type %d\n", __func__, key.c_str(), type); + ok = false; + } break; + } + } + + if (!ok) { + fprintf(stderr, "%s: failed to read key-value pairs\n", __func__); + gguf_free(ctx); + return nullptr; + } + GGML_ASSERT(int64_t(ctx->kv.size()) == n_kv); + + const int alignment_idx = gguf_find_key(ctx, GGUF_KEY_GENERAL_ALIGNMENT); + ctx->alignment = alignment_idx == -1 ? GGUF_DEFAULT_ALIGNMENT : gguf_get_val_u32(ctx, alignment_idx); + + if (ctx->alignment == 0 || (ctx->alignment & (ctx->alignment - 1)) != 0) { + fprintf(stderr, "%s: alignment %zu is not a power of 2\n", __func__, ctx->alignment); + gguf_free(ctx); + return nullptr; + } + } + + // read the tensor info + for (int64_t i = 0; ok && i < n_tensors; ++i) { + struct gguf_tensor_info info; + + // tensor name + { + std::string name; + try { + ok = ok && gr.read(name); + } catch (std::length_error &) { + fprintf(stderr, "%s: encountered length_error while reading tensor name %" PRIi64 "\n", __func__, i); + ok = false; + } catch (std::bad_alloc &) { + fprintf(stderr, "%s: encountered bad_alloc error while reading tensor name %" PRIi64 "\n", __func__, i); + ok = false; + } + if (name.length() >= GGML_MAX_NAME) { + fprintf(stderr, "%s: tensor name %" PRIi64 " is too long: %zu >= %d\n", __func__, i, name.length(), GGML_MAX_NAME); + ok = false; + break; + } + ggml_set_name(&info.t, name.c_str()); + + // make sure there are no duplicate tensor names + for (int64_t j = 0; ok && j < i; ++j) { + if (strcmp(info.t.name, ctx->info[j].t.name) == 0) { + fprintf(stderr, "%s: duplicate tensor name '%s' for tensors %" PRIi64 " and %" PRIi64 "\n", __func__, info.t.name, j, i); + ok = false; + break; + } + } + } + if (!ok) { + break; + } + + // tensor shape + { + uint32_t n_dims = -1; + ok = ok && gr.read(n_dims); + if (n_dims > GGML_MAX_DIMS) { + fprintf(stderr, "%s: tensor '%s' has invalid number of dimensions: %" PRIu32 " > %" PRIu32 "\n", + __func__, info.t.name, n_dims, GGML_MAX_DIMS); + ok = false; + break; + } + for (uint32_t j = 0; ok && j < GGML_MAX_DIMS; ++j) { + info.t.ne[j] = 1; + if (j < n_dims) { + ok = ok && gr.read(info.t.ne[j]); + } + + // check that all ne are non-negative + if (info.t.ne[j] < 0) { + fprintf(stderr, "%s: tensor '%s' dimension %" PRIu32 " has invalid number of elements: %" PRIi64 " < 0\n", + __func__, info.t.name, j, info.t.ne[j]); + ok = false; + break; + } + } + + // check that the total number of elements is representable + if (ok && ((INT64_MAX/info.t.ne[1] <= info.t.ne[0]) || + (INT64_MAX/info.t.ne[2] <= info.t.ne[0]*info.t.ne[1]) || + (INT64_MAX/info.t.ne[3] <= info.t.ne[0]*info.t.ne[1]*info.t.ne[2]))) { + + fprintf(stderr, "%s: total number of elements in tensor '%s' with shape " + "(%" PRIi64 ", %" PRIi64 ", %" PRIi64 ", %" PRIi64 ") is >= %" PRIi64 "\n", + __func__, info.t.name, info.t.ne[0], info.t.ne[1], info.t.ne[2], info.t.ne[3], INT64_MAX); + ok = false; + break; + } + } + if (!ok) { + break; + } + + // tensor type + { + ok = ok && gr.read(info.t.type); + + // check that tensor type is within defined range + if (info.t.type < 0 || info.t.type >= GGML_TYPE_COUNT) { + fprintf(stderr, "%s: tensor '%s' has invalid ggml type %d (%s)\n", + __func__, info.t.name, info.t.type, ggml_type_name(info.t.type)); + ok = false; + break; + } + const size_t type_size = ggml_type_size(info.t.type); + const int64_t blck_size = ggml_blck_size(info.t.type); + + // check that row size is divisible by block size + if (blck_size == 0 || info.t.ne[0] % blck_size != 0) { + fprintf(stderr, "%s: tensor '%s' of type %d (%s) has %" PRId64 " elements per row, " + "not a multiple of block size (%" PRId64 ")\n", + __func__, info.t.name, (int) info.t.type, ggml_type_name(info.t.type), info.t.ne[0], blck_size); + ok = false; + break; + } + + // calculate byte offsets given the tensor shape and type + info.t.nb[0] = type_size; + info.t.nb[1] = info.t.nb[0]*(info.t.ne[0]/blck_size); + for (int j = 2; j < GGML_MAX_DIMS; ++j) { + info.t.nb[j] = info.t.nb[j - 1]*info.t.ne[j - 1]; + } + } + if (!ok) { + break; + } + + // tensor data offset within buffer + ok = ok && gr.read(info.offset); + + ctx->info.push_back(info); + } + + if (!ok) { + fprintf(stderr, "%s: failed to read tensor info\n", __func__); + gguf_free(ctx); + return nullptr; + } + GGML_ASSERT(int64_t(ctx->info.size()) == n_tensors); + + // we require the data section to be aligned, so take into account any padding + if (fseek(file, GGML_PAD(ftell(file), ctx->alignment), SEEK_SET) != 0) { + fprintf(stderr, "%s: failed to seek to beginning of data section\n", __func__); + gguf_free(ctx); + return nullptr; + } + + // store the current file offset - this is where the data section starts + ctx->offset = ftell(file); + + // compute the total size of the data section, taking into account the alignment + { + ctx->size = 0; + for (size_t i = 0; i < ctx->info.size(); ++i) { + const gguf_tensor_info & ti = ctx->info[i]; + if (ti.offset != ctx->size) { + fprintf(stderr, "%s: tensor '%s' has offset %" PRIu64 ", expected %zu\n", + __func__, ti.t.name, ti.offset, ctx->size); + fprintf(stderr, "%s: failed to read tensor data\n", __func__); + gguf_free(ctx); + return nullptr; + } + ctx->size += GGML_PAD(ggml_nbytes(&ti.t), ctx->alignment); + } + } + + // load the tensor data only if requested + if (params.ctx != nullptr) { + // if the provided gguf_context is no_alloc, then we create "empty" tensors and do not read the binary blob + // otherwise, we load the binary blob into the created ggml_context as well, and point the "data" members of + // the ggml_tensor structs to the appropriate locations in the binary blob + + // compute the exact size needed for the new ggml_context + const size_t mem_size = + params.no_alloc ? + (n_tensors )*ggml_tensor_overhead() : + (n_tensors + 1)*ggml_tensor_overhead() + ctx->size; + + struct ggml_init_params pdata = { + /*mem_size =*/ mem_size, + /*mem_buffer =*/ nullptr, + /*no_alloc =*/ params.no_alloc, + }; + + *params.ctx = ggml_init(pdata); + if (*params.ctx == nullptr) { + fprintf(stderr, "%s: failed to initialize ggml context for storing tensors\n", __func__); + gguf_free(ctx); + return nullptr; + } + + struct ggml_context * ctx_data = *params.ctx; + + struct ggml_tensor * data = nullptr; + + if (!params.no_alloc) { + data = ggml_new_tensor_1d(ctx_data, GGML_TYPE_I8, ctx->size); + + ok = ok && data != nullptr; + + // read the binary blob with the tensor data + ok = ok && gr.read(data->data, ctx->size); + + if (!ok) { + fprintf(stderr, "%s: failed to read tensor data binary blob\n", __func__); + ggml_free(ctx_data); + *params.ctx = nullptr; + gguf_free(ctx); + return nullptr; + } + + ctx->data = data->data; + } + + ggml_set_no_alloc(ctx_data, true); + + // create the tensors + for (size_t i = 0; i < ctx->info.size(); ++i) { + const struct gguf_tensor_info & info = ctx->info[i]; + + struct ggml_tensor * cur = ggml_new_tensor(ctx_data, info.t.type, GGML_MAX_DIMS, info.t.ne); + + ok = ok && cur != nullptr; + + if (!ok) { + break; + } + + ggml_set_name(cur, info.t.name); + + // point the data member to the appropriate location in the binary blob using the tensor info + if (!params.no_alloc) { + cur->data = (char *) data->data + info.offset; + } + } + + if (!ok) { + fprintf(stderr, "%s: failed to create tensors\n", __func__); + ggml_free(ctx_data); + *params.ctx = nullptr; + gguf_free(ctx); + return nullptr; + } + + ggml_set_no_alloc(ctx_data, params.no_alloc); + } + + return ctx; +} + +struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_params params) { + FILE * file = ggml_fopen(fname, "rb"); + + if (!file) { + fprintf(stderr, "%s: failed to open GGUF file '%s'\n", __func__, fname); + return nullptr; + } + + struct gguf_context * result = gguf_init_from_file_impl(file, params); + fclose(file); + return result; +} + +void gguf_free(struct gguf_context * ctx) { + if (ctx == nullptr) { + return; + } + delete ctx; +} + +const char * gguf_type_name(enum gguf_type type) { + auto it = GGUF_TYPE_NAME.find(type); + return it == GGUF_TYPE_NAME.end() ? nullptr : it->second; +} + +uint32_t gguf_get_version(const struct gguf_context * ctx) { + return ctx->version; +} + +size_t gguf_get_alignment(const struct gguf_context * ctx) { + return ctx->alignment; +} + +size_t gguf_get_data_offset(const struct gguf_context * ctx) { + return ctx->offset; +} + +int64_t gguf_get_n_kv(const struct gguf_context * ctx) { + return ctx->kv.size(); +} + +int64_t gguf_find_key(const struct gguf_context * ctx, const char * key) { + // return -1 if key not found + int64_t keyfound = -1; + + const int64_t n_kv = gguf_get_n_kv(ctx); + + for (int64_t i = 0; i < n_kv; ++i) { + if (strcmp(key, gguf_get_key(ctx, i)) == 0) { + keyfound = i; + break; + } + } + + return keyfound; +} + +const char * gguf_get_key(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + return ctx->kv[key_id].get_key().c_str(); +} + +enum gguf_type gguf_get_kv_type(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + return ctx->kv[key_id].is_array ? GGUF_TYPE_ARRAY : ctx->kv[key_id].get_type(); +} + +enum gguf_type gguf_get_arr_type(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].is_array); + return ctx->kv[key_id].get_type(); +} + +const void * gguf_get_arr_data(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_type() != GGUF_TYPE_STRING); + return ctx->kv[key_id].data.data(); +} + +const char * gguf_get_arr_str(const struct gguf_context * ctx, int64_t key_id, size_t i) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_type() == GGUF_TYPE_STRING); + return ctx->kv[key_id].data_string[i].c_str(); +} + +size_t gguf_get_arr_n(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + + if (ctx->kv[key_id].type == GGUF_TYPE_STRING) { + return ctx->kv[key_id].data_string.size(); + } + + const size_t type_size = gguf_type_size(ctx->kv[key_id].type); + GGML_ASSERT(ctx->kv[key_id].data.size() % type_size == 0); + return ctx->kv[key_id].data.size() / type_size; +} + +uint8_t gguf_get_val_u8(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +int8_t gguf_get_val_i8(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +uint16_t gguf_get_val_u16(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +int16_t gguf_get_val_i16(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +uint32_t gguf_get_val_u32(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +int32_t gguf_get_val_i32(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +float gguf_get_val_f32(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +uint64_t gguf_get_val_u64(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +int64_t gguf_get_val_i64(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +double gguf_get_val_f64(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +bool gguf_get_val_bool(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val(); +} + +const char * gguf_get_val_str(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + return ctx->kv[key_id].get_val().c_str(); +} + +const void * gguf_get_val_data(const struct gguf_context * ctx, int64_t key_id) { + GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx)); + GGML_ASSERT(ctx->kv[key_id].get_ne() == 1); + GGML_ASSERT(ctx->kv[key_id].get_type() != GGUF_TYPE_STRING); + return ctx->kv[key_id].data.data(); +} + +int64_t gguf_get_n_tensors(const struct gguf_context * ctx) { + return ctx->info.size(); +} + +int64_t gguf_find_tensor(const struct gguf_context * ctx, const char * name) { + // return -1 if tensor not found + int64_t tensor_id = -1; + + const int64_t n_tensors = gguf_get_n_tensors(ctx); + + for (int64_t i = 0; i < n_tensors; ++i) { + if (strcmp(name, gguf_get_tensor_name(ctx, i)) == 0) { + tensor_id = i; + break; + } + } + + return tensor_id; +} + +size_t gguf_get_tensor_offset(const struct gguf_context * ctx, int64_t tensor_id) { + GGML_ASSERT(tensor_id >= 0 && tensor_id < gguf_get_n_tensors(ctx)); + return ctx->info[tensor_id].offset; +} + +const char * gguf_get_tensor_name(const struct gguf_context * ctx, int64_t tensor_id) { + GGML_ASSERT(tensor_id >= 0 && tensor_id < gguf_get_n_tensors(ctx)); + return ctx->info[tensor_id].t.name; +} + +enum ggml_type gguf_get_tensor_type(const struct gguf_context * ctx, int64_t tensor_id) { + GGML_ASSERT(tensor_id >= 0 && tensor_id < gguf_get_n_tensors(ctx)); + return ctx->info[tensor_id].t.type; +} + +size_t gguf_get_tensor_size(const struct gguf_context * ctx, int64_t tensor_id) { + GGML_ASSERT(tensor_id >= 0 && tensor_id < gguf_get_n_tensors(ctx)); + return ggml_nbytes(&ctx->info[tensor_id].t); +} + +int64_t gguf_remove_key(struct gguf_context * ctx, const char * key) { + const int64_t key_id = gguf_find_key(ctx, key); + if (key_id >= 0) { + ctx->kv.erase(ctx->kv.begin() + key_id); + } + return key_id; +} + +template +static void gguf_check_reserved_keys(const std::string & key, const T val) { + if (key == GGUF_KEY_GENERAL_ALIGNMENT) { + if constexpr (std::is_same::value) { + GGML_ASSERT(val > 0 && (val & (val - 1)) == 0 && GGUF_KEY_GENERAL_ALIGNMENT " must be power of 2"); + } else { + GGML_ABORT(GGUF_KEY_GENERAL_ALIGNMENT " must be type u32"); + } + } +} + +void gguf_set_val_u8(struct gguf_context * ctx, const char * key, uint8_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_i8(struct gguf_context * ctx, const char * key, int8_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_u16(struct gguf_context * ctx, const char * key, uint16_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_i16(struct gguf_context * ctx, const char * key, int16_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_u32(struct gguf_context * ctx, const char * key, uint32_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_i32(struct gguf_context * ctx, const char * key, int32_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_f32(struct gguf_context * ctx, const char * key, float val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_u64(struct gguf_context * ctx, const char * key, uint64_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_i64(struct gguf_context * ctx, const char * key, int64_t val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_f64(struct gguf_context * ctx, const char * key, double val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_bool(struct gguf_context * ctx, const char * key, bool val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, val); +} + +void gguf_set_val_str(struct gguf_context * ctx, const char * key, const char * val) { + gguf_check_reserved_keys(key, val); + gguf_remove_key(ctx, key); + ctx->kv.emplace_back(key, std::string(val)); +} + +void gguf_set_arr_data(struct gguf_context * ctx, const char * key, enum gguf_type type, const void * data, size_t n) { + gguf_check_reserved_keys(key, data); + gguf_remove_key(ctx, key); + + const size_t nbytes = n*gguf_type_size(type); + std::vector tmp(nbytes); + if (!tmp.empty()) { + memcpy(tmp.data(), data, nbytes); + } + ctx->kv.emplace_back(key, tmp); + ctx->kv.back().cast(type); +} + +void gguf_set_arr_str(struct gguf_context * ctx, const char * key, const char ** data, size_t n) { + gguf_check_reserved_keys(key, data); + gguf_remove_key(ctx, key); + + std::vector tmp(n); + for (size_t i = 0; i < n; ++i) { + tmp[i] = data[i]; + } + ctx->kv.emplace_back(key, tmp); +} + +// set or add KV pairs from another context +void gguf_set_kv(struct gguf_context * ctx, const struct gguf_context * src) { + const int64_t n_kv = gguf_get_n_kv(src); + for (int64_t i = 0; i < n_kv; ++i) { + const struct gguf_kv & kv = src->kv[i]; + + if (!kv.is_array) { + switch (kv.get_type()) { + case GGUF_TYPE_UINT8: gguf_set_val_u8 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_INT8: gguf_set_val_i8 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_UINT16: gguf_set_val_u16 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_INT16: gguf_set_val_i16 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_UINT32: gguf_set_val_u32 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_INT32: gguf_set_val_i32 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_FLOAT32: gguf_set_val_f32 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_UINT64: gguf_set_val_u64 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_INT64: gguf_set_val_i64 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_FLOAT64: gguf_set_val_f64 (ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_BOOL: gguf_set_val_bool(ctx, kv.get_key().c_str(), kv.get_val()); break; + case GGUF_TYPE_STRING: gguf_set_val_str (ctx, kv.get_key().c_str(), kv.get_val().c_str()); break; + case GGUF_TYPE_ARRAY: + default: GGML_ABORT("invalid type"); + } + continue; + } + + const size_t ne = kv.get_ne(); + + switch (kv.get_type()) { + case GGUF_TYPE_UINT8: + case GGUF_TYPE_INT8: + case GGUF_TYPE_UINT16: + case GGUF_TYPE_INT16: + case GGUF_TYPE_UINT32: + case GGUF_TYPE_INT32: + case GGUF_TYPE_FLOAT32: + case GGUF_TYPE_UINT64: + case GGUF_TYPE_INT64: + case GGUF_TYPE_FLOAT64: + case GGUF_TYPE_BOOL: { + gguf_set_arr_data(ctx, kv.get_key().c_str(), kv.get_type(), kv.data.data(), ne); + } break; + case GGUF_TYPE_STRING: { + std::vector tmp(ne); + for (size_t j = 0; j < ne; ++j) { + tmp[j] = kv.data_string[j].c_str(); + } + gguf_set_arr_str(ctx, kv.get_key().c_str(), tmp.data(), ne); + } break; + case GGUF_TYPE_ARRAY: + default: GGML_ABORT("invalid type"); + } + } +} + +void gguf_add_tensor( + struct gguf_context * ctx, + const struct ggml_tensor * tensor) { + GGML_ASSERT(tensor); + if (gguf_find_tensor(ctx, tensor->name) != -1) { + GGML_ABORT("duplicate tensor name: %s", tensor->name); + } + + struct gguf_tensor_info ti; + ti.t = *tensor; + ti.offset = ctx->info.empty() ? 0 : + ctx->info.back().offset + GGML_PAD(ggml_nbytes(&ctx->info.back().t), ctx->alignment); + ctx->info.push_back(ti); +} + +void gguf_set_tensor_type(struct gguf_context * ctx, const char * name, enum ggml_type type) { + const int64_t tensor_id = gguf_find_tensor(ctx, name); + if (tensor_id < 0) { + GGML_ABORT("tensor not found: %s", name); + } + struct ggml_tensor * tensor = &ctx->info[tensor_id].t; + const size_t type_size = ggml_type_size(type); + const int64_t blck_size = ggml_blck_size(type); + + tensor->type = type; + GGML_ASSERT(tensor->ne[0] % blck_size == 0 && "tensor row size not divisible by block size of new type"); + + tensor->nb[0] = type_size; + tensor->nb[1] = tensor->nb[0]*(tensor->ne[0]/blck_size); + for (int i = 2; i < GGML_MAX_DIMS; i++) { + tensor->nb[i] = tensor->nb[i - 1]*tensor->ne[i - 1]; + } + + // update offsets + const int64_t n_tensors = gguf_get_n_tensors(ctx); + for (int64_t i = tensor_id + 1; i < n_tensors; ++i) { + ctx->info[i].offset = ctx->info[i - 1].offset + GGML_PAD(ggml_nbytes(&ctx->info[i - 1].t), ctx->alignment); + } +} + +void gguf_set_tensor_data(struct gguf_context * ctx, const char * name, const void * data) { + const int64_t tensor_id = gguf_find_tensor(ctx, name); + if (tensor_id < 0) { + GGML_ABORT("tensor not found: %s", name); + } + + ctx->info[tensor_id].t.data = (void *)(uintptr_t)data; // double cast suppresses warning about casting away const +} + +struct gguf_writer { + std::vector & buf; + + gguf_writer(std::vector & buf) : buf(buf) {} + + template + void write(const T & val) const { + for (size_t i = 0; i < sizeof(val); ++i) { + buf.push_back(reinterpret_cast(&val)[i]); + } + } + + void write(const std::vector & val) const { + buf.insert(buf.end(), val.begin(), val.end()); + } + + void write(const bool & val) const { + const int8_t val8 = val ? 1 : 0; + write(val8); + } + + void write(const std::string & val) const { + { + const uint64_t n = val.length(); + write(n); + } + for (size_t i = 0; i < val.length(); ++i) { + buf.push_back(reinterpret_cast(val.data())[i]); + } + } + + void write(const char * val) const { + write(std::string(val)); + } + + void write(const enum ggml_type & val) const { + write(int32_t(val)); + } + + void write(const enum gguf_type & val) const { + write(int32_t(val)); + } + + void write(const struct gguf_kv & kv) const { + const uint64_t ne = kv.get_ne(); + + write(kv.get_key()); + + if (kv.is_array) { + write(GGUF_TYPE_ARRAY); + write(kv.get_type()); + write(ne); + } else { + write(kv.get_type()); + } + + switch (kv.get_type()) { + case GGUF_TYPE_UINT8: + case GGUF_TYPE_INT8: + case GGUF_TYPE_UINT16: + case GGUF_TYPE_INT16: + case GGUF_TYPE_UINT32: + case GGUF_TYPE_INT32: + case GGUF_TYPE_FLOAT32: + case GGUF_TYPE_UINT64: + case GGUF_TYPE_INT64: + case GGUF_TYPE_FLOAT64: { + write(kv.data); + } break; + case GGUF_TYPE_BOOL: { + for (size_t i = 0; i < ne; ++i) { + write(kv.get_val(i)); + } + } break; + case GGUF_TYPE_STRING: { + for (size_t i = 0; i < ne; ++i) { + write(kv.get_val(i)); + } + } break; + case GGUF_TYPE_ARRAY: + default: GGML_ABORT("invalid type"); + } + } + + void write_tensor_meta(const struct gguf_tensor_info & info) const { + write(info.t.name); + + const uint32_t n_dims = ggml_n_dims(&info.t); + write(n_dims); + + for (uint32_t j = 0; j < n_dims; ++j) { + write(info.t.ne[j]); + } + write(info.t.type); + write(info.offset); + } + + void pad(const size_t alignment) const { + while (buf.size() % alignment != 0) { + const int8_t zero = 0; + write(zero); + } + } + + void write_tensor_data(const struct gguf_tensor_info & info, const size_t offset_data, const size_t alignment) const { + GGML_ASSERT(buf.size() - offset_data == info.offset); + + GGML_ASSERT(ggml_is_contiguous(&info.t)); + const size_t offset = buf.size(); + const size_t nbytes = ggml_nbytes(&info.t); + + buf.resize(offset + nbytes); + if (info.t.buffer) { + ggml_backend_tensor_get(&info.t, buf.data() + offset, 0, nbytes); + } else { + GGML_ASSERT(info.t.data); + memcpy(buf.data() + offset, info.t.data, nbytes); + } + + pad(alignment); + } +}; + +void gguf_write_to_buf(const struct gguf_context * ctx, std::vector & buf, bool only_meta) { + const struct gguf_writer gw(buf); + + const int64_t n_kv = gguf_get_n_kv(ctx); + const int64_t n_tensors = gguf_get_n_tensors(ctx); + + // write header + gw.write(GGUF_MAGIC[0]); + gw.write(GGUF_MAGIC[1]); + gw.write(GGUF_MAGIC[2]); + gw.write(GGUF_MAGIC[3]); + gw.write(ctx->version); + gw.write(n_tensors); + gw.write(n_kv); + + // write key-value pairs + for (int64_t i = 0; i < n_kv; ++i) { + gw.write(ctx->kv[i]); + } + + // write tensor info + for (int64_t i = 0; i < n_tensors; ++i) { + gw.write_tensor_meta(ctx->info[i]); + } + + // we require the data section to be aligned + gw.pad(ctx->alignment); + + if (only_meta) { + return; + } + + const size_t offset_data = gw.buf.size(); + + // write tensor data + for (int64_t i = 0; i < n_tensors; ++i) { + gw.write_tensor_data(ctx->info[i], offset_data, ctx->alignment); + } +} + +bool gguf_write_to_file(const struct gguf_context * ctx, const char * fname, bool only_meta) { + FILE * file = ggml_fopen(fname, "wb"); + + if (!file) { + fprintf(stderr, "%s: failed to open file '%s' for writing GGUF data\n", __func__, fname); + return false; + } + + std::vector buf; + gguf_write_to_buf(ctx, buf, only_meta); + const bool ok = fwrite(buf.data(), 1, buf.size(), file) == buf.size(); + fclose(file); + return ok; +} + +size_t gguf_get_meta_size(const struct gguf_context * ctx) { + // only return size + std::vector buf; + gguf_write_to_buf(ctx, buf, /*only_meta =*/ true); + return buf.size(); +} + +void gguf_get_meta_data(const struct gguf_context * ctx, void * data) { + std::vector buf; + gguf_write_to_buf(ctx, buf, /*only_meta =*/ true); + memcpy(data, buf.data(), buf.size()); +} diff --git a/ggml/src/kompute-shaders/op_rope_f16.comp b/ggml/src/kompute-shaders/op_rope_f16.comp deleted file mode 100644 index 0ecfb2eab..000000000 --- a/ggml/src/kompute-shaders/op_rope_f16.comp +++ /dev/null @@ -1,73 +0,0 @@ -#version 450 - -#include "rope_common.comp" - -layout(binding = 0) buffer restrict readonly tensorInA { float16_t inA[]; }; -layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; -layout(binding = 2) buffer restrict writeonly tensorOut { float16_t out_[]; }; - -void main() { - const uint i3 = gl_WorkGroupID.z; - const uint i2 = gl_WorkGroupID.y; - const uint i1 = gl_WorkGroupID.x; - - const bool is_neox = (pcs.mode & GGML_ROPE_TYPE_NEOX) != 0; - - float corr_dims[2]; - rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); - - const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); - - const int p = inB[pcs.inBOff + i2]; - - float theta = float(p); - - if (!is_neox) { - for (uint i0 = 0; i0 < pcs.ne0; i0 += 2) { - float cos_theta, sin_theta; - rope_yarn(theta, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); - - theta *= theta_scale; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ - - const float x0 = float(inA[src]); - const float x1 = float(inA[src+1]); - - out_[dst_data] = float16_t(x0*cos_theta - x1*sin_theta); - out_[dst_data+1] = float16_t(x0*sin_theta + x1*cos_theta); - } - } else { - const float inv_ndims = -1.f/pcs.n_dims; - for (uint ic = 0; ic < pcs.n_dims; ic += 2) { - const uint cur_rot = ic; - - float cos_theta, sin_theta; - rope_yarn(theta, pcs.freq_scale, corr_dims, cur_rot, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); - - theta *= theta_scale; - - const uint i0 = ic/2; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ - - const float x0 = float(inA[src]); - const float x1 = float(inA[src+pcs.n_dims/2]); - - out_[dst_data] = float16_t(x0*cos_theta - x1*sin_theta); - out_[dst_data+pcs.n_dims/2] = float16_t(x0*sin_theta + x1*cos_theta); - } - - for (uint ic = pcs.n_dims; ic < pcs.ne0; ic += 2) { - const uint i0 = ic; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 2) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 2) + pcs.outOff; // Based from out_ - - out_[dst_data + 0] = inA[src + 0]; - out_[dst_data + 1] = inA[src + 1]; - } - } -} diff --git a/ggml/src/kompute-shaders/op_rope_f32.comp b/ggml/src/kompute-shaders/op_rope_f32.comp deleted file mode 100644 index cec0fd9a5..000000000 --- a/ggml/src/kompute-shaders/op_rope_f32.comp +++ /dev/null @@ -1,73 +0,0 @@ -#version 450 - -#include "rope_common.comp" - -layout(binding = 0) buffer restrict readonly tensorInA { float inA[]; }; -layout(binding = 1) buffer restrict readonly tensorInB { int inB[]; }; -layout(binding = 2) buffer restrict writeonly tensorOut { float out_[]; }; - -void main() { - const uint i3 = gl_WorkGroupID.z; - const uint i2 = gl_WorkGroupID.y; - const uint i1 = gl_WorkGroupID.x; - - const bool is_neox = (pcs.mode & GGML_ROPE_TYPE_NEOX) != 0; - - float corr_dims[2]; - rope_yarn_corr_dims(pcs.n_dims, pcs.n_ctx_orig, pcs.freq_base, pcs.beta_fast, pcs.beta_slow, corr_dims); - - const float theta_scale = pow(pcs.freq_base, -2.0/pcs.n_dims); - - const int p = inB[pcs.inBOff + i2]; - - float theta = float(p); - - if (!is_neox) { - for (uint i0 = 0; i0 < pcs.ne0; i0 += 2) { - float cos_theta, sin_theta; - rope_yarn(theta, pcs.freq_scale, corr_dims, i0, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); - - theta *= theta_scale; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ - - const float x0 = inA[src]; - const float x1 = inA[src+1]; - - out_[dst_data] = x0*cos_theta - x1*sin_theta; - out_[dst_data+1] = x0*sin_theta + x1*cos_theta; - } - } else { - const float inv_ndims = -1.f/pcs.n_dims; - for (uint ic = 0; ic < pcs.n_dims; ic += 2) { - const uint cur_rot = ic; - - float cos_theta, sin_theta; - rope_yarn(theta, pcs.freq_scale, corr_dims, cur_rot, pcs.ext_factor, pcs.attn_factor, cos_theta, sin_theta); - - theta *= theta_scale; - - const uint i0 = ic/2; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ - - const float x0 = inA[src]; - const float x1 = inA[src+pcs.n_dims/2]; - - out_[dst_data] = x0*cos_theta - x1*sin_theta; - out_[dst_data+pcs.n_dims/2] = x0*sin_theta + x1*cos_theta; - } - - for (uint ic = pcs.n_dims; ic < pcs.ne0; ic += 2) { - const uint i0 = ic; - - const uint src = uint((i3*pcs.nb03 + i2*pcs.nb02 + i1*pcs.nb01 + i0*pcs.nb00) / 4) + pcs.inAOff; // Based from in - const uint dst_data = uint((i3*pcs.nb3 + i2*pcs.nb2 + i1*pcs.nb1 + i0*pcs.nb0) / 4) + pcs.outOff; // Based from out_ - - out_[dst_data + 0] = inA[src + 0]; - out_[dst_data + 1] = inA[src + 1]; - } - } -} diff --git a/ggml/src/llamafile/sgemm.cpp b/ggml/src/llamafile/sgemm.cpp deleted file mode 100644 index 9eead3f61..000000000 --- a/ggml/src/llamafile/sgemm.cpp +++ /dev/null @@ -1,1276 +0,0 @@ -// Copyright 2024 Mozilla Foundation -// -// Permission is hereby granted, free of charge, to any person obtaining -// a copy of this software and associated documentation files (the -// "Software"), to deal in the Software without restriction, including -// without limitation the rights to use, copy, modify, merge, publish, -// distribute, sublicense, and/or sell copies of the Software, and to -// permit persons to whom the Software is furnished to do so, subject to -// the following conditions: -// -// The above copyright notice and this permission notice shall be -// included in all copies or substantial portions of the Software. -// -// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, -// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF -// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND -// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS -// BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN -// ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN -// CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE -// SOFTWARE. - -// -// _ _ ___ _ _ ___ -// | |_(_)_ _ _ _| _ ) | /_\ / __| -// | _| | ' \ || | _ \ |__ / _ \\__ \. -// \__|_|_||_\_, |___/____/_/ \_\___/ -// |__/ -// -// BASIC LINEAR ALGEBRA SUBPROGRAMS -// -// -// This file implements multithreaded CPU matrix multiplication for the -// common contiguous use case C = Aᵀ * B. These kernels are designed to -// have excellent performance[1] for matrices that fit in the CPU cache -// without imposing any overhead such as cache filling or malloc calls. -// -// This implementation does not guarantee any upper bound with rounding -// errors, which grow along with k. Our goal's to maximally exploit the -// hardware for performance, and then use whatever resources remain for -// improving numerical accuracy. -// -// [1] J. Tunney, ‘LLaMA Now Goes Faster on CPUs’, Mar. 2024. [Online]. -// Available: https://justine.lol/matmul/. [Accessed: 29-Mar-2024]. - -#if defined(__GNUC__) -#pragma GCC diagnostic ignored "-Wpedantic" -#pragma GCC diagnostic ignored "-Wignored-attributes" -#endif - -#include "sgemm.h" -#include "ggml-impl.h" -#include "ggml-cpu-impl.h" -#include "ggml-quants.h" - -#ifdef _MSC_VER -#define NOINLINE __declspec(noinline) -#else -#define NOINLINE __attribute__((__noinline__)) -#endif - -#if defined(__ARM_NEON) || defined(__AVX512F__) -#define VECTOR_REGISTERS 32 -#else -#define VECTOR_REGISTERS 16 -#endif - -#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1) - -namespace { - -inline float unhalf(ggml_fp16_t d) { - return GGML_FP16_TO_FP32(d); -} - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// VECTORIZED ARITHMETIC OPERATIONS - -#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -inline __m128 add(__m128 x, __m128 y) { return _mm_add_ps(x, y); } -inline __m128 sub(__m128 x, __m128 y) { return _mm_sub_ps(x, y); } -inline __m128 mul(__m128 x, __m128 y) { return _mm_mul_ps(x, y); } -#endif // __SSE__ - -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -inline __m256 add(__m256 x, __m256 y) { return _mm256_add_ps(x, y); } -inline __m256 sub(__m256 x, __m256 y) { return _mm256_sub_ps(x, y); } -inline __m256 mul(__m256 x, __m256 y) { return _mm256_mul_ps(x, y); } -#endif // __AVX__ - -#if defined(__AVX512F__) -inline __m512 add(__m512 x, __m512 y) { return _mm512_add_ps(x, y); } -inline __m512 sub(__m512 x, __m512 y) { return _mm512_sub_ps(x, y); } -inline __m512 mul(__m512 x, __m512 y) { return _mm512_mul_ps(x, y); } -#endif // __AVX512F__ - -#if defined(__ARM_NEON) -inline float32x4_t add(float32x4_t x, float32x4_t y) { return vaddq_f32(x, y); } -inline float32x4_t sub(float32x4_t x, float32x4_t y) { return vsubq_f32(x, y); } -inline float32x4_t mul(float32x4_t x, float32x4_t y) { return vmulq_f32(x, y); } -#endif // __ARM_NEON - -#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) -inline float16x8_t add(float16x8_t x, float16x8_t y) { return vaddq_f16(x, y); } -inline float16x8_t sub(float16x8_t x, float16x8_t y) { return vsubq_f16(x, y); } -inline float16x8_t mul(float16x8_t x, float16x8_t y) { return vmulq_f16(x, y); } -#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// VECTORIZED FUSED MULTIPLY ADD - -/** - * Computes a * b + c. - */ -template -inline U madd(T a, T b, U c) { - return add(mul(a, b), c); -} - -#if defined(__FMA__) -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -template <> -inline __m256 madd(__m256 a, __m256 b, __m256 c) { - return _mm256_fmadd_ps(a, b, c); -} -#endif -#if defined(__AVX512F__) -template <> -inline __m512 madd(__m512 a, __m512 b, __m512 c) { - return _mm512_fmadd_ps(a, b, c); -} -#endif -#endif - -#if defined(__ARM_FEATURE_FMA) -template <> -inline float32x4_t madd(float32x4_t a, float32x4_t b, float32x4_t c) { - return vfmaq_f32(c, b, a); -} -#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) -template <> -inline float16x8_t madd(float16x8_t a, float16x8_t b, float16x8_t c) { - return vfmaq_f16(c, b, a); -} -#endif -#endif - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// VECTORIZED HORIZONTAL SUM - -#if defined(__ARM_NEON) -inline float hsum(float32x4_t x) { - return vaddvq_f32(x); -} -#endif // __ARM_NEON - -#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) -inline float hsum(float16x8_t x) { - return vaddvq_f32(vaddq_f32(vcvt_f32_f16(vget_low_f16(x)), - vcvt_f32_f16(vget_high_f16(x)))); -} -#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC - -#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -inline float hsum(__m128 x) { -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) - x = _mm_add_ps(x, _mm_movehl_ps(x, x)); - x = _mm_add_ss(x, _mm_movehdup_ps(x)); -#else - __m128 t; - t = _mm_shuffle_ps(x, x, _MM_SHUFFLE(2, 3, 0, 1)); - x = _mm_add_ps(x, t); - t = _mm_movehl_ps(t, x); - x = _mm_add_ss(x, t); -#endif - return _mm_cvtss_f32(x); -} -#endif - -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -inline float hsum(__m256 x) { - return hsum(_mm_add_ps(_mm256_extractf128_ps(x, 1), - _mm256_castps256_ps128(x))); -} -#endif // __AVX__ - -#if defined(__AVX512F__) -inline float hsum(__m512 x) { - return _mm512_reduce_add_ps(x); -} -#endif // __AVX512F__ - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// VECTORIZED MEMORY LOADING - -template T load(const U *); - -#if defined(__ARM_NEON) -template <> inline float32x4_t load(const float *p) { - return vld1q_f32(p); -} -#if !defined(_MSC_VER) -template <> inline float16x8_t load(const ggml_fp16_t *p) { - return vld1q_f16((const float16_t *)p); -} -template <> inline float32x4_t load(const ggml_fp16_t *p) { - return vcvt_f32_f16(vld1_f16((const float16_t *)p)); -} -#endif // _MSC_VER -#endif // __ARM_NEON - -#if defined(__SSE__) || defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -template <> inline __m128 load(const float *p) { - return _mm_loadu_ps(p); -} -#endif // __SSE__ - -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -template <> inline __m256 load(const float *p) { - return _mm256_loadu_ps(p); -} -#endif // __AVX__ - -#if defined(__F16C__) -template <> inline __m256 load(const ggml_fp16_t *p) { - return _mm256_cvtph_ps(_mm_loadu_si128((const __m128i *)p)); -} -#endif // __F16C__ - -#if defined(__AVX512F__) -template <> inline __m512 load(const float *p) { - return _mm512_loadu_ps(p); -} -template <> inline __m512 load(const ggml_fp16_t *p) { - return _mm512_cvtph_ps(_mm256_loadu_si256((const __m256i *)p)); -} -#endif // __AVX512F__ - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// CONSTANTS - -#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) -static const int8_t kvalues_iq4nl[16] = {-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113}; -static const __m128i iq4nlt = _mm_loadu_si128((const __m128i *) kvalues_iq4nl); -#endif - -//////////////////////////////////////////////////////////////////////////////////////////////////// -// FLOATING POINT MATRIX MULTIPLICATION - -template -class tinyBLAS { - public: - tinyBLAS(int64_t k, - const TA *A, int64_t lda, - const TB *B, int64_t ldb, - TC *C, int64_t ldc, - int ith, int nth) - : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { - } - - void matmul(int64_t m, int64_t n) { - mnpack(0, m, 0, n); - } - - private: - NOINLINE void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t mc, nc, mp, np; - switch ((MIN(m - m0, 5) << 4) | MIN(n - n0, 5)) { -#if VECTOR_REGISTERS == 32 - case 0x55: - mc = 5; - nc = 5; - gemm<5, 5>(m0, m, n0, n); - break; - case 0x45: - mc = 4; - nc = 5; - gemm<4, 5>(m0, m, n0, n); - break; - case 0x54: - mc = 5; - nc = 4; - gemm<5, 4>(m0, m, n0, n); - break; - case 0x44: - mc = 4; - nc = 4; - gemm<4, 4>(m0, m, n0, n); - break; - case 0x53: - mc = 5; - nc = 3; - gemm<5, 3>(m0, m, n0, n); - break; - case 0x35: - mc = 3; - nc = 5; - gemm<3, 5>(m0, m, n0, n); - break; - case 0x43: - mc = 4; - nc = 3; - gemm<4, 3>(m0, m, n0, n); - break; -#else - case 0x55: - case 0x54: - case 0x53: - case 0x45: - case 0x44: - case 0x43: - mc = 4; - nc = 3; - gemm<4, 3>(m0, m, n0, n); - break; - case 0x35: -#endif - case 0x34: - mc = 3; - nc = 4; - gemm<3, 4>(m0, m, n0, n); - break; - case 0x52: - mc = 5; - nc = 2; - gemm<5, 2>(m0, m, n0, n); - break; - case 0x33: - mc = 3; - nc = 3; - gemm<3, 3>(m0, m, n0, n); - break; - case 0x25: - mc = 2; - nc = 5; - gemm<2, 5>(m0, m, n0, n); - break; - case 0x42: - mc = 4; - nc = 2; - gemm<4, 2>(m0, m, n0, n); - break; - case 0x24: - mc = 2; - nc = 4; - gemm<2, 4>(m0, m, n0, n); - break; - case 0x32: - mc = 3; - nc = 2; - gemm<3, 2>(m0, m, n0, n); - break; - case 0x23: - mc = 2; - nc = 3; - gemm<2, 3>(m0, m, n0, n); - break; - case 0x51: - mc = 5; - nc = 1; - gemm<5, 1>(m0, m, n0, n); - break; - case 0x41: - mc = 4; - nc = 1; - gemm<4, 1>(m0, m, n0, n); - break; - case 0x22: - mc = 2; - nc = 2; - gemm<2, 2>(m0, m, n0, n); - break; - case 0x15: - mc = 1; - nc = 5; - gemm<1, 5>(m0, m, n0, n); - break; - case 0x14: - mc = 1; - nc = 4; - gemm<1, 4>(m0, m, n0, n); - break; - case 0x31: - mc = 3; - nc = 1; - gemm<3, 1>(m0, m, n0, n); - break; - case 0x13: - mc = 1; - nc = 3; - gemm<1, 3>(m0, m, n0, n); - break; - case 0x21: - mc = 2; - nc = 1; - gemm<2, 1>(m0, m, n0, n); - break; - case 0x12: - mc = 1; - nc = 2; - gemm<1, 2>(m0, m, n0, n); - break; - case 0x11: - mc = 1; - nc = 1; - gemm<1, 1>(m0, m, n0, n); - break; - default: - return; - } - mp = m0 + (m - m0) / mc * mc; - np = n0 + (n - n0) / nc * nc; - mnpack(mp, m, n0, np); - mnpack(m0, m, np, n); - } - - template - NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t ytiles = (m - m0) / RM; - int64_t xtiles = (n - n0) / RN; - int64_t tiles = xtiles * ytiles; - int64_t duty = (tiles + nth - 1) / nth; - int64_t start = duty * ith; - int64_t end = start + duty; - if (end > tiles) - end = tiles; - for (int64_t job = start; job < end; ++job) { - int64_t ii = m0 + job / xtiles * RM; - int64_t jj = n0 + job % xtiles * RN; - D Cv[RN][RM] = {}; - for (int64_t l = 0; l < k; l += KN) - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) - Cv[j][i] = madd(load(A + lda * (ii + i) + l), - load(B + ldb * (jj + j) + l), - Cv[j][i]); - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) - C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); - } - } - - const TA *const A; - const TB *const B; - TC *const C; - const int64_t k; - const int64_t lda; - const int64_t ldb; - const int64_t ldc; - const int ith; - const int nth; -}; - -////////////////////////////////////////////////////////////////////////////////////////// -// QUANT ZERO MATRIX MULTIPLICATION - -#if defined(__ARM_FEATURE_DOTPROD) -template -class tinyBLAS_Q0_ARM { - public: - tinyBLAS_Q0_ARM(int64_t k, - const TA *A, int64_t lda, - const block_q8_0 *B, int64_t ldb, - float *C, int64_t ldc, - int ith, int nth) - : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { - } - - void matmul(int64_t m, int64_t n) { - mnpack(0, m, 0, n); - } - - private: - NOINLINE void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t mc, nc, mp, np; - switch ((MIN(m - m0, 3) << 4) | MIN(n - n0, 3ll)) { - case 0x33: - mc = 3; - nc = 3; - gemm<3, 3>(m0, m, n0, n); - break; - case 0x32: - mc = 3; - nc = 2; - gemm<3, 2>(m0, m, n0, n); - break; - case 0x23: - mc = 2; - nc = 3; - gemm<2, 3>(m0, m, n0, n); - break; - case 0x22: - mc = 2; - nc = 2; - gemm<2, 2>(m0, m, n0, n); - break; - case 0x31: - mc = 3; - nc = 1; - gemm<3, 1>(m0, m, n0, n); - break; - case 0x13: - mc = 1; - nc = 3; - gemm<1, 3>(m0, m, n0, n); - break; - case 0x21: - mc = 2; - nc = 1; - gemm<2, 1>(m0, m, n0, n); - break; - case 0x12: - mc = 1; - nc = 2; - gemm<1, 2>(m0, m, n0, n); - break; - case 0x11: - mc = 1; - nc = 1; - gemm<1, 1>(m0, m, n0, n); - break; - default: - return; - } - mp = m0 + (m - m0) / mc * mc; - np = n0 + (n - n0) / nc * nc; - mnpack(mp, m, n0, np); - mnpack(m0, m, np, n); - } - - template - NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t ytiles = (m - m0) / RM; - int64_t xtiles = (n - n0) / RN; - int64_t tiles = xtiles * ytiles; - int64_t duty = (tiles + nth - 1) / nth; - int64_t start = duty * ith; - int64_t end = start + duty; - if (end > tiles) - end = tiles; - for (int64_t job = start; job < end; ++job) { - int64_t ii = m0 + job / xtiles * RM; - int64_t jj = n0 + job % xtiles * RN; - float32x4_t Cv[RN][RM] = {}; - for (int64_t l = 0; l < k; ++l) - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) - Cv[j][i] = vmlaq_n_f32(Cv[j][i], - vcvtq_f32_s32(vdotq_s32( - vdotq_s32(vdupq_n_s32(0), - load_lo(A + lda * (ii + i) + l), - load_lo(B + ldb * (jj + j) + l)), - load_hi(A + lda * (ii + i) + l), - load_hi(B + ldb * (jj + j) + l))), - unhalf(A[lda * (ii + i) + l].d) * - unhalf(B[ldb * (jj + j) + l].d)); - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) - C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); - } - } - - inline int8x16_t load_lo(const block_q8_0 *b) { - return vld1q_s8(b->qs); - } - - inline int8x16_t load_hi(const block_q8_0 *b) { - return vld1q_s8(b->qs + 16); - } - - inline int8x16_t load_lo(const block_q4_0 *b) { - return vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vld1q_u8(b->qs), - vdupq_n_u8(0x0f))), - vdupq_n_s8(0x8)); - } - - inline int8x16_t load_hi(const block_q4_0 *b) { - return vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(vld1q_u8(b->qs), 4)), - vdupq_n_s8(0x8)); - } - - const TA *const A; - const block_q8_0 *const B; - float *const C; - const int64_t k; - const int64_t lda; - const int64_t ldb; - const int64_t ldc; - const int ith; - const int nth; -}; -#endif // __ARM_FEATURE_DOTPROD - -#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) -template -class tinyBLAS_Q0_AVX { - public: - tinyBLAS_Q0_AVX(int64_t k, - const TA *A, int64_t lda, - const TB *B, int64_t ldb, - TC *C, int64_t ldc, - int ith, int nth) - : A(A), B(B), C(C), k(k), lda(lda), ldb(ldb), ldc(ldc), ith(ith), nth(nth) { - } - - void matmul(int64_t m, int64_t n) { - mnpack(0, m, 0, n); - } - - private: - void mnpack(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t mc, nc, mp, np; - switch ((MIN(m - m0, 4) << 4) | MIN(n - n0, 4)) { -#if VECTOR_REGISTERS == 32 - case 0x44: - mc = 4; - nc = 4; -#if defined(__AVX2__) && defined(__F16C__) - gemm4xN<4>(m0, m, n0, n); -#else - gemm<4, 4>(m0, m, n0, n); -#endif - break; - case 0x43: - mc = 4; - nc = 3; -#if defined(__AVX2__) && defined(__F16C__) - gemm4xN<3>(m0, m, n0, n); -#else - gemm<4, 3>(m0, m, n0, n); -#endif - break; - case 0x34: - mc = 3; - nc = 4; -#if defined(__AVX2__) && defined(__F16C__) - gemmMx4<3>(m0, m, n0, n); -#else - gemm<3, 4>(m0, m, n0, n); -#endif - break; - case 0x33: - mc = 3; - nc = 3; - gemm<3, 3>(m0, m, n0, n); - break; - case 0x42: - mc = 4; - nc = 2; -#if defined(__AVX2__) && defined(__F16C__) - gemm4xN<2>(m0, m, n0, n); -#else - gemm<4, 2>(m0, m, n0, n); -#endif - break; - case 0x24: - mc = 2; - nc = 4; -#if defined(__AVX2__) && defined(__F16C__) - gemmMx4<2>(m0, m, n0, n); -#else - gemm<2, 4>(m0, m, n0, n); -#endif - break; -#else - case 0x44: - case 0x43: - case 0x42: - mc = 4; - nc = 2; -#if defined(__AVX2__) && defined(__F16C__) - gemm4xN<2>(m0, m, n0, n); -#else - gemm<4, 2>(m0, m, n0, n); -#endif - break; - case 0x34: - case 0x24: - mc = 2; - nc = 4; -#if defined(__AVX2__) && defined(__F16C__) - gemmMx4<2>(m0, m, n0, n); -#else - gemm<2, 4>(m0, m, n0, n); -#endif - break; - case 0x33: -#endif - case 0x32: - mc = 3; - nc = 2; - gemm<3, 2>(m0, m, n0, n); - break; - case 0x23: - mc = 2; - nc = 3; - gemm<2, 3>(m0, m, n0, n); - break; - case 0x41: - mc = 4; - nc = 1; -#if defined(__AVX2__) && defined(__F16C__) - gemm4xN<1>(m0, m, n0, n); -#else - gemm<4, 1>(m0, m, n0, n); -#endif - break; - case 0x22: - mc = 2; - nc = 2; - gemm<2, 2>(m0, m, n0, n); - break; - case 0x14: - mc = 1; - nc = 4; -#if defined(__AVX2__) && defined(__F16C__) - gemmMx4<1>(m0, m, n0, n); -#else - gemm<1, 4>(m0, m, n0, n); -#endif - break; - case 0x31: - mc = 3; - nc = 1; - gemm<3, 1>(m0, m, n0, n); - break; - case 0x13: - mc = 1; - nc = 3; - gemm<1, 3>(m0, m, n0, n); - break; - case 0x21: - mc = 2; - nc = 1; - gemm<2, 1>(m0, m, n0, n); - break; - case 0x12: - mc = 1; - nc = 2; - gemm<1, 2>(m0, m, n0, n); - break; - case 0x11: - mc = 1; - nc = 1; - gemm<1, 1>(m0, m, n0, n); - break; - default: - return; - } - mp = m0 + (m - m0) / mc * mc; - np = n0 + (n - n0) / nc * nc; - mnpack(mp, m, n0, np); - mnpack(m0, m, np, n); - } - -#if defined(__AVX2__) && defined(__F16C__) -// Templated functions for gemm of dimensions 4xN - template - NOINLINE void gemm4xN(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t ytiles = (m - m0) / 4; - int64_t xtiles = (n - n0) / RN; - int64_t tiles = xtiles * ytiles; - int64_t duty = (tiles + nth - 1) / nth; - int64_t start = duty * ith; - int64_t end = start + duty; - if (end > tiles) - end = tiles; - for (int64_t job = start; job < end; ++job) { - int64_t ii = m0 + job / xtiles * 4; - int64_t jj = n0 + job % xtiles * RN; - __m256 Cv[RN][4] = {}; - for (int64_t l = 0; l < k; ++l) { - uint64_t a_delta = ((uint64_t)A[lda * (ii + 3) + l].d << 48) | ((uint64_t)A[lda * (ii + 2) + l].d << 32) | ((uint64_t)A[lda * (ii + 1) + l].d << 16) | (A[lda * (ii + 0) + l].d); - // Convert delta values for four blocks to float values - __m128 da = _mm_cvtph_ps(_mm_set_epi64x(0, a_delta)); - __m256i avec0 = load(A + lda * (ii + 0) + l); - __m256i avec1 = load(A + lda * (ii + 1) + l); - __m256i avec2 = load(A + lda * (ii + 2) + l); - __m256i avec3 = load(A + lda * (ii + 3) + l); - for (int64_t j = 0; j < RN; ++j) { - __m128 db = _mm_set1_ps(unhalf(B[ldb * (jj + j) + l].d)); - // Computation of product of delta values for four blocks and replicate it across 256 bit lane - __m256 dvec = _mm256_castps128_ps256(_mm_mul_ps(da, db)); - dvec = _mm256_permute2f128_ps(dvec ,dvec, 0); - // Computation of dot product and multiplication with appropriate delta value products - Cv[j][0] = madd(_mm256_shuffle_ps(dvec, dvec, 0), - updot(_mm256_sign_epi8(avec0, avec0), - _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec0)), - Cv[j][0]); - Cv[j][1] = madd(_mm256_shuffle_ps(dvec, dvec, 85), - updot(_mm256_sign_epi8(avec1, avec1), - _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec1)), - Cv[j][1]); - Cv[j][2] = madd(_mm256_shuffle_ps(dvec, dvec, 170), - updot(_mm256_sign_epi8(avec2, avec2), - _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec2)), - Cv[j][2]); - Cv[j][3] = madd(_mm256_shuffle_ps(dvec, dvec, 255), - updot(_mm256_sign_epi8(avec3, avec3), - _mm256_sign_epi8(load(B + ldb * (jj + j) + l), avec3)), - Cv[j][3]); - } - } - - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < 4; ++i) - C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); - } - } - - // Templated functions for gemm of dimensions Mx4 - template - NOINLINE void gemmMx4(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t ytiles = (m - m0) / RM; - int64_t xtiles = (n - n0) / 4; - int64_t tiles = xtiles * ytiles; - int64_t duty = (tiles + nth - 1) / nth; - int64_t start = duty * ith; - int64_t end = start + duty; - if (end > tiles) - end = tiles; - for (int64_t job = start; job < end; ++job) { - int64_t ii = m0 + job / xtiles * RM; - int64_t jj = n0 + job % xtiles * 4; - __m256 Cv[4][RM] = {}; - for (int64_t l = 0; l < k; ++l) { - uint64_t b_delta = ((uint64_t)B[ldb * (jj + 3) + l].d << 48) | ((uint64_t)B[ldb * (jj + 2) + l].d << 32) | ((uint64_t)B[ldb * (jj + 1) + l].d << 16) | (B[ldb * (jj + 0) + l].d); - // Convert delta values for four blocks to float values - __m128 db = _mm_cvtph_ps(_mm_set_epi64x(0, b_delta)); - __m256i bvec0 = load(B + ldb * (jj + 0) + l); - __m256i bvec1 = load(B + ldb * (jj + 1) + l); - __m256i bvec2 = load(B + ldb * (jj + 2) + l); - __m256i bvec3 = load(B + ldb * (jj + 3) + l); - for (int64_t i = 0; i < RM; ++i) { - __m128 da = _mm_set1_ps(unhalf((A[lda * (ii + i) + l].d))); - // Computation of product of delta values for four blocks and replicate it across 256 bit lane - __m256 dvec = _mm256_castps128_ps256(_mm_mul_ps(da, db)); - dvec = _mm256_permute2f128_ps(dvec ,dvec, 0); - // Computation of dot product and multiplication with appropriate delta value products - Cv[0][i] = madd(_mm256_shuffle_ps(dvec, dvec, 0), - updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), - load(A + lda * (ii + i) + l)), - _mm256_sign_epi8(bvec0, load(A + lda * (ii + i) + l))), - Cv[0][i]); - Cv[1][i] = madd(_mm256_shuffle_ps(dvec, dvec, 85), - updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), - load(A + lda * (ii + i) + l)), - _mm256_sign_epi8(bvec1, load(A + lda * (ii + i) + l))), - Cv[1][i]); - Cv[2][i] = madd(_mm256_shuffle_ps(dvec, dvec, 170), - updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), - load(A + lda * (ii + i) + l)), - _mm256_sign_epi8(bvec2, load(A + lda * (ii + i) + l))), - Cv[2][i]); - Cv[3][i] = madd(_mm256_shuffle_ps(dvec, dvec, 255), - updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), - load(A + lda * (ii + i) + l)), - _mm256_sign_epi8(bvec3, load(A + lda * (ii + i) + l))), - Cv[3][i]); - } - } - for (int64_t j = 0; j < 4; ++j) - for (int64_t i = 0; i < RM; ++i) - C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); - } - } -#endif - - template - NOINLINE void gemm(int64_t m0, int64_t m, int64_t n0, int64_t n) { - int64_t ytiles = (m - m0) / RM; - int64_t xtiles = (n - n0) / RN; - int64_t tiles = xtiles * ytiles; - int64_t duty = (tiles + nth - 1) / nth; - int64_t start = duty * ith; - int64_t end = start + duty; - if (end > tiles) - end = tiles; - for (int64_t job = start; job < end; ++job) { - int64_t ii = m0 + job / xtiles * RM; - int64_t jj = n0 + job % xtiles * RN; - __m256 Cv[RN][RM] = {}; - for (int64_t l = 0; l < k; ++l) - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) { -#if defined(__AVX2__) - __m256 udTmp = updot(_mm256_sign_epi8(load(A + lda * (ii + i) + l), - load(A + lda * (ii + i) + l)), - _mm256_sign_epi8(load(B + ldb * (jj + j) + l), - load(A + lda * (ii + i) + l))); -#else - __m128i ali0 = load0(A + lda * (ii + i) + l); - __m128i ali1 = load1(A + lda * (ii + i) + l); - __m128i blj0 = load0(B + ldb * (jj + j) + l); - __m128i blj1 = load1(B + ldb * (jj + j) + l); - - __m128i sepAA0 = _mm_sign_epi8(ali0, ali0); - __m128i sepAA1 = _mm_sign_epi8(ali1, ali1); - __m128i sepBA0 = _mm_sign_epi8(blj0, ali0); - __m128i sepBA1 = _mm_sign_epi8(blj1, ali1); - - // updot - const __m128i oneFill = _mm_set1_epi16(1); - __m128i mad0 = _mm_maddubs_epi16(sepAA0, sepBA0); - __m128i mad1 = _mm_maddubs_epi16(sepAA1, sepBA1); - __m256 udTmp = _mm256_cvtepi32_ps(MM256_SET_M128I(_mm_madd_epi16(oneFill, mad1), _mm_madd_epi16(oneFill, mad0))); -#endif - Cv[j][i] = madd(_mm256_set1_ps(unhalf(A[lda * (ii + i) + l].d) * - unhalf(B[ldb * (jj + j) + l].d)), - udTmp, - Cv[j][i]); - } - for (int64_t j = 0; j < RN; ++j) - for (int64_t i = 0; i < RM; ++i) - C[ldc * (jj + j) + (ii + i)] = hsum(Cv[j][i]); - } - } - - inline __m256i load(const block_q8_0 *b) { - return _mm256_loadu_si256((const __m256i *)b->qs); - } - - inline __m128i load0(const block_q8_0 *b) { - return _mm_loadu_si128((const __m128i *)b->qs); - } - - inline __m128i load1(const block_q8_0 *b) { - return _mm_loadu_si128(((const __m128i *)b->qs) + 1); - } - - inline __m256i load(const block_q4_0 *b) { - return _mm256_sub_epi8(denibble(b->qs), _mm256_set1_epi8(8)); - } - - inline __m128i load0(const block_q4_0 *b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - return _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), x), _mm_set1_epi8(8)); - } - - inline __m128i load1(const block_q4_0 *b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - return _mm_sub_epi8(_mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4)), _mm_set1_epi8(8)); - } - - inline __m256i load(const block_q5_0 *b) { - return _mm256_or_si256(denibble(b->qs), bittobyte(b->qh)); - } - - inline __m128i load0(const block_q5_0* b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - uint32_t x32; - memcpy(&x32, b->qh, sizeof(uint32_t)); - __m128i qxl = _mm_and_si128(_mm_set1_epi8(15), x); - __m128i bytesl = _mm_cmpeq_epi8(_mm_set1_epi64x(-1), - _mm_or_si128(_mm_set1_epi64x(0x7fbfdfeff7fbfdfe), - _mm_shuffle_epi8(_mm_set1_epi32(x32), - _mm_set_epi64x(0x0101010101010101, 0x0000000000000000)))); - bytesl = _mm_andnot_si128(bytesl, _mm_set1_epi8((char)0xF0)); - return _mm_or_si128(qxl, bytesl); - } - - inline __m128i load1(const block_q5_0* b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - uint32_t x32; - memcpy(&x32, b->qh, sizeof(uint32_t)); - __m128i qxh = _mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4)); - __m128i bytesh = _mm_cmpeq_epi8(_mm_set1_epi64x(-1), - _mm_or_si128(_mm_set1_epi64x(0x7fbfdfeff7fbfdfe), - _mm_shuffle_epi8(_mm_set1_epi32(x32), - _mm_set_epi64x(0x0303030303030303, 0x0202020202020202)))); - bytesh = _mm_andnot_si128(bytesh, _mm_set1_epi8((char)0xF0)); - return _mm_or_si128(qxh, bytesh); - } - - inline __m256i load(const block_iq4_nl *b) { - return MM256_SET_M128I(load1(b), load0(b)); - } - - inline __m128i load0(const block_iq4_nl *b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - return _mm_shuffle_epi8(iq4nlt, _mm_and_si128(_mm_set1_epi8(15), x)); - } - - inline __m128i load1(const block_iq4_nl *b) { - const __m128i x = _mm_loadu_si128((const __m128i *)(b->qs)); - return _mm_shuffle_epi8(iq4nlt, _mm_and_si128(_mm_set1_epi8(15), _mm_srli_epi16(x, 4))); - } - - inline __m256 updot(__m256i u, __m256i s) { - __m256i res; -#if defined(__AVXVNNI__) || (defined(__AVX512VNNI__) && defined(__AVX512VL__)) - res = _mm256_dpbusd_epi32(_mm256_setzero_si256(), u, s); -#else - res = _mm256_madd_epi16(_mm256_set1_epi16(1), _mm256_maddubs_epi16(u, s)); -#endif - return _mm256_cvtepi32_ps(res); - } - - static inline __m256i denibble(const uint8_t *p) { - __m128i x = _mm_loadu_si128((const __m128i *)p); - return _mm256_and_si256(_mm256_set1_epi8(15), - _mm256_insertf128_si256(_mm256_castsi128_si256(x), - _mm_srli_epi16(x, 4), 1)); - } - - static inline __m256i bittobyte(const uint8_t *p) { - uint32_t x32; - memcpy(&x32, p, sizeof(uint32_t)); - __m256i bytes = _mm256_cmpeq_epi8(_mm256_set1_epi64x(-1), - _mm256_or_si256(_mm256_set1_epi64x(0x7fbfdfeff7fbfdfe), - _mm256_shuffle_epi8(_mm256_set1_epi32(x32), - _mm256_set_epi64x(0x0303030303030303, 0x0202020202020202, - 0x0101010101010101, 0x0000000000000000)))); - return _mm256_andnot_si256(bytes, _mm256_set1_epi8((char)0xF0)); - } - - const TA *const A; - const TB *const B; - TC *const C; - const int64_t k; - const int64_t lda; - const int64_t ldb; - const int64_t ldc; - const int ith; - const int nth; -}; -#endif // __AVX__ - -} // namespace - -/** - * Performs optimized matrix multiplication on CPU. - * - * This subroutine may compute C = Aᵀ * B with column major ordering. - * Despite its name, this isn't a generalized implementation. Work is - * only performed when a handwritten kernel is written and available. - * Otherwise the caller should fall back to a general matmul routine. - * - * For example, for single-threaded single-precision GEMM you can say - * - * llamafile_sgemm(m, n, k, A, lda, B, ldb, C, ldc, - * 0, 1, - * GGML_TYPE_F32, GGML_TYPE_F32, GGML_TYPE_F32); - * - * @param m is rows in `A` and `C` - * @param n is cols in `B` and `C` - * @param k is cols in `A` and rows in `B` - * @param A is first input matrix (always transposed) - * @param lda is row stride of `A` - * @param B is second input matrix (never transposed) - * @param ldb is row stride of `B` - * @param C is input/output array of output matrices - * @param ldc is row stride of `C` - * @param ith is thread id (must be less than `nth`) - * @param nth is number of threads (must be greater than zero) - * @param Atype is GGML data type of `A` - * @param Btype is GGML data type of `B` - * @param Ctype is GGML data type of `C` - * @return true if this function was able to service the matmul request - */ -bool llamafile_sgemm(int64_t m, int64_t n, int64_t k, const void *A, int64_t lda, const void *B, int64_t ldb, void *C, - int64_t ldc, int ith, int nth, int Atype, int Btype, int Ctype) { - - assert(m >= 0); - assert(n >= 0); - assert(k >= 0); - assert(lda >= k); - assert(ldb >= k); - assert(ldc >= m); - assert(nth > 0); - assert(ith < nth); - - // only enable sgemm for prompt processing - if (n < 2) - return false; - - if (Ctype != GGML_TYPE_F32) - return false; - - switch (Atype) { - - case GGML_TYPE_F32: { - if (Btype != GGML_TYPE_F32) - return false; -#if defined(__AVX512F__) - if (k % 16) - return false; - tinyBLAS<16, __m512, __m512, float, float, float> tb{ - k, (const float *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__AVX__) || defined(__AVX2__) - if (k % 8) - return false; - tinyBLAS<8, __m256, __m256, float, float, float> tb{ - k, (const float *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__ARM_NEON) - if (n < 4) - return false; - if (k % 4) - return false; - tinyBLAS<4, float32x4_t, float32x4_t, float, float, float> tb{ - k, (const float *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - case GGML_TYPE_F16: { -#if defined(__AVX512F__) - if (k % 16) - return false; - if (Btype != GGML_TYPE_F32) - return false; - tinyBLAS<16, __m512, __m512, ggml_fp16_t, float, float> tb{ - k, (const ggml_fp16_t *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif (defined(__AVX__) || defined(__AVX2__)) && defined(__F16C__) - if (k % 8) - return false; - if (Btype != GGML_TYPE_F32) - return false; - tinyBLAS<8, __m256, __m256, ggml_fp16_t, float, float> tb{ - k, (const ggml_fp16_t *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && !defined(_MSC_VER) - if (n < 8) - return false; - if (k % 8) - return false; - if (Btype != GGML_TYPE_F16) - return false; - tinyBLAS<8, float16x8_t, float16x8_t, ggml_fp16_t, ggml_fp16_t, float> tb{ - k, (const ggml_fp16_t *)A, lda, - (const ggml_fp16_t *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__ARM_NEON) && !defined(_MSC_VER) - if (k % 4) - return false; - if (Btype != GGML_TYPE_F32) - return false; - tinyBLAS<4, float32x4_t, float32x4_t, ggml_fp16_t, float, float> tb{ - k, (const ggml_fp16_t *)A, lda, - (const float *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - case GGML_TYPE_Q8_0: { - if (Btype != GGML_TYPE_Q8_0) - return false; -#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) - tinyBLAS_Q0_AVX tb{ - k, (const block_q8_0 *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__ARM_FEATURE_DOTPROD) - tinyBLAS_Q0_ARM tb{ - k, (const block_q8_0 *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - case GGML_TYPE_Q4_0: { - if (Btype != GGML_TYPE_Q8_0) - return false; -#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) - tinyBLAS_Q0_AVX tb{ - k, (const block_q4_0 *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#elif defined(__ARM_FEATURE_DOTPROD) - tinyBLAS_Q0_ARM tb{ - k, (const block_q4_0 *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - case GGML_TYPE_Q5_0: { - if (Btype != GGML_TYPE_Q8_0) - return false; -#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) - tinyBLAS_Q0_AVX tb{ - k, (const block_q5_0 *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - case GGML_TYPE_IQ4_NL: { - if (Btype != GGML_TYPE_Q8_0) - return false; -#if defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX__) - tinyBLAS_Q0_AVX tb{ - k, (const block_iq4_nl *)A, lda, - (const block_q8_0 *)B, ldb, - (float *)C, ldc, - ith, nth}; - tb.matmul(m, n); - return true; -#else - return false; -#endif - } - - default: - return false; - } - - (void)m; - (void)n; - (void)k; - (void)A; - (void)lda; - (void)B; - (void)ldb; - (void)C; - (void)ldc; - (void)ith; - (void)nth; - (void)Atype; - (void)Btype; - (void)Ctype; -} diff --git a/ggml/src/llamafile/sgemm.h b/ggml/src/llamafile/sgemm.h deleted file mode 100644 index caf6dd556..000000000 --- a/ggml/src/llamafile/sgemm.h +++ /dev/null @@ -1,14 +0,0 @@ -#pragma once -#include -#include -#ifdef __cplusplus -extern "C" { -#endif - -bool llamafile_sgemm(int64_t, int64_t, int64_t, const void *, int64_t, - const void *, int64_t, void *, int64_t, int, int, - int, int, int); - -#ifdef __cplusplus -} -#endif diff --git a/ggml/src/vulkan-shaders/acc.comp b/ggml/src/vulkan-shaders/acc.comp deleted file mode 100644 index 4c8739efe..000000000 --- a/ggml/src/vulkan-shaders/acc.comp +++ /dev/null @@ -1,24 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_binary_head.comp" - -void main() { - const uint idx = gl_GlobalInvocationID.x; - if (idx >= p.ne) { - return; - } - - const uint offset = p.param3; - const uint src1_i = idx - offset; - const uint oz = src1_i / p.nb02; - const uint oy = (src1_i - (oz * p.nb02)) / p.nb01; - const uint ox = src1_i % p.nb01; - - if (ox < p.ne10 && oy < p.ne11 && oz < p.ne12) { - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)]) + FLOAT_TYPE(data_b[ox + oy * p.ne10 + oz * p.ne10 * p.ne11])); - } else { - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)])); - } -} - diff --git a/ggml/src/vulkan-shaders/add.comp b/ggml/src/vulkan-shaders/add.comp deleted file mode 100644 index 3974845d6..000000000 --- a/ggml/src/vulkan-shaders/add.comp +++ /dev/null @@ -1,14 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_binary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)]) + FLOAT_TYPE(data_b[src1_idx(idx)])); -} diff --git a/ggml/src/vulkan-shaders/clamp.comp b/ggml/src/vulkan-shaders/clamp.comp deleted file mode 100644 index 7071302a4..000000000 --- a/ggml/src/vulkan-shaders/clamp.comp +++ /dev/null @@ -1,15 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - const FLOAT_TYPE val = FLOAT_TYPE(data_a[src0_idx(idx)]); - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(val < p.param1 ? p.param1 : (val > p.param2 ? p.param2 : val)); -} diff --git a/ggml/src/vulkan-shaders/copy.comp b/ggml/src/vulkan-shaders/copy.comp deleted file mode 100644 index c26917c0f..000000000 --- a/ggml/src/vulkan-shaders/copy.comp +++ /dev/null @@ -1,18 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - -#ifndef OPTIMIZATION_ERROR_WORKAROUND - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(data_a[src0_idx(idx)]); -#else - data_d[p.d_offset + dst_idx(idx)] = data_a[src0_idx(idx)]; -#endif -} diff --git a/ggml/src/vulkan-shaders/cos.comp b/ggml/src/vulkan-shaders/cos.comp deleted file mode 100644 index f9a858cbf..000000000 --- a/ggml/src/vulkan-shaders/cos.comp +++ /dev/null @@ -1,15 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - const FLOAT_TYPE val = FLOAT_TYPE(data_a[src0_idx(idx)]); - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(cos(val)); -} diff --git a/ggml/src/vulkan-shaders/dequant_funcs.comp b/ggml/src/vulkan-shaders/dequant_funcs.comp deleted file mode 100644 index d5b989735..000000000 --- a/ggml/src/vulkan-shaders/dequant_funcs.comp +++ /dev/null @@ -1,68 +0,0 @@ -#if !defined(DATA_A_F32) && !defined(DATA_A_F16) -#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require -#endif - -#if defined(DATA_A_F32) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - return vec2(data_a[a_offset + ib], data_a[a_offset + ib + 1]); -} -#endif - -#if defined(DATA_A_F16) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - return vec2(data_a[a_offset + ib], data_a[a_offset + ib + 1]); -} -#endif - -#if defined(DATA_A_Q4_0) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - const uint vui = uint(data_a[a_offset + ib].qs[iqs]); - return (vec2(vui & 0xF, vui >> 4) - 8.0f) * d; -} -#endif - -#if defined(DATA_A_Q4_1) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - const float m = float(data_a[a_offset + ib].m); - const uint vui = uint(data_a[a_offset + ib].qs[iqs]); - return vec2(vui & 0xF, vui >> 4) * d + m; -} -#endif - -#if defined(DATA_A_Q5_0) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - const uint uint_qh = uint(data_a[a_offset + ib].qh[1]) << 16 | data_a[a_offset + ib].qh[0]; - const ivec2 qh = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); - const uint vui = uint(data_a[a_offset + ib].qs[iqs]); - return (vec2((vui & 0xF) | qh.x, (vui >> 4) | qh.y) - 16.0f) * d; -} -#endif - -#if defined(DATA_A_Q5_1) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - const float m = float(data_a[a_offset + ib].m); - const uint uint_qh = data_a[a_offset + ib].qh; - const ivec2 qh = ivec2(((uint_qh >> iqs) << 4) & 0x10, (uint_qh >> (iqs + 12)) & 0x10); - const uint vui = uint(data_a[a_offset + ib].qs[iqs]); - return vec2((vui & 0xF) | qh.x, (vui >> 4) | qh.y) * d + m; -} -#endif - -#if defined(DATA_A_Q8_0) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - return vec2(int(data_a[a_offset + ib].qs[iqs]), int(data_a[a_offset + ib].qs[iqs + 1])) * d; -} -#endif - -#if defined(DATA_A_IQ4_NL) -vec2 dequantize(uint ib, uint iqs, uint a_offset) { - const float d = float(data_a[a_offset + ib].d); - const uint vui = uint(data_a[a_offset + ib].qs[iqs]); - return vec2(kvalues_iq4nl[vui & 0xF], kvalues_iq4nl[vui >> 4]) * d; -} -#endif diff --git a/ggml/src/vulkan-shaders/dequant_q4_k.comp b/ggml/src/vulkan-shaders/dequant_q4_k.comp deleted file mode 100644 index 92acb7540..000000000 --- a/ggml/src/vulkan-shaders/dequant_q4_k.comp +++ /dev/null @@ -1,56 +0,0 @@ -#version 450 - -#include "dequant_head.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; -layout (binding = 1) writeonly buffer D {D_TYPE data_b[];}; - -void main() { - [[unroll]] for (uint wgy = 0; wgy < 256; wgy++) { - const uint i = gl_WorkGroupID.x * 256 + wgy; - if (i >= p.M * p.K / QUANT_K) { - return; - } - - const uint tid = gl_LocalInvocationID.x; - const uint il = tid / 8; - const uint ir = tid % 8; - const uint is = 2 * il; - const uint n = 4; - - const FLOAT_TYPE dall = FLOAT_TYPE(data_a[i].d.x); - const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[i].d.y); - - const uint y_idx = i * QUANT_K + 64 * il + n * ir; - const uint qs_idx = 32*il + n * ir; - - uint8_t sc; - uint8_t m; - if (is < 4) { - sc = uint8_t(data_a[i].scales[is] & 63); - m = uint8_t(data_a[i].scales[is + 4] & 63); - } else { - sc = uint8_t((data_a[i].scales[is + 4] & 0xF) | ((data_a[i].scales[is - 4] >> 6) << 4)); - m = uint8_t((data_a[i].scales[is + 4] >> 4) | ((data_a[i].scales[is ] >> 6) << 4)); - } - const FLOAT_TYPE d1 = dall * sc; - const FLOAT_TYPE m1 = dmin * m; - - if (is < 4) { - sc = uint8_t(data_a[i].scales[is + 1] & 63); - m = uint8_t(data_a[i].scales[is + 5] & 63); - } else { - sc = uint8_t((data_a[i].scales[is + 5] & 0xF) | ((data_a[i].scales[is - 3] >> 6) << 4)); - m = uint8_t((data_a[i].scales[is + 5] >> 4) | ((data_a[i].scales[is + 1] >> 6) << 4)); - } - const FLOAT_TYPE d2 = dall * sc; - const FLOAT_TYPE m2 = dmin * m; - - [[unroll]] for (uint l = 0; l < n; ++l) { - data_b[y_idx + l ] = D_TYPE(d1 * FLOAT_TYPE(data_a[i].qs[qs_idx + l] & 0xF) - m1); - data_b[y_idx + l + 32] = D_TYPE(d2 * FLOAT_TYPE(data_a[i].qs[qs_idx + l] >> 4) - m2); - } - } -} diff --git a/ggml/src/vulkan-shaders/dequant_q5_k.comp b/ggml/src/vulkan-shaders/dequant_q5_k.comp deleted file mode 100644 index f314a76d1..000000000 --- a/ggml/src/vulkan-shaders/dequant_q5_k.comp +++ /dev/null @@ -1,58 +0,0 @@ -#version 450 - -#include "dequant_head.comp" - -layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; -layout (binding = 1) writeonly buffer D {D_TYPE data_b[];}; - -void main() { - [[unroll]] for (uint wgy = 0; wgy < 256; wgy++) { - const uint i = gl_WorkGroupID.x * 256 + wgy; - if (i >= p.M * p.K / QUANT_K) { - return; - } - - const uint tid = gl_LocalInvocationID.x; - const uint il = tid / 16; - const uint ir = tid % 16; - const uint is = 2 * il; - - const FLOAT_TYPE dall = FLOAT_TYPE(data_a[i].d.x); - const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[i].d.y); - - const uint y_idx = i * QUANT_K + 64 * il + 2 * ir; - const uint qs_idx = 32*il + 2 * ir; - const uint qh_idx = 2 * ir; - - uint8_t sc; - uint8_t m; - if (is < 4) { - sc = uint8_t(data_a[i].scales[is] & 63); - m = uint8_t(data_a[i].scales[is + 4] & 63); - } else { - sc = uint8_t((data_a[i].scales[is + 4] & 0xF) | ((data_a[i].scales[is - 4] >> 6) << 4)); - m = uint8_t((data_a[i].scales[is + 4] >> 4) | ((data_a[i].scales[is ] >> 6) << 4)); - } - const FLOAT_TYPE d1 = dall * sc; - const FLOAT_TYPE m1 = dmin * m; - - if (is < 4) { - sc = uint8_t(data_a[i].scales[is + 1] & 63); - m = uint8_t(data_a[i].scales[is + 5] & 63); - } else { - sc = uint8_t((data_a[i].scales[is + 5] & 0xF) | ((data_a[i].scales[is - 3] >> 6) << 4)); - m = uint8_t((data_a[i].scales[is + 5] >> 4) | ((data_a[i].scales[is + 1] >> 6) << 4)); - } - const FLOAT_TYPE d2 = dall * sc; - const FLOAT_TYPE m2 = dmin * m; - - const uint8_t hm1 = uint8_t(1 << (2 * il )); - const uint8_t hm2 = uint8_t(1 << (2 * il + 1)); - data_b[y_idx ] = D_TYPE(d1 * FLOAT_TYPE((data_a[i].qs[qs_idx ] & 0xF) + (((data_a[i].qh[qh_idx ] & hm1) != 0) ? 16 : 0)) - m1); - data_b[y_idx + 1] = D_TYPE(d1 * FLOAT_TYPE((data_a[i].qs[qs_idx + 1] & 0xF) + (((data_a[i].qh[qh_idx + 1] & hm1) != 0) ? 16 : 0)) - m1); - data_b[y_idx + 32] = D_TYPE(d2 * FLOAT_TYPE((data_a[i].qs[qs_idx ] >> 4) + (((data_a[i].qh[qh_idx ] & hm2) != 0) ? 16 : 0)) - m2); - data_b[y_idx + 33] = D_TYPE(d2 * FLOAT_TYPE((data_a[i].qs[qs_idx + 1] >> 4) + (((data_a[i].qh[qh_idx + 1] & hm2) != 0) ? 16 : 0)) - m2); - } -} diff --git a/ggml/src/vulkan-shaders/div.comp b/ggml/src/vulkan-shaders/div.comp deleted file mode 100644 index 8cfce58b1..000000000 --- a/ggml/src/vulkan-shaders/div.comp +++ /dev/null @@ -1,14 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_binary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)]) / FLOAT_TYPE(data_b[src1_idx(idx)])); -} diff --git a/ggml/src/vulkan-shaders/generic_binary_head.comp b/ggml/src/vulkan-shaders/generic_binary_head.comp deleted file mode 100644 index b6beaff1c..000000000 --- a/ggml/src/vulkan-shaders/generic_binary_head.comp +++ /dev/null @@ -1,52 +0,0 @@ -#extension GL_EXT_shader_16bit_storage : require - -layout (push_constant) uniform parameter -{ - uint ne; - uint ne00; uint ne01; uint ne02; uint ne03; uint nb00; uint nb01; uint nb02; uint nb03; - uint ne10; uint ne11; uint ne12; uint ne13; uint nb10; uint nb11; uint nb12; uint nb13; - uint ne20; uint ne21; uint ne22; uint ne23; uint nb20; uint nb21; uint nb22; uint nb23; - uint d_offset; - float param1; float param2; int param3; -} p; - -layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; -layout (binding = 1) readonly buffer B {B_TYPE data_b[];}; -layout (binding = 2) writeonly buffer D {D_TYPE data_d[];}; - -uint get_idx() { - return gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x; -} - -uint src0_idx(uint idx) { - const uint i03 = idx / (p.ne02*p.ne01*p.ne00); - const uint i03_offset = i03 * p.ne02*p.ne01*p.ne00; - const uint i02 = (idx - i03_offset) / (p.ne01*p.ne00); - const uint i02_offset = i02*p.ne01*p.ne00; - const uint i01 = (idx - i03_offset - i02_offset) / p.ne00; - const uint i00 = idx - i03_offset - i02_offset - i01*p.ne00; - return i03*p.nb03 + i02*p.nb02 + i01*p.nb01 + i00*p.nb00; -} - -uint src1_idx(uint idx) { - const uint i03 = idx / (p.ne02*p.ne01*p.ne00); - const uint i03_offset = i03 * p.ne02*p.ne01*p.ne00; - const uint i02 = (idx - i03_offset) / (p.ne01*p.ne00); - const uint i02_offset = i02*p.ne01*p.ne00; - const uint i01 = (idx - i03_offset - i02_offset) / p.ne00; - const uint i00 = idx - i03_offset - i02_offset - i01*p.ne00; - - return (i03 % p.ne13)*p.nb13 + (i02 % p.ne12)*p.nb12 + (i01 % p.ne11)*p.nb11 + (i00 % p.ne10)*p.nb10; -} - -uint dst_idx(uint idx) { - const uint i23 = idx / (p.ne22*p.ne21*p.ne20); - const uint i23_offset = i23 * p.ne22*p.ne21*p.ne20; - const uint i22 = (idx - i23_offset) / (p.ne21*p.ne20); - const uint i22_offset = i22*p.ne21*p.ne20; - const uint i21 = (idx - i23_offset - i22_offset) / p.ne20; - const uint i20 = idx - i23_offset - i22_offset - i21*p.ne20; - return i23*p.nb23 + i22*p.nb22 + i21*p.nb21 + i20*p.nb20; -} diff --git a/ggml/src/vulkan-shaders/im2col.comp b/ggml/src/vulkan-shaders/im2col.comp deleted file mode 100644 index 4d48610a3..000000000 --- a/ggml/src/vulkan-shaders/im2col.comp +++ /dev/null @@ -1,57 +0,0 @@ -#version 450 - -#extension GL_EXT_shader_16bit_storage : require - -layout (push_constant) uniform parameter -{ - uint batch_offset; uint offset_delta; - uint IC; - uint IW; uint IH; - uint OW; uint OH; - uint KW; uint KH; - uint pelements; - uint CHW; - int s0; int s1; - int p0; int p1; - int d0; int d1; -} p; - -#include "types.comp" - -#define BLOCK_SIZE 256 - -layout(local_size_x = BLOCK_SIZE, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer X {A_TYPE data_a[];}; -layout (binding = 1) writeonly buffer D {D_TYPE data_d[];}; - -void main() { - const uint i = gl_GlobalInvocationID.x; - if (i >= p.pelements) { - return; - } - - const uint ksize = p.OW * (p.KH > 1 ? p.KW : 1); - const uint kx = i / ksize; - const uint kd = kx * ksize; - const uint ky = (i - kd) / p.OW; - const uint ix = i % p.OW; - - const uint oh = gl_GlobalInvocationID.y; - const uint batch = gl_GlobalInvocationID.z / p.IC; - const uint ic = gl_GlobalInvocationID.z % p.IC; - - const uint iiw = ix * p.s0 + kx * p.d0 - p.p0; - const uint iih = oh * p.s1 + ky * p.d1 - p.p1; - - const uint offset_dst = - ((batch * p.OH + oh) * p.OW + ix) * p.CHW + - (ic * (p.KW * p.KH) + ky * p.KW + kx); - - if (iih < 0 || iih >= p.IH || iiw < 0 || iiw >= p.IW) { - data_d[offset_dst] = D_TYPE(0.0f); - } else { - const uint offset_src = ic * p.offset_delta + batch * p.batch_offset; - data_d[offset_dst] = D_TYPE(data_a[offset_src + iih * p.IW + iiw]); - } -} diff --git a/ggml/src/vulkan-shaders/mul.comp b/ggml/src/vulkan-shaders/mul.comp deleted file mode 100644 index bfb61c92d..000000000 --- a/ggml/src/vulkan-shaders/mul.comp +++ /dev/null @@ -1,14 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_binary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)]) * FLOAT_TYPE(data_b[src1_idx(idx)])); -} diff --git a/ggml/src/vulkan-shaders/mul_mat_split_k_reduce.comp b/ggml/src/vulkan-shaders/mul_mat_split_k_reduce.comp deleted file mode 100644 index 825b91031..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_split_k_reduce.comp +++ /dev/null @@ -1,29 +0,0 @@ -#version 450 - -#extension GL_EXT_control_flow_attributes : enable - -layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer A {float data_a[];}; -layout (binding = 1) writeonly buffer D {float data_d[];}; - -layout (push_constant) uniform parameter { - uint ne; - uint k_num; -} p; - -void main() { - const uint idx = gl_GlobalInvocationID.x; - - if (idx >= p.ne) { - return; - } - - float result = 0.0f; - - [[unroll]] for (uint i = 0; i < p.k_num; i++) { - result += data_a[i * p.ne + idx]; - } - - data_d[idx] = result; -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec.comp b/ggml/src/vulkan-shaders/mul_mat_vec.comp deleted file mode 100644 index d3ccba7fc..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec.comp +++ /dev/null @@ -1,56 +0,0 @@ -#version 450 - -#ifdef FLOAT16 -#extension GL_EXT_shader_explicit_arithmetic_types_float16 : require -#endif - -#include "mul_mat_vec_base.comp" - -layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; - -layout (constant_id = 0) const uint BLOCK_SIZE = 32; - -shared FLOAT_TYPE tmp[BLOCK_SIZE]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - const uint tid = gl_LocalInvocationID.x; - - // There are not enough cols to use all threads - if (tid >= p.ncols) { - return; - } - - const uint block_size = min(p.ncols, BLOCK_SIZE); - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint y_offset = QUANT_R == 1 ? 1 : QUANT_K/2; - - tmp[tid] = FLOAT_TYPE(0.0f); - - [[unroll]] for (uint i = 0; i < p.ncols/block_size; i += 2) { - const uint col = i*block_size + 2*tid; - const uint ib = (row*p.ncols + col)/QUANT_K; // block index - const uint iqs = (col%QUANT_K)/QUANT_R; // quant index - const uint iybs = col - col%QUANT_K; // y block start index - - vec2 v = dequantize(ib, iqs, a_offset / QUANT_K); - - // matrix multiplication - tmp[tid] = fma(FLOAT_TYPE(v.x), FLOAT_TYPE(data_b[b_offset + iybs + iqs]), fma(FLOAT_TYPE(v.y), FLOAT_TYPE(data_b[b_offset + iybs + iqs + y_offset]), tmp[tid])); - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = block_size/2; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_base.comp b/ggml/src/vulkan-shaders/mul_mat_vec_base.comp deleted file mode 100644 index 5920bc936..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_base.comp +++ /dev/null @@ -1,81 +0,0 @@ -#extension GL_EXT_control_flow_attributes : enable -#extension GL_EXT_shader_16bit_storage : require -#extension GL_EXT_shader_8bit_storage : require - -#define K_QUANTS_PER_ITERATION 2 - -#ifdef MUL_MAT_ID -#define EXPERT_COUNT 8 -#endif - -#include "types.comp" - -layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; -layout (binding = 1) readonly buffer B {B_TYPE data_b[];}; -layout (binding = 2) writeonly buffer D {D_TYPE data_d[];}; -#ifdef MUL_MAT_ID -layout (binding = 3) readonly buffer IDS {int data_ids[];}; -#endif - -#include "dequant_funcs.comp" - -layout (push_constant) uniform parameter -{ - uint ncols; - uint stride_a; - uint stride_b; - uint stride_d; - - uint batch_stride_a; - uint batch_stride_b; - uint batch_stride_d; - -#ifdef MUL_MAT_ID - uint nei0; - uint ne11; -#else - uint ne02; - uint ne12; - uint broadcast2; - uint broadcast3; -#endif -} p; - -void get_offsets(out uint a_offset, out uint b_offset, out uint d_offset) { -#ifdef MUL_MAT_ID - const uint expert_idx = gl_GlobalInvocationID.y; -#else - const uint batch_idx = gl_GlobalInvocationID.y; -#endif - -#ifndef MUL_MAT_ID - const uint i13 = batch_idx / p.ne12; - const uint i12 = batch_idx % p.ne12; - - const uint i03 = i13 / p.broadcast3; - const uint i02 = i12 / p.broadcast2; - - const uint batch_idx_a = i03 * p.ne02 + i02; -#else - const uint expert_id = data_ids[expert_idx]; -#endif - - a_offset = -#ifdef MUL_MAT_ID - expert_id * p.batch_stride_a; -#else - batch_idx_a * p.batch_stride_a; -#endif - b_offset = -#ifdef MUL_MAT_ID - (expert_idx % p.ne11) * p.stride_b; -#else - batch_idx * p.batch_stride_b; -#endif - d_offset = -#ifdef MUL_MAT_ID - expert_idx * p.stride_d; -#else - batch_idx * p.batch_stride_d; -#endif -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/vulkan-shaders/mul_mat_vec_q2_k.comp deleted file mode 100644 index ec8eadcd5..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_q2_k.comp +++ /dev/null @@ -1,74 +0,0 @@ -#version 450 - -#include "mul_mat_vec_base.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -shared FLOAT_TYPE tmp[32]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint num_blocks_per_row = p.ncols / QUANT_K; - const uint ib0 = a_offset / QUANT_K + row*num_blocks_per_row; - - const uint tid = gl_LocalInvocationID.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const uint ix = gl_LocalInvocationID.x%K_QUANTS_PER_ITERATION; // 0 or 0, 1 - - const uint step = 16/K_QUANTS_PER_ITERATION; // 16 or 8 - - const uint v_im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const uint v_in = tid - step*v_im; // 0...15 or 0...7 - - const uint l0 = K_QUANTS_PER_ITERATION*v_in; // 0...15 - const uint q_offset = 32*v_im + l0; - const uint s_offset = 8*v_im; - const uint y_offset = 128*v_im + l0; - - tmp[16 * ix + tid] = FLOAT_TYPE(0.0); // partial sum for thread in warp - - [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - const uint y_idx = i * QUANT_K + y_offset; - - const FLOAT_TYPE dall = FLOAT_TYPE(data_a[ib0 + i].d.x); - const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[ib0 + i].d.y); - - FLOAT_TYPE sum1 = FLOAT_TYPE(0.0); - FLOAT_TYPE sum2 = FLOAT_TYPE(0.0); - for (int l = 0; l < K_QUANTS_PER_ITERATION; ++l) { - sum1 = fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 0]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 0] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l + 0] >> 0) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 16]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 1] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l +16] >> 0) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 32]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 2] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l + 0] >> 2) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 48]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 3] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l +16] >> 2) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 64]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 4] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l + 0] >> 4) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 80]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 5] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l +16] >> 4) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 96]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 6] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l + 0] >> 6) & 3), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l +112]), FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 7] & 0xF) * FLOAT_TYPE((data_a[ib0 + i].qs[q_offset + l +16] >> 6) & 3), sum1)))))))); - sum2 = fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 0]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 0] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 16]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 1] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 32]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 2] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 48]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 3] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 64]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 4] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 80]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 5] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 96]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 6] >> 4) & 0xF), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l +112]), FLOAT_TYPE((data_a[ib0 + i].scales[s_offset + 7] >> 4) & 0xF), sum2)))))))); - } - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(dall, sum1, fma(-dmin, sum2, tmp[tmp_idx])); - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = 16; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/vulkan-shaders/mul_mat_vec_q3_k.comp deleted file mode 100644 index 3ca4ad85a..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_q3_k.comp +++ /dev/null @@ -1,67 +0,0 @@ -#version 450 - -#include "mul_mat_vec_base.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -shared FLOAT_TYPE tmp[32]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint num_blocks_per_row = p.ncols / QUANT_K; - const uint ib0 = a_offset / QUANT_K + row*num_blocks_per_row; - - const uint tid = gl_LocalInvocationID.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const uint ix = gl_LocalInvocationID.x%K_QUANTS_PER_ITERATION; // 0 or 0, 1 - - const uint step = 16/K_QUANTS_PER_ITERATION; // 16 or 8 - - const uint v_im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const uint v_in = tid - step*v_im; // 0...15 or 0...7 - - const uint8_t m = uint8_t(1 << (4 * v_im)); - - const uint l0 = K_QUANTS_PER_ITERATION*v_in; // 0...15 - const uint q_offset = 32*v_im + l0; - const uint y_offset = 128*v_im + l0; - - tmp[16 * ix + tid] = FLOAT_TYPE(0.0); // partial sum for thread in warp - - const uint s_shift = 4 * v_im; - - [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - const uint y_idx = i * QUANT_K + y_offset; - - const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d); - - FLOAT_TYPE sum = FLOAT_TYPE(0.0); - for (int l = 0; l < K_QUANTS_PER_ITERATION; ++l) { - sum = fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 0]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[0] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[ 8] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l ] ) & 3) - (((data_a[ib0 + i].hmask[l0 + l ] & (m << 0)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 32]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[2] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[10] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l ] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l ] & (m << 1)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 64]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[4] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[ 8] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l ] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l ] & (m << 2)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 96]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[6] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[10] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l ] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l ] & (m << 3)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 16]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[1] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[ 9] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 0)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 48]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[3] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[11] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 1)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l + 80]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[5] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[ 9] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 2)) != 0) ? 0 : 4)), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l +112]) * FLOAT_TYPE(int8_t(((data_a[ib0 + i].scales[7] >> s_shift) & 0xF) | ((data_a[ib0 + i].scales[11] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 3)) != 0) ? 0 : 4)), sum)))))))); - } - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(d, sum, tmp[tmp_idx]); - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = 16; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/vulkan-shaders/mul_mat_vec_q4_k.comp deleted file mode 100644 index d91e00e10..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_q4_k.comp +++ /dev/null @@ -1,118 +0,0 @@ -#version 450 - -#include "mul_mat_vec_base.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -shared FLOAT_TYPE tmp[32]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint num_blocks_per_row = p.ncols / QUANT_K; - const uint ib0 = a_offset / QUANT_K + row*num_blocks_per_row; - - const uint tid = gl_LocalInvocationID.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const uint ix = gl_LocalInvocationID.x%K_QUANTS_PER_ITERATION; // 0 or 0, 1 - - const uint step = 8/K_QUANTS_PER_ITERATION; // 8 or 4 - - const uint il = tid/step; // 0...3 - const uint ir = tid - step*il; // 0...7 or 0...3 - const uint n = 2 * K_QUANTS_PER_ITERATION; // 2 or 4 - - const uint v_im = il / 2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 - const uint v_in = il % 2; - - const uint l0 = n * (2 * ir + v_in); // 0...15 - const uint q_offset = 32*v_im + l0; - const uint y_offset = 64*v_im + l0; - - tmp[16 * ix + tid] = FLOAT_TYPE(0.0); // partial sum for thread in warp - - [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - const uint y1_idx = i * QUANT_K + y_offset; - const uint y2_idx = y1_idx + 128; - - const FLOAT_TYPE dall = FLOAT_TYPE(data_a[ib0 + i].d.x); - const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[ib0 + i].d.y); - - const uint8_t sc0 = uint8_t( data_a[ib0 + i].scales[v_im * 2 ] & 0x3f); - const uint8_t sc1 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 1] & 0x3f); - const uint8_t sc2 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 4] & 0x3f); - const uint8_t sc3 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 5] & 0x3f); - const uint8_t sc4 = uint8_t(( data_a[ib0 + i].scales[v_im * 2 + 8] & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 ] & 0xc0) >> 2)); - const uint8_t sc5 = uint8_t(( data_a[ib0 + i].scales[v_im * 2 + 9] & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 1] & 0xc0) >> 2)); - const uint8_t sc6 = uint8_t(((data_a[ib0 + i].scales[v_im * 2 + 8] >> 4) & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 4] & 0xc0) >> 2)); - const uint8_t sc7 = uint8_t(((data_a[ib0 + i].scales[v_im * 2 + 9] >> 4) & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 5] & 0xc0) >> 2)); - -#if K_QUANTS_PER_ITERATION == 2 - const uint8_t q4_0 = uint8_t(data_a[ib0 + i].qs[q_offset ] & 0xf); - const uint8_t q4_1 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] & 0xf); - const uint8_t q4_2 = uint8_t(data_a[ib0 + i].qs[q_offset + 2] & 0xf); - const uint8_t q4_3 = uint8_t(data_a[ib0 + i].qs[q_offset + 3] & 0xf); - const uint8_t q4_4 = uint8_t(data_a[ib0 + i].qs[q_offset ] >> 4); - const uint8_t q4_5 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] >> 4); - const uint8_t q4_6 = uint8_t(data_a[ib0 + i].qs[q_offset + 2] >> 4); - const uint8_t q4_7 = uint8_t(data_a[ib0 + i].qs[q_offset + 3] >> 4); - const uint8_t q4_8 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] & 0xf); - const uint8_t q4_9 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] & 0xf); - const uint8_t q4_10 = uint8_t(data_a[ib0 + i].qs[q_offset + 66] & 0xf); - const uint8_t q4_11 = uint8_t(data_a[ib0 + i].qs[q_offset + 67] & 0xf); - const uint8_t q4_12 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] >> 4); - const uint8_t q4_13 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] >> 4); - const uint8_t q4_14 = uint8_t(data_a[ib0 + i].qs[q_offset + 66] >> 4); - const uint8_t q4_15 = uint8_t(data_a[ib0 + i].qs[q_offset + 67] >> 4); - - const FLOAT_TYPE sx = fma(FLOAT_TYPE(data_b[b_offset + y1_idx]), q4_0, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 1]), q4_1, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 2]), q4_2, FLOAT_TYPE(data_b[b_offset + y1_idx + 3]) * q4_3))); - const FLOAT_TYPE sy = fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]), q4_4, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 33]), q4_5, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 34]), q4_6, FLOAT_TYPE(data_b[b_offset + y1_idx + 35]) * q4_7))); - const FLOAT_TYPE sz = fma(FLOAT_TYPE(data_b[b_offset + y2_idx]), q4_8, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 1]), q4_9, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 2]), q4_10, FLOAT_TYPE(data_b[b_offset + y2_idx + 3]) * q4_11))); - const FLOAT_TYPE sw = fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 32]), q4_12, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 33]), q4_13, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 34]), q4_14, FLOAT_TYPE(data_b[b_offset + y2_idx + 35]) * q4_15))); - const FLOAT_TYPE smin = - fma(FLOAT_TYPE(data_b[b_offset + y1_idx ]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx ]), sc6, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 32]), sc7, - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 1]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 33]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 1]), sc6, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 33]), sc7, - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 2]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 34]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 2]), sc6, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 34]), sc7, - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 3]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 35]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 3]), sc6, FLOAT_TYPE(data_b[b_offset + y2_idx + 35]) * sc7))))))))))))))); - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, tmp[tmp_idx])); -#else - const uint8_t q4_0 = uint8_t(data_a[ib0 + i].qs[q_offset ] & 0xf); - const uint8_t q4_1 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] & 0xf); - const uint8_t q4_2 = uint8_t(data_a[ib0 + i].qs[q_offset ] >> 4); - const uint8_t q4_3 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] >> 4); - const uint8_t q4_4 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] & 0xf); - const uint8_t q4_5 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] & 0xf); - const uint8_t q4_6 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] >> 4); - const uint8_t q4_7 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] >> 4); - - const FLOAT_TYPE sx = fma(FLOAT_TYPE(data_b[b_offset + y1_idx ]), q4_0, FLOAT_TYPE(data_b[b_offset + y1_idx + 1]) * q4_1); - const FLOAT_TYPE sy = fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]), q4_2, FLOAT_TYPE(data_b[b_offset + y1_idx + 33]) * q4_3); - const FLOAT_TYPE sz = fma(FLOAT_TYPE(data_b[b_offset + y2_idx ]), q4_4, FLOAT_TYPE(data_b[b_offset + y2_idx + 1]) * q4_5); - const FLOAT_TYPE sw = fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 32]), q4_6, FLOAT_TYPE(data_b[b_offset + y2_idx + 33]) * q4_7); - const FLOAT_TYPE smin = - fma(FLOAT_TYPE(data_b[b_offset + y1_idx ]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx ]), sc6, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 32]), sc7, - + fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 1]), sc2, fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 33]), sc3, fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 1]), sc6, FLOAT_TYPE(data_b[b_offset + y2_idx + 33]) * sc7))))))); - - tmp[16 * ix + tid] += FLOAT_TYPE(dall * (sx * FLOAT_TYPE(data_a[ib0 + i].scales[v_im] & 0x3f) + sy * FLOAT_TYPE(data_a[ib0 + i].scales[v_im + 1] & 0x3f) + - sz * FLOAT_TYPE((data_a[ib0 + i].scales[v_im + 4] & 0x0f) | ((data_a[ib0 + i].scales[v_im] & 0xc0) >> 2)) + sw * FLOAT_TYPE((data_a[ib0 + i].scales[v_im + 5] & 0x0f) | ((data_a[ib0 + i].scales[v_im + 1] & 0xc0) >> 2))) - dmin * smin); - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(dall, (fma(sx, FLOAT_TYPE(data_a[ib0 + i].scales[v_im] & 0x3f), fma(sy, FLOAT_TYPE(data_a[ib0 + i].scales[v_im + 1] & 0x3f), - fma(sz, FLOAT_TYPE((data_a[ib0 + i].scales[v_im + 4] & 0x0f) | ((data_a[ib0 + i].scales[v_im] & 0xc0) >> 2)), fma(sw, FLOAT_TYPE((data_a[ib0 + i].scales[v_im + 5] & 0x0f) | ((data_a[ib0 + i].scales[v_im + 1] & 0xc0) >> 2))))))), fma(-dmin, smin, tmp[tmp_idx])); -#endif - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = 16; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/vulkan-shaders/mul_mat_vec_q5_k.comp deleted file mode 100644 index 2306785af..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_q5_k.comp +++ /dev/null @@ -1,109 +0,0 @@ -#version 450 - -#include "mul_mat_vec_base.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -shared FLOAT_TYPE tmp[32]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint num_blocks_per_row = p.ncols / QUANT_K; - const uint ib0 = a_offset / QUANT_K + row*num_blocks_per_row; - - const uint tid = gl_LocalInvocationID.x/2; // 0...31 or 0...16 - const uint ix = gl_LocalInvocationID.x%2; // 0 or 0, 1 - - const uint il = tid/4; // 0...3 - const uint ir = tid - 4*il; // 0...7 or 0...3 - - const uint v_im = il / 2; // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224 - const uint v_in = il % 2; - - const uint l0 = 4*ir + 2*v_in; // 0...15 - const uint q_offset = 32*v_im + l0; - const uint y_offset = 64*v_im + l0; - - const uint8_t hm1 = uint8_t(1 << (2*v_im)); - const uint8_t hm2 = uint8_t(hm1 << 4); - - tmp[16 * ix + tid] = FLOAT_TYPE(0.0); // partial sum for thread in warp - - [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += 2) { - const uint y1_idx = i * QUANT_K + y_offset; - const uint y2_idx = y1_idx + 128; - - const FLOAT_TYPE dall = FLOAT_TYPE(data_a[ib0 + i].d.x); - const FLOAT_TYPE dmin = FLOAT_TYPE(data_a[ib0 + i].d.y); - - const uint8_t sc0 = uint8_t( data_a[ib0 + i].scales[v_im * 2 ] & 0x3f); - const uint8_t sc1 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 1] & 0x3f); - const uint8_t sc2 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 4] & 0x3f); - const uint8_t sc3 = uint8_t( data_a[ib0 + i].scales[v_im * 2 + 5] & 0x3f); - const uint8_t sc4 = uint8_t(( data_a[ib0 + i].scales[v_im * 2 + 8] & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 ] & 0xc0) >> 2)); - const uint8_t sc5 = uint8_t(( data_a[ib0 + i].scales[v_im * 2 + 9] & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 1] & 0xc0) >> 2)); - const uint8_t sc6 = uint8_t(((data_a[ib0 + i].scales[v_im * 2 + 8] >> 4) & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 4] & 0xc0) >> 2)); - const uint8_t sc7 = uint8_t(((data_a[ib0 + i].scales[v_im * 2 + 9] >> 4) & 0x0f) | ((data_a[ib0 + i].scales[v_im * 2 + 5] & 0xc0) >> 2)); - - const uint8_t q4_0 = uint8_t(data_a[ib0 + i].qs[q_offset ] & 0xf); - const uint8_t q4_1 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] & 0xf); - const uint8_t q4_2 = uint8_t(data_a[ib0 + i].qs[q_offset + 16] & 0xf); - const uint8_t q4_3 = uint8_t(data_a[ib0 + i].qs[q_offset + 17] & 0xf); - const uint8_t q4_4 = uint8_t(data_a[ib0 + i].qs[q_offset ] >> 4); - const uint8_t q4_5 = uint8_t(data_a[ib0 + i].qs[q_offset + 1] >> 4); - const uint8_t q4_6 = uint8_t(data_a[ib0 + i].qs[q_offset + 16] >> 4); - const uint8_t q4_7 = uint8_t(data_a[ib0 + i].qs[q_offset + 17] >> 4); - const uint8_t q4_8 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] & 0xf); - const uint8_t q4_9 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] & 0xf); - const uint8_t q4_10 = uint8_t(data_a[ib0 + i].qs[q_offset + 80] & 0xf); - const uint8_t q4_11 = uint8_t(data_a[ib0 + i].qs[q_offset + 81] & 0xf); - const uint8_t q4_12 = uint8_t(data_a[ib0 + i].qs[q_offset + 64] >> 4); - const uint8_t q4_13 = uint8_t(data_a[ib0 + i].qs[q_offset + 65] >> 4); - const uint8_t q4_14 = uint8_t(data_a[ib0 + i].qs[q_offset + 80] >> 4); - const uint8_t q4_15 = uint8_t(data_a[ib0 + i].qs[q_offset + 81] >> 4); - - const FLOAT_TYPE sx = - fma(FLOAT_TYPE(data_b[b_offset + y1_idx ]), (q4_0 + (((data_a[ib0 + i].qh[l0 ] & hm1) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 1]), (q4_1 + (((data_a[ib0 + i].qh[l0 + 1] & hm1) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 16]), (q4_2 + (((data_a[ib0 + i].qh[l0 + 16] & hm1) != 0) ? 16 : 0)), - FLOAT_TYPE(data_b[b_offset + y1_idx + 17]) * (q4_3 + (((data_a[ib0 + i].qh[l0 + 17] & hm1) != 0) ? 16 : 0))))); - const FLOAT_TYPE sy = - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]), (q4_4 + (((data_a[ib0 + i].qh[l0 ] & (hm1 << 1)) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 33]), (q4_5 + (((data_a[ib0 + i].qh[l0 + 1] & (hm1 << 1)) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 48]), (q4_6 + (((data_a[ib0 + i].qh[l0 + 16] & (hm1 << 1)) != 0) ? 16 : 0)), - FLOAT_TYPE(data_b[b_offset + y1_idx + 49]) * (q4_7 + (((data_a[ib0 + i].qh[l0 + 17] & (hm1 << 1)) != 0) ? 16 : 0))))); - const FLOAT_TYPE sz = - fma(FLOAT_TYPE(data_b[b_offset + y2_idx ]), (q4_8 + (((data_a[ib0 + i].qh[l0 ] & hm2) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 1]), (q4_9 + (((data_a[ib0 + i].qh[l0 + 1] & hm2) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 16]), (q4_10 + (((data_a[ib0 + i].qh[l0 + 16] & hm2) != 0) ? 16 : 0)), - FLOAT_TYPE(data_b[b_offset + y2_idx + 17]) * (q4_11 + (((data_a[ib0 + i].qh[l0 + 17] & hm2) != 0) ? 16 : 0))))); - const FLOAT_TYPE sw = - fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 32]), (q4_12 + (((data_a[ib0 + i].qh[l0 ] & (hm2 << 1)) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 33]), (q4_13 + (((data_a[ib0 + i].qh[l0 + 1] & (hm2 << 1)) != 0) ? 16 : 0)), - fma(FLOAT_TYPE(data_b[b_offset + y2_idx + 48]), (q4_14 + (((data_a[ib0 + i].qh[l0 + 16] & (hm2 << 1)) != 0) ? 16 : 0)), - FLOAT_TYPE(data_b[b_offset + y2_idx + 49]) * (q4_15 + (((data_a[ib0 + i].qh[l0 + 17] & (hm2 << 1)) != 0) ? 16 : 0))))); - const FLOAT_TYPE smin = - fma(FLOAT_TYPE(data_b[b_offset + y1_idx ]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 1 ]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 16]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 17]), sc2, - fma(FLOAT_TYPE(data_b[b_offset + y1_idx + 32]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 33]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 48]) + FLOAT_TYPE(data_b[b_offset + y1_idx + 49]), sc3, - fma(FLOAT_TYPE(data_b[b_offset + y2_idx ]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 1 ]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 16]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 17]), sc6, - (FLOAT_TYPE(data_b[b_offset + y2_idx + 32]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 33]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 48]) + FLOAT_TYPE(data_b[b_offset + y2_idx + 49])) * sc7))); - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, tmp[tmp_idx])); - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = 16; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/vulkan-shaders/mul_mat_vec_q6_k.comp deleted file mode 100644 index 95c286eeb..000000000 --- a/ggml/src/vulkan-shaders/mul_mat_vec_q6_k.comp +++ /dev/null @@ -1,79 +0,0 @@ -#version 450 - -#include "mul_mat_vec_base.comp" - -layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in; - -shared FLOAT_TYPE tmp[32]; - -void main() { - const uint row = gl_WorkGroupID.x + gl_NumWorkGroups.x * gl_WorkGroupID.z; - - uint a_offset, b_offset, d_offset; - get_offsets(a_offset, b_offset, d_offset); - - const uint num_blocks_per_row = p.ncols / QUANT_K; - const uint ib0 = a_offset / QUANT_K + row*num_blocks_per_row; - - const uint tid = gl_LocalInvocationID.x/K_QUANTS_PER_ITERATION; // 0...31 or 0...16 - const uint ix = gl_LocalInvocationID.x%K_QUANTS_PER_ITERATION; // 0 or 0, 1 - - const uint step = 16/K_QUANTS_PER_ITERATION; // 16 or 8 - - const uint v_im = tid/step; // 0 or 1. 0 computes 0..., 1 computes 128... - const uint v_in = tid - step*v_im; // 0...15 or 0...7 - -#if K_QUANTS_PER_ITERATION == 1 - const uint l0 = v_in; // 0...15 - const uint is = 0; -#else - const uint l0 = 4 * v_in; // 0, 4, 8, ..., 28 - const uint is = v_in / 4; -#endif - - const uint ql_offset = 64*v_im + l0; - const uint qh_offset = 32*v_im + l0; - const uint s_offset = 8*v_im + is; - const uint y_offset = 128*v_im + l0; - - tmp[16 * ix + tid] = FLOAT_TYPE(0.0); // partial sum for thread in warp - - [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += K_QUANTS_PER_ITERATION) { - const uint y_idx = i * QUANT_K + y_offset; - - const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d); - -#if K_QUANTS_PER_ITERATION == 1 - const uint tmp_idx = 16 * ix + tid; - tmp[tmp_idx] = fma(FLOAT_TYPE(data_b[b_offset + y_idx + 0]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 0]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 0] & 0xF) | ((data_a[ib0 + i].qh[qh_offset + 0] & 0x03) << 4)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 16]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 1]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 16] & 0xF) | ((data_a[ib0 + i].qh[qh_offset + 16] & 0x03) << 4)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 32]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 2]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 32] & 0xF) | ((data_a[ib0 + i].qh[qh_offset + 0] & 0x0c) << 2)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 48]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 3]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 48] & 0xF) | ((data_a[ib0 + i].qh[qh_offset + 16] & 0x0c) << 2)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 64]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 4]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 0] >> 4) | ((data_a[ib0 + i].qh[qh_offset + 0] & 0x30) >> 0)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 80]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 5]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 16] >> 4) | ((data_a[ib0 + i].qh[qh_offset + 16] & 0x30) >> 0)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + 96]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 6]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 32] >> 4) | ((data_a[ib0 + i].qh[qh_offset + 0] & 0xc0) >> 2)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx +112]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 7]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + 48] >> 4) | ((data_a[ib0 + i].qh[qh_offset + 16] & 0xc0) >> 2)) - 32), tmp[tmp_idx])))))))); -#else - FLOAT_TYPE sum = FLOAT_TYPE(0.0); - [[unroll]] for (int l = 0; l < 4; ++l) { - sum = fma(FLOAT_TYPE(data_b[b_offset + y_idx + l+ 0]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 0]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + l+ 0] & 0xF) | (((data_a[ib0 + i].qh[qh_offset + l] >> 0) & 3) << 4)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l+32]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 2]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + l+32] & 0xF) | (((data_a[ib0 + i].qh[qh_offset + l] >> 2) & 3) << 4)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l+64]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 4]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + l+ 0] >> 4) | (((data_a[ib0 + i].qh[qh_offset + l] >> 4) & 3) << 4)) - 32), - fma(FLOAT_TYPE(data_b[b_offset + y_idx + l+96]) * FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 6]) * d, FLOAT_TYPE(int8_t((data_a[ib0 + i].ql[ql_offset + l+32] >> 4) | (((data_a[ib0 + i].qh[qh_offset + l] >> 6) & 3) << 4)) - 32), sum)))); - } - tmp[16 * ix + tid] += sum; -#endif - } - - // sum up partial sums and write back result - barrier(); - [[unroll]] for (uint s = 16; s > 0; s >>= 1) { - if (tid < s) { - tmp[tid] += tmp[tid + s]; - } - barrier(); - } - if (tid == 0) { - data_d[d_offset + row] = D_TYPE(tmp[0]); - } -} diff --git a/ggml/src/vulkan-shaders/scale.comp b/ggml/src/vulkan-shaders/scale.comp deleted file mode 100644 index 5cd2f668d..000000000 --- a/ggml/src/vulkan-shaders/scale.comp +++ /dev/null @@ -1,14 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(FLOAT_TYPE(data_a[src0_idx(idx)]) * FLOAT_TYPE(p.param1)); -} diff --git a/ggml/src/vulkan-shaders/sin.comp b/ggml/src/vulkan-shaders/sin.comp deleted file mode 100644 index 7faf9be93..000000000 --- a/ggml/src/vulkan-shaders/sin.comp +++ /dev/null @@ -1,15 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - const FLOAT_TYPE val = FLOAT_TYPE(data_a[src0_idx(idx)]); - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(sin(val)); -} diff --git a/ggml/src/vulkan-shaders/soft_max.comp b/ggml/src/vulkan-shaders/soft_max.comp deleted file mode 100644 index 0bd51ecab..000000000 --- a/ggml/src/vulkan-shaders/soft_max.comp +++ /dev/null @@ -1,106 +0,0 @@ -#version 450 - -#extension GL_EXT_shader_16bit_storage : require - -layout (push_constant) uniform parameter -{ - uint KX; - uint KY; - float scale; - float max_bias; - float m0; - float m1; - uint n_head_log2; -} p; - -#include "types.comp" - -#extension GL_EXT_control_flow_attributes : enable -#define BLOCK_SIZE 512 - -layout(local_size_x = BLOCK_SIZE, local_size_y = 1, local_size_z = 1) in; - -layout (binding = 0) readonly buffer X {A_TYPE data_a[];}; -layout (binding = 1) readonly buffer Y {B_TYPE data_b[];}; -layout (binding = 2) buffer D {D_TYPE data_d[];}; - -shared FLOAT_TYPE vals[BLOCK_SIZE]; - -void main() { - const uint tid = gl_LocalInvocationID.x; - const uint rowx = gl_WorkGroupID.z * 262144 + gl_WorkGroupID.y * 512 + gl_WorkGroupID.x; - const uint rowy = rowx % p.KY; - - float slope = 1.0f; - - // ALiBi - if (p.max_bias > 0.0f) { - const uint h = rowx/p.KY; // head index - - const float base = h < p.n_head_log2 ? p.m0 : p.m1; - const uint exp = h < p.n_head_log2 ? h + 1 : 2*(h - p.n_head_log2) + 1; - - slope = pow(base, exp); - } - - // Find max - FLOAT_TYPE max_val = uintBitsToFloat(0xFF800000); - - [[unroll]] for (uint col0 = 0; col0 < p.KX; col0 += BLOCK_SIZE) { - const uint col = col0 + tid; - - if (col >= p.KX) { - break; - } - - max_val = max(max_val, FLOAT_TYPE(data_a[rowx * p.KX + col]) * p.scale + (p.KY > 0 ? slope * FLOAT_TYPE(data_b[rowy * p.KX + col]) : FLOAT_TYPE(0.0f))); - } - vals[tid] = max_val; - - barrier(); - [[unroll]] for (int s = BLOCK_SIZE / 2; s > 0; s >>= 1) { - if (tid < s) { - vals[tid] = max(vals[tid], vals[tid + s]); - } - barrier(); - } - - max_val = vals[0]; - barrier(); - - // Sum up values - vals[tid] = FLOAT_TYPE(0.0f); - - [[unroll]] for (uint col0 = 0; col0 < p.KX; col0 += BLOCK_SIZE) { - const uint col = col0 + tid; - - if (col >= p.KX) { - break; - } - - const uint i = rowx * p.KX + col; - const FLOAT_TYPE val = exp(FLOAT_TYPE(data_a[i]) * p.scale + (p.KY > 0 ? slope * FLOAT_TYPE(data_b[rowy * p.KX + col]) : FLOAT_TYPE(0.0f)) - max_val); - vals[tid] += val; - data_d[i] = D_TYPE(val); - } - - barrier(); - [[unroll]] for (int s = BLOCK_SIZE / 2; s > 0; s >>= 1) { - if (tid < s) { - vals[tid] += vals[tid + s]; - } - barrier(); - } - - const D_TYPE divisor = D_TYPE(vals[0]); - - [[unroll]] for (uint col0 = 0; col0 < p.KX; col0 += BLOCK_SIZE) { - const uint col = col0 + tid; - - if (col >= p.KX) { - break; - } - - data_d[rowx*p.KX + col] /= divisor; - } -} diff --git a/ggml/src/vulkan-shaders/square.comp b/ggml/src/vulkan-shaders/square.comp deleted file mode 100644 index 1fa118c99..000000000 --- a/ggml/src/vulkan-shaders/square.comp +++ /dev/null @@ -1,15 +0,0 @@ -#version 450 - -#include "types.comp" -#include "generic_unary_head.comp" - -void main() { - const uint idx = get_idx(); - - if (idx >= p.ne) { - return; - } - - const FLOAT_TYPE val = FLOAT_TYPE(data_a[src0_idx(idx)]); - data_d[p.d_offset + dst_idx(idx)] = D_TYPE(val * val); -} diff --git a/ggml/src/vulkan-shaders/types.comp b/ggml/src/vulkan-shaders/types.comp deleted file mode 100644 index 21dce72fc..000000000 --- a/ggml/src/vulkan-shaders/types.comp +++ /dev/null @@ -1,200 +0,0 @@ -#if !defined(DATA_A_F32) && !defined(DATA_A_F16) -#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require -#endif - -#if defined(DATA_A_F32) -#define QUANT_K 1 -#define QUANT_R 1 - -#if !defined(LOAD_VEC_A) || LOAD_VEC_A == 1 -#define A_TYPE float -#elif LOAD_VEC_A == 4 -#define A_TYPE vec4 -#elif LOAD_VEC_A == 8 -#define A_TYPE mat2x4 -#endif -#endif - -#if defined(DATA_A_F16) -#define QUANT_K 1 -#define QUANT_R 1 - -#if !defined(LOAD_VEC_A) || LOAD_VEC_A == 1 -#define A_TYPE float16_t -#elif LOAD_VEC_A == 4 -#define A_TYPE f16vec4 -#elif LOAD_VEC_A == 8 -#define A_TYPE f16mat2x4 -#endif -#endif - -#if defined(DATA_A_Q4_0) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 32 -#define QUANT_R 2 - -struct block_q4_0 -{ - float16_t d; - uint8_t qs[16]; -}; - -#define A_TYPE block_q4_0 -#endif - -#if defined(DATA_A_Q4_1) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 32 -#define QUANT_R 2 - -struct block_q4_1 -{ - float16_t d; - float16_t m; - uint8_t qs[16]; -}; - -#define A_TYPE block_q4_1 -#endif - -#if defined(DATA_A_Q5_0) -#extension GL_EXT_shader_16bit_storage : require -#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require -#define QUANT_K 32 -#define QUANT_R 2 - -struct block_q5_0 -{ - float16_t d; - uint16_t qh[2]; - uint8_t qs[16]; -}; - -#define A_TYPE block_q5_0 -#endif - -#if defined(DATA_A_Q5_1) -#extension GL_EXT_shader_16bit_storage : require -#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require -#define QUANT_K 32 -#define QUANT_R 2 - -struct block_q5_1 -{ - float16_t d; - float16_t m; - uint qh; - uint8_t qs[16]; -}; - -#define A_TYPE block_q5_1 -#endif - -#if defined(DATA_A_Q8_0) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 32 -#define QUANT_R 1 - -struct block_q8_0 -{ - float16_t d; - int8_t qs[32]; -}; - -#define A_TYPE block_q8_0 -#endif - -// K-quants -#if defined(DATA_A_Q2_K) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 256 - -struct block_q2_K -{ - uint8_t scales[QUANT_K/16]; - uint8_t qs[QUANT_K/4]; - f16vec2 d; -}; - -#define A_TYPE block_q2_K -#endif - -#if defined(DATA_A_Q3_K) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 256 - -struct block_q3_K -{ - uint8_t hmask[QUANT_K/8]; - uint8_t qs[QUANT_K/4]; - uint8_t scales[12]; - float16_t d; -}; - -#define A_TYPE block_q3_K -#endif - -#if defined(DATA_A_Q4_K) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 256 - -struct block_q4_K -{ - f16vec2 d; - uint8_t scales[3*QUANT_K/64]; - uint8_t qs[QUANT_K/2]; -}; - -#define A_TYPE block_q4_K -#endif - -#if defined(DATA_A_Q5_K) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 256 - -struct block_q5_K -{ - f16vec2 d; - uint8_t scales[12]; - uint8_t qh[QUANT_K/8]; - uint8_t qs[QUANT_K/2]; -}; - -#define A_TYPE block_q5_K -#endif - -#if defined(DATA_A_Q6_K) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 256 - -struct block_q6_K -{ - uint8_t ql[QUANT_K/2]; - uint8_t qh[QUANT_K/4]; - int8_t scales[QUANT_K/16]; - float16_t d; -}; - -#define A_TYPE block_q6_K -#endif - -// IQuants - -#if defined(DATA_A_IQ4_NL) -#extension GL_EXT_shader_16bit_storage : require -#define QUANT_K 32 -#define QUANT_R 2 - -struct block_iq4_nl -{ - float16_t d; - uint8_t qs[QUANT_K/2]; -}; - -#define A_TYPE block_iq4_nl - -const int8_t kvalues_iq4nl[16] = { - int8_t(-127), int8_t(-104), int8_t(-83), int8_t(-65), int8_t(-49), int8_t(-35), int8_t(-22), int8_t(-10), - int8_t(1), int8_t(13), int8_t(25), int8_t(38), int8_t(53), int8_t(69), int8_t(89), int8_t(113) -}; -#endif diff --git a/ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp b/ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp deleted file mode 100644 index 49759c593..000000000 --- a/ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp +++ /dev/null @@ -1,604 +0,0 @@ - - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#ifdef _WIN32 - #include - #include // For _mkdir on Windows - #include // For std::replace on w64devkit -#else - #include - #include - #include -#endif - -#define ASYNCIO_CONCURRENCY 64 - -std::mutex lock; -std::vector> shader_fnames; - -std::string GLSLC = "glslc"; -std::string input_dir = "vulkan-shaders"; -std::string output_dir = "/tmp"; -std::string target_hpp = "ggml-vulkan-shaders.hpp"; -std::string target_cpp = "ggml-vulkan-shaders.cpp"; -bool no_clean = false; - -const std::vector type_names = { - "f32", - "f16", - "q4_0", - "q4_1", - "q5_0", - "q5_1", - "q8_0", - "q2_k", - "q3_k", - "q4_k", - "q5_k", - "q6_k", - "iq4_nl" -}; - -void execute_command(const std::string& command, std::string& stdout_str, std::string& stderr_str) { -#ifdef _WIN32 - HANDLE stdout_read, stdout_write; - HANDLE stderr_read, stderr_write; - SECURITY_ATTRIBUTES sa = { sizeof(SECURITY_ATTRIBUTES), NULL, TRUE }; - - if (!CreatePipe(&stdout_read, &stdout_write, &sa, 0) || - !SetHandleInformation(stdout_read, HANDLE_FLAG_INHERIT, 0)) { - throw std::runtime_error("Failed to create stdout pipe"); - } - - if (!CreatePipe(&stderr_read, &stderr_write, &sa, 0) || - !SetHandleInformation(stderr_read, HANDLE_FLAG_INHERIT, 0)) { - throw std::runtime_error("Failed to create stderr pipe"); - } - - PROCESS_INFORMATION pi; - STARTUPINFOA si = { sizeof(STARTUPINFOA) }; - si.dwFlags = STARTF_USESTDHANDLES; - si.hStdOutput = stdout_write; - si.hStdError = stderr_write; - - std::vector cmd(command.begin(), command.end()); - cmd.push_back('\0'); - - if (!CreateProcessA(NULL, cmd.data(), NULL, NULL, TRUE, 0, NULL, NULL, &si, &pi)) { - throw std::runtime_error("Failed to create process"); - } - - CloseHandle(stdout_write); - CloseHandle(stderr_write); - - std::array buffer; - DWORD bytes_read; - - while (ReadFile(stdout_read, buffer.data(), buffer.size(), &bytes_read, NULL) && bytes_read > 0) { - stdout_str.append(buffer.data(), bytes_read); - } - - while (ReadFile(stderr_read, buffer.data(), buffer.size(), &bytes_read, NULL) && bytes_read > 0) { - stderr_str.append(buffer.data(), bytes_read); - } - - CloseHandle(stdout_read); - CloseHandle(stderr_read); - WaitForSingleObject(pi.hProcess, INFINITE); - CloseHandle(pi.hProcess); - CloseHandle(pi.hThread); -#else -int stdout_pipe[2]; - int stderr_pipe[2]; - - if (pipe(stdout_pipe) != 0 || pipe(stderr_pipe) != 0) { - throw std::runtime_error("Failed to create pipes"); - } - - pid_t pid = fork(); - if (pid < 0) { - throw std::runtime_error("Failed to fork process"); - } - - if (pid == 0) { - close(stdout_pipe[0]); - close(stderr_pipe[0]); - dup2(stdout_pipe[1], STDOUT_FILENO); - dup2(stderr_pipe[1], STDERR_FILENO); - close(stdout_pipe[1]); - close(stderr_pipe[1]); - execl("/bin/sh", "sh", "-c", command.c_str(), (char*) nullptr); - _exit(EXIT_FAILURE); - } else { - close(stdout_pipe[1]); - close(stderr_pipe[1]); - - std::array buffer; - ssize_t bytes_read; - - while ((bytes_read = read(stdout_pipe[0], buffer.data(), buffer.size())) > 0) { - stdout_str.append(buffer.data(), bytes_read); - } - - while ((bytes_read = read(stderr_pipe[0], buffer.data(), buffer.size())) > 0) { - stderr_str.append(buffer.data(), bytes_read); - } - - close(stdout_pipe[0]); - close(stderr_pipe[0]); - waitpid(pid, nullptr, 0); - } -#endif -} - -bool directory_exists(const std::string& path) { - struct stat info; - if (stat(path.c_str(), &info) != 0) { - return false; // Path doesn't exist or can't be accessed - } - return (info.st_mode & S_IFDIR) != 0; // Check if it is a directory -} - -bool create_directory(const std::string& path) { -#ifdef _WIN32 - return _mkdir(path.c_str()) == 0 || errno == EEXIST; // EEXIST means the directory already exists -#else - return mkdir(path.c_str(), 0755) == 0 || errno == EEXIST; // 0755 is the directory permissions -#endif -} - -std::string to_uppercase(const std::string& input) { - std::string result = input; - for (char& c : result) { - c = std::toupper(c); - } - return result; -} - -bool string_ends_with(const std::string& str, const std::string& suffix) { - if (suffix.size() > str.size()) { - return false; - } - return std::equal(suffix.rbegin(), suffix.rend(), str.rbegin()); -} - -static const char path_separator = '/'; - -std::string join_paths(const std::string& path1, const std::string& path2) { - return path1 + path_separator + path2; -} - -std::string basename(const std::string &path) { - return path.substr(path.find_last_of("/\\") + 1); -} - -void string_to_spv(const std::string& _name, const std::string& in_fname, const std::map& defines, bool fp16 = true) { - std::string name = _name + (fp16 ? "" : "_fp32"); - std::string out_fname = join_paths(output_dir, name + ".spv"); - std::string in_path = join_paths(input_dir, in_fname); - - #ifdef _WIN32 - std::vector cmd = {GLSLC, "-fshader-stage=compute", "--target-env=vulkan1.2", "-O", "\"" + in_path + "\"", "-o", "\"" + out_fname + "\""}; - #else - std::vector cmd = {GLSLC, "-fshader-stage=compute", "--target-env=vulkan1.2", "-O", in_path, "-o", out_fname}; - #endif - - #ifdef GGML_VULKAN_SHADER_DEBUG_INFO - cmd.push_back("-g"); - #endif - - for (const auto& define : defines) { - cmd.push_back("-D" + define.first + "=" + define.second); - } - - std::string command; - for (const auto& part : cmd) { - command += part + " "; - } - - std::string stdout_str, stderr_str; - try { - // std::cout << "Executing command: "; - // for (const auto& part : cmd) { - // std::cout << part << " "; - // } - // std::cout << std::endl; - - execute_command(command, stdout_str, stderr_str); - if (!stderr_str.empty()) { - std::cerr << "cannot compile " << name << "\n\n" << command << "\n\n" << stderr_str << std::endl; - return; - } - - std::lock_guard guard(lock); - shader_fnames.push_back(std::make_pair(name, out_fname)); - } catch (const std::exception& e) { - std::cerr << "Error executing command for " << name << ": " << e.what() << std::endl; - } -} - -std::map merge_maps(const std::map& a, const std::map& b) { - std::map result = a; - result.insert(b.begin(), b.end()); - return result; -} - -void matmul_shaders(std::vector>& tasks, bool fp16, bool matmul_id) { - std::string load_vec = fp16 ? "8" : "4"; - std::string aligned_b_type_f32 = fp16 ? "mat2x4" : "vec4"; - std::string aligned_b_type_f16 = fp16 ? "f16mat2x4" : "f16vec4"; - - std::map base_dict = {{"FLOAT_TYPE", fp16 ? "float16_t" : "float"}}; - std::string shader_name = "matmul"; - - if (matmul_id) { - base_dict["MUL_MAT_ID"] = "1"; - shader_name = "matmul_id"; - } - - if (fp16) { - base_dict["FLOAT16"] = "1"; - } - - // Shaders with f16 B_TYPE - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_f32_f16", "mul_mm.comp", merge_maps(base_dict, {{"DATA_A_F32", "1"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}}), fp16); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_f32_f16_aligned", "mul_mm.comp", merge_maps(base_dict, {{"DATA_A_F32", "1"}, {"LOAD_VEC_A", load_vec}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f16}, {"D_TYPE", "float"}}), fp16); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_f16", "mul_mm.comp", merge_maps(base_dict, {{"DATA_A_F16", "1"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}}), fp16); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_f16_aligned", "mul_mm.comp", merge_maps(base_dict, {{"DATA_A_F16", "1"}, {"LOAD_VEC_A", load_vec}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f16}, {"D_TYPE", "float"}}), fp16); - })); - - for (const auto& tname : type_names) { - std::string data_a_key = "DATA_A_" + to_uppercase(tname); - // For unaligned, load one at a time for f32/f16, or two at a time for quants - std::string load_vec_a_unaligned = (tname == "f32" || tname == "f16") ? "1" : "2"; - // For aligned matmul loads - std::string load_vec_a = (tname == "f32" || tname == "f16") ? load_vec : "2"; - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_" + tname + "_f32", "mul_mm.comp", merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a_unaligned}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}), fp16); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv(shader_name + "_" + tname + "_f32_aligned", "mul_mm.comp", merge_maps(base_dict, {{data_a_key, "1"}, {"LOAD_VEC_A", load_vec_a}, {"LOAD_VEC_B", load_vec}, {"B_TYPE", aligned_b_type_f32}, {"D_TYPE", "float"}}), fp16); - })); - } -} - -void process_shaders(std::vector>& tasks) { - std::cout << "ggml_vulkan: Generating and compiling shaders to SPIR-V" << std::endl; - std::map base_dict = {{"FLOAT_TYPE", "float"}}; - - for (const auto& fp16 : {false, true}) { - matmul_shaders(tasks, fp16, false); - matmul_shaders(tasks, fp16, true); - } - - for (const auto& tname : type_names) { - // mul mat vec - std::string data_a_key = "DATA_A_" + to_uppercase(tname); - std::string shader = (string_ends_with(tname, "_k")) ? "mul_mat_vec_" + tname + ".comp" : "mul_mat_vec.comp"; - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("mul_mat_vec_" + tname + "_f32_f32", shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("mul_mat_vec_" + tname + "_f16_f32", shader, merge_maps(base_dict, {{data_a_key, "1"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}})); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("mul_mat_vec_id_" + tname + "_f32", shader, merge_maps(base_dict, {{"MUL_MAT_ID", "1"}, {data_a_key, "1"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - - // Dequant shaders - if (tname != "f16") { - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("dequant_" + tname, "dequant_" + tname + ".comp", merge_maps(base_dict, {{data_a_key, "1"}, {"D_TYPE", "float16_t"}})); - })); - } - - if (!string_ends_with(tname, "_k")) { - shader = (tname == "f32" || tname == "f16") ? "get_rows.comp" : "get_rows_quant.comp"; - - if (tname == "f16") { - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("get_rows_" + tname, shader, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); - })); - } else { - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("get_rows_" + tname, shader, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float16_t"}}); - })); - } - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("get_rows_" + tname + "_f32", shader, {{data_a_key, "1"}, {"B_TYPE", "int"}, {"D_TYPE", "float"}}); - })); - } - } - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("mul_mat_vec_p021_f16_f32", "mul_mat_vec_p021.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("mul_mat_vec_nc_f16_f32", "mul_mat_vec_nc.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - // Norms - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("norm_f32", "norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("group_norm_f32", "group_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("rms_norm_f32", "rms_norm.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("cpy_f32_f32", "copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("cpy_f32_f16", "copy.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("cpy_f16_f16", "copy.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("add_f32", "add.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("add_f16_f32_f16", "add.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float"}, {"D_TYPE", "float16_t"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("acc_f32", "acc.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("split_k_reduce", "mul_mat_split_k_reduce.comp", {}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("mul_f32", "mul.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("div_f32", "div.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("repeat_f32", "repeat.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("scale_f32", "scale.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("sqr_f32", "square.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("sin_f32", "sin.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("cos_f32", "cos.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("clamp_f32", "clamp.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}, {"FLOAT_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("pad_f32", "pad.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("concat_f32", "concat.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("concat_f16", "concat.comp", {{"A_TYPE", "float16_t"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}, {"OPTIMIZATION_ERROR_WORKAROUND", "1"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("concat_i32", "concat.comp", {{"A_TYPE", "int"}, {"B_TYPE", "int"}, {"D_TYPE", "int"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("upscale_f32", "upscale.comp", {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("gelu_f32", "gelu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("gelu_quick_f32", "gelu_quick.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("silu_f32", "silu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("relu_f32", "relu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("leaky_relu_f32", "leaky_relu.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("tanh_f32", "tanh.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("diag_mask_inf_f32", "diag_mask_inf.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("soft_max_f32", "soft_max.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("soft_max_f32_f16", "soft_max.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"B_TYPE", "float16_t"}, {"D_TYPE", "float"}})); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("rope_norm_f32", "rope_norm.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("rope_norm_f16", "rope_norm.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("rope_neox_f32", "rope_neox.comp", {{"A_TYPE", "float"}, {"D_TYPE", "float"}}); - })); - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("rope_neox_f16", "rope_neox.comp", {{"A_TYPE", "float16_t"}, {"D_TYPE", "float16_t"}}); - })); - - tasks.push_back(std::async(std::launch::async, [] { - string_to_spv("argsort_f32", "argsort.comp", {{"A_TYPE", "float"}}); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("sum_rows_f32", "sum_rows.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("im2col_f32", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("im2col_f32_f16", "im2col.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float16_t"}})); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("timestep_embedding_f32", "timestep_embedding.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); - - tasks.push_back(std::async(std::launch::async, [=] { - string_to_spv("pool2d_f32", "pool2d.comp", merge_maps(base_dict, {{"A_TYPE", "float"}, {"D_TYPE", "float"}})); - })); -} - -void write_output_files() { - FILE* hdr = fopen(target_hpp.c_str(), "w"); - FILE* src = fopen(target_cpp.c_str(), "w"); - - fprintf(hdr, "#include \n\n"); - fprintf(src, "#include \"%s\"\n\n", basename(target_hpp).c_str()); - - for (const auto& pair : shader_fnames) { - const std::string& name = pair.first; - #ifdef _WIN32 - std::string path = pair.second; - std::replace(path.begin(), path.end(), '/', '\\' ); - #else - const std::string& path = pair.second; - #endif - - FILE* spv = fopen(path.c_str(), "rb"); - if (!spv) { - std::cerr << "Error opening SPIR-V file: " << path << " (" << strerror(errno) << ")\n"; - continue; - } - - fseek(spv, 0, SEEK_END); - size_t size = ftell(spv); - fseek(spv, 0, SEEK_SET); - - std::vector data(size); - size_t read_size = fread(data.data(), 1, size, spv); - fclose(spv); - if (read_size != size) { - std::cerr << "Error reading SPIR-V file: " << path << " (" << strerror(errno) << ")\n"; - continue; - } - - fprintf(hdr, "extern unsigned char %s_data[%zu];\n", name.c_str(), size); - fprintf(hdr, "const uint64_t %s_len = %zu;\n\n", name.c_str(), size); - - fprintf(src, "unsigned char %s_data[%zu] = {\n", name.c_str(), size); - for (size_t i = 0; i < size; ++i) { - fprintf(src, "0x%02x,", data[i]); - if ((i + 1) % 12 == 0) fprintf(src, "\n"); - } - fprintf(src, "\n};\n\n"); - - if (!no_clean) { - std::remove(path.c_str()); - } - } - - fclose(hdr); - fclose(src); -} - -int main(int argc, char** argv) { - std::map args; - for (int i = 1; i < argc; i += 2) { - if (i + 1 < argc) { - args[argv[i]] = argv[i + 1]; - } - } - - if (args.find("--glslc") != args.end()) { - GLSLC = args["--glslc"]; // Path to glslc - } - if (args.find("--input-dir") != args.end()) { - input_dir = args["--input-dir"]; // Directory containing shader sources - } - if (args.find("--output-dir") != args.end()) { - output_dir = args["--output-dir"]; // Directory for containing SPIR-V output - } - if (args.find("--target-hpp") != args.end()) { - target_hpp = args["--target-hpp"]; // Path to generated header file - } - if (args.find("--target-cpp") != args.end()) { - target_cpp = args["--target-cpp"]; // Path to generated cpp file - } - if (args.find("--no-clean") != args.end()) { - no_clean = true; // Keep temporary SPIR-V files in output-dir after build - } - - if (!directory_exists(input_dir)) { - std::cerr << "\"" << input_dir << "\" must be a valid directory containing shader sources" << std::endl; - return EXIT_FAILURE; - } - - if (!directory_exists(output_dir)) { - if (!create_directory(output_dir)) { - std::cerr << "Error creating output directory: " << output_dir << "\n"; - return EXIT_FAILURE; - } - } - - std::vector> tasks; - process_shaders(tasks); - - for (auto& task : tasks) { - task.get(); - } - - write_output_files(); - - return EXIT_SUCCESS; -} diff --git a/gguf-py/README.md b/gguf-py/README.md index 24af96a17..2e513633d 100644 --- a/gguf-py/README.md +++ b/gguf-py/README.md @@ -15,13 +15,15 @@ pip install gguf [examples/writer.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/examples/writer.py) — Generates `example.gguf` in the current directory to demonstrate generating a GGUF file. Note that this file cannot be used as a model. -[scripts/gguf_dump.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/scripts/gguf_dump.py) — Dumps a GGUF file's metadata to the console. +[examples/reader.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/examples/reader.py) — Extracts and displays key-value pairs and tensor details from a GGUF file in a readable format. -[scripts/gguf_set_metadata.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/scripts/gguf_set_metadata.py) — Allows changing simple metadata values in a GGUF file by key. +[gguf/scripts/gguf_dump.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/gguf/scripts/gguf_dump.py) — Dumps a GGUF file's metadata to the console. -[scripts/gguf_convert_endian.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/scripts/gguf_convert_endian.py) — Allows converting the endianness of GGUF files. +[gguf/scripts/gguf_set_metadata.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/gguf/scripts/gguf_set_metadata.py) — Allows changing simple metadata values in a GGUF file by key. -[scripts/gguf_new_metadata.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/scripts/gguf_new_metadata.py) — Copies a GGUF file with added/modified/removed metadata values. +[gguf/scripts/gguf_convert_endian.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/gguf/scripts/gguf_convert_endian.py) — Allows converting the endianness of GGUF files. + +[gguf/scripts/gguf_new_metadata.py](https://github.com/ggerganov/llama.cpp/blob/master/gguf-py/gguf/scripts/gguf_new_metadata.py) — Copies a GGUF file with added/modified/removed metadata values. ## Development Maintainers who participate in development of this package are advised to install it in editable mode: diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py index 7ab08b036..8fe84df21 100644 --- a/gguf-py/gguf/constants.py +++ b/gguf-py/gguf/constants.py @@ -64,20 +64,33 @@ class Keys: BASE_MODEL_AUTHOR = "general.base_model.{id}.author" BASE_MODEL_VERSION = "general.base_model.{id}.version" BASE_MODEL_ORGANIZATION = "general.base_model.{id}.organization" + BASE_MODEL_DESCRIPTION = "general.base_model.{id}.description" BASE_MODEL_URL = "general.base_model.{id}.url" # Model Website/Paper BASE_MODEL_DOI = "general.base_model.{id}.doi" BASE_MODEL_UUID = "general.base_model.{id}.uuid" BASE_MODEL_REPO_URL = "general.base_model.{id}.repo_url" # Model Source Repository (git/svn/etc...) + # Dataset Source + DATASET_COUNT = "general.dataset.count" + DATASET_NAME = "general.dataset.{id}.name" + DATASET_AUTHOR = "general.dataset.{id}.author" + DATASET_VERSION = "general.dataset.{id}.version" + DATASET_ORGANIZATION = "general.dataset.{id}.organization" + DATASET_DESCRIPTION = "general.dataset.{id}.description" + DATASET_URL = "general.dataset.{id}.url" # Model Website/Paper + DATASET_DOI = "general.dataset.{id}.doi" + DATASET_UUID = "general.dataset.{id}.uuid" + DATASET_REPO_URL = "general.dataset.{id}.repo_url" # Model Source Repository (git/svn/etc...) + # Array based KV stores TAGS = "general.tags" LANGUAGES = "general.languages" - DATASETS = "general.datasets" class LLM: VOCAB_SIZE = "{arch}.vocab_size" CONTEXT_LENGTH = "{arch}.context_length" EMBEDDING_LENGTH = "{arch}.embedding_length" + FEATURES_LENGTH = "{arch}.features_length" BLOCK_COUNT = "{arch}.block_count" LEADING_DENSE_BLOCK_COUNT = "{arch}.leading_dense_block_count" FEED_FORWARD_LENGTH = "{arch}.feed_forward_length" @@ -89,6 +102,8 @@ class Keys: EXPERT_USED_COUNT = "{arch}.expert_used_count" EXPERT_SHARED_COUNT = "{arch}.expert_shared_count" EXPERT_WEIGHTS_SCALE = "{arch}.expert_weights_scale" + EXPERT_WEIGHTS_NORM = "{arch}.expert_weights_norm" + EXPERT_GATING_FUNC = "{arch}.expert_gating_func" POOLING_TYPE = "{arch}.pooling_type" LOGIT_SCALE = "{arch}.logit_scale" DECODER_START_TOKEN_ID = "{arch}.decoder_start_token_id" @@ -100,6 +115,7 @@ class Keys: TIME_DECAY_EXTRA_DIM = "{arch}.time_decay_extra_dim" RESIDUAL_SCALE = "{arch}.residual_scale" EMBEDDING_SCALE = "{arch}.embedding_scale" + TOKEN_SHIFT_COUNT = "{arch}.token_shift_count" class Attention: HEAD_COUNT = "{arch}.attention.head_count" @@ -110,6 +126,8 @@ class Keys: VALUE_LENGTH = "{arch}.attention.value_length" LAYERNORM_EPS = "{arch}.attention.layer_norm_epsilon" LAYERNORM_RMS_EPS = "{arch}.attention.layer_norm_rms_epsilon" + GROUPNORM_EPS = "{arch}.attention.group_norm_epsilon" + GROUPNORM_GROUPS = "{arch}.attention.group_norm_groups" CAUSAL = "{arch}.attention.causal" Q_LORA_RANK = "{arch}.attention.q_lora_rank" KV_LORA_RANK = "{arch}.attention.kv_lora_rank" @@ -119,6 +137,7 @@ class Keys: class Rope: DIMENSION_COUNT = "{arch}.rope.dimension_count" + DIMENSION_SECTIONS = "{arch}.rope.dimension_sections" FREQ_BASE = "{arch}.rope.freq_base" SCALING_TYPE = "{arch}.rope.scaling.type" SCALING_FACTOR = "{arch}.rope.scaling.factor" @@ -142,6 +161,14 @@ class Keys: class WKV: HEAD_SIZE = "{arch}.wkv.head_size" + class PosNet: + EMBEDDING_LENGTH = "{arch}.posnet.embedding_length" + BLOCK_COUNT = "{arch}.posnet.block_count" + + class ConvNext: + EMBEDDING_LENGTH = "{arch}.convnext.embedding_length" + BLOCK_COUNT = "{arch}.convnext.block_count" + class Tokenizer: MODEL = "tokenizer.ggml.model" PRE = "tokenizer.ggml.pre" @@ -157,7 +184,6 @@ class Keys: UNK_ID = "tokenizer.ggml.unknown_token_id" SEP_ID = "tokenizer.ggml.seperator_token_id" PAD_ID = "tokenizer.ggml.padding_token_id" - CLS_ID = "tokenizer.ggml.cls_token_id" MASK_ID = "tokenizer.ggml.mask_token_id" ADD_BOS = "tokenizer.ggml.add_bos_token" ADD_EOS = "tokenizer.ggml.add_eos_token" @@ -196,55 +222,63 @@ class GGUFType: class MODEL_ARCH(IntEnum): - LLAMA = auto() - FALCON = auto() - BAICHUAN = auto() - GROK = auto() - GPT2 = auto() - GPTJ = auto() - GPTNEOX = auto() - MPT = auto() - STARCODER = auto() - REFACT = auto() - BERT = auto() - NOMIC_BERT = auto() - JINA_BERT_V2 = auto() - BLOOM = auto() - STABLELM = auto() - QWEN = auto() - QWEN2 = auto() - QWEN2MOE = auto() - PHI2 = auto() - PHI3 = auto() - PLAMO = auto() - CODESHELL = auto() - ORION = auto() - INTERNLM2 = auto() - MINICPM = auto() - MINICPM3 = auto() - GEMMA = auto() - GEMMA2 = auto() - STARCODER2 = auto() - RWKV6 = auto() - MAMBA = auto() - XVERSE = auto() - COMMAND_R = auto() - DBRX = auto() - OLMO = auto() - OLMOE = auto() - OPENELM = auto() - ARCTIC = auto() - DEEPSEEK2 = auto() - CHATGLM = auto() - BITNET = auto() - T5 = auto() - T5ENCODER = auto() - JAIS = auto() - NEMOTRON = auto() - EXAONE = auto() - GRANITE = auto() - GRANITE_MOE = auto() - CHAMELEON = auto() + LLAMA = auto() + DECI = auto() + FALCON = auto() + BAICHUAN = auto() + GROK = auto() + GPT2 = auto() + GPTJ = auto() + GPTNEOX = auto() + MPT = auto() + STARCODER = auto() + REFACT = auto() + BERT = auto() + NOMIC_BERT = auto() + JINA_BERT_V2 = auto() + BLOOM = auto() + STABLELM = auto() + QWEN = auto() + QWEN2 = auto() + QWEN2MOE = auto() + QWEN2VL = auto() + PHI2 = auto() + PHI3 = auto() + PHIMOE = auto() + PLAMO = auto() + CODESHELL = auto() + ORION = auto() + INTERNLM2 = auto() + MINICPM = auto() + MINICPM3 = auto() + GEMMA = auto() + GEMMA2 = auto() + STARCODER2 = auto() + RWKV6 = auto() + RWKV6QWEN2 = auto() + MAMBA = auto() + XVERSE = auto() + COMMAND_R = auto() + COHERE2 = auto() + DBRX = auto() + OLMO = auto() + OLMO2 = auto() + OLMOE = auto() + OPENELM = auto() + ARCTIC = auto() + DEEPSEEK = auto() + DEEPSEEK2 = auto() + CHATGLM = auto() + BITNET = auto() + T5 = auto() + T5ENCODER = auto() + JAIS = auto() + NEMOTRON = auto() + EXAONE = auto() + GRANITE = auto() + GRANITE_MOE = auto() + CHAMELEON = auto() + WAVTOKENIZER_DEC = auto() class MODEL_TENSOR(IntEnum): @@ -283,6 +317,7 @@ class MODEL_TENSOR(IntEnum): FFN_GATE_SHEXP = auto() FFN_DOWN_SHEXP = auto() FFN_UP_SHEXP = auto() + FFN_EXP_PROBS_B = auto() ATTN_Q_NORM = auto() ATTN_K_NORM = auto() LAYER_OUT_NORM = auto() @@ -300,6 +335,7 @@ class MODEL_TENSOR(IntEnum): TIME_MIX_LERP_V = auto() TIME_MIX_LERP_R = auto() TIME_MIX_LERP_G = auto() + TIME_MIX_LERP_FUSED = auto() TIME_MIX_LERP_W = auto() TIME_MIX_FIRST = auto() TIME_MIX_DECAY = auto() @@ -354,58 +390,82 @@ class MODEL_TENSOR(IntEnum): ENC_OUTPUT_NORM = auto() CLS = auto() # classifier CLS_OUT = auto() # classifier output projection + CONV1D = auto() + CONVNEXT_DW = auto() + CONVNEXT_NORM = auto() + CONVNEXT_PW1 = auto() + CONVNEXT_PW2 = auto() + CONVNEXT_GAMMA = auto() + POSNET_CONV1 = auto() + POSNET_CONV2 = auto() + POSNET_NORM = auto() + POSNET_NORM1 = auto() + POSNET_NORM2 = auto() + POSNET_ATTN_NORM = auto() + POSNET_ATTN_Q = auto() + POSNET_ATTN_K = auto() + POSNET_ATTN_V = auto() + POSNET_ATTN_OUT = auto() MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = { - MODEL_ARCH.LLAMA: "llama", - MODEL_ARCH.FALCON: "falcon", - MODEL_ARCH.BAICHUAN: "baichuan", - MODEL_ARCH.GROK: "grok", - MODEL_ARCH.GPT2: "gpt2", - MODEL_ARCH.GPTJ: "gptj", - MODEL_ARCH.GPTNEOX: "gptneox", - MODEL_ARCH.MPT: "mpt", - MODEL_ARCH.STARCODER: "starcoder", - MODEL_ARCH.REFACT: "refact", - MODEL_ARCH.BERT: "bert", - MODEL_ARCH.NOMIC_BERT: "nomic-bert", - MODEL_ARCH.JINA_BERT_V2: "jina-bert-v2", - MODEL_ARCH.BLOOM: "bloom", - MODEL_ARCH.STABLELM: "stablelm", - MODEL_ARCH.QWEN: "qwen", - MODEL_ARCH.QWEN2: "qwen2", - MODEL_ARCH.QWEN2MOE: "qwen2moe", - MODEL_ARCH.PHI2: "phi2", - MODEL_ARCH.PHI3: "phi3", - MODEL_ARCH.PLAMO: "plamo", - MODEL_ARCH.CODESHELL: "codeshell", - MODEL_ARCH.ORION: "orion", - MODEL_ARCH.INTERNLM2: "internlm2", - MODEL_ARCH.MINICPM: "minicpm", - MODEL_ARCH.MINICPM3: "minicpm3", - MODEL_ARCH.GEMMA: "gemma", - MODEL_ARCH.GEMMA2: "gemma2", - MODEL_ARCH.STARCODER2: "starcoder2", - MODEL_ARCH.RWKV6: "rwkv6", - MODEL_ARCH.MAMBA: "mamba", - MODEL_ARCH.XVERSE: "xverse", - MODEL_ARCH.COMMAND_R: "command-r", - MODEL_ARCH.DBRX: "dbrx", - MODEL_ARCH.OLMO: "olmo", - MODEL_ARCH.OLMOE: "olmoe", - MODEL_ARCH.OPENELM: "openelm", - MODEL_ARCH.ARCTIC: "arctic", - MODEL_ARCH.DEEPSEEK2: "deepseek2", - MODEL_ARCH.CHATGLM: "chatglm", - MODEL_ARCH.BITNET: "bitnet", - MODEL_ARCH.T5: "t5", - MODEL_ARCH.T5ENCODER: "t5encoder", - MODEL_ARCH.JAIS: "jais", - MODEL_ARCH.NEMOTRON: "nemotron", - MODEL_ARCH.EXAONE: "exaone", - MODEL_ARCH.GRANITE: "granite", - MODEL_ARCH.GRANITE_MOE: "granitemoe", - MODEL_ARCH.CHAMELEON: "chameleon", + MODEL_ARCH.LLAMA: "llama", + MODEL_ARCH.DECI: "deci", + MODEL_ARCH.FALCON: "falcon", + MODEL_ARCH.BAICHUAN: "baichuan", + MODEL_ARCH.GROK: "grok", + MODEL_ARCH.GPT2: "gpt2", + MODEL_ARCH.GPTJ: "gptj", + MODEL_ARCH.GPTNEOX: "gptneox", + MODEL_ARCH.MPT: "mpt", + MODEL_ARCH.STARCODER: "starcoder", + MODEL_ARCH.REFACT: "refact", + MODEL_ARCH.BERT: "bert", + MODEL_ARCH.NOMIC_BERT: "nomic-bert", + MODEL_ARCH.JINA_BERT_V2: "jina-bert-v2", + MODEL_ARCH.BLOOM: "bloom", + MODEL_ARCH.STABLELM: "stablelm", + MODEL_ARCH.QWEN: "qwen", + MODEL_ARCH.QWEN2: "qwen2", + MODEL_ARCH.QWEN2MOE: "qwen2moe", + MODEL_ARCH.QWEN2VL: "qwen2vl", + MODEL_ARCH.PHI2: "phi2", + MODEL_ARCH.PHI3: "phi3", + MODEL_ARCH.PHIMOE: "phimoe", + MODEL_ARCH.PLAMO: "plamo", + MODEL_ARCH.CODESHELL: "codeshell", + MODEL_ARCH.ORION: "orion", + MODEL_ARCH.INTERNLM2: "internlm2", + MODEL_ARCH.MINICPM: "minicpm", + MODEL_ARCH.MINICPM3: "minicpm3", + MODEL_ARCH.GEMMA: "gemma", + MODEL_ARCH.GEMMA2: "gemma2", + MODEL_ARCH.STARCODER2: "starcoder2", + MODEL_ARCH.RWKV6: "rwkv6", + MODEL_ARCH.RWKV6QWEN2: "rwkv6qwen2", + MODEL_ARCH.MAMBA: "mamba", + MODEL_ARCH.XVERSE: "xverse", + MODEL_ARCH.COMMAND_R: "command-r", + MODEL_ARCH.COHERE2: "cohere2", + MODEL_ARCH.DBRX: "dbrx", + MODEL_ARCH.OLMO: "olmo", + MODEL_ARCH.OLMO2: "olmo2", + MODEL_ARCH.OLMOE: "olmoe", + MODEL_ARCH.OPENELM: "openelm", + MODEL_ARCH.ARCTIC: "arctic", + MODEL_ARCH.DEEPSEEK: "deepseek", + MODEL_ARCH.DEEPSEEK2: "deepseek2", + MODEL_ARCH.CHATGLM: "chatglm", + MODEL_ARCH.BITNET: "bitnet", + MODEL_ARCH.T5: "t5", + MODEL_ARCH.T5ENCODER: "t5encoder", + MODEL_ARCH.JAIS: "jais", + MODEL_ARCH.NEMOTRON: "nemotron", + MODEL_ARCH.EXAONE: "exaone", + MODEL_ARCH.GRANITE: "granite", + MODEL_ARCH.GRANITE_MOE: "granitemoe", + MODEL_ARCH.CHAMELEON: "chameleon", + MODEL_ARCH.WAVTOKENIZER_DEC: "wavtokenizer-dec", } TENSOR_NAMES: dict[MODEL_TENSOR, str] = { @@ -446,6 +506,7 @@ TENSOR_NAMES: dict[MODEL_TENSOR, str] = { MODEL_TENSOR.FFN_GATE_EXP: "blk.{bid}.ffn_gate_exps", MODEL_TENSOR.FFN_DOWN_EXP: "blk.{bid}.ffn_down_exps", MODEL_TENSOR.FFN_UP_EXP: "blk.{bid}.ffn_up_exps", + MODEL_TENSOR.FFN_EXP_PROBS_B: "blk.{bid}.exp_probs_b", MODEL_TENSOR.LAYER_OUT_NORM: "blk.{bid}.layer_output_norm", MODEL_TENSOR.SSM_IN: "blk.{bid}.ssm_in", MODEL_TENSOR.SSM_CONV1D: "blk.{bid}.ssm_conv1d", @@ -461,6 +522,7 @@ TENSOR_NAMES: dict[MODEL_TENSOR, str] = { MODEL_TENSOR.TIME_MIX_LERP_V: "blk.{bid}.time_mix_lerp_v", MODEL_TENSOR.TIME_MIX_LERP_R: "blk.{bid}.time_mix_lerp_r", MODEL_TENSOR.TIME_MIX_LERP_G: "blk.{bid}.time_mix_lerp_g", + MODEL_TENSOR.TIME_MIX_LERP_FUSED: "blk.{bid}.time_mix_lerp_fused", MODEL_TENSOR.TIME_MIX_LERP_W: "blk.{bid}.time_mix_lerp_w", MODEL_TENSOR.TIME_MIX_FIRST: "blk.{bid}.time_mix_first", MODEL_TENSOR.TIME_MIX_DECAY: "blk.{bid}.time_mix_decay", @@ -515,6 +577,22 @@ TENSOR_NAMES: dict[MODEL_TENSOR, str] = { MODEL_TENSOR.ENC_OUTPUT_NORM: "enc.output_norm", MODEL_TENSOR.CLS: "cls", MODEL_TENSOR.CLS_OUT: "cls.output", + MODEL_TENSOR.CONV1D: "conv1d", + MODEL_TENSOR.CONVNEXT_DW: "convnext.{bid}.dw", + MODEL_TENSOR.CONVNEXT_NORM: "convnext.{bid}.norm", + MODEL_TENSOR.CONVNEXT_PW1: "convnext.{bid}.pw1", + MODEL_TENSOR.CONVNEXT_PW2: "convnext.{bid}.pw2", + MODEL_TENSOR.CONVNEXT_GAMMA: "convnext.{bid}.gamma", + MODEL_TENSOR.POSNET_CONV1: "posnet.{bid}.conv1", + MODEL_TENSOR.POSNET_CONV2: "posnet.{bid}.conv2", + MODEL_TENSOR.POSNET_NORM: "posnet.{bid}.norm", + MODEL_TENSOR.POSNET_NORM1: "posnet.{bid}.norm1", + MODEL_TENSOR.POSNET_NORM2: "posnet.{bid}.norm2", + MODEL_TENSOR.POSNET_ATTN_NORM: "posnet.{bid}.attn_norm", + MODEL_TENSOR.POSNET_ATTN_Q: "posnet.{bid}.attn_q", + MODEL_TENSOR.POSNET_ATTN_K: "posnet.{bid}.attn_k", + MODEL_TENSOR.POSNET_ATTN_V: "posnet.{bid}.attn_v", + MODEL_TENSOR.POSNET_ATTN_OUT: "posnet.{bid}.attn_output", } MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { @@ -538,6 +616,26 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN_EXP, MODEL_TENSOR.FFN_UP_EXP, ], + MODEL_ARCH.DECI: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_ROT_EMBD, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + ], MODEL_ARCH.GROK: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -744,6 +842,21 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_UP, ], MODEL_ARCH.QWEN2: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + ], + MODEL_ARCH.QWEN2VL: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, MODEL_TENSOR.OUTPUT, @@ -833,6 +946,24 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN, MODEL_TENSOR.FFN_UP, ], + MODEL_ARCH.PHIMOE: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ROPE_FACTORS_LONG, + MODEL_TENSOR.ROPE_FACTORS_SHORT, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_QKV, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + ], MODEL_ARCH.CODESHELL: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.POS_EMBD, @@ -882,6 +1013,8 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.OUTPUT, MODEL_TENSOR.OUTPUT_NORM, MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ROPE_FACTORS_LONG, + MODEL_TENSOR.ROPE_FACTORS_SHORT, MODEL_TENSOR.ATTN_NORM, MODEL_TENSOR.ATTN_Q, MODEL_TENSOR.ATTN_K, @@ -974,6 +1107,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.TIME_MIX_LERP_R, MODEL_TENSOR.TIME_MIX_LERP_G, MODEL_TENSOR.TIME_MIX_LERP_W, + MODEL_TENSOR.TIME_MIX_LERP_FUSED, MODEL_TENSOR.TIME_MIX_FIRST, MODEL_TENSOR.TIME_MIX_DECAY, MODEL_TENSOR.TIME_MIX_DECAY_W1, @@ -990,6 +1124,35 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.CHANNEL_MIX_RECEPTANCE, MODEL_TENSOR.CHANNEL_MIX_VALUE, ], + MODEL_ARCH.RWKV6QWEN2: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.TIME_MIX_W1, + MODEL_TENSOR.TIME_MIX_W2, + MODEL_TENSOR.TIME_MIX_LERP_X, + MODEL_TENSOR.TIME_MIX_LERP_K, + MODEL_TENSOR.TIME_MIX_LERP_V, + MODEL_TENSOR.TIME_MIX_LERP_R, + MODEL_TENSOR.TIME_MIX_LERP_G, + MODEL_TENSOR.TIME_MIX_LERP_W, + MODEL_TENSOR.TIME_MIX_LERP_FUSED, + MODEL_TENSOR.TIME_MIX_FIRST, + MODEL_TENSOR.TIME_MIX_DECAY, + MODEL_TENSOR.TIME_MIX_DECAY_W1, + MODEL_TENSOR.TIME_MIX_DECAY_W2, + MODEL_TENSOR.TIME_MIX_KEY, + MODEL_TENSOR.TIME_MIX_VALUE, + MODEL_TENSOR.TIME_MIX_RECEPTANCE, + MODEL_TENSOR.TIME_MIX_GATE, + MODEL_TENSOR.TIME_MIX_LN, + MODEL_TENSOR.TIME_MIX_OUTPUT, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + ], MODEL_ARCH.MAMBA: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -1033,6 +1196,18 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.ATTN_K_NORM, MODEL_TENSOR.ATTN_Q_NORM, ], + MODEL_ARCH.COHERE2: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + ], MODEL_ARCH.DBRX: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -1057,6 +1232,22 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN, MODEL_TENSOR.FFN_UP, ], + MODEL_ARCH.OLMO2: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_POST_NORM, + MODEL_TENSOR.ATTN_Q_NORM, + MODEL_TENSOR.ATTN_K_NORM, + MODEL_TENSOR.FFN_POST_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + ], MODEL_ARCH.OLMOE: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -1108,6 +1299,29 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN_EXP, MODEL_TENSOR.FFN_UP_EXP, ], + MODEL_ARCH.DEEPSEEK: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_ROT_EMBD, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + MODEL_TENSOR.FFN_GATE_SHEXP, + MODEL_TENSOR.FFN_DOWN_SHEXP, + MODEL_TENSOR.FFN_UP_SHEXP, + ], MODEL_ARCH.DEEPSEEK2: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -1134,6 +1348,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_GATE_SHEXP, MODEL_TENSOR.FFN_DOWN_SHEXP, MODEL_TENSOR.FFN_UP_SHEXP, + MODEL_TENSOR.FFN_EXP_PROBS_B, ], MODEL_ARCH.CHATGLM : [ MODEL_TENSOR.TOKEN_EMBD, @@ -1297,6 +1512,28 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN, MODEL_TENSOR.FFN_UP, ], + MODEL_ARCH.WAVTOKENIZER_DEC: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.TOKEN_EMBD_NORM, + MODEL_TENSOR.CONV1D, + MODEL_TENSOR.CONVNEXT_DW, + MODEL_TENSOR.CONVNEXT_NORM, + MODEL_TENSOR.CONVNEXT_PW1, + MODEL_TENSOR.CONVNEXT_PW2, + MODEL_TENSOR.CONVNEXT_GAMMA, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.POSNET_CONV1, + MODEL_TENSOR.POSNET_CONV2, + MODEL_TENSOR.POSNET_NORM, + MODEL_TENSOR.POSNET_NORM1, + MODEL_TENSOR.POSNET_NORM2, + MODEL_TENSOR.POSNET_ATTN_NORM, + MODEL_TENSOR.POSNET_ATTN_Q, + MODEL_TENSOR.POSNET_ATTN_K, + MODEL_TENSOR.POSNET_ATTN_V, + MODEL_TENSOR.POSNET_ATTN_OUT, + ], # TODO } @@ -1306,6 +1543,10 @@ MODEL_TENSOR_SKIP: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, ], + MODEL_ARCH.DECI: [ + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_ROT_EMBD, + ], MODEL_ARCH.BAICHUAN: [ MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, @@ -1330,6 +1571,10 @@ MODEL_TENSOR_SKIP: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, ], + MODEL_ARCH.DEEPSEEK: [ + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_ROT_EMBD, + ], MODEL_ARCH.DEEPSEEK2: [ MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, @@ -1358,9 +1603,10 @@ class TokenType(IntEnum): class RopeScalingType(Enum): - NONE = 'none' - LINEAR = 'linear' - YARN = 'yarn' + NONE = 'none' + LINEAR = 'linear' + YARN = 'yarn' + LONGROPE = 'longrope' class PoolingType(IntEnum): @@ -1399,13 +1645,15 @@ class GGMLQuantizationType(IntEnum): F64 = 28 IQ1_M = 29 BF16 = 30 - Q4_0_4_4 = 31 - Q4_0_4_8 = 32 - Q4_0_8_8 = 33 TQ1_0 = 34 TQ2_0 = 35 +class ExpertGatingFuncType(IntEnum): + SOFTMAX = 1 + SIGMOID = 2 + + # TODO: add GGMLFileType from ggml_ftype in ggml.h @@ -1445,9 +1693,9 @@ class LlamaFileType(IntEnum): MOSTLY_IQ4_XS = 30 # except 1d tensors MOSTLY_IQ1_M = 31 # except 1d tensors MOSTLY_BF16 = 32 # except 1d tensors - MOSTLY_Q4_0_4_4 = 33 # except 1d tensors - MOSTLY_Q4_0_4_8 = 34 # except 1d tensors - MOSTLY_Q4_0_8_8 = 35 # except 1d tensors + # MOSTLY_Q4_0_4_4 = 33 # removed from gguf files, use Q4_0 and runtime repack + # MOSTLY_Q4_0_4_8 = 34 # removed from gguf files, use Q4_0 and runtime repack + # MOSTLY_Q4_0_8_8 = 35 # removed from gguf files, use Q4_0 and runtime repack MOSTLY_TQ1_0 = 36 # except 1d tensors MOSTLY_TQ2_0 = 37 # except 1d tensors @@ -1523,9 +1771,6 @@ GGML_QUANT_SIZES: dict[GGMLQuantizationType, tuple[int, int]] = { GGMLQuantizationType.F64: (1, 8), GGMLQuantizationType.IQ1_M: (256, QK_K // 8 + QK_K // 16 + QK_K // 32), GGMLQuantizationType.BF16: (1, 2), - GGMLQuantizationType.Q4_0_4_4:(32, 2 + 16), - GGMLQuantizationType.Q4_0_4_8:(32, 2 + 16), - GGMLQuantizationType.Q4_0_8_8:(32, 2 + 16), GGMLQuantizationType.TQ1_0: (256, 2 + 4 * 13), GGMLQuantizationType.TQ2_0: (256, 2 + 64), } @@ -1591,7 +1836,6 @@ KEY_TOKENIZER_EOM_ID = Keys.Tokenizer.EOM_ID KEY_TOKENIZER_UNK_ID = Keys.Tokenizer.UNK_ID KEY_TOKENIZER_SEP_ID = Keys.Tokenizer.SEP_ID KEY_TOKENIZER_PAD_ID = Keys.Tokenizer.PAD_ID -KEY_TOKENIZER_CLS_ID = Keys.Tokenizer.CLS_ID KEY_TOKENIZER_MASK_ID = Keys.Tokenizer.MASK_ID KEY_TOKENIZER_HF_JSON = Keys.Tokenizer.HF_JSON KEY_TOKENIZER_RWKV = Keys.Tokenizer.RWKV diff --git a/gguf-py/gguf/gguf_reader.py b/gguf-py/gguf/gguf_reader.py index e8e61abf8..e17a4e831 100644 --- a/gguf-py/gguf/gguf_reader.py +++ b/gguf-py/gguf/gguf_reader.py @@ -145,11 +145,10 @@ class GGUFReader: count = int(count) itemsize = int(np.empty([], dtype = dtype).itemsize) end_offs = offset + itemsize * count - return ( - self.data[offset:end_offs] - .view(dtype = dtype)[:count] - .newbyteorder(override_order or self.byte_order) - ) + arr = self.data[offset:end_offs].view(dtype=dtype)[:count] + if override_order is None: + return arr + return arr.view(arr.dtype.newbyteorder(override_order)) def _push_field(self, field: ReaderField, skip_sum: bool = False) -> int: if field.name in self.fields: diff --git a/gguf-py/gguf/gguf_writer.py b/gguf-py/gguf/gguf_writer.py index 0d8d8a0b0..080d2b9dc 100644 --- a/gguf-py/gguf/gguf_writer.py +++ b/gguf-py/gguf/gguf_writer.py @@ -26,6 +26,7 @@ from .constants import ( RopeScalingType, PoolingType, TokenType, + ExpertGatingFuncType, ) from .quants import quant_shape_from_byte_shape @@ -568,6 +569,9 @@ class GGUFWriter: def add_base_model_organization(self, source_id: int, organization: str) -> None: self.add_string(Keys.General.BASE_MODEL_ORGANIZATION.format(id=source_id), organization) + def add_base_model_description(self, source_id: int, description: str) -> None: + self.add_string(Keys.General.BASE_MODEL_DESCRIPTION.format(id=source_id), description) + def add_base_model_url(self, source_id: int, url: str) -> None: self.add_string(Keys.General.BASE_MODEL_URL.format(id=source_id), url) @@ -580,15 +584,42 @@ class GGUFWriter: def add_base_model_repo_url(self, source_id: int, repo_url: str) -> None: self.add_string(Keys.General.BASE_MODEL_REPO_URL.format(id=source_id), repo_url) + def add_dataset_count(self, source_count: int) -> None: + self.add_uint32(Keys.General.DATASET_COUNT, source_count) + + def add_dataset_name(self, source_id: int, name: str) -> None: + self.add_string(Keys.General.DATASET_NAME.format(id=source_id), name) + + def add_dataset_author(self, source_id: int, author: str) -> None: + self.add_string(Keys.General.DATASET_AUTHOR.format(id=source_id), author) + + def add_dataset_version(self, source_id: int, version: str) -> None: + self.add_string(Keys.General.DATASET_VERSION.format(id=source_id), version) + + def add_dataset_organization(self, source_id: int, organization: str) -> None: + self.add_string(Keys.General.DATASET_ORGANIZATION.format(id=source_id), organization) + + def add_dataset_description(self, source_id: int, description: str) -> None: + self.add_string(Keys.General.DATASET_DESCRIPTION.format(id=source_id), description) + + def add_dataset_url(self, source_id: int, url: str) -> None: + self.add_string(Keys.General.DATASET_URL.format(id=source_id), url) + + def add_dataset_doi(self, source_id: int, doi: str) -> None: + self.add_string(Keys.General.DATASET_DOI.format(id=source_id), doi) + + def add_dataset_uuid(self, source_id: int, uuid: str) -> None: + self.add_string(Keys.General.DATASET_UUID.format(id=source_id), uuid) + + def add_dataset_repo_url(self, source_id: int, repo_url: str) -> None: + self.add_string(Keys.General.DATASET_REPO_URL.format(id=source_id), repo_url) + def add_tags(self, tags: Sequence[str]) -> None: self.add_array(Keys.General.TAGS, tags) def add_languages(self, languages: Sequence[str]) -> None: self.add_array(Keys.General.LANGUAGES, languages) - def add_datasets(self, datasets: Sequence[str]) -> None: - self.add_array(Keys.General.DATASETS, datasets) - def add_tensor_data_layout(self, layout: str) -> None: self.add_string(Keys.LLM.TENSOR_DATA_LAYOUT.format(arch=self.arch), layout) @@ -601,6 +632,21 @@ class GGUFWriter: def add_embedding_length(self, length: int) -> None: self.add_uint32(Keys.LLM.EMBEDDING_LENGTH.format(arch=self.arch), length) + def add_features_length(self, length: int) -> None: + self.add_uint32(Keys.LLM.FEATURES_LENGTH.format(arch=self.arch), length) + + def add_posnet_embedding_length(self, length: int) -> None: + self.add_uint32(Keys.PosNet.EMBEDDING_LENGTH.format(arch=self.arch), length) + + def add_posnet_block_count(self, length: int) -> None: + self.add_uint32(Keys.PosNet.BLOCK_COUNT.format(arch=self.arch), length) + + def add_convnext_embedding_length(self, length: int) -> None: + self.add_uint32(Keys.ConvNext.EMBEDDING_LENGTH.format(arch=self.arch), length) + + def add_convnext_block_count(self, length: int) -> None: + self.add_uint32(Keys.ConvNext.BLOCK_COUNT.format(arch=self.arch), length) + def add_block_count(self, length: int) -> None: self.add_uint32(Keys.LLM.BLOCK_COUNT.format(arch=self.arch), length) @@ -670,6 +716,12 @@ class GGUFWriter: def add_expert_weights_scale(self, value: float) -> None: self.add_float32(Keys.LLM.EXPERT_WEIGHTS_SCALE.format(arch=self.arch), value) + def add_expert_weights_norm(self, value: bool) -> None: + self.add_bool(Keys.LLM.EXPERT_WEIGHTS_NORM.format(arch=self.arch), value) + + def add_expert_gating_func(self, value: ExpertGatingFuncType) -> None: + self.add_uint32(Keys.LLM.EXPERT_GATING_FUNC.format(arch=self.arch), value.value) + def add_swin_norm(self, value: bool) -> None: self.add_bool(Keys.LLM.SWIN_NORM.format(arch=self.arch), value) @@ -691,12 +743,21 @@ class GGUFWriter: def add_wkv_head_size(self, size: int) -> None: self.add_uint32(Keys.WKV.HEAD_SIZE.format(arch=self.arch), size) + def add_token_shift_count(self, count: int) -> None: + self.add_uint32(Keys.LLM.TOKEN_SHIFT_COUNT.format(arch=self.arch), count) + def add_layer_norm_eps(self, value: float) -> None: self.add_float32(Keys.Attention.LAYERNORM_EPS.format(arch=self.arch), value) def add_layer_norm_rms_eps(self, value: float) -> None: self.add_float32(Keys.Attention.LAYERNORM_RMS_EPS.format(arch=self.arch), value) + def add_group_norm_eps(self, value: float) -> None: + self.add_float32(Keys.Attention.GROUPNORM_EPS.format(arch=self.arch), value) + + def add_group_norm_groups(self, value: int) -> None: + self.add_uint32(Keys.Attention.GROUPNORM_GROUPS.format(arch=self.arch), value) + def add_causal_attention(self, value: bool) -> None: self.add_bool(Keys.Attention.CAUSAL.format(arch=self.arch), value) @@ -721,6 +782,9 @@ class GGUFWriter: def add_rope_dimension_count(self, count: int) -> None: self.add_uint32(Keys.Rope.DIMENSION_COUNT.format(arch=self.arch), count) + def add_rope_dimension_sections(self, dims: Sequence[int]) -> None: + self.add_array(Keys.Rope.DIMENSION_SECTIONS.format(arch=self.arch), dims) + def add_rope_freq_base(self, value: float) -> None: self.add_float32(Keys.Rope.FREQ_BASE.format(arch=self.arch), value) @@ -793,9 +857,6 @@ class GGUFWriter: def add_pad_token_id(self, id: int) -> None: self.add_uint32(Keys.Tokenizer.PAD_ID, id) - def add_cls_token_id(self, id: int) -> None: - self.add_uint32(Keys.Tokenizer.CLS_ID, id) - def add_mask_token_id(self, id: int) -> None: self.add_uint32(Keys.Tokenizer.MASK_ID, id) diff --git a/gguf-py/gguf/metadata.py b/gguf-py/gguf/metadata.py index db318542a..962c27b20 100644 --- a/gguf-py/gguf/metadata.py +++ b/gguf-py/gguf/metadata.py @@ -41,7 +41,7 @@ class Metadata: base_models: Optional[list[dict]] = None tags: Optional[list[str]] = None languages: Optional[list[str]] = None - datasets: Optional[list[str]] = None + datasets: Optional[list[dict]] = None @staticmethod def load(metadata_override_path: Optional[Path] = None, model_path: Optional[Path] = None, model_name: Optional[str] = None, total_params: int = 0) -> Metadata: @@ -91,9 +91,11 @@ class Metadata: # Base Models is received here as an array of models metadata.base_models = metadata_override.get("general.base_models", metadata.base_models) + # Datasets is received here as an array of datasets + metadata.datasets = metadata_override.get("general.datasets", metadata.datasets) + metadata.tags = metadata_override.get(Keys.General.TAGS, metadata.tags) metadata.languages = metadata_override.get(Keys.General.LANGUAGES, metadata.languages) - metadata.datasets = metadata_override.get(Keys.General.DATASETS, metadata.datasets) # Direct Metadata Override (via direct cli argument) if model_name is not None: @@ -346,12 +348,12 @@ class Metadata: use_model_card_metadata("author", "model_creator") use_model_card_metadata("basename", "model_type") - if "base_model" in model_card: + if "base_model" in model_card or "base_models" in model_card or "base_model_sources" in model_card: # This represents the parent models that this is based on # Example: stabilityai/stable-diffusion-xl-base-1.0. Can also be a list (for merges) # Example of merges: https://huggingface.co/EmbeddedLLM/Mistral-7B-Merge-14-v0.1/blob/main/README.md metadata_base_models = [] - base_model_value = model_card.get("base_model", None) + base_model_value = model_card.get("base_model", model_card.get("base_models", model_card.get("base_model_sources", None))) if base_model_value is not None: if isinstance(base_model_value, str): @@ -364,18 +366,106 @@ class Metadata: for model_id in metadata_base_models: # NOTE: model size of base model is assumed to be similar to the size of the current model - model_full_name_component, org_component, basename, finetune, version, size_label = Metadata.get_model_id_components(model_id, total_params) base_model = {} - if model_full_name_component is not None: - base_model["name"] = Metadata.id_to_title(model_full_name_component) - if org_component is not None: - base_model["organization"] = Metadata.id_to_title(org_component) - if version is not None: - base_model["version"] = version - if org_component is not None and model_full_name_component is not None: - base_model["repo_url"] = f"https://huggingface.co/{org_component}/{model_full_name_component}" + if isinstance(model_id, str): + if model_id.startswith("http://") or model_id.startswith("https://") or model_id.startswith("ssh://"): + base_model["repo_url"] = model_id + + # Check if Hugging Face ID is present in URL + if "huggingface.co" in model_id: + match = re.match(r"https?://huggingface.co/([^/]+/[^/]+)$", model_id) + if match: + model_id_component = match.group(1) + model_full_name_component, org_component, basename, finetune, version, size_label = Metadata.get_model_id_components(model_id_component, total_params) + + # Populate model dictionary with extracted components + if model_full_name_component is not None: + base_model["name"] = Metadata.id_to_title(model_full_name_component) + if org_component is not None: + base_model["organization"] = Metadata.id_to_title(org_component) + if version is not None: + base_model["version"] = version + + else: + # Likely a Hugging Face ID + model_full_name_component, org_component, basename, finetune, version, size_label = Metadata.get_model_id_components(model_id, total_params) + + # Populate model dictionary with extracted components + if model_full_name_component is not None: + base_model["name"] = Metadata.id_to_title(model_full_name_component) + if org_component is not None: + base_model["organization"] = Metadata.id_to_title(org_component) + if version is not None: + base_model["version"] = version + if org_component is not None and model_full_name_component is not None: + base_model["repo_url"] = f"https://huggingface.co/{org_component}/{model_full_name_component}" + + elif isinstance(model_id, dict): + base_model = model_id + + else: + logger.error(f"base model entry '{str(model_id)}' not in a known format") + metadata.base_models.append(base_model) + if "datasets" in model_card or "dataset" in model_card or "dataset_sources" in model_card: + # This represents the datasets that this was trained from + metadata_datasets = [] + dataset_value = model_card.get("datasets", model_card.get("dataset", model_card.get("dataset_sources", None))) + + if dataset_value is not None: + if isinstance(dataset_value, str): + metadata_datasets.append(dataset_value) + elif isinstance(dataset_value, list): + metadata_datasets.extend(dataset_value) + + if metadata.datasets is None: + metadata.datasets = [] + + for dataset_id in metadata_datasets: + # NOTE: model size of base model is assumed to be similar to the size of the current model + dataset = {} + if isinstance(dataset_id, str): + if dataset_id.startswith(("http://", "https://", "ssh://")): + dataset["repo_url"] = dataset_id + + # Check if Hugging Face ID is present in URL + if "huggingface.co" in dataset_id: + match = re.match(r"https?://huggingface.co/([^/]+/[^/]+)$", dataset_id) + if match: + dataset_id_component = match.group(1) + dataset_name_component, org_component, basename, finetune, version, size_label = Metadata.get_model_id_components(dataset_id_component, total_params) + + # Populate dataset dictionary with extracted components + if dataset_name_component is not None: + dataset["name"] = Metadata.id_to_title(dataset_name_component) + if org_component is not None: + dataset["organization"] = Metadata.id_to_title(org_component) + if version is not None: + dataset["version"] = version + + else: + # Likely a Hugging Face ID + dataset_name_component, org_component, basename, finetune, version, size_label = Metadata.get_model_id_components(dataset_id, total_params) + + # Populate dataset dictionary with extracted components + if dataset_name_component is not None: + dataset["name"] = Metadata.id_to_title(dataset_name_component) + if org_component is not None: + dataset["organization"] = Metadata.id_to_title(org_component) + if version is not None: + dataset["version"] = version + if org_component is not None and dataset_name_component is not None: + dataset["repo_url"] = f"https://huggingface.co/{org_component}/{dataset_name_component}" + + elif isinstance(dataset_id, dict): + dataset = dataset_id + + else: + logger.error(f"dataset entry '{str(dataset_id)}' not in a known format") + + metadata.datasets.append(dataset) + use_model_card_metadata("license", "license") use_model_card_metadata("license_name", "license_name") use_model_card_metadata("license_link", "license_link") @@ -386,9 +476,6 @@ class Metadata: use_array_model_card_metadata("languages", "languages") use_array_model_card_metadata("languages", "language") - use_array_model_card_metadata("datasets", "datasets") - use_array_model_card_metadata("datasets", "dataset") - # Hugging Face Parameter Heuristics #################################### @@ -458,7 +545,10 @@ class Metadata: gguf_writer.add_size_label(self.size_label) if self.license is not None: - gguf_writer.add_license(self.license) + if isinstance(self.license, list): + gguf_writer.add_license(",".join(self.license)) + else: + gguf_writer.add_license(self.license) if self.license_name is not None: gguf_writer.add_license_name(self.license_name) if self.license_link is not None: @@ -493,6 +583,8 @@ class Metadata: gguf_writer.add_base_model_version(key, base_model_entry["version"]) if "organization" in base_model_entry: gguf_writer.add_base_model_organization(key, base_model_entry["organization"]) + if "description" in base_model_entry: + gguf_writer.add_base_model_description(key, base_model_entry["description"]) if "url" in base_model_entry: gguf_writer.add_base_model_url(key, base_model_entry["url"]) if "doi" in base_model_entry: @@ -502,9 +594,29 @@ class Metadata: if "repo_url" in base_model_entry: gguf_writer.add_base_model_repo_url(key, base_model_entry["repo_url"]) + if self.datasets is not None: + gguf_writer.add_dataset_count(len(self.datasets)) + for key, dataset_entry in enumerate(self.datasets): + if "name" in dataset_entry: + gguf_writer.add_dataset_name(key, dataset_entry["name"]) + if "author" in dataset_entry: + gguf_writer.add_dataset_author(key, dataset_entry["author"]) + if "version" in dataset_entry: + gguf_writer.add_dataset_version(key, dataset_entry["version"]) + if "organization" in dataset_entry: + gguf_writer.add_dataset_organization(key, dataset_entry["organization"]) + if "description" in dataset_entry: + gguf_writer.add_dataset_description(key, dataset_entry["description"]) + if "url" in dataset_entry: + gguf_writer.add_dataset_url(key, dataset_entry["url"]) + if "doi" in dataset_entry: + gguf_writer.add_dataset_doi(key, dataset_entry["doi"]) + if "uuid" in dataset_entry: + gguf_writer.add_dataset_uuid(key, dataset_entry["uuid"]) + if "repo_url" in dataset_entry: + gguf_writer.add_dataset_repo_url(key, dataset_entry["repo_url"]) + if self.tags is not None: gguf_writer.add_tags(self.tags) if self.languages is not None: gguf_writer.add_languages(self.languages) - if self.datasets is not None: - gguf_writer.add_datasets(self.datasets) diff --git a/gguf-py/scripts/__init__.py b/gguf-py/gguf/scripts/__init__.py similarity index 100% rename from gguf-py/scripts/__init__.py rename to gguf-py/gguf/scripts/__init__.py diff --git a/gguf-py/scripts/gguf_convert_endian.py b/gguf-py/gguf/scripts/gguf_convert_endian.py similarity index 97% rename from gguf-py/scripts/gguf_convert_endian.py rename to gguf-py/gguf/scripts/gguf_convert_endian.py index b698af0fe..f97e91bd4 100755 --- a/gguf-py/scripts/gguf_convert_endian.py +++ b/gguf-py/gguf/scripts/gguf_convert_endian.py @@ -11,8 +11,8 @@ from pathlib import Path import numpy as np # Necessary to load the local gguf package -if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent / 'gguf-py').exists(): - sys.path.insert(0, str(Path(__file__).parent.parent)) +if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent.parent / 'gguf-py').exists(): + sys.path.insert(0, str(Path(__file__).parent.parent.parent)) import gguf diff --git a/gguf-py/scripts/gguf_dump.py b/gguf-py/gguf/scripts/gguf_dump.py similarity index 99% rename from gguf-py/scripts/gguf_dump.py rename to gguf-py/gguf/scripts/gguf_dump.py index 1b6546541..f95b4fd48 100755 --- a/gguf-py/scripts/gguf_dump.py +++ b/gguf-py/gguf/scripts/gguf_dump.py @@ -12,8 +12,8 @@ from typing import Any import numpy as np # Necessary to load the local gguf package -if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent / 'gguf-py').exists(): - sys.path.insert(0, str(Path(__file__).parent.parent)) +if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent.parent / 'gguf-py').exists(): + sys.path.insert(0, str(Path(__file__).parent.parent.parent)) from gguf import GGUFReader, GGUFValueType, ReaderTensor # noqa: E402 diff --git a/gguf-py/scripts/gguf_hash.py b/gguf-py/gguf/scripts/gguf_hash.py similarity index 97% rename from gguf-py/scripts/gguf_hash.py rename to gguf-py/gguf/scripts/gguf_hash.py index ee34d09bf..3ef989921 100755 --- a/gguf-py/scripts/gguf_hash.py +++ b/gguf-py/gguf/scripts/gguf_hash.py @@ -13,8 +13,8 @@ from pathlib import Path from tqdm import tqdm # Necessary to load the local gguf package -if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent / 'gguf-py').exists(): - sys.path.insert(0, str(Path(__file__).parent.parent)) +if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent.parent / 'gguf-py').exists(): + sys.path.insert(0, str(Path(__file__).parent.parent.parent)) from gguf import GGUFReader # noqa: E402 diff --git a/gguf-py/scripts/gguf_new_metadata.py b/gguf-py/gguf/scripts/gguf_new_metadata.py similarity index 98% rename from gguf-py/scripts/gguf_new_metadata.py rename to gguf-py/gguf/scripts/gguf_new_metadata.py index fce52a8c1..a8cfc9d58 100755 --- a/gguf-py/scripts/gguf_new_metadata.py +++ b/gguf-py/gguf/scripts/gguf_new_metadata.py @@ -13,8 +13,8 @@ from tqdm import tqdm from typing import Any, Sequence, NamedTuple # Necessary to load the local gguf package -if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent / 'gguf-py').exists(): - sys.path.insert(0, str(Path(__file__).parent.parent)) +if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent.parent / 'gguf-py').exists(): + sys.path.insert(0, str(Path(__file__).parent.parent.parent)) import gguf diff --git a/gguf-py/scripts/gguf_set_metadata.py b/gguf-py/gguf/scripts/gguf_set_metadata.py similarity index 97% rename from gguf-py/scripts/gguf_set_metadata.py rename to gguf-py/gguf/scripts/gguf_set_metadata.py index e35b651b8..f5809c35c 100755 --- a/gguf-py/scripts/gguf_set_metadata.py +++ b/gguf-py/gguf/scripts/gguf_set_metadata.py @@ -6,8 +6,8 @@ import sys from pathlib import Path # Necessary to load the local gguf package -if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent / 'gguf-py').exists(): - sys.path.insert(0, str(Path(__file__).parent.parent)) +if "NO_LOCAL_GGUF" not in os.environ and (Path(__file__).parent.parent.parent.parent / 'gguf-py').exists(): + sys.path.insert(0, str(Path(__file__).parent.parent.parent)) from gguf import GGUFReader # noqa: E402 diff --git a/gguf-py/gguf/tensor_mapping.py b/gguf-py/gguf/tensor_mapping.py index f4a787c56..617791e24 100644 --- a/gguf-py/gguf/tensor_mapping.py +++ b/gguf-py/gguf/tensor_mapping.py @@ -13,7 +13,7 @@ class TensorNameMap: "transformer.wte", # gpt2 gpt-j mpt refact qwen dbrx jais exaone "transformer.word_embeddings", # falcon "word_embeddings", # bloom - "model.embed_tokens", # llama-hf nemotron olmoe + "model.embed_tokens", # llama-hf nemotron olmoe olmo2 rwkv6qwen2 "tok_embeddings", # llama-pth "embeddings.word_embeddings", # bert nomic-bert "language_model.embedding.word_embeddings", # persimmon @@ -42,6 +42,7 @@ class TensorNameMap: "emb_ln", # nomic-bert "transformer.norm", # openelm "rwkv.blocks.0.pre_ln", # rwkv + "backbone.norm", # wavtokenizer ), # Position embeddings @@ -54,19 +55,20 @@ class TensorNameMap: # Output MODEL_TENSOR.OUTPUT: ( "embed_out", # gptneox - "lm_head", # gpt2 mpt falcon llama-hf baichuan qwen mamba dbrx jais nemotron exaone olmoe + "lm_head", # gpt2 mpt falcon llama-hf baichuan qwen mamba dbrx jais nemotron exaone olmoe olmo2 phimoe "output", # llama-pth bloom internlm2 "word_embeddings_for_head", # persimmon "lm_head.linear", # phi2 "output_layer", # chatglm "head", # rwkv + "head.out", # wavtokenizer ), # Output norm MODEL_TENSOR.OUTPUT_NORM: ( "gpt_neox.final_layer_norm", # gptneox "transformer.ln_f", # gpt2 gpt-j falcon jais exaone - "model.norm", # llama-hf baichuan internlm2 olmoe + "model.norm", # llama-hf baichuan internlm2 olmoe olmo2 phimoe "norm", # llama-pth "transformer.norm_f", # mpt dbrx "ln_f", # refact bloom qwen gpt2 @@ -80,6 +82,7 @@ class TensorNameMap: "transformer.norm", # openelm "model.norm", # nemotron "rwkv.ln_out", # rwkv + "backbone.final_layer_norm", # wavtokenizer ), # Rope frequencies @@ -90,6 +93,10 @@ class TensorNameMap: MODEL_TENSOR.ROPE_FACTORS_LONG: (), MODEL_TENSOR.ROPE_FACTORS_SHORT: (), + + MODEL_TENSOR.CONV1D: ( + "backbone.embed", # roberta + ), } block_mappings_cfg: dict[MODEL_TENSOR, tuple[str, ...]] = { @@ -101,7 +108,7 @@ class TensorNameMap: "transformer.h.{bid}.input_layernorm", # falcon7b "h.{bid}.input_layernorm", # bloom "transformer.h.{bid}.ln_mlp", # falcon40b - "model.layers.{bid}.input_layernorm", # llama-hf nemotron olmoe + "model.layers.{bid}.input_layernorm", # llama-hf nemotron olmoe phimoe "layers.{bid}.attention_norm", # llama-pth "language_model.encoder.layers.{bid}.input_layernorm", # persimmon "model.layers.{bid}.ln1", # yi @@ -145,7 +152,8 @@ class TensorNameMap: # Attention query MODEL_TENSOR.ATTN_Q: ( - "model.layers.{bid}.self_attn.q_proj", # llama-hf nemotron olmoe + "model.layers.{bid}.self_attn.q_proj", # llama-hf nemotron olmoe olmo2 phimoe + "model.layers.{bid}.self_attn.q_proj_no_perm", # llama-custom "layers.{bid}.attention.wq", # llama-pth "encoder.layer.{bid}.attention.self.query", # bert "transformer.h.{bid}.attn.q_proj", # gpt-j @@ -157,7 +165,8 @@ class TensorNameMap: # Attention key MODEL_TENSOR.ATTN_K: ( - "model.layers.{bid}.self_attn.k_proj", # llama-hf nemotron olmoe + "model.layers.{bid}.self_attn.k_proj", # llama-hf nemotron olmoe olmo2 phimoe + "model.layers.{bid}.self_attn.k_proj_no_perm", # llama-custom "layers.{bid}.attention.wk", # llama-pth "encoder.layer.{bid}.attention.self.key", # bert "transformer.h.{bid}.attn.k_proj", # gpt-j @@ -170,7 +179,7 @@ class TensorNameMap: # Attention value MODEL_TENSOR.ATTN_V: ( - "model.layers.{bid}.self_attn.v_proj", # llama-hf nemotron olmoe + "model.layers.{bid}.self_attn.v_proj", # llama-hf nemotron olmoe olmo2 phimoe "layers.{bid}.attention.wv", # llama-pth "encoder.layer.{bid}.attention.self.value", # bert "transformer.h.{bid}.attn.v_proj", # gpt-j @@ -188,7 +197,8 @@ class TensorNameMap: "transformer.blocks.{bid}.attn.out_proj", # mpt "transformer.h.{bid}.self_attention.dense", # falcon "h.{bid}.self_attention.dense", # bloom - "model.layers.{bid}.self_attn.o_proj", # llama-hf nemotron olmoe + "model.layers.{bid}.self_attn.o_proj", # llama-hf nemotron olmoe olmo2 phimoe + "model.layers.{bid}.self_attn.linear_attn", # deci "layers.{bid}.attention.wo", # llama-pth "encoder.layer.{bid}.attention.output.dense", # bert "transformer.h.{bid}.attn.out_proj", # gpt-j @@ -215,7 +225,7 @@ class TensorNameMap: ), MODEL_TENSOR.ATTN_POST_NORM: ( - "model.layers.{bid}.post_attention_layernorm", # gemma2 + "model.layers.{bid}.post_attention_layernorm", # gemma2 olmo2 ), # Rotary embeddings @@ -232,7 +242,7 @@ class TensorNameMap: "transformer.h.{bid}.ln_2", # gpt2 refact qwen jais exaone "h.{bid}.post_attention_layernorm", # bloom "transformer.blocks.{bid}.norm_2", # mpt - "model.layers.{bid}.post_attention_layernorm", # llama-hf nemotron olmoe + "model.layers.{bid}.post_attention_layernorm", # llama-hf nemotron olmoe phimoe "layers.{bid}.ffn_norm", # llama-pth "language_model.encoder.layers.{bid}.post_attention_layernorm", # persimmon "model.layers.{bid}.ln2", # yi @@ -250,12 +260,12 @@ class TensorNameMap: # Post feed-forward norm MODEL_TENSOR.FFN_POST_NORM: ( - "model.layers.{bid}.post_feedforward_layernorm", # gemma2 + "model.layers.{bid}.post_feedforward_layernorm", # gemma2 olmo2 ), MODEL_TENSOR.FFN_GATE_INP: ( "layers.{bid}.feed_forward.gate", # mixtral - "model.layers.{bid}.block_sparse_moe.gate", # mixtral + "model.layers.{bid}.block_sparse_moe.gate", # mixtral phimoe "model.layers.{bid}.mlp.gate", # qwen2moe olmoe "transformer.decoder_layer.{bid}.router", # Grok "transformer.blocks.{bid}.ffn.router.layer", # dbrx @@ -266,6 +276,10 @@ class TensorNameMap: "model.layers.{bid}.mlp.shared_expert_gate", # qwen2moe ), + MODEL_TENSOR.FFN_EXP_PROBS_B: ( + "model.layers.{bid}.mlp.gate.e_score_correction", # deepseek-v3 + ), + # Feed-forward up MODEL_TENSOR.FFN_UP: ( "gpt_neox.layers.{bid}.mlp.dense_h_to_4h", # gptneox @@ -273,7 +287,7 @@ class TensorNameMap: "transformer.blocks.{bid}.ffn.up_proj", # mpt "transformer.h.{bid}.mlp.dense_h_to_4h", # falcon "h.{bid}.mlp.dense_h_to_4h", # bloom - "model.layers.{bid}.mlp.up_proj", # llama-hf refact nemotron + "model.layers.{bid}.mlp.up_proj", # llama-hf refact nemotron olmo2 "layers.{bid}.feed_forward.w3", # llama-pth "encoder.layer.{bid}.intermediate.dense", # bert "transformer.h.{bid}.mlp.fc_in", # gpt-j @@ -296,15 +310,16 @@ class TensorNameMap: ), MODEL_TENSOR.FFN_UP_EXP: ( - "layers.{bid}.feed_forward.experts.w3", # mixtral (merged) - "transformer.decoder_layer.{bid}.moe.linear_v", # Grok (merged) - "transformer.blocks.{bid}.ffn.experts.mlp.v1", # dbrx - "model.layers.{bid}.mlp.experts.up_proj", # qwen2moe olmoe (merged) + "layers.{bid}.feed_forward.experts.w3", # mixtral (merged) + "transformer.decoder_layer.{bid}.moe.linear_v", # Grok (merged) + "transformer.blocks.{bid}.ffn.experts.mlp.v1", # dbrx + "model.layers.{bid}.mlp.experts.up_proj", # qwen2moe olmoe (merged) + "model.layers.{bid}.block_sparse_moe.experts.w3", # phimoe (merged) ), MODEL_TENSOR.FFN_UP_SHEXP: ( "model.layers.{bid}.mlp.shared_expert.up_proj", # qwen2moe - "model.layers.{bid}.mlp.shared_experts.up_proj", # deepseek2 + "model.layers.{bid}.mlp.shared_experts.up_proj", # deepseek deepseek2 ), # AWQ-activation gate @@ -314,7 +329,7 @@ class TensorNameMap: # Feed-forward gate MODEL_TENSOR.FFN_GATE: ( - "model.layers.{bid}.mlp.gate_proj", # llama-hf refact + "model.layers.{bid}.mlp.gate_proj", # llama-hf refact olmo2 "layers.{bid}.feed_forward.w1", # llama-pth "transformer.h.{bid}.mlp.w2", # qwen "transformer.h.{bid}.mlp.c_fc2", # jais @@ -328,15 +343,16 @@ class TensorNameMap: ), MODEL_TENSOR.FFN_GATE_EXP: ( - "layers.{bid}.feed_forward.experts.w1", # mixtral (merged) - "transformer.decoder_layer.{bid}.moe.linear", # Grok (merged) - "transformer.blocks.{bid}.ffn.experts.mlp.w1", # dbrx - "model.layers.{bid}.mlp.experts.gate_proj", # qwen2moe olmoe (merged) + "layers.{bid}.feed_forward.experts.w1", # mixtral (merged) + "transformer.decoder_layer.{bid}.moe.linear", # Grok (merged) + "transformer.blocks.{bid}.ffn.experts.mlp.w1", # dbrx + "model.layers.{bid}.mlp.experts.gate_proj", # qwen2moe olmoe (merged) + "model.layers.{bid}.block_sparse_moe.experts.w1", # phimoe (merged) ), MODEL_TENSOR.FFN_GATE_SHEXP: ( "model.layers.{bid}.mlp.shared_expert.gate_proj", # qwen2moe - "model.layers.{bid}.mlp.shared_experts.gate_proj", # deepseek2 + "model.layers.{bid}.mlp.shared_experts.gate_proj", # deepseek deepseek2 ), # Feed-forward down @@ -346,7 +362,7 @@ class TensorNameMap: "transformer.blocks.{bid}.ffn.down_proj", # mpt "transformer.h.{bid}.mlp.dense_4h_to_h", # falcon "h.{bid}.mlp.dense_4h_to_h", # bloom - "model.layers.{bid}.mlp.down_proj", # llama-hf nemotron + "model.layers.{bid}.mlp.down_proj", # llama-hf nemotron olmo2 "layers.{bid}.feed_forward.w2", # llama-pth "encoder.layer.{bid}.output.dense", # bert "transformer.h.{bid}.mlp.fc_out", # gpt-j @@ -373,17 +389,18 @@ class TensorNameMap: "transformer.blocks.{bid}.ffn.experts.mlp.w2", # dbrx "model.layers.{bid}.mlp.experts.down_proj", # qwen2moe olmoe (merged) "model.layers.{bid}.block_sparse_moe.output_linear", # granitemoe + "model.layers.{bid}.block_sparse_moe.experts.w2", # phimoe (merged) ), MODEL_TENSOR.FFN_DOWN_SHEXP: ( "model.layers.{bid}.mlp.shared_expert.down_proj", # qwen2moe - "model.layers.{bid}.mlp.shared_experts.down_proj", # deepseek2 + "model.layers.{bid}.mlp.shared_experts.down_proj", # deepseek deepseek2 ), MODEL_TENSOR.ATTN_Q_NORM: ( "language_model.encoder.layers.{bid}.self_attention.q_layernorm", "model.layers.{bid}.self_attn.q_layernorm", # persimmon - "model.layers.{bid}.self_attn.q_norm", # cohere olmoe chameleon + "model.layers.{bid}.self_attn.q_norm", # cohere olmoe chameleon olmo2 "transformer.blocks.{bid}.attn.q_ln", # sea-lion "encoder.layer.{bid}.attention.self.layer_norm_q", # jina-bert-v2 "transformer.layers.{bid}.attn.q_norm", # openelm @@ -392,7 +409,7 @@ class TensorNameMap: MODEL_TENSOR.ATTN_K_NORM: ( "language_model.encoder.layers.{bid}.self_attention.k_layernorm", "model.layers.{bid}.self_attn.k_layernorm", # persimmon - "model.layers.{bid}.self_attn.k_norm", # cohere olmoe chameleon + "model.layers.{bid}.self_attn.k_norm", # cohere olmoe chameleon olmo2 "transformer.blocks.{bid}.attn.k_ln", # sea-lion "encoder.layer.{bid}.attention.self.layer_norm_k", # jina-bert-v2 "transformer.layers.{bid}.attn.k_norm", # openelm @@ -447,34 +464,42 @@ class TensorNameMap: MODEL_TENSOR.TIME_MIX_W1: ( "rwkv.blocks.{bid}.attention.time_maa_w1", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_w1", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_W2: ( "rwkv.blocks.{bid}.attention.time_maa_w2", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_w2", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_X: ( "rwkv.blocks.{bid}.attention.time_maa_x", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_x", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_K: ( "rwkv.blocks.{bid}.attention.time_maa_k", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_k", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_V: ( "rwkv.blocks.{bid}.attention.time_maa_v", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_v", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_R: ( "rwkv.blocks.{bid}.attention.time_maa_r", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_r", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_G: ( "rwkv.blocks.{bid}.attention.time_maa_g", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_g", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LERP_W: ( "rwkv.blocks.{bid}.attention.time_maa_w", # rwkv v6 + "model.layers.{bid}.self_attn.time_maa_w", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_FIRST: ( @@ -483,30 +508,37 @@ class TensorNameMap: MODEL_TENSOR.TIME_MIX_DECAY: ( "rwkv.blocks.{bid}.attention.time_decay", # rwkv v6 + "model.layers.{bid}.self_attn.time_decay", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_DECAY_W1: ( "rwkv.blocks.{bid}.attention.time_decay_w1", # rwkv v6 + "model.layers.{bid}.self_attn.time_decay_w1", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_DECAY_W2: ( "rwkv.blocks.{bid}.attention.time_decay_w2", # rwkv v6 + "model.layers.{bid}.self_attn.time_decay_w2", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_KEY: ( - "rwkv.blocks.{bid}.attention.key", # rwkv + "rwkv.blocks.{bid}.attention.key", # rwkv + "model.layers.{bid}.self_attn.k_proj", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_VALUE: ( - "rwkv.blocks.{bid}.attention.value", # rwkv + "rwkv.blocks.{bid}.attention.value", # rwkv + "model.layers.{bid}.self_attn.v_proj", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_RECEPTANCE: ( "rwkv.blocks.{bid}.attention.receptance", # rwkv + "model.layers.{bid}.self_attn.q_proj", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_GATE: ( - "rwkv.blocks.{bid}.attention.gate", # rwkv + "rwkv.blocks.{bid}.attention.gate", # rwkv + "model.layers.{bid}.self_attn.gate", # rwkv6qwen2 ), MODEL_TENSOR.TIME_MIX_LN: ( @@ -514,7 +546,8 @@ class TensorNameMap: ), MODEL_TENSOR.TIME_MIX_OUTPUT: ( - "rwkv.blocks.{bid}.attention.output", # rwkv + "rwkv.blocks.{bid}.attention.output", # rwkv + "model.layers.{bid}.self_attn.o_proj", # rwkv6qwen2 ), MODEL_TENSOR.CHANNEL_MIX_LERP_K: ( @@ -679,6 +712,8 @@ class TensorNameMap: "encoder.block.{bid}.layer.1.DenseReluDense.wo", # t5 ), + ############################################################################ + # TODO: these do not belong to block_mappings_cfg - move them to mappings_cfg MODEL_TENSOR.ENC_OUTPUT_NORM: ( "encoder.final_layer_norm", # t5 ), @@ -691,6 +726,67 @@ class TensorNameMap: MODEL_TENSOR.CLS_OUT: ( "classifier.out_proj", # roberta ), + ############################################################################# + + MODEL_TENSOR.CONVNEXT_DW: ( + "backbone.convnext.{bid}.dwconv", # wavtokenizer + ), + + MODEL_TENSOR.CONVNEXT_NORM: ( + "backbone.convnext.{bid}.norm", # wavtokenizer + ), + + MODEL_TENSOR.CONVNEXT_PW1: ( + "backbone.convnext.{bid}.pwconv1", # wavtokenizer + ), + + MODEL_TENSOR.CONVNEXT_PW2: ( + "backbone.convnext.{bid}.pwconv2", # wavtokenizer + ), + + MODEL_TENSOR.CONVNEXT_GAMMA: ( + "backbone.convnext.{bid}.gamma", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_CONV1: ( + "backbone.posnet.{bid}.conv1", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_CONV2: ( + "backbone.posnet.{bid}.conv2", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_NORM: ( + "backbone.posnet.{bid}.norm", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_NORM1: ( + "backbone.posnet.{bid}.norm1", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_NORM2: ( + "backbone.posnet.{bid}.norm2", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_ATTN_NORM: ( + "backbone.posnet.{bid}.norm", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_ATTN_Q: ( + "backbone.posnet.{bid}.q", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_ATTN_K: ( + "backbone.posnet.{bid}.k", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_ATTN_V: ( + "backbone.posnet.{bid}.v", # wavtokenizer + ), + + MODEL_TENSOR.POSNET_ATTN_OUT: ( + "backbone.posnet.{bid}.proj_out", # wavtokenizer + ), } # architecture-specific block mappings diff --git a/gguf-py/pyproject.toml b/gguf-py/pyproject.toml index 33cfe26b7..78c6baa64 100644 --- a/gguf-py/pyproject.toml +++ b/gguf-py/pyproject.toml @@ -1,12 +1,11 @@ [tool.poetry] name = "gguf" -version = "0.10.0" +version = "0.15.0" description = "Read and write ML models in GGUF for GGML" authors = ["GGML "] packages = [ {include = "gguf"}, {include = "gguf/py.typed"}, - {include = "scripts"}, ] readme = "README.md" homepage = "https://ggml.ai" @@ -33,7 +32,7 @@ requires = ["poetry-core>=1.0.0"] build-backend = "poetry.core.masonry.api" [tool.poetry.scripts] -gguf-convert-endian = "scripts:gguf_convert_endian_entrypoint" -gguf-dump = "scripts:gguf_dump_entrypoint" -gguf-set-metadata = "scripts:gguf_set_metadata_entrypoint" -gguf-new-metadata = "scripts:gguf_new_metadata_entrypoint" +gguf-convert-endian = "gguf.scripts:gguf_convert_endian_entrypoint" +gguf-dump = "gguf.scripts:gguf_dump_entrypoint" +gguf-set-metadata = "gguf.scripts:gguf_set_metadata_entrypoint" +gguf-new-metadata = "gguf.scripts:gguf_new_metadata_entrypoint" diff --git a/gguf-py/tests/test_metadata.py b/gguf-py/tests/test_metadata.py index 81a2a30ae..40d484f4e 100755 --- a/gguf-py/tests/test_metadata.py +++ b/gguf-py/tests/test_metadata.py @@ -182,8 +182,43 @@ class TestMetadataMethod(unittest.TestCase): expect.base_models=[{'name': 'Mistral 7B Merge 14 v0', 'organization': 'EmbeddedLLM', 'version': '14-v0', 'repo_url': 'https://huggingface.co/EmbeddedLLM/Mistral-7B-Merge-14-v0'}, {'name': 'Trinity v1', 'organization': 'Janai Hq', 'version': 'v1', 'repo_url': 'https://huggingface.co/janai-hq/trinity-v1'}] expect.tags=['Llama-3', 'instruct', 'finetune', 'chatml', 'DPO', 'RLHF', 'gpt4', 'synthetic data', 'distillation', 'function calling', 'json mode', 'axolotl'] expect.languages=['en'] - expect.datasets=['teknium/OpenHermes-2.5'] + expect.datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}] + self.assertEqual(got, expect) + # Base Model spec is inferred from model id + model_card = {'base_models': 'teknium/OpenHermes-2.5'} + expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) + self.assertEqual(got, expect) + + # Base Model spec is only url + model_card = {'base_models': ['https://huggingface.co/teknium/OpenHermes-2.5']} + expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) + self.assertEqual(got, expect) + + # Base Model spec is given directly + model_card = {'base_models': [{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]} + expect = gguf.Metadata(base_models=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) + self.assertEqual(got, expect) + + # Dataset spec is inferred from model id + model_card = {'datasets': 'teknium/OpenHermes-2.5'} + expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) + self.assertEqual(got, expect) + + # Dataset spec is only url + model_card = {'datasets': ['https://huggingface.co/teknium/OpenHermes-2.5']} + expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) + self.assertEqual(got, expect) + + # Dataset spec is given directly + model_card = {'datasets': [{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]} + expect = gguf.Metadata(datasets=[{'name': 'OpenHermes 2.5', 'organization': 'Teknium', 'version': '2.5', 'repo_url': 'https://huggingface.co/teknium/OpenHermes-2.5'}]) + got = gguf.Metadata.apply_metadata_heuristic(gguf.Metadata(), model_card, None, None) self.assertEqual(got, expect) def test_apply_metadata_heuristic_from_hf_parameters(self): diff --git a/gguf-py/tests/test_quants.py b/gguf-py/tests/test_quants.py index 762067814..f04d5acce 100755 --- a/gguf-py/tests/test_quants.py +++ b/gguf-py/tests/test_quants.py @@ -136,7 +136,7 @@ def compare_tensors(t1: np.ndarray, t2: np.ndarray, qtype: GGMLQuantizationType) logger.debug(f"Sample bad block ({diff_bits[bad_block_id]} differing bits):\n{t1[bad_block_id]}\nReference:\n{t2[bad_block_id]}") sum_diff_bits = np.sum(diff_bits) - logger.debug(f"{sum_diff_bits} bits differ ({100 * sum_diff_bits/(x.size * 8):.6f}%)") + logger.debug(f"{sum_diff_bits} bits differ ({100 * sum_diff_bits / (x.size * 8):.6f}%)") return False diff --git a/grammars/README.md b/grammars/README.md index 4e57bca5f..976954091 100644 --- a/grammars/README.md +++ b/grammars/README.md @@ -46,7 +46,7 @@ Terminals support the full range of Unicode. Unicode characters can be specified Character ranges can be negated with `^`: ``` -single-line ::= [^\n]+ "\n"` +single-line ::= [^\n]+ "\n" ``` ## Sequences and Alternatives diff --git a/grammars/english.gbnf b/grammars/english.gbnf new file mode 100644 index 000000000..2e53686c8 --- /dev/null +++ b/grammars/english.gbnf @@ -0,0 +1,6 @@ +# note: this might be incomplete, mostly an example +root ::= en-char+ ([ \t\n] en-char+)* +en-char ::= letter | digit | punctuation +letter ::= [a-zA-Z] +digit ::= [0-9] +punctuation ::= [!"#$%&'()*+,-./:;<=>?@[\\\]^_`{|}~] diff --git a/include/llama-cpp.h b/include/llama-cpp.h new file mode 100644 index 000000000..8f6368177 --- /dev/null +++ b/include/llama-cpp.h @@ -0,0 +1,30 @@ +#pragma once + +#ifndef __cplusplus +#error "This header is for C++ only" +#endif + +#include + +#include "llama.h" + +struct llama_model_deleter { + void operator()(llama_model * model) { llama_model_free(model); } +}; + +struct llama_context_deleter { + void operator()(llama_context * context) { llama_free(context); } +}; + +struct llama_sampler_deleter { + void operator()(llama_sampler * sampler) { llama_sampler_free(sampler); } +}; + +struct llama_adapter_lora_deleter { + void operator()(llama_adapter_lora * adapter) { llama_adapter_lora_free(adapter); } +}; + +typedef std::unique_ptr llama_model_ptr; +typedef std::unique_ptr llama_context_ptr; +typedef std::unique_ptr llama_sampler_ptr; +typedef std::unique_ptr llama_adapter_lora_ptr; diff --git a/include/llama.h b/include/llama.h index ccb48f73c..a184884c7 100644 --- a/include/llama.h +++ b/include/llama.h @@ -34,7 +34,6 @@ #define LLAMA_DEFAULT_SEED 0xFFFFFFFF -// TODO: use everywhere in the implementation #define LLAMA_TOKEN_NULL -1 #define LLAMA_FILE_MAGIC_GGLA 0x67676c61u // 'ggla' @@ -57,7 +56,7 @@ extern "C" { // TODO: show sample usage // - // struct llama_vocab; // TODO: add in the future + struct llama_vocab; struct llama_model; struct llama_context; struct llama_sampler; @@ -104,12 +103,16 @@ extern "C" { LLAMA_VOCAB_PRE_TYPE_GPT3_FINNISH = 24, LLAMA_VOCAB_PRE_TYPE_EXAONE = 25, LLAMA_VOCAB_PRE_TYPE_CHAMELEON = 26, + LLAMA_VOCAB_PRE_TYPE_MINERVA = 27, + LLAMA_VOCAB_PRE_TYPE_DEEPSEEK3_LLM = 28, }; enum llama_rope_type { - LLAMA_ROPE_TYPE_NONE = -1, - LLAMA_ROPE_TYPE_NORM = 0, - LLAMA_ROPE_TYPE_NEOX = GGML_ROPE_TYPE_NEOX, + LLAMA_ROPE_TYPE_NONE = -1, + LLAMA_ROPE_TYPE_NORM = 0, + LLAMA_ROPE_TYPE_NEOX = GGML_ROPE_TYPE_NEOX, + LLAMA_ROPE_TYPE_MROPE = GGML_ROPE_TYPE_MROPE, + LLAMA_ROPE_TYPE_VISION = GGML_ROPE_TYPE_VISION, }; enum llama_token_type { //TODO: remove, required until per token attributes are available from GGUF file @@ -171,9 +174,9 @@ extern "C" { LLAMA_FTYPE_MOSTLY_IQ4_XS = 30, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ1_M = 31, // except 1d tensors LLAMA_FTYPE_MOSTLY_BF16 = 32, // except 1d tensors - LLAMA_FTYPE_MOSTLY_Q4_0_4_4 = 33, // except 1d tensors - LLAMA_FTYPE_MOSTLY_Q4_0_4_8 = 34, // except 1d tensors - LLAMA_FTYPE_MOSTLY_Q4_0_8_8 = 35, // except 1d tensors + //LLAMA_FTYPE_MOSTLY_Q4_0_4_4 = 33, // removed from gguf files, use Q4_0 and runtime repack + //LLAMA_FTYPE_MOSTLY_Q4_0_4_8 = 34, // removed from gguf files, use Q4_0 and runtime repack + //LLAMA_FTYPE_MOSTLY_Q4_0_8_8 = 35, // removed from gguf files, use Q4_0 and runtime repack LLAMA_FTYPE_MOSTLY_TQ1_0 = 36, // except 1d tensors LLAMA_FTYPE_MOSTLY_TQ2_0 = 37, // except 1d tensors @@ -185,7 +188,8 @@ extern "C" { LLAMA_ROPE_SCALING_TYPE_NONE = 0, LLAMA_ROPE_SCALING_TYPE_LINEAR = 1, LLAMA_ROPE_SCALING_TYPE_YARN = 2, - LLAMA_ROPE_SCALING_TYPE_MAX_VALUE = LLAMA_ROPE_SCALING_TYPE_YARN, + LLAMA_ROPE_SCALING_TYPE_LONGROPE = 3, + LLAMA_ROPE_SCALING_TYPE_MAX_VALUE = LLAMA_ROPE_SCALING_TYPE_LONGROPE, }; enum llama_pooling_type { @@ -272,6 +276,9 @@ extern "C" { }; struct llama_model_params { + // NULL-terminated list of devices to use for offloading (if NULL, all available devices are used) + ggml_backend_dev_t * devices; + int32_t n_gpu_layers; // number of layers to store in VRAM enum llama_split_mode split_mode; // how to split the model across multiple GPUs @@ -378,7 +385,7 @@ extern "C" { } llama_chat_message; // lora adapter - struct llama_lora_adapter; + struct llama_adapter_lora; // Helpers for getting default parameters // TODO: update API to start accepting pointers to params structs (https://github.com/ggerganov/llama.cpp/discussions/9172) @@ -392,30 +399,43 @@ extern "C" { // Call once at the start of the program LLAMA_API void llama_backend_init(void); + // Call once at the end of the program - currently only used for MPI + LLAMA_API void llama_backend_free(void); + //optional: LLAMA_API void llama_numa_init(enum ggml_numa_strategy numa); // Optional: an auto threadpool gets created in ggml if not passed explicitly LLAMA_API void llama_attach_threadpool( - struct llama_context * ctx, - ggml_threadpool_t threadpool, - ggml_threadpool_t threadpool_batch); + struct llama_context * ctx, + ggml_threadpool_t threadpool, + ggml_threadpool_t threadpool_batch); + LLAMA_API void llama_detach_threadpool(struct llama_context * ctx); - // Call once at the end of the program - currently only used for MPI - LLAMA_API void llama_backend_free(void); + DEPRECATED(LLAMA_API struct llama_model * llama_load_model_from_file( + const char * path_model, + struct llama_model_params params), + "use llama_model_load_from_file instead"); - LLAMA_API struct llama_model * llama_load_model_from_file( + LLAMA_API struct llama_model * llama_model_load_from_file( const char * path_model, struct llama_model_params params); - LLAMA_API void llama_free_model(struct llama_model * model); + DEPRECATED(LLAMA_API void llama_free_model(struct llama_model * model), + "use llama_model_free instead"); - // TODO: rename to llama_init_from_model - LLAMA_API struct llama_context * llama_new_context_with_model( + LLAMA_API void llama_model_free(struct llama_model * model); + + LLAMA_API struct llama_context * llama_init_from_model( struct llama_model * model, struct llama_context_params params); + DEPRECATED(LLAMA_API struct llama_context * llama_new_context_with_model( + struct llama_model * model, + struct llama_context_params params), + "use llama_init_from_model instead"); + // Frees all allocated memory LLAMA_API void llama_free(struct llama_context * ctx); @@ -433,24 +453,35 @@ extern "C" { LLAMA_API uint32_t llama_n_ubatch (const struct llama_context * ctx); LLAMA_API uint32_t llama_n_seq_max (const struct llama_context * ctx); - LLAMA_API int32_t llama_n_vocab (const struct llama_model * model); - LLAMA_API int32_t llama_n_ctx_train(const struct llama_model * model); - LLAMA_API int32_t llama_n_embd (const struct llama_model * model); - LLAMA_API int32_t llama_n_layer (const struct llama_model * model); - LLAMA_API int32_t llama_n_head (const struct llama_model * model); + DEPRECATED(LLAMA_API int32_t llama_n_ctx_train(const struct llama_model * model), "use llama_model_n_ctx_train instead"); + DEPRECATED(LLAMA_API int32_t llama_n_embd (const struct llama_model * model), "use llama_model_n_embd instead"); + DEPRECATED(LLAMA_API int32_t llama_n_layer (const struct llama_model * model), "use llama_model_n_layer instead"); + DEPRECATED(LLAMA_API int32_t llama_n_head (const struct llama_model * model), "use llama_model_n_head instead"); - LLAMA_API const struct llama_model * llama_get_model(const struct llama_context * ctx); + DEPRECATED(LLAMA_API int32_t llama_n_vocab (const struct llama_vocab * vocab), "use llama_vocab_n_tokens instead"); - LLAMA_API enum llama_pooling_type llama_pooling_type(const struct llama_context * ctx); - LLAMA_API enum llama_vocab_type llama_vocab_type (const struct llama_model * model); - LLAMA_API enum llama_rope_type llama_rope_type (const struct llama_model * model); + LLAMA_API const struct llama_model * llama_get_model (const struct llama_context * ctx); + LLAMA_API enum llama_pooling_type llama_pooling_type(const struct llama_context * ctx); + + LLAMA_API const struct llama_vocab * llama_model_get_vocab(const struct llama_model * model); + LLAMA_API enum llama_rope_type llama_model_rope_type(const struct llama_model * model); + + LLAMA_API int32_t llama_model_n_ctx_train(const struct llama_model * model); + LLAMA_API int32_t llama_model_n_embd (const struct llama_model * model); + LLAMA_API int32_t llama_model_n_layer (const struct llama_model * model); + LLAMA_API int32_t llama_model_n_head (const struct llama_model * model); // Get the model's RoPE frequency scaling factor - LLAMA_API float llama_rope_freq_scale_train(const struct llama_model * model); + LLAMA_API float llama_model_rope_freq_scale_train(const struct llama_model * model); + + LLAMA_API enum llama_vocab_type llama_vocab_type(const struct llama_vocab * vocab); + + LLAMA_API int32_t llama_vocab_n_tokens(const struct llama_vocab * vocab); // Functions to access the model's GGUF metadata scalar values // - The functions return the length of the string on success, or -1 on failure // - The output string is always null-terminated and cleared on failure + // - When retrieving a string, an extra byte must be allocated to account for the null terminator // - GGUF array values are not supported by these functions // Get metadata value as a string by key name @@ -471,12 +502,12 @@ extern "C" { // Returns the total size of all the tensors in the model in bytes LLAMA_API uint64_t llama_model_size(const struct llama_model * model); + // Get the default chat template. Returns nullptr if not available + LLAMA_API const char * llama_model_chat_template(const struct llama_model * model); + // Returns the total number of parameters in the model LLAMA_API uint64_t llama_model_n_params(const struct llama_model * model); - // Get a llama model tensor - LLAMA_API struct ggml_tensor * llama_get_model_tensor(struct llama_model * model, const char * name); - // Returns true if the model contains an encoder that requires llama_encode() call LLAMA_API bool llama_model_has_encoder(const struct llama_model * model); @@ -496,32 +527,36 @@ extern "C" { const char * fname_out, const llama_model_quantize_params * params); + // + // Adapters + // + // Load a LoRA adapter from file - // The loaded adapter will be associated to the given model, and will be free when the model is deleted - LLAMA_API struct llama_lora_adapter * llama_lora_adapter_init( + LLAMA_API struct llama_adapter_lora * llama_adapter_lora_init( struct llama_model * model, const char * path_lora); + // Manually free a LoRA adapter + // Note: loaded adapters will be free when the associated model is deleted + LLAMA_API void llama_adapter_lora_free(struct llama_adapter_lora * adapter); + + // The following functions operate on a llama_context, hence the naming: llama_verb_... + // Add a loaded LoRA adapter to given context // This will not modify model's weight - LLAMA_API int32_t llama_lora_adapter_set( + LLAMA_API int32_t llama_set_adapter_lora( struct llama_context * ctx, - struct llama_lora_adapter * adapter, + struct llama_adapter_lora * adapter, float scale); // Remove a specific LoRA adapter from given context // Return -1 if the adapter is not present in the context - LLAMA_API int32_t llama_lora_adapter_remove( + LLAMA_API int32_t llama_rm_adapter_lora( struct llama_context * ctx, - struct llama_lora_adapter * adapter); + struct llama_adapter_lora * adapter); // Remove all LoRA adapters from given context - LLAMA_API void llama_lora_adapter_clear( - struct llama_context * ctx); - - // Manually free a LoRA adapter - // Note: loaded adapters will be free when the associated model is deleted - LLAMA_API void llama_lora_adapter_free(struct llama_lora_adapter * adapter); + LLAMA_API void llama_clear_adapter_lora(struct llama_context * ctx); // Apply a loaded control vector to a llama_context, or if data is NULL, clear // the currently loaded vector. @@ -529,8 +564,8 @@ extern "C" { // to an n_embd x n_layers buffer starting from layer 1. // il_start and il_end are the layer range the vector should apply to (both inclusive) // See llama_control_vector_load in common to load a control vector. - LLAMA_API int32_t llama_control_vector_apply( - struct llama_context * lctx, + LLAMA_API int32_t llama_apply_adapter_cvec( + struct llama_context * ctx, const float * data, size_t len, int32_t n_embd, @@ -541,6 +576,8 @@ extern "C" { // KV cache // + // TODO: remove llama_kv_cache_view_* API + // Information associated with an individual cell in the KV cache view. struct llama_kv_cache_view_cell { // The position for this cell. Takes KV cache shifts into account. @@ -587,8 +624,11 @@ extern "C" { LLAMA_API void llama_kv_cache_view_free(struct llama_kv_cache_view * view); // Update the KV cache view structure with the current state of the KV cache. (use only for debugging purposes) + // TODO: change signature to llama_kv_cache_view_update(struct llama_kv_cache_view * view, const struct llama_context * ctx) LLAMA_API void llama_kv_cache_view_update(const struct llama_context * ctx, struct llama_kv_cache_view * view); + /// + // Returns the number of tokens in the KV cache (slow, use only for debug) // If a KV cell has multiple sequences assigned to it, it will be counted multiple times LLAMA_API int32_t llama_get_kv_cache_token_count(const struct llama_context * ctx); @@ -658,6 +698,9 @@ extern "C" { struct llama_context * ctx, llama_seq_id seq_id); + // TODO: the llama_kv_cache_defrag and llama_kv_cache_update API tightly couples llama_context with llama_kv_cache + // how to avoid this? + // Defragment the KV cache // This will be applied: // - lazily on next llama_decode() @@ -667,6 +710,9 @@ extern "C" { // Apply the KV cache updates (such as K-shifts, defragmentation, etc.) LLAMA_API void llama_kv_cache_update(struct llama_context * ctx); + // Check if the context supports KV cache shifting + LLAMA_API bool llama_kv_cache_can_shift(struct llama_context * ctx); + // // State / sessions // @@ -797,7 +843,7 @@ extern "C" { // Processes a batch of tokens with the ecoder part of the encoder-decoder model. // Stores the encoder output internally for later use by the decoder cross-attention layers. // 0 - success - // < 0 - error + // < 0 - error. the KV cache state is restored to the state before this call LLAMA_API int32_t llama_encode( struct llama_context * ctx, struct llama_batch batch); @@ -805,7 +851,7 @@ extern "C" { // Positive return values does not mean a fatal error, but rather a warning. // 0 - success // 1 - could not find a KV slot for the batch (try reducing the size of the batch or increase the context) - // < 0 - error + // < 0 - error. the KV cache state is restored to the state before this call LLAMA_API int32_t llama_decode( struct llama_context * ctx, struct llama_batch batch); @@ -875,41 +921,60 @@ extern "C" { // Vocab // - LLAMA_API const char * llama_token_get_text(const struct llama_model * model, llama_token token); + LLAMA_API const char * llama_vocab_get_text(const struct llama_vocab * vocab, llama_token token); - LLAMA_API float llama_token_get_score(const struct llama_model * model, llama_token token); + LLAMA_API float llama_vocab_get_score(const struct llama_vocab * vocab, llama_token token); - LLAMA_API enum llama_token_attr llama_token_get_attr(const struct llama_model * model, llama_token token); + LLAMA_API enum llama_token_attr llama_vocab_get_attr(const struct llama_vocab * vocab, llama_token token); // Check if the token is supposed to end generation (end-of-generation, eg. EOS, EOT, etc.) - LLAMA_API bool llama_token_is_eog(const struct llama_model * model, llama_token token); + LLAMA_API bool llama_vocab_is_eog(const struct llama_vocab * vocab, llama_token token); // Identify if Token Id is a control token or a render-able token - LLAMA_API bool llama_token_is_control(const struct llama_model * model, llama_token token); + LLAMA_API bool llama_vocab_is_control(const struct llama_vocab * vocab, llama_token token); // Special tokens - LLAMA_API llama_token llama_token_bos(const struct llama_model * model); // beginning-of-sentence - LLAMA_API llama_token llama_token_eos(const struct llama_model * model); // end-of-sentence - LLAMA_API llama_token llama_token_eot(const struct llama_model * model); // end-of-turn - LLAMA_API llama_token llama_token_cls(const struct llama_model * model); // classification - LLAMA_API llama_token llama_token_sep(const struct llama_model * model); // sentence separator - LLAMA_API llama_token llama_token_nl (const struct llama_model * model); // next-line - LLAMA_API llama_token llama_token_pad(const struct llama_model * model); // padding + LLAMA_API llama_token llama_vocab_bos(const struct llama_vocab * vocab); // beginning-of-sentence + LLAMA_API llama_token llama_vocab_eos(const struct llama_vocab * vocab); // end-of-sentence + LLAMA_API llama_token llama_vocab_eot(const struct llama_vocab * vocab); // end-of-turn + LLAMA_API llama_token llama_vocab_sep(const struct llama_vocab * vocab); // sentence separator + LLAMA_API llama_token llama_vocab_nl (const struct llama_vocab * vocab); // next-line + LLAMA_API llama_token llama_vocab_pad(const struct llama_vocab * vocab); // padding - LLAMA_API bool llama_add_bos_token(const struct llama_model * model); - LLAMA_API bool llama_add_eos_token(const struct llama_model * model); + LLAMA_API bool llama_vocab_get_add_bos(const struct llama_vocab * vocab); + LLAMA_API bool llama_vocab_get_add_eos(const struct llama_vocab * vocab); - // infill tokens - DEPRECATED(LLAMA_API llama_token llama_token_prefix(const struct llama_model * model), "use llama_token_fim_pre instead"); - DEPRECATED(LLAMA_API llama_token llama_token_middle(const struct llama_model * model), "use llama_token_fim_mid instead"); - DEPRECATED(LLAMA_API llama_token llama_token_suffix(const struct llama_model * model), "use llama_token_fim_suf instead"); + LLAMA_API llama_token llama_vocab_fim_pre(const struct llama_vocab * vocab); + LLAMA_API llama_token llama_vocab_fim_suf(const struct llama_vocab * vocab); + LLAMA_API llama_token llama_vocab_fim_mid(const struct llama_vocab * vocab); + LLAMA_API llama_token llama_vocab_fim_pad(const struct llama_vocab * vocab); + LLAMA_API llama_token llama_vocab_fim_rep(const struct llama_vocab * vocab); + LLAMA_API llama_token llama_vocab_fim_sep(const struct llama_vocab * vocab); - LLAMA_API llama_token llama_token_fim_pre(const struct llama_model * model); - LLAMA_API llama_token llama_token_fim_suf(const struct llama_model * model); - LLAMA_API llama_token llama_token_fim_mid(const struct llama_model * model); - LLAMA_API llama_token llama_token_fim_pad(const struct llama_model * model); - LLAMA_API llama_token llama_token_fim_rep(const struct llama_model * model); - LLAMA_API llama_token llama_token_fim_sep(const struct llama_model * model); + DEPRECATED(LLAMA_API const char * llama_token_get_text(const struct llama_vocab * vocab, llama_token token), "use llama_vocabable_get_text instead"); + DEPRECATED(LLAMA_API float llama_token_get_score(const struct llama_vocab * vocab, llama_token token), "use llama_vocab_get_score instead"); + DEPRECATED(LLAMA_API enum llama_token_attr llama_token_get_attr(const struct llama_vocab * vocab, llama_token token), "use llama_vocab_get_attr instead"); + DEPRECATED(LLAMA_API bool llama_token_is_eog(const struct llama_vocab * vocab, llama_token token), "use llama_vocab_is_eog instead"); + DEPRECATED(LLAMA_API bool llama_token_is_control(const struct llama_vocab * vocab, llama_token token), "use llama_vocab_is_control instead"); + DEPRECATED(LLAMA_API llama_token llama_token_bos(const struct llama_vocab * vocab), "use llama_vocab_bos instead"); + DEPRECATED(LLAMA_API llama_token llama_token_eos(const struct llama_vocab * vocab), "use llama_vocab_eos instead"); + DEPRECATED(LLAMA_API llama_token llama_token_eot(const struct llama_vocab * vocab), "use llama_vocab_eot instead"); + DEPRECATED(LLAMA_API llama_token llama_token_cls(const struct llama_vocab * vocab), "use llama_vocab_cls instead"); + DEPRECATED(LLAMA_API llama_token llama_token_sep(const struct llama_vocab * vocab), "use llama_vocab_sep instead"); + DEPRECATED(LLAMA_API llama_token llama_token_nl (const struct llama_vocab * vocab), "use llama_vocab_nl instead"); + DEPRECATED(LLAMA_API llama_token llama_token_pad(const struct llama_vocab * vocab), "use llama_vocab_pad instead"); + DEPRECATED(LLAMA_API bool llama_add_bos_token(const struct llama_vocab * vocab), "use llama_vocab_get_add_bos instead"); + DEPRECATED(LLAMA_API bool llama_add_eos_token(const struct llama_vocab * vocab), "use llama_vocab_get_add_eos instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_pre(const struct llama_vocab * vocab), "use llama_vocab_fim_pre instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_suf(const struct llama_vocab * vocab), "use llama_vocab_fim_suf instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_mid(const struct llama_vocab * vocab), "use llama_vocab_fim_mid instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_pad(const struct llama_vocab * vocab), "use llama_vocab_fim_pad instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_rep(const struct llama_vocab * vocab), "use llama_vocab_fim_rep instead"); + DEPRECATED(LLAMA_API llama_token llama_token_fim_sep(const struct llama_vocab * vocab), "use llama_vocab_fim_sep instead"); + + // CLS is equivalent to BOS + DEPRECATED(LLAMA_API llama_token llama_vocab_cls(const struct llama_vocab * vocab), // classification + "use llama_vocab_bos instead"); // // Tokenization @@ -925,7 +990,7 @@ extern "C" { /// @param parse_special Allow tokenizing special and/or control tokens which otherwise are not exposed and treated /// as plaintext. Does not insert a leading space. LLAMA_API int32_t llama_tokenize( - const struct llama_model * model, + const struct llama_vocab * vocab, const char * text, int32_t text_len, llama_token * tokens, @@ -939,7 +1004,7 @@ extern "C" { // User can skip up to 'lstrip' leading spaces before copying (useful when encoding/decoding multiple tokens with 'add_space_prefix') // @param special If true, special tokens are rendered in the output. LLAMA_API int32_t llama_token_to_piece( - const struct llama_model * model, + const struct llama_vocab * vocab, llama_token token, char * buf, int32_t length, @@ -953,7 +1018,7 @@ extern "C" { /// @param remove_special Allow to remove BOS and EOS tokens if model is configured to do so. /// @param unparse_special If true, special tokens are rendered in the output. LLAMA_API int32_t llama_detokenize( - const struct llama_model * model, + const struct llama_vocab * vocab, const llama_token * tokens, int32_t n_tokens, char * text, @@ -976,7 +1041,6 @@ extern "C" { /// @param length The size of the allocated buffer /// @return The total number of bytes of the formatted prompt. If is it larger than the size of buffer, you may need to re-alloc it and then re-apply the template. LLAMA_API int32_t llama_chat_apply_template( - const struct llama_model * model, const char * tmpl, const struct llama_chat_message * chat, size_t n_msg, @@ -984,6 +1048,9 @@ extern "C" { char * buf, int32_t length); + // Get list of built-in chat templates + LLAMA_API int32_t llama_chat_builtin_templates(const char ** output, size_t len); + // // Sampling API // @@ -1021,7 +1088,6 @@ extern "C" { // llama_sampler_free(smpl); // // TODO: In the future, llama_sampler will be utilized to offload the sampling to the backends (e.g. GPU). - // TODO: in the future, the entire sampling API that uses llama_model should start using llama_vocab // typedef void * llama_sampler_context_t; @@ -1121,24 +1187,21 @@ extern "C" { float eta); LLAMA_API struct llama_sampler * llama_sampler_init_grammar( - const struct llama_model * model, + const struct llama_vocab * vocab, const char * grammar_str, const char * grammar_root); + /// NOTE: Avoid using on the full vocabulary as searching for repeated tokens can become slow. For example, apply top-k or top-p sampling first. LLAMA_API struct llama_sampler * llama_sampler_init_penalties( - int32_t n_vocab, // llama_n_vocab() - llama_token special_eos_id, // llama_token_eos() - llama_token linefeed_id, // llama_token_nl() - int32_t penalty_last_n, // last n tokens to penalize (0 = disable penalty, -1 = context size) - float penalty_repeat, // 1.0 = disabled - float penalty_freq, // 0.0 = disabled - float penalty_present, // 0.0 = disabled - bool penalize_nl, // consider newlines as a repeatable token - bool ignore_eos); // ignore the end-of-sequence token + int32_t penalty_last_n, // last n tokens to penalize (0 = disable penalty, -1 = context size) + float penalty_repeat, // 1.0 = disabled + float penalty_freq, // 0.0 = disabled + float penalty_present); // 0.0 = disabled /// @details DRY sampler, designed by p-e-w, as described in: https://github.com/oobabooga/text-generation-webui/pull/5677, porting Koboldcpp implementation authored by pi6am: https://github.com/LostRuins/koboldcpp/pull/982 - LLAMA_API struct llama_sampler * llama_sampler_init_dry( - const struct llama_model * model, + LLAMA_API struct llama_sampler * llama_sampler_init_dry( + const struct llama_vocab * vocab, + int32_t n_ctx_train, float dry_multiplier, float dry_base, int32_t dry_allowed_length, @@ -1172,7 +1235,7 @@ extern "C" { // 3. discard non-EOG tokens with low prob // 4. if no tokens are left -> pick EOT // - LLAMA_API struct llama_sampler * llama_sampler_init_infill(const struct llama_model * model); + LLAMA_API struct llama_sampler * llama_sampler_init_infill(const struct llama_vocab * vocab); // Returns the seed used by the sampler if applicable, LLAMA_DEFAULT_SEED otherwise LLAMA_API uint32_t llama_sampler_get_seed(const struct llama_sampler * smpl); @@ -1244,8 +1307,6 @@ extern "C" { LLAMA_API void llama_perf_sampler_print(const struct llama_sampler * chain); LLAMA_API void llama_perf_sampler_reset( struct llama_sampler * chain); - LLAMA_API void llama_perf_dump_yaml(FILE * stream, const struct llama_context * ctx); - #ifdef __cplusplus } #endif diff --git a/media/llama-leader.jpeg b/media/llama-leader.jpeg deleted file mode 100644 index 0b4e6e1cf..000000000 Binary files a/media/llama-leader.jpeg and /dev/null differ diff --git a/models/ggml-vocab-roberta-bpe.gguf.inp b/models/ggml-vocab-roberta-bpe.gguf.inp new file mode 100644 index 000000000..9baf7d77a --- /dev/null +++ b/models/ggml-vocab-roberta-bpe.gguf.inp @@ -0,0 +1,112 @@ +ied 4 ½ months +__ggml_vocab_test__ +Führer +__ggml_vocab_test__ + +__ggml_vocab_test__ + +__ggml_vocab_test__ + +__ggml_vocab_test__ + +__ggml_vocab_test__ + +__ggml_vocab_test__ + + +__ggml_vocab_test__ + + + +__ggml_vocab_test__ + + + + +__ggml_vocab_test__ + + +__ggml_vocab_test__ +Hello world +__ggml_vocab_test__ + Hello world +__ggml_vocab_test__ +Hello World +__ggml_vocab_test__ + Hello World +__ggml_vocab_test__ + Hello World! +__ggml_vocab_test__ +Hello, world! +__ggml_vocab_test__ + Hello, world! +__ggml_vocab_test__ + this is 🦙.cpp +__ggml_vocab_test__ +w048 7tuijk dsdfhu +__ggml_vocab_test__ +нещо на Български +__ggml_vocab_test__ +កាន់តែពិសេសអាចខលចេញ +__ggml_vocab_test__ +🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token) +__ggml_vocab_test__ +Hello +__ggml_vocab_test__ + Hello +__ggml_vocab_test__ + Hello +__ggml_vocab_test__ + Hello +__ggml_vocab_test__ + Hello +__ggml_vocab_test__ + Hello + Hello +__ggml_vocab_test__ + ( +__ggml_vocab_test__ + + = +__ggml_vocab_test__ +' era +__ggml_vocab_test__ +Hello, y'all! How are you 😁 ?我想在apple工作1314151天～ +__ggml_vocab_test__ +!!!!!! +__ggml_vocab_test__ +3 +__ggml_vocab_test__ +33 +__ggml_vocab_test__ +333 +__ggml_vocab_test__ +3333 +__ggml_vocab_test__ +33333 +__ggml_vocab_test__ +333333 +__ggml_vocab_test__ +3333333 +__ggml_vocab_test__ +33333333 +__ggml_vocab_test__ +333333333 +__ggml_vocab_test__ +Cửa Việt +__ggml_vocab_test__ + discards +__ggml_vocab_test__ + + + + + + + + + + + +🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL +__ggml_vocab_test__ diff --git a/models/ggml-vocab-roberta-bpe.gguf.out b/models/ggml-vocab-roberta-bpe.gguf.out new file mode 100644 index 000000000..f181ac3dc --- /dev/null +++ b/models/ggml-vocab-roberta-bpe.gguf.out @@ -0,0 +1,46 @@ + 2550 204 18430 377 + 597 2768 298 8564 + + 1437 + 1437 1437 + 1437 1437 1437 + 50117 + 50118 + 50140 + 50140 50118 + 50117 50118 + 31414 232 + 20920 232 + 31414 623 + 20920 623 + 20920 623 328 + 31414 6 232 328 + 20920 6 232 328 + 42 16 8103 18164 27 4 49317 + 605 40976 262 10109 18474 385 29 36807 6455 + 36765 25482 22063 23171 34251 18697 10809 26161 18697 3602 22063 27969 40966 25417 15264 26161 24269 36709 41171 35328 + 1376 17772 7471 1376 17772 19002 1376 17772 9085 1376 4333 13859 1376 17772 9357 1376 4333 9264 1376 17772 25448 1376 17772 18400 1376 17772 4333 1376 4333 10172 1376 17772 4333 1376 17772 7258 1376 17772 19002 1376 17772 5782 1376 17772 10172 1376 17772 3726 1376 17772 5782 1376 4333 10172 1376 17772 23171 + 6569 15113 7471 36 21113 43 17841 19002 17 8384 6569 14285 4958 12605 36 34654 2841 4203 354 10146 26511 1070 43 36174 5782 36 8338 21554 14 34 63 308 19233 43 + 31414 + 20920 + 1437 20920 + 1437 1437 20920 + 1437 1437 1437 20920 + 1437 1437 1437 20920 50118 1437 1437 1437 20920 + 36 + 50118 5457 + 108 3567 + 31414 6 1423 108 1250 328 1336 32 47 17841 10172 17487 47876 3602 48617 15264 46537 11423 27326 48494 8210 49233 1558 1570 27761 49429 43251 10809 17772 + 32376 12846 + 246 + 3103 + 25631 + 46152 + 3103 25631 + 46152 3103 + 46152 25631 + 46152 46152 + 46152 3103 25631 + 347 1376 2023 12410 102 16376 1376 2023 6382 90 + 9553 5954 + 50118 1437 50140 1437 50140 50118 1437 50117 1437 50117 50117 1437 50117 50118 1437 1437 50118 1437 1437 1437 50118 1437 1437 1437 1437 50118 1437 1437 1437 1437 1437 50118 6569 15113 7471 36 21113 43 17841 19002 17 8384 6569 14285 4958 12605 36 34654 2841 4203 354 10146 26511 1070 43 36174 5782 8103 18164 27 6569 18164 27 155 2357 30242 155 25631 30242 3103 30242 25631 30242 46152 30242 3103 25631 155 4 246 155 7586 246 155 734 246 25974 17772 7471 1376 17772 19002 1376 17772 9085 1376 4333 13859 1376 17772 9357 1376 4333 9264 1376 17772 25448 1376 17772 18400 1376 17772 4333 1376 4333 10172 1376 17772 4333 1376 17772 7258 1376 17772 19002 1376 17772 5782 18636 10172 17487 47876 3602 48617 15264 46537 11423 27326 48494 8210 49233 1558 1570 27761 49429 43251 10809 17772 36738 48332 47463 18697 10809 25482 22063 23171 34251 18697 10809 26161 18697 3602 22063 27969 40966 25417 15264 26161 24269 36709 41171 35328 128 49690 108 49972 49519 12905 48149 48149 43796 32376 12846 27282 28749 38 348 57 128 41042 37 18 89 6 128 4629 47 686 116 128 448 45 686 38 581 146 24 6 128 495 47 101 103 6845 116 166 108 30660 10 108 462 574 diff --git a/pocs/CMakeLists.txt b/pocs/CMakeLists.txt index 03e1d2c04..d49d14dee 100644 --- a/pocs/CMakeLists.txt +++ b/pocs/CMakeLists.txt @@ -8,5 +8,7 @@ include_directories(${CMAKE_CURRENT_SOURCE_DIR}) if (EMSCRIPTEN) else() - add_subdirectory(vdot) + if (NOT GGML_BACKEND_DL) + add_subdirectory(vdot) + endif() endif() diff --git a/pocs/vdot/CMakeLists.txt b/pocs/vdot/CMakeLists.txt index d5405ad29..6235aec1f 100644 --- a/pocs/vdot/CMakeLists.txt +++ b/pocs/vdot/CMakeLists.txt @@ -1,9 +1,9 @@ set(TARGET llama-vdot) add_executable(${TARGET} vdot.cpp) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) set(TARGET llama-q8dot) add_executable(${TARGET} q8dot.cpp) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) +target_compile_features(${TARGET} PRIVATE cxx_std_17) diff --git a/pocs/vdot/vdot.cpp b/pocs/vdot/vdot.cpp index e9af8a363..2dca62848 100644 --- a/pocs/vdot/vdot.cpp +++ b/pocs/vdot/vdot.cpp @@ -237,7 +237,6 @@ int main(int argc, char** argv) { int n4 = useQ4_1 ? kVecSize / QK4_1 : kVecSize / QK4_0; n4 = 64*((n4 + 63)/64); int n8 = kVecSize / QK8_0; n8 = 64*((n8 + 63)/64); - const auto * funcs = ggml_get_type_traits(useQ4_1 ? GGML_TYPE_Q4_1 : GGML_TYPE_Q4_0); const auto * funcs_cpu = ggml_get_type_traits_cpu(useQ4_1 ? GGML_TYPE_Q4_1 : GGML_TYPE_Q4_0); std::vector q40; @@ -263,9 +262,9 @@ int main(int argc, char** argv) { // Note, we do not include this in the timing as in practical application // we already have the quantized model weights. if (useQ4_1) { - funcs->from_float(x1.data(), q41.data(), kVecSize); + funcs_cpu->from_float(x1.data(), q41.data(), kVecSize); } else { - funcs->from_float(x1.data(), q40.data(), kVecSize); + funcs_cpu->from_float(x1.data(), q40.data(), kVecSize); } // Now measure time the dot product needs using the "scalar" version above @@ -284,7 +283,7 @@ int main(int argc, char** argv) { dot_q4_q8(kVecSize, &result, q40.data(), q8.data()); } else { - const auto * vdot = ggml_get_type_traits(funcs_cpu->vec_dot_type); + const auto * vdot = ggml_get_type_traits_cpu(funcs_cpu->vec_dot_type); vdot->from_float(y1.data(), q8.data(), kVecSize); if (useQ4_1) funcs_cpu->vec_dot(kVecSize, &result, 0, q41.data(), 0, q8.data(), 0, 1); else funcs_cpu->vec_dot(kVecSize, &result, 0, q40.data(), 0, q8.data(), 0, 1); diff --git a/scripts/compare-commits.sh b/scripts/compare-commits.sh index 8b9b1ad39..e40d1cc6d 100755 --- a/scripts/compare-commits.sh +++ b/scripts/compare-commits.sh @@ -16,15 +16,23 @@ bench_args="${@:3}" rm -f llama-bench.sqlite > /dev/null # to test a backend, call the script with the corresponding environment variable (e.g. GGML_CUDA=1 ./scripts/compare-commits.sh ...) +if [ -n "$GGML_CUDA" ]; then + cmake_opts="-DGGML_CUDA=ON" +fi + +dir="build-bench" + +function run { + rm -fr ${dir} > /dev/null + cmake -B ${dir} -S . $cmake_opts > /dev/null + cmake --build ${dir} -t llama-bench > /dev/null + ${dir}/bin/llama-bench -o sql -oe md $bench_args | sqlite3 llama-bench.sqlite +} git checkout $1 > /dev/null -make clean > /dev/null -make -j$(nproc) $make_opts llama-bench > /dev/null -./llama-bench -o sql -oe md $bench_args | sqlite3 llama-bench.sqlite +run git checkout $2 > /dev/null -make clean > /dev/null -make -j$(nproc) $make_opts llama-bench > /dev/null -./llama-bench -o sql -oe md $bench_args | sqlite3 llama-bench.sqlite +run ./scripts/compare-llama-bench.py -b $1 -c $2 diff --git a/scripts/compare-llama-bench.py b/scripts/compare-llama-bench.py index 4ac6b5fc0..239c458d8 100755 --- a/scripts/compare-llama-bench.py +++ b/scripts/compare-llama-bench.py @@ -19,22 +19,22 @@ logger = logging.getLogger("compare-llama-bench") # Properties by which to differentiate results per commit: KEY_PROPERTIES = [ - "cpu_info", "gpu_info", "n_gpu_layers", "cuda", "vulkan", "kompute", "metal", "sycl", "rpc", "gpu_blas", - "blas", "model_filename", "model_type", "n_batch", "n_ubatch", "embeddings", "n_threads", - "type_k", "type_v", "use_mmap", "no_kv_offload", "split_mode", "main_gpu", "tensor_split", "flash_attn", "n_prompt", "n_gen" + "cpu_info", "gpu_info", "backends", "n_gpu_layers", "model_filename", "model_type", "n_batch", "n_ubatch", + "embeddings", "cpu_mask", "cpu_strict", "poll", "n_threads", "type_k", "type_v", "use_mmap", "no_kv_offload", + "split_mode", "main_gpu", "tensor_split", "flash_attn", "n_prompt", "n_gen" ] # Properties that are boolean and are converted to Yes/No for the table: -BOOL_PROPERTIES = ["cuda", "vulkan", "kompute", "metal", "sycl", "gpu_blas", "blas", "embeddings", "use_mmap", "no_kv_offload", "flash_attn"] +BOOL_PROPERTIES = ["embeddings", "cpu_strict", "use_mmap", "no_kv_offload", "flash_attn"] # Header names for the table: PRETTY_NAMES = { - "cuda": "CUDA", "vulkan": "Vulkan", "kompute": "Kompute", "metal": "Metal", "sycl": "SYCL", "rpc": "RPC", - "gpu_blas": "GPU BLAS", "blas": "BLAS", "cpu_info": "CPU", "gpu_info": "GPU", "model_filename": "File", "model_type": "Model", - "model_size": "Model Size [GiB]", "model_n_params": "Num. of Par.", "n_batch": "Batch size", "n_ubatch": "Microbatch size", - "n_threads": "Threads", "type_k": "K type", "type_v": "V type", "n_gpu_layers": "GPU layers", "split_mode": "Split mode", - "main_gpu": "Main GPU", "no_kv_offload": "NKVO", "flash_attn": "FlashAttention", "tensor_split": "Tensor split", - "use_mmap": "Use mmap", "embeddings": "Embeddings", + "cpu_info": "CPU", "gpu_info": "GPU", "backends": "Backends", "n_gpu_layers": "GPU layers", + "model_filename": "File", "model_type": "Model", "model_size": "Model size [GiB]", + "model_n_params": "Num. of par.", "n_batch": "Batch size", "n_ubatch": "Microbatch size", + "embeddings": "Embeddings", "cpu_mask": "CPU mask", "cpu_strict": "CPU strict", "poll": "Poll", + "n_threads": "Threads", "type_k": "K type", "type_v": "V type", "split_mode": "Split mode", "main_gpu": "Main GPU", + "no_kv_offload": "NKVO", "flash_attn": "FlashAttention", "tensor_split": "Tensor split", "use_mmap": "Use mmap", } DEFAULT_SHOW = ["model_type"] # Always show these properties by default. @@ -126,6 +126,8 @@ connection = sqlite3.connect(input_file) cursor = connection.cursor() builds = cursor.execute("SELECT DISTINCT build_commit FROM test;").fetchall() +commit_short_len = len(builds[0][0]) + try: repo = git.Repo(".", search_parent_directories=True) except git.InvalidGitRepositoryError: @@ -138,11 +140,11 @@ def find_parent_in_data(commit: git.Commit): seen_hexsha8 = set() while heap: depth, current_commit = heapq.heappop(heap) - current_hexsha8 = commit.hexsha[:8] + current_hexsha8 = commit.hexsha[:commit_short_len] if (current_hexsha8,) in builds: return current_hexsha8 for parent in commit.parents: - parent_hexsha8 = parent.hexsha[:8] + parent_hexsha8 = parent.hexsha[:commit_short_len] if parent_hexsha8 not in seen_hexsha8: seen_hexsha8.add(parent_hexsha8) heapq.heappush(heap, (depth + 1, parent)) @@ -156,9 +158,9 @@ def get_all_parent_hexsha8s(commit: git.Commit): while unvisited: current_commit = unvisited.pop(0) - visited.append(current_commit.hexsha[:8]) + visited.append(current_commit.hexsha[:commit_short_len]) for parent in current_commit.parents: - if parent.hexsha[:8] not in visited: + if parent.hexsha[:commit_short_len] not in visited: unvisited.append(parent) return visited @@ -169,10 +171,10 @@ def get_commit_name(hexsha8): if repo is None: return hexsha8 for h in repo.heads: - if h.commit.hexsha[:8] == hexsha8: + if h.commit.hexsha[:commit_short_len] == hexsha8: return h.name for t in repo.tags: - if t.commit.hexsha[:8] == hexsha8: + if t.commit.hexsha[:commit_short_len] == hexsha8: return t.name return hexsha8 @@ -183,13 +185,13 @@ def get_commit_hexsha8(name): return None for h in repo.heads: if h.name == name: - return h.commit.hexsha[:8] + return h.commit.hexsha[:commit_short_len] for t in repo.tags: if t.name == name: - return t.commit.hexsha[:8] + return t.commit.hexsha[:commit_short_len] for c in repo.iter_commits("--all"): - if c.hexsha[:8] == name[:8]: - return c.hexsha[:8] + if c.hexsha[:commit_short_len] == name[:commit_short_len]: + return c.hexsha[:commit_short_len] return None @@ -303,14 +305,11 @@ else: show = [] # Show CPU and/or GPU by default even if the hardware for all results is the same: - if "gpu_blas" not in properties_different and "n_gpu_layers" not in properties_different: - gpu_blas = bool(rows_full[0][KEY_PROPERTIES.index("gpu_blas")]) + if "n_gpu_layers" not in properties_different: ngl = int(rows_full[0][KEY_PROPERTIES.index("n_gpu_layers")]) - if not gpu_blas or ngl != 99 and "cpu_info" not in properties_different: + if ngl != 99 and "cpu_info" not in properties_different: show.append("cpu_info") - if gpu_blas and "gpu_info" not in properties_different: - show.append("gpu_info") show += properties_different diff --git a/scripts/hf.sh b/scripts/hf.sh deleted file mode 100755 index 85c2c4d9a..000000000 --- a/scripts/hf.sh +++ /dev/null @@ -1,112 +0,0 @@ -#!/bin/bash -# -# Shortcut for downloading HF models -# -# Usage: -# ./llama-cli -m $(./scripts/hf.sh https://huggingface.co/TheBloke/Mixtral-8x7B-v0.1-GGUF/resolve/main/mixtral-8x7b-v0.1.Q4_K_M.gguf) -# ./llama-cli -m $(./scripts/hf.sh --url https://huggingface.co/TheBloke/Mixtral-8x7B-v0.1-GGUF/blob/main/mixtral-8x7b-v0.1.Q4_K_M.gguf) -# ./llama-cli -m $(./scripts/hf.sh --repo TheBloke/Mixtral-8x7B-v0.1-GGUF --file mixtral-8x7b-v0.1.Q4_K_M.gguf) -# - -# all logs go to stderr -function log { - echo "$@" 1>&2 -} - -function usage { - log "Usage: $0 [[--url] ] [--repo ] [--file ] [--outdir [-h|--help]" - exit 1 -} - -# check for curl or wget -function has_cmd { - if ! [ -x "$(command -v $1)" ]; then - return 1 - fi -} - -if has_cmd wget; then - cmd="wget -q --show-progress -c -O %s/%s %s" -elif has_cmd curl; then - cmd="curl -C - -f --output-dir %s -o %s -L %s" -else - log "[E] curl or wget not found" - exit 1 -fi - -url="" -repo="" -file="" -outdir="." - -# parse args -while [[ $# -gt 0 ]]; do - case "$1" in - --url) - url="$2" - shift 2 - ;; - --repo) - repo="$2" - shift 2 - ;; - --file) - file="$2" - shift 2 - ;; - --outdir) - outdir="$2" - shift 2 - ;; - -h|--help) - usage - ;; - *) - url="$1" - shift - ;; - esac -done - -if [ -n "$repo" ] && [ -n "$file" ]; then - url="https://huggingface.co/$repo/resolve/main/$file" -fi - -if [ -z "$url" ]; then - log "[E] missing --url" - usage -fi - -# check if the URL is a HuggingFace model, and if so, try to download it -is_url=false - -if [[ ${#url} -gt 22 ]]; then - if [[ ${url:0:22} == "https://huggingface.co" ]]; then - is_url=true - fi -fi - -if [ "$is_url" = false ]; then - log "[E] invalid URL, must start with https://huggingface.co" - exit 0 -fi - -# replace "blob/main" with "resolve/main" -url=${url/blob\/main/resolve\/main} - -basename=$(basename $url) - -log "[+] attempting to download $basename" - -if [ -n "$cmd" ]; then - cmd=$(printf "$cmd" "$outdir" "$basename" "$url") - log "[+] $cmd" - if $cmd; then - echo $outdir/$basename - exit 0 - fi -fi - -log "[-] failed to download" - -exit 1 diff --git a/scripts/pod-llama.sh b/scripts/pod-llama.sh deleted file mode 100644 index 6e56e1ed0..000000000 --- a/scripts/pod-llama.sh +++ /dev/null @@ -1,212 +0,0 @@ -#!/bin/bash -# -# Use this script only on fresh pods (runpod.io)! -# Otherwise, it can break your environment! -# - -if [ -z "$1" ]; then - echo "Usage: $0 " - echo " 0: no models" - echo " 1: tinyllama-1b" - echo " 2: codellama-7b" - echo " 3: codellama-13b" - echo " 4: codellama-34b" - echo " 5: codellama-7b-instruct" - echo " 6: codellama-13b-instruct" - echo " 7: codellama-34b-instruct" - - exit 1 -fi - -set -x - -# setup deps -apt-get update -apt-get install -y git-lfs cmake cmake-curses-gui vim ruby -git-lfs install - -if [ ! -d "/workspace" ]; then - ln -sfn $(pwd) /workspace -fi - -# download data -cd /workspace - -# this is useful to git clone repos without doubling the disk size due to .git -git clone https://github.com/iboB/git-lfs-download -ln -sfn /workspace/git-lfs-download/git-lfs-download /usr/local/bin/git-lfs-download - -# llama.cpp -cd /workspace -git clone https://github.com/ggerganov/llama.cpp - -cd llama.cpp - -GGML_CUDA=1 make -j - -ln -sfn /workspace/TinyLlama-1.1B-Chat-v0.3 ./models/tinyllama-1b -ln -sfn /workspace/CodeLlama-7b-hf ./models/codellama-7b -ln -sfn /workspace/CodeLlama-13b-hf ./models/codellama-13b -ln -sfn /workspace/CodeLlama-34b-hf ./models/codellama-34b -ln -sfn /workspace/CodeLlama-7b-Instruct-hf ./models/codellama-7b-instruct -ln -sfn /workspace/CodeLlama-13b-Instruct-hf ./models/codellama-13b-instruct -ln -sfn /workspace/CodeLlama-34b-Instruct-hf ./models/codellama-34b-instruct - -pip install -r requirements.txt - -# cmake -cd /workspace/llama.cpp - -mkdir build-cublas -cd build-cublas - -cmake -DGGML_CUDA=1 ../ -make -j - -if [ "$1" -eq "0" ]; then - exit 0 -fi - -# more models -if [ "$1" -eq "1" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/PY007/TinyLlama-1.1B-Chat-v0.3 - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/tinyllama-1b --outfile ./models/tinyllama-1b/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/tinyllama-1b/ggml-model-f16.gguf ./models/tinyllama-1b/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/tinyllama-1b/ggml-model-f16.gguf ./models/tinyllama-1b/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/tinyllama-1b/ggml-model-f16.gguf ./models/tinyllama-1b/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "2" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-7b-hf --without *safetensors* - rm -v ./CodeLlama-7b-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-7b --outfile ./models/codellama-7b/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-7b/ggml-model-f16.gguf ./models/codellama-7b/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-7b/ggml-model-f16.gguf ./models/codellama-7b/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-7b/ggml-model-f16.gguf ./models/codellama-7b/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "3" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-13b-hf --without *safetensors* - rm -v ./CodeLlama-13b-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-13b --outfile ./models/codellama-13b/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-13b/ggml-model-f16.gguf ./models/codellama-13b/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-13b/ggml-model-f16.gguf ./models/codellama-13b/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-13b/ggml-model-f16.gguf ./models/codellama-13b/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "4" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-34b-hf --without *safetensors* - rm -v ./CodeLlama-34b-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-34b --outfile ./models/codellama-34b/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-34b/ggml-model-f16.gguf ./models/codellama-34b/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-34b/ggml-model-f16.gguf ./models/codellama-34b/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-34b/ggml-model-f16.gguf ./models/codellama-34b/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "5" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-7b-Instruct-hf --without *safetensors* - rm -v ./CodeLlama-7b-Instruct-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-7b-instruct --outfile ./models/codellama-7b-instruct/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-7b-instruct/ggml-model-f16.gguf ./models/codellama-7b-instruct/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-7b-instruct/ggml-model-f16.gguf ./models/codellama-7b-instruct/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-7b-instruct/ggml-model-f16.gguf ./models/codellama-7b-instruct/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "6" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-13b-Instruct-hf --without *safetensors* - rm -v ./CodeLlama-13b-Instruct-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-13b-instruct --outfile ./models/codellama-13b-instruct/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-13b-instruct/ggml-model-f16.gguf ./models/codellama-13b-instruct/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-13b-instruct/ggml-model-f16.gguf ./models/codellama-13b-instruct/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-13b-instruct/ggml-model-f16.gguf ./models/codellama-13b-instruct/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "7" ]; then - cd /workspace - - git-lfs-download https://huggingface.co/codellama/CodeLlama-34b-Instruct-hf --without *safetensors* - rm -v ./CodeLlama-34b-Instruct-hf/*safetensors* - - cd /workspace/llama.cpp - - python3 examples/convert_legacy_llama.py ./models/codellama-34b-instruct --outfile ./models/codellama-34b-instruct/ggml-model-f16.gguf --outtype f16 - - ./llama-quantize ./models/codellama-34b-instruct/ggml-model-f16.gguf ./models/codellama-34b-instruct/ggml-model-q4_0.gguf q4_0 - ./llama-quantize ./models/codellama-34b-instruct/ggml-model-f16.gguf ./models/codellama-34b-instruct/ggml-model-q4_k.gguf q4_k - ./llama-quantize ./models/codellama-34b-instruct/ggml-model-f16.gguf ./models/codellama-34b-instruct/ggml-model-q8_0.gguf q8_0 -fi - -if [ "$1" -eq "1" ]; then - # perf + perplexity - cd /workspace/llama.cpp/build-cublas - - make -j && ../scripts/run-all-perf.sh tinyllama-1b "f16" "-ngl 99 -t 1 -p 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,32,64,128,256,512,1024,2048 -n 128" - - ../scripts/get-wikitext-2.sh - unzip wikitext-2-raw-v1.zip - - make -j && ./bin/llama-perplexity -m ../models/tinyllama-1b/ggml-model-f16.gguf -f ./wikitext-2-raw/wiki.test.raw -ngl 100 --chunks 32 - - # batched - cd /workspace/llama.cpp - - GGML_CUDA=1 make -j && ./llama-batched ./models/tinyllama-1b/ggml-model-f16.gguf "Hello, my name is" 8 128 999 - - # batched-bench - cd /workspace/llama.cpp - - GGML_CUDA=1 make -j && ./llama-batched-bench ./models/tinyllama-1b/ggml-model-f16.gguf 4608 1 99 0 512 128 1,2,3,4,5,6,7,8,16,32 - - # parallel - cd /workspace/llama.cpp - - GGML_CUDA=1 make -j && ./llama-parallel -m ./models/tinyllama-1b/ggml-model-f16.gguf -t 1 -ngl 100 -c 4096 -b 512 -s 1 -np 8 -ns 128 -n 100 -cb - -fi - -# speculative -#if [ "$1" -eq "7" ]; then -# cd /workspace/llama.cpp -# -# GGML_CUDA=1 make -j && ./llama-speculative -m ./models/codellama-34b-instruct/ggml-model-f16.gguf -md ./models/codellama-7b-instruct/ggml-model-q4_0.gguf -p "# Dijkstra's shortest path algorithm in Python (4 spaces indentation) + complexity analysis:\n\n" -e -ngl 999 -ngld 999 -t 4 -n 512 -c 4096 -s 21 --draft 16 -np 1 --temp 0.0 -#fi - -# more benches -#GGML_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-7b/ggml-model-q4_k.gguf 4096 1 99 1 512,3200 128,128,800 1 -#GGML_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-13b/ggml-model-q4_k.gguf 4096 1 99 1 512,3200 128,128,800 1 diff --git a/scripts/run-with-preset.py b/scripts/run-with-preset.py deleted file mode 100755 index 8f0bf8ca8..000000000 --- a/scripts/run-with-preset.py +++ /dev/null @@ -1,146 +0,0 @@ -#!/usr/bin/env python3 - -import logging -import argparse -import os -import subprocess -import sys - -import yaml - -logger = logging.getLogger("run-with-preset") - -CLI_ARGS_LLAMA_CLI_PERPLEXITY = [ - "batch-size", "cfg-negative-prompt", "cfg-scale", "chunks", "color", "ctx-size", "escape", - "export", "file", "frequency-penalty", "grammar", "grammar-file", "hellaswag", - "hellaswag-tasks", "ignore-eos", "in-prefix", "in-prefix-bos", "in-suffix", - "interactive", "interactive-first", "keep", "logdir", "logit-bias", "lora", "lora-base", - "low-vram", "main-gpu", "mirostat", "mirostat-ent", "mirostat-lr", "mlock", - "model", "multiline-input", "n-gpu-layers", "n-predict", "no-mmap", "no-mul-mat-q", - "np-penalize-nl", "numa", "ppl-output-type", "ppl-stride", "presence-penalty", "prompt", - "prompt-cache", "prompt-cache-all", "prompt-cache-ro", "repeat-last-n", - "repeat-penalty", "reverse-prompt", "rope-freq-base", "rope-freq-scale", "rope-scale", "seed", - "simple-io", "tensor-split", "threads", "temp", "top-k", "top-p", "typical", - "verbose-prompt" -] - -CLI_ARGS_LLAMA_BENCH = [ - "batch-size", "low-vram", "model", "mul-mat-q", "n-gen", "n-gpu-layers", - "n-prompt", "output", "repetitions", "tensor-split", "threads", "verbose" -] - -CLI_ARGS_LLAMA_SERVER = [ - "alias", "batch-size", "ctx-size", "embedding", "host", "lora", "lora-base", - "low-vram", "main-gpu", "mlock", "model", "n-gpu-layers", "n-probs", "no-mmap", "no-mul-mat-q", - "numa", "path", "port", "rope-freq-base", "timeout", "rope-freq-scale", "tensor-split", - "threads", "verbose" -] - -description = """Run llama.cpp binaries with presets from YAML file(s). -To specify which binary should be run, specify the "binary" property (llama-cli, llama-perplexity, llama-bench, and llama-server are supported). -To get a preset file template, run a llama.cpp binary with the "--logdir" CLI argument. - -Formatting considerations: -- The YAML property names are the same as the CLI argument names of the corresponding binary. -- Properties must use the long name of their corresponding llama.cpp CLI arguments. -- Like the llama.cpp binaries the property names do not differentiate between hyphens and underscores. -- Flags must be defined as ": true" to be effective. -- To define the logit_bias property, the expected format is ": " in the "logit_bias" namespace. -- To define multiple "reverse_prompt" properties simultaneously the expected format is a list of strings. -- To define a tensor split, pass a list of floats. -""" -usage = "run-with-preset.py [-h] [yaml_files ...] [-- ...]" -epilog = (" -- specify additional CLI ars to be passed to the binary (override all preset files). " - "Unknown args will be ignored.") - -parser = argparse.ArgumentParser( - description=description, usage=usage, epilog=epilog, formatter_class=argparse.RawTextHelpFormatter) -parser.add_argument("-bin", "--binary", help="The binary to run.") -parser.add_argument("yaml_files", nargs="*", - help="Arbitrary number of YAML files from which to read preset values. " - "If two files specify the same values the later one will be used.") -parser.add_argument("--verbose", action="store_true", help="increase output verbosity") - -known_args, unknown_args = parser.parse_known_args() - -if not known_args.yaml_files and not unknown_args: - parser.print_help() - sys.exit(0) - -logging.basicConfig(level=logging.DEBUG if known_args.verbose else logging.INFO) - -props = dict() - -for yaml_file in known_args.yaml_files: - with open(yaml_file, "r") as f: - props.update(yaml.load(f, yaml.SafeLoader)) - -props = {prop.replace("_", "-"): val for prop, val in props.items()} - -binary = props.pop("binary", "llama-cli") -if known_args.binary: - binary = known_args.binary - -if os.path.exists(f"./{binary}"): - binary = f"./{binary}" - -if binary.lower().endswith("llama-cli") or binary.lower().endswith("llama-perplexity"): - cli_args = CLI_ARGS_LLAMA_CLI_PERPLEXITY -elif binary.lower().endswith("llama-bench"): - cli_args = CLI_ARGS_LLAMA_BENCH -elif binary.lower().endswith("llama-server"): - cli_args = CLI_ARGS_LLAMA_SERVER -else: - logger.error(f"Unknown binary: {binary}") - sys.exit(1) - -command_list = [binary] - -for cli_arg in cli_args: - value = props.pop(cli_arg, None) - - if not value or value == -1: - continue - - if cli_arg == "logit-bias": - for token, bias in value.items(): - command_list.append("--logit-bias") - command_list.append(f"{token}{bias:+}") - continue - - if cli_arg == "reverse-prompt" and not isinstance(value, str): - for rp in value: - command_list.append("--reverse-prompt") - command_list.append(str(rp)) - continue - - command_list.append(f"--{cli_arg}") - - if cli_arg == "tensor-split": - command_list.append(",".join([str(v) for v in value])) - continue - - value = str(value) - - if value != "True": - command_list.append(str(value)) - -num_unused = len(props) -if num_unused > 10: - logger.info(f"The preset file contained a total of {num_unused} unused properties.") -elif num_unused > 0: - logger.info("The preset file contained the following unused properties:") - for prop, value in props.items(): - logger.info(f" {prop}: {value}") - -command_list += unknown_args - -sp = subprocess.Popen(command_list) - -while sp.returncode is None: - try: - sp.wait() - except KeyboardInterrupt: - pass - -sys.exit(sp.returncode) diff --git a/scripts/server-llm.sh b/scripts/server-llm.sh deleted file mode 100644 index 802592a3e..000000000 --- a/scripts/server-llm.sh +++ /dev/null @@ -1,418 +0,0 @@ -#!/bin/bash -# -# Helper script for deploying llama.cpp server with a single Bash command -# -# - Works on Linux and macOS -# - Supports: CPU, CUDA, Metal -# - Can run all GGUF models from HuggingFace -# - Can serve requests in parallel -# - Always builds latest llama.cpp from GitHub -# -# Limitations -# -# - Chat templates are poorly supported (base models recommended) -# - Might be unstable! -# -# Usage: -# ./server-llm.sh [--port] [--repo] [--wtype] [--backend] [--gpu-id] [--n-parallel] [--n-kv] [--verbose] [-non-interactive] -# -# --port: port number, default is 8888 -# --repo: path to a repo containing GGUF model files -# --wtype: weights type (f16, q8_0, q4_0, q4_1), default is user-input -# --backend: cpu, cuda, metal, depends on the OS -# --gpu-id: gpu id, default is 0 -# --n-parallel: number of parallel requests, default is 8 -# --n-kv: KV cache size, default is 4096 -# --verbose: verbose output -# --non-interactive: run without asking a permission to run -# -# Example: -# -# bash -c "$(curl -s https://ggml.ai/server-llm.sh)" -# - -set -e - -# required utils: curl, git, make -if ! command -v curl &> /dev/null; then - printf "[-] curl not found\n" - exit 1 -fi -if ! command -v git &> /dev/null; then - printf "[-] git not found\n" - exit 1 -fi -if ! command -v make &> /dev/null; then - printf "[-] make not found\n" - exit 1 -fi - -# parse arguments -is_interactive=1 -port=8888 -repo="" -wtype="" -backend="cpu" - -# if macOS, use metal backend by default -if [[ "$OSTYPE" == "darwin"* ]]; then - backend="metal" -elif command -v nvcc &> /dev/null; then - backend="cuda" -fi - -gpu_id=0 -n_parallel=8 -n_kv=4096 -verbose=0 - -function print_usage { - printf "Usage:\n" - printf " ./server-llm.sh [--port] [--repo] [--wtype] [--backend] [--gpu-id] [--n-parallel] [--n-kv] [--verbose] [-non-interactive]\n\n" - printf " --port: port number, default is 8888\n" - printf " --repo: path to a repo containing GGUF model files\n" - printf " --wtype: weights type (f16, q8_0, q4_0, q4_1), default is user-input\n" - printf " --backend: cpu, cuda, metal, depends on the OS\n" - printf " --gpu-id: gpu id, default is 0\n" - printf " --n-parallel: number of parallel requests, default is 8\n" - printf " --n-kv: KV cache size, default is 4096\n" - printf " --verbose: verbose output\n\n" - printf " --non-interactive: run without asking a permission to run\n" - printf "Example:\n\n" - printf ' bash -c "$(curl -s https://ggml.ai/server-llm.sh)"\n\n' -} - -while [[ $# -gt 0 ]]; do - key="$1" - case $key in - --non-interactive) - is_interactive=0 - shift - ;; - --port) - port="$2" - shift - shift - ;; - --repo) - repo="$2" - shift - shift - ;; - --wtype) - wtype="$2" - shift - shift - ;; - --backend) - backend="$2" - shift - shift - ;; - --gpu-id) - gpu_id="$2" - shift - shift - ;; - --n-parallel) - n_parallel="$2" - shift - shift - ;; - --n-kv) - n_kv="$2" - shift - shift - ;; - --verbose) - verbose=1 - shift - ;; - --help) - print_usage - exit 0 - ;; - *) - echo "Unknown argument: $key" - print_usage - exit 1 - ;; - esac -done - -# available weights types -wtypes=("F16" "Q8_0" "Q4_0" "Q4_1" "Q5_0" "Q5_1" "Q6_K" "Q5_K_M" "Q5_K_S" "Q4_K_M" "Q4_K_S" "Q3_K_L" "Q3_K_M" "Q3_K_S" "Q2_K") - -wfiles=() -for wt in "${wtypes[@]}"; do - wfiles+=("") -done - -# map wtype input to index -if [[ ! -z "$wtype" ]]; then - iw=-1 - is=0 - for wt in "${wtypes[@]}"; do - # uppercase - uwt=$(echo "$wt" | tr '[:lower:]' '[:upper:]') - if [[ "$uwt" == "$wtype" ]]; then - iw=$is - break - fi - is=$((is+1)) - done - - if [[ $iw -eq -1 ]]; then - printf "[-] Invalid weight type: %s\n" "$wtype" - exit 1 - fi - - wtype="$iw" -fi - -# sample repos -repos=( - "https://huggingface.co/TheBloke/Llama-2-7B-GGUF" - "https://huggingface.co/TheBloke/Llama-2-13B-GGUF" - "https://huggingface.co/TheBloke/Llama-2-70B-GGUF" - "https://huggingface.co/TheBloke/CodeLlama-7B-GGUF" - "https://huggingface.co/TheBloke/CodeLlama-13B-GGUF" - "https://huggingface.co/TheBloke/CodeLlama-34B-GGUF" - "https://huggingface.co/TheBloke/Mistral-7B-v0.1-GGUF" - "https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF" - "https://huggingface.co/TheBloke/OpenHermes-2-Mistral-7B-GGUF" - "https://huggingface.co/TheBloke/CausalLM-7B-GGUF" -) -if [ $is_interactive -eq 1 ]; then - printf "\n" - printf "[I] This is a helper script for deploying llama.cpp's server on this machine.\n\n" - printf " Based on the options that follow, the script might download a model file\n" - printf " from the internet, which can be a few GBs in size. The script will also\n" - printf " build the latest llama.cpp source code from GitHub, which can be unstable.\n" - printf "\n" - printf " Upon success, an HTTP server will be started and it will serve the selected\n" - printf " model using llama.cpp for demonstration purposes.\n" - printf "\n" - printf " Please note:\n" - printf "\n" - printf " - All new data will be stored in the current folder\n" - printf " - The server will be listening on all network interfaces\n" - printf " - The server will run with default settings which are not always optimal\n" - printf " - Do not judge the quality of a model based on the results from this script\n" - printf " - Do not use this script to benchmark llama.cpp\n" - printf " - Do not use this script in production\n" - printf " - This script is only for demonstration purposes\n" - printf "\n" - printf " If you don't know what you are doing, please press Ctrl-C to abort now\n" - printf "\n" - printf " Press Enter to continue ...\n\n" - - read -fi - -if [[ -z "$repo" ]]; then - printf "[+] No repo provided from the command line\n" - printf " Please select a number from the list below or enter an URL:\n\n" - - is=0 - for r in "${repos[@]}"; do - printf " %2d) %s\n" $is "$r" - is=$((is+1)) - done - - # ask for repo until index of sample repo is provided or an URL - while [[ -z "$repo" ]]; do - printf "\n Or choose one from: https://huggingface.co/models?sort=trending&search=gguf\n\n" - read -p "[+] Select repo: " repo - - # check if the input is a number - if [[ "$repo" =~ ^[0-9]+$ ]]; then - if [[ "$repo" -ge 0 && "$repo" -lt ${#repos[@]} ]]; then - repo="${repos[$repo]}" - else - printf "[-] Invalid repo index: %s\n" "$repo" - repo="" - fi - elif [[ "$repo" =~ ^https?:// ]]; then - repo="$repo" - else - printf "[-] Invalid repo URL: %s\n" "$repo" - repo="" - fi - done -fi - -# remove suffix -repo=$(echo "$repo" | sed -E 's/\/tree\/main$//g') - -printf "[+] Checking for GGUF model files in %s\n" "$repo" - -# find GGUF files in the source -# TODO: better logic -model_tree="${repo%/}/tree/main" -model_files=$(curl -s "$model_tree" | grep -i "\\.gguf" | sed -E 's/.*(.*)<\/span><\/a>/\1/g') - -# list all files in the provided git repo -printf "[+] Model files:\n\n" -for file in $model_files; do - # determine iw by grepping the filename with wtypes - iw=-1 - is=0 - for wt in "${wtypes[@]}"; do - # uppercase - ufile=$(echo "$file" | tr '[:lower:]' '[:upper:]') - if [[ "$ufile" =~ "$wt" ]]; then - iw=$is - break - fi - is=$((is+1)) - done - - if [[ $iw -eq -1 ]]; then - continue - fi - - wfiles[$iw]="$file" - - have=" " - if [[ -f "$file" ]]; then - have="*" - fi - - printf " %2d) %s %s\n" $iw "$have" "$file" -done - -wfile="${wfiles[$wtype]}" - -# ask for weights type until provided and available -while [[ -z "$wfile" ]]; do - printf "\n" - read -p "[+] Select weight type: " wtype - wfile="${wfiles[$wtype]}" - - if [[ -z "$wfile" ]]; then - printf "[-] Invalid weight type: %s\n" "$wtype" - wtype="" - fi -done - -printf "[+] Selected weight type: %s (%s)\n" "$wtype" "$wfile" - -url="${repo%/}/resolve/main/$wfile" - -# check file if the model has been downloaded before -chk="$wfile.chk" - -# check if we should download the file -# - if $wfile does not exist -# - if $wfile exists but $chk does not exist -# - if $wfile exists and $chk exists but $wfile is newer than $chk -# TODO: better logic using git lfs info - -do_download=0 - -if [[ ! -f "$wfile" ]]; then - do_download=1 -elif [[ ! -f "$chk" ]]; then - do_download=1 -elif [[ "$wfile" -nt "$chk" ]]; then - do_download=1 -fi - -if [[ $do_download -eq 1 ]]; then - printf "[+] Downloading weights from %s\n" "$url" - - # download the weights file - curl -o "$wfile" -# -L "$url" - - # create a check file if successful - if [[ $? -eq 0 ]]; then - printf "[+] Creating check file %s\n" "$chk" - touch "$chk" - fi -else - printf "[+] Using cached weights %s\n" "$wfile" -fi - -# get latest llama.cpp and build - -printf "[+] Downloading latest llama.cpp\n" - -llama_cpp_dir="__llama_cpp_port_${port}__" - -if [[ -d "$llama_cpp_dir" && ! -f "$llama_cpp_dir/__ggml_script__" ]]; then - # if the dir exists and there isn't a file "__ggml_script__" in it, abort - printf "[-] Directory %s already exists\n" "$llama_cpp_dir" - printf "[-] Please remove it and try again\n" - exit 1 -elif [[ -d "$llama_cpp_dir" ]]; then - printf "[+] Directory %s already exists\n" "$llama_cpp_dir" - printf "[+] Using cached llama.cpp\n" - - cd "$llama_cpp_dir" - git reset --hard - git fetch - git checkout origin/master - - cd .. -else - printf "[+] Cloning llama.cpp\n" - - git clone https://github.com/ggerganov/llama.cpp "$llama_cpp_dir" -fi - -# mark that that the directory is made by this script -touch "$llama_cpp_dir/__ggml_script__" - -if [[ $verbose -eq 1 ]]; then - set -x -fi - -# build -cd "$llama_cpp_dir" - -make clean - -log="--silent" -if [[ $verbose -eq 1 ]]; then - log="" -fi - -if [[ "$backend" == "cuda" ]]; then - printf "[+] Building with CUDA backend\n" - GGML_CUDA=1 make -j llama-server $log -elif [[ "$backend" == "cpu" ]]; then - printf "[+] Building with CPU backend\n" - make -j llama-server $log -elif [[ "$backend" == "metal" ]]; then - printf "[+] Building with Metal backend\n" - make -j llama-server $log -else - printf "[-] Unknown backend: %s\n" "$backend" - exit 1 -fi - -# run the server - -printf "[+] Running server\n" - -args="" -if [[ "$backend" == "cuda" ]]; then - export CUDA_VISIBLE_DEVICES=$gpu_id - args="-ngl 999" -elif [[ "$backend" == "cpu" ]]; then - args="-ngl 0" -elif [[ "$backend" == "metal" ]]; then - args="-ngl 999" -else - printf "[-] Unknown backend: %s\n" "$backend" - exit 1 -fi - -if [[ $verbose -eq 1 ]]; then - args="$args --verbose" -fi - -./llama-server -m "../$wfile" --host 0.0.0.0 --port "$port" -c $n_kv -np "$n_parallel" $args - -exit 0 diff --git a/scripts/sync-ggml-am.sh b/scripts/sync-ggml-am.sh index 06a04745b..ec4f4b0a2 100755 --- a/scripts/sync-ggml-am.sh +++ b/scripts/sync-ggml-am.sh @@ -73,17 +73,22 @@ while read c; do src/ggml*.h \ src/ggml*.c \ src/ggml*.cpp \ - src/ggml*.m \ - src/ggml*.metal \ - src/ggml*.cu \ - src/ggml-amx/* \ + src/gguf*.cpp \ + src/ggml-blas/* \ src/ggml-cann/* \ + src/ggml-cpu/* \ src/ggml-cuda/* \ + src/ggml-hip/* \ + src/ggml-kompute/* \ + src/ggml-metal/* \ + src/ggml-musa/* \ + src/ggml-opencl/* \ + src/ggml-rpc/* \ src/ggml-sycl/* \ - src/vulkan-shaders/* \ + src/ggml-vulkan/* \ include/ggml*.h \ + include/gguf*.h \ tests/test-opt.cpp \ - tests/test-grad0.cpp \ tests/test-quantize-fns.cpp \ tests/test-quantize-perf.cpp \ tests/test-backend-ops.cpp \ @@ -121,21 +126,24 @@ if [ -f $SRC_LLAMA/ggml-src.patch ]; then # src/ggml*.c -> ggml/src/ggml*.c # src/ggml*.cpp -> ggml/src/ggml*.cpp # src/ggml*.h -> ggml/src/ggml*.h - # src/ggml*.cu -> ggml/src/ggml*.cu - # src/ggml*.m -> ggml/src/ggml*.m - # src/ggml-amx/* -> ggml/src/ggml-amx/ - # src/ggml-cann/* -> ggml/src/ggml-cann/ - # src/ggml-cuda/* -> ggml/src/ggml-cuda/ - # src/ggml-sycl/* -> ggml/src/ggml-sycl/ - # src/vulkan-shaders/* -> ggml/src/vulkan-shaders/ + # src/gguf*.cpp -> ggml/src/gguf*.cpp + # src/ggml-blas/* -> ggml/src/ggml-blas/* + # src/ggml-cann/* -> ggml/src/ggml-cann/* + # src/ggml-cpu/* -> ggml/src/ggml-cpu/* + # src/ggml-cuda/* -> ggml/src/ggml-cuda/* + # src/ggml-hip/* -> ggml/src/ggml-hip/* + # src/ggml-kompute/* -> ggml/src/ggml-kompute/* + # src/ggml-metal/* -> ggml/src/ggml-metal/* + # src/ggml-musa/* -> ggml/src/ggml-musa/* + # src/ggml-opencl/* -> ggml/src/ggml-opencl/* + # src/ggml-rpc/* -> ggml/src/ggml-rpc/* + # src/ggml-sycl/* -> ggml/src/ggml-sycl/* + # src/ggml-vulkan/* -> ggml/src/ggml-vulkan/* # # include/ggml*.h -> ggml/include/ggml*.h + # include/gguf*.h -> ggml/include/gguf*.h # - # tests/test-opt.cpp -> tests/test-opt.cpp - # tests/test-grad0.cpp -> tests/test-grad0.cpp - # tests/test-quantize-fns.cpp -> tests/test-quantize-fns.cpp - # tests/test-quantize-perf.cpp -> tests/test-quantize-perf.cpp - # tests/test-backend-ops.cpp -> tests/test-backend-ops.cpp + # tests/test*.cpp -> tests/ # # LICENSE -> LICENSE # scripts/gen-authors.sh -> scripts/gen-authors.sh @@ -144,21 +152,24 @@ if [ -f $SRC_LLAMA/ggml-src.patch ]; then -e 's/([[:space:]]|[ab]\/)CMakeLists.txt/\1ggml\/CMakeLists.txt/g' \ -e 's/([[:space:]]|[ab]\/)src\/CMakeLists.txt/\1ggml\/src\/CMakeLists.txt/g' \ -e 's/([[:space:]]|[ab]\/)cmake\/FindSIMD.cmake/\1ggml\/cmake\/FindSIMD.cmake/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.c/\1ggml\/src\/ggml\1.c/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.cpp/\1ggml\/src\/ggml\1.cpp/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.h/\1ggml\/src\/ggml\1.h/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.cu/\1ggml\/src\/ggml\1.cu/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.m/\1ggml\/src\/ggml\1.m/g' \ - -e 's/([[:space:]]|[ab]\/)src\/ggml-amx\//\1ggml\/src\/ggml-amx\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.c/\1ggml\/src\/ggml\2.c/g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.cpp/\1ggml\/src\/ggml\2.cpp/g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml(.*)\.h/\1ggml\/src\/ggml\2.h/g' \ + -e 's/([[:space:]]|[ab]\/)src\/gguf(.*)\.cpp/\1ggml\/src\/gguf\2.cpp/g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-blas\//\1ggml\/src\/ggml-blas\//g' \ -e 's/([[:space:]]|[ab]\/)src\/ggml-cann\//\1ggml\/src\/ggml-cann\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-cpu\//\1ggml\/src\/ggml-cpu\//g' \ -e 's/([[:space:]]|[ab]\/)src\/ggml-cuda\//\1ggml\/src\/ggml-cuda\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-hip\//\1ggml\/src\/ggml-hip\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-kompute\//\1ggml\/src\/ggml-kompute\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-metal\//\1ggml\/src\/ggml-metal\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-opencl\//\1ggml\/src\/ggml-opencl\//g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-rpc\//\1ggml\/src\/ggml-rpc\//g' \ -e 's/([[:space:]]|[ab]\/)src\/ggml-sycl\//\1ggml\/src\/ggml-sycl\//g' \ - -e 's/([[:space:]]|[ab]\/)src\/vulkan-shaders\//\1ggml\/src\/vulkan-shaders\//g' \ - -e 's/([[:space:]]|[ab]\/)include\/ggml(.*)\.h/\1ggml\/include\/ggml\1.h/g' \ - -e 's/([[:space:]]|[ab]\/)examples\/common\.h/\1examples\/common.h/g' \ - -e 's/([[:space:]]|[ab]\/)examples\/common\.cpp/\1examples\/common.cpp/g' \ - -e 's/([[:space:]]|[ab]\/)examples\/common-ggml\.h/\1examples\/common-ggml.h/g' \ - -e 's/([[:space:]]|[ab]\/)examples\/common-ggml\.cpp/\1examples\/common-ggml.cpp/g' \ + -e 's/([[:space:]]|[ab]\/)src\/ggml-vulkan\//\1ggml\/src\/ggml-vulkan\//g' \ + -e 's/([[:space:]]|[ab]\/)include\/ggml(.*)\.h/\1ggml\/include\/ggml\2.h/g' \ + -e 's/([[:space:]]|[ab]\/)include\/gguf(.*)\.h/\1ggml\/include\/gguf\2.h/g' \ + -e 's/([[:space:]]|[ab]\/)tests\/(.*)\.cpp/\1tests\/\2.cpp/g' \ -e 's/([[:space:]]|[ab]\/)LICENSE/\1LICENSE/g' \ -e 's/([[:space:]]|[ab]\/)scripts\/gen-authors\.sh/\1scripts\/gen-authors.sh/g' \ > ggml-src.patch.tmp diff --git a/scripts/sync-ggml.last b/scripts/sync-ggml.last index e82984f49..cfba59d32 100644 --- a/scripts/sync-ggml.last +++ b/scripts/sync-ggml.last @@ -1 +1 @@ -89952d649e0c5cabbb9ff8c4906f5a843a789fb2 +d92321c0d151fe73a47d89738c7c3091ac904297 diff --git a/scripts/sync-ggml.sh b/scripts/sync-ggml.sh index f29554c82..e83d415c0 100755 --- a/scripts/sync-ggml.sh +++ b/scripts/sync-ggml.sh @@ -4,21 +4,27 @@ cp -rpv ../ggml/CMakeLists.txt ./ggml/CMakeLists.txt cp -rpv ../ggml/src/CMakeLists.txt ./ggml/src/CMakeLists.txt cp -rpv ../ggml/cmake/FindSIMD.cmake ./ggml/cmake/FindSIMD.cmake -cp -rpv ../ggml/src/ggml*.c ./ggml/src/ -cp -rpv ../ggml/src/ggml*.cpp ./ggml/src/ -cp -rpv ../ggml/src/ggml*.h ./ggml/src/ -cp -rpv ../ggml/src/ggml*.cu ./ggml/src/ -cp -rpv ../ggml/src/ggml*.m ./ggml/src/ -cp -rpv ../ggml/src/ggml-amx/* ./ggml/src/ggml-amx/ -cp -rpv ../ggml/src/ggml-cann/* ./ggml/src/ggml-cann/ -cp -rpv ../ggml/src/ggml-cuda/* ./ggml/src/ggml-cuda/ -cp -rpv ../ggml/src/ggml-sycl/* ./ggml/src/ggml-sycl/ -cp -rpv ../ggml/src/vulkan-shaders/* ./ggml/src/vulkan-shaders/ +cp -rpv ../ggml/src/ggml*.c ./ggml/src/ +cp -rpv ../ggml/src/ggml*.cpp ./ggml/src/ +cp -rpv ../ggml/src/ggml*.h ./ggml/src/ +cp -rpv ../ggml/src/gguf*.cpp ./ggml/src/ +cp -rpv ../ggml/src/ggml-blas/* ./ggml/src/ggml-blas/ +cp -rpv ../ggml/src/ggml-cann/* ./ggml/src/ggml-cann/ +cp -rpv ../ggml/src/ggml-cpu/* ./ggml/src/ggml-cpu/ +cp -rpv ../ggml/src/ggml-cuda/* ./ggml/src/ggml-cuda/ +cp -rpv ../ggml/src/ggml-hip/* ./ggml/src/ggml-hip/ +cp -rpv ../ggml/src/ggml-kompute/* ./ggml/src/ggml-kompute/ +cp -rpv ../ggml/src/ggml-metal/* ./ggml/src/ggml-metal/ +cp -rpv ../ggml/src/ggml-musa/* ./ggml/src/ggml-musa/ +cp -rpv ../ggml/src/ggml-opencl/* ./ggml/src/ggml-opencl/ +cp -rpv ../ggml/src/ggml-rpc/* ./ggml/src/ggml-rpc/ +cp -rpv ../ggml/src/ggml-sycl/* ./ggml/src/ggml-sycl/ +cp -rpv ../ggml/src/ggml-vulkan/* ./ggml/src/ggml-vulkan/ cp -rpv ../ggml/include/ggml*.h ./ggml/include/ +cp -rpv ../ggml/include/gguf*.h ./ggml/include/ cp -rpv ../ggml/tests/test-opt.cpp ./tests/test-opt.cpp -cp -rpv ../ggml/tests/test-grad0.cpp ./tests/test-grad0.cpp cp -rpv ../ggml/tests/test-quantize-fns.cpp ./tests/test-quantize-fns.cpp cp -rpv ../ggml/tests/test-quantize-perf.cpp ./tests/test-quantize-perf.cpp cp -rpv ../ggml/tests/test-backend-ops.cpp ./tests/test-backend-ops.cpp diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index 46a6ad562..aeb75bf3e 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -1,9 +1,4 @@ -# TODO: should not use this -if (WIN32) - if (BUILD_SHARED_LIBS) - set(CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS ON) - endif() -endif() +llama_add_compile_flags() # # libraries @@ -14,20 +9,33 @@ endif() add_library(llama ../include/llama.h llama.cpp - llama-vocab.cpp + llama-adapter.cpp + llama-arch.cpp + llama-batch.cpp + llama-chat.cpp + llama-context.cpp llama-grammar.cpp + llama-hparams.cpp + llama-impl.cpp + llama-kv-cache.cpp + llama-mmap.cpp + llama-model-loader.cpp + llama-model.cpp + llama-quant.cpp llama-sampling.cpp + llama-vocab.cpp unicode.h unicode.cpp unicode-data.cpp ) target_include_directories(llama PUBLIC . ../include) -target_compile_features (llama PUBLIC cxx_std_11) # don't bump +target_compile_features (llama PUBLIC cxx_std_17) # don't bump target_link_libraries(llama PUBLIC ggml) if (BUILD_SHARED_LIBS) set_target_properties(llama PROPERTIES POSITION_INDEPENDENT_CODE ON) - target_compile_definitions(llama PRIVATE LLAMA_SHARED LLAMA_BUILD) + target_compile_definitions(llama PRIVATE LLAMA_BUILD) + target_compile_definitions(llama PUBLIC LLAMA_SHARED) endif() diff --git a/src/llama-adapter.cpp b/src/llama-adapter.cpp new file mode 100644 index 000000000..8a0800463 --- /dev/null +++ b/src/llama-adapter.cpp @@ -0,0 +1,347 @@ +#include "llama-adapter.h" + +#include "llama-impl.h" +#include "llama-mmap.h" +#include "llama-model.h" + +#include +#include +#include +#include + +// vec + +struct ggml_tensor * llama_adapter_cvec::tensor_for(int il) const { + if (il < 0 || il < layer_start || il > layer_end || (size_t) il >= tensors.size()) { + return nullptr; + } + + return tensors[il]; +} + +struct ggml_tensor * llama_adapter_cvec::apply_to(struct ggml_context * ctx, struct ggml_tensor * cur, int il) const { + ggml_tensor * layer_dir = tensor_for(il); + if (layer_dir != nullptr) { + cur = ggml_add(ctx, cur, layer_dir); + } + + return cur; +} + +bool llama_adapter_cvec::init(const llama_model & model) { + const auto & hparams = model.hparams; + + GGML_ASSERT(tensors.empty()); + GGML_ASSERT(ctxs.empty()); + GGML_ASSERT(bufs.empty()); + + // create a context for each buffer type + std::map ctx_map; + auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * { + auto it = ctx_map.find(buft); + if (it == ctx_map.end()) { + struct ggml_init_params params = { + /*.mem_size =*/ hparams.n_layer*ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + + ggml_context * ctx = ggml_init(params); + if (!ctx) { + return nullptr; + } + + ctx_map[buft] = ctx; + ctxs.emplace_back(ctx); + + return ctx; + } + + return it->second; + }; + + // make tensors + tensors.reserve(hparams.n_layer); + tensors.push_back(nullptr); // there's never a tensor for layer 0 + for (size_t il = 1; il < hparams.n_layer; il++) { + ggml_backend_buffer_type_t buft = model.select_buft(il); + ggml_context * ctx = ctx_for_buft(buft); + if (!ctx) { + LLAMA_LOG_ERROR("%s: failed to allocate context for control vector\n", __func__); + return false; + } + ggml_tensor * tensor = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, hparams.n_embd); + tensors.push_back(tensor); + } + + // allocate tensors / buffers and zero + bufs.reserve(ctx_map.size()); + for (auto it : ctx_map) { + ggml_backend_buffer_type_t buft = it.first; + ggml_context * ctx = it.second; + ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft); + if (!buf) { + LLAMA_LOG_ERROR("%s: failed to allocate buffer for control vector\n", __func__); + return false; + } + ggml_backend_buffer_clear(buf, 0); + bufs.emplace_back(buf); + } + + return true; +} + +int32_t llama_adapter_cvec::apply( + const llama_model & model, + const float * data, + size_t len, + int32_t n_embd, + int32_t il_start, + int32_t il_end) { + const auto & hparams = model.hparams; + + if (data == nullptr) { + // disable the current control vector (but leave allocated for later) + layer_start = -1; + layer_end = -1; + return 0; + } + + if (n_embd != (int) hparams.n_embd) { + LLAMA_LOG_ERROR("%s: control vector n_embd does not match model\n", __func__); + return 1; + } + + if (tensors.empty()) { + if (!init(model)) { + return 1; + } + } + + layer_start = il_start; + layer_end = il_end; + + for (size_t il = 1; il < hparams.n_layer; il++) { + assert(tensors[il] != nullptr); + + const size_t off = n_embd * (il - 1); // buffer doesn't have data for layer 0, since it's never present + if (off + n_embd <= len) { + ggml_backend_tensor_set(tensors[il], data + off, 0, n_embd * ggml_element_size(tensors[il])); + } + } + + return 0; +} + +// lora + +llama_adapter_lora_weight * llama_adapter_lora::get_weight(struct ggml_tensor * w) { + const std::string name(w->name); + + const auto pos = ab_map.find(name); + if (pos != ab_map.end()) { + return &pos->second; + } + + return nullptr; +} + +static void llama_adapter_lora_init_impl(struct llama_model & model, const char * path_lora, struct llama_adapter_lora & adapter) { + LLAMA_LOG_INFO("%s: loading lora adapter from '%s' ...\n", __func__, path_lora); + + ggml_context * ctx_init; + struct gguf_init_params meta_gguf_params = { + /* .no_alloc = */ true, + /* .ctx = */ &ctx_init, + }; + + gguf_context_ptr ctx_gguf { gguf_init_from_file(path_lora, meta_gguf_params) }; + if (!ctx_gguf) { + throw std::runtime_error("failed to load lora adapter file from " + std::string(path_lora)); + } + + ggml_context_ptr ctx { ctx_init }; + + // check metadata + { + auto get_kv_str = [&](const std::string & key) -> std::string { + int id = gguf_find_key(ctx_gguf.get(), key.c_str()); + return id < 0 ? "" : std::string(gguf_get_val_str(ctx_gguf.get(), id)); + }; + auto get_kv_f32 = [&](const std::string & key) -> float { + int id = gguf_find_key(ctx_gguf.get(), key.c_str()); + return id < 0 ? 0.0f : gguf_get_val_f32(ctx_gguf.get(), id); + }; + LLM_KV llm_kv = LLM_KV(LLM_ARCH_UNKNOWN); + + auto general_type = get_kv_str(llm_kv(LLM_KV_GENERAL_TYPE)); + if (general_type != "adapter") { + throw std::runtime_error("expect general.type to be 'adapter', but got: " + general_type); + } + + auto general_arch_str = get_kv_str(llm_kv(LLM_KV_GENERAL_ARCHITECTURE)); + auto general_arch = llm_arch_from_string(general_arch_str); + if (general_arch != model.arch) { + throw std::runtime_error("model arch and LoRA arch mismatch"); + } + + auto adapter_type = get_kv_str(llm_kv(LLM_KV_ADAPTER_TYPE)); + if (adapter_type != "lora") { + throw std::runtime_error("expect adapter.type to be 'lora', but got: " + adapter_type); + } + + adapter.alpha = get_kv_f32(llm_kv(LLM_KV_ADAPTER_LORA_ALPHA)); + } + + int n_tensors = gguf_get_n_tensors(ctx_gguf.get()); + + // contexts for each buffer type + std::map ctx_map; + auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * { + auto it = ctx_map.find(buft); + if (it == ctx_map.end()) { + // add a new context + struct ggml_init_params params = { + /*.mem_size =*/ n_tensors*ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + ggml_context * buft_ctx = ggml_init(params); + if (!buft_ctx) { + return nullptr; + } + ctx_map[buft] = buft_ctx; + adapter.ctxs.emplace_back(buft_ctx); + return buft_ctx; + }; + return it->second; + }; + + // bundle lora_a and lora_b into pairs + std::map ab_map; + auto str_endswith = [](const std::string & str, const std::string & suffix) { + return str.size() >= suffix.size() && str.compare(str.size()-suffix.size(), suffix.size(), suffix) == 0; + }; + + for (ggml_tensor * cur = ggml_get_first_tensor(ctx.get()); cur; cur = ggml_get_next_tensor(ctx.get(), cur)) { + std::string name(cur->name); + if (str_endswith(name, ".lora_a")) { + replace_all(name, ".lora_a", ""); + if (ab_map.find(name) == ab_map.end()) { + ab_map[name] = llama_adapter_lora_weight(cur, nullptr); + } else { + ab_map[name].a = cur; + } + } else if (str_endswith(name, ".lora_b")) { + replace_all(name, ".lora_b", ""); + if (ab_map.find(name) == ab_map.end()) { + ab_map[name] = llama_adapter_lora_weight(nullptr, cur); + } else { + ab_map[name].b = cur; + } + } else if (str_endswith(name, "_norm.weight")) { + // TODO: add support for norm vector + // for now, we don't really care because most adapters still work fine without it + continue; + } else { + throw std::runtime_error("LoRA tensor '" + name + "' has unexpected suffix"); + } + } + + // add tensors + for (auto & it : ab_map) { + const std::string & name = it.first; + llama_adapter_lora_weight & w = it.second; + bool is_token_embd = str_endswith(name, "token_embd.weight"); + + if (!w.a || !w.b) { + throw std::runtime_error("LoRA tensor pair for '" + name + "' is missing one component"); + } + + // device buft and device ctx + const auto * model_tensor = model.get_tensor(name.c_str()); + if (!model_tensor) { + throw std::runtime_error("LoRA tensor '" + name + "' does not exist in base model (hint: maybe wrong base model?)"); + } + + struct ggml_context * dev_ctx = ctx_for_buft(ggml_backend_buffer_get_type(model_tensor->buffer)); + // validate tensor shape + if (is_token_embd) { + // expect B to be non-transposed, A and B are flipped; see llm_build_inp_embd() + if (model_tensor->ne[0] != w.b->ne[1] || model_tensor->ne[1] != w.a->ne[1]) { + throw std::runtime_error("tensor '" + name + "' has incorrect shape (hint: maybe wrong base model?)"); + } + } else { + if (model_tensor->ne[0] != w.a->ne[0] || model_tensor->ne[1] != w.b->ne[1]) { + throw std::runtime_error("tensor '" + name + "' has incorrect shape (hint: maybe wrong base model?)"); + } + if (w.a->ne[1] != w.b->ne[0]) { + throw std::runtime_error("lora_a tensor is not transposed (hint: adapter from \"finetune\" example is no longer supported)"); + } + } + + // save tensor to adapter + struct ggml_tensor * tensor_a = ggml_dup_tensor(dev_ctx, w.a); + struct ggml_tensor * tensor_b = ggml_dup_tensor(dev_ctx, w.b); + ggml_set_name(tensor_a, w.a->name); + ggml_set_name(tensor_b, w.b->name); + adapter.ab_map[name] = llama_adapter_lora_weight(tensor_a, tensor_b); + } + + // allocate tensors / buffers and zero + { + adapter.ctxs.reserve(ctx_map.size()); + adapter.bufs.reserve(ctx_map.size()); + for (auto & it : ctx_map) { + ggml_backend_buffer_type_t buft = it.first; + ggml_context * ctx_dev = it.second; + ggml_backend_buffer_ptr buf { ggml_backend_alloc_ctx_tensors_from_buft(ctx_dev, buft) }; + if (!buf) { + throw std::runtime_error("failed to allocate buffer for lora adapter\n"); + } + LLAMA_LOG_INFO("%s: %10s LoRA buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get())/1024.0/1024.0); + adapter.bufs.emplace_back(std::move(buf)); + } + } + + // set tensor data + { + llama_file gguf_file(path_lora, "rb"); + std::vector read_buf; + auto set_tensor = [&](struct ggml_tensor * orig, struct ggml_tensor * dev) { + size_t offs = gguf_get_data_offset(ctx_gguf.get()) + gguf_get_tensor_offset(ctx_gguf.get(), gguf_find_tensor(ctx_gguf.get(), orig->name)); + size_t size = ggml_nbytes(orig); + read_buf.resize(size); + gguf_file.seek(offs, SEEK_SET); + gguf_file.read_raw(read_buf.data(), size); + ggml_backend_tensor_set(dev, read_buf.data(), 0, size); + }; + for (auto & it : adapter.ab_map) { + auto orig = ab_map[it.first]; + auto dev = it.second; + set_tensor(orig.a, dev.a); + set_tensor(orig.b, dev.b); + } + } + + LLAMA_LOG_INFO("%s: loaded %zu tensors from lora file\n", __func__, adapter.ab_map.size()*2); +} + +struct llama_adapter_lora * llama_adapter_lora_init(struct llama_model * model, const char * path_lora) { + struct llama_adapter_lora * adapter = new llama_adapter_lora(); + + try { + llama_adapter_lora_init_impl(*model, path_lora, *adapter); + return adapter; + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: failed to apply lora adapter: %s\n", __func__, err.what()); + + delete adapter; + } + + return nullptr; +} + +void llama_adapter_lora_free(struct llama_adapter_lora * adapter) { + delete adapter; +} diff --git a/src/llama-adapter.h b/src/llama-adapter.h new file mode 100644 index 000000000..603fa08f6 --- /dev/null +++ b/src/llama-adapter.h @@ -0,0 +1,74 @@ +#pragma once + +#include "llama.h" + +#include "ggml-cpp.h" + +#include +#include +#include + +// TODO: pimpl + +// +// llama_adapter_cvec +// + +struct llama_adapter_cvec { + struct ggml_tensor * tensor_for(int il) const; + + struct ggml_tensor * apply_to(struct ggml_context * ctx, struct ggml_tensor * cur, int il) const; + + int32_t apply( + const llama_model & model, + const float * data, + size_t len, + int32_t n_embd, + int32_t il_start, + int32_t il_end); + +private: + bool init(const llama_model & model); + + int32_t layer_start = -1; + int32_t layer_end = -1; + + std::vector ctxs; + std::vector bufs; + + std::vector tensors; // per layer +}; + +// +// llama_adapter_lora +// + +struct llama_adapter_lora_weight { + struct ggml_tensor * a = nullptr; + struct ggml_tensor * b = nullptr; + + // get actual scale based on rank and alpha + float get_scale(float alpha, float adapter_scale) const { + const float rank = (float) b->ne[0]; + const float scale = alpha ? adapter_scale * alpha / rank : adapter_scale; + return scale; + } + + llama_adapter_lora_weight() = default; + llama_adapter_lora_weight(struct ggml_tensor * a, struct ggml_tensor * b) : a(a), b(b) {} +}; + +struct llama_adapter_lora { + // map tensor name to lora_a_b + std::unordered_map ab_map; + + std::vector ctxs; + std::vector bufs; + + float alpha; + + llama_adapter_lora() = default; + ~llama_adapter_lora() = default; + + llama_adapter_lora_weight * get_weight(struct ggml_tensor * w); +}; diff --git a/src/llama-arch.cpp b/src/llama-arch.cpp new file mode 100644 index 000000000..d7d277e72 --- /dev/null +++ b/src/llama-arch.cpp @@ -0,0 +1,1487 @@ +#include "llama-arch.h" + +#include "llama-impl.h" + +#include + +static const std::map LLM_ARCH_NAMES = { + { LLM_ARCH_LLAMA, "llama" }, + { LLM_ARCH_DECI, "deci" }, + { LLM_ARCH_FALCON, "falcon" }, + { LLM_ARCH_GROK, "grok" }, + { LLM_ARCH_GPT2, "gpt2" }, + { LLM_ARCH_GPTJ, "gptj" }, + { LLM_ARCH_GPTNEOX, "gptneox" }, + { LLM_ARCH_MPT, "mpt" }, + { LLM_ARCH_BAICHUAN, "baichuan" }, + { LLM_ARCH_STARCODER, "starcoder" }, + { LLM_ARCH_REFACT, "refact" }, + { LLM_ARCH_BERT, "bert" }, + { LLM_ARCH_NOMIC_BERT, "nomic-bert" }, + { LLM_ARCH_JINA_BERT_V2, "jina-bert-v2" }, + { LLM_ARCH_BLOOM, "bloom" }, + { LLM_ARCH_STABLELM, "stablelm" }, + { LLM_ARCH_QWEN, "qwen" }, + { LLM_ARCH_QWEN2, "qwen2" }, + { LLM_ARCH_QWEN2MOE, "qwen2moe" }, + { LLM_ARCH_QWEN2VL, "qwen2vl" }, + { LLM_ARCH_PHI2, "phi2" }, + { LLM_ARCH_PHI3, "phi3" }, + { LLM_ARCH_PHIMOE, "phimoe" }, + { LLM_ARCH_PLAMO, "plamo" }, + { LLM_ARCH_CODESHELL, "codeshell" }, + { LLM_ARCH_ORION, "orion" }, + { LLM_ARCH_INTERNLM2, "internlm2" }, + { LLM_ARCH_MINICPM, "minicpm" }, + { LLM_ARCH_MINICPM3, "minicpm3" }, + { LLM_ARCH_GEMMA, "gemma" }, + { LLM_ARCH_GEMMA2, "gemma2" }, + { LLM_ARCH_STARCODER2, "starcoder2" }, + { LLM_ARCH_MAMBA, "mamba" }, + { LLM_ARCH_XVERSE, "xverse" }, + { LLM_ARCH_COMMAND_R, "command-r" }, + { LLM_ARCH_COHERE2, "cohere2" }, + { LLM_ARCH_DBRX, "dbrx" }, + { LLM_ARCH_OLMO, "olmo" }, + { LLM_ARCH_OLMO2, "olmo2" }, + { LLM_ARCH_OLMOE, "olmoe" }, + { LLM_ARCH_OPENELM, "openelm" }, + { LLM_ARCH_ARCTIC, "arctic" }, + { LLM_ARCH_DEEPSEEK, "deepseek" }, + { LLM_ARCH_DEEPSEEK2, "deepseek2" }, + { LLM_ARCH_CHATGLM, "chatglm" }, + { LLM_ARCH_BITNET, "bitnet" }, + { LLM_ARCH_T5, "t5" }, + { LLM_ARCH_T5ENCODER, "t5encoder" }, + { LLM_ARCH_JAIS, "jais" }, + { LLM_ARCH_NEMOTRON, "nemotron" }, + { LLM_ARCH_EXAONE, "exaone" }, + { LLM_ARCH_RWKV6, "rwkv6" }, + { LLM_ARCH_RWKV6QWEN2, "rwkv6qwen2" }, + { LLM_ARCH_GRANITE, "granite" }, + { LLM_ARCH_GRANITE_MOE, "granitemoe" }, + { LLM_ARCH_CHAMELEON, "chameleon" }, + { LLM_ARCH_WAVTOKENIZER_DEC, "wavtokenizer-dec" }, + { LLM_ARCH_UNKNOWN, "(unknown)" }, +}; + +static const std::map LLM_KV_NAMES = { + { LLM_KV_GENERAL_TYPE, "general.type" }, + { LLM_KV_GENERAL_ARCHITECTURE, "general.architecture" }, + { LLM_KV_GENERAL_QUANTIZATION_VERSION, "general.quantization_version" }, + { LLM_KV_GENERAL_ALIGNMENT, "general.alignment" }, + { LLM_KV_GENERAL_NAME, "general.name" }, + { LLM_KV_GENERAL_AUTHOR, "general.author" }, + { LLM_KV_GENERAL_VERSION, "general.version" }, + { LLM_KV_GENERAL_URL, "general.url" }, + { LLM_KV_GENERAL_DESCRIPTION, "general.description" }, + { LLM_KV_GENERAL_LICENSE, "general.license" }, + { LLM_KV_GENERAL_SOURCE_URL, "general.source.url" }, + { LLM_KV_GENERAL_SOURCE_HF_REPO, "general.source.huggingface.repository" }, + + { LLM_KV_VOCAB_SIZE, "%s.vocab_size" }, + { LLM_KV_CONTEXT_LENGTH, "%s.context_length" }, + { LLM_KV_EMBEDDING_LENGTH, "%s.embedding_length" }, + { LLM_KV_FEATURES_LENGTH, "%s.features_length" }, + { LLM_KV_BLOCK_COUNT, "%s.block_count" }, + { LLM_KV_LEADING_DENSE_BLOCK_COUNT, "%s.leading_dense_block_count" }, + { LLM_KV_FEED_FORWARD_LENGTH, "%s.feed_forward_length" }, + { LLM_KV_EXPERT_FEED_FORWARD_LENGTH, "%s.expert_feed_forward_length" }, + { LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, "%s.expert_shared_feed_forward_length" }, + { LLM_KV_USE_PARALLEL_RESIDUAL, "%s.use_parallel_residual" }, + { LLM_KV_TENSOR_DATA_LAYOUT, "%s.tensor_data_layout" }, + { LLM_KV_EXPERT_COUNT, "%s.expert_count" }, + { LLM_KV_EXPERT_USED_COUNT, "%s.expert_used_count" }, + { LLM_KV_EXPERT_SHARED_COUNT, "%s.expert_shared_count" }, + { LLM_KV_EXPERT_WEIGHTS_SCALE, "%s.expert_weights_scale" }, + { LLM_KV_EXPERT_WEIGHTS_NORM, "%s.expert_weights_norm" }, + { LLM_KV_EXPERT_GATING_FUNC, "%s.expert_gating_func" }, + { LLM_KV_POOLING_TYPE, "%s.pooling_type" }, + { LLM_KV_LOGIT_SCALE, "%s.logit_scale" }, + { LLM_KV_DECODER_START_TOKEN_ID, "%s.decoder_start_token_id" }, + { LLM_KV_ATTN_LOGIT_SOFTCAPPING, "%s.attn_logit_softcapping" }, + { LLM_KV_FINAL_LOGIT_SOFTCAPPING, "%s.final_logit_softcapping" }, + { LLM_KV_SWIN_NORM, "%s.swin_norm" }, + { LLM_KV_RESCALE_EVERY_N_LAYERS, "%s.rescale_every_n_layers" }, + { LLM_KV_TIME_MIX_EXTRA_DIM, "%s.time_mix_extra_dim" }, + { LLM_KV_TIME_DECAY_EXTRA_DIM, "%s.time_decay_extra_dim" }, + { LLM_KV_RESIDUAL_SCALE, "%s.residual_scale" }, + { LLM_KV_EMBEDDING_SCALE, "%s.embedding_scale" }, + { LLM_KV_TOKEN_SHIFT_COUNT, "%s.token_shift_count" }, + + { LLM_KV_ATTENTION_HEAD_COUNT, "%s.attention.head_count" }, + { LLM_KV_ATTENTION_HEAD_COUNT_KV, "%s.attention.head_count_kv" }, + { LLM_KV_ATTENTION_MAX_ALIBI_BIAS, "%s.attention.max_alibi_bias" }, + { LLM_KV_ATTENTION_CLAMP_KQV, "%s.attention.clamp_kqv" }, + { LLM_KV_ATTENTION_KEY_LENGTH, "%s.attention.key_length" }, + { LLM_KV_ATTENTION_VALUE_LENGTH, "%s.attention.value_length" }, + { LLM_KV_ATTENTION_LAYERNORM_EPS, "%s.attention.layer_norm_epsilon" }, + { LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, "%s.attention.layer_norm_rms_epsilon" }, + { LLM_KV_ATTENTION_GROUPNORM_EPS, "%s.attention.group_norm_epsilon" }, + { LLM_KV_ATTENTION_GROUPNORM_GROUPS, "%s.attention.group_norm_groups" }, + { LLM_KV_ATTENTION_CAUSAL, "%s.attention.causal" }, + { LLM_KV_ATTENTION_Q_LORA_RANK, "%s.attention.q_lora_rank" }, + { LLM_KV_ATTENTION_KV_LORA_RANK, "%s.attention.kv_lora_rank" }, + { LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, "%s.attention.relative_buckets_count" }, + { LLM_KV_ATTENTION_SLIDING_WINDOW, "%s.attention.sliding_window" }, + { LLM_KV_ATTENTION_SCALE, "%s.attention.scale" }, + + { LLM_KV_ROPE_DIMENSION_COUNT, "%s.rope.dimension_count" }, + { LLM_KV_ROPE_DIMENSION_SECTIONS, "%s.rope.dimension_sections" }, + { LLM_KV_ROPE_FREQ_BASE, "%s.rope.freq_base" }, + { LLM_KV_ROPE_SCALE_LINEAR, "%s.rope.scale_linear" }, + { LLM_KV_ROPE_SCALING_TYPE, "%s.rope.scaling.type" }, + { LLM_KV_ROPE_SCALING_FACTOR, "%s.rope.scaling.factor" }, + { LLM_KV_ROPE_SCALING_ATTN_FACTOR, "%s.rope.scaling.attn_factor" }, + { LLM_KV_ROPE_SCALING_ORIG_CTX_LEN, "%s.rope.scaling.original_context_length" }, + { LLM_KV_ROPE_SCALING_FINETUNED, "%s.rope.scaling.finetuned" }, + { LLM_KV_ROPE_SCALING_YARN_LOG_MUL, "%s.rope.scaling.yarn_log_multiplier" }, + + { LLM_KV_SPLIT_NO, "split.no" }, + { LLM_KV_SPLIT_COUNT, "split.count" }, + { LLM_KV_SPLIT_TENSORS_COUNT, "split.tensors.count" }, + + { LLM_KV_SSM_CONV_KERNEL, "%s.ssm.conv_kernel" }, + { LLM_KV_SSM_INNER_SIZE, "%s.ssm.inner_size" }, + { LLM_KV_SSM_STATE_SIZE, "%s.ssm.state_size" }, + { LLM_KV_SSM_TIME_STEP_RANK, "%s.ssm.time_step_rank" }, + { LLM_KV_SSM_DT_B_C_RMS, "%s.ssm.dt_b_c_rms" }, + + { LLM_KV_WKV_HEAD_SIZE, "%s.wkv.head_size" }, + + { LLM_KV_POSNET_EMBEDDING_LENGTH, "%s.posnet.embedding_length" }, + { LLM_KV_POSNET_BLOCK_COUNT, "%s.posnet.block_count" }, + + { LLM_KV_CONVNEXT_EMBEDDING_LENGTH, "%s.convnext.embedding_length" }, + { LLM_KV_CONVNEXT_BLOCK_COUNT, "%s.convnext.block_count" }, + + { LLM_KV_TOKENIZER_MODEL, "tokenizer.ggml.model" }, + { LLM_KV_TOKENIZER_PRE, "tokenizer.ggml.pre" }, + { LLM_KV_TOKENIZER_LIST, "tokenizer.ggml.tokens" }, + { LLM_KV_TOKENIZER_TOKEN_TYPE, "tokenizer.ggml.token_type" }, + { LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, "tokenizer.ggml.token_type_count" }, + { LLM_KV_TOKENIZER_SCORES, "tokenizer.ggml.scores" }, + { LLM_KV_TOKENIZER_MERGES, "tokenizer.ggml.merges" }, + { LLM_KV_TOKENIZER_BOS_ID, "tokenizer.ggml.bos_token_id" }, + { LLM_KV_TOKENIZER_EOS_ID, "tokenizer.ggml.eos_token_id" }, + { LLM_KV_TOKENIZER_EOT_ID, "tokenizer.ggml.eot_token_id" }, + { LLM_KV_TOKENIZER_EOM_ID, "tokenizer.ggml.eom_token_id" }, + { LLM_KV_TOKENIZER_UNK_ID, "tokenizer.ggml.unknown_token_id" }, + { LLM_KV_TOKENIZER_SEP_ID, "tokenizer.ggml.seperator_token_id" }, + { LLM_KV_TOKENIZER_PAD_ID, "tokenizer.ggml.padding_token_id" }, + { LLM_KV_TOKENIZER_CLS_ID, "tokenizer.ggml.cls_token_id" }, + { LLM_KV_TOKENIZER_MASK_ID, "tokenizer.ggml.mask_token_id" }, + { LLM_KV_TOKENIZER_ADD_BOS, "tokenizer.ggml.add_bos_token" }, + { LLM_KV_TOKENIZER_ADD_EOS, "tokenizer.ggml.add_eos_token" }, + { LLM_KV_TOKENIZER_ADD_PREFIX, "tokenizer.ggml.add_space_prefix" }, + { LLM_KV_TOKENIZER_REMOVE_EXTRA_WS, "tokenizer.ggml.remove_extra_whitespaces" }, + { LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP, "tokenizer.ggml.precompiled_charsmap" }, + { LLM_KV_TOKENIZER_HF_JSON, "tokenizer.huggingface.json" }, + { LLM_KV_TOKENIZER_RWKV, "tokenizer.rwkv.world" }, + { LLM_KV_TOKENIZER_CHAT_TEMPLATE, "tokenizer.chat_template" }, + { LLM_KV_TOKENIZER_FIM_PRE_ID, "tokenizer.ggml.fim_pre_token_id" }, + { LLM_KV_TOKENIZER_FIM_SUF_ID, "tokenizer.ggml.fim_suf_token_id" }, + { LLM_KV_TOKENIZER_FIM_MID_ID, "tokenizer.ggml.fim_mid_token_id" }, + { LLM_KV_TOKENIZER_FIM_PAD_ID, "tokenizer.ggml.fim_pad_token_id" }, + { LLM_KV_TOKENIZER_FIM_REP_ID, "tokenizer.ggml.fim_rep_token_id" }, + { LLM_KV_TOKENIZER_FIM_SEP_ID, "tokenizer.ggml.fim_sep_token_id" }, + + { LLM_KV_ADAPTER_TYPE, "adapter.type" }, + { LLM_KV_ADAPTER_LORA_ALPHA, "adapter.lora.alpha" }, + + // deprecated + { LLM_KV_TOKENIZER_PREFIX_ID, "tokenizer.ggml.prefix_token_id" }, + { LLM_KV_TOKENIZER_SUFFIX_ID, "tokenizer.ggml.suffix_token_id" }, + { LLM_KV_TOKENIZER_MIDDLE_ID, "tokenizer.ggml.middle_token_id" }, +}; + +static const std::map> LLM_TENSOR_NAMES = { + { + LLM_ARCH_LLAMA, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_GATE_EXP, "blk.%d.ffn_gate.%d" }, + { LLM_TENSOR_FFN_DOWN_EXP, "blk.%d.ffn_down.%d" }, + { LLM_TENSOR_FFN_UP_EXP, "blk.%d.ffn_up.%d" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_DECI, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_GATE_EXP, "blk.%d.ffn_gate.%d" }, + { LLM_TENSOR_FFN_DOWN_EXP, "blk.%d.ffn_down.%d" }, + { LLM_TENSOR_FFN_UP_EXP, "blk.%d.ffn_up.%d" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_BAICHUAN, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_FALCON, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_NORM_2, "blk.%d.attn_norm_2" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_GROK, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE_EXP, "blk.%d.ffn_gate.%d" }, + { LLM_TENSOR_FFN_DOWN_EXP, "blk.%d.ffn_down.%d" }, + { LLM_TENSOR_FFN_UP_EXP, "blk.%d.ffn_up.%d" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + { LLM_TENSOR_LAYER_OUT_NORM, "blk.%d.layer_output_norm" }, + { LLM_TENSOR_ATTN_OUT_NORM, "blk.%d.attn_output_norm" }, + }, + }, + { + LLM_ARCH_GPT2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_POS_EMBD, "position_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_GPTJ, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + }, + }, + { + LLM_ARCH_GPTNEOX, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_MPT, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output"}, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_ACT, "blk.%d.ffn.act" }, + { LLM_TENSOR_POS_EMBD, "position_embd" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm"}, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm"}, + }, + }, + { + LLM_ARCH_STARCODER, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_POS_EMBD, "position_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_REFACT, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_BERT, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_TOKEN_TYPES, "token_types" }, + { LLM_TENSOR_POS_EMBD, "position_embd" }, + { LLM_TENSOR_ATTN_OUT_NORM, "blk.%d.attn_output_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_LAYER_OUT_NORM, "blk.%d.layer_output_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_CLS, "cls" }, + { LLM_TENSOR_CLS_OUT, "cls.output" }, + }, + }, + { + LLM_ARCH_NOMIC_BERT, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_TOKEN_TYPES, "token_types" }, + { LLM_TENSOR_ATTN_OUT_NORM, "blk.%d.attn_output_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_LAYER_OUT_NORM, "blk.%d.layer_output_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_JINA_BERT_V2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_TOKEN_TYPES, "token_types" }, + { LLM_TENSOR_ATTN_NORM_2, "blk.%d.attn_norm_2" }, + { LLM_TENSOR_ATTN_OUT_NORM, "blk.%d.attn_output_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_LAYER_OUT_NORM, "blk.%d.layer_output_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_CLS, "cls" }, + }, + }, + { + LLM_ARCH_BLOOM, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_STABLELM, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + }, + }, + { + LLM_ARCH_QWEN, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_QWEN2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_QWEN2VL, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_QWEN2MOE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + { LLM_TENSOR_FFN_GATE_INP_SHEXP, "blk.%d.ffn_gate_inp_shexp" }, + { LLM_TENSOR_FFN_GATE_SHEXP, "blk.%d.ffn_gate_shexp" }, + { LLM_TENSOR_FFN_DOWN_SHEXP, "blk.%d.ffn_down_shexp" }, + { LLM_TENSOR_FFN_UP_SHEXP, "blk.%d.ffn_up_shexp" }, + }, + }, + { + LLM_ARCH_PHI2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_PHI3, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FACTORS_LONG, "rope_factors_long" }, + { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_PHIMOE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FACTORS_LONG, "rope_factors_long" }, + { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_PLAMO, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_CODESHELL, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_ORION, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_INTERNLM2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_MINICPM, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ROPE_FACTORS_LONG, "rope_factors_long" }, + { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_GATE_EXP, "blk.%d.ffn_gate.%d" }, + { LLM_TENSOR_FFN_DOWN_EXP, "blk.%d.ffn_down.%d" }, + { LLM_TENSOR_FFN_UP_EXP, "blk.%d.ffn_up.%d" }, + }, + }, + { + LLM_ARCH_MINICPM3, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FACTORS_LONG, "rope_factors_long" }, + { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q_A_NORM, "blk.%d.attn_q_a_norm" }, + { LLM_TENSOR_ATTN_KV_A_NORM, "blk.%d.attn_kv_a_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_Q_A, "blk.%d.attn_q_a" }, + { LLM_TENSOR_ATTN_Q_B, "blk.%d.attn_q_b" }, + { LLM_TENSOR_ATTN_KV_A_MQA, "blk.%d.attn_kv_a_mqa" }, + { LLM_TENSOR_ATTN_KV_B, "blk.%d.attn_kv_b" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_GEMMA, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_GEMMA2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_POST_NORM, "blk.%d.post_attention_norm" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_POST_NORM, "blk.%d.post_ffw_norm" }, + }, + }, + { + LLM_ARCH_STARCODER2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_MAMBA, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_SSM_IN, "blk.%d.ssm_in" }, + { LLM_TENSOR_SSM_CONV1D, "blk.%d.ssm_conv1d" }, + { LLM_TENSOR_SSM_X, "blk.%d.ssm_x" }, + { LLM_TENSOR_SSM_DT, "blk.%d.ssm_dt" }, + { LLM_TENSOR_SSM_A, "blk.%d.ssm_a" }, + { LLM_TENSOR_SSM_D, "blk.%d.ssm_d" }, + { LLM_TENSOR_SSM_OUT, "blk.%d.ssm_out" }, + }, + }, + { + LLM_ARCH_XVERSE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_COMMAND_R, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + }, + }, + { + LLM_ARCH_COHERE2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_DBRX, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_OUT_NORM, "blk.%d.attn_output_norm" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_OLMO, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_OLMO2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_POST_NORM, "blk.%d.post_attention_norm" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + { LLM_TENSOR_FFN_POST_NORM, "blk.%d.post_ffw_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_OLMOE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_OPENELM, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_ARCTIC, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_NORM_EXPS, "blk.%d.ffn_norm_exps" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_DEEPSEEK, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + { LLM_TENSOR_FFN_GATE_INP_SHEXP, "blk.%d.ffn_gate_inp_shexp" }, + { LLM_TENSOR_FFN_GATE_SHEXP, "blk.%d.ffn_gate_shexp" }, + { LLM_TENSOR_FFN_DOWN_SHEXP, "blk.%d.ffn_down_shexp" }, + { LLM_TENSOR_FFN_UP_SHEXP, "blk.%d.ffn_up_shexp" }, + }, + }, + { + LLM_ARCH_DEEPSEEK2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q_A_NORM, "blk.%d.attn_q_a_norm" }, + { LLM_TENSOR_ATTN_KV_A_NORM, "blk.%d.attn_kv_a_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_Q_A, "blk.%d.attn_q_a" }, + { LLM_TENSOR_ATTN_Q_B, "blk.%d.attn_q_b" }, + { LLM_TENSOR_ATTN_KV_A_MQA, "blk.%d.attn_kv_a_mqa" }, + { LLM_TENSOR_ATTN_KV_B, "blk.%d.attn_kv_b" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + { LLM_TENSOR_FFN_GATE_INP_SHEXP, "blk.%d.ffn_gate_inp_shexp" }, + { LLM_TENSOR_FFN_GATE_SHEXP, "blk.%d.ffn_gate_shexp" }, + { LLM_TENSOR_FFN_DOWN_SHEXP, "blk.%d.ffn_down_shexp" }, + { LLM_TENSOR_FFN_UP_SHEXP, "blk.%d.ffn_up_shexp" }, + { LLM_TENSOR_FFN_EXP_PROBS_B, "blk.%d.exp_probs_b" }, + }, + }, + { + LLM_ARCH_CHATGLM, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_BITNET, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_SUB_NORM, "blk.%d.attn_sub_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_SUB_NORM, "blk.%d.ffn_sub_norm" }, + }, + }, + { + LLM_ARCH_T5, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_DEC_OUTPUT_NORM, "dec.output_norm" }, + { LLM_TENSOR_DEC_ATTN_NORM, "dec.blk.%d.attn_norm" }, + { LLM_TENSOR_DEC_ATTN_Q, "dec.blk.%d.attn_q" }, + { LLM_TENSOR_DEC_ATTN_K, "dec.blk.%d.attn_k" }, + { LLM_TENSOR_DEC_ATTN_V, "dec.blk.%d.attn_v" }, + { LLM_TENSOR_DEC_ATTN_OUT, "dec.blk.%d.attn_o" }, + { LLM_TENSOR_DEC_ATTN_REL_B, "dec.blk.%d.attn_rel_b" }, + { LLM_TENSOR_DEC_CROSS_ATTN_NORM, "dec.blk.%d.cross_attn_norm" }, + { LLM_TENSOR_DEC_CROSS_ATTN_Q, "dec.blk.%d.cross_attn_q" }, + { LLM_TENSOR_DEC_CROSS_ATTN_K, "dec.blk.%d.cross_attn_k" }, + { LLM_TENSOR_DEC_CROSS_ATTN_V, "dec.blk.%d.cross_attn_v" }, + { LLM_TENSOR_DEC_CROSS_ATTN_OUT, "dec.blk.%d.cross_attn_o" }, + { LLM_TENSOR_DEC_CROSS_ATTN_REL_B, "dec.blk.%d.cross_attn_rel_b" }, + { LLM_TENSOR_DEC_FFN_NORM, "dec.blk.%d.ffn_norm" }, + { LLM_TENSOR_DEC_FFN_GATE, "dec.blk.%d.ffn_gate" }, + { LLM_TENSOR_DEC_FFN_DOWN, "dec.blk.%d.ffn_down" }, + { LLM_TENSOR_DEC_FFN_UP, "dec.blk.%d.ffn_up" }, + { LLM_TENSOR_ENC_OUTPUT_NORM, "enc.output_norm" }, + { LLM_TENSOR_ENC_ATTN_NORM, "enc.blk.%d.attn_norm" }, + { LLM_TENSOR_ENC_ATTN_Q, "enc.blk.%d.attn_q" }, + { LLM_TENSOR_ENC_ATTN_K, "enc.blk.%d.attn_k" }, + { LLM_TENSOR_ENC_ATTN_V, "enc.blk.%d.attn_v" }, + { LLM_TENSOR_ENC_ATTN_OUT, "enc.blk.%d.attn_o" }, + { LLM_TENSOR_ENC_ATTN_REL_B, "enc.blk.%d.attn_rel_b" }, + { LLM_TENSOR_ENC_FFN_NORM, "enc.blk.%d.ffn_norm" }, + { LLM_TENSOR_ENC_FFN_GATE, "enc.blk.%d.ffn_gate" }, + { LLM_TENSOR_ENC_FFN_DOWN, "enc.blk.%d.ffn_down" }, + { LLM_TENSOR_ENC_FFN_UP, "enc.blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_T5ENCODER, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ENC_OUTPUT_NORM, "enc.output_norm" }, + { LLM_TENSOR_ENC_ATTN_NORM, "enc.blk.%d.attn_norm" }, + { LLM_TENSOR_ENC_ATTN_Q, "enc.blk.%d.attn_q" }, + { LLM_TENSOR_ENC_ATTN_K, "enc.blk.%d.attn_k" }, + { LLM_TENSOR_ENC_ATTN_V, "enc.blk.%d.attn_v" }, + { LLM_TENSOR_ENC_ATTN_OUT, "enc.blk.%d.attn_o" }, + { LLM_TENSOR_ENC_ATTN_REL_B, "enc.blk.%d.attn_rel_b" }, + { LLM_TENSOR_ENC_FFN_NORM, "enc.blk.%d.ffn_norm" }, + { LLM_TENSOR_ENC_FFN_GATE, "enc.blk.%d.ffn_gate" }, + { LLM_TENSOR_ENC_FFN_DOWN, "enc.blk.%d.ffn_down" }, + { LLM_TENSOR_ENC_FFN_UP, "enc.blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_JAIS, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_QKV, "blk.%d.attn_qkv" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + }, + }, + { + LLM_ARCH_NEMOTRON, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_EXAONE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ROPE_FREQS, "rope_freqs" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_ATTN_ROT_EMBD, "blk.%d.attn_rot_embd" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_RWKV6, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_NORM_2, "blk.%d.attn_norm_2" }, + { LLM_TENSOR_TIME_MIX_W1, "blk.%d.time_mix_w1" }, + { LLM_TENSOR_TIME_MIX_W2, "blk.%d.time_mix_w2" }, + { LLM_TENSOR_TIME_MIX_LERP_X, "blk.%d.time_mix_lerp_x" }, + { LLM_TENSOR_TIME_MIX_LERP_W, "blk.%d.time_mix_lerp_w" }, + { LLM_TENSOR_TIME_MIX_LERP_K, "blk.%d.time_mix_lerp_k" }, + { LLM_TENSOR_TIME_MIX_LERP_V, "blk.%d.time_mix_lerp_v" }, + { LLM_TENSOR_TIME_MIX_LERP_R, "blk.%d.time_mix_lerp_r" }, + { LLM_TENSOR_TIME_MIX_LERP_G, "blk.%d.time_mix_lerp_g" }, + { LLM_TENSOR_TIME_MIX_LERP_FUSED, "blk.%d.time_mix_lerp_fused" }, + { LLM_TENSOR_TIME_MIX_FIRST, "blk.%d.time_mix_first" }, + { LLM_TENSOR_TIME_MIX_DECAY, "blk.%d.time_mix_decay" }, + { LLM_TENSOR_TIME_MIX_DECAY_W1, "blk.%d.time_mix_decay_w1" }, + { LLM_TENSOR_TIME_MIX_DECAY_W2, "blk.%d.time_mix_decay_w2" }, + { LLM_TENSOR_TIME_MIX_KEY, "blk.%d.time_mix_key" }, + { LLM_TENSOR_TIME_MIX_VALUE, "blk.%d.time_mix_value" }, + { LLM_TENSOR_TIME_MIX_RECEPTANCE, "blk.%d.time_mix_receptance" }, + { LLM_TENSOR_TIME_MIX_GATE, "blk.%d.time_mix_gate" }, + { LLM_TENSOR_TIME_MIX_LN, "blk.%d.time_mix_ln" }, + { LLM_TENSOR_TIME_MIX_OUTPUT, "blk.%d.time_mix_output" }, + { LLM_TENSOR_CHANNEL_MIX_LERP_K, "blk.%d.channel_mix_lerp_k" }, + { LLM_TENSOR_CHANNEL_MIX_LERP_R, "blk.%d.channel_mix_lerp_r" }, + { LLM_TENSOR_CHANNEL_MIX_KEY, "blk.%d.channel_mix_key" }, + { LLM_TENSOR_CHANNEL_MIX_VALUE, "blk.%d.channel_mix_value" }, + { LLM_TENSOR_CHANNEL_MIX_RECEPTANCE, "blk.%d.channel_mix_receptance" }, + }, + }, + { + LLM_ARCH_RWKV6QWEN2, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_TIME_MIX_W1, "blk.%d.time_mix_w1" }, + { LLM_TENSOR_TIME_MIX_W2, "blk.%d.time_mix_w2" }, + { LLM_TENSOR_TIME_MIX_LERP_X, "blk.%d.time_mix_lerp_x" }, + { LLM_TENSOR_TIME_MIX_LERP_FUSED, "blk.%d.time_mix_lerp_fused" }, + { LLM_TENSOR_TIME_MIX_FIRST, "blk.%d.time_mix_first" }, + { LLM_TENSOR_TIME_MIX_DECAY, "blk.%d.time_mix_decay" }, + { LLM_TENSOR_TIME_MIX_DECAY_W1, "blk.%d.time_mix_decay_w1" }, + { LLM_TENSOR_TIME_MIX_DECAY_W2, "blk.%d.time_mix_decay_w2" }, + { LLM_TENSOR_TIME_MIX_KEY, "blk.%d.time_mix_key" }, + { LLM_TENSOR_TIME_MIX_VALUE, "blk.%d.time_mix_value" }, + { LLM_TENSOR_TIME_MIX_RECEPTANCE, "blk.%d.time_mix_receptance" }, + { LLM_TENSOR_TIME_MIX_GATE, "blk.%d.time_mix_gate" }, + { LLM_TENSOR_TIME_MIX_OUTPUT, "blk.%d.time_mix_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_GRANITE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + }, + }, + { + LLM_ARCH_GRANITE_MOE, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE_INP, "blk.%d.ffn_gate_inp" }, + { LLM_TENSOR_FFN_GATE_EXPS, "blk.%d.ffn_gate_exps" }, + { LLM_TENSOR_FFN_DOWN_EXPS, "blk.%d.ffn_down_exps" }, + { LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" }, + }, + }, + { + LLM_ARCH_CHAMELEON, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, + { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, + { LLM_TENSOR_ATTN_K, "blk.%d.attn_k" }, + { LLM_TENSOR_ATTN_V, "blk.%d.attn_v" }, + { LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" }, + { LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" }, + { LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" }, + { LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" }, + { LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" }, + { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, + { LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" }, + }, + }, + { + LLM_ARCH_WAVTOKENIZER_DEC, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" }, + { LLM_TENSOR_CONV1D, "conv1d" }, + { LLM_TENSOR_CONVNEXT_DW, "convnext.%d.dw" }, + { LLM_TENSOR_CONVNEXT_NORM, "convnext.%d.norm" }, + { LLM_TENSOR_CONVNEXT_PW1, "convnext.%d.pw1" }, + { LLM_TENSOR_CONVNEXT_PW2, "convnext.%d.pw2" }, + { LLM_TENSOR_CONVNEXT_GAMMA, "convnext.%d.gamma" }, + { LLM_TENSOR_OUTPUT_NORM, "output_norm" }, + { LLM_TENSOR_OUTPUT, "output" }, + { LLM_TENSOR_POS_NET_CONV1, "posnet.%d.conv1" }, + { LLM_TENSOR_POS_NET_CONV2, "posnet.%d.conv2" }, + { LLM_TENSOR_POS_NET_NORM, "posnet.%d.norm" }, + { LLM_TENSOR_POS_NET_NORM1, "posnet.%d.norm1" }, + { LLM_TENSOR_POS_NET_NORM2, "posnet.%d.norm2" }, + { LLM_TENSOR_POS_NET_ATTN_NORM, "posnet.%d.attn_norm" }, + { LLM_TENSOR_POS_NET_ATTN_Q, "posnet.%d.attn_q" }, + { LLM_TENSOR_POS_NET_ATTN_K, "posnet.%d.attn_k" }, + { LLM_TENSOR_POS_NET_ATTN_V, "posnet.%d.attn_v" }, + { LLM_TENSOR_POS_NET_ATTN_OUT, "posnet.%d.attn_output" }, + }, + }, + { + LLM_ARCH_UNKNOWN, + { + { LLM_TENSOR_TOKEN_EMBD, "token_embd" }, + }, + }, +}; + +static const std::map LLM_TENSOR_INFOS = { + {LLM_TENSOR_TOKEN_EMBD, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_POS_EMBD, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_TOKEN_EMBD_NORM, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_TOKEN_TYPES, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_OUTPUT, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CLS, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CLS_OUT, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_OUTPUT_NORM, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}}, + {LLM_TENSOR_DEC_OUTPUT_NORM, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}}, + {LLM_TENSOR_ENC_OUTPUT_NORM, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}}, + {LLM_TENSOR_ROPE_FREQS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}}, + {LLM_TENSOR_ROPE_FACTORS_LONG, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}}, + {LLM_TENSOR_ROPE_FACTORS_SHORT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}}, + {LLM_TENSOR_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_QKV, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_DOWN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_UP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_DOWN_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_UP_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_Q_A, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_Q_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV_A_MQA, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_QKV, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_DOWN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_UP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_DOWN_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_UP_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_Q_A, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_Q_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV_A_MQA, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ATTN_KV_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_CROSS_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_CROSS_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_CROSS_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_CROSS_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_FFN_GATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_FFN_DOWN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_DEC_FFN_UP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_FFN_GATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_FFN_DOWN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_ENC_FFN_UP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE_INP_SHEXP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_GATE_INP, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_SSM_IN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_SSM_X, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_SSM_DT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_SSM_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_W1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_W2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_DECAY_W1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_DECAY_W2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_KEY, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_VALUE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_RECEPTANCE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_GATE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_TIME_MIX_OUTPUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CHANNEL_MIX_KEY, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CHANNEL_MIX_RECEPTANCE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CHANNEL_MIX_VALUE, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_FFN_ACT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_DIV}}, + {LLM_TENSOR_SSM_CONV1D, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_SSM_CONV}}, + {LLM_TENSOR_SSM_A, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_SSM_SCAN}}, + {LLM_TENSOR_SSM_D, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_TIME_MIX_LERP_X, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_TIME_MIX_LN, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_CHANNEL_MIX_LERP_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_CHANNEL_MIX_LERP_R, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_TIME_MIX_LERP_W, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_LERP_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_LERP_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_LERP_R, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_LERP_G, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_LERP_FUSED, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_DECAY, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + {LLM_TENSOR_TIME_MIX_FIRST, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_RWKV_WKV6}}, + {LLM_TENSOR_ATTN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_NORM_2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_OUT_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_POST_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_FFN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_FFN_POST_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_FFN_NORM_EXPS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_Q_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_K_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_LAYER_OUT_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_Q_A_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_KV_A_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ATTN_SUB_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_FFN_SUB_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_DEC_ATTN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_DEC_CROSS_ATTN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_DEC_FFN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ENC_ATTN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_ENC_FFN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_DEC_ATTN_REL_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_ENC_ATTN_REL_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_GET_ROWS}}, + {LLM_TENSOR_FFN_DOWN_EXPS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}}, + {LLM_TENSOR_FFN_GATE_EXPS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}}, + {LLM_TENSOR_FFN_UP_EXPS, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}}, + {LLM_TENSOR_FFN_EXP_PROBS_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}}, + // this tensor is loaded for T5, but never used + {LLM_TENSOR_DEC_CROSS_ATTN_REL_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_NONE}}, + {LLM_TENSOR_CONV1D, {LLM_TENSOR_LAYER_INPUT, GGML_OP_IM2COL}}, + {LLM_TENSOR_POS_NET_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_POS_NET_NORM1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_POS_NET_NORM2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_POS_NET_CONV1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_IM2COL}}, + {LLM_TENSOR_POS_NET_CONV2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_IM2COL}}, + {LLM_TENSOR_POS_NET_ATTN_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_POS_NET_ATTN_Q, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_POS_NET_ATTN_K, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_POS_NET_ATTN_V, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_POS_NET_ATTN_OUT, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CONVNEXT_DW, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_IM2COL}}, + {LLM_TENSOR_CONVNEXT_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, + {LLM_TENSOR_CONVNEXT_PW1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CONVNEXT_PW2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}}, + {LLM_TENSOR_CONVNEXT_GAMMA, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, +}; + +LLM_KV::LLM_KV(llm_arch arch) : arch(arch) {} + +std::string LLM_KV::operator()(llm_kv kv) const { + return ::format(LLM_KV_NAMES.at(kv), LLM_ARCH_NAMES.at(arch)); +} + +std::string LLM_TN_IMPL::str() const { + if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) { + return "__missing__"; + } + + std::string name = ::format(LLM_TENSOR_NAMES.at(arch).at(tensor), bid, xid); + + if (suffix != nullptr) { + name += "."; + name += suffix; + } + + return name; +} + +const char * llm_arch_name(llm_arch arch) { + auto it = LLM_ARCH_NAMES.find(arch); + if (it == LLM_ARCH_NAMES.end()) { + return "unknown"; + } + return it->second; +} + +llm_arch llm_arch_from_string(const std::string & name) { + for (const auto & kv : LLM_ARCH_NAMES) { // NOLINT + if (kv.second == name) { + return kv.first; + } + } + + return LLM_ARCH_UNKNOWN; +} + +const llm_tensor_info & llm_tensor_info_for(llm_tensor tensor) { + return LLM_TENSOR_INFOS.at(tensor); +} diff --git a/src/llama-arch.h b/src/llama-arch.h new file mode 100644 index 000000000..349844790 --- /dev/null +++ b/src/llama-arch.h @@ -0,0 +1,400 @@ +#pragma once + +#include "ggml.h" // ggml_op + +#include + +// +// gguf constants (sync with gguf.py) +// + +enum llm_arch { + LLM_ARCH_LLAMA, + LLM_ARCH_DECI, + LLM_ARCH_FALCON, + LLM_ARCH_BAICHUAN, + LLM_ARCH_GROK, + LLM_ARCH_GPT2, + LLM_ARCH_GPTJ, + LLM_ARCH_GPTNEOX, + LLM_ARCH_MPT, + LLM_ARCH_STARCODER, + LLM_ARCH_REFACT, + LLM_ARCH_BERT, + LLM_ARCH_NOMIC_BERT, + LLM_ARCH_JINA_BERT_V2, + LLM_ARCH_BLOOM, + LLM_ARCH_STABLELM, + LLM_ARCH_QWEN, + LLM_ARCH_QWEN2, + LLM_ARCH_QWEN2MOE, + LLM_ARCH_QWEN2VL, + LLM_ARCH_PHI2, + LLM_ARCH_PHI3, + LLM_ARCH_PHIMOE, + LLM_ARCH_PLAMO, + LLM_ARCH_CODESHELL, + LLM_ARCH_ORION, + LLM_ARCH_INTERNLM2, + LLM_ARCH_MINICPM, + LLM_ARCH_MINICPM3, + LLM_ARCH_GEMMA, + LLM_ARCH_GEMMA2, + LLM_ARCH_STARCODER2, + LLM_ARCH_MAMBA, + LLM_ARCH_XVERSE, + LLM_ARCH_COMMAND_R, + LLM_ARCH_COHERE2, + LLM_ARCH_DBRX, + LLM_ARCH_OLMO, + LLM_ARCH_OLMO2, + LLM_ARCH_OLMOE, + LLM_ARCH_OPENELM, + LLM_ARCH_ARCTIC, + LLM_ARCH_DEEPSEEK, + LLM_ARCH_DEEPSEEK2, + LLM_ARCH_CHATGLM, + LLM_ARCH_BITNET, + LLM_ARCH_T5, + LLM_ARCH_T5ENCODER, + LLM_ARCH_JAIS, + LLM_ARCH_NEMOTRON, + LLM_ARCH_EXAONE, + LLM_ARCH_RWKV6, + LLM_ARCH_RWKV6QWEN2, + LLM_ARCH_GRANITE, + LLM_ARCH_GRANITE_MOE, + LLM_ARCH_CHAMELEON, + LLM_ARCH_WAVTOKENIZER_DEC, + LLM_ARCH_UNKNOWN, +}; + +enum llm_kv { + LLM_KV_GENERAL_TYPE, + LLM_KV_GENERAL_ARCHITECTURE, + LLM_KV_GENERAL_QUANTIZATION_VERSION, + LLM_KV_GENERAL_ALIGNMENT, + LLM_KV_GENERAL_NAME, + LLM_KV_GENERAL_AUTHOR, + LLM_KV_GENERAL_VERSION, + LLM_KV_GENERAL_URL, + LLM_KV_GENERAL_DESCRIPTION, + LLM_KV_GENERAL_LICENSE, + LLM_KV_GENERAL_SOURCE_URL, + LLM_KV_GENERAL_SOURCE_HF_REPO, + + LLM_KV_VOCAB_SIZE, + LLM_KV_CONTEXT_LENGTH, + LLM_KV_EMBEDDING_LENGTH, + LLM_KV_FEATURES_LENGTH, + LLM_KV_BLOCK_COUNT, + LLM_KV_LEADING_DENSE_BLOCK_COUNT, + LLM_KV_FEED_FORWARD_LENGTH, + LLM_KV_EXPERT_FEED_FORWARD_LENGTH, + LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, + LLM_KV_USE_PARALLEL_RESIDUAL, + LLM_KV_TENSOR_DATA_LAYOUT, + LLM_KV_EXPERT_COUNT, + LLM_KV_EXPERT_USED_COUNT, + LLM_KV_EXPERT_SHARED_COUNT, + LLM_KV_EXPERT_WEIGHTS_SCALE, + LLM_KV_EXPERT_WEIGHTS_NORM, + LLM_KV_EXPERT_GATING_FUNC, + LLM_KV_POOLING_TYPE, + LLM_KV_LOGIT_SCALE, + LLM_KV_DECODER_START_TOKEN_ID, + LLM_KV_ATTN_LOGIT_SOFTCAPPING, + LLM_KV_FINAL_LOGIT_SOFTCAPPING, + LLM_KV_SWIN_NORM, + LLM_KV_RESCALE_EVERY_N_LAYERS, + LLM_KV_TIME_MIX_EXTRA_DIM, + LLM_KV_TIME_DECAY_EXTRA_DIM, + LLM_KV_RESIDUAL_SCALE, + LLM_KV_EMBEDDING_SCALE, + LLM_KV_TOKEN_SHIFT_COUNT, + + LLM_KV_ATTENTION_HEAD_COUNT, + LLM_KV_ATTENTION_HEAD_COUNT_KV, + LLM_KV_ATTENTION_MAX_ALIBI_BIAS, + LLM_KV_ATTENTION_CLAMP_KQV, + LLM_KV_ATTENTION_KEY_LENGTH, + LLM_KV_ATTENTION_VALUE_LENGTH, + LLM_KV_ATTENTION_LAYERNORM_EPS, + LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, + LLM_KV_ATTENTION_GROUPNORM_EPS, + LLM_KV_ATTENTION_GROUPNORM_GROUPS, + LLM_KV_ATTENTION_CAUSAL, + LLM_KV_ATTENTION_Q_LORA_RANK, + LLM_KV_ATTENTION_KV_LORA_RANK, + LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, + LLM_KV_ATTENTION_SLIDING_WINDOW, + LLM_KV_ATTENTION_SCALE, + + LLM_KV_ROPE_DIMENSION_COUNT, + LLM_KV_ROPE_DIMENSION_SECTIONS, + LLM_KV_ROPE_FREQ_BASE, + LLM_KV_ROPE_SCALE_LINEAR, + LLM_KV_ROPE_SCALING_TYPE, + LLM_KV_ROPE_SCALING_FACTOR, + LLM_KV_ROPE_SCALING_ATTN_FACTOR, + LLM_KV_ROPE_SCALING_ORIG_CTX_LEN, + LLM_KV_ROPE_SCALING_FINETUNED, + LLM_KV_ROPE_SCALING_YARN_LOG_MUL, + + LLM_KV_SPLIT_NO, + LLM_KV_SPLIT_COUNT, + LLM_KV_SPLIT_TENSORS_COUNT, + + LLM_KV_SSM_INNER_SIZE, + LLM_KV_SSM_CONV_KERNEL, + LLM_KV_SSM_STATE_SIZE, + LLM_KV_SSM_TIME_STEP_RANK, + LLM_KV_SSM_DT_B_C_RMS, + + LLM_KV_WKV_HEAD_SIZE, + + LLM_KV_TOKENIZER_MODEL, + LLM_KV_TOKENIZER_PRE, + LLM_KV_TOKENIZER_LIST, + LLM_KV_TOKENIZER_TOKEN_TYPE, + LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, + LLM_KV_TOKENIZER_SCORES, + LLM_KV_TOKENIZER_MERGES, + LLM_KV_TOKENIZER_BOS_ID, + LLM_KV_TOKENIZER_EOS_ID, + LLM_KV_TOKENIZER_EOT_ID, + LLM_KV_TOKENIZER_EOM_ID, + LLM_KV_TOKENIZER_UNK_ID, + LLM_KV_TOKENIZER_SEP_ID, + LLM_KV_TOKENIZER_PAD_ID, + LLM_KV_TOKENIZER_CLS_ID, + LLM_KV_TOKENIZER_MASK_ID, + LLM_KV_TOKENIZER_ADD_BOS, + LLM_KV_TOKENIZER_ADD_EOS, + LLM_KV_TOKENIZER_ADD_PREFIX, + LLM_KV_TOKENIZER_REMOVE_EXTRA_WS, + LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP, + LLM_KV_TOKENIZER_HF_JSON, + LLM_KV_TOKENIZER_RWKV, + LLM_KV_TOKENIZER_CHAT_TEMPLATE, + LLM_KV_TOKENIZER_FIM_PRE_ID, + LLM_KV_TOKENIZER_FIM_SUF_ID, + LLM_KV_TOKENIZER_FIM_MID_ID, + LLM_KV_TOKENIZER_FIM_PAD_ID, + LLM_KV_TOKENIZER_FIM_REP_ID, + LLM_KV_TOKENIZER_FIM_SEP_ID, + + LLM_KV_ADAPTER_TYPE, + LLM_KV_ADAPTER_LORA_ALPHA, + + LLM_KV_POSNET_EMBEDDING_LENGTH, + LLM_KV_POSNET_BLOCK_COUNT, + + LLM_KV_CONVNEXT_EMBEDDING_LENGTH, + LLM_KV_CONVNEXT_BLOCK_COUNT, + + // deprecated: + LLM_KV_TOKENIZER_PREFIX_ID, + LLM_KV_TOKENIZER_SUFFIX_ID, + LLM_KV_TOKENIZER_MIDDLE_ID, +}; + +enum llm_tensor { + LLM_TENSOR_TOKEN_EMBD, + LLM_TENSOR_TOKEN_EMBD_NORM, + LLM_TENSOR_TOKEN_TYPES, + LLM_TENSOR_POS_EMBD, + LLM_TENSOR_OUTPUT, + LLM_TENSOR_OUTPUT_NORM, + LLM_TENSOR_ROPE_FREQS, + LLM_TENSOR_ROPE_FACTORS_LONG, + LLM_TENSOR_ROPE_FACTORS_SHORT, + LLM_TENSOR_ATTN_Q, + LLM_TENSOR_ATTN_K, + LLM_TENSOR_ATTN_V, + LLM_TENSOR_ATTN_QKV, + LLM_TENSOR_ATTN_OUT, + LLM_TENSOR_ATTN_NORM, + LLM_TENSOR_ATTN_NORM_2, + LLM_TENSOR_ATTN_OUT_NORM, + LLM_TENSOR_ATTN_POST_NORM, + LLM_TENSOR_ATTN_ROT_EMBD, + LLM_TENSOR_FFN_GATE_INP, + LLM_TENSOR_FFN_GATE_INP_SHEXP, + LLM_TENSOR_FFN_NORM, + LLM_TENSOR_FFN_POST_NORM, + LLM_TENSOR_FFN_GATE, + LLM_TENSOR_FFN_DOWN, + LLM_TENSOR_FFN_UP, + LLM_TENSOR_FFN_ACT, + LLM_TENSOR_FFN_DOWN_EXP, // split experts for backward compatibility + LLM_TENSOR_FFN_GATE_EXP, + LLM_TENSOR_FFN_UP_EXP, + LLM_TENSOR_FFN_NORM_EXPS, + LLM_TENSOR_FFN_DOWN_EXPS, // merged experts + LLM_TENSOR_FFN_GATE_EXPS, + LLM_TENSOR_FFN_UP_EXPS, + LLM_TENSOR_FFN_DOWN_SHEXP, + LLM_TENSOR_FFN_GATE_SHEXP, + LLM_TENSOR_FFN_UP_SHEXP, + LLM_TENSOR_FFN_EXP_PROBS_B, + LLM_TENSOR_ATTN_Q_NORM, + LLM_TENSOR_ATTN_K_NORM, + LLM_TENSOR_LAYER_OUT_NORM, + LLM_TENSOR_SSM_IN, + LLM_TENSOR_SSM_CONV1D, + LLM_TENSOR_SSM_X, + LLM_TENSOR_SSM_DT, + LLM_TENSOR_SSM_A, + LLM_TENSOR_SSM_D, + LLM_TENSOR_SSM_OUT, + LLM_TENSOR_TIME_MIX_W1, + LLM_TENSOR_TIME_MIX_W2, + LLM_TENSOR_TIME_MIX_LERP_X, + LLM_TENSOR_TIME_MIX_LERP_W, + LLM_TENSOR_TIME_MIX_LERP_K, + LLM_TENSOR_TIME_MIX_LERP_V, + LLM_TENSOR_TIME_MIX_LERP_R, + LLM_TENSOR_TIME_MIX_LERP_G, + LLM_TENSOR_TIME_MIX_LERP_FUSED, + LLM_TENSOR_TIME_MIX_FIRST, + LLM_TENSOR_TIME_MIX_DECAY, + LLM_TENSOR_TIME_MIX_DECAY_W1, + LLM_TENSOR_TIME_MIX_DECAY_W2, + LLM_TENSOR_TIME_MIX_KEY, + LLM_TENSOR_TIME_MIX_VALUE, + LLM_TENSOR_TIME_MIX_RECEPTANCE, + LLM_TENSOR_TIME_MIX_GATE, + LLM_TENSOR_TIME_MIX_LN, + LLM_TENSOR_TIME_MIX_OUTPUT, + LLM_TENSOR_CHANNEL_MIX_LERP_K, + LLM_TENSOR_CHANNEL_MIX_LERP_R, + LLM_TENSOR_CHANNEL_MIX_KEY, + LLM_TENSOR_CHANNEL_MIX_RECEPTANCE, + LLM_TENSOR_CHANNEL_MIX_VALUE, + LLM_TENSOR_ATTN_Q_A, + LLM_TENSOR_ATTN_Q_B, + LLM_TENSOR_ATTN_KV_A_MQA, + LLM_TENSOR_ATTN_KV_B, + LLM_TENSOR_ATTN_Q_A_NORM, + LLM_TENSOR_ATTN_KV_A_NORM, + LLM_TENSOR_ATTN_SUB_NORM, + LLM_TENSOR_FFN_SUB_NORM, + LLM_TENSOR_DEC_ATTN_NORM, + LLM_TENSOR_DEC_ATTN_Q, + LLM_TENSOR_DEC_ATTN_K, + LLM_TENSOR_DEC_ATTN_V, + LLM_TENSOR_DEC_ATTN_OUT, + LLM_TENSOR_DEC_ATTN_REL_B, + LLM_TENSOR_DEC_CROSS_ATTN_NORM, + LLM_TENSOR_DEC_CROSS_ATTN_Q, + LLM_TENSOR_DEC_CROSS_ATTN_K, + LLM_TENSOR_DEC_CROSS_ATTN_V, + LLM_TENSOR_DEC_CROSS_ATTN_OUT, + LLM_TENSOR_DEC_CROSS_ATTN_REL_B, + LLM_TENSOR_DEC_FFN_NORM, + LLM_TENSOR_DEC_FFN_GATE, + LLM_TENSOR_DEC_FFN_DOWN, + LLM_TENSOR_DEC_FFN_UP, + LLM_TENSOR_DEC_OUTPUT_NORM, + LLM_TENSOR_ENC_ATTN_NORM, + LLM_TENSOR_ENC_ATTN_Q, + LLM_TENSOR_ENC_ATTN_K, + LLM_TENSOR_ENC_ATTN_V, + LLM_TENSOR_ENC_ATTN_OUT, + LLM_TENSOR_ENC_ATTN_REL_B, + LLM_TENSOR_ENC_FFN_NORM, + LLM_TENSOR_ENC_FFN_GATE, + LLM_TENSOR_ENC_FFN_DOWN, + LLM_TENSOR_ENC_FFN_UP, + LLM_TENSOR_ENC_OUTPUT_NORM, + LLM_TENSOR_CLS, + LLM_TENSOR_CLS_OUT, + LLM_TENSOR_CONV1D, + LLM_TENSOR_CONVNEXT_DW, + LLM_TENSOR_CONVNEXT_NORM, + LLM_TENSOR_CONVNEXT_PW1, + LLM_TENSOR_CONVNEXT_PW2, + LLM_TENSOR_CONVNEXT_GAMMA, + LLM_TENSOR_POS_NET_CONV1, + LLM_TENSOR_POS_NET_CONV2, + LLM_TENSOR_POS_NET_NORM, + LLM_TENSOR_POS_NET_NORM1, + LLM_TENSOR_POS_NET_NORM2, + LLM_TENSOR_POS_NET_ATTN_NORM, + LLM_TENSOR_POS_NET_ATTN_Q, + LLM_TENSOR_POS_NET_ATTN_K, + LLM_TENSOR_POS_NET_ATTN_V, + LLM_TENSOR_POS_NET_ATTN_OUT, +}; + +enum llm_tensor_layer { + LLM_TENSOR_LAYER_INPUT, + LLM_TENSOR_LAYER_REPEATING, + LLM_TENSOR_LAYER_OUTPUT, +}; + +struct LLM_KV { + LLM_KV(llm_arch arch); + + llm_arch arch; + + std::string operator()(llm_kv kv) const; +}; + +// helper to handle gguf constants +// usage: +// +// const auto tn = LLM_TN(LLM_ARCH_LLAMA); +// +// std::string name = tn(LLM_TENSOR_OUTPUT); -> "output" +// std::string name = tn(LLM_TENSOR_TOKEN_EMBD, "bias"); -> "token_embd.bias" +// std::string name = tn(LLM_TENSOR_ATTN_NORM, "weight", 3); -> "blk.3.attn_norm.weight" +// +struct LLM_TN_IMPL { + const llm_arch arch; + const llm_tensor tensor; + const char * const suffix; + const int bid; + const int xid; + + std::string str() const; + + operator std::string() const { + return str(); + } + + friend bool operator==(const std::string & str, const LLM_TN_IMPL & tn) { + return str == tn.str(); + } + + friend bool operator!=(const std::string & str, const LLM_TN_IMPL & tn) { + return str != tn.str(); + } +}; + +struct LLM_TN { + LLM_TN(llm_arch arch) : arch(arch) {} + + llm_arch arch; + + LLM_TN_IMPL operator()(llm_tensor tensor, const char * suffix, int bid = -1, int xid = -1) const { + return { arch, tensor, suffix, bid, xid }; + } + + LLM_TN_IMPL operator()(llm_tensor tensor, int bid = -1, int xid = -1) const { + return { arch, tensor, nullptr, bid, xid }; + } +}; + + +struct llm_tensor_info { + llm_tensor_layer layer; + ggml_op op; +}; + +const char * llm_arch_name(llm_arch arch); + +llm_arch llm_arch_from_string(const std::string & name); + +const llm_tensor_info & llm_tensor_info_for(llm_tensor tensor); diff --git a/src/llama-batch.cpp b/src/llama-batch.cpp new file mode 100644 index 000000000..01d5ca57f --- /dev/null +++ b/src/llama-batch.cpp @@ -0,0 +1,368 @@ +#include "llama-batch.h" + +#include +#include + +llama_ubatch llama_sbatch::reserve_ubatch(size_t n_ubatch, bool has_embd) { + // clear empty sequences + // the previous ubatch is assumed to be gone, + // so nothing should refer to values in these sequences anymore. + for (size_t i = seq.size(); i-- > 0;) { + if (seq[i].length == 0) { + seq.pop_back(); + } else { + break; + } + } + ubatch_token.resize(!has_embd ? n_ubatch : 0); + ubatch_embd.resize(has_embd ? n_embd * n_ubatch : 0); + ubatch_pos.resize(n_ubatch); + ubatch_n_seq_id.resize(n_ubatch); + ubatch_seq_id.resize(n_ubatch); + ubatch_output.resize(n_ubatch); + llama_ubatch ubatch = { + /*equal_seqs =*/ true, + /*n_tokens =*/ 0, + /*n_seq_tokens =*/ 0, + /*n_seqs =*/ 0, + /*token =*/ !has_embd ? ubatch_token.data() : nullptr, + /*embd =*/ has_embd ? ubatch_embd.data() : nullptr, + /*pos =*/ ubatch_pos.data(), + /*n_seq_id =*/ ubatch_n_seq_id.data(), + /*seq_id =*/ ubatch_seq_id.data(), + /*output =*/ ubatch_output.data(), + }; + return ubatch; +} + +void llama_sbatch::add_seq_to_ubatch(llama_ubatch & ubatch, llama_sbatch_seq & seq, size_t length) { + GGML_ASSERT(batch != nullptr); + GGML_ASSERT(length <= seq.length); + // Can only add sequences of equal lengths to a batch, + // otherwise it isn't clear to which sequence a token belongs + GGML_ASSERT(seq.n_seq_id == 0 || ubatch.n_seqs == 0 || length == (size_t) ubatch.n_tokens / ubatch.n_seqs); + GGML_ASSERT((seq.n_seq_id != 0) == ubatch.equal_seqs); + // NOTE: loops are separated for cache-friendliness + if (batch->token) { + if (ubatch.equal_seqs) { + for (size_t i = 0; i < length; ++i) { + ubatch.token[ubatch.n_tokens + i] = batch->token[ids[seq.offset + i]]; + } + } else { + // simple split + ubatch.token = batch->token + seq.offset; + } + } else { + ubatch.token = nullptr; + } + if (batch->embd) { + if (ubatch.equal_seqs) { + for (size_t i = 0; i < length; ++i) { + memcpy( + ubatch.embd + (n_embd * (ubatch.n_tokens + i)), + batch->embd + (n_embd * ids[seq.offset + i]), + n_embd * sizeof(float) + ); + } + } else { + // simple split + ubatch.embd = batch->embd + (n_embd * seq.offset); + } + } else { + ubatch.embd = nullptr; + } + if (ubatch.equal_seqs) { + for (size_t i = 0; i < length; ++i) { + ubatch.pos[ubatch.n_tokens + i] = batch->pos[ids[seq.offset + i]]; + } + } else { + // simple split + ubatch.pos = batch->pos + seq.offset; + } + if (ubatch.equal_seqs) { + ubatch.n_seq_id[ubatch.n_seqs] = seq.n_seq_id; + if (seq.seq_id) { + ubatch.seq_id[ubatch.n_seqs] = seq.seq_id; + } + } else { + // simple split + if (batch->n_seq_id) { + ubatch.n_seq_id = batch->n_seq_id + seq.offset; + } else { + for (size_t i = 0; i < length; ++i) { + ubatch.n_seq_id[ubatch.n_seqs + i] = 1; + } + } + if (batch->seq_id) { + ubatch.seq_id = batch->seq_id + seq.offset; + } + } + if (logits_all) { + for (size_t i = 0; i < length; ++i) { + ubatch.output[ubatch.n_tokens + i] = 1; + out_ids.push_back(ids[seq.offset + i]); + } + } else if (batch->logits) { + if (ubatch.equal_seqs) { + for (size_t i = 0; i < length; ++i) { + size_t id = ids[seq.offset + i]; + int8_t is_output = batch->logits[id]; + ubatch.output[ubatch.n_tokens + i] = is_output; + if (is_output) { out_ids.push_back(id); } + } + } else { + // simple split + ubatch.output = batch->logits + seq.offset; + for (size_t i = 0; i < length; ++i) { + if (ubatch.output[i] != 0) { out_ids.push_back(seq.offset + i); } + } + } + } else { + // only get last output + for (size_t i = 0; i < length; ++i) { + size_t id = ids[seq.offset + i]; + int8_t is_last = id == ids.size() - 1; + ubatch.output[ubatch.n_tokens + i] = is_last; + if (is_last) { out_ids.push_back(id); } + } + } + if (ubatch.n_tokens == 0 && ubatch.n_seqs == 0) { + ubatch.n_seq_tokens = ubatch.equal_seqs ? length : 1; + } + ubatch.n_tokens += length; + ubatch.n_seqs += ubatch.equal_seqs ? 1 : length; // virtual sequences for simple splits + seq.offset += length; + seq.length -= length; + n_tokens -= length; + GGML_ASSERT(ubatch.n_tokens == ubatch.n_seq_tokens * ubatch.n_seqs); +} + +llama_ubatch llama_sbatch::split_simple(size_t n_ubatch) { + n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch; + llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr); + ubatch.equal_seqs = false; + if (!seq.empty()) { + llama_sbatch_seq & s = seq[0]; + size_t length = s.length < n_ubatch ? s.length : n_ubatch; + GGML_ASSERT(seq.size() == 1 && s.n_seq_id == 0); // don't mix with other splits + add_seq_to_ubatch(ubatch, s, length); + } + return ubatch; +} + +llama_ubatch llama_sbatch::split_equal(size_t n_ubatch) { + n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch; + llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr); + if (!seq.empty()) { + size_t length = 0; + size_t n_tokens_in_ubatch = 0; + GGML_ASSERT(seq[0].n_seq_id > 0); // should not be mixed with simple splits + // smallest first, because it's easier to split this way; + // starting from the end to pop in constant time. + for (size_t i = seq.size(); i-- > 0;) { + llama_sbatch_seq & s = seq[i]; + GGML_ASSERT(s.length > 0); + if (length == 0) { + length = s.length < n_ubatch ? s.length : n_ubatch; + } + add_seq_to_ubatch(ubatch, s, length); + n_tokens_in_ubatch += length; + // shared prompts can't be mixed with any of their sequences, + // so it's safer to compute them in their own ubatch + if (s.n_seq_id > 1) { break; } + // stop when there isn't enough space for another sequence + if (length + n_tokens_in_ubatch > n_ubatch) { break; } + } + } + return ubatch; +} + +llama_ubatch llama_sbatch::split_seq(size_t n_ubatch) { + n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch; + llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr); + if (!seq.empty()) { + llama_sbatch_seq & s = seq[seq.size() - 1]; + size_t length = s.length < n_ubatch ? s.length : n_ubatch; + GGML_ASSERT(s.n_seq_id > 0); // should not be mixed with simple splits + add_seq_to_ubatch(ubatch, s, length); + } + return ubatch; +} + +void llama_sbatch::from_batch(const llama_batch & batch, size_t n_embd, bool simple_split, bool logits_all) { + GGML_ASSERT(batch.n_tokens >= 0); + this->batch = &batch; + this->n_embd = n_embd; + this->logits_all = logits_all; + + n_tokens = batch.n_tokens; + ids.resize(n_tokens); + out_ids.clear(); + // TODO: reserve out_ids and seq + + for (size_t i = 0; i < n_tokens; ++i) { + ids[i] = i; + } + if (simple_split) { + seq.resize(1); + llama_sbatch_seq & s = seq[0]; + s.n_seq_id = 0; + s.seq_id = nullptr; + s.offset = 0; + s.length = n_tokens; + return; + } + std::sort(ids.begin(), ids.end(), + [&batch](size_t a, size_t b) { + int32_t n_seq_a = batch.n_seq_id ? batch.n_seq_id[a] : 1; + int32_t n_seq_b = batch.n_seq_id ? batch.n_seq_id[b] : 1; + // sort by seq_id, then by pos + if (n_seq_a == n_seq_b) { + if (batch.seq_id) { + for (int32_t i = 0; i < n_seq_a; ++i) { + llama_seq_id seq_id_a = batch.seq_id[a][i]; + llama_seq_id seq_id_b = batch.seq_id[b][i]; + // smaller seq_ids go first + if (seq_id_a != seq_id_b) { + return seq_id_a < seq_id_b; + } + } + } + // when all else is equal, sort by pos + if (batch.pos) { + return batch.pos[a] < batch.pos[b]; + } + // no pos, sort by id + return a < b; + } + // shared prompts go first + return n_seq_a > n_seq_b; + } + ); + // init seq + llama_sbatch_seq * last_seq = nullptr; + + for (size_t i = 0; i < n_tokens; ++i) { + const size_t bi = ids[i]; + const int32_t n_seqs = batch.n_seq_id[bi]; + llama_seq_id * seq_ids = batch.seq_id[bi]; + if (last_seq != nullptr) { + bool same = n_seqs == last_seq->n_seq_id; + for (int32_t j = 0; same && j < n_seqs; ++j) { + if (seq_ids[j] != last_seq->seq_id[j]) { + same = false; + } + } + if (same) { + last_seq->length += 1; + continue; + } + } + llama_sbatch_seq new_seq = {n_seqs, seq_ids, i, 1}; + seq.push_back(new_seq); + last_seq = &seq.back(); + } + // keep shared prompts first at the end, then sort by length descending. + std::sort(seq.begin(), seq.end(), + [](llama_sbatch_seq & a, llama_sbatch_seq & b) { + if (a.n_seq_id == b.n_seq_id) { + return a.length > b.length; + } + return a.n_seq_id < b.n_seq_id; + } + ); +} + +llama_batch_allocr::llama_batch_allocr(struct llama_batch in_batch, llama_pos p0) { + batch = in_batch; + GGML_ASSERT(batch.n_tokens > 0); + if (!batch.pos) { + pos.resize(batch.n_tokens); + for (int32_t i = 0; i < batch.n_tokens; i++) { + pos[i] = i + p0; + } + batch.pos = pos.data(); + } + if (!batch.n_seq_id) { + n_seq_id.resize(batch.n_tokens); + for (int32_t i = 0; i < batch.n_tokens; i++) { + n_seq_id[i] = seq_id_0.size(); + } + batch.n_seq_id = n_seq_id.data(); + } + if (!batch.seq_id) { + seq_id.resize(batch.n_tokens + 1); + seq_id[batch.n_tokens] = NULL; + for (int32_t i = 0; i < batch.n_tokens; i++) { + seq_id[i] = seq_id_0.data(); + } + batch.seq_id = seq_id.data(); + } + if (!batch.logits) { + logits.resize(batch.n_tokens); + logits[logits.size() - 1] = true; + batch.logits = logits.data(); + } +} + +// +// interface implementation +// + +struct llama_batch llama_batch_get_one( + llama_token * tokens, + int32_t n_tokens) { + return { + /*n_tokens =*/ n_tokens, + /*tokens =*/ tokens, + /*embd =*/ nullptr, + /*pos =*/ nullptr, + /*n_seq_id =*/ nullptr, + /*seq_id =*/ nullptr, + /*logits =*/ nullptr, + }; +} + +struct llama_batch llama_batch_init(int32_t n_tokens_alloc, int32_t embd, int32_t n_seq_max) { + llama_batch batch = { + /*n_tokens =*/ 0, + /*tokens =*/ nullptr, + /*embd =*/ nullptr, + /*pos =*/ nullptr, + /*n_seq_id =*/ nullptr, + /*seq_id =*/ nullptr, + /*logits =*/ nullptr, + }; + + if (embd) { + batch.embd = (float *) malloc(sizeof(float) * n_tokens_alloc * embd); + } else { + batch.token = (llama_token *) malloc(sizeof(llama_token) * n_tokens_alloc); + } + + batch.pos = (llama_pos *) malloc(sizeof(llama_pos) * n_tokens_alloc); + batch.n_seq_id = (int32_t *) malloc(sizeof(int32_t) * n_tokens_alloc); + batch.seq_id = (llama_seq_id **) malloc(sizeof(llama_seq_id *) * (n_tokens_alloc + 1)); + for (int i = 0; i < n_tokens_alloc; ++i) { + batch.seq_id[i] = (llama_seq_id *) malloc(sizeof(llama_seq_id) * n_seq_max); + } + batch.seq_id[n_tokens_alloc] = nullptr; + + batch.logits = (int8_t *) malloc(sizeof(int8_t) * n_tokens_alloc); + + return batch; +} + +void llama_batch_free(struct llama_batch batch) { + if (batch.token) free(batch.token); + if (batch.embd) free(batch.embd); + if (batch.pos) free(batch.pos); + if (batch.n_seq_id) free(batch.n_seq_id); + if (batch.seq_id) { + for (int i = 0; batch.seq_id[i] != nullptr; ++i) { + free(batch.seq_id[i]); + } + free(batch.seq_id); + } + if (batch.logits) free(batch.logits); +} diff --git a/src/llama-batch.h b/src/llama-batch.h new file mode 100644 index 000000000..773c3808b --- /dev/null +++ b/src/llama-batch.h @@ -0,0 +1,88 @@ +#pragma once + +#include "llama.h" + +#include +#include + +// very similar to llama_batch, +// but has more metadata about sequences +struct llama_ubatch { + bool equal_seqs; + // TODO: whole_seqs for embeddings? + + uint32_t n_tokens; // total tokens (n_seq_tokens * n_seqs) + uint32_t n_seq_tokens; // tokens per sequence + uint32_t n_seqs; + + llama_token * token; // [n_tokens] + float * embd; // [n_embd, n_tokens] + llama_pos * pos; // [n_tokens] + int32_t * n_seq_id; // [n_seqs] + llama_seq_id ** seq_id; // [n_seqs] + int8_t * output; // [n_tokens] +}; + +struct llama_sbatch_seq { + int32_t n_seq_id; + + llama_seq_id * seq_id; + + size_t offset; + size_t length; +}; + +// sequence-length-aware batch splitting +struct llama_sbatch { + // tokens left in this batch + size_t n_tokens; + + size_t n_embd; + + bool logits_all; // TODO: remove once lctx.logits_all is removed too + + // sorted indices into the batch + std::vector ids; + // batch indices of the output + std::vector out_ids; + std::vector seq; + + const llama_batch * batch = nullptr; + + // buffers for the ubatch + std::vector ubatch_token; + std::vector ubatch_embd; + std::vector ubatch_pos; + std::vector ubatch_n_seq_id; + std::vector ubatch_seq_id; + std::vector ubatch_output; + + llama_ubatch reserve_ubatch(size_t n_ubatch, bool has_embd = false); + + void add_seq_to_ubatch(llama_ubatch & ubatch, llama_sbatch_seq & seq, size_t length); + + // simple split, unknown number of sequences of unequal lengths + llama_ubatch split_simple(size_t n_ubatch); + + // make batches of equal-length sequences + llama_ubatch split_equal(size_t n_ubatch); + + // sequence-wise split + llama_ubatch split_seq(size_t n_ubatch); + + void from_batch(const llama_batch & batch, size_t n_embd, bool simple_split = false, bool logits_all = false); +}; + +// temporary allocate memory for the input batch if needed +struct llama_batch_allocr { + struct llama_batch batch; + + std::array seq_id_0 = { 0 }; // default sequence id + std::vector pos; + std::vector n_seq_id; + std::vector seq_id; + std::vector logits; + + // optionally fulfill the batch returned by llama_batch_get_one + llama_batch_allocr(struct llama_batch in_batch, llama_pos p0); +}; diff --git a/src/llama-chat.cpp b/src/llama-chat.cpp new file mode 100644 index 000000000..1347ec156 --- /dev/null +++ b/src/llama-chat.cpp @@ -0,0 +1,578 @@ +#include "llama-chat.h" + +#include "llama.h" + +#include +#include + +#if __cplusplus >= 202000L + #define LU8(x) (const char*)(u8##x) +#else + #define LU8(x) u8##x +#endif + +// trim whitespace from the beginning and end of a string +static std::string trim(const std::string & str) { + size_t start = 0; + size_t end = str.size(); + while (start < end && isspace(str[start])) { + start += 1; + } + while (end > start && isspace(str[end - 1])) { + end -= 1; + } + return str.substr(start, end - start); +} + +static const std::map LLM_CHAT_TEMPLATES = { + { "chatml", LLM_CHAT_TEMPLATE_CHATML }, + { "llama2", LLM_CHAT_TEMPLATE_LLAMA_2 }, + { "llama2-sys", LLM_CHAT_TEMPLATE_LLAMA_2_SYS }, + { "llama2-sys-bos", LLM_CHAT_TEMPLATE_LLAMA_2_SYS_BOS }, + { "llama2-sys-strip", LLM_CHAT_TEMPLATE_LLAMA_2_SYS_STRIP }, + { "mistral-v1", LLM_CHAT_TEMPLATE_MISTRAL_V1 }, + { "mistral-v3", LLM_CHAT_TEMPLATE_MISTRAL_V3 }, + { "mistral-v3-tekken", LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN }, + { "mistral-v7", LLM_CHAT_TEMPLATE_MISTRAL_V7 }, + { "phi3", LLM_CHAT_TEMPLATE_PHI_3 }, + { "phi4", LLM_CHAT_TEMPLATE_PHI_4 }, + { "falcon3", LLM_CHAT_TEMPLATE_FALCON_3 }, + { "zephyr", LLM_CHAT_TEMPLATE_ZEPHYR }, + { "monarch", LLM_CHAT_TEMPLATE_MONARCH }, + { "gemma", LLM_CHAT_TEMPLATE_GEMMA }, + { "orion", LLM_CHAT_TEMPLATE_ORION }, + { "openchat", LLM_CHAT_TEMPLATE_OPENCHAT }, + { "vicuna", LLM_CHAT_TEMPLATE_VICUNA }, + { "vicuna-orca", LLM_CHAT_TEMPLATE_VICUNA_ORCA }, + { "deepseek", LLM_CHAT_TEMPLATE_DEEPSEEK }, + { "deepseek2", LLM_CHAT_TEMPLATE_DEEPSEEK_2 }, + { "deepseek3", LLM_CHAT_TEMPLATE_DEEPSEEK_3 }, + { "command-r", LLM_CHAT_TEMPLATE_COMMAND_R }, + { "llama3", LLM_CHAT_TEMPLATE_LLAMA_3 }, + { "chatglm3", LLM_CHAT_TEMPLATE_CHATGML_3 }, + { "chatglm4", LLM_CHAT_TEMPLATE_CHATGML_4 }, + { "minicpm", LLM_CHAT_TEMPLATE_MINICPM }, + { "exaone3", LLM_CHAT_TEMPLATE_EXAONE_3 }, + { "rwkv-world", LLM_CHAT_TEMPLATE_RWKV_WORLD }, + { "granite", LLM_CHAT_TEMPLATE_GRANITE }, + { "gigachat", LLM_CHAT_TEMPLATE_GIGACHAT }, + { "megrez", LLM_CHAT_TEMPLATE_MEGREZ }, +}; + +llm_chat_template llm_chat_template_from_str(const std::string & name) { + return LLM_CHAT_TEMPLATES.at(name); +} + +llm_chat_template llm_chat_detect_template(const std::string & tmpl) { + try { + return llm_chat_template_from_str(tmpl); + } catch (const std::out_of_range &) { + // ignore + } + + auto tmpl_contains = [&tmpl](const char * haystack) -> bool { + return tmpl.find(haystack) != std::string::npos; + }; + if (tmpl_contains("<|im_start|>")) { + return tmpl_contains("<|im_sep|>") + ? LLM_CHAT_TEMPLATE_PHI_4 + : LLM_CHAT_TEMPLATE_CHATML; + } else if (tmpl.find("mistral") == 0 || tmpl_contains("[INST]")) { + if (tmpl_contains("[SYSTEM_PROMPT]")) { + return LLM_CHAT_TEMPLATE_MISTRAL_V7; + } else if ( + // catches official 'v1' template + tmpl_contains("' [INST] ' + system_message") + // catches official 'v3' and 'v3-tekken' templates + || tmpl_contains("[AVAILABLE_TOOLS]") + ) { + // Official mistral 'v1', 'v3' and 'v3-tekken' templates + // See: https://github.com/mistralai/cookbook/blob/main/concept-deep-dive/tokenization/chat_templates.md + // See: https://github.com/mistralai/cookbook/blob/main/concept-deep-dive/tokenization/templates.md + if (tmpl_contains(" [INST]")) { + return LLM_CHAT_TEMPLATE_MISTRAL_V1; + } else if (tmpl_contains("\"[INST]\"")) { + return LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN; + } + return LLM_CHAT_TEMPLATE_MISTRAL_V3; + } else { + // llama2 template and its variants + // [variant] support system message + // See: https://huggingface.co/blog/llama2#how-to-prompt-llama-2 + bool support_system_message = tmpl_contains("<>"); + bool add_bos_inside_history = tmpl_contains("bos_token + '[INST]"); + bool strip_message = tmpl_contains("content.strip()"); + if (strip_message) { + return LLM_CHAT_TEMPLATE_LLAMA_2_SYS_STRIP; + } else if (add_bos_inside_history) { + return LLM_CHAT_TEMPLATE_LLAMA_2_SYS_BOS; + } else if (support_system_message) { + return LLM_CHAT_TEMPLATE_LLAMA_2_SYS; + } else { + return LLM_CHAT_TEMPLATE_LLAMA_2; + } + } + } else if (tmpl_contains("<|assistant|>") && tmpl_contains("<|end|>")) { + return LLM_CHAT_TEMPLATE_PHI_3; + } else if (tmpl_contains("<|assistant|>") && tmpl_contains("<|user|>")) { + return LLM_CHAT_TEMPLATE_FALCON_3; + } else if (tmpl_contains("<|user|>") && tmpl_contains("<|endoftext|>")) { + return LLM_CHAT_TEMPLATE_ZEPHYR; + } else if (tmpl_contains("bos_token + message['role']")) { + return LLM_CHAT_TEMPLATE_MONARCH; + } else if (tmpl_contains("")) { + return LLM_CHAT_TEMPLATE_GEMMA; + } else if (tmpl_contains("'\\n\\nAssistant: ' + eos_token")) { + // OrionStarAI/Orion-14B-Chat + return LLM_CHAT_TEMPLATE_ORION; + } else if (tmpl_contains("GPT4 Correct ")) { + // openchat/openchat-3.5-0106 + return LLM_CHAT_TEMPLATE_OPENCHAT; + } else if (tmpl_contains("USER: ") && tmpl_contains("ASSISTANT: ")) { + // eachadea/vicuna-13b-1.1 (and Orca variant) + if (tmpl_contains("SYSTEM: ")) { + return LLM_CHAT_TEMPLATE_VICUNA_ORCA; + } + return LLM_CHAT_TEMPLATE_VICUNA; + } else if (tmpl_contains("### Instruction:") && tmpl_contains("<|EOT|>")) { + // deepseek-ai/deepseek-coder-33b-instruct + return LLM_CHAT_TEMPLATE_DEEPSEEK; + } else if (tmpl_contains("<|START_OF_TURN_TOKEN|>") && tmpl_contains("<|USER_TOKEN|>")) { + // CohereForAI/c4ai-command-r-plus + return LLM_CHAT_TEMPLATE_COMMAND_R; + } else if (tmpl_contains("<|start_header_id|>") && tmpl_contains("<|end_header_id|>")) { + return LLM_CHAT_TEMPLATE_LLAMA_3; + } else if (tmpl_contains("[gMASK]sop")) { + // chatglm3-6b + return LLM_CHAT_TEMPLATE_CHATGML_3; + } else if (tmpl_contains("[gMASK]")) { + return LLM_CHAT_TEMPLATE_CHATGML_4; + } else if (tmpl_contains(LU8("<用户>"))) { + // MiniCPM-3B-OpenHermes-2.5-v2-GGUF + return LLM_CHAT_TEMPLATE_MINICPM; + } else if (tmpl_contains("'Assistant: ' + message['content'] + eos_token")) { + return LLM_CHAT_TEMPLATE_DEEPSEEK_2; + } else if (tmpl_contains(LU8("'<｜Assistant｜>' + message['content'] + '<｜end▁of▁sentence｜>'"))) { + return LLM_CHAT_TEMPLATE_DEEPSEEK_3; + } else if (tmpl_contains("[|system|]") && tmpl_contains("[|assistant|]") && tmpl_contains("[|endofturn|]")) { + // ref: https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct/discussions/8#66bae61b1893d14ee8ed85bb + // EXAONE-3.0-7.8B-Instruct + return LLM_CHAT_TEMPLATE_EXAONE_3; + } else if (tmpl_contains("rwkv-world")) { + return LLM_CHAT_TEMPLATE_RWKV_WORLD; + } else if (tmpl_contains("<|start_of_role|>")) { + return LLM_CHAT_TEMPLATE_GRANITE; + } else if (tmpl_contains("message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1]")) { + return LLM_CHAT_TEMPLATE_GIGACHAT; + } else if (tmpl_contains("<|role_start|>")) { + return LLM_CHAT_TEMPLATE_MEGREZ; + } + return LLM_CHAT_TEMPLATE_UNKNOWN; +} + +// Simple version of "llama_apply_chat_template" that only works with strings +// This function uses heuristic checks to determine commonly used template. It is not a jinja parser. +int32_t llm_chat_apply_template( + llm_chat_template tmpl, + const std::vector & chat, + std::string & dest, bool add_ass) { + // Taken from the research: https://github.com/ggerganov/llama.cpp/issues/5527 + std::stringstream ss; + if (tmpl == LLM_CHAT_TEMPLATE_CHATML) { + // chatml template + for (auto message : chat) { + ss << "<|im_start|>" << message->role << "\n" << message->content << "<|im_end|>\n"; + } + if (add_ass) { + ss << "<|im_start|>assistant\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V7) { + // Official mistral 'v7' template + // See: https://huggingface.co/mistralai/Mistral-Large-Instruct-2411#basic-instruct-template-v7 + for (auto message : chat) { + std::string role(message->role); + std::string content(message->content); + if (role == "system") { + ss << "[SYSTEM_PROMPT] " << content << "[/SYSTEM_PROMPT]"; + } else if (role == "user") { + ss << "[INST] " << content << "[/INST]"; + } + else { + ss << " " << content << ""; + } + } + } else if (tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V1 + || tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V3 + || tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN) { + // See: https://github.com/mistralai/cookbook/blob/main/concept-deep-dive/tokenization/chat_templates.md + // See: https://github.com/mistralai/cookbook/blob/main/concept-deep-dive/tokenization/templates.md + std::string leading_space = tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V1 ? " " : ""; + std::string trailing_space = tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN ? "" : " "; + bool trim_assistant_message = tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V3; + bool is_inside_turn = false; + for (auto message : chat) { + if (!is_inside_turn) { + ss << leading_space << "[INST]" << trailing_space; + is_inside_turn = true; + } + std::string role(message->role); + std::string content(message->content); + if (role == "system") { + ss << content << "\n\n"; + } else if (role == "user") { + ss << content << leading_space << "[/INST]"; + } else { + ss << trailing_space << (trim_assistant_message ? trim(content) : content) << ""; + is_inside_turn = false; + } + } + } else if ( + tmpl == LLM_CHAT_TEMPLATE_LLAMA_2 + || tmpl == LLM_CHAT_TEMPLATE_LLAMA_2_SYS + || tmpl == LLM_CHAT_TEMPLATE_LLAMA_2_SYS_BOS + || tmpl == LLM_CHAT_TEMPLATE_LLAMA_2_SYS_STRIP) { + // llama2 template and its variants + // [variant] support system message + // See: https://huggingface.co/blog/llama2#how-to-prompt-llama-2 + bool support_system_message = tmpl != LLM_CHAT_TEMPLATE_LLAMA_2; + // [variant] add BOS inside history + bool add_bos_inside_history = tmpl == LLM_CHAT_TEMPLATE_LLAMA_2_SYS_BOS; + // [variant] trim spaces from the input message + bool strip_message = tmpl == LLM_CHAT_TEMPLATE_LLAMA_2_SYS_STRIP; + // construct the prompt + bool is_inside_turn = true; // skip BOS at the beginning + ss << "[INST] "; + for (auto message : chat) { + std::string content = strip_message ? trim(message->content) : message->content; + std::string role(message->role); + if (!is_inside_turn) { + is_inside_turn = true; + ss << (add_bos_inside_history ? "[INST] " : "[INST] "); + } + if (role == "system") { + if (support_system_message) { + ss << "<>\n" << content << "\n<>\n\n"; + } else { + // if the model does not support system message, we still include it in the first message, but without <> + ss << content << "\n"; + } + } else if (role == "user") { + ss << content << " [/INST]"; + } else { + ss << content << ""; + is_inside_turn = false; + } + } + } else if (tmpl == LLM_CHAT_TEMPLATE_PHI_3) { + // Phi 3 + for (auto message : chat) { + std::string role(message->role); + ss << "<|" << role << "|>\n" << message->content << "<|end|>\n"; + } + if (add_ass) { + ss << "<|assistant|>\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_PHI_4) { + // chatml template + for (auto message : chat) { + ss << "<|im_start|>" << message->role << "<|im_sep|>" << message->content << "<|im_end|>"; + } + if (add_ass) { + ss << "<|im_start|>assistant<|im_sep|>"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_FALCON_3) { + // Falcon 3 + for (auto message : chat) { + std::string role(message->role); + ss << "<|" << role << "|>\n" << message->content << "\n"; + } + if (add_ass) { + ss << "<|assistant|>\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_ZEPHYR) { + // zephyr template + for (auto message : chat) { + ss << "<|" << message->role << "|>" << "\n" << message->content << "<|endoftext|>\n"; + } + if (add_ass) { + ss << "<|assistant|>\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_MONARCH) { + // mlabonne/AlphaMonarch-7B template (the is included inside history) + for (auto message : chat) { + std::string bos = (message == chat.front()) ? "" : ""; // skip BOS for first message + ss << bos << message->role << "\n" << message->content << "\n"; + } + if (add_ass) { + ss << "assistant\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_GEMMA) { + // google/gemma-7b-it + std::string system_prompt = ""; + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + // there is no system message for gemma, but we will merge it with user prompt, so nothing is broken + system_prompt = trim(message->content); + continue; + } + // in gemma, "assistant" is "model" + role = role == "assistant" ? "model" : message->role; + ss << "" << role << "\n"; + if (!system_prompt.empty() && role != "model") { + ss << system_prompt << "\n\n"; + system_prompt = ""; + } + ss << trim(message->content) << "\n"; + } + if (add_ass) { + ss << "model\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_ORION) { + // OrionStarAI/Orion-14B-Chat + std::string system_prompt = ""; + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + // there is no system message support, we will merge it with user prompt + system_prompt = message->content; + continue; + } else if (role == "user") { + ss << "Human: "; + if (!system_prompt.empty()) { + ss << system_prompt << "\n\n"; + system_prompt = ""; + } + ss << message->content << "\n\nAssistant: "; + } else { + ss << message->content << ""; + } + } + } else if (tmpl == LLM_CHAT_TEMPLATE_OPENCHAT) { + // openchat/openchat-3.5-0106, + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << message->content << "<|end_of_turn|>"; + } else { + role[0] = toupper(role[0]); + ss << "GPT4 Correct " << role << ": " << message->content << "<|end_of_turn|>"; + } + } + if (add_ass) { + ss << "GPT4 Correct Assistant:"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_VICUNA || tmpl == LLM_CHAT_TEMPLATE_VICUNA_ORCA) { + // eachadea/vicuna-13b-1.1 (and Orca variant) + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + // Orca-Vicuna variant uses a system prefix + if (tmpl == LLM_CHAT_TEMPLATE_VICUNA_ORCA) { + ss << "SYSTEM: " << message->content << "\n"; + } else { + ss << message->content << "\n\n"; + } + } else if (role == "user") { + ss << "USER: " << message->content << "\n"; + } else if (role == "assistant") { + ss << "ASSISTANT: " << message->content << "\n"; + } + } + if (add_ass) { + ss << "ASSISTANT:"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_DEEPSEEK) { + // deepseek-ai/deepseek-coder-33b-instruct + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << message->content; + } else if (role == "user") { + ss << "### Instruction:\n" << message->content << "\n"; + } else if (role == "assistant") { + ss << "### Response:\n" << message->content << "\n<|EOT|>\n"; + } + } + if (add_ass) { + ss << "### Response:\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_COMMAND_R) { + // CohereForAI/c4ai-command-r-plus + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << "<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>"; + } else if (role == "user") { + ss << "<|START_OF_TURN_TOKEN|><|USER_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>"; + } else if (role == "assistant") { + ss << "<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>"; + } + } + if (add_ass) { + ss << "<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_LLAMA_3) { + // Llama 3 + for (auto message : chat) { + std::string role(message->role); + ss << "<|start_header_id|>" << role << "<|end_header_id|>\n\n" << trim(message->content) << "<|eot_id|>"; + } + if (add_ass) { + ss << "<|start_header_id|>assistant<|end_header_id|>\n\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_CHATGML_3) { + // chatglm3-6b + ss << "[gMASK]" << "sop"; + for (auto message : chat) { + std::string role(message->role); + ss << "<|" << role << "|>" << "\n " << message->content; + } + if (add_ass) { + ss << "<|assistant|>"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_CHATGML_4) { + ss << "[gMASK]" << ""; + for (auto message : chat) { + std::string role(message->role); + ss << "<|" << role << "|>" << "\n" << message->content; + } + if (add_ass) { + ss << "<|assistant|>"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_MINICPM) { + // MiniCPM-3B-OpenHermes-2.5-v2-GGUF + for (auto message : chat) { + std::string role(message->role); + if (role == "user") { + ss << LU8("<用户>"); + ss << trim(message->content); + ss << ""; + } else { + ss << trim(message->content); + } + } + } else if (tmpl == LLM_CHAT_TEMPLATE_DEEPSEEK_2) { + // DeepSeek-V2 + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << message->content << "\n\n"; + } else if (role == "user") { + ss << "User: " << message->content << "\n\n"; + } else if (role == "assistant") { + ss << "Assistant: " << message->content << LU8("<｜end▁of▁sentence｜>"); + } + } + if (add_ass) { + ss << "Assistant:"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_DEEPSEEK_3) { + // DeepSeek-V3 + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << message->content << "\n\n"; + } else if (role == "user") { + ss << LU8("<｜User｜>") << message->content; + } else if (role == "assistant") { + ss << LU8("<｜Assistant｜>") << message->content << LU8("<｜end▁of▁sentence｜>"); + } + } + if (add_ass) { + ss << LU8("<｜Assistant｜>"); + } + } else if (tmpl == LLM_CHAT_TEMPLATE_EXAONE_3) { + // ref: https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct/discussions/8#66bae61b1893d14ee8ed85bb + // EXAONE-3.0-7.8B-Instruct + for (auto message : chat) { + std::string role(message->role); + if (role == "system") { + ss << "[|system|]" << trim(message->content) << "[|endofturn|]\n"; + } else if (role == "user") { + ss << "[|user|]" << trim(message->content) << "\n"; + } else if (role == "assistant") { + ss << "[|assistant|]" << trim(message->content) << "[|endofturn|]\n"; + } + } + if (add_ass) { + ss << "[|assistant|]"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_RWKV_WORLD) { + // this template requires the model to have "\n\n" as EOT token + for (auto message : chat) { + std::string role(message->role); + if (role == "user") { + ss << "User: " << message->content << "\n\nAssistant:"; + } else { + ss << message->content << "\n\n"; + } + } + } else if (tmpl == LLM_CHAT_TEMPLATE_GRANITE) { + // IBM Granite template + for (const auto & message : chat) { + std::string role(message->role); + ss << "<|start_of_role|>" << role << "<|end_of_role|>"; + if (role == "assistant_tool_call") { + ss << "<|tool_call|>"; + } + ss << message->content << "<|end_of_text|>\n"; + } + if (add_ass) { + ss << "<|start_of_role|>assistant<|end_of_role|>\n"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_GIGACHAT) { + // GigaChat template + bool has_system = !chat.empty() && std::string(chat[0]->role) == "system"; + + // Handle system message if present + if (has_system) { + ss << "" << chat[0]->content << "<|message_sep|>"; + } else { + ss << ""; + } + + // Process remaining messages + for (size_t i = has_system ? 1 : 0; i < chat.size(); i++) { + std::string role(chat[i]->role); + if (role == "user") { + ss << "user<|role_sep|>" << chat[i]->content << "<|message_sep|>" + << "available functions<|role_sep|>[]<|message_sep|>"; + } else if (role == "assistant") { + ss << "assistant<|role_sep|>" << chat[i]->content << "<|message_sep|>"; + } + } + + // Add generation prompt if needed + if (add_ass) { + ss << "assistant<|role_sep|>"; + } + } else if (tmpl == LLM_CHAT_TEMPLATE_MEGREZ) { + // Megrez template + for (auto message : chat) { + std::string role(message->role); + ss << "<|role_start|>" << role << "<|role_end|>" << message->content << "<|turn_end|>"; + } + + if (add_ass) { + ss << "<|role_start|>assistant<|role_end|>"; + } + } else { + // template not supported + return -1; + } + dest = ss.str(); + return dest.size(); +} + +// public interface + +int32_t llama_chat_builtin_templates(const char ** output, size_t len) { + auto it = LLM_CHAT_TEMPLATES.begin(); + for (size_t i = 0; i < std::min(len, LLM_CHAT_TEMPLATES.size()); i++) { + output[i] = it->first.c_str(); + std::advance(it, 1); + } + return (int32_t) LLM_CHAT_TEMPLATES.size(); +} + diff --git a/src/llama-chat.h b/src/llama-chat.h new file mode 100644 index 000000000..3a4d07ce3 --- /dev/null +++ b/src/llama-chat.h @@ -0,0 +1,52 @@ +#pragma once + +#include +#include +#include + +enum llm_chat_template { + LLM_CHAT_TEMPLATE_CHATML, + LLM_CHAT_TEMPLATE_LLAMA_2, + LLM_CHAT_TEMPLATE_LLAMA_2_SYS, + LLM_CHAT_TEMPLATE_LLAMA_2_SYS_BOS, + LLM_CHAT_TEMPLATE_LLAMA_2_SYS_STRIP, + LLM_CHAT_TEMPLATE_MISTRAL_V1, + LLM_CHAT_TEMPLATE_MISTRAL_V3, + LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN, + LLM_CHAT_TEMPLATE_MISTRAL_V7, + LLM_CHAT_TEMPLATE_PHI_3, + LLM_CHAT_TEMPLATE_PHI_4, + LLM_CHAT_TEMPLATE_FALCON_3, + LLM_CHAT_TEMPLATE_ZEPHYR, + LLM_CHAT_TEMPLATE_MONARCH, + LLM_CHAT_TEMPLATE_GEMMA, + LLM_CHAT_TEMPLATE_ORION, + LLM_CHAT_TEMPLATE_OPENCHAT, + LLM_CHAT_TEMPLATE_VICUNA, + LLM_CHAT_TEMPLATE_VICUNA_ORCA, + LLM_CHAT_TEMPLATE_DEEPSEEK, + LLM_CHAT_TEMPLATE_DEEPSEEK_2, + LLM_CHAT_TEMPLATE_DEEPSEEK_3, + LLM_CHAT_TEMPLATE_COMMAND_R, + LLM_CHAT_TEMPLATE_LLAMA_3, + LLM_CHAT_TEMPLATE_CHATGML_3, + LLM_CHAT_TEMPLATE_CHATGML_4, + LLM_CHAT_TEMPLATE_MINICPM, + LLM_CHAT_TEMPLATE_EXAONE_3, + LLM_CHAT_TEMPLATE_RWKV_WORLD, + LLM_CHAT_TEMPLATE_GRANITE, + LLM_CHAT_TEMPLATE_GIGACHAT, + LLM_CHAT_TEMPLATE_MEGREZ, + LLM_CHAT_TEMPLATE_UNKNOWN, +}; + +struct llama_chat_message; + +llm_chat_template llm_chat_template_from_str(const std::string & name); + +llm_chat_template llm_chat_detect_template(const std::string & tmpl); + +int32_t llm_chat_apply_template( + llm_chat_template tmpl, + const std::vector & chat, + std::string & dest, bool add_ass); diff --git a/src/llama-context.cpp b/src/llama-context.cpp new file mode 100644 index 000000000..671d2a81a --- /dev/null +++ b/src/llama-context.cpp @@ -0,0 +1,1775 @@ +#include "llama-context.h" + +#include "llama-impl.h" +#include "llama-mmap.h" + +#include +#include +#include +#include + +void llama_set_k_shift(struct llama_context & lctx) { + const int64_t kv_size = lctx.kv_self.size; + + assert(ggml_backend_buffer_is_host(lctx.inp_K_shift->buffer)); + + int32_t * data = (int32_t *) lctx.inp_K_shift->data; + + for (int i = 0; i < kv_size; ++i) { + data[i] = lctx.kv_self.cells[i].delta; + } +} + +void llama_set_s_copy(struct llama_context & lctx) { + const int64_t kv_size = lctx.kv_self.size; + + assert(ggml_backend_buffer_is_host(lctx.inp_s_copy->buffer)); + + int32_t * data = (int32_t *) lctx.inp_s_copy->data; + + for (int i = 0; i < kv_size; ++i) { + data[i] = lctx.kv_self.cells[i].src; + } +} + +// llama input + +static int32_t llama_relative_position_bucket(llama_pos x, llama_pos y, uint64_t n_buckets, bool bidirectional) { + // TODO move to hparams if a T5 variant appears that uses a different value + const int64_t max_distance = 128; + + if (bidirectional) { + n_buckets >>= 1; + } + + const int64_t max_exact = n_buckets >> 1; + + int32_t relative_position = x - y; + int32_t relative_bucket = 0; + if (bidirectional) { + relative_bucket += (relative_position > 0) * n_buckets; + relative_position = abs(relative_position); + } else { + relative_position = -std::min(relative_position, 0); + } + int32_t relative_position_if_large = floorf(max_exact + logf(1.0 * relative_position / max_exact) * (n_buckets - max_exact) / log(1.0 * max_distance / max_exact)); + relative_position_if_large = std::min(relative_position_if_large, n_buckets - 1); + relative_bucket += (relative_position < max_exact ? relative_position : relative_position_if_large); + return relative_bucket; +} + +void llama_set_inputs(llama_context & lctx, const llama_ubatch & ubatch) { + // + // set input data + // + + const auto & hparams = lctx.model.hparams; + const auto & cparams = lctx.cparams; + const auto & kv_self = lctx.kv_self; + + if (ubatch.token) { + const int64_t n_tokens = ubatch.n_tokens; + + ggml_backend_tensor_set(lctx.inp_tokens, ubatch.token, 0, n_tokens*ggml_element_size(lctx.inp_tokens)); + } + + if (ubatch.embd) { + const int64_t n_embd = hparams.n_embd; + const int64_t n_tokens = ubatch.n_tokens; + + ggml_backend_tensor_set(lctx.inp_embd, ubatch.embd, 0, n_tokens*n_embd*ggml_element_size(lctx.inp_embd)); + } + + if (ubatch.pos && lctx.inp_pos) { + const int64_t n_tokens = ubatch.n_tokens; + auto n_pos = lctx.n_pos_per_token; + ggml_backend_tensor_set(lctx.inp_pos, ubatch.pos, 0, n_tokens*n_pos*ggml_element_size(lctx.inp_pos)); + } + + if (hparams.causal_attn || cparams.pooling_type == LLAMA_POOLING_TYPE_NONE) { + //GGML_ASSERT(lctx.inp_out_ids && "every model that can must skip unused outputs"); + + if (!lctx.inp_out_ids) { + LLAMA_LOG_WARN("%s: 'lctx.inp_out_ids' is not created\n", __func__); + } else { + const int64_t n_tokens = ubatch.n_tokens; + + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_out_ids->buffer)); + int32_t * data = (int32_t *) lctx.inp_out_ids->data; + + if (lctx.n_outputs == n_tokens) { + for (int i = 0; i < n_tokens; ++i) { + data[i] = i; + } + } else if (ubatch.output) { + int32_t n_outputs = 0; + for (int i = 0; i < n_tokens; ++i) { + if (ubatch.output[i]) { + data[n_outputs++] = i; + } + } + // the graph needs to have been passed the correct number of outputs + GGML_ASSERT(lctx.n_outputs == n_outputs); + } else if (lctx.n_outputs == 1) { + // only keep last output + data[0] = n_tokens - 1; + } else { + GGML_ASSERT(lctx.n_outputs == 0); + } + } + } + + GGML_ASSERT( + // (!a || b) is a logical implication (a -> b) + // !hparams.causal_attn -> !cparams.causal_attn + (hparams.causal_attn || !cparams.causal_attn) && + "causal attention is not supported by this model" + ); + + if (lctx.inp_KQ_mask || lctx.inp_KQ_mask_swa) { + // NOTE: hparams.causal_attn indicates the model is capable of generation and uses the kv cache. + if (cparams.causal_attn && !lctx.is_encoding) { + const int64_t n_kv = kv_self.n; + const int64_t n_tokens = ubatch.n_tokens; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + const int64_t n_seqs = ubatch.n_seqs; + + + float * data = nullptr; + float * data_swa = nullptr; + + if (lctx.inp_KQ_mask) { + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask->buffer)); + data = (float *) lctx.inp_KQ_mask->data; + } + + if (lctx.inp_KQ_mask_swa) { + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask_swa->buffer)); + data_swa = (float *) lctx.inp_KQ_mask_swa->data; + } + + // For causal attention, use only the previous KV cells + // of the correct sequence for each token of the ubatch. + // It's assumed that if a token in the batch has multiple sequences, they are equivalent. + for (int h = 0; h < 1; ++h) { + for (int s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + + for (int j = 0; j < n_seq_tokens; ++j) { + const llama_pos pos = ubatch.pos[s*n_seq_tokens + j]; + + for (int i = 0; i < n_kv; ++i) { + float f; + if (!kv_self.cells[i].has_seq_id(seq_id) || kv_self.cells[i].pos > pos) { + f = -INFINITY; + } else { + if (hparams.use_alibi) { + f = -std::abs(kv_self.cells[i].pos - pos); + } else { + f = 0.0f; + } + } + + if (data) { + data[h*(n_kv*n_tokens) + s*(n_kv*n_seq_tokens) + j*n_kv + i] = f; + } + + // may need to cut off old tokens for sliding window + if (data_swa) { + if (pos - kv_self.cells[i].pos >= (int32_t)hparams.n_swa) { + f = -INFINITY; + } + data_swa[h*(n_kv*n_tokens) + s*(n_kv*n_seq_tokens) + j*n_kv + i] = f; + } + } + } + } + + if (data) { + for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) { + for (int j = 0; j < n_kv; ++j) { + data[h*(n_kv*n_tokens) + i*n_kv + j] = -INFINITY; + } + } + } + + if (data_swa) { + for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) { + for (int j = 0; j < n_kv; ++j) { + data_swa[h*(n_kv*n_tokens) + i*n_kv + j] = -INFINITY; + } + } + } + } + } else { + const int64_t n_tokens = ubatch.n_tokens; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + const int64_t n_seqs = ubatch.n_seqs; + // when using kv cache, the mask needs to match the kv cache size + const int64_t n_stride = hparams.causal_attn && !lctx.is_encoding ? kv_self.n : n_tokens; + + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask->buffer)); + + float * data = (float *) lctx.inp_KQ_mask->data; + + for (int h = 0; h < 1; ++h) { + for (int s1 = 0; s1 < n_seqs; ++s1) { + const llama_seq_id seq_id = ubatch.seq_id[s1][0]; + + for (int j = 0; j < n_seq_tokens; ++j) { + const int32_t tj = s1*n_seq_tokens + j; + + for (int s0 = 0; s0 < n_seqs; ++s0) { + for (int i = 0; i < n_seq_tokens; ++i) { + const int32_t ti = s0*n_seq_tokens + i; + float f = -INFINITY; + + for (int s = 0; s < ubatch.n_seq_id[s0]; ++s) { + if (ubatch.seq_id[s0][s] == seq_id) { + if (hparams.use_alibi) { + f = -std::abs(ubatch.pos[ti] - ubatch.pos[tj]); + } else { + f = 0.0f; + } + break; + } + } + + data[h*(n_tokens*n_tokens) + tj*n_stride + ti] = f; + } + } + + for (int i = n_tokens; i < n_stride; ++i) { + data[h*(n_tokens*n_tokens) + tj*n_stride + i] = -INFINITY; + } + } + } + } + } + } + + if (cparams.embeddings && cparams.pooling_type == LLAMA_POOLING_TYPE_MEAN) { + const int64_t n_tokens = ubatch.n_tokens; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + const int64_t n_seqs = ubatch.n_seqs; + + GGML_ASSERT(lctx.inp_mean); + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_mean->buffer)); + + float * data = (float *) lctx.inp_mean->data; + memset(lctx.inp_mean->data, 0, n_tokens * n_tokens * ggml_element_size(lctx.inp_mean)); + + std::vector sum(n_tokens, 0); + + for (int s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + + // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true + GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == MEAN"); + + sum[seq_id] += ubatch.n_seq_tokens; + } + + std::vector div(n_tokens, 0.0f); + for (int i = 0; i < n_tokens; ++i) { + const uint64_t s = sum[i]; + if (s > 0) { + div[i] = 1.0f/float(s); + } + } + + for (int s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + + for (int i = 0; i < n_seq_tokens; ++i) { + data[seq_id*n_tokens + s*n_seq_tokens + i] = div[seq_id]; + } + } + } + + if (cparams.embeddings && ( + cparams.pooling_type == LLAMA_POOLING_TYPE_CLS || + cparams.pooling_type == LLAMA_POOLING_TYPE_RANK)) { + const int64_t n_tokens = ubatch.n_tokens; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + const int64_t n_seqs = ubatch.n_seqs; + + GGML_ASSERT(lctx.inp_cls); + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_cls->buffer)); + + uint32_t * data = (uint32_t *) lctx.inp_cls->data; + memset(lctx.inp_cls->data, 0, n_tokens * ggml_element_size(lctx.inp_cls)); + + for (int s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + + // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true + GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == CLS or RANK"); + + for (int i = 0; i < n_seq_tokens; ++i) { + const llama_pos pos = ubatch.pos[s*n_seq_tokens + i]; + + if (pos == 0) { + data[seq_id] = s*n_seq_tokens + i; + } + } + } + } + + if (cparams.embeddings && cparams.pooling_type == LLAMA_POOLING_TYPE_LAST) { + const int64_t n_tokens = ubatch.n_tokens; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + const int64_t n_seqs = ubatch.n_seqs; + + GGML_ASSERT(lctx.inp_cls); + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_cls->buffer)); + + uint32_t * data = (uint32_t *) lctx.inp_cls->data; + memset(lctx.inp_cls->data, 0, n_tokens * ggml_element_size(lctx.inp_cls)); + + std::vector last_pos(n_tokens, -1); + std::vector last_row(n_tokens, -1); + + for (int s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + + // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true + GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == LAST"); + + for (int i = 0; i < n_seq_tokens; ++i) { + const llama_pos pos = ubatch.pos[s*n_seq_tokens + i]; + + if (pos >= last_pos[seq_id]) { + last_pos[seq_id] = pos; + last_row[seq_id] = s*n_seq_tokens + i; + } + } + } + + for (int i = 0; i < n_tokens; ++i) { + if (last_row[i] >= 0) { + data[i] = last_row[i]; + } + } + } + + if (kv_self.recurrent) { + const int64_t n_kv = kv_self.n; + + if (lctx.inp_s_mask) { + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_s_mask->buffer)); + float * data = (float *) lctx.inp_s_mask->data; + + // clear unused states + for (int i = 0; i < n_kv; ++i) { + const uint32_t cell_id = i + kv_self.head; + llama_kv_cell & kv_cell = lctx.kv_self.cells[cell_id]; + + data[i] = (float) (kv_cell.src >= 0); + + // only clear once + if (kv_cell.src < 0) { + kv_cell.src = cell_id; + } + } + } + + if (lctx.inp_s_copy) { + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_s_copy->buffer)); + int32_t * data = (int32_t *) lctx.inp_s_copy->data; + + // assuming copy destinations ALWAYS happen ONLY on the cells between head and head+n + for (uint32_t i = 0; i < n_kv; ++i) { + const uint32_t cell_id = i + kv_self.head; + llama_kv_cell & kv_cell = lctx.kv_self.cells[cell_id]; + + // prevent out-of-bound sources + if (kv_cell.src < 0 || (uint32_t) kv_cell.src >= kv_self.size) { + kv_cell.src = cell_id; + } + + data[i] = kv_cell.src; + + // ensure copy only happens once + if (kv_cell.src != (int32_t) cell_id) { + kv_cell.src = cell_id; + } + } + } + } + + if (lctx.inp_pos_bucket) { + const int64_t n_tokens = ubatch.n_tokens; + + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_pos_bucket->buffer)); + GGML_ASSERT(!ubatch.equal_seqs); // TODO: use ubatch.n_seqs instead of failing + + int32_t * data = (int32_t *) lctx.inp_pos_bucket->data; + + if (!lctx.is_encoding) { + const int64_t n_kv = kv_self.n; + for (int h = 0; h < 1; ++h) { + for (int j = 0; j < n_tokens; ++j) { + for (int i = 0; i < n_kv; ++i) { + data[h*(n_kv*n_tokens) + j*n_kv + i] = llama_relative_position_bucket(lctx.kv_self.cells[i].pos, ubatch.pos[j], hparams.n_rel_attn_bkts, lctx.is_encoding); + } + } + } + } else { + for (int h = 0; h < 1; ++h) { + for (int j = 0; j < n_tokens; ++j) { + for (int i = 0; i < n_tokens; ++i) { + data[h*(n_tokens*n_tokens) + j*n_tokens + i] = llama_relative_position_bucket(ubatch.pos[i], ubatch.pos[j], hparams.n_rel_attn_bkts, lctx.is_encoding); + } + } + } + } + } + + if (!lctx.is_encoding && lctx.inp_embd_enc) { + assert(lctx.inp_embd_enc->type == GGML_TYPE_F32); + assert((size_t) ggml_nelements(lctx.inp_embd_enc) == lctx.embd_enc.size()); + + ggml_backend_tensor_set(lctx.inp_embd_enc, lctx.embd_enc.data(), 0, ggml_nbytes(lctx.inp_embd_enc)); + } + + if (!lctx.is_encoding && lctx.inp_KQ_mask_cross) { + const int64_t n_output_enc = lctx.embd_enc.size() / hparams.n_embd; + const int64_t n_tokens = ubatch.n_tokens; + + GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask_cross->buffer)); + GGML_ASSERT(!ubatch.equal_seqs); // TODO: use ubatch.n_seqs instead of failing + + float * data = (float *) lctx.inp_KQ_mask_cross->data; + + for (int h = 0; h < 1; ++h) { + for (int j = 0; j < n_tokens; ++j) { + for (int i = 0; i < n_output_enc; ++i) { + float f = -INFINITY; + for (int s = 0; s < ubatch.n_seq_id[j]; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[j][s]; + if (lctx.seq_ids_enc[i].find(seq_id) != lctx.seq_ids_enc[i].end()) { + f = 0.0f; + } + } + data[h*(n_output_enc*n_tokens) + j*n_output_enc + i] = f; + } + } + + for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) { + for (int j = 0; j < n_output_enc; ++j) { + data[h*(n_output_enc*n_tokens) + i*n_output_enc + j] = -INFINITY; + } + } + } + } +} + +// llama output + +size_t llama_output_reserve(struct llama_context & lctx, size_t n_outputs) { + const auto & cparams = lctx.cparams; + const auto & hparams = lctx.model.hparams; + const auto & vocab = lctx.model.vocab; + + const size_t n_outputs_max = std::max(n_outputs, (size_t) cparams.n_seq_max); + + const auto n_batch = cparams.n_batch; + const auto n_vocab = vocab.n_tokens(); + const auto n_embd = hparams.n_embd; + + // TODO: use a per-batch flag for logits presence instead + const bool has_logits = !cparams.embeddings; + const bool has_embd = cparams.embeddings && (cparams.pooling_type == LLAMA_POOLING_TYPE_NONE); + + const size_t logits_size = has_logits ? n_vocab*n_outputs_max : 0; + const size_t embd_size = has_embd ? n_embd*n_outputs_max : 0; + + if (lctx.output_ids.empty()) { + // init, never resized afterwards + lctx.output_ids.resize(n_batch); + } + + const size_t prev_size = lctx.buf_output ? ggml_backend_buffer_get_size(lctx.buf_output.get()) : 0; + const size_t new_size = (logits_size + embd_size) * sizeof(float); + + // alloc only when more than the current capacity is required + // TODO: also consider shrinking the buffer + if (!lctx.buf_output || prev_size < new_size) { + if (lctx.buf_output) { +#ifndef NDEBUG + // This doesn't happen often, but may be annoying in some cases (like the HellaSwag benchmark) + LLAMA_LOG_INFO("%s: reallocating output buffer from size %.02f MiB to %.02f MiB\n", __func__, prev_size / 1024.0 / 1024.0, new_size / 1024.0 / 1024.0); +#endif + lctx.buf_output = nullptr; + lctx.logits = nullptr; + lctx.embd = nullptr; + } + + auto * buft = ggml_backend_cpu_buffer_type(); + // try to use the host buffer of the device where the output tensor is allocated for faster transfer to system memory + auto * output_dev = lctx.model.dev_output(); + auto * output_dev_host_buft = output_dev ? ggml_backend_dev_host_buffer_type(output_dev) : nullptr; + if (output_dev_host_buft) { + buft = output_dev_host_buft; + } + lctx.buf_output.reset(ggml_backend_buft_alloc_buffer(buft, new_size)); + if (lctx.buf_output == nullptr) { + LLAMA_LOG_ERROR("%s: failed to allocate output buffer of size %.2f MiB\n", __func__, new_size / (1024.0 * 1024.0)); + return 0; + } + } + + float * output_base = (float *) ggml_backend_buffer_get_base(lctx.buf_output.get()); + + lctx.logits = has_logits ? output_base : nullptr; + lctx.embd = has_embd ? output_base + logits_size : nullptr; + + lctx.output_size = n_outputs_max; + lctx.logits_size = logits_size; + lctx.embd_size = embd_size; + + // set all ids as invalid (negative) + std::fill(lctx.output_ids.begin(), lctx.output_ids.end(), -1); + + ggml_backend_buffer_clear(lctx.buf_output.get(), 0); + + lctx.n_outputs = 0; + + return n_outputs_max; +} + +void llama_output_reorder(struct llama_context & ctx) { + std::vector & out_ids = ctx.sbatch.out_ids; + if (!out_ids.empty()) { + const uint32_t n_vocab = ctx.model.vocab.n_tokens(); + const uint32_t n_embd = ctx.model.hparams.n_embd; + + const int32_t n_outputs = ctx.n_outputs; + GGML_ASSERT((size_t) n_outputs == out_ids.size()); + + // TODO: is there something more efficient which also minimizes swaps? + // selection sort, to minimize swaps (from https://en.wikipedia.org/wiki/Selection_sort) + for (int32_t i = 0; i < n_outputs - 1; ++i) { + int32_t j_min = i; + for (int32_t j = i + 1; j < n_outputs; ++j) { + if (out_ids[j] < out_ids[j_min]) { + j_min = j; + } + } + if (j_min == i) { continue; } + std::swap(out_ids[i], out_ids[j_min]); + if (ctx.logits_size > 0) { + for (uint32_t k = 0; k < n_vocab; k++) { + std::swap(ctx.logits[i*n_vocab + k], ctx.logits[j_min*n_vocab + k]); + } + } + if (ctx.embd_size > 0) { + for (uint32_t k = 0; k < n_embd; k++) { + std::swap(ctx.embd[i*n_embd + k], ctx.embd[j_min*n_embd + k]); + } + } + } + std::fill(ctx.output_ids.begin(), ctx.output_ids.end(), -1); + for (int32_t i = 0; i < n_outputs; ++i) { + ctx.output_ids[out_ids[i]] = i; + } + out_ids.clear(); + } +} + +// +// interface implementation +// + +void llama_free(struct llama_context * ctx) { + delete ctx; +} + +uint32_t llama_n_ctx(const struct llama_context * ctx) { + return ctx->cparams.n_ctx; +} + +uint32_t llama_n_batch(const struct llama_context * ctx) { + return ctx->cparams.n_batch; +} + +uint32_t llama_n_ubatch(const struct llama_context * ctx) { + return ctx->cparams.n_ubatch; +} + +uint32_t llama_n_seq_max(const struct llama_context * ctx) { + return ctx->kv_self.size; +} + +const struct llama_model * llama_get_model(const struct llama_context * ctx) { + return &ctx->model; +} + +enum llama_pooling_type llama_pooling_type(const struct llama_context * ctx) { + return ctx->cparams.pooling_type; +} + +void llama_attach_threadpool( + struct llama_context * ctx, + ggml_threadpool_t threadpool, + ggml_threadpool_t threadpool_batch) { + ctx->threadpool = threadpool; + ctx->threadpool_batch = threadpool_batch ? threadpool_batch : threadpool; +} + +void llama_detach_threadpool(struct llama_context * ctx) { + ctx->threadpool = nullptr; + ctx->threadpool_batch = nullptr; +} + +void llama_set_n_threads(struct llama_context * ctx, int32_t n_threads, int32_t n_threads_batch) { + ctx->cparams.n_threads = n_threads; + ctx->cparams.n_threads_batch = n_threads_batch; +} + +int32_t llama_n_threads(struct llama_context * ctx) { + return ctx->cparams.n_threads; +} + +int32_t llama_n_threads_batch(struct llama_context * ctx) { + return ctx->cparams.n_threads_batch; +} + +void llama_set_abort_callback(struct llama_context * ctx, bool (*abort_callback)(void * data), void * abort_callback_data) { + ctx->abort_callback = abort_callback; + ctx->abort_callback_data = abort_callback_data; + + for (auto & backend : ctx->backends) { + auto * reg = ggml_backend_dev_backend_reg(ggml_backend_get_device(backend.get())); + auto * set_abort_callback_fn = (ggml_backend_set_abort_callback_t) ggml_backend_reg_get_proc_address(reg, "ggml_backend_set_abort_callback"); + if (set_abort_callback_fn) { + set_abort_callback_fn(backend.get(), ctx->abort_callback, ctx->abort_callback_data); + } + } +} + +void llama_set_embeddings(struct llama_context * ctx, bool embeddings) { + ctx->cparams.embeddings = embeddings; +} + +void llama_set_causal_attn(struct llama_context * ctx, bool causal_attn) { + ctx->cparams.causal_attn = causal_attn; +} + +void llama_synchronize(struct llama_context * ctx) { + ggml_backend_sched_synchronize(ctx->sched.get()); + + // FIXME: if multiple single tokens are evaluated without a synchronization, + // the stats will be added to the prompt evaluation stats + // this should only happen when using batch size 1 to evaluate a batch + + // add the evaluation to the stats + if (ctx->n_queued_tokens == 1) { + if (!ctx->cparams.no_perf) { + ctx->t_eval_us += ggml_time_us() - ctx->t_compute_start_us; + } + ctx->n_eval++; + } else if (ctx->n_queued_tokens > 1) { + if (!ctx->cparams.no_perf) { + ctx->t_p_eval_us += ggml_time_us() - ctx->t_compute_start_us; + } + ctx->n_p_eval += ctx->n_queued_tokens; + } + + // get a more accurate load time, upon first eval + if (ctx->n_queued_tokens > 0 && !ctx->has_evaluated_once) { + ctx->t_load_us = ggml_time_us() - ctx->t_start_us; + ctx->has_evaluated_once = true; + } + + ctx->n_queued_tokens = 0; + ctx->t_compute_start_us = 0; +} + +float * llama_get_logits(struct llama_context * ctx) { + llama_synchronize(ctx); + + // reorder logits for backward compatibility + // TODO: maybe deprecate this + llama_output_reorder(*ctx); + + return ctx->logits; +} + +float * llama_get_logits_ith(struct llama_context * ctx, int32_t i) { + int32_t j = -1; + + llama_synchronize(ctx); + + try { + if (ctx->logits == nullptr) { + throw std::runtime_error("no logits"); + } + + if (i < 0) { + j = ctx->n_outputs + i; + if (j < 0) { + throw std::runtime_error(format("negative index out of range [0, %d)", ctx->n_outputs)); + } + } else if ((size_t) i >= ctx->output_ids.size()) { + throw std::runtime_error(format("out of range [0, %zu)", ctx->output_ids.size())); + } else { + j = ctx->output_ids[i]; + } + + if (j < 0) { + throw std::runtime_error(format("batch.logits[%d] != true", i)); + } + if (j >= ctx->n_outputs) { + // This should not happen + throw std::runtime_error(format("corrupt output buffer (j=%d, n_outputs=%d)", j, ctx->n_outputs)); + } + + return ctx->logits + j*ctx->model.vocab.n_tokens(); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: invalid logits id %d, reason: %s\n", __func__, i, err.what()); +#ifndef NDEBUG + GGML_ABORT("fatal error"); +#else + return nullptr; +#endif + } +} + +float * llama_get_embeddings(struct llama_context * ctx) { + llama_synchronize(ctx); + + // reorder embeddings for backward compatibility + // TODO: maybe deprecate this + llama_output_reorder(*ctx); + + return ctx->embd; +} + +float * llama_get_embeddings_ith(struct llama_context * ctx, int32_t i) { + int32_t j = -1; + + llama_synchronize(ctx); + + try { + if (ctx->embd == nullptr) { + throw std::runtime_error("no embeddings"); + } + + if (i < 0) { + j = ctx->n_outputs + i; + if (j < 0) { + throw std::runtime_error(format("negative index out of range [0, %d)", ctx->n_outputs)); + } + } else if ((size_t) i >= ctx->output_ids.size()) { + throw std::runtime_error(format("out of range [0, %zu)", ctx->output_ids.size())); + } else { + j = ctx->output_ids[i]; + } + + if (j < 0) { + throw std::runtime_error(format("batch.logits[%d] != true", i)); + } + if (j >= ctx->n_outputs) { + // This should not happen + throw std::runtime_error(format("corrupt output buffer (j=%d, n_outputs=%d)", j, ctx->n_outputs)); + } + + return ctx->embd + j*ctx->model.hparams.n_embd; + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: invalid embeddings id %d, reason: %s\n", __func__, i, err.what()); +#ifndef NDEBUG + GGML_ABORT("fatal error"); +#else + return nullptr; +#endif + } +} + +float * llama_get_embeddings_seq(struct llama_context * ctx, llama_seq_id seq_id) { + llama_synchronize(ctx); + + auto it = ctx->embd_seq.find(seq_id); + if (it == ctx->embd_seq.end()) { + return nullptr; + } + + return it->second.data(); +} + +// llama state API + +// deprecated +size_t llama_get_state_size(struct llama_context * ctx) { + return llama_state_get_size(ctx); +} + +// deprecated +size_t llama_copy_state_data(struct llama_context * ctx, uint8_t * dst) { + return llama_state_get_data(ctx, dst, -1); +} + +// deprecated +size_t llama_set_state_data(struct llama_context * ctx, const uint8_t * src) { + return llama_state_set_data(ctx, src, -1); +} + +// deprecated +bool llama_load_session_file(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) { + return llama_state_load_file(ctx, path_session, tokens_out, n_token_capacity, n_token_count_out); +} + +// deprecated +bool llama_save_session_file(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) { + return llama_state_save_file(ctx, path_session, tokens, n_token_count); +} + +// TODO: replace all non-fatal assertions with returned errors or exceptions +struct llama_data_write { + virtual void write(const void * src, size_t size) = 0; + virtual void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) = 0; + virtual size_t get_size_written() = 0; + virtual ~llama_data_write() = default; + + void write_string(const std::string & str) { + uint32_t str_size = str.size(); + + write(&str_size, sizeof(str_size)); + write(str.data(), str_size); + } + + void write_model_info(const struct llama_context * ctx) { + const std::string arch_str = llm_arch_name(ctx->model.arch); + write_string(arch_str); + // TODO: add more model-specific info which should prevent loading the session file if not identical + } + + //void write_rng(const std::mt19937 & rng) { + // std::ostringstream rng_ss; + // rng_ss << rng; + + // const std::string & rng_str = rng_ss.str(); + + // write_string(rng_str); + //} + + void write_output_ids(struct llama_context * ctx) { + llama_output_reorder(*ctx); + + const uint32_t n_outputs = ctx->n_outputs; + + std::vector output_pos; + + const size_t n_batch = ctx->cparams.n_batch; + const auto & output_ids = ctx->output_ids; + + GGML_ASSERT(n_outputs <= ctx->output_size); + + output_pos.resize(n_outputs); + + // build a more compact representation of the output ids + for (size_t i = 0; i < n_batch; ++i) { + // map an output id to a position in the batch + int32_t pos = output_ids[i]; + if (pos >= 0) { + GGML_ASSERT((uint32_t) pos < n_outputs); + output_pos[pos] = i; + } + } + + write(&n_outputs, sizeof(n_outputs)); + + if (n_outputs) { + write(output_pos.data(), n_outputs * sizeof(int32_t)); + } + } + + void write_logits(const struct llama_context * ctx) { + const uint64_t logits_size = std::min((uint64_t) ctx->logits_size, (uint64_t) ctx->n_outputs * ctx->model.vocab.n_tokens()); + + write(&logits_size, sizeof(logits_size)); + + if (logits_size) { + write(ctx->logits, logits_size * sizeof(float)); + } + } + + void write_embeddings(const struct llama_context * ctx) { + const uint64_t embeddings_size = std::min((uint64_t) ctx->embd_size, (uint64_t) ctx->n_outputs * ctx->model.hparams.n_embd); + + write(&embeddings_size, sizeof(embeddings_size)); + + if (embeddings_size) { + write(ctx->embd, embeddings_size * sizeof(float)); + } + } + + void write_kv_cache_meta(const llama_kv_cache & kv_self, const std::vector> & cell_ranges, llama_seq_id seq_id = -1) { + for (const auto & range : cell_ranges) { + for (uint32_t i = range.first; i < range.second; ++i) { + const auto & cell = kv_self.cells[i]; + const llama_pos pos = cell.pos; + const uint32_t n_seq_id = seq_id == -1 ? cell.seq_id.size() : 0; + + write(&pos, sizeof(pos)); + write(&n_seq_id, sizeof(n_seq_id)); + + if (n_seq_id) { + for (auto seq_id : cell.seq_id) { + write(&seq_id, sizeof(seq_id)); + } + } + } + } + } + + void write_kv_cache_data(const struct llama_context * ctx, const std::vector> & cell_ranges) { + const struct llama_kv_cache & kv_self = ctx->kv_self; + const struct llama_hparams & hparams = ctx->model.hparams; + + const uint32_t v_trans = kv_self.v_trans ? 1 : 0; + const uint32_t n_layer = hparams.n_layer; + + write(&v_trans, sizeof(v_trans)); + write(&n_layer, sizeof(n_layer)); + + std::vector tmp_buf; + + // Iterate and write all the keys first, each row is a cell + // Get whole range at a time + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(il) + hparams.n_embd_k_s(); + + // Write key type + const int32_t k_type_i = (int32_t)kv_self.k_l[il]->type; + write(&k_type_i, sizeof(k_type_i)); + + // Write row size of key + const uint64_t k_size_row = ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa); + write(&k_size_row, sizeof(k_size_row)); + + // Read each range of cells of k_size length each into tmp_buf and write out + for (const auto & range : cell_ranges) { + const size_t range_size = range.second - range.first; + const size_t buf_size = range_size * k_size_row; + write_tensor_data(kv_self.k_l[il], range.first * k_size_row, buf_size); + } + } + + if (!kv_self.v_trans) { + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s(); + + // Write value type + const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type; + write(&v_type_i, sizeof(v_type_i)); + + // Write row size of value + const uint64_t v_size_row = ggml_row_size(kv_self.v_l[il]->type, n_embd_v_gqa); + write(&v_size_row, sizeof(v_size_row)); + + // Read each range of cells of v_size length each into tmp_buf and write out + for (const auto & range : cell_ranges) { + const size_t range_size = range.second - range.first; + const size_t buf_size = range_size * v_size_row; + write_tensor_data(kv_self.v_l[il], range.first * v_size_row, buf_size); + } + } + } else { + // When v is transposed, we also need the element size and get the element ranges from each row + const uint32_t kv_size = kv_self.size; + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s(); + + // Write value type + const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type; + write(&v_type_i, sizeof(v_type_i)); + + // Write element size + const uint32_t v_size_el = ggml_type_size(kv_self.v_l[il]->type); + write(&v_size_el, sizeof(v_size_el)); + + // Write GQA embedding size + write(&n_embd_v_gqa, sizeof(n_embd_v_gqa)); + + // For each row, we get the element values of each cell + for (uint32_t j = 0; j < n_embd_v_gqa; ++j) { + // Read each range of cells of v_size_el length each into tmp_buf and write out + for (const auto & range : cell_ranges) { + const size_t range_size = range.second - range.first; + const size_t src_offset = (range.first + j * kv_size) * v_size_el; + const size_t buf_size = range_size * v_size_el; + write_tensor_data(kv_self.v_l[il], src_offset, buf_size); + } + } + } + } + } + + void write_kv_cache(const struct llama_context * ctx, llama_seq_id seq_id = -1) { + const struct llama_kv_cache & kv_self = ctx->kv_self; + std::vector> cell_ranges; // ranges, from inclusive, to exclusive + uint32_t cell_count = 0; + + // Count the number of cells with the specified seq_id + // Find all the ranges of cells with this seq id (or all, when -1) + uint32_t cell_range_begin = kv_self.size; + for (uint32_t i = 0; i < kv_self.size; ++i) { + const auto & cell = kv_self.cells[i]; + if ((seq_id == -1 && !cell.is_empty()) || cell.has_seq_id(seq_id)) { + ++cell_count; + if (cell_range_begin == kv_self.size) { + cell_range_begin = i; + } + } else { + if (cell_range_begin != kv_self.size) { + cell_ranges.emplace_back(cell_range_begin, i); + cell_range_begin = kv_self.size; + } + } + } + if (cell_range_begin != kv_self.size) { + cell_ranges.emplace_back(cell_range_begin, kv_self.size); + } + + // DEBUG CHECK: Sum of cell counts in ranges should equal the total cell count + uint32_t cell_count_check = 0; + for (const auto & range : cell_ranges) { + cell_count_check += range.second - range.first; + } + GGML_ASSERT(cell_count == cell_count_check); + + write(&cell_count, sizeof(cell_count)); + + write_kv_cache_meta(kv_self, cell_ranges, seq_id); + write_kv_cache_data(ctx, cell_ranges); + } +}; + +struct llama_data_read { + virtual const uint8_t * read(size_t size) = 0; + virtual void read_to(void * dst, size_t size) = 0; + virtual size_t get_size_read() = 0; + virtual ~llama_data_read() = default; + + void read_string(std::string & str) { + uint32_t str_size; + read_to(&str_size, sizeof(str_size)); + + str.assign((const char *) read(str_size), str_size); + } + + // validate model information + void read_model_info(const struct llama_context * ctx) { + const std::string cur_arch_str = llm_arch_name(ctx->model.arch); + + std::string arch_str; + read_string(arch_str); + if (cur_arch_str != arch_str) { + throw std::runtime_error(format("wrong model arch: '%s' instead of '%s'", arch_str.c_str(), cur_arch_str.c_str())); + } + // TODO: add more info which needs to be identical but which is not verified otherwise + } + + //void read_rng(std::mt19937 & rng) { + // std::string rng_str; + // read_string(rng_str); + + // std::istringstream rng_ss(rng_str); + // rng_ss >> rng; + + // if (rng_ss.fail()) { + // throw std::runtime_error("failed to load RNG state"); + // } + //} + + void read_output_ids(struct llama_context * ctx) { + std::vector output_pos; + + uint32_t n_outputs; + read_to(&n_outputs, sizeof(n_outputs)); + + if (n_outputs > llama_output_reserve(*ctx, n_outputs)) { + throw std::runtime_error("could not reserve outputs"); + } + + if (n_outputs) { + output_pos.resize(n_outputs); + read_to(output_pos.data(), n_outputs * sizeof(int32_t)); + + for (int32_t i = 0; i < (int32_t) output_pos.size(); ++i) { + int32_t id = output_pos[i]; + if ((uint32_t) id >= ctx->cparams.n_batch) { + throw std::runtime_error(format("invalid output id, %d does not fit in batch size of %u", id, ctx->cparams.n_batch)); + } + ctx->output_ids[id] = i; + } + + ctx->n_outputs = n_outputs; + } + } + + void read_logits(struct llama_context * ctx) { + uint64_t logits_size; + read_to(&logits_size, sizeof(logits_size)); + + if (ctx->logits_size < logits_size) { + throw std::runtime_error("logits buffer too small"); + } + + if (logits_size) { + read_to(ctx->logits, logits_size * sizeof(float)); + } + } + + void read_embeddings(struct llama_context * ctx) { + uint64_t embeddings_size; + read_to(&embeddings_size, sizeof(embeddings_size)); + + if (ctx->embd_size < embeddings_size) { + throw std::runtime_error("embeddings buffer too small"); + } + + if (embeddings_size) { + read_to(ctx->embd, embeddings_size * sizeof(float)); + } + } + + bool read_kv_cache_meta(struct llama_context * ctx, uint32_t cell_count, llama_seq_id dest_seq_id = -1) { + struct llama_kv_cache & kv_self = ctx->kv_self; + + if (dest_seq_id != -1) { + // single sequence + + llama_kv_cache_seq_rm(kv_self, dest_seq_id, -1, -1); + + llama_ubatch batch = ctx->sbatch.reserve_ubatch(cell_count, /* has_embd */ false); + batch.n_tokens = cell_count; + batch.n_seq_tokens = cell_count; + batch.n_seqs = 1; + + for (uint32_t i = 0; i < cell_count; ++i) { + llama_pos pos; + uint32_t n_seq_id; + + read_to(&pos, sizeof(pos)); + read_to(&n_seq_id, sizeof(n_seq_id)); + + if (n_seq_id != 0) { + LLAMA_LOG_ERROR("%s: invalid seq_id-agnostic kv cell\n", __func__); + return false; + } + + batch.pos[i] = pos; + } + batch.n_seq_id[0] = 1; + batch.seq_id[0] = &dest_seq_id; + if (!llama_kv_cache_find_slot(kv_self, batch)) { + LLAMA_LOG_ERROR("%s: failed to find available cells in kv cache\n", __func__); + return false; + } + + // DEBUG CHECK: kv_self.head should be our first cell, kv_self.head + cell_count - 1 should be our last cell (verify seq_id and pos values) + // Assume that this is one contiguous block of cells + GGML_ASSERT(kv_self.head + cell_count <= kv_self.size); + GGML_ASSERT(kv_self.cells[kv_self.head].pos == batch.pos[0]); + GGML_ASSERT(kv_self.cells[kv_self.head + cell_count - 1].pos == batch.pos[cell_count - 1]); + GGML_ASSERT(kv_self.cells[kv_self.head].has_seq_id(dest_seq_id)); + GGML_ASSERT(kv_self.cells[kv_self.head + cell_count - 1].has_seq_id(dest_seq_id)); + } else { + // whole KV cache restore + + if (cell_count > kv_self.size) { + LLAMA_LOG_ERROR("%s: not enough cells in kv cache\n", __func__); + return false; + } + + llama_kv_cache_clear(kv_self); + + for (uint32_t i = 0; i < cell_count; ++i) { + llama_kv_cell & cell = kv_self.cells[i]; + + llama_pos pos; + uint32_t n_seq_id; + + read_to(&pos, sizeof(pos)); + read_to(&n_seq_id, sizeof(n_seq_id)); + + cell.pos = pos; + + for (uint32_t j = 0; j < n_seq_id; ++j) { + llama_seq_id seq_id; + read_to(&seq_id, sizeof(seq_id)); + + if (seq_id < 0 || (uint32_t) seq_id >= llama_n_seq_max(ctx)) { + LLAMA_LOG_ERROR("%s: invalid seq_id, %d is out of range [0, %u)\n", __func__, seq_id, llama_n_seq_max(ctx)); + return false; + } + + cell.seq_id.insert(seq_id); + + if (kv_self.recurrent) { + int32_t & tail = kv_self.cells[seq_id].tail; + if (tail != -1) { + LLAMA_LOG_ERROR("%s: duplicate tail for seq_id %d in cell %d and %d\n", __func__, seq_id, i, tail); + return false; + } + tail = i; + } + } + } + + kv_self.head = 0; + kv_self.used = cell_count; + } + + if (kv_self.recurrent) { + for (uint32_t i = 0; i < cell_count; ++i) { + uint32_t cell_id = kv_self.head + i; + // make sure the recurrent states will keep their restored state + kv_self.cells[cell_id].src = cell_id; + } + } + + return true; + } + + bool read_kv_cache_data(struct llama_context * ctx, uint32_t cell_count) { + const struct llama_hparams & hparams = ctx->model.hparams; + struct llama_kv_cache & kv_self = ctx->kv_self; + uint32_t v_trans; + uint32_t n_layer; + read_to(&v_trans, sizeof(v_trans)); + read_to(&n_layer, sizeof(n_layer)); + + if (n_layer != hparams.n_layer) { + LLAMA_LOG_ERROR("%s: mismatched layer count (%u instead of %u)\n", __func__, n_layer, hparams.n_layer); + return false; + } + if (cell_count > kv_self.size) { + LLAMA_LOG_ERROR("%s: not enough cells in kv cache to restore state (%u > %u)\n", __func__, cell_count, kv_self.size); + return false; + } + if (kv_self.v_trans != (bool) v_trans) { + LLAMA_LOG_ERROR("%s: incompatible V transposition\n", __func__); + return false; + } + + // For each layer, read the keys for each cell, one row is one cell, read as one contiguous block + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(il) + hparams.n_embd_k_s(); + + // Read type of key + int32_t k_type_i_ref; + read_to(&k_type_i_ref, sizeof(k_type_i_ref)); + const int32_t k_type_i = (int32_t)kv_self.k_l[il]->type; + if (k_type_i != k_type_i_ref) { + LLAMA_LOG_ERROR("%s: mismatched key type (%d != %d, layer %d)\n", __func__, k_type_i, k_type_i_ref, il); + return false; + } + + // Read row size of key + uint64_t k_size_row_ref; + read_to(&k_size_row_ref, sizeof(k_size_row_ref)); + const size_t k_size_row = ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa); + if (k_size_row != k_size_row_ref) { + LLAMA_LOG_ERROR("%s: mismatched key row size (%zu != %zu, layer %d)\n", __func__, k_size_row, (size_t) k_size_row_ref, il); + return false; + } + + if (cell_count) { + // Read and set the keys for the whole cell range + ggml_backend_tensor_set(kv_self.k_l[il], read(cell_count * k_size_row), kv_self.head * k_size_row, cell_count * k_size_row); + } + } + + if (!kv_self.v_trans) { + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s(); + + // Read type of value + int32_t v_type_i_ref; + read_to(&v_type_i_ref, sizeof(v_type_i_ref)); + const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type; + if (v_type_i != v_type_i_ref) { + LLAMA_LOG_ERROR("%s: mismatched value type (%d != %d, layer %d)\n", __func__, v_type_i, v_type_i_ref, il); + return false; + } + + // Read row size of value + uint64_t v_size_row_ref; + read_to(&v_size_row_ref, sizeof(v_size_row_ref)); + const size_t v_size_row = ggml_row_size(kv_self.v_l[il]->type, n_embd_v_gqa); + if (v_size_row != v_size_row_ref) { + LLAMA_LOG_ERROR("%s: mismatched value row size (%zu != %zu, layer %d)\n", __func__, v_size_row, (size_t) v_size_row_ref, il); + return false; + } + + if (cell_count) { + // Read and set the values for the whole cell range + ggml_backend_tensor_set(kv_self.v_l[il], read(cell_count * v_size_row), kv_self.head * v_size_row, cell_count * v_size_row); + } + } + } else { + // For each layer, read the values for each cell (transposed) + for (uint32_t il = 0; il < n_layer; ++il) { + const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s(); + + // Read type of value + int32_t v_type_i_ref; + read_to(&v_type_i_ref, sizeof(v_type_i_ref)); + const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type; + if (v_type_i != v_type_i_ref) { + LLAMA_LOG_ERROR("%s: mismatched value type (%d != %d, layer %d)\n", __func__, v_type_i, v_type_i_ref, il); + return false; + } + + // Read element size of value + uint32_t v_size_el_ref; + read_to(&v_size_el_ref, sizeof(v_size_el_ref)); + const size_t v_size_el = ggml_type_size(kv_self.v_l[il]->type); + if (v_size_el != v_size_el_ref) { + LLAMA_LOG_ERROR("%s: mismatched value element size (%zu != %zu, layer %d)\n", __func__, v_size_el, (size_t) v_size_el_ref, il); + return false; + } + + // Read GQA embedding size + uint32_t n_embd_v_gqa_ref; + read_to(&n_embd_v_gqa_ref, sizeof(n_embd_v_gqa_ref)); + if (n_embd_v_gqa != n_embd_v_gqa_ref) { + LLAMA_LOG_ERROR("%s: mismatched GQA embedding size (%u != %u, layer %d)\n", __func__, n_embd_v_gqa, n_embd_v_gqa_ref, il); + return false; + } + + if (cell_count) { + // For each row in the transposed matrix, read the values for the whole cell range + for (uint32_t j = 0; j < n_embd_v_gqa; ++j) { + const size_t dst_offset = (kv_self.head + j * kv_self.size) * v_size_el; + ggml_backend_tensor_set(kv_self.v_l[il], read(cell_count * v_size_el), dst_offset, cell_count * v_size_el); + } + } + } + } + return true; + } + + void read_kv_cache(struct llama_context * ctx, llama_seq_id seq_id = -1) { + uint32_t cell_count; + read_to(&cell_count, sizeof(cell_count)); + + bool res = read_kv_cache_meta(ctx, cell_count, seq_id) && read_kv_cache_data(ctx, cell_count); + + if (!res) { + if (seq_id == -1) { + llama_kv_cache_clear(ctx); + } else { + llama_kv_cache_seq_rm(ctx, seq_id, -1, -1); + } + throw std::runtime_error("failed to restore kv cache"); + } + } +}; + +struct llama_data_write_dummy : llama_data_write { + size_t size_written = 0; + + llama_data_write_dummy() {} + + void write(const void * /* src */, size_t size) override { + size_written += size; + } + + void write_tensor_data(const struct ggml_tensor * /* tensor */, size_t /* offset */, size_t size) override { + size_written += size; + } + + size_t get_size_written() override { + return size_written; + } +}; + +struct llama_data_write_buffer : llama_data_write { + uint8_t * ptr; + size_t buf_size = 0; + size_t size_written = 0; + + llama_data_write_buffer(uint8_t * p, size_t len) : ptr(p), buf_size(len) {} + + void write(const void * src, size_t size) override { + if (size > buf_size) { + throw std::runtime_error("unexpectedly reached end of buffer"); + } + memcpy(ptr, src, size); + ptr += size; + size_written += size; + buf_size -= size; + } + + void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) override { + if (size > buf_size) { + throw std::runtime_error("unexpectedly reached end of buffer"); + } + ggml_backend_tensor_get(tensor, ptr, offset, size); + ptr += size; + size_written += size; + buf_size -= size; + } + + size_t get_size_written() override { + return size_written; + } +}; + +struct llama_data_read_buffer : llama_data_read { + const uint8_t * ptr; + size_t buf_size = 0; + size_t size_read = 0; + + llama_data_read_buffer(const uint8_t * p, size_t len) : ptr(p), buf_size(len) {} + + const uint8_t * read(size_t size) override { + const uint8_t * base_ptr = ptr; + if (size > buf_size) { + throw std::runtime_error("unexpectedly reached end of buffer"); + } + ptr += size; + size_read += size; + buf_size -= size; + return base_ptr; + } + + void read_to(void * dst, size_t size) override { + memcpy(dst, read(size), size); + } + + size_t get_size_read() override { + return size_read; + } +}; + +struct llama_data_write_file : llama_data_write { + llama_file * file; + size_t size_written = 0; + std::vector temp_buffer; + + llama_data_write_file(llama_file * f) : file(f) {} + + void write(const void * src, size_t size) override { + file->write_raw(src, size); + size_written += size; + } + + void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) override { + temp_buffer.resize(size); + ggml_backend_tensor_get(tensor, temp_buffer.data(), offset, size); + write(temp_buffer.data(), temp_buffer.size()); + } + + size_t get_size_written() override { + return size_written; + } +}; + +struct llama_data_read_file : llama_data_read { + llama_file * file; + size_t size_read = 0; + std::vector temp_buffer; + + llama_data_read_file(llama_file * f) : file(f) {} + + void read_to(void * dst, size_t size) override { + file->read_raw(dst, size); + size_read += size; + } + + const uint8_t * read(size_t size) override { + temp_buffer.resize(size); + read_to(temp_buffer.data(), size); + return temp_buffer.data(); + } + + size_t get_size_read() override { + return size_read; + } +}; + +/** copy state data into either a buffer or file depending on the passed in context + * + * file context: + * llama_file file("/path", "wb"); + * llama_data_write_file data_ctx(&file); + * llama_state_get_data_internal(ctx, data_ctx); + * + * buffer context: + * std::vector buf(max_size, 0); + * llama_data_write_buffer data_ctx(buf.data(), max_size); + * llama_state_get_data_internal(ctx, data_ctx); + * +*/ +static size_t llama_state_get_data_internal(struct llama_context * ctx, llama_data_write & data_ctx) { + llama_synchronize(ctx); + + data_ctx.write_model_info(ctx); + + // copy outputs + data_ctx.write_output_ids(ctx); + data_ctx.write_logits(ctx); + data_ctx.write_embeddings(ctx); + + data_ctx.write_kv_cache(ctx); + + return data_ctx.get_size_written(); +} + +size_t llama_state_get_data(struct llama_context * ctx, uint8_t * dst, size_t size) { + llama_data_write_buffer data_ctx(dst, size); + try { + return llama_state_get_data_internal(ctx, data_ctx); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error saving state: %s\n", __func__, err.what()); + return 0; + } +} + +// Returns the *actual* size of the state. +// Intended to be used when saving to state to a buffer. +size_t llama_state_get_size(struct llama_context * ctx) { + llama_data_write_dummy data_ctx; + try { + return llama_state_get_data_internal(ctx, data_ctx); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error getting state size: %s\n", __func__, err.what()); + return 0; + } +} + +static size_t llama_state_set_data_internal(struct llama_context * ctx, llama_data_read & data_ctx) { + llama_synchronize(ctx); + + data_ctx.read_model_info(ctx); + + // set outputs + data_ctx.read_output_ids(ctx); + data_ctx.read_logits(ctx); + data_ctx.read_embeddings(ctx); + + data_ctx.read_kv_cache(ctx); + + return data_ctx.get_size_read(); +} + +// Sets the state reading from the specified source address +size_t llama_state_set_data(struct llama_context * ctx, const uint8_t * src, size_t size) { + llama_data_read_buffer data_ctx(src, size); + try { + return llama_state_set_data_internal(ctx, data_ctx); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error loading state: %s\n", __func__, err.what()); + return 0; + } +} + +static bool llama_state_load_file_internal(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) { + llama_file file(path_session, "rb"); + + // sanity checks + { + const uint32_t magic = file.read_u32(); + const uint32_t version = file.read_u32(); + + if (magic != LLAMA_SESSION_MAGIC || version != LLAMA_SESSION_VERSION) { + LLAMA_LOG_ERROR("%s: unknown (magic, version) for session file: %08x, %08x\n", __func__, magic, version); + return false; + } + } + + // load the prompt + { + const uint32_t n_token_count = file.read_u32(); + + if (n_token_count > n_token_capacity) { + LLAMA_LOG_ERROR("%s: token count in session file exceeded capacity! %u > %zu\n", __func__, n_token_count, n_token_capacity); + return false; + } + + file.read_raw(tokens_out, sizeof(llama_token) * n_token_count); + *n_token_count_out = n_token_count; + } + + // restore the context state + { + const size_t n_state_size_cur = file.size() - file.tell(); + + llama_data_read_file data_ctx(&file); + const size_t n_read = llama_state_set_data_internal(ctx, data_ctx); + + if (n_read != n_state_size_cur) { + LLAMA_LOG_ERROR("%s: did not read all of the session file data! size %zu, got %zu\n", __func__, n_state_size_cur, n_read); + return false; + } + } + return true; +} + +bool llama_state_load_file(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) { + try { + return llama_state_load_file_internal(ctx, path_session, tokens_out, n_token_capacity, n_token_count_out); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error loading session file: %s\n", __func__, err.what()); + return false; + } +} + +static bool llama_state_save_file_internal(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) { + llama_file file(path_session, "wb"); + + file.write_u32(LLAMA_SESSION_MAGIC); + file.write_u32(LLAMA_SESSION_VERSION); + + // save the prompt + file.write_u32((uint32_t) n_token_count); + file.write_raw(tokens, sizeof(llama_token) * n_token_count); + + // save the context state using stream saving + llama_data_write_file data_ctx(&file); + llama_state_get_data_internal(ctx, data_ctx); + + return true; +} + +bool llama_state_save_file(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) { + try { + return llama_state_save_file_internal(ctx, path_session, tokens, n_token_count); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error saving session file: %s\n", __func__, err.what()); + return false; + } +} + +static size_t llama_state_seq_get_data_internal(struct llama_context * ctx, llama_data_write & data_ctx, llama_seq_id seq_id) { + llama_synchronize(ctx); + + data_ctx.write_kv_cache(ctx, seq_id); + + return data_ctx.get_size_written(); +} + +size_t llama_state_seq_get_size(struct llama_context * ctx, llama_seq_id seq_id) { + llama_data_write_dummy data_ctx; + return llama_state_seq_get_data_internal(ctx, data_ctx, seq_id); +} + +size_t llama_state_seq_get_data(struct llama_context * ctx, uint8_t * dst, size_t size, llama_seq_id seq_id) { + llama_data_write_buffer data_ctx(dst, size); + try { + return llama_state_seq_get_data_internal(ctx, data_ctx, seq_id); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error saving sequence state: %s\n", __func__, err.what()); + return 0; + } +} + +static size_t llama_state_seq_set_data_internal(struct llama_context * ctx, llama_data_read & data_ctx, llama_seq_id dest_seq_id) { + llama_synchronize(ctx); + + data_ctx.read_kv_cache(ctx, dest_seq_id); + + return data_ctx.get_size_read(); +} + +size_t llama_state_seq_set_data(struct llama_context * ctx, const uint8_t * src, size_t size, llama_seq_id dest_seq_id) { + llama_data_read_buffer data_ctx(src, size); + try { + return llama_state_seq_set_data_internal(ctx, data_ctx, dest_seq_id); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error loading sequence state: %s\n", __func__, err.what()); + return 0; + } +} + +static size_t llama_state_seq_save_file_internal(struct llama_context * ctx, const char * filepath, llama_seq_id seq_id, const llama_token * tokens, size_t n_token_count) { + llama_file file(filepath, "wb"); + + file.write_u32(LLAMA_STATE_SEQ_MAGIC); + file.write_u32(LLAMA_STATE_SEQ_VERSION); + + // save the prompt + file.write_u32((uint32_t) n_token_count); + file.write_raw(tokens, sizeof(llama_token) * n_token_count); + + // save the context state using stream saving + llama_data_write_file data_ctx(&file); + llama_state_seq_get_data_internal(ctx, data_ctx, seq_id); + + const size_t res = file.tell(); + GGML_ASSERT(res == sizeof(uint32_t) * 3 + sizeof(llama_token) * n_token_count + data_ctx.get_size_written()); + return res; +} + +static size_t llama_state_seq_load_file_internal(struct llama_context * ctx, const char * filepath, llama_seq_id dest_seq_id, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) { + llama_file file(filepath, "rb"); + + // version checks + { + const uint32_t magic = file.read_u32(); + const uint32_t version = file.read_u32(); + + if (magic != LLAMA_STATE_SEQ_MAGIC || version != LLAMA_STATE_SEQ_VERSION) { + LLAMA_LOG_ERROR("%s: unknown (magic, version) for sequence state file: %08x, %08x\n", __func__, magic, version); + return 0; + } + } + + // load the prompt + { + const uint32_t n_token_count = file.read_u32(); + + if (n_token_count > n_token_capacity) { + LLAMA_LOG_ERROR("%s: token count in sequence state file exceeded capacity! %u > %zu\n", __func__, n_token_count, n_token_capacity); + return 0; + } + + file.read_raw(tokens_out, sizeof(llama_token) * n_token_count); + *n_token_count_out = n_token_count; + } + + // restore the context state + { + const size_t state_size = file.size() - file.tell(); + llama_data_read_file data_ctx(&file); + const size_t nread = llama_state_seq_set_data_internal(ctx, data_ctx, dest_seq_id); + if (!nread) { + LLAMA_LOG_ERROR("%s: failed to restore sequence state\n", __func__); + return 0; + } + GGML_ASSERT(nread <= state_size); + GGML_ASSERT(nread + sizeof(uint32_t) * 3 + sizeof(llama_token) * *n_token_count_out == file.tell()); + } + + return file.tell(); +} + +size_t llama_state_seq_save_file(struct llama_context * ctx, const char * filepath, llama_seq_id seq_id, const llama_token * tokens, size_t n_token_count) { + try { + return llama_state_seq_save_file_internal(ctx, filepath, seq_id, tokens, n_token_count); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error saving sequence state file: %s\n", __func__, err.what()); + return 0; + } +} + +size_t llama_state_seq_load_file(struct llama_context * ctx, const char * filepath, llama_seq_id dest_seq_id, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) { + try { + return llama_state_seq_load_file_internal(ctx, filepath, dest_seq_id, tokens_out, n_token_capacity, n_token_count_out); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: error loading sequence state file: %s\n", __func__, err.what()); + return 0; + } +} + +const std::vector> & llama_internal_get_tensor_map( + struct llama_context * ctx +) { + return ctx->model.tensors_by_name; +} diff --git a/src/llama-context.h b/src/llama-context.h new file mode 100644 index 000000000..a9268b292 --- /dev/null +++ b/src/llama-context.h @@ -0,0 +1,128 @@ +#pragma once + +#include "llama.h" +#include "llama-batch.h" +#include "llama-cparams.h" +#include "llama-model.h" +#include "llama-kv-cache.h" +#include "llama-adapter.h" + +#include "ggml-cpp.h" + +#include +#include +#include +#include + +struct llama_context { + llama_context(const llama_model & model) + : model(model) + , t_start_us(model.t_start_us) + , t_load_us(model.t_load_us) {} + + const struct llama_model & model; + + struct llama_cparams cparams; + struct llama_sbatch sbatch; // TODO: revisit if needed + struct llama_kv_cache kv_self; + struct llama_adapter_cvec cvec; + + std::unordered_map lora; + + std::vector backends; + std::vector> set_n_threads_fns; + + ggml_backend_t backend_cpu = nullptr; + + ggml_threadpool_t threadpool = nullptr; + ggml_threadpool_t threadpool_batch = nullptr; + + bool has_evaluated_once = false; + + mutable int64_t t_start_us; + mutable int64_t t_load_us; + mutable int64_t t_p_eval_us = 0; + mutable int64_t t_eval_us = 0; + + mutable int64_t t_compute_start_us = 0; + mutable int64_t n_queued_tokens = 0; + + mutable int32_t n_p_eval = 0; // number of tokens in eval calls for the prompt (with batch size > 1) + mutable int32_t n_eval = 0; // number of eval calls + + // host buffer for the model output (logits and embeddings) + ggml_backend_buffer_ptr buf_output; + + // decode output (2-dimensional array: [n_outputs][n_vocab]) + size_t logits_size = 0; // capacity (of floats) for logits + float * logits = nullptr; + + std::vector output_ids; // map batch token positions to ids of the logits and embd buffers + size_t output_size = 0; // capacity (of tokens positions) for the output buffers + int32_t n_outputs = 0; // number of actually-used outputs in the current ubatch or last logical batch + + bool logits_all = false; + + // embeddings output (2-dimensional array: [n_outputs][n_embd]) + // populated only when pooling_type == LLAMA_POOLING_TYPE_NONE + size_t embd_size = 0; // capacity (of floats) for embeddings + float * embd = nullptr; + + // sequence embeddings output (map of [n_embd] vectors) + // populated only when pooling_type != LLAMA_POOLING_TYPE_NONE + std::map> embd_seq; + + // whether we are computing encoder output or decoder output + bool is_encoding = false; + + // TODO: find a better way to accommodate mutli-dimension position encoding methods + // number of position id each token get, 1 for each token in most cases. + // when using m-rope, it will be 3 position ids per token to representing 3 dimension coordinate. + int n_pos_per_token = 1; + + // output of the encoder part of the encoder-decoder models + std::vector embd_enc; + std::vector> seq_ids_enc; + + // memory buffers used to evaluate the model + std::vector buf_compute_meta; + ggml_backend_sched_ptr sched; + + ggml_abort_callback abort_callback = nullptr; + void * abort_callback_data = nullptr; + + // input tensors + struct ggml_tensor * inp_tokens; // I32 [n_batch] + struct ggml_tensor * inp_embd; // F32 [n_embd, n_batch] + struct ggml_tensor * inp_pos; // I32 [n_batch] + struct ggml_tensor * inp_out_ids; // I32 [n_outputs] + struct ggml_tensor * inp_KQ_mask; // F32 [kv_size, n_batch] + struct ggml_tensor * inp_KQ_mask_swa; // F32 [kv_size, n_batch] + struct ggml_tensor * inp_K_shift; // I32 [kv_size] + struct ggml_tensor * inp_mean; // F32 [n_batch, n_batch] + struct ggml_tensor * inp_cls; // I32 [n_batch] + struct ggml_tensor * inp_s_copy; // I32 [kv_size] + struct ggml_tensor * inp_s_mask; // F32 [1, n_kv] + struct ggml_tensor * inp_s_seq; // I32 [n_kv, n_batch] + struct ggml_tensor * inp_pos_bucket; // I32 [n_batch|n_kv, n_batch] + struct ggml_tensor * inp_embd_enc; // F32 [n_embd, n_outputs_enc] + struct ggml_tensor * inp_KQ_mask_cross; // F32 [n_outputs_enc, n_batch] +}; + +// TODO: make these methods of llama_context +void llama_set_k_shift(struct llama_context & lctx); + +void llama_set_s_copy(struct llama_context & lctx); + +void llama_set_inputs(llama_context & lctx, const llama_ubatch & ubatch); + +// Make sure enough space is available for outputs. +// Returns max number of outputs for which space was reserved. +size_t llama_output_reserve(struct llama_context & lctx, size_t n_outputs); + +// make the outputs have the same order they had in the user-provided batch +void llama_output_reorder(struct llama_context & ctx); + +// For internal test use +// TODO: remove +const std::vector> & llama_internal_get_tensor_map(struct llama_context * ctx); diff --git a/src/llama-cparams.cpp b/src/llama-cparams.cpp new file mode 100644 index 000000000..28369be36 --- /dev/null +++ b/src/llama-cparams.cpp @@ -0,0 +1 @@ +#include "llama-cparams.h" diff --git a/src/llama-cparams.h b/src/llama-cparams.h new file mode 100644 index 000000000..252012f3d --- /dev/null +++ b/src/llama-cparams.h @@ -0,0 +1,37 @@ +#pragma once + +#include "llama.h" + +#include + +struct llama_cparams { + uint32_t n_ctx; // context size used during inference + uint32_t n_batch; + uint32_t n_ubatch; + uint32_t n_seq_max; + int n_threads; // number of threads to use for generation + int n_threads_batch; // number of threads to use for batch processing + + float rope_freq_base; + float rope_freq_scale; + + uint32_t n_ctx_orig_yarn; + // These hyperparameters are not exposed in GGUF, because all + // existing YaRN models use the same values for them. + float yarn_ext_factor; + float yarn_attn_factor; + float yarn_beta_fast; + float yarn_beta_slow; + float defrag_thold; + + bool embeddings; + bool causal_attn; + bool offload_kqv; + bool flash_attn; + bool no_perf; + + enum llama_pooling_type pooling_type; + + ggml_backend_sched_eval_callback cb_eval; + void * cb_eval_user_data; +}; diff --git a/src/llama-grammar.cpp b/src/llama-grammar.cpp index 74e9f64b3..bebe4e9a3 100644 --- a/src/llama-grammar.cpp +++ b/src/llama-grammar.cpp @@ -1,5 +1,6 @@ #include "llama-grammar.h" +#include "llama-impl.h" #include "llama-vocab.h" #include "llama-sampling.h" @@ -822,15 +823,11 @@ llama_grammar_stacks & llama_grammar_get_stacks(struct llama_grammar * grammar) return grammar->stacks; } -void llama_grammar_accept( - const llama_grammar_rules & rules, - const llama_grammar_stacks & stacks, - const uint32_t chr, - llama_grammar_stacks & stacks_new) { - stacks_new.clear(); - stacks_new.reserve(stacks.size()); +void llama_grammar_accept(struct llama_grammar * grammar, uint32_t chr) { + llama_grammar_stacks stacks_new; + stacks_new.reserve(grammar->stacks.size()); - for (const auto & stack : stacks) { + for (const auto & stack : grammar->stacks) { if (stack.empty()) { continue; } @@ -844,9 +841,11 @@ void llama_grammar_accept( if (!llama_grammar_is_end_of_sequence(pos)) { new_stack.push_back(pos); } - llama_grammar_advance_stack(rules, new_stack, stacks_new); + llama_grammar_advance_stack(grammar->rules, new_stack, stacks_new); } } + + grammar->stacks = std::move(stacks_new); } llama_grammar_candidates llama_grammar_reject_candidates_for_stack( @@ -1051,7 +1050,12 @@ void llama_grammar_free_impl(struct llama_grammar * grammar) { } struct llama_grammar * llama_grammar_clone_impl(const struct llama_grammar & grammar) { - llama_grammar * result = new llama_grammar { grammar.vocab, grammar.rules, grammar.stacks, grammar.partial_utf8, }; + llama_grammar * result = new llama_grammar { + grammar.vocab, + grammar.rules, + grammar.stacks, + grammar.partial_utf8, + }; // redirect elements in stacks to point to new rules for (size_t is = 0; is < result->stacks.size(); is++) { @@ -1059,7 +1063,7 @@ struct llama_grammar * llama_grammar_clone_impl(const struct llama_grammar & gra for (size_t ir0 = 0; ir0 < grammar.rules.size(); ir0++) { for (size_t ir1 = 0; ir1 < grammar.rules[ir0].size(); ir1++) { if (grammar.stacks[is][ie] == &grammar.rules[ir0][ir1]) { - result->stacks[is][ie] = &result->rules[ir0][ir1]; + result->stacks[is][ie] = &result->rules[ir0][ir1]; } } } @@ -1088,9 +1092,9 @@ void llama_grammar_apply_impl(const struct llama_grammar & grammar, llama_token_ for (size_t i = 0; i < cur_p->size; ++i) { const llama_token id = cur_p->data[i].id; - const std::string & piece = grammar.vocab->cache_token_to_piece.at(id); + const std::string & piece = grammar.vocab->token_to_piece(id); - if (llama_token_is_eog_impl(*grammar.vocab, id)) { + if (grammar.vocab->is_eog(id)) { if (!allow_eog) { cur_p->data[i].logit = -INFINITY; } @@ -1111,7 +1115,7 @@ void llama_grammar_apply_impl(const struct llama_grammar & grammar, llama_token_ void llama_grammar_accept_impl(struct llama_grammar & grammar, llama_token token) { GGML_ASSERT(grammar.vocab != nullptr); - if (llama_token_is_eog_impl(*grammar.vocab, token)) { + if (grammar.vocab->is_eog(token)) { for (const auto & stack : grammar.stacks) { if (stack.empty()) { return; @@ -1120,17 +1124,14 @@ void llama_grammar_accept_impl(struct llama_grammar & grammar, llama_token token GGML_ABORT("fatal error"); } - const std::string & piece = grammar.vocab->cache_token_to_piece.at(token); + const std::string & piece = grammar.vocab->token_to_piece(token); // Note terminating 0 in decoded string const auto decoded = decode_utf8(piece, grammar.partial_utf8); const auto & code_points = decoded.first; - llama_grammar_stacks stacks_new; - for (auto it = code_points.begin(), end = code_points.end() - 1; it != end; ++it) { - llama_grammar_accept(grammar.rules, grammar.stacks, *it, stacks_new); - grammar.stacks = std::move(stacks_new); + llama_grammar_accept(&grammar, *it); } grammar.partial_utf8 = decoded.second; diff --git a/src/llama-grammar.h b/src/llama-grammar.h index f529ce351..f8b40c651 100644 --- a/src/llama-grammar.h +++ b/src/llama-grammar.h @@ -1,8 +1,10 @@ #pragma once -#include "llama-impl.h" +#include "llama.h" #include +#include +#include struct llama_vocab; @@ -58,6 +60,7 @@ using llama_grammar_rules = std::vector; using llama_grammar_stacks = std::vector; using llama_grammar_candidates = std::vector; +// TODO: remove, needed for tests atm const llama_grammar_rules & llama_grammar_get_rules (const struct llama_grammar * grammar); llama_grammar_stacks & llama_grammar_get_stacks( struct llama_grammar * grammar); @@ -65,11 +68,7 @@ const llama_grammar_rules & llama_grammar_get_rules (const struct llama_grammar // be positioned at a character range (see `llama_grammar_advance_stack`), and // produces the N possible stacks if the given char is accepted at those // positions -void llama_grammar_accept( - const llama_grammar_rules & rules, - const llama_grammar_stacks & stacks, - uint32_t chr, - llama_grammar_stacks & stacks_new); +void llama_grammar_accept(struct llama_grammar * grammar, uint32_t chr); std::vector llama_grammar_reject_candidates_for_stack( const llama_grammar_rules & rules, diff --git a/src/llama-hparams.cpp b/src/llama-hparams.cpp new file mode 100644 index 000000000..ea87b2953 --- /dev/null +++ b/src/llama-hparams.cpp @@ -0,0 +1,71 @@ +#include "llama-hparams.h" + +#include "ggml.h" + +uint32_t llama_hparams::n_head(uint32_t il) const { + if (il < n_layer) { + return n_head_arr[il]; + } + + GGML_ABORT("fatal error"); +} + +uint32_t llama_hparams::n_head_kv(uint32_t il) const { + if (il < n_layer) { + return n_head_kv_arr[il]; + } + + GGML_ABORT("fatal error"); +} + +uint32_t llama_hparams::n_ff(uint32_t il) const { + if (il < n_layer) { + return n_ff_arr[il]; + } + + GGML_ABORT("fatal error"); +} + +uint32_t llama_hparams::n_gqa(uint32_t il) const { + const uint32_t n_head = this->n_head(il); + const uint32_t n_head_kv = this->n_head_kv(il); + + if (n_head_kv == 0) { + return 0; + } + + return n_head/n_head_kv; +} + +uint32_t llama_hparams::n_embd_k_gqa(uint32_t il) const { + const uint32_t n_head_kv = this->n_head_kv(il); + + return n_embd_head_k * n_head_kv; +} + +uint32_t llama_hparams::n_embd_v_gqa(uint32_t il) const { + const uint32_t n_head_kv = this->n_head_kv(il); + + return n_embd_head_v * n_head_kv; +} + +uint32_t llama_hparams::n_embd_k_s() const { + if (wkv_head_size != 0) { + // for RWKV models + return token_shift_count * n_embd; + } + + // TODO: maybe support other convolution strides than 1 + // NOTE: since the first column of the conv_state is shifted out each time, it's not actually needed + return (ssm_d_conv > 0 ? ssm_d_conv - 1 : 0) * ssm_d_inner; +} + +uint32_t llama_hparams::n_embd_v_s() const { + if (wkv_head_size != 0) { + // corresponds to RWKV's wkv_states size + return n_embd * wkv_head_size; + } + + // corresponds to Mamba's ssm_states size + return ssm_d_state * ssm_d_inner; +} diff --git a/src/llama-hparams.h b/src/llama-hparams.h new file mode 100644 index 000000000..1fe454103 --- /dev/null +++ b/src/llama-hparams.h @@ -0,0 +1,139 @@ +#pragma once + +#include "llama.h" + +#include + +// bump if necessary +#define LLAMA_MAX_LAYERS 512 +#define LLAMA_MAX_EXPERTS 256 // DeepSeekV3 + +enum llama_expert_gating_func_type { + LLAMA_EXPERT_GATING_FUNC_TYPE_NONE = 0, + LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX = 1, + LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID = 2, +}; + +struct llama_hparams_posnet { + uint32_t n_embd; + uint32_t n_layer; +}; + +struct llama_hparams_convnext { + uint32_t n_embd; + uint32_t n_layer; +}; + +struct llama_hparams { + bool vocab_only; + bool rope_finetuned; + bool use_par_res; + bool swin_norm; + + uint32_t n_ctx_train; // context size the model was trained on + uint32_t n_embd; + uint32_t n_embd_features = 0; + uint32_t n_layer; + uint32_t n_rot; + uint32_t n_swa = 0; // sliding window attention (SWA) + uint32_t n_embd_head_k; // dimension of keys (d_k). d_q is assumed to be the same, but there are n_head q heads, and only n_head_kv k-v heads + uint32_t n_embd_head_v; // dimension of values (d_v) aka n_embd_head + uint32_t n_expert = 0; + uint32_t n_expert_used = 0; + uint32_t n_rel_attn_bkts = 0; + + // for WavTokenizer + struct llama_hparams_posnet posnet; + struct llama_hparams_convnext convnext; + + std::array n_head_arr; + std::array n_head_kv_arr; + std::array n_ff_arr; + + uint32_t n_layer_dense_lead = 0; + uint32_t n_lora_q = 0; + uint32_t n_lora_kv = 0; + uint32_t n_ff_exp = 0; + uint32_t n_ff_shexp = 0; + uint32_t n_expert_shared = 0; + uint32_t n_norm_groups = 0; + + float expert_weights_scale = 0.0; + bool expert_weights_norm = false; + uint32_t expert_gating_func = LLAMA_EXPERT_GATING_FUNC_TYPE_NONE; + + float f_norm_eps; + float f_norm_rms_eps; + float f_norm_group_eps; + + float f_attn_logit_softcapping = 50.0f; + float f_final_logit_softcapping = 30.0f; + + // for RWKV + uint32_t rescale_every_n_layers = 0; + uint32_t time_mix_extra_dim = 0; + uint32_t time_decay_extra_dim = 0; + uint32_t wkv_head_size = 0; + uint32_t token_shift_count = 2; + + float rope_attn_factor = 1.0f; + float rope_freq_base_train; + float rope_freq_scale_train; + uint32_t n_ctx_orig_yarn; + float rope_yarn_log_mul; + + std::array rope_sections; + + // for State Space Models + uint32_t ssm_d_conv = 0; + uint32_t ssm_d_inner = 0; + uint32_t ssm_d_state = 0; + uint32_t ssm_dt_rank = 0; + + bool ssm_dt_b_c_rms = false; + + float f_clamp_kqv = 0.0f; + float f_max_alibi_bias = 0.0f; + float f_logit_scale = 0.0f; + + // Additional scale factors (Granite/Granite MoE) + float f_residual_scale = 0.0f; + float f_embedding_scale = 0.0f; + float f_attention_scale = 0.0f; + + bool causal_attn = true; + bool use_alibi = false; + bool attn_soft_cap = false; + + // needed by encoder-decoder models (e.g. T5, FLAN-T5) + // ref: https://github.com/ggerganov/llama.cpp/pull/8141 + llama_token dec_start_token_id = LLAMA_TOKEN_NULL; + + enum llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_NONE; + enum llama_rope_type rope_type = LLAMA_ROPE_TYPE_NONE; + enum llama_rope_scaling_type rope_scaling_type_train = LLAMA_ROPE_SCALING_TYPE_NONE; + + uint32_t n_head(uint32_t il = 0) const; + + uint32_t n_head_kv(uint32_t il = 0) const; + + uint32_t n_ff(uint32_t il = 0) const; + + uint32_t n_gqa(uint32_t il = 0) const; + + // dimension of key embeddings across all k-v heads + uint32_t n_embd_k_gqa(uint32_t il = 0) const; + + // dimension of value embeddings across all k-v heads + uint32_t n_embd_v_gqa(uint32_t il = 0) const; + + // dimension of the rolling state embeddings + // corresponds to Mamba's conv_states size or RWKV's token_shift states size + uint32_t n_embd_k_s() const; + + // dimension of the recurrent state embeddings + uint32_t n_embd_v_s() const; +}; + +static_assert(std::is_trivially_copyable::value, "llama_hparams must be trivially copyable"); + diff --git a/src/llama-impl.cpp b/src/llama-impl.cpp new file mode 100644 index 000000000..6ec709dd3 --- /dev/null +++ b/src/llama-impl.cpp @@ -0,0 +1,167 @@ +#include "llama-impl.h" + +#include "gguf.h" +#include "llama.h" + +#include +#include +#include +#include +#include +#include + +struct llama_logger_state { + ggml_log_callback log_callback = llama_log_callback_default; + void * log_callback_user_data = nullptr; +}; + +static llama_logger_state g_logger_state; + +time_meas::time_meas(int64_t & t_acc, bool disable) : t_start_us(disable ? -1 : ggml_time_us()), t_acc(t_acc) {} + +time_meas::~time_meas() { + if (t_start_us >= 0) { + t_acc += ggml_time_us() - t_start_us; + } + } + +void llama_log_set(ggml_log_callback log_callback, void * user_data) { + ggml_log_set(log_callback, user_data); + g_logger_state.log_callback = log_callback ? log_callback : llama_log_callback_default; + g_logger_state.log_callback_user_data = user_data; +} + +static void llama_log_internal_v(ggml_log_level level, const char * format, va_list args) { + va_list args_copy; + va_copy(args_copy, args); + char buffer[128]; + int len = vsnprintf(buffer, 128, format, args); + if (len < 128) { + g_logger_state.log_callback(level, buffer, g_logger_state.log_callback_user_data); + } else { + char * buffer2 = new char[len + 1]; + vsnprintf(buffer2, len + 1, format, args_copy); + buffer2[len] = 0; + g_logger_state.log_callback(level, buffer2, g_logger_state.log_callback_user_data); + delete[] buffer2; + } + va_end(args_copy); +} + +void llama_log_internal(ggml_log_level level, const char * format, ...) { + va_list args; + va_start(args, format); + llama_log_internal_v(level, format, args); + va_end(args); +} + +void llama_log_callback_default(ggml_log_level level, const char * text, void * user_data) { + (void) level; + (void) user_data; + fputs(text, stderr); + fflush(stderr); +} + +void replace_all(std::string & s, const std::string & search, const std::string & replace) { + if (search.empty()) { + return; + } + std::string builder; + builder.reserve(s.length()); + size_t pos = 0; + size_t last_pos = 0; + while ((pos = s.find(search, last_pos)) != std::string::npos) { + builder.append(s, last_pos, pos - last_pos); + builder.append(replace); + last_pos = pos + search.length(); + } + builder.append(s, last_pos, std::string::npos); + s = std::move(builder); +} + +std::string format(const char * fmt, ...) { + va_list ap; + va_list ap2; + va_start(ap, fmt); + va_copy(ap2, ap); + int size = vsnprintf(NULL, 0, fmt, ap); + GGML_ASSERT(size >= 0 && size < INT_MAX); // NOLINT + std::vector buf(size + 1); + int size2 = vsnprintf(buf.data(), size + 1, fmt, ap2); + GGML_ASSERT(size2 == size); + va_end(ap2); + va_end(ap); + return std::string(buf.data(), size); +} + +std::string llama_format_tensor_shape(const std::vector & ne) { + char buf[256]; + snprintf(buf, sizeof(buf), "%5" PRId64, ne.at(0)); + for (size_t i = 1; i < ne.size(); i++) { + snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %5" PRId64, ne.at(i)); + } + return buf; +} + +std::string llama_format_tensor_shape(const struct ggml_tensor * t) { + char buf[256]; + snprintf(buf, sizeof(buf), "%5" PRId64, t->ne[0]); + for (int i = 1; i < GGML_MAX_DIMS; i++) { + snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %5" PRId64, t->ne[i]); + } + return buf; +} + +static std::string gguf_data_to_str(enum gguf_type type, const void * data, int i) { + switch (type) { + case GGUF_TYPE_UINT8: return std::to_string(((const uint8_t *)data)[i]); + case GGUF_TYPE_INT8: return std::to_string(((const int8_t *)data)[i]); + case GGUF_TYPE_UINT16: return std::to_string(((const uint16_t *)data)[i]); + case GGUF_TYPE_INT16: return std::to_string(((const int16_t *)data)[i]); + case GGUF_TYPE_UINT32: return std::to_string(((const uint32_t *)data)[i]); + case GGUF_TYPE_INT32: return std::to_string(((const int32_t *)data)[i]); + case GGUF_TYPE_UINT64: return std::to_string(((const uint64_t *)data)[i]); + case GGUF_TYPE_INT64: return std::to_string(((const int64_t *)data)[i]); + case GGUF_TYPE_FLOAT32: return std::to_string(((const float *)data)[i]); + case GGUF_TYPE_FLOAT64: return std::to_string(((const double *)data)[i]); + case GGUF_TYPE_BOOL: return ((const bool *)data)[i] ? "true" : "false"; + default: return format("unknown type %d", type); + } +} + +std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i) { + const enum gguf_type type = gguf_get_kv_type(ctx_gguf, i); + + switch (type) { + case GGUF_TYPE_STRING: + return gguf_get_val_str(ctx_gguf, i); + case GGUF_TYPE_ARRAY: + { + const enum gguf_type arr_type = gguf_get_arr_type(ctx_gguf, i); + int arr_n = gguf_get_arr_n(ctx_gguf, i); + const void * data = arr_type == GGUF_TYPE_STRING ? nullptr : gguf_get_arr_data(ctx_gguf, i); + std::stringstream ss; + ss << "["; + for (int j = 0; j < arr_n; j++) { + if (arr_type == GGUF_TYPE_STRING) { + std::string val = gguf_get_arr_str(ctx_gguf, i, j); + // escape quotes + replace_all(val, "\\", "\\\\"); + replace_all(val, "\"", "\\\""); + ss << '"' << val << '"'; + } else if (arr_type == GGUF_TYPE_ARRAY) { + ss << "???"; + } else { + ss << gguf_data_to_str(arr_type, data, j); + } + if (j < arr_n - 1) { + ss << ", "; + } + } + ss << "]"; + return ss.str(); + } + default: + return gguf_data_to_str(type, gguf_get_val_data(ctx_gguf, i), 0); + } +} diff --git a/src/llama-impl.h b/src/llama-impl.h index 70f16b61c..12d1fb082 100644 --- a/src/llama-impl.h +++ b/src/llama-impl.h @@ -1,10 +1,9 @@ #pragma once -#include "llama.h" +#include "ggml.h" // for ggml_log_level #include #include -#include #ifdef __GNUC__ #ifdef __MINGW32__ @@ -35,147 +34,28 @@ void llama_log_callback_default(ggml_log_level level, const char * text, void * // helpers // -struct time_meas { - time_meas(int64_t & t_acc, bool disable = false) : t_start_us(disable ? -1 : ggml_time_us()), t_acc(t_acc) {} +template +struct no_init { + T value; + no_init() { /* do nothing */ } +}; - ~time_meas() { - if (t_start_us >= 0) { - t_acc += ggml_time_us() - t_start_us; - } - } +struct time_meas { + time_meas(int64_t & t_acc, bool disable = false); + ~time_meas(); const int64_t t_start_us; int64_t & t_acc; }; -static void replace_all(std::string & s, const std::string & search, const std::string & replace) { - if (search.empty()) { - return; - } - std::string builder; - builder.reserve(s.length()); - size_t pos = 0; - size_t last_pos = 0; - while ((pos = s.find(search, last_pos)) != std::string::npos) { - builder.append(s, last_pos, pos - last_pos); - builder.append(replace); - last_pos = pos + search.length(); - } - builder.append(s, last_pos, std::string::npos); - s = std::move(builder); -} +void replace_all(std::string & s, const std::string & search, const std::string & replace); -const std::vector> & llama_internal_get_tensor_map( - struct llama_context * ctx -); +// TODO: rename to llama_format ? +LLAMA_ATTRIBUTE_FORMAT(1, 2) +std::string format(const char * fmt, ...); -// the ring buffer works similarly to std::deque, but with a fixed capacity -template -struct ring_buffer { - ring_buffer(size_t cap) : capacity(cap), data(cap) {} +std::string llama_format_tensor_shape(const std::vector & ne); +std::string llama_format_tensor_shape(const struct ggml_tensor * t); - T & front() { - if (sz == 0) { - throw std::runtime_error("ring buffer is empty"); - } - return data[first]; - } - - const T & front() const { - if (sz == 0) { - throw std::runtime_error("ring buffer is empty"); - } - return data[first]; - } - - T & back() { - if (sz == 0) { - throw std::runtime_error("ring buffer is empty"); - } - return data[pos]; - } - - const T & back() const { - if (sz == 0) { - throw std::runtime_error("ring buffer is empty"); - } - return data[pos]; - } - - void push_back(const T & value) { - if (capacity == 0) { - throw std::runtime_error("ring buffer: capacity is zero"); - } - - if (sz == capacity) { - // advance the start when buffer is full - first = (first + 1) % capacity; - } else { - sz++; - } - data[pos] = value; - pos = (pos + 1) % capacity; - } - - T pop_front() { - if (sz == 0) { - throw std::runtime_error("ring buffer is empty"); - } - T value = data[first]; - first = (first + 1) % capacity; - sz--; - return value; - } - - //T & operator[](size_t i) { - // if (i >= sz) { - // throw std::runtime_error("ring buffer: index out of bounds"); - // } - // return data[(first + i) % capacity]; - //} - - //const T & at(size_t i) const { - // if (i >= sz) { - // throw std::runtime_error("ring buffer: index out of bounds"); - // } - // return data[(first + i) % capacity]; - //} - - const T & rat(size_t i) const { - if (i >= sz) { - throw std::runtime_error("ring buffer: index out of bounds"); - } - return data[(first + sz - i - 1) % capacity]; - } - - std::vector to_vector() const { - std::vector result; - result.reserve(sz); - for (size_t i = 0; i < sz; i++) { - result.push_back(data[(first + i) % capacity]); - } - return result; - } - - void clear() { - // here only reset the status of the buffer - sz = 0; - first = 0; - pos = 0; - } - - bool empty() const { - return sz == 0; - } - - size_t size() const { - return sz; - } - - size_t capacity = 0; - size_t sz = 0; - size_t first = 0; - size_t pos = 0; - std::vector data; -}; +std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i); diff --git a/src/llama-kv-cache.cpp b/src/llama-kv-cache.cpp new file mode 100644 index 000000000..feffdf0de --- /dev/null +++ b/src/llama-kv-cache.cpp @@ -0,0 +1,718 @@ +#include "llama-kv-cache.h" + +#include "llama-impl.h" +#include "llama-batch.h" +#include "llama-cparams.h" +#include "llama-model.h" + +#include +#include +#include + +static const llama_kv_cache_slot_info llama_kv_cache_slot_info_failed{false}; + +uint32_t llama_kv_cache_get_padding(const struct llama_cparams & cparams) { + // the FA kernels require padding to avoid extra runtime boundary checks + return cparams.flash_attn ? 256u : 32u; +} + +bool llama_kv_cache_init( + struct llama_kv_cache & cache, + const llama_model & model, + const llama_cparams & cparams, + ggml_type type_k, + ggml_type type_v, + uint32_t kv_size, + bool offload) { + const struct llama_hparams & hparams = model.hparams; + + const int32_t n_layer = hparams.n_layer; + + cache.has_shift = false; + + cache.recurrent = llama_model_is_recurrent(&model); + cache.v_trans = !cache.recurrent && !cparams.flash_attn; + cache.can_shift = !cache.recurrent && model.arch != LLM_ARCH_DEEPSEEK2; // not supported due to MLA + + LLAMA_LOG_INFO("%s: kv_size = %d, offload = %d, type_k = '%s', type_v = '%s', n_layer = %d, can_shift = %d\n", + __func__, kv_size, offload, ggml_type_name(type_k), ggml_type_name(type_v), n_layer, cache.can_shift); + + cache.head = 0; + cache.size = kv_size; + cache.used = 0; + + cache.type_k = type_k; + cache.type_v = type_v; + + cache.cells.clear(); + cache.cells.resize(kv_size); + + // create a context for each buffer type + std::map ctx_map; + auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * { + auto it = ctx_map.find(buft); + if (it == ctx_map.end()) { + struct ggml_init_params params = { + /*.mem_size =*/ size_t(2u*n_layer*ggml_tensor_overhead()), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + ggml_context * ctx = ggml_init(params); + if (!ctx) { + return nullptr; + } + ctx_map[buft] = ctx; + cache.ctxs.emplace_back(ctx); + return ctx; + } + return it->second; + }; + + cache.k_l.reserve(n_layer); + cache.v_l.reserve(n_layer); + + for (int i = 0; i < n_layer; i++) { + const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(i) + hparams.n_embd_k_s(); + const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(i) + hparams.n_embd_v_s(); + + LLAMA_LOG_DEBUG("%s: layer %d: n_embd_k_gqa = %d, n_embd_v_gqa = %d\n", __func__, i, n_embd_k_gqa, n_embd_v_gqa); + + ggml_backend_buffer_type_t buft; + if (offload) { + auto * dev = model.dev_layer(i); + buft = ggml_backend_dev_buffer_type(dev); + } else { + buft = ggml_backend_cpu_buffer_type(); + } + ggml_context * ctx = ctx_for_buft(buft); + + if (!ctx) { + LLAMA_LOG_ERROR("%s: failed to create ggml context for kv cache\n", __func__); + return false; + } + + ggml_tensor * k = ggml_new_tensor_1d(ctx, type_k, n_embd_k_gqa*kv_size); + ggml_tensor * v = ggml_new_tensor_1d(ctx, type_v, n_embd_v_gqa*kv_size); + ggml_format_name(k, "cache_k_l%d", i); + ggml_format_name(v, "cache_v_l%d", i); + cache.k_l.push_back(k); + cache.v_l.push_back(v); + } + + // allocate tensors and initialize the buffers to avoid NaNs in the padding + for (auto it : ctx_map) { + auto * buft = it.first; + auto * ctx = it.second; + + ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft); + if (!buf) { + LLAMA_LOG_ERROR("%s: failed to allocate buffer for kv cache\n", __func__); + return false; + } + ggml_backend_buffer_clear(buf, 0); + LLAMA_LOG_INFO("%s: %10s KV buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf), ggml_backend_buffer_get_size(buf)/1024.0/1024.0); + cache.bufs.emplace_back(buf); + } + + return true; +} + +struct llama_kv_cache_slot_info llama_kv_cache_find_slot( + struct llama_kv_cache & cache, + const struct llama_ubatch & ubatch) { + const uint32_t n_tokens = ubatch.n_tokens; + const uint32_t n_seqs = ubatch.n_seqs; + const uint32_t n_seq_tokens = ubatch.n_seq_tokens; + + if (cache.recurrent) { + // For recurrent state architectures (like Mamba or RWKV), + // each cache cell can store the state for a whole sequence. + // A slot should be always be contiguous. + + // can only process batches with an equal number of new tokens in each sequence + GGML_ASSERT(ubatch.equal_seqs); + + int32_t min = cache.size - 1; + int32_t max = 0; + + // everything should fit if all seq_ids are smaller than the max + for (uint32_t s = 0; s < n_seqs; ++s) { + const uint32_t n_seq_id = ubatch.n_seq_id[s]; + for (uint32_t j = 0; j < n_seq_id; ++j) { + const llama_seq_id seq_id = ubatch.seq_id[s][j]; + + if (seq_id < 0 || (uint32_t) seq_id >= cache.size) { + // too big seq_id + // TODO: would it be possible to resize the cache instead? + LLAMA_LOG_ERROR("%s: seq_id=%d >= n_seq_max=%d Try using a bigger --parallel value\n", __func__, seq_id, cache.size); + return llama_kv_cache_slot_info_failed; + } + if (j > 0) { + llama_kv_cell & seq = cache.cells[seq_id]; + if (seq.tail >= 0) { + llama_kv_cell & cell = cache.cells[seq.tail]; + // clear cells from seq_ids that become shared + // (should not normally happen, but let's handle it anyway) + cell.seq_id.erase(seq_id); + seq.tail = -1; + if (cell.seq_id.empty()) { + cell.pos = -1; + cell.src = -1; + cache.used -= 1; + } + } + } + } + } + +#ifndef NDEBUG + { + std::vector tails_verif; + tails_verif.assign(cache.size, -1); + for (uint32_t i = 0; i < cache.size; ++i) { + llama_kv_cell & cell = cache.cells[i]; + for (llama_seq_id seq_id : cell.seq_id) { + if (tails_verif[seq_id] != -1) { + LLAMA_LOG_ERROR("%s: duplicate tail for seq_id %d in cell %d and %d\n", __func__, seq_id, i, tails_verif[seq_id]); + } + tails_verif[seq_id] = i; + } + } + for (uint32_t i = 0; i < cache.size; ++i) { + if (tails_verif[i] != cache.cells[i].tail) { + LLAMA_LOG_ERROR("%s: wrong tail for seq_id %d, (%d instead of %d)\n", __func__, i, cache.cells[i].tail, tails_verif[i]); + } + } + } +#endif + + // find next empty cell + uint32_t next_empty_cell = cache.head; + + for (uint32_t i = 0; i < cache.size; ++i) { + if (next_empty_cell >= cache.size) { next_empty_cell -= cache.size; } + llama_kv_cell & cell = cache.cells[next_empty_cell]; + if (cell.is_empty()) { break; } + next_empty_cell += 1; + } + + // find usable cell range + for (uint32_t s = 0; s < n_seqs; ++s) { + const llama_seq_id seq_id = ubatch.seq_id[s][0]; + llama_kv_cell & seq_meta = cache.cells[seq_id]; + bool has_cell = false; + if (seq_meta.tail >= 0) { + llama_kv_cell & cell = cache.cells[seq_meta.tail]; + GGML_ASSERT(cell.has_seq_id(seq_id)); + // does this seq_id "own" the cell? + if (cell.seq_id.size() == 1) { has_cell = true; } + } + if (!has_cell) { + llama_kv_cell & empty_cell = cache.cells[next_empty_cell]; + GGML_ASSERT(empty_cell.is_empty()); + // copy old tail into the empty cell + if (seq_meta.tail >= 0) { + llama_kv_cell & orig_cell = cache.cells[seq_meta.tail]; + empty_cell.pos = orig_cell.pos; + empty_cell.src = orig_cell.src; + orig_cell.seq_id.erase(seq_id); + empty_cell.seq_id.insert(seq_id); // will be overwritten + } + seq_meta.tail = next_empty_cell; + // find next empty cell + if (s + 1 < n_seqs) { + next_empty_cell += 1; + for (uint32_t i = 0; i < cache.size; ++i) { + if (next_empty_cell >= cache.size) { next_empty_cell -= cache.size; } + llama_kv_cell & cell = cache.cells[next_empty_cell]; + if (cell.is_empty()) { break; } + next_empty_cell += 1; + } + } + } + if (min > seq_meta.tail) { min = seq_meta.tail; } + if (max < seq_meta.tail) { max = seq_meta.tail; } + } + + // gather and re-order + for (uint32_t s = 0; s < n_seqs; ++s) { + int32_t dst_id = s + min; + int32_t src_id = cache.cells[ubatch.seq_id[s][0]].tail; + if (dst_id != src_id) { + llama_kv_cell & dst_cell = cache.cells[dst_id]; + llama_kv_cell & src_cell = cache.cells[src_id]; + + std::swap(dst_cell.pos, src_cell.pos); + std::swap(dst_cell.src, src_cell.src); + std::swap(dst_cell.seq_id, src_cell.seq_id); + + // swap tails (assuming they NEVER overlap) + for (const llama_seq_id seq_id : src_cell.seq_id) { + cache.cells[seq_id].tail = src_id; + } + for (const llama_seq_id seq_id : dst_cell.seq_id) { + cache.cells[seq_id].tail = dst_id; + } + } + } + + // update the pos of the used seqs + for (uint32_t s = 0; s < n_seqs; ++s) { + const llama_pos last_pos = ubatch.pos[n_seq_tokens * s + n_seq_tokens - 1]; + int32_t cell_id = s + min; + llama_kv_cell & cell = cache.cells[cell_id]; + + if (cell.pos >= 0 && last_pos != cell.pos + (llama_pos) n_seq_tokens) { + // What should happen when the pos backtracks or skips a value? + // Clearing the state mid-batch would require special-casing which isn't done. + LLAMA_LOG_WARN("%s: non-consecutive token position %d after %d for sequence %d with %u new tokens\n", + __func__, last_pos, cell.pos, ubatch.seq_id[s][0], n_seq_tokens); + } + cell.pos = last_pos; + cell.seq_id.clear(); + for (int32_t j = 0; j < ubatch.n_seq_id[s]; ++j) { + const llama_seq_id seq_id = ubatch.seq_id[s][j]; + cell.seq_id.insert(seq_id); + cache.cells[seq_id].tail = cell_id; + } + } + + // allow getting the range of used cells, from head to head + n + cache.head = min; + cache.n = max - min + 1; + cache.used = std::count_if(cache.cells.begin(), cache.cells.end(), + [](const llama_kv_cell& cell){ return !cell.is_empty(); }); + + // sanity check + return llama_kv_cache_slot_info(cache.n >= n_seqs); + } + // otherwise, one cell per token. + + if (n_tokens > cache.size) { + LLAMA_LOG_ERROR("%s: n_tokens=%d > cache.size=%d\n", __func__, n_tokens, cache.size); + return llama_kv_cache_slot_info_failed; + } + + uint32_t n_tested = 0; + + while (true) { + if (cache.head + n_tokens > cache.size) { + n_tested += cache.size - cache.head; + cache.head = 0; + continue; + } + + bool found = true; + for (uint32_t i = 0; i < n_tokens; i++) { + if (cache.cells[cache.head + i].pos >= 0) { + found = false; + cache.head += i + 1; + n_tested += i + 1; + break; + } + } + + if (found) { + break; + } + + if (n_tested >= cache.size) { + //LLAMA_LOG_ERROR("%s: failed to find a slot for %d tokens\n", __func__, n_tokens); + return llama_kv_cache_slot_info_failed; + } + } + + for (uint32_t s = 0; s < n_seqs; s++) { + for (uint32_t i = 0; i < n_seq_tokens; ++i) { + uint32_t k = s*n_seq_tokens + i; + cache.cells[cache.head + k].pos = ubatch.pos[k]; + + for (int32_t j = 0; j < ubatch.n_seq_id[s]; j++) { + cache.cells[cache.head + k].seq_id.insert(ubatch.seq_id[s][j]); + } + } + } + + cache.used += n_tokens; + + return llama_kv_cache_slot_info(cache.head, cache.head + n_tokens); +} + +uint32_t llama_kv_cache_cell_max(const struct llama_kv_cache & cache) { + for (uint32_t i = cache.size; i > 0; --i) { + const llama_kv_cell & cell = cache.cells[i - 1]; + + if (cell.pos >= 0 && !cell.is_empty()) { + return i; + } + } + + return 0; +} + +void llama_kv_cache_clear(struct llama_kv_cache & cache) { + for (int32_t i = 0; i < (int32_t) cache.size; ++i) { + cache.cells[i].pos = -1; + cache.cells[i].seq_id.clear(); + cache.cells[i].src = -1; + cache.cells[i].tail = -1; + } + cache.head = 0; + cache.used = 0; + + for (auto & buf : cache.bufs) { + ggml_backend_buffer_clear(buf.get(), 0); + } +} + +bool llama_kv_cache_seq_rm( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1) { + uint32_t new_head = cache.size; + + if (p0 < 0) p0 = 0; + if (p1 < 0) p1 = std::numeric_limits::max(); + + // models like Mamba or RWKV can't have a state partially erased + if (cache.recurrent) { + if (seq_id >= (int64_t) cache.size) { + // could be fatal + return false; + } + if (0 <= seq_id) { + int32_t & tail_id = cache.cells[seq_id].tail; + if (tail_id >= 0) { + const llama_kv_cell & cell = cache.cells[tail_id]; + // partial intersection is invalid + if ((0 < p0 && p0 <= cell.pos) || (0 < p1 && p1 <= cell.pos)) { + return false; + } + // invalidate tails which will be cleared + if (p0 <= cell.pos && cell.pos < p1) { + tail_id = -1; + } + } + } else { + // seq_id is negative, then the range should include everything or nothing + if (p0 != p1 && (p0 != 0 || p1 != std::numeric_limits::max())) { + return false; + } + } + } + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) { + if (seq_id < 0) { + cache.cells[i].seq_id.clear(); + } else if (cache.cells[i].has_seq_id(seq_id)) { + cache.cells[i].seq_id.erase(seq_id); + } else { + continue; + } + if (cache.cells[i].is_empty()) { + // keep count of the number of used cells + if (cache.cells[i].pos >= 0) cache.used--; + + cache.cells[i].pos = -1; + cache.cells[i].src = -1; + if (new_head == cache.size) new_head = i; + } + } + } + + // If we freed up a slot, set head to it so searching can start there. + if (new_head != cache.size && new_head < cache.head) cache.head = new_head; + + return true; +} + +void llama_kv_cache_seq_cp( + struct llama_kv_cache & cache, + llama_seq_id seq_id_src, + llama_seq_id seq_id_dst, + llama_pos p0, + llama_pos p1) { + if (p0 < 0) p0 = 0; + if (p1 < 0) p1 = std::numeric_limits::max(); + + if (cache.recurrent) { + if ((uint32_t) seq_id_dst < cache.size && (uint32_t) seq_id_src < cache.size) { + llama_kv_cell & tail_src = cache.cells[seq_id_src]; + llama_kv_cell & tail_dst = cache.cells[seq_id_dst]; + if (tail_dst.tail >= 0) { + // clear destination seq_id if it wasn't empty + llama_kv_cell & cell_dst = cache.cells[tail_dst.tail]; + + cell_dst.seq_id.erase(seq_id_dst); + tail_dst.tail = -1; + if (cell_dst.seq_id.empty()) { + cell_dst.pos = -1; + cell_dst.delta = -1; + cell_dst.src = -1; + cache.used -= 1; + } + } + if (tail_src.tail >= 0) { + llama_kv_cell & cell_src = cache.cells[tail_src.tail]; + + cell_src.seq_id.insert(seq_id_dst); + tail_dst.tail = tail_src.tail; + } + } + + return; + } + // otherwise, this is the KV cache of a Transformer-like model + + cache.head = 0; + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.cells[i].has_seq_id(seq_id_src) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) { + cache.cells[i].seq_id.insert(seq_id_dst); + } + } +} + +void llama_kv_cache_seq_keep(struct llama_kv_cache & cache, llama_seq_id seq_id) { + uint32_t new_head = cache.size; + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.recurrent && (llama_seq_id) i != seq_id) { + cache.cells[i].tail = -1; + } + if (!cache.cells[i].has_seq_id(seq_id)) { + if (cache.cells[i].pos >= 0) cache.used--; + cache.cells[i].pos = -1; + cache.cells[i].src = -1; + cache.cells[i].seq_id.clear(); + if (new_head == cache.size) new_head = i; + } else { + cache.cells[i].seq_id.clear(); + cache.cells[i].seq_id.insert(seq_id); + } + } + + // If we freed up a slot, set head to it so searching can start there. + if (new_head != cache.size && new_head < cache.head) cache.head = new_head; +} + +void llama_kv_cache_seq_add( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1, + llama_pos delta) { + uint32_t new_head = cache.size; + + if (p0 < 0) p0 = 0; + if (p1 < 0) p1 = std::numeric_limits::max(); + // If there is no range then return early to avoid looping over the cache. + if (p0 == p1) return; + + if (cache.recurrent) { + // for Mamba-like or RWKV models, only the pos needs to be shifted + if (0 <= seq_id && seq_id < (int64_t) cache.size) { + const int32_t tail_id = cache.cells[seq_id].tail; + if (tail_id >= 0) { + llama_kv_cell & cell = cache.cells[tail_id]; + if (cell.has_seq_id(seq_id) && p0 <= cell.pos && cell.pos < p1) { + cell.pos += delta; + } + } + } + return; + } + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.cells[i].has_seq_id(seq_id) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) { + cache.has_shift = true; + cache.cells[i].pos += delta; + cache.cells[i].delta += delta; + + if (cache.cells[i].pos < 0) { + if (!cache.cells[i].is_empty()) { + cache.used--; + } + cache.cells[i].pos = -1; + cache.cells[i].seq_id.clear(); + if (new_head == cache.size) { + new_head = i; + } + } + } + } + + // If we freed up a slot, set head to it so searching can start there. + // Otherwise we just start the next search from the beginning. + cache.head = new_head != cache.size ? new_head : 0; +} + +void llama_kv_cache_seq_div( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1, + int d) { + if (p0 < 0) p0 = 0; + if (p1 < 0) p1 = std::numeric_limits::max(); + // If there is no range then return early to avoid looping over the cache. + if (p0 == p1) return; + + if (cache.recurrent) { + // for Mamba-like or RWKV models, only the pos needs to be changed + if (0 <= seq_id && seq_id < (int64_t) cache.size) { + const int32_t tail_id = cache.cells[seq_id].tail; + if (tail_id >= 0) { + llama_kv_cell & cell = cache.cells[tail_id]; + if (cell.has_seq_id(seq_id) && p0 <= cell.pos && cell.pos < p1) { + cell.pos /= d; + } + } + } + return; + } + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.cells[i].has_seq_id(seq_id) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) { + cache.has_shift = true; + + { + llama_pos p_old = cache.cells[i].pos; + cache.cells[i].pos /= d; + cache.cells[i].delta += cache.cells[i].pos - p_old; + } + } + } +} + +llama_pos llama_kv_cache_seq_pos_max(struct llama_kv_cache & cache, llama_seq_id seq_id) { + llama_pos result = 0; + + for (uint32_t i = 0; i < cache.size; ++i) { + if (cache.cells[i].has_seq_id(seq_id)) { + result = std::max(result, cache.cells[i].pos); + } + } + + return result; +} + +void llama_kv_cache_defrag(struct llama_kv_cache & cache) { + if (!cache.recurrent) { + cache.do_defrag = true; + } +} + +int32_t llama_get_kv_cache_token_count(const struct llama_kv_cache & kv) { + int result = 0; + + for (uint32_t i = 0; i < kv.size; i++) { + result += kv.cells[i].seq_id.size(); + } + + return result; +} + +int32_t llama_get_kv_cache_used_cells(const struct llama_kv_cache & kv) { + return kv.used; +} + +bool llama_kv_cache_can_shift(const struct llama_kv_cache & kv) { + return kv.can_shift; +} + +// +// kv cache view +// + +struct llama_kv_cache_view llama_kv_cache_view_init(const struct llama_kv_cache & kv, int32_t n_seq_max) { + struct llama_kv_cache_view result = { + /*.n_cells = */ 0, + /*.n_seq_max = */ n_seq_max, + /*.token_count = */ 0, + /*.used_cells = */ llama_get_kv_cache_used_cells(kv), + /*.max_contiguous = */ 0, + /*.max_contiguous_idx = */ -1, + /*.cells = */ nullptr, + /*.cells_sequences = */ nullptr, + }; + + return result; +} + +void llama_kv_cache_view_free(struct llama_kv_cache_view * view) { + if (view->cells != nullptr) { + free(view->cells); + view->cells = nullptr; + } + if (view->cells_sequences != nullptr) { + free(view->cells_sequences); + view->cells_sequences = nullptr; + } +} + +void llama_kv_cache_view_update(struct llama_kv_cache_view * view, const struct llama_kv_cache & kv) { + if (uint32_t(view->n_cells) < kv.size || view->cells == nullptr) { + view->n_cells = int32_t(kv.size); + void * p = realloc(view->cells, sizeof(struct llama_kv_cache_view_cell) * view->n_cells); + GGML_ASSERT(p != nullptr && "Failed to alloc kv_cache_view cells"); + view->cells = (struct llama_kv_cache_view_cell *)p; + p = realloc(view->cells_sequences, sizeof(llama_seq_id) * view->n_seq_max * view->n_cells); + GGML_ASSERT(p != nullptr && "Failed to alloc kv_cache_view cells sequences"); + view->cells_sequences = (llama_seq_id *)p; + } + + const std::vector & kv_cells = kv.cells; + llama_kv_cache_view_cell * c_curr = view->cells; + llama_seq_id * cs_curr = view->cells_sequences; + int32_t used_cells = 0; + int32_t token_count = 0; + int32_t curr_contig_idx = -1; + uint32_t max_contig = 0; + int32_t max_contig_idx = -1; + + for (int32_t i = 0; i < int32_t(kv.size); i++, c_curr++, cs_curr += view->n_seq_max) { + const size_t curr_size = kv_cells[i].seq_id.size(); + token_count += curr_size; + c_curr->pos = kv_cells[i].pos + kv_cells[i].delta; + + if (curr_size > 0) { + if (curr_contig_idx >= 0 && uint32_t(i - curr_contig_idx) > max_contig) { + max_contig = i - curr_contig_idx; + max_contig_idx = curr_contig_idx; + } + curr_contig_idx = -1; + } else if (curr_contig_idx < 0) { + curr_contig_idx = i; + } + + int seq_idx = 0; + for (const llama_seq_id it : kv_cells[i].seq_id) { + if (seq_idx >= view->n_seq_max) { + break; + } + cs_curr[seq_idx] = it; + seq_idx++; + } + if (seq_idx != 0) { + used_cells++; + } + for (; seq_idx < view->n_seq_max; seq_idx++) { + cs_curr[seq_idx] = -1; + } + } + if (curr_contig_idx >= 0 && kv_cells.size() - curr_contig_idx > max_contig) { + max_contig_idx = curr_contig_idx; + max_contig = kv_cells.size() - curr_contig_idx; + } + view->max_contiguous = max_contig; + view->max_contiguous_idx = max_contig_idx; + view->token_count = token_count; + view->used_cells = used_cells; + if (uint32_t(used_cells) != kv.used) { + LLAMA_LOG_ERROR("%s: used cells mismatch. kv_cache says %d but we calculated %d\n", + __func__, kv.used, used_cells); + } +} diff --git a/src/llama-kv-cache.h b/src/llama-kv-cache.h new file mode 100644 index 000000000..dca6f3998 --- /dev/null +++ b/src/llama-kv-cache.h @@ -0,0 +1,218 @@ +#pragma once + +#include "llama.h" + +#include "ggml-cpp.h" + +#include +#include + +struct llama_kv_cell { + llama_pos pos = -1; + llama_pos delta = 0; + int32_t src = -1; // used by recurrent state models to copy states + int32_t tail = -1; + + std::set seq_id; + + bool has_seq_id(const llama_seq_id & id) const { + return seq_id.find(id) != seq_id.end(); + } + + bool is_empty() const { + return seq_id.empty(); + } + + bool is_same_seq(const llama_kv_cell & other) const { + return seq_id == other.seq_id; + } +}; + +// ring-buffer of cached KV data +struct llama_kv_cache { + bool has_shift = false; + bool do_defrag = false; + bool recurrent = false; // with recurrent state models, a cell can hold the state for more than one past token + bool v_trans = true; // the value tensor is transposed + bool can_shift = false; + + // Note: The value of head isn't only used to optimize searching + // for a free KV slot. llama_decode_internal also uses it, so it + // cannot be freely changed after a slot has been allocated. + uint32_t head = 0; + uint32_t size = 0; + uint32_t used = 0; // used cells (i.e. at least one seq_id) + + // computed before each graph build + uint32_t n = 0; + + ggml_type type_k = GGML_TYPE_F16; + ggml_type type_v = GGML_TYPE_F16; + + std::vector cells; + + std::vector k_l; // per layer + std::vector v_l; + + std::vector ctxs; + std::vector bufs; + + size_t total_size() const { + size_t size = 0; + for (const auto & buf : bufs) { + size += ggml_backend_buffer_get_size(buf.get()); + } + + return size; + } + + // TODO: better data structures to reduce the cost of this operation + llama_pos max_pos() const { + llama_pos max_pos = -1; + for (const auto & cell : cells) { + max_pos = std::max(max_pos, cell.pos); + } + + return max_pos; + } +}; + +// a structure holds information about the slot found in llama_kv_cache_find_slot +struct llama_kv_cache_slot_info { + std::pair boundaries; // slot boundaries [begin, end) + bool found = false; // the slot was found + + explicit llama_kv_cache_slot_info(bool found_) : found{found_} {} + llama_kv_cache_slot_info(uint32_t begin, uint32_t end) : boundaries{begin, end}, found{true} {} + + operator bool() const { return found; } +}; + +// TODO: maybe not needed +uint32_t llama_kv_cache_get_padding(const struct llama_cparams & cparams); + +bool llama_kv_cache_init( + struct llama_kv_cache & cache, + const llama_model & model, + const llama_cparams & cparams, + ggml_type type_k, + ggml_type type_v, + uint32_t kv_size, + bool offload); + +// find an empty slot of size "n_tokens" in the cache +// updates the cache head +// returns a structure holding information about the slot found +// Note: On success, it's important that cache.head points +// to the first cell of the slot. +struct llama_kv_cache_slot_info llama_kv_cache_find_slot( + struct llama_kv_cache & cache, + const struct llama_ubatch & batch); + +// find how many cells are currently in use +uint32_t llama_kv_cache_cell_max(const struct llama_kv_cache & cache); + +void llama_kv_cache_clear(struct llama_kv_cache & cache); + +bool llama_kv_cache_seq_rm( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1); + +void llama_kv_cache_seq_cp( + struct llama_kv_cache & cache, + llama_seq_id seq_id_src, + llama_seq_id seq_id_dst, + llama_pos p0, + llama_pos p1); + +void llama_kv_cache_seq_keep( + struct llama_kv_cache & cache, + llama_seq_id seq_id); + +void llama_kv_cache_seq_add( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1, + llama_pos delta); + +void llama_kv_cache_seq_div( + struct llama_kv_cache & cache, + llama_seq_id seq_id, + llama_pos p0, + llama_pos p1, + int d); + +llama_pos llama_kv_cache_seq_pos_max( + struct llama_kv_cache & cache, + llama_seq_id seq_id); + +void llama_kv_cache_defrag(struct llama_kv_cache & cache); + +int32_t llama_get_kv_cache_token_count(const struct llama_kv_cache & kv); + +int32_t llama_get_kv_cache_used_cells(const struct llama_kv_cache & kv); + +bool llama_kv_cache_can_shift(const struct llama_kv_cache & kv); + +// +// kv cache view +// + +struct llama_kv_cache_view llama_kv_cache_view_init(const struct llama_kv_cache & kv, int32_t n_seq_max); + +void llama_kv_cache_view_update(struct llama_kv_cache_view * view, const struct llama_kv_cache & kv); + +// +// kv cache restore +// + +// saves the kv_cache state for future recovery. +// used to rollback llama_kv_cache_find_slot changes. +struct llama_kv_slot_restorer { + struct llama_kv_cache_state { + uint32_t head = 0; + uint32_t n = 0; + } old_state; + + // for non-recurrent models only + // list of slots to restore + std::vector> slot_boundaries; + + bool do_restore = false; + + explicit llama_kv_slot_restorer(const struct llama_kv_cache & cache) { + old_state.head = cache.head; + old_state.n = cache.n; + } + + // saves a slot information for future restoration + void save(const struct llama_kv_cache_slot_info & slot) { + if (slot) { + do_restore = true; + if (slot.boundaries.first != slot.boundaries.second) { + slot_boundaries.push_back(slot.boundaries); + } + } + } + + // must be explicitly called to restore the kv_cache state + // and rollback changes from all llama_kv_cache_find_slot calls + void restore(struct llama_kv_cache & cache) { + if (do_restore) { + cache.head = old_state.head; + cache.n = old_state.n; + + if (cache.recurrent) { // recurrent models like Mamba or RWKV can't have a state partially erased + llama_kv_cache_seq_rm(cache, -1, -1, -1); + } else { + for (auto & slot : slot_boundaries) { + llama_kv_cache_seq_rm(cache, -1, slot.first, slot.second); + } + } + } + } +}; + diff --git a/src/llama-mmap.cpp b/src/llama-mmap.cpp new file mode 100644 index 000000000..57c6e4f51 --- /dev/null +++ b/src/llama-mmap.cpp @@ -0,0 +1,589 @@ +#include "llama-mmap.h" + +#include "llama-impl.h" + +#include "ggml.h" + +#include +#include +#include + +#ifdef __has_include + #if __has_include() + #include + #if defined(_POSIX_MAPPED_FILES) + #include + #include + #endif + #if defined(_POSIX_MEMLOCK_RANGE) + #include + #endif + #endif +#endif + +#if defined(_WIN32) + #define WIN32_LEAN_AND_MEAN + #ifndef NOMINMAX + #define NOMINMAX + #endif + #include + #ifndef PATH_MAX + #define PATH_MAX MAX_PATH + #endif + #include +#endif + +// TODO: consider moving to llama-impl.h if needed in more places +#if defined(_WIN32) +static std::string llama_format_win_err(DWORD err) { + LPSTR buf; + size_t size = FormatMessageA(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS, + NULL, err, MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), (LPSTR)&buf, 0, NULL); + if (!size) { + return "FormatMessageA failed"; + } + std::string ret(buf, size); + LocalFree(buf); + return ret; +} +#endif + +// llama_file + +struct llama_file::impl { +#if defined(_WIN32) + HANDLE fp_win32; + std::string GetErrorMessageWin32(DWORD error_code) const { + std::string ret; + LPSTR lpMsgBuf = NULL; + DWORD bufLen = FormatMessageA(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS, + NULL, error_code, MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), (LPSTR)&lpMsgBuf, 0, NULL); + if (!bufLen) { + ret = format("Win32 error code: %lx", error_code); + } else { + ret = lpMsgBuf; + LocalFree(lpMsgBuf); + } + + return ret; + } + + impl(const char * fname, const char * mode) { + fp = ggml_fopen(fname, mode); + if (fp == NULL) { + throw std::runtime_error(format("failed to open %s: %s", fname, strerror(errno))); + } + fp_win32 = (HANDLE) _get_osfhandle(_fileno(fp)); + seek(0, SEEK_END); + size = tell(); + seek(0, SEEK_SET); + } + + size_t tell() const { + LARGE_INTEGER li; + li.QuadPart = 0; + BOOL ret = SetFilePointerEx(fp_win32, li, &li, FILE_CURRENT); + if (!ret) { + throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str())); + } + + return li.QuadPart; + } + + void seek(size_t offset, int whence) const { + static_assert(SEEK_SET == FILE_BEGIN, "SEEK_SET != FILE_BEGIN"); + static_assert(SEEK_CUR == FILE_CURRENT, "SEEK_CUR != FILE_CURRENT"); + static_assert(SEEK_END == FILE_END, "SEEK_END != FILE_END"); + + LARGE_INTEGER li; + li.QuadPart = offset; + BOOL ret = SetFilePointerEx(fp_win32, li, NULL, whence); + if (!ret) { + throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str())); + } + } + + void read_raw(void * ptr, size_t len) const { + size_t bytes_read = 0; + while (bytes_read < len) { + size_t chunk_size = std::min(len - bytes_read, 64*1024*1024); + DWORD chunk_read = 0; + BOOL result = ReadFile(fp_win32, reinterpret_cast(ptr) + bytes_read, chunk_size, &chunk_read, NULL); + if (!result) { + throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str())); + } + if (chunk_read < chunk_size || chunk_read == 0) { + throw std::runtime_error("unexpectedly reached end of file"); + } + + bytes_read += chunk_read; + } + } + + uint32_t read_u32() const { + uint32_t val; + read_raw(&val, sizeof(val)); + return val; + } + + void write_raw(const void * ptr, size_t len) const { + size_t bytes_written = 0; + while (bytes_written < len) { + size_t chunk_size = std::min(len - bytes_written, 64*1024*1024); + DWORD chunk_written = 0; + BOOL result = WriteFile(fp_win32, reinterpret_cast(ptr) + bytes_written, chunk_size, &chunk_written, NULL); + if (!result) { + throw std::runtime_error(format("write error: %s", GetErrorMessageWin32(GetLastError()).c_str())); + } + if (chunk_written < chunk_size || chunk_written == 0) { + throw std::runtime_error("unexpectedly failed to write bytes"); + } + + bytes_written += chunk_written; + } + } + + void write_u32(uint32_t val) const { + write_raw(&val, sizeof(val)); + } + + ~impl() { + if (fp) { + std::fclose(fp); + } + } +#else + impl(const char * fname, const char * mode) { + fp = ggml_fopen(fname, mode); + if (fp == NULL) { + throw std::runtime_error(format("failed to open %s: %s", fname, strerror(errno))); + } + seek(0, SEEK_END); + size = tell(); + seek(0, SEEK_SET); + } + + size_t tell() const { +// TODO: this ifdef is never true? +#ifdef _WIN32 + __int64 ret = _ftelli64(fp); +#else + long ret = std::ftell(fp); +#endif + if (ret == -1) { + throw std::runtime_error(format("ftell error: %s", strerror(errno))); + } + + return (size_t) ret; + } + + void seek(size_t offset, int whence) const { +// TODO: this ifdef is never true? +#ifdef _WIN32 + int ret = _fseeki64(fp, (__int64) offset, whence); +#else + int ret = std::fseek(fp, (long) offset, whence); +#endif + if (ret != 0) { + throw std::runtime_error(format("seek error: %s", strerror(errno))); + } + } + + void read_raw(void * ptr, size_t len) const { + if (len == 0) { + return; + } + errno = 0; + std::size_t ret = std::fread(ptr, len, 1, fp); + if (ferror(fp)) { + throw std::runtime_error(format("read error: %s", strerror(errno))); + } + if (ret != 1) { + throw std::runtime_error("unexpectedly reached end of file"); + } + } + + uint32_t read_u32() const { + uint32_t ret; + read_raw(&ret, sizeof(ret)); + return ret; + } + + void write_raw(const void * ptr, size_t len) const { + if (len == 0) { + return; + } + errno = 0; + size_t ret = std::fwrite(ptr, len, 1, fp); + if (ret != 1) { + throw std::runtime_error(format("write error: %s", strerror(errno))); + } + } + + void write_u32(uint32_t val) const { + write_raw(&val, sizeof(val)); + } + + ~impl() { + if (fp) { + std::fclose(fp); + } + } +#endif + + FILE * fp; + size_t size; +}; + +llama_file::llama_file(const char * fname, const char * mode) : pimpl(std::make_unique(fname, mode)) {} +llama_file::~llama_file() = default; + +size_t llama_file::tell() const { return pimpl->tell(); } +size_t llama_file::size() const { return pimpl->size; } + +int llama_file::file_id() const { +#ifdef _WIN32 + return _fileno(pimpl->fp); +#else +#if defined(fileno) + return fileno(pimpl->fp); +#else + return ::fileno(pimpl->fp); +#endif +#endif +} + +void llama_file::seek(size_t offset, int whence) const { pimpl->seek(offset, whence); } +void llama_file::read_raw(void * ptr, size_t len) const { pimpl->read_raw(ptr, len); } + +uint32_t llama_file::read_u32() const { return pimpl->read_u32(); } + +void llama_file::write_raw(const void * ptr, size_t len) const { pimpl->write_raw(ptr, len); } +void llama_file::write_u32(uint32_t val) const { pimpl->write_u32(val); } + +// llama_mmap + +struct llama_mmap::impl { +#ifdef _POSIX_MAPPED_FILES + std::vector> mapped_fragments; + + impl(struct llama_file * file, size_t prefetch, bool numa) { + size = file->size(); + int fd = file->file_id(); + int flags = MAP_SHARED; + if (numa) { prefetch = 0; } +#ifdef __linux__ + if (posix_fadvise(fd, 0, 0, POSIX_FADV_SEQUENTIAL)) { + LLAMA_LOG_WARN("warning: posix_fadvise(.., POSIX_FADV_SEQUENTIAL) failed: %s\n", + strerror(errno)); + } + if (prefetch) { flags |= MAP_POPULATE; } +#endif + addr = mmap(NULL, file->size(), PROT_READ, flags, fd, 0); + if (addr == MAP_FAILED) { + throw std::runtime_error(format("mmap failed: %s", strerror(errno))); + } + + if (prefetch > 0) { + if (posix_madvise(addr, std::min(file->size(), prefetch), POSIX_MADV_WILLNEED)) { + LLAMA_LOG_WARN("warning: posix_madvise(.., POSIX_MADV_WILLNEED) failed: %s\n", + strerror(errno)); + } + } + if (numa) { + if (posix_madvise(addr, file->size(), POSIX_MADV_RANDOM)) { + LLAMA_LOG_WARN("warning: posix_madvise(.., POSIX_MADV_RANDOM) failed: %s\n", + strerror(errno)); + } + } + + mapped_fragments.emplace_back(0, file->size()); + } + + static void align_range(size_t * first, size_t * last, size_t page_size) { + size_t offset_in_page = *first & (page_size - 1); + size_t offset_to_page = offset_in_page == 0 ? 0 : page_size - offset_in_page; + *first += offset_to_page; + + *last = *last & ~(page_size - 1); + + if (*last <= *first) { + *last = *first; + } + } + + void unmap_fragment(size_t first, size_t last) { + int page_size = sysconf(_SC_PAGESIZE); + align_range(&first, &last, page_size); + size_t len = last - first; + + if (len == 0) { + return; + } + + GGML_ASSERT(first % page_size == 0); + GGML_ASSERT(last % page_size == 0); + GGML_ASSERT(last > first); + + void * next_page_start = (uint8_t *) addr + first; + + if (munmap(next_page_start, len)) { + LLAMA_LOG_WARN("warning: munmap failed: %s\n", strerror(errno)); + } + + std::vector> new_mapped_fragments; + for (const auto & frag : mapped_fragments) { + if (frag.first < first && frag.second > last) { + new_mapped_fragments.emplace_back(frag.first, first); + new_mapped_fragments.emplace_back(last, frag.second); + } else if (frag.first < first && frag.second > first) { + new_mapped_fragments.emplace_back(frag.first, first); + } else if (frag.first < last && frag.second > last) { + new_mapped_fragments.emplace_back(last, frag.second); + } else if (frag.first >= first && frag.second <= last) { + } else { + new_mapped_fragments.push_back(frag); + } + } + mapped_fragments = std::move(new_mapped_fragments); + } + + ~impl() { + for (const auto & frag : mapped_fragments) { + if (munmap((char *) addr + frag.first, frag.second - frag.first)) { + LLAMA_LOG_WARN("warning: munmap failed: %s\n", strerror(errno)); + } + } + } +#elif defined(_WIN32) + impl(struct llama_file * file, size_t prefetch, bool numa) { + GGML_UNUSED(numa); + + size = file->size(); + + HANDLE hFile = (HANDLE) _get_osfhandle(file->file_id()); + + HANDLE hMapping = CreateFileMappingA(hFile, NULL, PAGE_READONLY, 0, 0, NULL); + + if (hMapping == NULL) { + DWORD error = GetLastError(); + throw std::runtime_error(format("CreateFileMappingA failed: %s", llama_format_win_err(error).c_str())); + } + + addr = MapViewOfFile(hMapping, FILE_MAP_READ, 0, 0, 0); + DWORD error = GetLastError(); + CloseHandle(hMapping); + + if (addr == NULL) { + throw std::runtime_error(format("MapViewOfFile failed: %s", llama_format_win_err(error).c_str())); + } + + if (prefetch > 0) { +#if _WIN32_WINNT >= 0x602 + BOOL (WINAPI *pPrefetchVirtualMemory) (HANDLE, ULONG_PTR, PWIN32_MEMORY_RANGE_ENTRY, ULONG); + HMODULE hKernel32 = GetModuleHandleW(L"kernel32.dll"); + + pPrefetchVirtualMemory = (decltype(pPrefetchVirtualMemory))(void *) GetProcAddress(hKernel32, "PrefetchVirtualMemory"); + + if (pPrefetchVirtualMemory) { + WIN32_MEMORY_RANGE_ENTRY range; + range.VirtualAddress = addr; + range.NumberOfBytes = (SIZE_T) std::min(size, prefetch); + if (!pPrefetchVirtualMemory(GetCurrentProcess(), 1, &range, 0)) { + LLAMA_LOG_WARN("warning: PrefetchVirtualMemory failed: %s\n", + llama_format_win_err(GetLastError()).c_str()); + } + } +#else + throw std::runtime_error("PrefetchVirtualMemory unavailable"); +#endif + } + } + + void unmap_fragment(size_t first, size_t last) { + GGML_UNUSED(first); + GGML_UNUSED(last); + } + + ~impl() { + if (!UnmapViewOfFile(addr)) { + LLAMA_LOG_WARN("warning: UnmapViewOfFile failed: %s\n", + llama_format_win_err(GetLastError()).c_str()); + } + } +#else + impl(struct llama_file * file, size_t prefetch, bool numa) { + GGML_UNUSED(file); + GGML_UNUSED(prefetch); + GGML_UNUSED(numa); + + throw std::runtime_error("mmap not supported"); + } + + void unmap_fragment(size_t first, size_t last) { + GGML_UNUSED(first); + GGML_UNUSED(last); + + throw std::runtime_error("mmap not supported"); + } +#endif + + void * addr; + size_t size; +}; + +llama_mmap::llama_mmap(struct llama_file * file, size_t prefetch, bool numa) : pimpl(std::make_unique(file, prefetch, numa)) {} +llama_mmap::~llama_mmap() = default; + +size_t llama_mmap::size() const { return pimpl->size; } +void * llama_mmap::addr() const { return pimpl->addr; } + +void llama_mmap::unmap_fragment(size_t first, size_t last) { pimpl->unmap_fragment(first, last); } + +#if defined(_POSIX_MEMLOCK_RANGE) || defined(_WIN32) +const bool llama_mmap::SUPPORTED = true; +#else +const bool llama_mmap::SUPPORTED = false; +#endif + +// llama_mlock + +struct llama_mlock::impl { +#ifdef _POSIX_MEMLOCK_RANGE + static size_t lock_granularity() { + return (size_t) sysconf(_SC_PAGESIZE); + } + + bool raw_lock(const void * addr, size_t size) const { + if (!mlock(addr, size)) { + return true; + } + +#ifdef __APPLE__ +#define MLOCK_SUGGESTION \ + "Try increasing the sysctl values 'vm.user_wire_limit' and 'vm.global_user_wire_limit' and/or " \ + "decreasing 'vm.global_no_user_wire_amount'. Also try increasing RLIMIT_MEMLOCK (ulimit -l).\n" +#else +#define MLOCK_SUGGESTION \ + "Try increasing RLIMIT_MEMLOCK ('ulimit -l' as root).\n" +#endif + + char* errmsg = std::strerror(errno); + bool suggest = (errno == ENOMEM); + + struct rlimit lock_limit; + if (suggest && getrlimit(RLIMIT_MEMLOCK, &lock_limit)) { + suggest = false; + } + if (suggest && (lock_limit.rlim_max > lock_limit.rlim_cur + size)) { + suggest = false; + } + + LLAMA_LOG_WARN("warning: failed to mlock %zu-byte buffer (after previously locking %zu bytes): %s\n%s", + size, this->size, errmsg, suggest ? MLOCK_SUGGESTION : ""); + return false; + } + + static void raw_unlock(void * addr, size_t size) { + if (munlock(addr, size)) { + LLAMA_LOG_WARN("warning: failed to munlock buffer: %s\n", std::strerror(errno)); + } + } +#elif defined(_WIN32) + static size_t lock_granularity() { + SYSTEM_INFO si; + GetSystemInfo(&si); + return (size_t) si.dwPageSize; + } + + bool raw_lock(void * ptr, size_t len) const { + for (int tries = 1; ; tries++) { + if (VirtualLock(ptr, len)) { + return true; + } + if (tries == 2) { + LLAMA_LOG_WARN("warning: failed to VirtualLock %zu-byte buffer (after previously locking %zu bytes): %s\n", + len, size, llama_format_win_err(GetLastError()).c_str()); + return false; + } + + SIZE_T min_ws_size, max_ws_size; + if (!GetProcessWorkingSetSize(GetCurrentProcess(), &min_ws_size, &max_ws_size)) { + LLAMA_LOG_WARN("warning: GetProcessWorkingSetSize failed: %s\n", + llama_format_win_err(GetLastError()).c_str()); + return false; + } + size_t increment = len + 1048576; + min_ws_size += increment; + max_ws_size += increment; + if (!SetProcessWorkingSetSize(GetCurrentProcess(), min_ws_size, max_ws_size)) { + LLAMA_LOG_WARN("warning: SetProcessWorkingSetSize failed: %s\n", + llama_format_win_err(GetLastError()).c_str()); + return false; + } + } + } + + static void raw_unlock(void * ptr, size_t len) { + if (!VirtualUnlock(ptr, len)) { + LLAMA_LOG_WARN("warning: failed to VirtualUnlock buffer: %s\n", + llama_format_win_err(GetLastError()).c_str()); + } + } +#else + static size_t lock_granularity() { + return (size_t) 65536; + } + + bool raw_lock(const void * addr, size_t len) const { + LLAMA_LOG_WARN("warning: mlock not supported on this system\n"); + return false; + } + + static void raw_unlock(const void * addr, size_t len) {} +#endif + + impl() : addr(NULL), size(0), failed_already(false) {} + + void init(void * ptr) { + GGML_ASSERT(addr == NULL && size == 0); + addr = ptr; + } + + void grow_to(size_t target_size) { + GGML_ASSERT(addr); + if (failed_already) { + return; + } + size_t granularity = lock_granularity(); + target_size = (target_size + granularity - 1) & ~(granularity - 1); + if (target_size > size) { + if (raw_lock((uint8_t *) addr + size, target_size - size)) { + size = target_size; + } else { + failed_already = true; + } + } + } + + void * addr; + size_t size; + + bool failed_already; +}; + +llama_mlock::llama_mlock() : pimpl(std::make_unique()) {} +llama_mlock::~llama_mlock() = default; + +void llama_mlock::init(void * ptr) { pimpl->init(ptr); } +void llama_mlock::grow_to(size_t target_size) { pimpl->grow_to(target_size); } + +#if defined(_POSIX_MEMLOCK_RANGE) || defined(_WIN32) +const bool llama_mlock::SUPPORTED = true; +#else +const bool llama_mlock::SUPPORTED = false; +#endif + +size_t llama_path_max() { + return PATH_MAX; +} diff --git a/src/llama-mmap.h b/src/llama-mmap.h new file mode 100644 index 000000000..1da9ecb6b --- /dev/null +++ b/src/llama-mmap.h @@ -0,0 +1,67 @@ +#pragma once + +#include +#include + +struct llama_file; +struct llama_mmap; +struct llama_mlock; + +using llama_files = std::vector>; +using llama_mmaps = std::vector>; +using llama_mlocks = std::vector>; + +struct llama_file { + llama_file(const char * fname, const char * mode); + ~llama_file(); + + size_t tell() const; + size_t size() const; + + int file_id() const; // fileno overload + + void seek(size_t offset, int whence) const; + + void read_raw(void * ptr, size_t len) const; + uint32_t read_u32() const; + + void write_raw(const void * ptr, size_t len) const; + void write_u32(uint32_t val) const; + +private: + struct impl; + std::unique_ptr pimpl; +}; + +struct llama_mmap { + llama_mmap(const llama_mmap &) = delete; + llama_mmap(struct llama_file * file, size_t prefetch = (size_t) -1, bool numa = false); + ~llama_mmap(); + + size_t size() const; + void * addr() const; + + void unmap_fragment(size_t first, size_t last); + + static const bool SUPPORTED; + +private: + struct impl; + std::unique_ptr pimpl; +}; + +struct llama_mlock { + llama_mlock(); + ~llama_mlock(); + + void init(void * ptr); + void grow_to(size_t target_size); + + static const bool SUPPORTED; + +private: + struct impl; + std::unique_ptr pimpl; +}; + +size_t llama_path_max(); diff --git a/src/llama-model-loader.cpp b/src/llama-model-loader.cpp new file mode 100644 index 000000000..53175f0e0 --- /dev/null +++ b/src/llama-model-loader.cpp @@ -0,0 +1,1072 @@ +#include "llama-model-loader.h" + +#include "ggml.h" + +#include +#include +#include +#include + +static const size_t kiB = 1024; +static const size_t MiB = 1024*kiB; +static const size_t GiB = 1024*MiB; + +const char * llama_file_version_name(llama_fver version) { + switch (version) { + case GGUF_FILE_VERSION_V1: return "GGUF V1 (support until nov 2023)"; + case GGUF_FILE_VERSION_V2: return "GGUF V2"; + case GGUF_FILE_VERSION_V3: return "GGUF V3 (latest)"; + } + + return "unknown"; +} + +static std::string llama_model_ftype_name(llama_ftype ftype) { + if (ftype & LLAMA_FTYPE_GUESSED) { + return llama_model_ftype_name((enum llama_ftype) (ftype & ~LLAMA_FTYPE_GUESSED)) + " (guessed)"; + } + + switch (ftype) { + case LLAMA_FTYPE_ALL_F32: return "all F32"; + case LLAMA_FTYPE_MOSTLY_F16: return "F16"; + case LLAMA_FTYPE_MOSTLY_BF16: return "BF16"; + case LLAMA_FTYPE_MOSTLY_Q4_0: return "Q4_0"; + case LLAMA_FTYPE_MOSTLY_Q4_1: return "Q4_1"; + case LLAMA_FTYPE_MOSTLY_Q5_0: return "Q5_0"; + case LLAMA_FTYPE_MOSTLY_Q5_1: return "Q5_1"; + case LLAMA_FTYPE_MOSTLY_Q8_0: return "Q8_0"; + case LLAMA_FTYPE_MOSTLY_Q2_K: return "Q2_K - Medium"; + case LLAMA_FTYPE_MOSTLY_Q2_K_S: return "Q2_K - Small"; + case LLAMA_FTYPE_MOSTLY_Q3_K_S: return "Q3_K - Small"; + case LLAMA_FTYPE_MOSTLY_Q3_K_M: return "Q3_K - Medium"; + case LLAMA_FTYPE_MOSTLY_Q3_K_L: return "Q3_K - Large"; + case LLAMA_FTYPE_MOSTLY_Q4_K_S: return "Q4_K - Small"; + case LLAMA_FTYPE_MOSTLY_Q4_K_M: return "Q4_K - Medium"; + case LLAMA_FTYPE_MOSTLY_Q5_K_S: return "Q5_K - Small"; + case LLAMA_FTYPE_MOSTLY_Q5_K_M: return "Q5_K - Medium"; + case LLAMA_FTYPE_MOSTLY_Q6_K: return "Q6_K"; + case LLAMA_FTYPE_MOSTLY_TQ1_0: return "TQ1_0 - 1.69 bpw ternary"; + case LLAMA_FTYPE_MOSTLY_TQ2_0: return "TQ2_0 - 2.06 bpw ternary"; + case LLAMA_FTYPE_MOSTLY_IQ2_XXS: return "IQ2_XXS - 2.0625 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ2_XS: return "IQ2_XS - 2.3125 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ2_S: return "IQ2_S - 2.5 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ2_M: return "IQ2_M - 2.7 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ3_XS: return "IQ3_XS - 3.3 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ3_XXS: return "IQ3_XXS - 3.0625 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ1_S: return "IQ1_S - 1.5625 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ1_M: return "IQ1_M - 1.75 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ4_NL: return "IQ4_NL - 4.5 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ4_XS: return "IQ4_XS - 4.25 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ3_S: return "IQ3_S - 3.4375 bpw"; + case LLAMA_FTYPE_MOSTLY_IQ3_M: return "IQ3_S mix - 3.66 bpw"; + + default: return "unknown, may not work"; + } +} + +namespace GGUFMeta { + template + struct GKV_Base_Type { + static constexpr gguf_type gt = gt_; + + static T getter(const gguf_context * ctx, const int kid) { + return gfun(ctx, kid); + } + }; + + template struct GKV_Base; + + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + template<> struct GKV_Base: GKV_Base_Type {}; + + template<> struct GKV_Base { + static constexpr gguf_type gt = GGUF_TYPE_STRING; + + static std::string getter(const gguf_context * ctx, const int kid) { + return gguf_get_val_str(ctx, kid); + } + }; + + struct ArrayInfo { + const gguf_type gt; + const size_t length; + const void * data; + }; + + template<> struct GKV_Base { + public: + static constexpr gguf_type gt = GGUF_TYPE_ARRAY; + static ArrayInfo getter(const gguf_context *ctx, const int k) { + const enum gguf_type arr_type = gguf_get_arr_type(ctx, k); + return ArrayInfo { + arr_type, + size_t(gguf_get_arr_n(ctx, k)), + arr_type == GGUF_TYPE_STRING ? nullptr : gguf_get_arr_data(ctx, k), + }; + } + }; + + template + class GKV : public GKV_Base { + GKV() = delete; + + public: + static T get_kv(const gguf_context * ctx, const int k) { + const enum gguf_type kt = gguf_get_kv_type(ctx, k); + + if (kt != GKV::gt) { + throw std::runtime_error(format("key %s has wrong type %s but expected type %s", + gguf_get_key(ctx, k), gguf_type_name(kt), gguf_type_name(GKV::gt))); + } + return GKV::getter(ctx, k); + } + + static const char * override_type_to_str(const llama_model_kv_override_type ty) { + switch (ty) { + case LLAMA_KV_OVERRIDE_TYPE_BOOL: return "bool"; + case LLAMA_KV_OVERRIDE_TYPE_INT: return "int"; + case LLAMA_KV_OVERRIDE_TYPE_FLOAT: return "float"; + case LLAMA_KV_OVERRIDE_TYPE_STR: return "str"; + } + return "unknown"; + } + + static bool validate_override(const llama_model_kv_override_type expected_type, const struct llama_model_kv_override * ovrd) { + if (!ovrd) { return false; } + if (ovrd->tag == expected_type) { + LLAMA_LOG_INFO("%s: Using metadata override (%5s) '%s' = ", + __func__, override_type_to_str(ovrd->tag), ovrd->key); + switch (ovrd->tag) { + case LLAMA_KV_OVERRIDE_TYPE_BOOL: { + LLAMA_LOG_INFO("%s\n", ovrd->val_bool ? "true" : "false"); + } break; + case LLAMA_KV_OVERRIDE_TYPE_INT: { + LLAMA_LOG_INFO("%" PRId64 "\n", ovrd->val_i64); + } break; + case LLAMA_KV_OVERRIDE_TYPE_FLOAT: { + LLAMA_LOG_INFO("%.6f\n", ovrd->val_f64); + } break; + case LLAMA_KV_OVERRIDE_TYPE_STR: { + LLAMA_LOG_INFO("%s\n", ovrd->val_str); + } break; + default: + // Shouldn't be possible to end up here, but just in case... + throw std::runtime_error( + format("Unsupported attempt to override %s type for metadata key %s\n", + override_type_to_str(ovrd->tag), ovrd->key)); + } + return true; + } + LLAMA_LOG_WARN("%s: Warning: Bad metadata override type for key '%s', expected %s but got %s\n", + __func__, ovrd->key, override_type_to_str(expected_type), override_type_to_str(ovrd->tag)); + return false; + } + + template + static typename std::enable_if::value, bool>::type + try_override(OT & target, const struct llama_model_kv_override * ovrd) { + if (validate_override(LLAMA_KV_OVERRIDE_TYPE_BOOL, ovrd)) { + target = ovrd->val_bool; + return true; + } + return false; + } + + template + static typename std::enable_if::value && std::is_integral::value, bool>::type + try_override(OT & target, const struct llama_model_kv_override * ovrd) { + if (validate_override(LLAMA_KV_OVERRIDE_TYPE_INT, ovrd)) { + target = ovrd->val_i64; + return true; + } + return false; + } + + template + static typename std::enable_if::value, bool>::type + try_override(T & target, const struct llama_model_kv_override * ovrd) { + if (validate_override(LLAMA_KV_OVERRIDE_TYPE_FLOAT, ovrd)) { + target = ovrd->val_f64; + return true; + } + return false; + } + + template + static typename std::enable_if::value, bool>::type + try_override(T & target, const struct llama_model_kv_override * ovrd) { + if (validate_override(LLAMA_KV_OVERRIDE_TYPE_STR, ovrd)) { + target = ovrd->val_str; + return true; + } + return false; + } + + static bool set(const gguf_context * ctx, const int k, T & target, const struct llama_model_kv_override * ovrd = nullptr) { + if (try_override(target, ovrd)) { + return true; + } + if (k < 0) { return false; } + target = get_kv(ctx, k); + return true; + } + + static bool set(const gguf_context * ctx, const char * key, T & target, const struct llama_model_kv_override * ovrd = nullptr) { + return set(ctx, gguf_find_key(ctx, key), target, ovrd); + } + + static bool set(const gguf_context * ctx, const std::string & key, T & target, const struct llama_model_kv_override * ovrd = nullptr) { + return set(ctx, key.c_str(), target, ovrd); + } + }; +} + + template + typename std::enable_if::value, bool>::type + llama_model_loader::get_arr_n(const std::string & key, T & result, bool required) { + const int kid = gguf_find_key(meta.get(), key.c_str()); + + if (kid < 0) { + if (required) { + throw std::runtime_error(format("key not found in model: %s", key.c_str())); + } + return false; + } + + struct GGUFMeta::ArrayInfo arr_info = + GGUFMeta::GKV::get_kv(meta.get(), kid); + + + result = arr_info.length; + return true; + } + + template + typename std::enable_if::value, bool>::type + llama_model_loader::get_arr_n(enum llm_kv kid, T & result, bool required) { + return get_arr_n(llm_kv(kid), result, required); + } + + template bool llama_model_loader::get_arr_n(enum llm_kv kid, uint32_t & result, bool required); + + template + bool llama_model_loader::get_arr(const std::string & key, std::vector & result, bool required) { + const int kid = gguf_find_key(meta.get(), key.c_str()); + + if (kid < 0 || gguf_get_kv_type(meta.get(), kid) != GGUF_TYPE_ARRAY) { + if (required) { + throw std::runtime_error(format("array key not found in model: %s", key.c_str())); + } + return false; + } + + struct GGUFMeta::ArrayInfo arr_info = + GGUFMeta::GKV::get_kv(meta.get(), kid); + + switch (arr_info.gt) { + case GGUF_TYPE_FLOAT32: GGML_ASSERT((std::is_same::value)); break; + case GGUF_TYPE_INT32: GGML_ASSERT( + (std::is_same::value) || + (std::is_same::value)); break; + default: + throw std::runtime_error(format("%s is not a float32, int32 array", key.c_str())); + } + + result.resize(arr_info.length); + result.assign((const T*)arr_info.data, (const T *)arr_info.data + arr_info.length); + + return true; + } + + template + bool llama_model_loader::get_arr(const std::string & key, std::array & result, bool required) { + const int kid = gguf_find_key(meta.get(), key.c_str()); + + if (kid < 0 || gguf_get_kv_type(meta.get(), kid) != GGUF_TYPE_ARRAY) { + if (required) { + throw std::runtime_error(format("array key not found in model: %s", key.c_str())); + } + return false; + } + + struct GGUFMeta::ArrayInfo arr_info = + GGUFMeta::GKV::get_kv(meta.get(), kid); + + switch (arr_info.gt) { + case GGUF_TYPE_FLOAT32: GGML_ASSERT((std::is_same::value)); break; + case GGUF_TYPE_INT32: GGML_ASSERT( + (std::is_same::value) || + (std::is_same::value)); break; + default: + throw std::runtime_error(format("%s is not a float32, int32 array", key.c_str())); + } + + if (arr_info.length > N_MAX) { + throw std::runtime_error(format("array length %u for key %s exceeds max %u", (uint32_t) arr_info.length, key.c_str(), (uint32_t) N_MAX)); + } + + std::copy((const T*)arr_info.data, (const T *)arr_info.data + arr_info.length, result.begin()); + + return true; + } + + template + bool llama_model_loader::get_arr(enum llm_kv kid, T & result, bool required) { + return get_arr(llm_kv(kid), result, required); + } + + template + bool llama_model_loader::get_key(const std::string & key, T & result, bool required) { + auto it = kv_overrides.find(key); + + const struct llama_model_kv_override * override = + it != kv_overrides.end() ? &it->second : nullptr; + + const bool found = GGUFMeta::GKV::set(meta.get(), key, result, override); + + if (required && !found) { + throw std::runtime_error(format("key not found in model: %s", key.c_str())); + } + + return found; + } + + template + bool llama_model_loader::get_key(enum llm_kv kid, T & result, bool required) { + return get_key(llm_kv(kid), result, required); + } + + template bool llama_model_loader::get_key (enum llm_kv kid, bool & result, bool required); + template bool llama_model_loader::get_key (enum llm_kv kid, float & result, bool required); + template bool llama_model_loader::get_key (enum llm_kv kid, uint32_t & result, bool required); + template bool llama_model_loader::get_key(enum llm_kv kid, std::string & result, bool required); + + template<> + bool llama_model_loader::get_key(enum llm_kv kid, enum llama_pooling_type & result, bool required) { + uint32_t tmp; + const bool found = get_key(kid, tmp, required); + if (found) { + result = (enum llama_pooling_type) tmp; + } else { + result = LLAMA_POOLING_TYPE_UNSPECIFIED; + } + return found; + } + + // get array of n <= N_MAX elements, or a single element repeated n times + template + bool llama_model_loader::get_key_or_arr(const std::string & key, std::array & result, uint32_t n, bool required) { + const int kid = gguf_find_key(meta.get(), key.c_str()); + + if (kid < 0) { + if (required) { + throw std::runtime_error(format("key not found in model: %s", key.c_str())); + } + return false; + } + + if (n > N_MAX) { + throw std::runtime_error(format("n > N_MAX: %u > %u for key %s", (uint32_t) n, (uint32_t) N_MAX, key.c_str())); + } + + if (gguf_get_kv_type(meta.get(), kid) == GGUF_TYPE_ARRAY) { + struct GGUFMeta::ArrayInfo arr_info = + GGUFMeta::GKV::get_kv(meta.get(), kid); + + if (n != arr_info.length) { + throw std::runtime_error(format("key %s has wrong array length; expected %u, got %u", key.c_str(), n, (uint32_t) arr_info.length)); + } + + return get_arr(key, result, required); + } + + T value; + + bool ok = get_key(key, value, required); + if (!ok) { + return false; + } + + for (uint32_t i = 0; i < n; i++) { + result[i] = value; + } + + return true; + } + + template + bool llama_model_loader::get_key_or_arr(enum llm_kv kid, T & result, uint32_t n, bool required) { + return get_key_or_arr(llm_kv(kid), result, n, required); + } + + // TODO: this is not very clever - figure out something better + template bool llama_model_loader::get_key_or_arr>(enum llm_kv kid, std::array & result, uint32_t n, bool required); + template bool llama_model_loader::get_key_or_arr>(enum llm_kv kid, std::array & result, uint32_t n, bool required); + +llama_model_loader::llama_model_loader(const std::string & fname, bool use_mmap, bool check_tensors, const struct llama_model_kv_override * param_overrides_p) { + int trace = 0; + if (getenv("LLAMA_TRACE")) { + trace = atoi(getenv("LLAMA_TRACE")); + } + + if (param_overrides_p != nullptr) { + for (const struct llama_model_kv_override * p = param_overrides_p; p->key[0] != 0; p++) { + kv_overrides.insert({std::string(p->key), *p}); + } + } + + struct ggml_context * ctx = NULL; + struct gguf_init_params params = { + /*.no_alloc = */ true, + /*.ctx = */ &ctx, + }; + + meta.reset(gguf_init_from_file(fname.c_str(), params)); + if (!meta) { + throw std::runtime_error(format("%s: failed to load model from %s\n", __func__, fname.c_str())); + } + + get_key(llm_kv(LLM_KV_GENERAL_ARCHITECTURE), arch_name, false); + llm_kv = LLM_KV(llm_arch_from_string(arch_name)); + + files.emplace_back(new llama_file(fname.c_str(), "rb")); + contexts.emplace_back(ctx); + + // Save tensors data offset of the main file. + // For subsidiary files, `meta` tensor data offset must not be used, + // so we build a unified tensors index for weights. + for (ggml_tensor * cur = ggml_get_first_tensor(ctx); cur; cur = ggml_get_next_tensor(ctx, cur)) { + std::string tensor_name = std::string(cur->name); + // make sure there is no duplicated tensor names + if (weights_map.find(tensor_name) != weights_map.end()) { + throw std::runtime_error(format("invalid model: tensor '%s' is duplicated", ggml_get_name(cur))); + } + n_elements += ggml_nelements(cur); + n_bytes += ggml_nbytes(cur); + weights_map.emplace(tensor_name, llama_tensor_weight(files.back().get(), 0, meta.get(), cur)); + } + uint16_t n_split = 0; + get_key(llm_kv(LLM_KV_SPLIT_COUNT), n_split, false); + + // Load additional GGML contexts + if (n_split > 1) { + uint16_t idx = 0; + get_key(llm_kv(LLM_KV_SPLIT_NO), idx); + if (idx != 0) { + throw std::runtime_error(format("illegal split file: %d, model must be loaded with the first split", idx)); + } + + std::vector split_prefix(llama_path_max(), 0); + if (!llama_split_prefix(split_prefix.data(), split_prefix.size(), fname.c_str(), idx, n_split)) { + throw std::runtime_error(format("invalid split file: %s", fname.c_str())); + } + + if (trace > 0) { + LLAMA_LOG_INFO("%s: loading additional %d GGUFs\n", __func__, n_split); + } + + std::vector split_path(llama_path_max(), 0); + for (idx = 1; idx < n_split; idx++) { + llama_split_path(split_path.data(), split_path.size(), split_prefix.data(), idx, n_split); + + struct gguf_init_params split_params = { + /*.no_alloc = */ true, + /*.ctx = */ &ctx, + }; + gguf_context_ptr ctx_gguf { gguf_init_from_file(split_path.data(), split_params) }; + if (!ctx_gguf) { + throw std::runtime_error(format("%s: failed to load GGUF split from %s\n", __func__, split_path.data())); + } + + files.emplace_back(new llama_file(split_path.data(), "rb")); + contexts.emplace_back(ctx); + + // Save tensors data offset info of the shard. + for (ggml_tensor * cur = ggml_get_first_tensor(ctx); cur; cur = ggml_get_next_tensor(ctx, cur)) { + std::string tensor_name = std::string(cur->name); + // make sure there is no duplicated tensor names + if (weights_map.find(tensor_name) != weights_map.end()) { + throw std::runtime_error(format("invalid model: tensor '%s' is duplicated", ggml_get_name(cur))); + } + n_elements += ggml_nelements(cur); + n_bytes += ggml_nbytes(cur); + weights_map.emplace(tensor_name, llama_tensor_weight(files.back().get(), idx, ctx_gguf.get(), cur)); + } + } + + get_key(llm_kv(LLM_KV_SPLIT_TENSORS_COUNT), n_tensors); + + // sanity check + { + const int n_tensors_loaded = (int) weights_map.size(); + if (n_tensors != n_tensors_loaded) { + throw std::runtime_error(format("corrupted model: %d tensors expected but %d found", n_tensors, n_tensors_loaded)); + } + } + + LLAMA_LOG_INFO("%s: additional %d GGUFs metadata loaded.\n", __func__, n_split - 1); + } + + n_kv = gguf_get_n_kv(meta.get()); + n_tensors = weights_map.size(); + + fver = (enum llama_fver) gguf_get_version(meta.get()); + + LLAMA_LOG_INFO("%s: loaded meta data with %d key-value pairs and %d tensors from %s (version %s)\n", + __func__, n_kv, n_tensors, fname.c_str(), llama_file_version_name(fver)); + + // determine file type based on the number of tensors for each quantization and print meta data + // TODO: make optional + { + std::map n_type; + + uint32_t n_type_max = 0; + enum ggml_type type_max = GGML_TYPE_F32; + + for (const auto & it : weights_map) { + const llama_tensor_weight & w = it.second; + const ggml_tensor * tensor = w.tensor; + + enum ggml_type type = tensor->type; + + n_type[type]++; + + if (n_type_max < n_type[type]) { + n_type_max = n_type[type]; + type_max = type; + } + + if (trace > 0) { + const uint16_t sid = w.idx; + LLAMA_LOG_INFO("%s: - tensor split %2d: %32s %-8s [ %s ]\n", __func__, sid, ggml_get_name(tensor), ggml_type_name(type), llama_format_tensor_shape(tensor).c_str()); + } + } + + switch (type_max) { + case GGML_TYPE_F32: ftype = LLAMA_FTYPE_ALL_F32; break; + case GGML_TYPE_F16: ftype = LLAMA_FTYPE_MOSTLY_F16; break; + case GGML_TYPE_BF16: ftype = LLAMA_FTYPE_MOSTLY_BF16; break; + case GGML_TYPE_Q4_0: ftype = LLAMA_FTYPE_MOSTLY_Q4_0; break; + case GGML_TYPE_Q4_1: ftype = LLAMA_FTYPE_MOSTLY_Q4_1; break; + case GGML_TYPE_Q5_0: ftype = LLAMA_FTYPE_MOSTLY_Q5_0; break; + case GGML_TYPE_Q5_1: ftype = LLAMA_FTYPE_MOSTLY_Q5_1; break; + case GGML_TYPE_Q8_0: ftype = LLAMA_FTYPE_MOSTLY_Q8_0; break; + case GGML_TYPE_Q2_K: ftype = LLAMA_FTYPE_MOSTLY_Q2_K; break; + case GGML_TYPE_Q3_K: ftype = LLAMA_FTYPE_MOSTLY_Q3_K_M; break; + case GGML_TYPE_Q4_K: ftype = LLAMA_FTYPE_MOSTLY_Q4_K_M; break; + case GGML_TYPE_Q5_K: ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M; break; + case GGML_TYPE_Q6_K: ftype = LLAMA_FTYPE_MOSTLY_Q6_K; break; + case GGML_TYPE_TQ1_0: ftype = LLAMA_FTYPE_MOSTLY_TQ1_0; break; + case GGML_TYPE_TQ2_0: ftype = LLAMA_FTYPE_MOSTLY_TQ2_0; break; + case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break; + case GGML_TYPE_IQ2_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XS; break; + case GGML_TYPE_IQ2_S: ftype = LLAMA_FTYPE_MOSTLY_IQ2_S; break; + case GGML_TYPE_IQ3_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ3_XXS; break; + case GGML_TYPE_IQ1_S: ftype = LLAMA_FTYPE_MOSTLY_IQ1_S; break; + case GGML_TYPE_IQ1_M: ftype = LLAMA_FTYPE_MOSTLY_IQ1_M; break; + case GGML_TYPE_IQ4_NL: ftype = LLAMA_FTYPE_MOSTLY_IQ4_NL; break; + case GGML_TYPE_IQ4_XS: ftype = LLAMA_FTYPE_MOSTLY_IQ4_XS; break; + case GGML_TYPE_IQ3_S: ftype = LLAMA_FTYPE_MOSTLY_IQ3_S; break; + default: + { + LLAMA_LOG_WARN("%s: unknown type %s\n", __func__, ggml_type_name(type_max)); + ftype = LLAMA_FTYPE_ALL_F32; + } break; + } + + // this is a way to mark that we have "guessed" the file type + ftype = (llama_ftype) (ftype | LLAMA_FTYPE_GUESSED); + + { + const int kid = gguf_find_key(meta.get(), "general.file_type"); // TODO: use LLM_KV + if (kid >= 0) { + ftype = (llama_ftype) gguf_get_val_u32(meta.get(), kid); + } + } + + LLAMA_LOG_INFO("%s: Dumping metadata keys/values. Note: KV overrides do not apply in this output.\n", __func__); + + for (int i = 0; i < n_kv; i++) { + const char * name = gguf_get_key(meta.get(), i); + const enum gguf_type type = gguf_get_kv_type(meta.get(), i); + const std::string type_name = + type == GGUF_TYPE_ARRAY + ? format("%s[%s,%zu]", gguf_type_name(type), gguf_type_name(gguf_get_arr_type(meta.get(), i)), gguf_get_arr_n(meta.get(), i)) + : gguf_type_name(type); + + std::string value = gguf_kv_to_str(meta.get(), i); + const size_t MAX_VALUE_LEN = 40; + if (value.size() > MAX_VALUE_LEN) { + value = format("%s...", value.substr(0, MAX_VALUE_LEN - 3).c_str()); + } + replace_all(value, "\n", "\\n"); + + LLAMA_LOG_INFO("%s: - kv %3d: %42s %-16s = %s\n", __func__, i, name, type_name.c_str(), value.c_str()); + } + + // print type counts + for (auto & kv : n_type) { + if (kv.second == 0) { + continue; + } + + LLAMA_LOG_INFO("%s: - type %4s: %4d tensors\n", __func__, ggml_type_name(kv.first), kv.second); + } + } + + if (!llama_mmap::SUPPORTED) { + LLAMA_LOG_WARN("%s: mmap is not supported on this platform\n", __func__); + use_mmap = false; + } + + this->use_mmap = use_mmap; + this->check_tensors = check_tensors; +} + +std::string llama_model_loader::get_arch_name() const { + return arch_name; +} + +enum llm_arch llama_model_loader::get_arch() const { + return llm_kv.arch; +} + +const llama_model_loader::llama_tensor_weight * llama_model_loader::get_weight(const char * name) const { + auto pos = weights_map.find(name); + if (pos != weights_map.end()) { + return &pos->second; + } + + return nullptr; +} + +const llama_model_loader::llama_tensor_weight & llama_model_loader::require_weight(const char * name) const { + const llama_tensor_weight * weight = get_weight(name); + if (!weight) { + throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name)); + } + return *weight; +} + +struct ggml_tensor * llama_model_loader::get_tensor_meta(const char * name) const { + const auto * weight = get_weight(name); + if (!weight) { + return nullptr; + } + return weight->tensor; +} + +struct ggml_tensor * llama_model_loader::require_tensor_meta(const std::string & name) const { + struct ggml_tensor * tensor = get_tensor_meta(name.c_str()); + if (!tensor) { + throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name.c_str())); + } + return tensor; +} + +const struct ggml_tensor * llama_model_loader::check_tensor_dims(const std::string & name, const std::vector & ne, bool required) const { + const struct ggml_tensor * cur = get_tensor_meta(name.c_str()); + + if (cur == NULL) { + if (!required) { + return NULL; + } + throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name.c_str())); + } + + { + bool is_ok = true; + for (size_t i = 0; i < GGML_MAX_DIMS; ++i) { + if ((i < ne.size() && ne[i] != cur->ne[i]) || (i >= ne.size() && cur->ne[i] != 1)) { + is_ok = false; + break; + } + } + if (!is_ok) { + throw std::runtime_error( + format("%s: tensor '%s' has wrong shape; expected %s, got %s", + __func__, name.c_str(), + llama_format_tensor_shape(ne).c_str(), + llama_format_tensor_shape(cur).c_str())); + } + } + + return cur; +} + +struct ggml_tensor * llama_model_loader::create_tensor(struct ggml_context * ctx, const std::string & name, const std::initializer_list & ne, int flags) { + const struct ggml_tensor * cur = check_tensor_dims(name, ne, !(flags & TENSOR_NOT_REQUIRED)); + + if (cur == NULL) { + return NULL; + } + + bool duplicated = flags & TENSOR_DUPLICATED; + + struct ggml_tensor * tensor = ggml_dup_tensor(ctx, cur); + ggml_set_name(tensor, ggml_get_name(cur)); + + if (duplicated) { + size_data += ggml_nbytes(cur); + } else { + n_created++; + } + + return tensor; + +} + +struct ggml_tensor * llama_model_loader::create_tensor_as_view(struct ggml_context * ctx, struct ggml_tensor * base, const std::string & name, const std::initializer_list & ne, size_t offset, bool required) { + const struct ggml_tensor * cur = check_tensor_dims(name, ne, required); + + if (cur == NULL) { + return NULL; + } + + if (cur->type != base->type) { + throw std::runtime_error(format("%s: tensor '%s' has wrong type; expected %s, got %s", __func__, name.c_str(), ggml_type_name(base->type), ggml_type_name(cur->type))); + } + + std::array dims; + for (size_t i = 0; i < GGML_MAX_DIMS; ++i) { + dims[i] = i < ne.size() ? ne.begin()[i] : 1; + } + + struct ggml_tensor * tensor = ggml_view_4d(ctx, base, + dims[0], dims[1], dims[2], dims[3], + cur->nb[1], cur->nb[2], cur->nb[3], + offset); + + ggml_set_name(tensor, name.c_str()); + + n_created++; + + return tensor; +} + +void llama_model_loader::done_getting_tensors() const { + if (n_created != n_tensors) { + throw std::runtime_error(format("%s: wrong number of tensors; expected %d, got %d", __func__, n_tensors, n_created)); + } +} + +void llama_model_loader::init_mappings(bool prefetch, llama_mlocks * mlock_mmaps) { + if (use_mmap) { + mappings.reserve(files.size()); + mmaps_used.reserve(files.size()); + for (const auto & file : files) { + auto * reg = ggml_backend_dev_backend_reg(ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU)); + auto * is_numa_fn = (decltype(ggml_is_numa) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_is_numa"); + std::unique_ptr mapping(new llama_mmap(file.get(), prefetch ? -1 : 0, is_numa_fn())); + mmaps_used.emplace_back(mapping->size(), 0); + if (mlock_mmaps) { + std::unique_ptr mlock_mmap(new llama_mlock()); + mlock_mmap->init(mapping->addr()); + mlock_mmaps->emplace_back(std::move(mlock_mmap)); + } + mappings.emplace_back(std::move(mapping)); + } + } + + // compute the total size of all tensors for progress reporting + for (const auto & it : weights_map) { + size_data += ggml_nbytes(it.second.tensor); + } +} + +void llama_model_loader::get_mapping_range(size_t * first, size_t * last, void ** addr, int idx, ggml_context * ctx) const { + GGML_ASSERT(!mappings.empty()); + const auto & mapping = mappings.at(idx); + + *first = mapping->size(); + *last = 0; + *addr = mapping->addr(); + for (ggml_tensor * tensor = ggml_get_first_tensor(ctx); tensor; tensor = ggml_get_next_tensor(ctx, tensor)) { + const auto * weight = get_weight(ggml_get_name(tensor)); + if (!weight || weight->idx != idx) { + continue; + } + *first = std::min(*first, weight->offs); + *last = std::max(*last, weight->offs + ggml_nbytes(tensor)); + } +} + +void llama_model_loader::load_data_for(struct ggml_tensor * cur) const { + const auto & w = require_weight(ggml_get_name(cur)); + + if (use_mmap) { + const auto & mapping = mappings.at(w.idx); + if (cur->data == nullptr) { + cur->data = (uint8_t *)mapping->addr() + w.offs; + } else { + memcpy(cur->data, (uint8_t *)mapping->addr() + w.offs, ggml_nbytes(cur)); + } + } else { + GGML_ASSERT(cur->data != nullptr); + GGML_ASSERT(w.idx < files.size()); + const auto & file = files.at(w.idx); + file->seek(w.offs, SEEK_SET); + file->read_raw(cur->data, ggml_nbytes(cur)); + } + + if (check_tensors && !ggml_validate_row_data(cur->type, cur->data, ggml_nbytes(cur))) { + throw std::runtime_error(format("tensor '%s' has invalid data", ggml_get_name(cur))); + } +} + +bool llama_model_loader::load_all_data( + struct ggml_context * ctx, + llama_buf_map & bufs, + llama_mlocks * lmlocks, + llama_progress_callback progress_callback, + void * progress_callback_user_data) { + GGML_ASSERT(size_data != 0 && "call init_mappings() first"); + + std::vector> read_buf; + std::vector>> validation_result; + + // 4 staging buffers for async uploads, each sized 1MB seems to be a good default for single NVMe drives. + // NVMe raid configurations might require more / larger buffers. + constexpr size_t n_buffers = 4; + constexpr size_t buffer_size = 1 * 1024 * 1024; // 1MB + + std::vector host_buffers; + std::vector events; + std::vector host_ptrs; + size_t buffer_idx = 0; // buffer to use for async loads + ggml_backend_t upload_backend = [&](const char * func) -> ggml_backend_t { + if (use_mmap || check_tensors) { + return nullptr; + } + // When not using mmaped io use async uploads from pinned memory to GPU memory. + // First determine if the backend supports the necessary features for async uploads. + auto * buf = bufs.count(0) ? bufs.at(0) : nullptr; + if (!buf) { + LLAMA_LOG_DEBUG("%s: no buffer found for async uploads\n", func); + return nullptr; + } + + auto * buft = ggml_backend_buffer_get_type(buf); + auto * dev = ggml_backend_buft_get_device(buft); + if (!dev) { + LLAMA_LOG_DEBUG("%s: no device found for buffer type %s for async uploads\n", func, + ggml_backend_buft_name(buft)); + return nullptr; + } + + if (buft != ggml_backend_dev_buffer_type(dev)) { + LLAMA_LOG_DEBUG("%s: buffer type %s is not the default buffer type for device %s for async uploads\n", func, + ggml_backend_buft_name(buft), ggml_backend_dev_name(dev)); + return nullptr; + } + + ggml_backend_dev_props props; + ggml_backend_dev_get_props(dev, &props); + if (!props.caps.async || !props.caps.host_buffer || !props.caps.events) { + LLAMA_LOG_DEBUG("%s: device %s does not support async, host buffers or events\n", func, + ggml_backend_dev_name(dev)); + return nullptr; + } + + auto * host_buft = ggml_backend_dev_host_buffer_type(dev); + if (!host_buft) { + LLAMA_LOG_DEBUG("%s: no host buffer type found for device %s\n", func, + ggml_backend_dev_name(dev)); + return nullptr; + } + + // If the backend is supported, create pinned memory buffers and events for synchronisation. + for (size_t idx = 0; idx < n_buffers; ++idx) { + auto * buf = ggml_backend_buft_alloc_buffer(host_buft, buffer_size); + if (!buf) { + LLAMA_LOG_DEBUG("%s: failed to allocate host buffer for async uploads for device %s\n", func, + ggml_backend_dev_name(dev)); + return nullptr; + } + + host_buffers.emplace_back(buf); + host_ptrs.emplace_back(ggml_backend_buffer_get_base(buf)); + + auto * event = ggml_backend_event_new(dev); + if (!event) { + LLAMA_LOG_DEBUG("%s: failed to create event for async uploads for device %s\n", func, + ggml_backend_dev_name(dev)); + return nullptr; + } + + events.emplace_back(event); + } + + ggml_backend_t backend = ggml_backend_dev_init(dev, nullptr); + if (!backend) { + LLAMA_LOG_DEBUG("%s: failed to initialize backend for device %s for async uploads\n", func, + ggml_backend_dev_name(dev)); + return nullptr; + } + + return backend; + }(__func__); + + if (upload_backend) { + LLAMA_LOG_DEBUG("%s: using async uploads for device %s, buffer type %s, backend %s\n", __func__, + ggml_backend_dev_name(ggml_backend_get_device(upload_backend)), + ggml_backend_buft_name(ggml_backend_buffer_get_type(bufs.at(0))), + ggml_backend_name(upload_backend)); + } + + for (struct ggml_tensor * cur = ggml_get_first_tensor(ctx); cur != NULL; cur = ggml_get_next_tensor(ctx, cur)) { + const auto * weight = get_weight(ggml_get_name(cur)); + if (weight == nullptr) { + // this can happen with split experts models + continue; + } + + if (progress_callback) { + if (!progress_callback((float) size_done / size_data, progress_callback_user_data)) { + return false; + } + } + + size_t n_size = ggml_nbytes(cur); + + if (use_mmap) { + const auto & mapping = mappings.at(weight->idx); + ggml_backend_buffer_t buf_mmap = nullptr; + if (bufs.count(weight->idx)) { + buf_mmap = bufs.at(weight->idx); + } + uint8_t * data = (uint8_t *) mapping->addr() + weight->offs; + + if (check_tensors) { + validation_result.emplace_back(std::async(std::launch::async, [cur, data, n_size] { + return std::make_pair(cur, ggml_validate_row_data(cur->type, data, n_size)); + })); + } + + GGML_ASSERT(buf_mmap || cur->data); // either we have a buffer to allocate the tensor in, or it is already allocated + if (buf_mmap && cur->data == nullptr) { + ggml_backend_tensor_alloc(buf_mmap, cur, data); + if (lmlocks) { + const auto & lmlock = lmlocks->at(weight->idx); + lmlock->grow_to(weight->offs + n_size); + } + + auto & mmap_used = mmaps_used[weight->idx]; + mmap_used.first = std::min(mmap_used.first, weight->offs); + mmap_used.second = std::max(mmap_used.second, weight->offs + n_size); + } else { + ggml_backend_tensor_set(cur, data, 0, n_size); + } + } else { + const auto & file = files.at(weight->idx); + if (ggml_backend_buffer_is_host(cur->buffer)) { + file->seek(weight->offs, SEEK_SET); + file->read_raw(cur->data, n_size); + if (check_tensors) { + validation_result.emplace_back(std::async(std::launch::async, [cur, n_size] { + return std::make_pair(cur, ggml_validate_row_data(cur->type, cur->data, n_size)); + })); + } + } else { + // If upload_backend is valid load the tensor in chunks to pinned memory and upload the buffers asynchronously to the GPU. + if (upload_backend) { + file->seek(weight->offs, SEEK_SET); + + size_t bytes_read = 0; + + while (bytes_read < n_size) { + size_t read_iteration = std::min(buffer_size, n_size - bytes_read); + + ggml_backend_event_synchronize(events[buffer_idx]); + file->read_raw(host_ptrs[buffer_idx], read_iteration); + ggml_backend_tensor_set_async(upload_backend, cur, host_ptrs[buffer_idx], bytes_read, read_iteration); + ggml_backend_event_record(events[buffer_idx], upload_backend); + + bytes_read += read_iteration; + ++buffer_idx; + buffer_idx %= n_buffers; + } + } else { + read_buf.resize(n_size); + file->seek(weight->offs, SEEK_SET); + file->read_raw(read_buf.data(), n_size); + ggml_backend_tensor_set(cur, read_buf.data(), 0, n_size); + if (check_tensors && !ggml_validate_row_data(cur->type, read_buf.data(), n_size)) { + throw std::runtime_error(format("tensor '%s' has invalid data", ggml_get_name(cur))); + } + } + } + } + + size_done += n_size; + } + + // free temporary resources used for async uploads + for (auto * event : events) { + ggml_backend_event_synchronize(event); + ggml_backend_event_free(event); + } + for (auto * buf : host_buffers) { + ggml_backend_buffer_free(buf); + } + ggml_backend_free(upload_backend); + + // check validation results + bool validation_failed = false; + for (auto & future : validation_result) { + auto result = future.get(); + if (!result.second) { + LLAMA_LOG_ERROR("%s: tensor '%s' has invalid data\n", __func__, ggml_get_name(result.first)); + validation_failed = true; + } + } + if (validation_failed) { + throw std::runtime_error("found tensors with invalid data"); + } + + // check if this is the last call and do final cleanup + if (size_done >= size_data) { + // unmap offloaded tensors and metadata + if (use_mmap) { + for (uint32_t idx = 0; idx < mappings.size(); idx++) { + const auto & mmap_used = mmaps_used.at(idx); + auto & mapping = mappings.at(idx); + mapping->unmap_fragment(0, mmap_used.first); + if (mmap_used.second != 0) { + mapping->unmap_fragment(mmap_used.second, mapping->size()); + } + } + } + if (progress_callback) { + // Even though the model is done loading, we still honor + // cancellation since we need to free allocations. + return progress_callback(1.0f, progress_callback_user_data); + } + } + + return true; +} + +std::string llama_model_loader::ftype_name() const { + return llama_model_ftype_name(ftype); +} + +void llama_model_loader::print_info() const { + LLAMA_LOG_INFO("%s: file format = %s\n", __func__, llama_file_version_name(fver)); + LLAMA_LOG_INFO("%s: file type = %s\n", __func__, llama_model_ftype_name(ftype).c_str()); + if (n_bytes < GiB) { + LLAMA_LOG_INFO("%s: file size = %.2f MiB (%.2f BPW) \n", __func__, n_bytes/1024.0/1024.0, n_bytes*8.0/n_elements); + } else { + LLAMA_LOG_INFO("%s: file size = %.2f GiB (%.2f BPW) \n", __func__, n_bytes/1024.0/1024.0/1024.0, n_bytes*8.0/n_elements); + } +} diff --git a/src/llama-model-loader.h b/src/llama-model-loader.h new file mode 100644 index 000000000..b63d158d9 --- /dev/null +++ b/src/llama-model-loader.h @@ -0,0 +1,162 @@ +#pragma once + +#include "llama.h" + +#include "llama-impl.h" +#include "llama-arch.h" +#include "llama-mmap.h" + +#include "ggml-cpp.h" + +#include +#include +#include +#include + +using llama_buf_map = std::unordered_map; + +enum llama_fver { + GGUF_FILE_VERSION_V1 = 1, + GGUF_FILE_VERSION_V2 = 2, + GGUF_FILE_VERSION_V3 = 3, +}; + +const char * llama_file_version_name(llama_fver version); + +struct llama_model_loader { + // Holds information on a model weight + struct llama_tensor_weight { + uint16_t idx; // source file index + size_t offs; // tensor data offset in the original file + + ggml_tensor * tensor; + + llama_tensor_weight(const llama_file * file, uint16_t idx, const struct gguf_context * gguf_ctx, ggml_tensor * tensor) : idx(idx), tensor(tensor) { + const int tensor_idx = gguf_find_tensor(gguf_ctx, ggml_get_name(tensor)); + if (tensor_idx < 0) { + throw std::runtime_error(format("tensor '%s' not found in the model", ggml_get_name(tensor))); + } + + offs = gguf_get_data_offset(gguf_ctx) + gguf_get_tensor_offset(gguf_ctx, tensor_idx); + if (offs + ggml_nbytes(tensor) < offs || offs + ggml_nbytes(tensor) > file->size()) { + throw std::runtime_error(format("tensor '%s' data is not within the file bounds, model is corrupted or incomplete", ggml_get_name(tensor))); + } + } + }; + + // custom comparator to sort weights more nicely by layer + struct weight_name_comparer { + bool operator()(const std::string & a, const std::string & b) const { + int a_layer = -1; + int b_layer = -1; + sscanf(a.c_str(), "blk.%d.", &a_layer); + sscanf(b.c_str(), "blk.%d.", &b_layer); + if (a_layer != b_layer) { + return a_layer < b_layer; + } + return a < b; + } + }; + + static const int TENSOR_NOT_REQUIRED = 1; + static const int TENSOR_DUPLICATED = 2; + + int n_kv = 0; + int n_tensors = 0; + int n_created = 0; + + uint64_t n_elements = 0; + size_t n_bytes = 0; + + bool use_mmap = false; + bool check_tensors; + + llama_files files; + llama_ftype ftype; + llama_fver fver; + + llama_mmaps mappings; + + std::map weights_map; + std::unordered_map kv_overrides; + + gguf_context_ptr meta; + std::vector contexts; + + std::string arch_name; + LLM_KV llm_kv = LLM_KV(LLM_ARCH_UNKNOWN); + + size_t size_done = 0; + size_t size_data = 0; + std::vector> mmaps_used; + + llama_model_loader(const std::string & fname, bool use_mmap, bool check_tensors, const struct llama_model_kv_override * param_overrides_p); + + template + typename std::enable_if::value, bool>::type + get_arr_n(const std::string & key, T & result, bool required = true); + + template + typename std::enable_if::value, bool>::type + get_arr_n(enum llm_kv kid, T & result, bool required = true); + + template + bool get_arr(const std::string & key, std::vector & result, bool required = true); + + template + bool get_arr(const std::string & key, std::array & result, bool required = true); + + template + bool get_arr(enum llm_kv kid, T & result, bool required = true); + + template + bool get_key(const std::string & key, T & result, bool required = true); + + template + bool get_key(enum llm_kv kid, T & result, bool required = true); + + template + bool get_key_or_arr(const std::string & key, std::array & result, uint32_t n, bool required = true); + + template + bool get_key_or_arr(enum llm_kv kid, T & result, uint32_t n, bool required = true); + + std::string get_arch_name() const; + + enum llm_arch get_arch() const; + + const llama_tensor_weight * get_weight(const char * name) const; + + const llama_tensor_weight & require_weight(const char * name) const; + + struct ggml_tensor * get_tensor_meta(const char * name) const; + + struct ggml_tensor * require_tensor_meta(const std::string & name) const; + + const struct ggml_tensor * check_tensor_dims(const std::string & name, const std::vector & ne, bool required) const; + + struct ggml_tensor * create_tensor(struct ggml_context * ctx, const std::string & name, const std::initializer_list & ne, int flags = 0); + + struct ggml_tensor * create_tensor_as_view(struct ggml_context * ctx, struct ggml_tensor * base, const std::string & name, const std::initializer_list & ne, size_t offset, bool required = true); + + void done_getting_tensors() const; + + void init_mappings(bool prefetch = true, llama_mlocks * mlock_mmaps = nullptr); + + void get_mapping_range(size_t * first, size_t * last, void ** addr, int idx, ggml_context * ctx) const; + + // for backwards compatibility, does not support ggml-backend + void load_data_for(struct ggml_tensor * cur) const; + + // Returns false if cancelled by progress_callback + bool load_all_data( + struct ggml_context * ctx, + llama_buf_map & bufs, + llama_mlocks * lmlocks, + llama_progress_callback progress_callback, + void * progress_callback_user_data); + + std::string ftype_name() const; + + void print_info() const; +}; diff --git a/src/llama-model.cpp b/src/llama-model.cpp new file mode 100644 index 000000000..f90f5e746 --- /dev/null +++ b/src/llama-model.cpp @@ -0,0 +1,3954 @@ +#include "llama-model.h" + +#include "llama-impl.h" +#include "llama-mmap.h" +#include "llama-model-loader.h" + +#include "ggml-cpp.h" + +#include +#include +#include +#include +#include +#include +#include + +const char * llm_type_name(llm_type type) { + switch (type) { + case LLM_TYPE_14M: return "14M"; + case LLM_TYPE_17M: return "17M"; + case LLM_TYPE_22M: return "22M"; + case LLM_TYPE_33M: return "33M"; + case LLM_TYPE_60M: return "60M"; + case LLM_TYPE_70M: return "70M"; + case LLM_TYPE_80M: return "80M"; + case LLM_TYPE_109M: return "109M"; + case LLM_TYPE_137M: return "137M"; + case LLM_TYPE_160M: return "160M"; + case LLM_TYPE_220M: return "220M"; + case LLM_TYPE_250M: return "250M"; + case LLM_TYPE_270M: return "270M"; + case LLM_TYPE_335M: return "335M"; + case LLM_TYPE_410M: return "410M"; + case LLM_TYPE_450M: return "450M"; + case LLM_TYPE_770M: return "770M"; + case LLM_TYPE_780M: return "780M"; + case LLM_TYPE_0_5B: return "0.5B"; + case LLM_TYPE_1B: return "1B"; + case LLM_TYPE_1_3B: return "1.3B"; + case LLM_TYPE_1_4B: return "1.4B"; + case LLM_TYPE_1_5B: return "1.5B"; + case LLM_TYPE_1_6B: return "1.6B"; + case LLM_TYPE_2B: return "2B"; + case LLM_TYPE_2_8B: return "2.8B"; + case LLM_TYPE_3B: return "3B"; + case LLM_TYPE_4B: return "4B"; + case LLM_TYPE_6B: return "6B"; + case LLM_TYPE_6_9B: return "6.9B"; + case LLM_TYPE_7B: return "7B"; + case LLM_TYPE_8B: return "8B"; + case LLM_TYPE_9B: return "9B"; + case LLM_TYPE_11B: return "11B"; + case LLM_TYPE_12B: return "12B"; + case LLM_TYPE_13B: return "13B"; + case LLM_TYPE_14B: return "14B"; + case LLM_TYPE_15B: return "15B"; + case LLM_TYPE_16B: return "16B"; + case LLM_TYPE_20B: return "20B"; + case LLM_TYPE_30B: return "30B"; + case LLM_TYPE_32B: return "32B"; + case LLM_TYPE_34B: return "34B"; + case LLM_TYPE_35B: return "35B"; + case LLM_TYPE_40B: return "40B"; + case LLM_TYPE_65B: return "65B"; + case LLM_TYPE_70B: return "70B"; + case LLM_TYPE_236B: return "236B"; + case LLM_TYPE_314B: return "314B"; + case LLM_TYPE_671B: return "671B"; + case LLM_TYPE_SMALL: return "0.1B"; + case LLM_TYPE_MEDIUM: return "0.4B"; + case LLM_TYPE_LARGE: return "0.8B"; + case LLM_TYPE_XL: return "1.5B"; + case LLM_TYPE_A1_7B: return "A1.7B"; + case LLM_TYPE_A2_7B: return "A2.7B"; + case LLM_TYPE_8x7B: return "8x7B"; + case LLM_TYPE_8x22B: return "8x22B"; + case LLM_TYPE_16x12B: return "16x12B"; + case LLM_TYPE_16x3_8B: return "16x3.8B"; + case LLM_TYPE_10B_128x3_66B: return "10B+128x3.66B"; + case LLM_TYPE_57B_A14B: return "57B.A14B"; + case LLM_TYPE_27B: return "27B"; + default: return "?B"; + } +} + +static const char * llama_expert_gating_func_name(llama_expert_gating_func_type type) { + switch (type) { + case LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX: return "softmax"; + case LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID: return "sigmoid"; + default: return "unknown"; + } +} + +static const std::map LLAMA_ROPE_SCALING_TYPES = { + { LLAMA_ROPE_SCALING_TYPE_NONE, "none" }, + { LLAMA_ROPE_SCALING_TYPE_LINEAR, "linear" }, + { LLAMA_ROPE_SCALING_TYPE_YARN, "yarn" }, + { LLAMA_ROPE_SCALING_TYPE_LONGROPE, "longrope" }, +}; + +static llama_rope_scaling_type llama_rope_scaling_type_from_string(const std::string & name) { + for (const auto & kv : LLAMA_ROPE_SCALING_TYPES) { + if (kv.second == name) { + return (llama_rope_scaling_type) kv.first; + } + } + + return LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED; +} + +// checks if the weight tensor can be used with the specified buffer type and device +static bool weight_buft_supported(const llama_hparams & hparams, ggml_tensor * w, ggml_op op, ggml_backend_buffer_type_t buft, ggml_backend_dev_t dev) { + GGML_ASSERT(w != nullptr); + + if (op == GGML_OP_NONE) { + return true; + } + + ggml_init_params params = { + /*.mem_size =*/ ggml_tensor_overhead()*8, + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + ggml_context_ptr ctx_ptr { ggml_init(params) }; + if (!ctx_ptr) { + throw std::runtime_error(format("failed to create ggml context")); + } + ggml_context * ctx = ctx_ptr.get(); + + ggml_tensor * op_tensor = nullptr; + + switch (op) { + case GGML_OP_GET_ROWS: + { + ggml_tensor * b = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 512); + op_tensor = ggml_get_rows(ctx, w, b); + } break; + case GGML_OP_MUL_MAT: + { + ggml_tensor * b = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], 512, w->ne[2], w->ne[3]); + op_tensor = ggml_mul_mat(ctx, w, b); + } break; + case GGML_OP_MUL_MAT_ID: + { + int n_expert_used = hparams.n_expert_used; + ggml_tensor * b = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, w->ne[0], n_expert_used, 512); + ggml_tensor * ids = ggml_new_tensor_2d(ctx, GGML_TYPE_I32, n_expert_used, 512); + op_tensor = ggml_mul_mat_id(ctx, w, b, ids); + } break; + case GGML_OP_ADD: + { + ggml_tensor * a = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], w->ne[1], w->ne[2], w->ne[3]); + op_tensor = ggml_add(ctx, a, w); + } break; + case GGML_OP_MUL: + { + ggml_tensor * a = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], w->ne[1], w->ne[2], w->ne[3]); + op_tensor = ggml_mul(ctx, a, w); + } break; + case GGML_OP_DIV: + { + ggml_tensor * a = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, w->ne[0]); + op_tensor = ggml_div(ctx, a, w); + } break; + case GGML_OP_ROPE: + { + int n_embd_head = hparams.n_embd_head_v; + int n_head = hparams.n_head(); + ggml_tensor * a = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, n_embd_head, n_head, 512); + ggml_tensor * b = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 512); + op_tensor = ggml_rope_ext( + ctx, a, b, w, + 0, 0, 0, 0, 0, + 0, 0, 0, 0 + ); + + } break; + case GGML_OP_SSM_CONV: + { + // FIXME + ggml_tensor * conv_x = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, 12345, w->ne[1], 6789); + op_tensor = ggml_ssm_conv(ctx, conv_x, w); + } break; + case GGML_OP_SSM_SCAN: + { + // FIXME + const int64_t d_state = w->ne[0]; + const int64_t d_inner = w->ne[1]; + const int64_t n_seq_tokens = 512; + const int64_t n_seqs = 1; + ggml_tensor * s = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, d_inner, n_seqs); + ggml_tensor * x = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_inner, n_seq_tokens, n_seqs); + ggml_tensor * dt = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_inner, n_seq_tokens, n_seqs); + ggml_tensor * B = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, n_seq_tokens, n_seqs); + ggml_tensor * C = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, n_seq_tokens, n_seqs); + op_tensor = ggml_ssm_scan(ctx, s, x, dt, w, B, C); + } break; + case GGML_OP_RWKV_WKV6: + { + // FIXME + const int64_t S = 123; + const int64_t H = 123; + const int64_t n_tokens = 123; + const int64_t n_seqs = 123; + ggml_tensor * k = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, S, H, n_tokens); + ggml_tensor * v = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, S, H, n_tokens); + ggml_tensor * r = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, S, H, n_tokens); + ggml_tensor * tf = w; + ggml_tensor * td = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, S, H, n_tokens); + ggml_tensor * state = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, S, n_seqs, S, H); + op_tensor = ggml_rwkv_wkv6(ctx, k, v, r, tf, td, state); + } break; + case GGML_OP_IM2COL: + { + const int n_embd = hparams.n_embd; + ggml_tensor * b = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, n_embd, w->ne[1], 1, 1); + op_tensor = ggml_im2col(ctx, w, b, 1, 0, 0, 0, 1, 0, false, GGML_TYPE_F16); + } break; + default: + GGML_ABORT("%s: missing test for op %s for tensor %s", __func__, ggml_op_name(op), w->name); + } + + // create a temporary dummy buffer for the weight so that supports_op can check the buffer type + GGML_ASSERT(w->buffer == nullptr); + w->buffer = ggml_backend_buft_alloc_buffer(buft, 0); + bool op_supported = ggml_backend_dev_supports_op(dev, op_tensor); + ggml_backend_buffer_free(w->buffer); + w->buffer = nullptr; + + return op_supported; +} + +// lists of buffer types used for each layer +using buft_list_t = std::vector>; + +// find the first buffer type in the list that can use the tensor +static ggml_backend_buffer_type_t select_weight_buft(const llama_hparams & hparams, ggml_tensor * tensor, ggml_op op, const buft_list_t & buft_list) { + GGML_ASSERT(!buft_list.empty()); + for (const auto & cur : buft_list) { + ggml_backend_dev_t cur_dev = cur.first; + ggml_backend_buffer_type_t cur_buft = cur.second; + if (weight_buft_supported(hparams, tensor, op, cur_buft, cur_dev)) { + return cur_buft; + } + } + return nullptr; +} + +// CPU: ACCEL -> CPU extra -> GPU host -> CPU +static buft_list_t make_cpu_buft_list(const std::vector & devices) { + buft_list_t buft_list; + + // add ACCEL buffer types + for (size_t i = 0; i < ggml_backend_dev_count(); ++i) { + ggml_backend_dev_t dev = ggml_backend_dev_get(i); + if (ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_ACCEL) { + auto * buft = ggml_backend_dev_buffer_type(dev); + // skip + if (buft != ggml_backend_cpu_buffer_type()) { + buft_list.emplace_back(dev, buft); + } + } + } + + // add extra buffer types + auto * cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + auto * cpu_reg = ggml_backend_dev_backend_reg(cpu_dev); + auto ggml_backend_dev_get_extra_bufts_fn = (ggml_backend_dev_get_extra_bufts_t) + ggml_backend_reg_get_proc_address(cpu_reg, "ggml_backend_dev_get_extra_bufts"); + if (ggml_backend_dev_get_extra_bufts_fn) { + ggml_backend_buffer_type_t * extra_bufts = ggml_backend_dev_get_extra_bufts_fn(cpu_dev); + while (extra_bufts && *extra_bufts) { + buft_list.emplace_back(cpu_dev, *extra_bufts); + ++extra_bufts; + } + } + + // add a host buffer type + // storing the tensors in a host buffer is useful when the processing of large batches + // is offloaded to a GPU device, since it reduces the time spent on data transfers + // generally, this will be done using the first device in the list + // a better approach would be to handle this on a weight-by-weight basis using the offload_op + // function of the device to determine if it would benefit from being stored in a host buffer + for (auto * dev : devices) { + ggml_backend_buffer_type_t buft = ggml_backend_dev_host_buffer_type(dev); + if (buft) { + buft_list.emplace_back(dev, buft); + break; + } + } + + // add the CPU buffer type + for (size_t i = 0; i < ggml_backend_dev_count(); ++i) { + ggml_backend_dev_t dev = ggml_backend_dev_get(i); + if (ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_CPU) { + buft_list.emplace_back(dev, ggml_backend_dev_buffer_type(dev)); + } + } + + return buft_list; +} + +// GPU: split if LLAMA_SPLIT_MODE_ROW -> GPU +static buft_list_t make_gpu_buft_list(ggml_backend_dev_t dev, enum llama_split_mode split_mode, const float * tensor_split) { + buft_list_t buft_list; + + // add the device split buffer type if requested and available + if (split_mode == LLAMA_SPLIT_MODE_ROW) { + ggml_backend_reg_t reg = ggml_backend_dev_backend_reg(dev); + auto ggml_backend_split_buffer_type_fn = (ggml_backend_split_buffer_type_t) + ggml_backend_reg_get_proc_address(reg, "ggml_backend_split_buffer_type"); + if (ggml_backend_split_buffer_type_fn) { + size_t dev_index = [&]() { + auto * reg = ggml_backend_dev_backend_reg(dev); + for (size_t i = 0; i < ggml_backend_reg_dev_count(reg); ++i) { + if (ggml_backend_reg_dev_get(reg, i) == dev) { + return i; + } + } + throw std::runtime_error(format("device %s not found in its backend reg", ggml_backend_dev_name(dev))); + }(); + auto * buft = ggml_backend_split_buffer_type_fn(dev_index, tensor_split); + if (buft != nullptr) { + buft_list.emplace_back(dev, buft); + } + } + } + + // add the device default buffer type + buft_list.emplace_back(dev, ggml_backend_dev_buffer_type(dev)); + + return buft_list; +} + +struct llama_model::impl { + impl() {} + ~impl() {} + + uint64_t n_elements = 0; + + size_t n_bytes = 0; + + std::string desc_str; + + // model memory mapped files + llama_mmaps mappings; + + // objects representing data potentially being locked in memory + llama_mlocks mlock_bufs; + llama_mlocks mlock_mmaps; + + // contexts where the model tensors metadata is stored + std::vector ctxs; + + // the model memory buffers for the tensor data + std::vector bufs; + + buft_list_t cpu_buft_list; + std::map gpu_buft_list; + + struct layer_dev { + ggml_backend_dev_t dev; + buft_list_t * buft_list; + }; + + layer_dev dev_input = {}; + layer_dev dev_output = {}; + std::vector dev_layer; +}; + +llama_model::llama_model(const struct llama_model_params & params) : params(params), pimpl(std::make_unique()) { +} + +llama_model::~llama_model() {} + +void llama_model::load_stats(llama_model_loader & ml) { + pimpl->n_elements = ml.n_elements; + pimpl->n_bytes = ml.n_bytes; +} + +void llama_model::load_arch(llama_model_loader & ml) { + arch = ml.get_arch(); + if (arch == LLM_ARCH_UNKNOWN) { + throw std::runtime_error("unknown model architecture: '" + ml.get_arch_name() + "'"); + } +} + +void llama_model::load_hparams(llama_model_loader & ml) { + const gguf_context * ctx = ml.meta.get(); + + // get metadata as string + for (int i = 0; i < gguf_get_n_kv(ctx); i++) { + enum gguf_type type = gguf_get_kv_type(ctx, i); + if (type == GGUF_TYPE_ARRAY) { + continue; + } + const char * name = gguf_get_key(ctx, i); + const std::string value = gguf_kv_to_str(ctx, i); + gguf_kv.emplace(name, value); + } + + // get general kv + ml.get_key(LLM_KV_GENERAL_NAME, name, false); + + // everything past this point is not vocab-related + if (hparams.vocab_only) { + return; + } + + ml.get_key(LLM_KV_CONTEXT_LENGTH, hparams.n_ctx_train); + ml.get_key(LLM_KV_EMBEDDING_LENGTH, hparams.n_embd); + ml.get_key(LLM_KV_BLOCK_COUNT, hparams.n_layer); + ml.get_key(LLM_KV_EXPERT_COUNT, hparams.n_expert, false); + ml.get_key(LLM_KV_EXPERT_USED_COUNT, hparams.n_expert_used, false); + + if (arch == LLM_ARCH_WAVTOKENIZER_DEC) { + ml.get_key(LLM_KV_FEATURES_LENGTH, hparams.n_embd_features); + + ml.get_key(LLM_KV_POSNET_EMBEDDING_LENGTH, hparams.posnet.n_embd); + ml.get_key(LLM_KV_POSNET_BLOCK_COUNT, hparams.posnet.n_layer); + + ml.get_key(LLM_KV_CONVNEXT_EMBEDDING_LENGTH, hparams.convnext.n_embd); + ml.get_key(LLM_KV_CONVNEXT_BLOCK_COUNT, hparams.convnext.n_layer); + } + + GGML_ASSERT(hparams.n_expert <= LLAMA_MAX_EXPERTS); + GGML_ASSERT(hparams.n_expert_used <= hparams.n_expert); + if (hparams.n_expert > 0) { + GGML_ASSERT(hparams.n_expert_used > 0); + } else { + GGML_ASSERT(hparams.n_expert_used == 0); + } + + // zero-out the array hparams + std::fill(hparams.n_head_arr.begin(), hparams.n_head_arr.end(), 0); + std::fill(hparams.n_head_kv_arr.begin(), hparams.n_head_kv_arr.end(), 0); + std::fill(hparams.n_ff_arr.begin(), hparams.n_ff_arr.end(), 0); + + ml.get_key_or_arr(LLM_KV_FEED_FORWARD_LENGTH, hparams.n_ff_arr, hparams.n_layer, false); + ml.get_key_or_arr(LLM_KV_ATTENTION_HEAD_COUNT, hparams.n_head_arr, hparams.n_layer, false); + + // n_head_kv is optional, default to n_head + hparams.n_head_kv_arr = hparams.n_head_arr; + + ml.get_key_or_arr(LLM_KV_ATTENTION_HEAD_COUNT_KV, hparams.n_head_kv_arr, hparams.n_layer, false); + + bool rope_finetuned = false; + ml.get_key(LLM_KV_ROPE_SCALING_FINETUNED, rope_finetuned, false); + hparams.rope_finetuned = rope_finetuned; + + hparams.n_ctx_orig_yarn = hparams.n_ctx_train; + ml.get_key(LLM_KV_ROPE_SCALING_ORIG_CTX_LEN, hparams.n_ctx_orig_yarn, false); + + // rope_freq_base (optional) + hparams.rope_freq_base_train = 10000.0f; + ml.get_key(LLM_KV_ROPE_FREQ_BASE, hparams.rope_freq_base_train, false); + + std::string rope_scaling("linear"); + ml.get_key(LLM_KV_ROPE_SCALING_TYPE, rope_scaling, false); + hparams.rope_scaling_type_train = llama_rope_scaling_type_from_string(rope_scaling); + GGML_ASSERT(hparams.rope_scaling_type_train != LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED); + + // rope_freq_scale (inverse of the kv) is optional + float ropescale = 0.0f; + if (!ml.get_key(LLM_KV_ROPE_SCALING_FACTOR, ropescale, false)) { + // try the old key name + ml.get_key(LLM_KV_ROPE_SCALE_LINEAR, ropescale, false); + } + hparams.rope_freq_scale_train = ropescale == 0.0f ? 1.0f : 1.0f/ropescale; + + ml.get_key(LLM_KV_ROPE_SCALING_ATTN_FACTOR, hparams.rope_attn_factor, false); + + // non-transformer models do not have attention heads + if (hparams.n_head() > 0) { + // gpt-neox n_rot = rotary_pct * (n_embd / n_head) + // gpt-j n_rot = rotary_dim + + hparams.n_embd_head_k = hparams.n_embd / hparams.n_head(); + ml.get_key(LLM_KV_ATTENTION_KEY_LENGTH, hparams.n_embd_head_k, false); + + hparams.n_embd_head_v = hparams.n_embd / hparams.n_head(); + ml.get_key(LLM_KV_ATTENTION_VALUE_LENGTH, hparams.n_embd_head_v, false); + + // sanity check for n_rot (optional) + hparams.n_rot = hparams.n_embd_head_k; + + ml.get_key(LLM_KV_ROPE_DIMENSION_COUNT, hparams.n_rot, false); + + if (arch == LLM_ARCH_LLAMA || arch == LLM_ARCH_DECI || arch == LLM_ARCH_FALCON) { + if (hparams.n_rot != hparams.n_embd_head_k) { + throw std::runtime_error(format("invalid n_rot: %u, expected %u", hparams.n_rot, hparams.n_embd_head_k)); + } + } + } else { + hparams.n_rot = 0; + hparams.n_embd_head_k = 0; + hparams.n_embd_head_v = 0; + } + + // for differentiating model types + uint32_t n_vocab = 0; + ml.get_key(LLM_KV_VOCAB_SIZE, n_vocab, false) || ml.get_arr_n(LLM_KV_TOKENIZER_LIST, n_vocab, false); + + // arch-specific KVs + switch (arch) { + case LLM_ARCH_LLAMA: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + if (hparams.n_expert == 8) { + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_8x7B; break; + case 56: type = LLM_TYPE_8x22B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } else { + switch (hparams.n_layer) { + case 16: type = LLM_TYPE_1B; break; // Llama 3.2 1B + case 22: type = LLM_TYPE_1B; break; + case 26: type = LLM_TYPE_3B; break; + case 28: type = LLM_TYPE_3B; break; // Llama 3.2 3B + // granite uses a vocab with len 49152 + case 32: type = n_vocab == 49152 ? LLM_TYPE_3B : (n_vocab < 40000 ? LLM_TYPE_7B : LLM_TYPE_8B); break; + case 36: type = LLM_TYPE_8B; break; // granite + case 40: type = LLM_TYPE_13B; break; + case 48: type = LLM_TYPE_34B; break; + case 60: type = LLM_TYPE_30B; break; + case 80: type = hparams.n_head() == hparams.n_head_kv() ? LLM_TYPE_65B : LLM_TYPE_70B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } + } break; + case LLM_ARCH_DECI: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 80: type = LLM_TYPE_70B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_MINICPM: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_EMBEDDING_SCALE, hparams.f_embedding_scale); + ml.get_key(LLM_KV_RESIDUAL_SCALE, hparams.f_residual_scale); + ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale); + + switch (hparams.n_layer) { + case 52: type = LLM_TYPE_1B; break; + case 40: type = LLM_TYPE_2B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_MINICPM3: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q); + ml.get_key(LLM_KV_ATTENTION_KV_LORA_RANK, hparams.n_lora_kv); + + switch (hparams.n_layer) { + case 62: type = LLM_TYPE_4B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GROK: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 64: type = LLM_TYPE_314B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_FALCON: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 60: type = LLM_TYPE_40B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_BAICHUAN: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_13B; break; + default: type = LLM_TYPE_UNKNOWN; + } + + if (type == LLM_TYPE_13B) { + // TODO: become GGUF KV parameter + hparams.f_max_alibi_bias = 8.0f; + } + } break; + case LLM_ARCH_STARCODER: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1B; break; + case 36: type = LLM_TYPE_3B; break; + case 42: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_15B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_REFACT: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_1B; break; + default: type = LLM_TYPE_UNKNOWN; + } + + // TODO: become GGUF KV parameter + hparams.f_max_alibi_bias = 8.0f; + } break; + case LLM_ARCH_BERT: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CAUSAL, hparams.causal_attn); + ml.get_key(LLM_KV_POOLING_TYPE, hparams.pooling_type, false); + + switch (hparams.n_layer) { + case 3: + type = LLM_TYPE_17M; break; // bge-micro + case 6: + type = LLM_TYPE_22M; break; // MiniLM-L6 + case 12: + switch (hparams.n_embd) { + case 384: type = LLM_TYPE_33M; break; // MiniLM-L12, bge-small + case 768: type = LLM_TYPE_109M; break; // bge-base + default: type = LLM_TYPE_UNKNOWN; + } break; + case 24: + type = LLM_TYPE_335M; break; // bge-large + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_JINA_BERT_V2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CAUSAL, hparams.causal_attn); + ml.get_key(LLM_KV_POOLING_TYPE, hparams.pooling_type, false); + hparams.f_max_alibi_bias = 8.0f; + + switch (hparams.n_layer) { + case 4: type = LLM_TYPE_33M; break; // jina-embeddings-small + case 12: type = LLM_TYPE_137M; break; // jina-embeddings-base + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_NOMIC_BERT: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CAUSAL, hparams.causal_attn); + ml.get_key(LLM_KV_POOLING_TYPE, hparams.pooling_type); + + if (hparams.n_layer == 12 && hparams.n_embd == 768) { + type = LLM_TYPE_137M; + } + } break; + case LLM_ARCH_BLOOM: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1B; break; + case 30: + switch (hparams.n_embd) { + case 2560: type = LLM_TYPE_3B; break; + case 4096: type = LLM_TYPE_7B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + default: type = LLM_TYPE_UNKNOWN; + } + + // TODO: become GGUF KV parameter + hparams.f_max_alibi_bias = 8.0f; + } break; + case LLM_ARCH_MPT: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV, hparams.f_clamp_kqv, false); + ml.get_key(LLM_KV_ATTENTION_MAX_ALIBI_BIAS, hparams.f_max_alibi_bias); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 48: type = LLM_TYPE_30B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_STABLELM: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1B; break; + case 32: type = LLM_TYPE_3B; break; + case 40: type = LLM_TYPE_12B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_QWEN: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_13B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_QWEN2VL: + { + ml.get_key_or_arr(LLM_KV_ROPE_DIMENSION_SECTIONS, hparams.rope_sections, 4, true); + } + // fall through + case LLM_ARCH_QWEN2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 24: type = hparams.n_embd == 1024 ? LLM_TYPE_0_5B : LLM_TYPE_1B; break; + case 28: type = hparams.n_embd == 1536 ? LLM_TYPE_1_5B : LLM_TYPE_7B; break; + case 32: type = LLM_TYPE_7B; break; + case 36: type = LLM_TYPE_3B; break; + case 40: type = hparams.n_head() == 20 ? LLM_TYPE_4B : LLM_TYPE_13B; break; + case 48: type = LLM_TYPE_14B; break; + case 64: type = LLM_TYPE_32B; break; + case 80: type = LLM_TYPE_70B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_QWEN2MOE: + { + ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp, false); + ml.get_key(LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, hparams.n_ff_shexp, false); + + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_A2_7B; break; + case 28: type = LLM_TYPE_57B_A14B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_PHI2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1B; break; + case 32: type = LLM_TYPE_3B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_PHI3: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1B; break; + case 32: type = LLM_TYPE_3B; break; + case 40: type = LLM_TYPE_14B; break; + default: type = LLM_TYPE_UNKNOWN; + } + + // for backward compatibility ; see: https://github.com/ggerganov/llama.cpp/pull/8931 + if ((hparams.n_layer == 32 || hparams.n_layer == 40) && hparams.n_ctx_train == 4096) { + // default value for Phi-3-mini-4k-instruct and Phi-3-medium-4k-instruct + hparams.n_swa = 2047; + } else if (hparams.n_layer == 32 && hparams.n_head_kv(0) == 32 && hparams.n_ctx_train == 131072) { + // default value for Phi-3-mini-128k-instruct + hparams.n_swa = 262144; + } else if (hparams.n_layer == 40 && hparams.n_ctx_train == 131072) { + // default value for Phi-3-medium-128k-instruct + hparams.n_swa = 131072; + } + bool found_swa = ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false); + if (!found_swa && hparams.n_swa == 0) { + throw std::runtime_error("invalid value for sliding_window"); + } + } break; + case LLM_ARCH_PHIMOE: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_16x3_8B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_PLAMO: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 40: type = LLM_TYPE_13B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GPT2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 12: type = LLM_TYPE_SMALL; break; + case 24: type = LLM_TYPE_MEDIUM; break; + case 36: type = LLM_TYPE_LARGE; break; + case 48: type = LLM_TYPE_XL; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_CODESHELL: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 42: type = LLM_TYPE_7B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_ORION: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + + switch (hparams.n_layer) { + case 40: type = LLM_TYPE_14B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_INTERNLM2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 48: type = LLM_TYPE_20B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GEMMA: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 18: type = LLM_TYPE_2B; break; + case 28: type = LLM_TYPE_7B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GEMMA2: + { + hparams.n_swa = 4096; // default value of gemma 2 + ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_ATTN_LOGIT_SOFTCAPPING, hparams.f_attn_logit_softcapping, false); + ml.get_key(LLM_KV_FINAL_LOGIT_SOFTCAPPING, hparams.f_final_logit_softcapping, false); + hparams.attn_soft_cap = true; + + switch (hparams.n_layer) { + case 26: type = LLM_TYPE_2B; break; + case 42: type = LLM_TYPE_9B; break; + case 46: type = LLM_TYPE_27B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_STARCODER2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 30: type = LLM_TYPE_3B; break; + case 32: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_15B; break; + case 52: type = LLM_TYPE_20B; break; // granite + case 88: type = LLM_TYPE_34B; break; // granite + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_MAMBA: + { + ml.get_key(LLM_KV_SSM_CONV_KERNEL, hparams.ssm_d_conv); + ml.get_key(LLM_KV_SSM_INNER_SIZE, hparams.ssm_d_inner); + ml.get_key(LLM_KV_SSM_STATE_SIZE, hparams.ssm_d_state); + ml.get_key(LLM_KV_SSM_TIME_STEP_RANK, hparams.ssm_dt_rank); + ml.get_key(LLM_KV_SSM_DT_B_C_RMS, hparams.ssm_dt_b_c_rms, false); + + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 24: + switch (hparams.n_embd) { + case 768: type = LLM_TYPE_SMALL; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 48: + switch (hparams.n_embd) { + case 1024: type = LLM_TYPE_MEDIUM; break; + case 1536: type = LLM_TYPE_LARGE; break; + case 2048: type = LLM_TYPE_XL; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 64: + switch (hparams.n_embd) { + case 2560: type = LLM_TYPE_3B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_XVERSE: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_13B; break; + case 80: type = LLM_TYPE_65B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_COMMAND_R: + { + ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 40: type = LLM_TYPE_35B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_COHERE2: + { + ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa); + ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_8B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_DBRX: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV, hparams.f_clamp_kqv); + + switch (hparams.n_layer) { + case 40: type = LLM_TYPE_16x12B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_OLMO: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV, hparams.f_clamp_kqv, false); + + switch (hparams.n_layer) { + case 22: type = LLM_TYPE_1B; break; + case 32: type = LLM_TYPE_7B; break; + case 80: type = LLM_TYPE_70B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_OLMO2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 16: type = LLM_TYPE_1B; break; + case 32: type = LLM_TYPE_7B; break; + case 40: type = LLM_TYPE_13B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_OLMOE: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 16: type = LLM_TYPE_A1_7B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_OPENELM: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 16: type = LLM_TYPE_270M; break; + case 20: type = LLM_TYPE_450M; break; + case 28: type = LLM_TYPE_1B; break; + case 36: type = LLM_TYPE_3B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GPTNEOX: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_USE_PARALLEL_RESIDUAL, hparams.use_par_res); + switch (hparams.n_layer) { + case 6: + switch (hparams.n_ff()) { + case 512: type = LLM_TYPE_14M; break; + case 2048: type = LLM_TYPE_70M; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 12: + switch (hparams.n_ff()) { + case 3072: type = LLM_TYPE_160M; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 16: + switch (hparams.n_ff()) { + case 8192: type = LLM_TYPE_1B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 24: + switch (hparams.n_ff()) { + case 4096: type = LLM_TYPE_410M; break; + case 8192: type = LLM_TYPE_1_4B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 32: + switch (hparams.n_ff()) { + case 10240: type = LLM_TYPE_2_8B; break; + case 16384: type = LLM_TYPE_6_9B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 36: + switch (hparams.n_ff()) { + case 20480: type = LLM_TYPE_12B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 44: + switch (hparams.n_ff()) { + case 24576: type = LLM_TYPE_20B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_ARCTIC: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + if (hparams.n_expert == 128) { + switch (hparams.n_layer) { + case 35: type = LLM_TYPE_10B_128x3_66B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } else { + type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_DEEPSEEK: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT, hparams.n_layer_dense_lead); + ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp); + ml.get_key(LLM_KV_EXPERT_SHARED_COUNT, hparams.n_expert_shared); + ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE, hparams.expert_weights_scale); + + switch (hparams.n_layer) { + case 28: type = LLM_TYPE_20B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_DEEPSEEK2: + { + bool is_lite = (hparams.n_layer == 27); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT, hparams.n_layer_dense_lead); + if (!is_lite) { + ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q); + } + ml.get_key(LLM_KV_ATTENTION_KV_LORA_RANK, hparams.n_lora_kv); + ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp); + ml.get_key(LLM_KV_EXPERT_SHARED_COUNT, hparams.n_expert_shared); + ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE, hparams.expert_weights_scale); + ml.get_key(LLM_KV_EXPERT_WEIGHTS_NORM, hparams.expert_weights_norm, false); + ml.get_key(LLM_KV_EXPERT_GATING_FUNC, hparams.expert_gating_func, false); + if (hparams.expert_gating_func == LLAMA_EXPERT_GATING_FUNC_TYPE_NONE) { + // for compatibility with existing DeepSeek V2 and V2.5 GGUFs + // that have no expert_gating_func model parameter set + hparams.expert_gating_func = LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX; + } + ml.get_key(LLM_KV_ROPE_SCALING_YARN_LOG_MUL, hparams.rope_yarn_log_mul); + + switch (hparams.n_layer) { + case 27: type = LLM_TYPE_16B; break; + case 60: type = LLM_TYPE_236B; break; + case 61: type = LLM_TYPE_671B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_CHATGLM: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + switch (hparams.n_layer) { + case 28: type = LLM_TYPE_6B; break; + case 40: type = LLM_TYPE_9B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_BITNET: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 26: type = LLM_TYPE_3B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_T5: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, hparams.n_rel_attn_bkts); + + uint32_t dec_start_token_id; + if (ml.get_key(LLM_KV_DECODER_START_TOKEN_ID, dec_start_token_id, false)) { + hparams.dec_start_token_id = dec_start_token_id; + } + + switch (hparams.n_layer) { + case 6: type = LLM_TYPE_60M; break; // t5-small + case 8: type = LLM_TYPE_80M; break; // flan-t5-small + case 12: + switch (hparams.n_ff()) { + case 3072: type = LLM_TYPE_220M; break; // t5-base + case 2048: type = LLM_TYPE_250M; break; // flan-t5-base + default: type = LLM_TYPE_UNKNOWN; + } break; + case 24: + switch (hparams.n_ff()) { + case 4096: type = LLM_TYPE_770M; break; // t5-large + case 2816: type = LLM_TYPE_780M; break; // flan-t5-large + case 16384: type = LLM_TYPE_3B; break; // t5-3b + case 5120: type = LLM_TYPE_3B; break; // flan-t5-xl + case 65536: type = LLM_TYPE_11B; break; // t5-11b + case 10240: type = LLM_TYPE_11B; break; // flan-t5-xxl + default: type = LLM_TYPE_UNKNOWN; + } break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_T5ENCODER: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, hparams.n_rel_attn_bkts); + type = LLM_TYPE_UNKNOWN; + } break; + case LLM_ARCH_JAIS: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_MAX_ALIBI_BIAS, hparams.f_max_alibi_bias); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1_3B; break; + case 40: type = LLM_TYPE_13B; break; + /* TODO: add variants */ + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_NEMOTRON: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_4B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_EXAONE: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_8B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_RWKV6: + case LLM_ARCH_RWKV6QWEN2: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps, false); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps, false); + ml.get_key(LLM_KV_WKV_HEAD_SIZE, hparams.wkv_head_size); + ml.get_key(LLM_KV_TIME_MIX_EXTRA_DIM, hparams.time_mix_extra_dim); + ml.get_key(LLM_KV_TIME_DECAY_EXTRA_DIM, hparams.time_decay_extra_dim); + ml.get_key(LLM_KV_RESCALE_EVERY_N_LAYERS, hparams.rescale_every_n_layers, false); + ml.get_key(LLM_KV_TOKEN_SHIFT_COUNT, hparams.token_shift_count, false); + + switch (hparams.n_layer) { + case 24: type = LLM_TYPE_1_6B; break; + case 32: + switch (hparams.n_embd) { + case 2560: type = LLM_TYPE_3B; break; + case 4096: type = LLM_TYPE_7B; break; + default: type = LLM_TYPE_UNKNOWN; + } break; + case 61: type = LLM_TYPE_14B; break; + case 64: type = LLM_TYPE_32B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_GRANITE: + case LLM_ARCH_GRANITE_MOE: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale); + ml.get_key(LLM_KV_RESIDUAL_SCALE, hparams.f_residual_scale); + ml.get_key(LLM_KV_EMBEDDING_SCALE, hparams.f_embedding_scale); + ml.get_key(LLM_KV_ATTENTION_SCALE, hparams.f_attention_scale); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_3B; break; + case 40: type = LLM_TYPE_3B; break; + // Add additional layer/vocab/etc checks here for other model sizes + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_CHAMELEON: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + hparams.f_norm_eps = 1e-5; // eps for qk-norm, torch default + ml.get_key(LLM_KV_SWIN_NORM, hparams.swin_norm); + + switch (hparams.n_layer) { + case 32: type = LLM_TYPE_7B; break; + case 48: type = LLM_TYPE_34B; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; + case LLM_ARCH_WAVTOKENIZER_DEC: + { + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps); + ml.get_key(LLM_KV_ATTENTION_GROUPNORM_EPS, hparams.f_norm_group_eps); + ml.get_key(LLM_KV_ATTENTION_GROUPNORM_GROUPS, hparams.n_norm_groups); + ml.get_key(LLM_KV_ATTENTION_CAUSAL, hparams.causal_attn); + } break; + default: throw std::runtime_error("unsupported model architecture"); + } + + pimpl->n_bytes = ml.n_bytes; + + pimpl->desc_str = arch_name() + " " + type_name() + " " + ml.ftype_name(); + + if (hparams.f_max_alibi_bias > 0.0f) { + hparams.use_alibi = true; + } + + hparams.rope_type = llama_model_rope_type(this); +} + +void llama_model::load_vocab(llama_model_loader & ml) { + const auto kv = LLM_KV(arch); + + vocab.load(ml, kv); +} + +bool llama_model::load_tensors(llama_model_loader & ml) { + const auto & split_mode = params.split_mode; + const auto & n_gpu_layers = params.n_gpu_layers; + const auto & use_mlock = params.use_mlock; + const auto & tensor_split = params.tensor_split; + + const int n_layer = hparams.n_layer; + + const bool use_mmap_buffer = true; + + // build a list of buffer types for the CPU and GPU devices + pimpl->cpu_buft_list = make_cpu_buft_list(devices); + for (auto * dev : devices) { + buft_list_t buft_list = make_gpu_buft_list(dev, split_mode, tensor_split); + // add CPU buffer types as a fallback + buft_list.insert(buft_list.end(), pimpl->cpu_buft_list.begin(), pimpl->cpu_buft_list.end()); + pimpl->gpu_buft_list.emplace(dev, std::move(buft_list)); + } + + // calculate the split points + bool all_zero = tensor_split == nullptr || std::all_of(tensor_split, tensor_split + n_devices(), [](float x) { return x == 0.0f; }); + std::vector splits(n_devices()); + if (all_zero) { + // default split, by free memory + for (size_t i = 0; i < n_devices(); ++i) { + ggml_backend_dev_t dev = devices[i]; + size_t total; + size_t free; + ggml_backend_dev_memory(dev, &free, &total); + splits[i] = free; + } + } else { + std::copy(tensor_split, tensor_split + n_devices(), splits.begin()); + } + + // sum and normalize the splits to get the split points + float split_sum = 0.0f; + for (size_t i = 0; i < n_devices(); ++i) { + split_sum += splits[i]; + splits[i] = split_sum; + } + for (size_t i = 0; i < n_devices(); ++i) { + splits[i] /= split_sum; + } + + ggml_backend_dev_t cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + const int i_gpu_start = std::max((int) hparams.n_layer - n_gpu_layers, (int) 0); + const int act_gpu_layers = devices.empty() ? 0 : std::min(n_gpu_layers, (int)n_layer + 1); + auto get_layer_buft_list = [&](int il) -> llama_model::impl::layer_dev { + if (il < i_gpu_start || (il - i_gpu_start) >= act_gpu_layers) { + return {cpu_dev, &pimpl->cpu_buft_list}; + } + const int layer_gpu = std::upper_bound(splits.begin(), splits.begin() + n_devices(), float(il - i_gpu_start)/act_gpu_layers) - splits.begin(); + auto * dev = devices.at(layer_gpu); + return {dev, &pimpl->gpu_buft_list.at(dev)}; + }; + + // assign the input layer + // there is very little benefit to offloading the input layer, so always keep it on the CPU + pimpl->dev_input = { cpu_dev, &pimpl->cpu_buft_list }; + + // assign the repeating layers to the devices according to the splits + pimpl->dev_layer.resize(n_layer); + for (int il = 0; il < n_layer; ++il) { + pimpl->dev_layer[il] = get_layer_buft_list(il); + } + + // assign the output layer + pimpl->dev_output = get_layer_buft_list(n_layer); + + // one ggml context per buffer type + int max_n_tensors = ml.n_tensors; + max_n_tensors += 1; // duplicated output tensor + max_n_tensors += n_layer*2; // duplicated rope freq tensors + const size_t ctx_size = ggml_tensor_overhead()*max_n_tensors; + + std::map ctx_map; + auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * { + auto it = ctx_map.find(buft); + if (it == ctx_map.end()) { + ggml_init_params params = { + /*.mem_size =*/ ctx_size, + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + + ggml_context * ctx = ggml_init(params); + if (!ctx) { + throw std::runtime_error(format("failed to create ggml context")); + } + + ctx_map[buft] = ctx; + pimpl->ctxs.emplace_back(ctx); + + return ctx; + } + return it->second; + }; + + const auto TENSOR_DUPLICATED = llama_model_loader::TENSOR_DUPLICATED; + const auto TENSOR_NOT_REQUIRED = llama_model_loader::TENSOR_NOT_REQUIRED; + + // create tensors for the weights + { + // note: cast to int64_t since we will use these for the tensor dimensions + const int64_t n_head = hparams.n_head(); + const int64_t n_head_kv = hparams.n_head_kv(); + const int64_t n_embd = hparams.n_embd; + const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(); + const int64_t n_embd_v_gqa = hparams.n_embd_v_gqa(); + const int64_t n_embd_head_k = hparams.n_embd_head_k; + const int64_t n_embd_head_v = hparams.n_embd_head_v; + const int64_t n_ff = hparams.n_ff(); + const int64_t n_embd_gqa = n_embd_v_gqa; + const int64_t n_vocab = vocab.n_tokens(); + const int64_t n_token_types = vocab.n_token_types(); + const int64_t n_rot = hparams.n_rot; + const int64_t n_expert = hparams.n_expert; + const int64_t n_expert_used = hparams.n_expert_used; + const int64_t n_ctx_train = hparams.n_ctx_train; + + if (n_expert > 0 && hparams.n_expert_used == 0) { + throw std::runtime_error("model has expert layers but no expert layers are used"); + } + + int n_moved_tensors = 0; + ggml_tensor * first_moved_tensor = nullptr; + ggml_backend_buffer_type_t first_moved_from_buft = nullptr; + ggml_backend_buffer_type_t first_moved_to_buft = nullptr; + + auto create_tensor = [&](const LLM_TN_IMPL & tn, const std::initializer_list & ne, int flags) -> ggml_tensor * { + ggml_tensor * t_meta = ml.get_tensor_meta(tn.str().c_str()); + + if (!t_meta) { + if (flags & TENSOR_NOT_REQUIRED) { + return nullptr; + } + throw std::runtime_error(format("missing tensor '%s'", tn.str().c_str())); + } + + // some models use the token embedding tensor as the output, but since these are used in different layers and with different ops + // the tensor is duplicated + // to handle this, we check if the tensor is duplicated, and if so, we assume that it is being loaded as the output tensor + llm_tensor tn_tensor = tn.tensor; + if (tn.tensor == LLM_TENSOR_TOKEN_EMBD && flags & TENSOR_DUPLICATED) { + tn_tensor = LLM_TENSOR_OUTPUT; + } + + llm_tensor_info info; + try { + info = llm_tensor_info_for(tn_tensor); + } catch (const std::out_of_range & e) { + throw std::runtime_error(format("missing tensor info mapping for %s", tn.str().c_str())); + } + + // tensors with "bias" suffix are always used with GGML_OP_ADD + ggml_op op; + bool bias = tn.suffix != nullptr && strcmp(tn.suffix, "bias") == 0; + if (bias) { + op = GGML_OP_ADD; + } else { + op = info.op; + } + + // sanity checks + if (info.layer == LLM_TENSOR_LAYER_INPUT || info.layer == LLM_TENSOR_LAYER_OUTPUT) { + if (tn.bid != -1) { + GGML_ABORT("input/output layer tensor %s used with a layer number", tn.str().c_str()); + } + } else { + if (tn.bid == -1) { + GGML_ABORT("repeating layer tensor %s used without a layer number", tn.str().c_str()); + } + } + + // select the buffer type for this tensor + buft_list_t * buft_list; + switch (info.layer) { + case LLM_TENSOR_LAYER_INPUT: + buft_list = pimpl->dev_input.buft_list; + break; + case LLM_TENSOR_LAYER_OUTPUT: + buft_list = pimpl->dev_output.buft_list; + break; + case LLM_TENSOR_LAYER_REPEATING: + buft_list = pimpl->dev_layer.at(tn.bid).buft_list; + break; + default: + GGML_ABORT("invalid layer %d for tensor %s", info.layer, tn.str().c_str()); + } + + ggml_backend_buffer_type_t buft = select_weight_buft(hparams, t_meta, op, *buft_list); + if (!buft) { + throw std::runtime_error(format("failed to find a compatible buffer type for tensor %s", tn.str().c_str())); + } + + // avoid using a host buffer when using mmap + auto * buft_dev = ggml_backend_buft_get_device(buft); + if (ml.use_mmap && buft_dev && buft == ggml_backend_dev_host_buffer_type(buft_dev)) { + auto * cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + buft = ggml_backend_dev_buffer_type(cpu_dev); + } + + if (buft != buft_list->front().second) { + n_moved_tensors++; + if (!first_moved_tensor) { + first_moved_tensor = t_meta; + first_moved_from_buft = buft_list->front().second; + first_moved_to_buft = buft; + } + } + + ggml_context * ctx = ctx_for_buft(buft); + + // if duplicated, check if the original tensor was allocated in the same buffer type context and avoid creating a new one + if (flags & TENSOR_DUPLICATED) { + ggml_tensor * t = ggml_get_tensor(ctx, tn.str().c_str()); + if (t) { + return t; + } + } + return ml.create_tensor(ctx, tn, ne, flags); + }; + + layers.resize(n_layer); + + // TODO: move to a separate function + const auto tn = LLM_TN(arch); + switch (arch) { + case LLM_ARCH_LLAMA: + case LLM_ARCH_REFACT: + case LLM_ARCH_MINICPM: + case LLM_ARCH_GRANITE: + case LLM_ARCH_GRANITE_MOE: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0); + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + if (hparams.rope_scaling_type_train == LLAMA_ROPE_SCALING_TYPE_LONGROPE) { + layer.rope_long = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + else { + layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + + if (n_expert == 0) { + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + // optional MLP bias + layer.ffn_gate_b = create_tensor(tn(LLM_TENSOR_FFN_GATE, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + } else { + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, TENSOR_NOT_REQUIRED); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + } + } + } break; + case LLM_ARCH_DECI: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(i); + const int64_t n_embd_v_gqa = hparams.n_embd_v_gqa(i); + const int64_t n_embd_gqa = hparams.n_embd_v_gqa(i); + const int64_t n_ff = hparams.n_ff(i); + const int64_t n_head = hparams.n_head(i); + const int64_t n_head_kv = hparams.n_head_kv(i); + + if (n_head_kv == 0 && n_head > 0) { + // linear attention for DeciLMCausalModel + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + } + else if (n_head_kv > 0) { + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0); + } + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + if (hparams.rope_scaling_type_train == LLAMA_ROPE_SCALING_TYPE_LONGROPE) { + layer.rope_long = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + else { + layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + // optional MLP bias + layer.ffn_gate_b = create_tensor(tn(LLM_TENSOR_FFN_GATE, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + } + } break; + case LLM_ARCH_MINICPM3: + { + const int64_t n_embd_head_qk_rope = hparams.n_rot; + const int64_t n_embd_head_qk_nope = hparams.n_embd_head_k - hparams.n_rot; + + const int64_t q_lora_rank = hparams.n_lora_q; + const int64_t kv_lora_rank = hparams.n_lora_kv; + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0); + + layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, 0); + + layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0); + layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k}, 0); + + layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + (n_embd_head_qk_rope)}, 0); + layer.wkv_b = create_tensor(tn(LLM_TENSOR_ATTN_KV_B, "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_head * ( n_embd_head_v), n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + layer.rope_long = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG, "weight", i), { n_embd_head_qk_rope/2 }, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), { n_embd_head_qk_rope/2 }, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + } break; + case LLM_ARCH_GROK: + { + if (n_expert == 0) { + throw std::runtime_error("Grok model cannot have zero experts"); + } + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.attn_out_norm = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, TENSOR_NOT_REQUIRED); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + + layer.layer_out_norm = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0); + } + } break; + case LLM_ARCH_DBRX: + { + if (n_expert == 0) { + throw std::runtime_error("DBRX model cannot have zero experts"); + } + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.attn_out_norm = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + } + } break; + case LLM_ARCH_BAICHUAN: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + { + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_FALCON: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + { + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + if (!output) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); // needs to be on GPU + } + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.attn_norm_2 = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_STARCODER: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD, "weight"), {n_embd, n_ctx_train}, 0); + + // output + { + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + if (!output) { + // needs to be on GPU + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_BERT: + case LLM_ARCH_NOMIC_BERT: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + type_embd = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_token_types}, 0); + + if (arch == LLM_ARCH_BERT) { + pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD, "weight"), {n_embd, n_ctx_train}, 0); + + cls = create_tensor(tn(LLM_TENSOR_CLS, "weight"), {n_embd, n_embd}, TENSOR_NOT_REQUIRED); + cls_b = create_tensor(tn(LLM_TENSOR_CLS, "bias"), {n_embd}, TENSOR_NOT_REQUIRED); + + cls_out = create_tensor(tn(LLM_TENSOR_CLS_OUT, "weight"), {n_embd, 1}, TENSOR_NOT_REQUIRED); + cls_out_b = create_tensor(tn(LLM_TENSOR_CLS_OUT, "bias"), {1}, TENSOR_NOT_REQUIRED); + } + + tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0); + tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {n_embd}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + if (arch == LLM_ARCH_BERT) { + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + } else { + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + } + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.attn_out_norm = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0); + layer.attn_out_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + + if (arch == LLM_ARCH_BERT) { + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + } else { + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + } + + layer.layer_out_norm = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0); + layer.layer_out_norm_b = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "bias", i), {n_embd}, 0); + } + } break; + case LLM_ARCH_JINA_BERT_V2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); // word_embeddings + type_embd = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_token_types}, 0); // token_type_embeddings + + tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0); // LayerNorm + tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {n_embd}, 0); //LayerNorm bias + + cls = create_tensor(tn(LLM_TENSOR_CLS, "weight"), {n_embd, 1}, TENSOR_NOT_REQUIRED); + cls_b = create_tensor(tn(LLM_TENSOR_CLS, "bias"), {1}, TENSOR_NOT_REQUIRED); + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; // JinaBertLayer + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); //output_dens + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); //output_dens + + layer.attn_out_norm = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0); //output_norm + layer.attn_out_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "bias", i), {n_embd}, 0); + + layer.attn_norm_2 = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.layer_out_norm = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0); + layer.layer_out_norm_b = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "bias", i), {n_embd}, 0); + } + } break; + case LLM_ARCH_BLOOM: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0); + tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {n_embd}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_MPT: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD, "weight"), {n_embd, n_ctx_train}, TENSOR_NOT_REQUIRED); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, TENSOR_NOT_REQUIRED); + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + if (!output) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); // needs to be on GPU + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, TENSOR_NOT_REQUIRED); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + // AWQ ScaleActivation layer + layer.ffn_act = create_tensor(tn(LLM_TENSOR_FFN_ACT, "scales", i), {n_ff}, TENSOR_NOT_REQUIRED); + } + } break; + case LLM_ARCH_STABLELM: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + // optional bias tensors, present in Stable LM 2 1.6B + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + + // optional q and k layernorms, present in StableLM 2 12B + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head}, TENSOR_NOT_REQUIRED); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, TENSOR_NOT_REQUIRED); + + // optional FFN norm, not present in StableLM 2 12B which uses parallel residual + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_QWEN: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd*3}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd*3}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff/2}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff/2, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff/2}, 0); + } + } break; + case LLM_ARCH_QWEN2: + case LLM_ARCH_QWEN2VL: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_QWEN2MOE: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + + if (n_expert == 0) { + throw std::runtime_error("n_expert must be > 0 for QWEN2MOE"); + } + if (n_expert_used == 0) { + throw std::runtime_error("n_expert_used must be > 0 for QWEN2MOE"); + } + + // MoE branch + const int64_t n_ff_exp = hparams.n_ff_exp ? hparams.n_ff_exp : n_ff / n_expert_used; + + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + + // Shared expert branch + const int64_t n_ff_shexp = hparams.n_ff_shexp ? hparams.n_ff_shexp : n_ff; + + layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", i), {n_embd}, 0); + layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), { n_embd, n_ff_shexp}, 0); + layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {n_ff_shexp, n_embd}, 0); + layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), { n_embd, n_ff_shexp}, 0); + } + } break; + case LLM_ARCH_PHI2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + output_b = create_tensor(tn(LLM_TENSOR_OUTPUT, "bias"), {n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, TENSOR_NOT_REQUIRED); + + if (layer.wqkv == nullptr) { + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + } + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_PHI3: + { + const int64_t n_embd_head = n_embd / n_head; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), { n_embd }, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), { n_embd, n_vocab }, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), { n_embd }, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), { n_embd, n_embd + 2 * n_embd_gqa }, TENSOR_NOT_REQUIRED); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_embd, n_embd }, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), { n_embd }, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd }, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), { n_embd, 2 * n_ff }, 0); + + layer.rope_long = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG, "weight", i), { n_embd_head/2 }, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), { n_embd_head/2 }, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + } + } break; + case LLM_ARCH_PLAMO: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_GPT2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD, "weight"), {n_embd, n_ctx_train}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_CODESHELL: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_ORION: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_INTERNLM2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + // layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_GEMMA: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); // same as tok_embd, duplicated to allow offloading + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + } + } break; + case LLM_ARCH_GEMMA2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); // same as tok_embd, duplicated to allow offloading + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0); + layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_post_norm = create_tensor(tn(LLM_TENSOR_FFN_POST_NORM, "weight", i), {n_embd}, 0); + } + } break; + case LLM_ARCH_STARCODER2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, 0); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, 0); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + // optional bias tensors + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP , "bias", i), { n_ff}, 0); + } + } break; + case LLM_ARCH_MAMBA: + { + const int64_t d_conv = hparams.ssm_d_conv; + const int64_t d_inner = hparams.ssm_d_inner; + const int64_t d_state = hparams.ssm_d_state; + const int64_t dt_rank = hparams.ssm_dt_rank; + + // only an expansion factor of 2 is supported for now + if (2 * n_embd != d_inner) { + throw std::runtime_error("only an expansion factor of 2 is supported for now"); + } + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed, duplicated to allow offloading + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + // norm + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.ssm_in = create_tensor(tn(LLM_TENSOR_SSM_IN, "weight", i), {n_embd, 2*d_inner}, 0); + + layer.ssm_conv1d = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "weight", i), {d_conv, d_inner}, 0); + layer.ssm_conv1d_b = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "bias", i), {d_inner}, 0); + + layer.ssm_x = create_tensor(tn(LLM_TENSOR_SSM_X, "weight", i), {d_inner, dt_rank + 2*d_state}, 0); + + layer.ssm_dt = create_tensor(tn(LLM_TENSOR_SSM_DT, "weight", i), {dt_rank, d_inner}, 0); + layer.ssm_dt_b = create_tensor(tn(LLM_TENSOR_SSM_DT, "bias", i), {d_inner}, 0); + + // no "weight" suffix for these + layer.ssm_a = create_tensor(tn(LLM_TENSOR_SSM_A, i), {d_state, d_inner}, 0); + layer.ssm_d = create_tensor(tn(LLM_TENSOR_SSM_D, i), {d_inner}, 0); + + // out_proj + layer.ssm_out = create_tensor(tn(LLM_TENSOR_SSM_OUT, "weight", i), {d_inner, n_embd}, 0); + } + } break; + case LLM_ARCH_XVERSE: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_COMMAND_R: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + // init output from the input tok embed + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + if (n_layer >= 64){ + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head}, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, 0); + } + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_COHERE2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), { n_embd }, 0); + // init output from the input tok embed + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, + TENSOR_DUPLICATED); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), { n_embd }, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), { n_embd, n_embd }, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), { n_embd, n_embd_gqa }, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), { n_embd, n_embd_gqa }, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_embd, n_embd }, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), { n_embd, n_ff }, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd }, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), { n_embd, n_ff }, 0); + } + } + break; + case LLM_ARCH_OLMO: // adapted from LLM_ARCH_LLAMA with norm params removed + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_OLMO2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, 0); + layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_post_norm = create_tensor(tn(LLM_TENSOR_FFN_POST_NORM, "weight", i), {n_embd}, 0); + } + } break; + case LLM_ARCH_OLMOE: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + + if (n_expert == 0) { + throw std::runtime_error("n_expert must be > 0"); + } + if (n_expert_used == 0) { + throw std::runtime_error("n_expert_used must be > 0"); + } + + // MoE branch + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + } + } break; + case LLM_ARCH_OPENELM: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + // init output from the input tok embed + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + + for (int i = 0; i < n_layer; ++i) { + const int64_t n_head = hparams.n_head(i); + const int64_t n_head_qkv = 2*hparams.n_head_kv(i) + n_head; + const int64_t n_ff = hparams.n_ff(i); + + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_head_qkv*n_embd_head_k}, 0); + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k}, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_head*n_embd_head_k, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_GPTNEOX: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_ARCTIC: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_embd}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_embd, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + layer.ffn_norm_exps = create_tensor(tn(LLM_TENSOR_FFN_NORM_EXPS, "weight", i), {n_embd}, 0); + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, false); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert}, 0); + } + } break; + case LLM_ARCH_DEEPSEEK: + { + + const int64_t n_ff_exp = hparams.n_ff_exp; + const int64_t n_expert_shared = hparams.n_expert_shared; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + if (i < (int) hparams.n_layer_dense_lead) { + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } else { + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + + if (n_expert == 0) { + throw std::runtime_error("n_expert must be > 0"); + } + if (n_expert_used == 0) { + throw std::runtime_error("n_expert_used must be > 0"); + } + + // MoE branch + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + + // Shared expert branch + layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), { n_ff_exp * n_expert_shared, n_embd}, 0); + layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + } + } + } break; + case LLM_ARCH_DEEPSEEK2: + { + const bool is_lite = (hparams.n_layer == 27); + + const int64_t n_embd_head_qk_rope = hparams.n_rot; + const int64_t n_embd_head_qk_nope = hparams.n_embd_head_k - hparams.n_rot; + + const int64_t q_lora_rank = hparams.n_lora_q; + const int64_t kv_lora_rank = hparams.n_lora_kv; + + const int64_t n_ff_exp = hparams.n_ff_exp; + const int64_t n_expert_shared = hparams.n_expert_shared; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + if (!is_lite) { + layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0); + } + + layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, 0); + + if (!is_lite) { + layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0); + layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k}, 0); + } else { + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0); + } + + layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + (n_embd_head_qk_rope)}, 0); + layer.wkv_b = create_tensor(tn(LLM_TENSOR_ATTN_KV_B, "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_head * ( n_embd_head_v), n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + if (i < (int) hparams.n_layer_dense_lead) { + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } else { + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0); + layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, TENSOR_NOT_REQUIRED); + + if (n_expert == 0) { + throw std::runtime_error("n_expert must be > 0"); + } + if (n_expert_used == 0) { + throw std::runtime_error("n_expert_used must be > 0"); + } + + // MoE branch + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert}, 0); + + // Shared expert branch + layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), { n_ff_exp * n_expert_shared, n_embd}, 0); + layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0); + } + } + } break; + case LLM_ARCH_BITNET: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_sub_norm = create_tensor(tn(LLM_TENSOR_ATTN_SUB_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wq_scale = create_tensor(tn(LLM_TENSOR_ATTN_Q, "scale", i), {1}, TENSOR_NOT_REQUIRED); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wk_scale = create_tensor(tn(LLM_TENSOR_ATTN_K, "scale", i), {1}, TENSOR_NOT_REQUIRED); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv_scale = create_tensor(tn(LLM_TENSOR_ATTN_V, "scale", i), {1}, TENSOR_NOT_REQUIRED); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.wo_scale = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "scale", i), {1}, TENSOR_NOT_REQUIRED); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_sub_norm = create_tensor(tn(LLM_TENSOR_FFN_SUB_NORM, "weight", i), {n_ff}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_gate_scale = create_tensor(tn(LLM_TENSOR_FFN_GATE, "scale", i), {1}, TENSOR_NOT_REQUIRED); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_scale = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "scale", i), {1}, TENSOR_NOT_REQUIRED); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_scale = create_tensor(tn(LLM_TENSOR_FFN_UP, "scale", i), {1}, TENSOR_NOT_REQUIRED); + } + } break; + case LLM_ARCH_T5: + { + const auto n_rel_attn_bkts = hparams.n_rel_attn_bkts; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm = create_tensor(tn(LLM_TENSOR_DEC_OUTPUT_NORM, "weight"), {n_embd}, 0); + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_rel_b_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, TENSOR_NOT_REQUIRED); + + layer.wq_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wk_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0); + + layer.ffn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_GATE, "weight", i), {n_embd, n_ff}, TENSOR_NOT_REQUIRED); + layer.ffn_down_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_DEC_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_rel_b = create_tensor(tn(LLM_TENSOR_DEC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, TENSOR_NOT_REQUIRED); + + layer.wq = create_tensor(tn(LLM_TENSOR_DEC_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_DEC_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_DEC_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_DEC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0); + + layer.attn_norm_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_NORM, "weight", i), {n_embd}, 0); + // this tensor seems to be unused in HF transformers implementation + layer.attn_rel_b_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, TENSOR_NOT_REQUIRED); + + layer.wq_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wk_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_DEC_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_DEC_FFN_GATE, "weight", i), {n_embd, n_ff}, TENSOR_NOT_REQUIRED); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_DEC_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_DEC_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_T5ENCODER: + { + const auto n_rel_attn_bkts = hparams.n_rel_attn_bkts; + + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_rel_b_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, TENSOR_NOT_REQUIRED); + + layer.wq_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wk_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0); + + layer.ffn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_gate_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_GATE, "weight", i), {n_embd, n_ff}, TENSOR_NOT_REQUIRED); + layer.ffn_down_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_JAIS: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_gate_b = create_tensor(tn(LLM_TENSOR_FFN_GATE, "bias", i), {n_ff}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, 0); + } + } break; + case LLM_ARCH_CHATGLM: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0); + layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i), {n_embd + 2*n_embd_gqa}, 0); + + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff * 2}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0); + } + } break; + case LLM_ARCH_NEMOTRON: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + // optional bias tensors + layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED); + layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0); + + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + + // optional MLP bias + layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED); + layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i), {n_ff}, TENSOR_NOT_REQUIRED); + } + } break; + case LLM_ARCH_EXAONE: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_k_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_v_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0)); + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_RWKV6: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // Block 0, LN0 + tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0); + tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {n_embd}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + const int time_mix_extra_dim = hparams.time_mix_extra_dim; + const int time_decay_extra_dim = hparams.time_decay_extra_dim; + const int head_size = hparams.wkv_head_size; + const int attn_hidden_size = n_embd; + const int ffn_size = hparams.n_ff_arr[0]; + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0); + + layer.attn_norm_2 = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, 0); + layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias", i), {n_embd}, 0); + + layer.time_mix_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W1, "weight", i), {n_embd, time_mix_extra_dim * 5}, 0); + layer.time_mix_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W2, "weight", i), {time_mix_extra_dim, n_embd, 5}, 0); + + layer.time_mix_lerp_x = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_X, "weight", i), {n_embd, 1, 1}, 0); + layer.time_mix_lerp_w = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_W, "weight", i), {n_embd, 1, 1}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_lerp_k = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_K, "weight", i), {n_embd, 1, 1}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_lerp_v = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_V, "weight", i), {n_embd, 1, 1}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_lerp_r = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_R, "weight", i), {n_embd, 1, 1}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_lerp_g = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_G, "weight", i), {n_embd, 1, 1}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_lerp_fused = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_FUSED, "weight", i), {n_embd, 1, 1, 5}, llama_model_loader::TENSOR_NOT_REQUIRED); + GGML_ASSERT(!(layer.time_mix_lerp_fused == NULL && layer.time_mix_lerp_w == NULL)); + + layer.time_mix_first = create_tensor(tn(LLM_TENSOR_TIME_MIX_FIRST, "weight", i), {head_size, n_embd / head_size}, 0); + layer.time_mix_decay = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY, "weight", i), {n_embd}, 0); + layer.time_mix_decay_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W1, "weight", i), {n_embd, time_decay_extra_dim}, 0); + layer.time_mix_decay_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W2, "weight", i), {time_decay_extra_dim, attn_hidden_size}, 0); + layer.time_mix_key = create_tensor(tn(LLM_TENSOR_TIME_MIX_KEY, "weight", i), {attn_hidden_size, n_embd}, 0); + layer.time_mix_value = create_tensor(tn(LLM_TENSOR_TIME_MIX_VALUE, "weight", i), {attn_hidden_size, n_embd}, 0); + layer.time_mix_receptance = create_tensor(tn(LLM_TENSOR_TIME_MIX_RECEPTANCE, "weight", i), {attn_hidden_size, n_embd}, 0); + layer.time_mix_gate = create_tensor(tn(LLM_TENSOR_TIME_MIX_GATE, "weight", i), {attn_hidden_size, n_embd}, 0); + + layer.time_mix_ln = create_tensor(tn(LLM_TENSOR_TIME_MIX_LN, "weight", i), {n_embd}, 0); + layer.time_mix_ln_b = create_tensor(tn(LLM_TENSOR_TIME_MIX_LN, "bias", i), {n_embd}, 0); + layer.time_mix_output = create_tensor(tn(LLM_TENSOR_TIME_MIX_OUTPUT, "weight", i), {n_embd, attn_hidden_size}, 0); + + layer.channel_mix_lerp_k = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_LERP_K, "weight", i), {n_embd, 1, 1}, 0); + layer.channel_mix_lerp_r = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_LERP_R, "weight", i), {n_embd, 1, 1}, 0); + + layer.channel_mix_key = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_KEY, "weight", i), {n_embd, ffn_size}, 0); + layer.channel_mix_value = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_VALUE, "weight", i), {ffn_size, n_embd}, 0); + layer.channel_mix_receptance = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_RECEPTANCE, "weight", i), {n_embd, n_embd}, 0); + } + + } break; + case LLM_ARCH_RWKV6QWEN2: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0); + + const int time_mix_extra_dim = hparams.time_mix_extra_dim; + const int time_decay_extra_dim = hparams.time_decay_extra_dim; + const int head_size = hparams.wkv_head_size; + const int attn_hidden_size = n_embd; + const int n_head_kv = hparams.n_head_kv(); + int attn_key_value_size; + if (n_head_kv == 0 || attn_hidden_size / head_size == n_head_kv) { + attn_key_value_size = attn_hidden_size; + } else { + attn_key_value_size = n_head_kv * head_size; + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + + layer.time_mix_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W1, "weight", i), {n_embd, time_mix_extra_dim * 5}, 0); + layer.time_mix_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W2, "weight", i), {time_mix_extra_dim, n_embd, 5}, 0); + + layer.time_mix_lerp_x = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_X, "weight", i), {n_embd, 1, 1}, 0); + layer.time_mix_lerp_fused = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_FUSED, "weight", i), {n_embd, 1, 1, 5}, 0); + + layer.time_mix_first = create_tensor(tn(LLM_TENSOR_TIME_MIX_FIRST, "weight", i), {head_size, n_embd / head_size}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_decay = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY, "weight", i), {n_embd}, 0); + layer.time_mix_decay_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W1, "weight", i), {n_embd, time_decay_extra_dim}, 0); + layer.time_mix_decay_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W2, "weight", i), {time_decay_extra_dim, attn_hidden_size}, 0); + layer.time_mix_key = create_tensor(tn(LLM_TENSOR_TIME_MIX_KEY, "weight", i), {n_embd, attn_key_value_size}, 0); + layer.time_mix_value = create_tensor(tn(LLM_TENSOR_TIME_MIX_VALUE, "weight", i), {n_embd, attn_key_value_size}, 0); + layer.time_mix_receptance = create_tensor(tn(LLM_TENSOR_TIME_MIX_RECEPTANCE, "weight", i), {attn_hidden_size, n_embd}, 0); + layer.time_mix_gate = create_tensor(tn(LLM_TENSOR_TIME_MIX_GATE, "weight", i), {attn_hidden_size, n_embd}, 0); + // optional bias tensors + layer.time_mix_key_b = create_tensor(tn(LLM_TENSOR_TIME_MIX_KEY, "bias", i), {attn_key_value_size}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_value_b = create_tensor(tn(LLM_TENSOR_TIME_MIX_VALUE, "bias", i), {attn_key_value_size}, llama_model_loader::TENSOR_NOT_REQUIRED); + layer.time_mix_receptance_b = create_tensor(tn(LLM_TENSOR_TIME_MIX_RECEPTANCE, "bias", i), {attn_hidden_size}, llama_model_loader::TENSOR_NOT_REQUIRED); + + layer.time_mix_output = create_tensor(tn(LLM_TENSOR_TIME_MIX_OUTPUT, "weight", i), {n_embd, attn_hidden_size}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_CHAMELEON: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0); + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED); + // if output is NULL, init from the input tok embed + if (output == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); + } + + for (int i = 0; i < n_layer; ++i) { + auto & layer = layers[i]; + + layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0); + layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head}, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, 0); + layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i), {n_embd_head_k, n_head}, TENSOR_NOT_REQUIRED); + layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias", i), {n_embd_head_k, n_head_kv}, TENSOR_NOT_REQUIRED); + + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0); + layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); + + layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0); + + layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0); + layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0); + layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0); + } + } break; + case LLM_ARCH_WAVTOKENIZER_DEC: + { + tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {hparams.n_embd_features, n_vocab}, 0); + + conv1d = create_tensor(tn(LLM_TENSOR_CONV1D, "weight"), {7, hparams.n_embd_features, hparams.posnet.n_embd}, 0); + conv1d_b = create_tensor(tn(LLM_TENSOR_CONV1D, "bias"), {1, hparams.posnet.n_embd}, 0); + + // posnet + { + const int64_t n_embd = hparams.posnet.n_embd; + + for (uint32_t i = 0; i < hparams.posnet.n_layer; ++i) { + auto & layer = layers[i].posnet; + + // posnet: + // + // - resnet + // - resnet + // - attn + // - resnet + // - resnet + // - norm + // + switch (i) { + case 0: + case 1: + case 3: + case 4: + { + layer.norm1 = create_tensor(tn(LLM_TENSOR_POS_NET_NORM1, "weight", i), {1, n_embd}, 0); + layer.norm1_b = create_tensor(tn(LLM_TENSOR_POS_NET_NORM1, "bias", i), {1, n_embd}, 0); + + layer.conv1 = create_tensor(tn(LLM_TENSOR_POS_NET_CONV1, "weight", i), {3, n_embd, n_embd}, 0); + layer.conv1_b = create_tensor(tn(LLM_TENSOR_POS_NET_CONV1, "bias", i), {1, n_embd}, 0); + + layer.norm2 = create_tensor(tn(LLM_TENSOR_POS_NET_NORM2, "weight", i), {1, n_embd}, 0); + layer.norm2_b = create_tensor(tn(LLM_TENSOR_POS_NET_NORM2, "bias", i), {1, n_embd}, 0); + + layer.conv2 = create_tensor(tn(LLM_TENSOR_POS_NET_CONV2, "weight", i), {3, n_embd, n_embd}, 0); + layer.conv2_b = create_tensor(tn(LLM_TENSOR_POS_NET_CONV2, "bias", i), {1, n_embd}, 0); + } break; + case 2: + { + layer.attn_norm = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_NORM, "weight", i), {1, n_embd}, 0); + layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_NORM, "bias", i), {1, n_embd}, 0); + + layer.attn_q = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_Q, "weight", i), {1, n_embd, n_embd}, 0); + layer.attn_q_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_Q, "bias", i), {1, n_embd}, 0); + + layer.attn_k = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_K, "weight", i), {1, n_embd, n_embd}, 0); + layer.attn_k_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_K, "bias", i), {1, n_embd}, 0); + + layer.attn_v = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_V, "weight", i), {1, n_embd, n_embd}, 0); + layer.attn_v_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_V, "bias", i), {1, n_embd}, 0); + + layer.attn_o = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_OUT, "weight", i), {1, n_embd, n_embd}, 0); + layer.attn_o_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_OUT, "bias", i), {1, n_embd}, 0); + } break; + case 5: + { + layer.norm = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_NORM, "weight", i), {1, n_embd}, 0); + layer.norm_b = create_tensor(tn(LLM_TENSOR_POS_NET_ATTN_NORM, "bias", i), {1, n_embd}, 0); + } break; + default: GGML_ABORT("unknown posnet layer"); + }; + } + } + + GGML_ASSERT(hparams.posnet.n_embd == hparams.convnext.n_embd); + + tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {hparams.posnet.n_embd}, 0); + tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {hparams.posnet.n_embd}, 0); + + // convnext + { + const int64_t n_embd = hparams.convnext.n_embd; + + for (uint32_t i = 0; i < hparams.convnext.n_layer; ++i) { + auto & layer = layers[i].convnext; + + layer.dw = create_tensor(tn(LLM_TENSOR_CONVNEXT_DW, "weight", i), {7, 1, n_embd}, 0); + layer.dw_b = create_tensor(tn(LLM_TENSOR_CONVNEXT_DW, "bias", i), {1, n_embd}, 0); + + layer.norm = create_tensor(tn(LLM_TENSOR_CONVNEXT_NORM, "weight", i), {n_embd}, 0); + layer.norm_b = create_tensor(tn(LLM_TENSOR_CONVNEXT_NORM, "bias", i), {n_embd}, 0); + + layer.pw1 = create_tensor(tn(LLM_TENSOR_CONVNEXT_PW1, "weight", i), {n_embd, n_ff}, 0); + layer.pw1_b = create_tensor(tn(LLM_TENSOR_CONVNEXT_PW1, "bias", i), {n_ff}, 0); + + layer.pw2 = create_tensor(tn(LLM_TENSOR_CONVNEXT_PW2, "weight", i), {n_ff, n_embd}, 0); + layer.pw2_b = create_tensor(tn(LLM_TENSOR_CONVNEXT_PW2, "bias", i), {n_embd}, 0); + + layer.gamma = create_tensor(tn(LLM_TENSOR_CONVNEXT_GAMMA, "weight", i), {n_embd}, 0); + } + + // output + output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); + output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0); + } + + output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {hparams.convnext.n_embd, n_embd}, 0); + output_b = create_tensor(tn(LLM_TENSOR_OUTPUT, "bias"), {n_embd}, 0); + } break; + default: + throw std::runtime_error("unknown architecture"); + } + + if (n_moved_tensors > 0) { + LLAMA_LOG_DEBUG("%s: tensor '%s' (%s) (and %d others) cannot be used with preferred buffer type %s, using %s instead\n", + __func__, first_moved_tensor->name, ggml_type_name(first_moved_tensor->type), n_moved_tensors - 1, + ggml_backend_buft_name(first_moved_from_buft), ggml_backend_buft_name(first_moved_to_buft)); + } + } + + ml.done_getting_tensors(); + + ml.init_mappings(true, use_mlock ? &pimpl->mlock_mmaps : nullptr); + pimpl->mappings.reserve(ml.mappings.size()); + + // create the backend buffers + std::vector> ctx_bufs; + ctx_bufs.reserve(ctx_map.size()); + + // Ensure we have enough capacity for the maximum backend buffer we will potentially create + const size_t n_max_backend_buffer = ctx_map.size() * ml.files.size(); + pimpl->bufs.reserve(n_max_backend_buffer); + + for (auto & it : ctx_map) { + ggml_backend_buffer_type_t buft = it.first; + ggml_context * ctx = it.second; + + // skip contexts without tensors + if (ggml_get_first_tensor(ctx) == nullptr) { + continue; + } + + llama_buf_map buf_map; + buf_map.reserve(n_max_backend_buffer); + + // check if it is possible to use buffer_from_host_ptr with this buffer type + ggml_backend_dev_t dev = ggml_backend_buft_get_device(buft); + if (!dev) { + // FIXME: workaround for CPU backend buft having a NULL device + dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU); + } + ggml_backend_dev_props props; + ggml_backend_dev_get_props(dev, &props); + bool buffer_from_host_ptr_supported = props.caps.buffer_from_host_ptr; + bool is_default_buft = buft == ggml_backend_dev_buffer_type(dev); + + if (ml.use_mmap && use_mmap_buffer && buffer_from_host_ptr_supported && is_default_buft) { + for (uint32_t idx = 0; idx < ml.files.size(); idx++) { + // only the mmap region containing the tensors in the model is mapped to the backend buffer + // this is important for metal with apple silicon: if the entire model could be mapped to a metal buffer, then we could just use metal for all layers + // this allows using partial offloading when the model size exceeds the metal buffer size, but not the RAM size + void * addr = nullptr; + size_t first, last; // NOLINT + ml.get_mapping_range(&first, &last, &addr, idx, ctx); + if (first >= last) { + continue; + } + const size_t max_size = ggml_get_max_tensor_size(ctx); + ggml_backend_buffer_t buf = ggml_backend_dev_buffer_from_host_ptr(dev, (char *) addr + first, last - first, max_size); + if (buf == nullptr) { + throw std::runtime_error(format("unable to allocate %s buffer", ggml_backend_buft_name(buft))); + } + pimpl->bufs.emplace_back(buf); + buf_map.emplace(idx, buf); + } + } + else { + ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft); + if (buf == nullptr) { + throw std::runtime_error(format("unable to allocate %s buffer", ggml_backend_buft_name(buft))); + } + pimpl->bufs.emplace_back(buf); + if (use_mlock && ggml_backend_buffer_is_host(buf)) { + pimpl->mlock_bufs.emplace_back(new llama_mlock); + auto & mlock_buf = pimpl->mlock_bufs.back(); + mlock_buf->init (ggml_backend_buffer_get_base(buf)); + mlock_buf->grow_to(ggml_backend_buffer_get_size(buf)); + } + for (uint32_t idx = 0; idx < ml.files.size(); idx++) { + buf_map.emplace(idx, buf); + } + } + + if (pimpl->bufs.empty()) { + throw std::runtime_error("failed to allocate buffer"); + } + + for (auto & buf : buf_map) { + // indicate that this buffer contains weights + // this is used by ggml_backend_sched to improve op scheduling: ops that use a weight are preferably scheduled to the backend that contains the weight + ggml_backend_buffer_set_usage(buf.second, GGML_BACKEND_BUFFER_USAGE_WEIGHTS); + } + + ctx_bufs.emplace_back(ctx, buf_map); + } + + if (llama_supports_gpu_offload()) { + const int n_gpu = std::min(n_gpu_layers, int(hparams.n_layer)); + + LLAMA_LOG_INFO("%s: offloading %d repeating layers to GPU\n", __func__, n_gpu); + if (n_gpu_layers > (int) hparams.n_layer) { + LLAMA_LOG_INFO("%s: offloading output layer to GPU\n", __func__); + } + + const int max_backend_supported_layers = hparams.n_layer + 1; + const int max_offloadable_layers = hparams.n_layer + 1; + + LLAMA_LOG_INFO("%s: offloaded %d/%d layers to GPU\n", __func__, std::min(n_gpu_layers, max_offloadable_layers), max_backend_supported_layers); + } + + // print memory requirements per buffer type + for (auto & buf : pimpl->bufs) { + LLAMA_LOG_INFO("%s: %12s model buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get()) / 1024.0 / 1024.0); + } + + // populate tensors_by_name + for (auto & ctx : pimpl->ctxs) { + for (auto * cur = ggml_get_first_tensor(ctx.get()); cur != NULL; cur = ggml_get_next_tensor(ctx.get(), cur)) { + tensors_by_name.emplace_back(ggml_get_name(cur), cur); + } + } + + // load tensor data + for (auto & it : ctx_bufs) { + ggml_context * ctx = it.first; + auto & bufs = it.second; + if (!ml.load_all_data(ctx, bufs, use_mlock ? &pimpl->mlock_mmaps : NULL, params.progress_callback, params.progress_callback_user_data)) { + return false; + } + } + + if (use_mmap_buffer) { + for (auto & mapping : ml.mappings) { + pimpl->mappings.emplace_back(std::move(mapping)); + } + } + + return true; +} + +std::string llama_model::arch_name() const { + return llm_arch_name(arch); +} + +std::string llama_model::type_name() const { + return llm_type_name(type); +} + +std::string llama_model::desc() const { + return pimpl->desc_str; +} + +size_t llama_model::size() const { + return pimpl->n_bytes; +} + +size_t llama_model::max_nodes() const { + return std::max(8192, tensors_by_name.size()*5); +} + +size_t llama_model::n_devices() const { + return devices.size(); +} + +uint64_t llama_model::n_elements() const { + return pimpl->n_elements; +} + +void llama_model::print_info() const { + const char * rope_scaling_type = LLAMA_ROPE_SCALING_TYPES.at(hparams.rope_scaling_type_train); + + auto print_f = [](const std::function & f, uint32_t n) { + bool is_var = false; + + std::vector v; + for (uint32_t i = 0; i < n; ++i) { + v.push_back(f(i)); + if (v[i] != v[0]) { + is_var = true; + } + } + + std::stringstream ss; + + if (is_var) { + ss << "["; + for (uint32_t i = 0; i < n; ++i) { + ss << v[i]; + if (i < n - 1) { + ss << ", "; + } + } + ss << "]"; + } else { + ss << v[0]; + } + + return ss.str(); + }; + + // hparams + LLAMA_LOG_INFO("%s: arch = %s\n", __func__, arch_name().c_str()); + LLAMA_LOG_INFO("%s: vocab_only = %d\n", __func__, hparams.vocab_only); + + if (!hparams.vocab_only) { + LLAMA_LOG_INFO("%s: n_ctx_train = %u\n", __func__, hparams.n_ctx_train); + LLAMA_LOG_INFO("%s: n_embd = %u\n", __func__, hparams.n_embd); + LLAMA_LOG_INFO("%s: n_layer = %u\n", __func__, hparams.n_layer); + LLAMA_LOG_INFO("%s: n_head = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_head(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: n_head_kv = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_head_kv(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: n_rot = %u\n", __func__, hparams.n_rot); + LLAMA_LOG_INFO("%s: n_swa = %u\n", __func__, hparams.n_swa); + LLAMA_LOG_INFO("%s: n_embd_head_k = %u\n", __func__, hparams.n_embd_head_k); + LLAMA_LOG_INFO("%s: n_embd_head_v = %u\n", __func__, hparams.n_embd_head_v); + LLAMA_LOG_INFO("%s: n_gqa = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_gqa(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: n_embd_k_gqa = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_embd_k_gqa(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: n_embd_v_gqa = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_embd_v_gqa(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: f_norm_eps = %.1e\n", __func__, hparams.f_norm_eps); + LLAMA_LOG_INFO("%s: f_norm_rms_eps = %.1e\n", __func__, hparams.f_norm_rms_eps); + LLAMA_LOG_INFO("%s: f_clamp_kqv = %.1e\n", __func__, hparams.f_clamp_kqv); + LLAMA_LOG_INFO("%s: f_max_alibi_bias = %.1e\n", __func__, hparams.f_max_alibi_bias); + LLAMA_LOG_INFO("%s: f_logit_scale = %.1e\n", __func__, hparams.f_logit_scale); + LLAMA_LOG_INFO("%s: n_ff = %s\n", __func__, print_f([&](uint32_t il) { return hparams.n_ff(il); }, hparams.n_layer).c_str()); + LLAMA_LOG_INFO("%s: n_expert = %u\n", __func__, hparams.n_expert); + LLAMA_LOG_INFO("%s: n_expert_used = %u\n", __func__, hparams.n_expert_used); + LLAMA_LOG_INFO("%s: causal attn = %d\n", __func__, hparams.causal_attn); + LLAMA_LOG_INFO("%s: pooling type = %d\n", __func__, hparams.pooling_type); + LLAMA_LOG_INFO("%s: rope type = %d\n", __func__, hparams.rope_type); + LLAMA_LOG_INFO("%s: rope scaling = %s\n", __func__, rope_scaling_type); + LLAMA_LOG_INFO("%s: freq_base_train = %.1f\n", __func__, hparams.rope_freq_base_train); + LLAMA_LOG_INFO("%s: freq_scale_train = %g\n", __func__, hparams.rope_freq_scale_train); + LLAMA_LOG_INFO("%s: n_ctx_orig_yarn = %u\n", __func__, hparams.n_ctx_orig_yarn); + LLAMA_LOG_INFO("%s: rope_finetuned = %s\n", __func__, hparams.rope_finetuned ? "yes" : "unknown"); + LLAMA_LOG_INFO("%s: ssm_d_conv = %u\n", __func__, hparams.ssm_d_conv); + LLAMA_LOG_INFO("%s: ssm_d_inner = %u\n", __func__, hparams.ssm_d_inner); + LLAMA_LOG_INFO("%s: ssm_d_state = %u\n", __func__, hparams.ssm_d_state); + LLAMA_LOG_INFO("%s: ssm_dt_rank = %u\n", __func__, hparams.ssm_dt_rank); + LLAMA_LOG_INFO("%s: ssm_dt_b_c_rms = %d\n", __func__, hparams.ssm_dt_b_c_rms); + } + + LLAMA_LOG_INFO("%s: model type = %s\n", __func__, type_name().c_str()); + if (pimpl->n_elements >= 1e12) { + LLAMA_LOG_INFO("%s: model params = %.2f T\n", __func__, pimpl->n_elements*1e-12); + } else if (pimpl->n_elements >= 1e9) { + LLAMA_LOG_INFO("%s: model params = %.2f B\n", __func__, pimpl->n_elements*1e-9); + } else if (pimpl->n_elements >= 1e6) { + LLAMA_LOG_INFO("%s: model params = %.2f M\n", __func__, pimpl->n_elements*1e-6); + } else { + LLAMA_LOG_INFO("%s: model params = %.2f K\n", __func__, pimpl->n_elements*1e-3); + } + + // general kv + LLAMA_LOG_INFO("%s: general.name = %s\n", __func__, name.c_str()); + + if (arch == LLM_ARCH_DEEPSEEK) { + LLAMA_LOG_INFO("%s: n_layer_dense_lead = %d\n", __func__, hparams.n_layer_dense_lead); + LLAMA_LOG_INFO("%s: n_ff_exp = %d\n", __func__, hparams.n_ff_exp); + LLAMA_LOG_INFO("%s: n_expert_shared = %d\n", __func__, hparams.n_expert_shared); + LLAMA_LOG_INFO("%s: expert_weights_scale = %.1f\n", __func__, hparams.expert_weights_scale); + } + + if (arch == LLM_ARCH_DEEPSEEK2) { + LLAMA_LOG_INFO("%s: n_layer_dense_lead = %d\n", __func__, hparams.n_layer_dense_lead); + LLAMA_LOG_INFO("%s: n_lora_q = %d\n", __func__, hparams.n_lora_q); + LLAMA_LOG_INFO("%s: n_lora_kv = %d\n", __func__, hparams.n_lora_kv); + LLAMA_LOG_INFO("%s: n_ff_exp = %d\n", __func__, hparams.n_ff_exp); + LLAMA_LOG_INFO("%s: n_expert_shared = %d\n", __func__, hparams.n_expert_shared); + LLAMA_LOG_INFO("%s: expert_weights_scale = %.1f\n", __func__, hparams.expert_weights_scale); + LLAMA_LOG_INFO("%s: expert_weights_norm = %d\n", __func__, hparams.expert_weights_norm); + LLAMA_LOG_INFO("%s: expert_gating_func = %s\n", __func__, llama_expert_gating_func_name((enum llama_expert_gating_func_type) hparams.expert_gating_func)); + LLAMA_LOG_INFO("%s: rope_yarn_log_mul = %.4f\n", __func__, hparams.rope_yarn_log_mul); + } + + if (arch == LLM_ARCH_QWEN2MOE) { + LLAMA_LOG_INFO("%s: n_ff_exp = %d\n", __func__, hparams.n_ff_exp); + LLAMA_LOG_INFO("%s: n_ff_shexp = %d\n", __func__, hparams.n_ff_shexp); + } + + if (arch == LLM_ARCH_MINICPM || arch == LLM_ARCH_GRANITE || arch == LLM_ARCH_GRANITE_MOE) { + LLAMA_LOG_INFO("%s: f_embedding_scale = %f\n", __func__, hparams.f_embedding_scale); + LLAMA_LOG_INFO("%s: f_residual_scale = %f\n", __func__, hparams.f_residual_scale); + LLAMA_LOG_INFO("%s: f_attention_scale = %f\n", __func__, hparams.f_attention_scale); + } + + vocab.print_info(); +} + +ggml_backend_dev_t llama_model::dev_layer(int il) const { + return pimpl->dev_layer.at(il).dev; +} + +ggml_backend_dev_t llama_model::dev_output() const { + return pimpl->dev_output.dev; +} + +template +static bool buft_supported(ggml_backend_buffer_type_t buft, ggml_backend_dev_t dev, F & fn) { + ggml_init_params params = { + /*.mem_size =*/ ggml_tensor_overhead()*8, + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + + ggml_context_ptr ctx { ggml_init(params) }; + if (!ctx) { + throw std::runtime_error(format("failed to create ggml context")); + } + + ggml_backend_buffer_ptr buf { ggml_backend_buft_alloc_buffer(buft, 0) }; + ggml_tensor * op_tensor = fn(ctx.get()); + for (int i = 0; i < GGML_MAX_SRC; i++) { + if (op_tensor->src[i] != nullptr) { + assert(op_tensor->src[i]->buffer == nullptr); + op_tensor->src[i]->buffer = buf.get(); + } + } + + bool op_supported = ggml_backend_dev_supports_op(dev, op_tensor); + + return op_supported; +} + +template +static ggml_backend_buffer_type_t select_buft(const buft_list_t & buft_list, const F & fn) { + for (const auto & cur : buft_list) { + ggml_backend_dev_t cur_dev = cur.first; + ggml_backend_buffer_type_t cur_buft = cur.second; + if (buft_supported(cur_buft, cur_dev, fn)) { + return cur_buft; + } + } + + throw std::runtime_error(format("no suitable buffer type found")); +} + +ggml_backend_buffer_type_t llama_model::select_buft(int il) const { + return ::select_buft( + *pimpl->dev_layer.at(il).buft_list, + [&](ggml_context * ctx) { + ggml_tensor * cur = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, hparams.n_embd); + ggml_tensor * layer_dir = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, hparams.n_embd); + return ggml_add(ctx, cur, layer_dir); + }); +} + +const struct ggml_tensor * llama_model::get_tensor(const char * name) const { + auto it = std::find_if(tensors_by_name.begin(), tensors_by_name.end(), + [name](const std::pair & it) { + return it.first == name; + }); + if (it == tensors_by_name.end()) { + return nullptr; + } + + return it->second; +} + +// +// interface implementation +// + +struct llama_model_params llama_model_default_params() { + struct llama_model_params result = { + /*.devices =*/ nullptr, + /*.n_gpu_layers =*/ 0, + /*.split_mode =*/ LLAMA_SPLIT_MODE_LAYER, + /*.main_gpu =*/ 0, + /*.tensor_split =*/ nullptr, + /*.rpc_servers =*/ nullptr, + /*.progress_callback =*/ nullptr, + /*.progress_callback_user_data =*/ nullptr, + /*.kv_overrides =*/ nullptr, + /*.vocab_only =*/ false, + /*.use_mmap =*/ true, + /*.use_mlock =*/ false, + /*.check_tensors =*/ false, + }; + +#ifdef GGML_USE_METAL + // note: we usually have plenty of VRAM, so by default offload all layers to the GPU + result.n_gpu_layers = 999; +#endif + + return result; +} + +const struct llama_vocab * llama_model_get_vocab(const struct llama_model * model) { + return &model->vocab; +} + +void llama_free_model(struct llama_model * model) { + llama_model_free(model); +} + +void llama_model_free(struct llama_model * model) { + delete model; +} + +int32_t llama_model_n_ctx_train(const struct llama_model * model) { + return model->hparams.n_ctx_train; +} + +int32_t llama_model_n_embd(const struct llama_model * model) { + return model->hparams.n_embd; +} + +int32_t llama_model_n_layer(const struct llama_model * model) { + return model->hparams.n_layer; +} + +int32_t llama_model_n_head(const struct llama_model * model) { + return model->hparams.n_head(); +} + +// deprecated +int32_t llama_n_ctx_train(const struct llama_model * model) { + return llama_model_n_ctx_train(model); +} + +// deprecated +int32_t llama_n_embd(const struct llama_model * model) { + return llama_model_n_embd(model); +} + +// deprecated +int32_t llama_n_layer(const struct llama_model * model) { + return llama_model_n_layer(model); +} + +// deprecated +int32_t llama_n_head(const struct llama_model * model) { + return llama_model_n_head(model); +} + +enum llama_rope_type llama_model_rope_type(const struct llama_model * model) { + switch (model->arch) { + // these models do not use RoPE + case LLM_ARCH_GPT2: + case LLM_ARCH_GPTJ: + case LLM_ARCH_MPT: + case LLM_ARCH_REFACT: + case LLM_ARCH_BLOOM: + case LLM_ARCH_MAMBA: + case LLM_ARCH_JINA_BERT_V2: + case LLM_ARCH_T5: + case LLM_ARCH_T5ENCODER: + case LLM_ARCH_JAIS: + case LLM_ARCH_RWKV6: + case LLM_ARCH_RWKV6QWEN2: + case LLM_ARCH_WAVTOKENIZER_DEC: + return LLAMA_ROPE_TYPE_NONE; + + // use what we call a normal RoPE, operating on pairs of consecutive head values + case LLM_ARCH_LLAMA: + case LLM_ARCH_DECI: + case LLM_ARCH_BAICHUAN: + case LLM_ARCH_STARCODER: + case LLM_ARCH_PLAMO: + case LLM_ARCH_ORION: + case LLM_ARCH_INTERNLM2: + case LLM_ARCH_MINICPM: + case LLM_ARCH_XVERSE: + case LLM_ARCH_COMMAND_R: + case LLM_ARCH_COHERE2: + case LLM_ARCH_OLMO: + case LLM_ARCH_ARCTIC: + case LLM_ARCH_DEEPSEEK: + case LLM_ARCH_DEEPSEEK2: + case LLM_ARCH_CHATGLM: + case LLM_ARCH_GRANITE: + case LLM_ARCH_GRANITE_MOE: + case LLM_ARCH_CHAMELEON: + return LLAMA_ROPE_TYPE_NORM; + + // the pairs of head values are offset by n_rot/2 + case LLM_ARCH_FALCON: + case LLM_ARCH_GROK: + case LLM_ARCH_DBRX: + case LLM_ARCH_BERT: + case LLM_ARCH_NOMIC_BERT: + case LLM_ARCH_STABLELM: + case LLM_ARCH_BITNET: + case LLM_ARCH_QWEN: + case LLM_ARCH_QWEN2: + case LLM_ARCH_QWEN2MOE: + case LLM_ARCH_OLMO2: + case LLM_ARCH_OLMOE: + case LLM_ARCH_PHI2: + case LLM_ARCH_PHI3: + case LLM_ARCH_PHIMOE: + case LLM_ARCH_GEMMA: + case LLM_ARCH_GEMMA2: + case LLM_ARCH_STARCODER2: + case LLM_ARCH_OPENELM: + case LLM_ARCH_GPTNEOX: + case LLM_ARCH_CODESHELL: + case LLM_ARCH_NEMOTRON: + case LLM_ARCH_EXAONE: + case LLM_ARCH_MINICPM3: + return LLAMA_ROPE_TYPE_NEOX; + + case LLM_ARCH_QWEN2VL: + return LLAMA_ROPE_TYPE_MROPE; + + // all model arches should be listed explicitly here + case LLM_ARCH_UNKNOWN: + GGML_ABORT("unknown architecture"); + } + + return LLAMA_ROPE_TYPE_NONE; +} + +float llama_model_rope_freq_scale_train(const struct llama_model * model) { + return model->hparams.rope_freq_scale_train; +} + +int32_t llama_model_meta_val_str(const struct llama_model * model, const char * key, char * buf, size_t buf_size) { + const auto & it = model->gguf_kv.find(key); + if (it == model->gguf_kv.end()) { + if (buf_size > 0) { + buf[0] = '\0'; + } + return -1; + } + return snprintf(buf, buf_size, "%s", it->second.c_str()); +} + +int32_t llama_model_meta_count(const struct llama_model * model) { + return (int)model->gguf_kv.size(); +} + +int32_t llama_model_meta_key_by_index(const struct llama_model * model, int i, char * buf, size_t buf_size) { + if (i < 0 || i >= (int)model->gguf_kv.size()) { + if (buf_size > 0) { + buf[0] = '\0'; + } + return -1; + } + auto it = model->gguf_kv.begin(); + std::advance(it, i); + return snprintf(buf, buf_size, "%s", it->first.c_str()); +} + +int32_t llama_model_meta_val_str_by_index(const struct llama_model * model, int32_t i, char * buf, size_t buf_size) { + if (i < 0 || i >= (int)model->gguf_kv.size()) { + if (buf_size > 0) { + buf[0] = '\0'; + } + return -1; + } + auto it = model->gguf_kv.begin(); + std::advance(it, i); + return snprintf(buf, buf_size, "%s", it->second.c_str()); +} + +int32_t llama_model_desc(const struct llama_model * model, char * buf, size_t buf_size) { + return snprintf(buf, buf_size, "%s", model->desc().c_str()); +} + +uint64_t llama_model_size(const struct llama_model * model) { + return model->size(); +} + +const char * llama_model_chat_template(const struct llama_model * model) { + const auto & it = model->gguf_kv.find(LLM_KV(model->arch)(LLM_KV_TOKENIZER_CHAT_TEMPLATE)); + if (it == model->gguf_kv.end()) { + return nullptr; + } + + return it->second.c_str(); +} + +uint64_t llama_model_n_params(const struct llama_model * model) { + return model->n_elements(); +} + +bool llama_model_has_encoder(const struct llama_model * model) { + switch (model->arch) { + case LLM_ARCH_T5: return true; + case LLM_ARCH_T5ENCODER: return true; + default: return false; + } +} + +bool llama_model_has_decoder(const struct llama_model * model) { + switch (model->arch) { + case LLM_ARCH_T5ENCODER: return false; + default: return true; + } +} + +llama_token llama_model_decoder_start_token(const struct llama_model * model) { + return model->hparams.dec_start_token_id; +} + +bool llama_model_is_recurrent(const struct llama_model * model) { + switch (model->arch) { + case LLM_ARCH_MAMBA: return true; + case LLM_ARCH_RWKV6: return true; + case LLM_ARCH_RWKV6QWEN2: return true; + default: return false; + } +} diff --git a/src/llama-model.h b/src/llama-model.h new file mode 100644 index 000000000..4cc8abb75 --- /dev/null +++ b/src/llama-model.h @@ -0,0 +1,372 @@ +#pragma once + +#include "llama.h" +#include "llama-arch.h" +#include "llama-hparams.h" +#include "llama-vocab.h" + +#include +#include +#include +#include + +struct llama_model_loader; + +// available models +enum llm_type { + LLM_TYPE_UNKNOWN, + LLM_TYPE_14M, + LLM_TYPE_17M, + LLM_TYPE_22M, + LLM_TYPE_33M, + LLM_TYPE_60M, + LLM_TYPE_70M, + LLM_TYPE_80M, + LLM_TYPE_109M, + LLM_TYPE_137M, + LLM_TYPE_160M, + LLM_TYPE_220M, + LLM_TYPE_250M, + LLM_TYPE_270M, + LLM_TYPE_335M, + LLM_TYPE_410M, + LLM_TYPE_450M, + LLM_TYPE_770M, + LLM_TYPE_780M, + LLM_TYPE_0_5B, + LLM_TYPE_1B, + LLM_TYPE_1_3B, + LLM_TYPE_1_4B, + LLM_TYPE_1_5B, + LLM_TYPE_1_6B, + LLM_TYPE_2B, + LLM_TYPE_2_8B, + LLM_TYPE_3B, + LLM_TYPE_4B, + LLM_TYPE_6B, + LLM_TYPE_6_9B, + LLM_TYPE_7B, + LLM_TYPE_8B, + LLM_TYPE_9B, + LLM_TYPE_11B, + LLM_TYPE_12B, + LLM_TYPE_13B, + LLM_TYPE_14B, + LLM_TYPE_15B, + LLM_TYPE_16B, + LLM_TYPE_20B, + LLM_TYPE_30B, + LLM_TYPE_32B, + LLM_TYPE_34B, + LLM_TYPE_35B, + LLM_TYPE_40B, + LLM_TYPE_65B, + LLM_TYPE_70B, + LLM_TYPE_236B, + LLM_TYPE_314B, + LLM_TYPE_671B, + LLM_TYPE_SMALL, + LLM_TYPE_MEDIUM, + LLM_TYPE_LARGE, + LLM_TYPE_XL, + LLM_TYPE_A1_7B, + LLM_TYPE_A2_7B, + LLM_TYPE_8x7B, + LLM_TYPE_8x22B, + LLM_TYPE_16x12B, + LLM_TYPE_16x3_8B, + LLM_TYPE_10B_128x3_66B, + LLM_TYPE_57B_A14B, + LLM_TYPE_27B, +}; + +struct llama_layer_posnet { + // resnet + struct ggml_tensor * norm1 = nullptr; + struct ggml_tensor * norm1_b = nullptr; + + struct ggml_tensor * conv1 = nullptr; + struct ggml_tensor * conv1_b = nullptr; + + struct ggml_tensor * norm2 = nullptr; + struct ggml_tensor * norm2_b = nullptr; + + struct ggml_tensor * conv2 = nullptr; + struct ggml_tensor * conv2_b = nullptr; + + // attention + struct ggml_tensor * attn_norm = nullptr; + struct ggml_tensor * attn_norm_b = nullptr; + + struct ggml_tensor * attn_q = nullptr; + struct ggml_tensor * attn_q_b = nullptr; + + struct ggml_tensor * attn_k = nullptr; + struct ggml_tensor * attn_k_b = nullptr; + + struct ggml_tensor * attn_v = nullptr; + struct ggml_tensor * attn_v_b = nullptr; + + struct ggml_tensor * attn_o = nullptr; + struct ggml_tensor * attn_o_b = nullptr; + + // normalize + struct ggml_tensor * norm = nullptr; + struct ggml_tensor * norm_b = nullptr; +}; + +struct llama_layer_convnext { + struct ggml_tensor * dw = nullptr; + struct ggml_tensor * dw_b = nullptr; + + struct ggml_tensor * norm = nullptr; + struct ggml_tensor * norm_b = nullptr; + + struct ggml_tensor * pw1 = nullptr; + struct ggml_tensor * pw1_b = nullptr; + + struct ggml_tensor * pw2 = nullptr; + struct ggml_tensor * pw2_b = nullptr; + + struct ggml_tensor * gamma = nullptr; +}; + +struct llama_layer { + // normalization + struct ggml_tensor * attn_norm = nullptr; + struct ggml_tensor * attn_norm_b = nullptr; + struct ggml_tensor * attn_norm_2 = nullptr; + struct ggml_tensor * attn_norm_2_b = nullptr; + struct ggml_tensor * attn_q_norm = nullptr; + struct ggml_tensor * attn_q_norm_b = nullptr; + struct ggml_tensor * attn_k_norm = nullptr; + struct ggml_tensor * attn_k_norm_b = nullptr; + struct ggml_tensor * attn_out_norm = nullptr; + struct ggml_tensor * attn_out_norm_b = nullptr; + struct ggml_tensor * attn_q_a_norm = nullptr; + struct ggml_tensor * attn_kv_a_norm = nullptr; + struct ggml_tensor * attn_sub_norm = nullptr; + struct ggml_tensor * attn_post_norm = nullptr; + struct ggml_tensor * ffn_sub_norm = nullptr; + struct ggml_tensor * attn_norm_cross = nullptr; + struct ggml_tensor * attn_norm_enc = nullptr; + + // attention + struct ggml_tensor * wq = nullptr; + struct ggml_tensor * wk = nullptr; + struct ggml_tensor * wv = nullptr; + struct ggml_tensor * wo = nullptr; + struct ggml_tensor * wqkv = nullptr; + struct ggml_tensor * wq_a = nullptr; + struct ggml_tensor * wq_b = nullptr; + struct ggml_tensor * wkv_a_mqa = nullptr; + struct ggml_tensor * wkv_b = nullptr; + struct ggml_tensor * wq_cross = nullptr; + struct ggml_tensor * wk_cross = nullptr; + struct ggml_tensor * wv_cross = nullptr; + struct ggml_tensor * wo_cross = nullptr; + struct ggml_tensor * wq_enc = nullptr; + struct ggml_tensor * wk_enc = nullptr; + struct ggml_tensor * wv_enc = nullptr; + struct ggml_tensor * wo_enc = nullptr; + + // attention bias + struct ggml_tensor * bq = nullptr; + struct ggml_tensor * bk = nullptr; + struct ggml_tensor * bv = nullptr; + struct ggml_tensor * bo = nullptr; + struct ggml_tensor * bqkv = nullptr; + + // relative position bias + struct ggml_tensor * attn_rel_b = nullptr; + struct ggml_tensor * attn_rel_b_enc = nullptr; + struct ggml_tensor * attn_rel_b_cross = nullptr; + + // normalization + struct ggml_tensor * ffn_norm = nullptr; + struct ggml_tensor * ffn_norm_b = nullptr; + struct ggml_tensor * ffn_post_norm = nullptr; + struct ggml_tensor * layer_out_norm = nullptr; + struct ggml_tensor * layer_out_norm_b = nullptr; + struct ggml_tensor * ffn_norm_exps = nullptr; + struct ggml_tensor * ffn_norm_enc = nullptr; + + // ff + struct ggml_tensor * ffn_gate = nullptr; // w1 + struct ggml_tensor * ffn_down = nullptr; // w2 + struct ggml_tensor * ffn_up = nullptr; // w3 + struct ggml_tensor * ffn_gate_enc = nullptr; + struct ggml_tensor * ffn_down_enc = nullptr; + struct ggml_tensor * ffn_up_enc = nullptr; + + // ff MoE + struct ggml_tensor * ffn_gate_inp = nullptr; + struct ggml_tensor * ffn_gate_exps = nullptr; + struct ggml_tensor * ffn_down_exps = nullptr; + struct ggml_tensor * ffn_up_exps = nullptr; + + // ff shared expert (shexp) + struct ggml_tensor * ffn_gate_inp_shexp = nullptr; + struct ggml_tensor * ffn_gate_shexp = nullptr; + struct ggml_tensor * ffn_down_shexp = nullptr; + struct ggml_tensor * ffn_up_shexp = nullptr; + + // ff bias + struct ggml_tensor * ffn_gate_b = nullptr; + struct ggml_tensor * ffn_down_b = nullptr; // b2 + struct ggml_tensor * ffn_up_b = nullptr; // b3 + struct ggml_tensor * ffn_act = nullptr; + struct ggml_tensor * ffn_exp_probs_b = nullptr; + + // mamba proj + struct ggml_tensor * ssm_in = nullptr; + struct ggml_tensor * ssm_x = nullptr; + struct ggml_tensor * ssm_dt = nullptr; + struct ggml_tensor * ssm_out = nullptr; + + // mamba + struct ggml_tensor * ssm_conv1d = nullptr; + struct ggml_tensor * ssm_a = nullptr; + struct ggml_tensor * ssm_d = nullptr; + + // mamba bias + struct ggml_tensor * ssm_conv1d_b = nullptr; + struct ggml_tensor * ssm_dt_b = nullptr; + + // rwkv + struct ggml_tensor * time_mix_w1 = nullptr; + struct ggml_tensor * time_mix_w2 = nullptr; + struct ggml_tensor * time_mix_lerp_x = nullptr; + struct ggml_tensor * time_mix_lerp_w = nullptr; + struct ggml_tensor * time_mix_lerp_k = nullptr; + struct ggml_tensor * time_mix_lerp_v = nullptr; + struct ggml_tensor * time_mix_lerp_r = nullptr; + struct ggml_tensor * time_mix_lerp_g = nullptr; + struct ggml_tensor * time_mix_lerp_fused = nullptr; + + struct ggml_tensor * time_mix_first = nullptr; + struct ggml_tensor * time_mix_decay = nullptr; + struct ggml_tensor * time_mix_decay_w1 = nullptr; + struct ggml_tensor * time_mix_decay_w2 = nullptr; + struct ggml_tensor * time_mix_key = nullptr; + struct ggml_tensor * time_mix_key_b = nullptr; + struct ggml_tensor * time_mix_value = nullptr; + struct ggml_tensor * time_mix_value_b = nullptr; + struct ggml_tensor * time_mix_receptance = nullptr; + struct ggml_tensor * time_mix_receptance_b = nullptr; + struct ggml_tensor * time_mix_gate = nullptr; + + struct ggml_tensor * time_mix_ln = nullptr; + struct ggml_tensor * time_mix_ln_b = nullptr; + struct ggml_tensor * time_mix_output = nullptr; + + struct ggml_tensor * channel_mix_lerp_k = nullptr; + struct ggml_tensor * channel_mix_lerp_r = nullptr; + + struct ggml_tensor * channel_mix_key = nullptr; + struct ggml_tensor * channel_mix_receptance = nullptr; + struct ggml_tensor * channel_mix_value = nullptr; + + // long rope factors + struct ggml_tensor * rope_long = nullptr; + struct ggml_tensor * rope_short = nullptr; + struct ggml_tensor * rope_freqs = nullptr; + + // bitnet scale + struct ggml_tensor * wq_scale = nullptr; + struct ggml_tensor * wk_scale = nullptr; + struct ggml_tensor * wv_scale = nullptr; + struct ggml_tensor * wo_scale = nullptr; + struct ggml_tensor * ffn_gate_scale = nullptr; + struct ggml_tensor * ffn_up_scale = nullptr; + struct ggml_tensor * ffn_down_scale = nullptr; + + struct llama_layer_posnet posnet; + + struct llama_layer_convnext convnext; +}; + +struct llama_model { + llm_type type = LLM_TYPE_UNKNOWN; + llm_arch arch = LLM_ARCH_UNKNOWN; + + std::string name = "n/a"; + + llama_hparams hparams = {}; + llama_vocab vocab; + + struct ggml_tensor * tok_embd = nullptr; + struct ggml_tensor * type_embd = nullptr; + struct ggml_tensor * pos_embd = nullptr; + struct ggml_tensor * tok_norm = nullptr; + struct ggml_tensor * tok_norm_b = nullptr; + + struct ggml_tensor * output_norm = nullptr; + struct ggml_tensor * output_norm_b = nullptr; + struct ggml_tensor * output = nullptr; + struct ggml_tensor * output_b = nullptr; + struct ggml_tensor * output_norm_enc = nullptr; + + // classifier + struct ggml_tensor * cls = nullptr; + struct ggml_tensor * cls_b = nullptr; + struct ggml_tensor * cls_out = nullptr; + struct ggml_tensor * cls_out_b = nullptr; + + struct ggml_tensor * conv1d = nullptr; + struct ggml_tensor * conv1d_b = nullptr; + + std::vector layers; + + llama_model_params params; + + // gguf metadata + std::unordered_map gguf_kv; + + std::vector rpc_servers; + + // list of devices used in this model + std::vector devices; + + // for quantize-stats only + std::vector> tensors_by_name; + + int64_t t_load_us = 0; + int64_t t_start_us = 0; + + explicit llama_model(const struct llama_model_params & params); + ~llama_model(); + + void load_stats (llama_model_loader & ml); + void load_arch (llama_model_loader & ml); + void load_hparams(llama_model_loader & ml); + void load_vocab (llama_model_loader & ml); + bool load_tensors(llama_model_loader & ml); // returns false if cancelled by progress_callback + + std::string arch_name() const; + std::string type_name() const; + + std::string desc() const; + + size_t size() const; + size_t max_nodes() const; + size_t n_devices() const; + + // total number of parameters in the model + uint64_t n_elements() const; + + void print_info() const; + + ggml_backend_dev_t dev_layer(int il) const; + ggml_backend_dev_t dev_output() const; + + ggml_backend_buffer_type_t select_buft(int il) const; + + const struct ggml_tensor * get_tensor(const char * name) const; + +private: + struct impl; + std::unique_ptr pimpl; +}; + +const char * llm_type_name(llm_type type); diff --git a/src/llama-quant.cpp b/src/llama-quant.cpp new file mode 100644 index 000000000..d4947a780 --- /dev/null +++ b/src/llama-quant.cpp @@ -0,0 +1,933 @@ +#include "llama-quant.h" + +#include "llama-impl.h" +#include "llama-model.h" +#include "llama-model-loader.h" + +#include +#include +#include +#include +#include +#include +#include +#include + +static void zeros(std::ofstream & file, size_t n) { + char zero = 0; + for (size_t i = 0; i < n; ++i) { + file.write(&zero, 1); + } +} + +struct quantize_state_impl { + const llama_model & model; + const llama_model_quantize_params * params; + + int n_attention_wv = 0; + int n_ffn_down = 0; + int n_ffn_gate = 0; + int n_ffn_up = 0; + int i_attention_wv = 0; + int i_ffn_down = 0; + int i_ffn_gate = 0; + int i_ffn_up = 0; + + int n_k_quantized = 0; + int n_fallback = 0; + + bool has_imatrix = false; + + // used to figure out if a model shares tok_embd with the output weight + bool has_output = false; + + quantize_state_impl(const llama_model & model, const llama_model_quantize_params * params) + : model(model) + , params(params) + {} +}; + +static void llama_tensor_dequantize_impl( + struct ggml_tensor * tensor, std::vector> & output, std::vector & workers, + const size_t nelements, const int nthread +) { + if (output.size() < nelements) { + output.resize(nelements); + } + float * f32_output = (float *) output.data(); + + const ggml_type_traits * qtype = ggml_get_type_traits(tensor->type); + if (ggml_is_quantized(tensor->type)) { + if (qtype->to_float == NULL) { + throw std::runtime_error(format("type %s unsupported for integer quantization: no dequantization available", ggml_type_name(tensor->type))); + } + } else if (tensor->type != GGML_TYPE_F16 && + tensor->type != GGML_TYPE_BF16) { + throw std::runtime_error(format("cannot dequantize/convert tensor type %s", ggml_type_name(tensor->type))); + } + + if (nthread < 2) { + if (tensor->type == GGML_TYPE_F16) { + ggml_fp16_to_fp32_row((ggml_fp16_t *)tensor->data, f32_output, nelements); + } else if (tensor->type == GGML_TYPE_BF16) { + ggml_bf16_to_fp32_row((ggml_bf16_t *)tensor->data, f32_output, nelements); + } else if (ggml_is_quantized(tensor->type)) { + qtype->to_float(tensor->data, f32_output, nelements); + } else { + GGML_ABORT("fatal error"); // unreachable + } + return; + } + + size_t block_size; + if (tensor->type == GGML_TYPE_F16 || + tensor->type == GGML_TYPE_BF16) { + block_size = 1; + } else { + block_size = (size_t)ggml_blck_size(tensor->type); + } + + size_t block_size_bytes = ggml_type_size(tensor->type); + + GGML_ASSERT(nelements % block_size == 0); + size_t nblocks = nelements / block_size; + size_t blocks_per_thread = nblocks / nthread; + size_t spare_blocks = nblocks - (blocks_per_thread * nthread); // if blocks aren't divisible by thread count + + size_t in_buff_offs = 0; + size_t out_buff_offs = 0; + + for (int tnum = 0; tnum < nthread; tnum++) { + size_t thr_blocks = blocks_per_thread + (tnum == nthread - 1 ? spare_blocks : 0); // num blocks for this thread + size_t thr_elems = thr_blocks * block_size; // number of elements for this thread + size_t thr_block_bytes = thr_blocks * block_size_bytes; // number of input bytes for this thread + + auto compute = [qtype] (ggml_type typ, uint8_t * inbuf, float * outbuf, int nels) { + if (typ == GGML_TYPE_F16) { + ggml_fp16_to_fp32_row((ggml_fp16_t *)inbuf, outbuf, nels); + } else if (typ == GGML_TYPE_BF16) { + ggml_bf16_to_fp32_row((ggml_bf16_t *)inbuf, outbuf, nels); + } else { + qtype->to_float(inbuf, outbuf, nels); + } + }; + workers.emplace_back(compute, tensor->type, (uint8_t *) tensor->data + in_buff_offs, f32_output + out_buff_offs, thr_elems); + in_buff_offs += thr_block_bytes; + out_buff_offs += thr_elems; + } + for (auto & w : workers) { w.join(); } + workers.clear(); +} + +static ggml_type llama_tensor_get_type(quantize_state_impl & qs, ggml_type new_type, const ggml_tensor * tensor, llama_ftype ftype) { + const std::string name = ggml_get_name(tensor); + + // TODO: avoid hardcoded tensor names - use the TN_* constants + const llm_arch arch = qs.model.arch; + const auto tn = LLM_TN(arch); + + auto use_more_bits = [](int i_layer, int n_layers) -> bool { + return i_layer < n_layers/8 || i_layer >= 7*n_layers/8 || (i_layer - n_layers/8)%3 == 2; + }; + const int n_expert = std::max(1, (int)qs.model.hparams.n_expert); + auto layer_info = [n_expert] (int i_layer, int n_layer, const char * name) { + if (n_expert > 1) { + // Believe it or not, "experts" in the FFN of Mixtral-8x7B are not consecutive, but occasionally randomly + // sprinkled in the model. Hence, simply dividing i_ffn_down by n_expert does not work + // for getting the current layer as I initially thought, and we need to resort to parsing the + // tensor name. + if (sscanf(name, "blk.%d.", &i_layer) != 1) { + throw std::runtime_error(format("Failed to determine layer for tensor %s", name)); + } + if (i_layer < 0 || i_layer >= n_layer) { + throw std::runtime_error(format("Bad layer %d for tensor %s. Must be in [0, %d)", i_layer, name, n_layer)); + } + } + return std::make_pair(i_layer, n_layer); + }; + + // for arches that share the same tensor between the token embeddings and the output, we quantize the token embeddings + // with the quantization of the output tensor + if (name == tn(LLM_TENSOR_OUTPUT, "weight") || (!qs.has_output && name == tn(LLM_TENSOR_TOKEN_EMBD, "weight"))) { + if (qs.params->output_tensor_type < GGML_TYPE_COUNT) { + new_type = qs.params->output_tensor_type; + } else { + const int64_t nx = tensor->ne[0]; + const int64_t qk_k = ggml_blck_size(new_type); + + if (arch == LLM_ARCH_FALCON || nx % qk_k != 0) { + new_type = GGML_TYPE_Q8_0; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS || + ftype == LLAMA_FTYPE_MOSTLY_IQ1_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M || + ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) { + new_type = GGML_TYPE_Q5_K; + } + else if (new_type != GGML_TYPE_Q8_0) { + new_type = GGML_TYPE_Q6_K; + } + } + } else if (name == "token_embd.weight") { + if (qs.params->token_embedding_type < GGML_TYPE_COUNT) { + new_type = qs.params->token_embedding_type; + } else { + if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || + ftype == LLAMA_FTYPE_MOSTLY_IQ1_S || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) { + new_type = GGML_TYPE_Q2_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M) { + new_type = GGML_TYPE_IQ3_S; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) { + new_type = GGML_TYPE_IQ3_S; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_TQ1_0 || ftype == LLAMA_FTYPE_MOSTLY_TQ2_0) { + new_type = GGML_TYPE_Q4_K; + } + } + } else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ1_S || + ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) { + if (name.find("attn_v.weight") != std::string::npos) { + if (qs.model.hparams.n_gqa() >= 4 || qs.model.hparams.n_expert >= 4) new_type = GGML_TYPE_Q4_K; + else new_type = ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M ? GGML_TYPE_IQ3_S : GGML_TYPE_Q2_K; + ++qs.i_attention_wv; + } + else if (qs.model.hparams.n_expert == 8 && name.find("attn_k.weight") != std::string::npos) { + new_type = GGML_TYPE_Q4_K; + } + else if (name.find("ffn_down") != std::string::npos) { + if (qs.i_ffn_down < qs.n_ffn_down/8) { + new_type = ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M ? GGML_TYPE_IQ3_S : GGML_TYPE_Q2_K; + } + ++qs.i_ffn_down; + } + else if (name.find("attn_output.weight") != std::string::npos) { + if (qs.model.hparams.n_expert == 8) { + new_type = GGML_TYPE_Q5_K; + } else { + if (ftype == LLAMA_FTYPE_MOSTLY_IQ1_S || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) new_type = GGML_TYPE_IQ2_XXS; + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M) new_type = GGML_TYPE_IQ3_S; + } + } + } else if (name.find("attn_v.weight") != std::string::npos) { + if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) { + new_type = qs.model.hparams.n_gqa() >= 4 ? GGML_TYPE_Q4_K : GGML_TYPE_Q3_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S && qs.model.hparams.n_gqa() >= 4) { + new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) { + new_type = qs.model.hparams.n_gqa() >= 4 ? GGML_TYPE_Q4_K : !qs.has_imatrix ? GGML_TYPE_IQ3_S : GGML_TYPE_IQ3_XXS; + } + else if ((ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_S) && qs.model.hparams.n_gqa() >= 4) { + new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M) { + new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) { + new_type = qs.i_attention_wv < 2 ? GGML_TYPE_Q5_K : GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q5_K; + else if ((ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) && qs.model.hparams.n_gqa() >= 4) { + new_type = GGML_TYPE_Q5_K; + } + else if ((ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M || ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M) && + use_more_bits(qs.i_attention_wv, qs.n_attention_wv)) new_type = GGML_TYPE_Q6_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S && qs.i_attention_wv < 4) new_type = GGML_TYPE_Q5_K; + if (qs.model.type == LLM_TYPE_70B) { + // In the 70B model we have 8 heads sharing the same attn_v weights. As a result, the attn_v.weight tensor is + // 8x smaller compared to attn_q.weight. Hence, we can get a nice boost in quantization accuracy with + // nearly negligible increase in model size by quantizing this tensor with more bits: + if (new_type == GGML_TYPE_Q3_K || new_type == GGML_TYPE_Q4_K) new_type = GGML_TYPE_Q5_K; + } + if (qs.model.hparams.n_expert == 8) { + // for the 8-expert model, bumping this to Q8_0 trades just ~128MB + // TODO: explore better strategies + new_type = GGML_TYPE_Q8_0; + } + ++qs.i_attention_wv; + } else if (name.find("attn_k.weight") != std::string::npos) { + if (qs.model.hparams.n_expert == 8) { + // for the 8-expert model, bumping this to Q8_0 trades just ~128MB + // TODO: explore better strategies + new_type = GGML_TYPE_Q8_0; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS) { + new_type = GGML_TYPE_IQ3_XXS; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) { + new_type = GGML_TYPE_IQ2_S; + } + } else if (name.find("attn_q.weight") != std::string::npos) { + if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS) { + new_type = GGML_TYPE_IQ3_XXS; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) { + new_type = GGML_TYPE_IQ2_S; + } + } else if (name.find("ffn_down") != std::string::npos) { + auto info = layer_info(qs.i_ffn_down, qs.n_ffn_down, name.c_str()); + int i_layer = info.first, n_layer = info.second; + if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S) { + if (i_layer < n_layer/8) new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS && !qs.has_imatrix) { + new_type = i_layer < n_layer/8 ? GGML_TYPE_Q4_K : GGML_TYPE_Q3_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) { + new_type = i_layer < n_layer/16 ? GGML_TYPE_Q5_K + : arch != LLM_ARCH_FALCON || use_more_bits(i_layer, n_layer) ? GGML_TYPE_Q4_K + : GGML_TYPE_Q3_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M && (i_layer < n_layer/8 || + (qs.model.hparams.n_expert == 8 && use_more_bits(i_layer, n_layer)))) { + new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) { + new_type = arch == LLM_ARCH_FALCON ? GGML_TYPE_Q4_K : GGML_TYPE_Q5_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M) { + if (arch == LLM_ARCH_FALCON) { + new_type = i_layer < n_layer/16 ? GGML_TYPE_Q6_K : + use_more_bits(i_layer, n_layer) ? GGML_TYPE_Q5_K : GGML_TYPE_Q4_K; + } else { + if (use_more_bits(i_layer, n_layer)) new_type = GGML_TYPE_Q6_K; + } + } + else if (i_layer < n_layer/8 && (ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) && !qs.has_imatrix) { + new_type = GGML_TYPE_Q5_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M && use_more_bits(i_layer, n_layer)) new_type = GGML_TYPE_Q6_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S && arch != LLM_ARCH_FALCON && i_layer < n_layer/8) { + new_type = GGML_TYPE_Q5_K; + } + else if ((ftype == LLAMA_FTYPE_MOSTLY_Q4_0 || ftype == LLAMA_FTYPE_MOSTLY_Q5_0) + && qs.has_imatrix && i_layer < n_layer/8) { + // Guard against craziness in the first few ffn_down layers that can happen even with imatrix for Q4_0/Q5_0. + // We only do it when an imatrix is provided because a) we want to make sure that one can always get the + // same quantization as before imatrix stuff, and b) Q4_1/Q5_1 do go crazy on ffn_down without an imatrix. + new_type = ftype == LLAMA_FTYPE_MOSTLY_Q4_0 ? GGML_TYPE_Q4_1 : GGML_TYPE_Q5_1; + } + ++qs.i_ffn_down; + } else if (name.find("attn_output.weight") != std::string::npos) { + if (arch != LLM_ARCH_FALCON) { + if (qs.model.hparams.n_expert == 8) { + if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS || + ftype == LLAMA_FTYPE_MOSTLY_Q3_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M || ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL || + ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M || ftype == LLAMA_FTYPE_MOSTLY_IQ3_S || + ftype == LLAMA_FTYPE_MOSTLY_IQ3_M || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) { + new_type = GGML_TYPE_Q5_K; + } + } else { + if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K ) new_type = GGML_TYPE_Q3_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) new_type = GGML_TYPE_IQ3_S; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M ) new_type = GGML_TYPE_Q4_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L ) new_type = GGML_TYPE_Q5_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M ) new_type = GGML_TYPE_Q4_K; + } + } else { + if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q4_K; + } + } + else if (name.find("attn_qkv.weight") != std::string::npos) { + if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L || ftype == LLAMA_FTYPE_MOSTLY_IQ3_M) { + new_type = GGML_TYPE_Q4_K; + } + else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M) new_type = GGML_TYPE_Q5_K; + else if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M) new_type = GGML_TYPE_Q6_K; + } + else if (name.find("ffn_gate") != std::string::npos) { + auto info = layer_info(qs.i_ffn_gate, qs.n_ffn_gate, name.c_str()); + int i_layer = info.first, n_layer = info.second; + if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS && (i_layer >= n_layer/8 && i_layer < 7*n_layer/8)) { + new_type = GGML_TYPE_IQ3_XXS; + } + ++qs.i_ffn_gate; + } + else if (name.find("ffn_up") != std::string::npos) { + auto info = layer_info(qs.i_ffn_up, qs.n_ffn_up, name.c_str()); + int i_layer = info.first, n_layer = info.second; + if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS && (i_layer >= n_layer/8 && i_layer < 7*n_layer/8)) { + new_type = GGML_TYPE_IQ3_XXS; + } + ++qs.i_ffn_up; + } + + // if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K; + //} + // IK: let's remove this, else Q2_K is almost the same as Q3_K_S + //else if (name.find("ffn_gate") != std::string::npos || name.find("ffn_up") != std::string::npos) { + // if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K; + //} + // This can be used to reduce the size of the Q5_K_S model. + // The associated PPL increase is fully in line with the size reduction + //else { + // if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_S) new_type = GGML_TYPE_Q4_K; + //} + bool convert_incompatible_tensor = false; + { + const int64_t nx = tensor->ne[0]; + const int64_t ny = tensor->ne[1]; + const int64_t qk_k = ggml_blck_size(new_type); + + if (nx % qk_k != 0) { + LLAMA_LOG_WARN("\n\n%s : tensor cols %" PRId64 " x %" PRId64 " are not divisible by %" PRId64 ", required for %s", __func__, nx, ny, qk_k, ggml_type_name(new_type)); + convert_incompatible_tensor = true; + } else { + ++qs.n_k_quantized; + } + } + + if (convert_incompatible_tensor) { + switch (new_type) { + case GGML_TYPE_TQ1_0: + case GGML_TYPE_TQ2_0: new_type = GGML_TYPE_Q4_0; break; // TODO: use a symmetric type instead + case GGML_TYPE_IQ2_XXS: + case GGML_TYPE_IQ2_XS: + case GGML_TYPE_IQ2_S: + case GGML_TYPE_IQ3_XXS: + case GGML_TYPE_IQ3_S: + case GGML_TYPE_IQ1_S: + case GGML_TYPE_IQ1_M: + case GGML_TYPE_Q2_K: + case GGML_TYPE_Q3_K: + case GGML_TYPE_IQ4_XS: new_type = GGML_TYPE_IQ4_NL; break; + case GGML_TYPE_Q4_K: new_type = GGML_TYPE_Q5_0; break; + case GGML_TYPE_Q5_K: new_type = GGML_TYPE_Q5_1; break; + case GGML_TYPE_Q6_K: new_type = GGML_TYPE_Q8_0; break; + default: throw std::runtime_error("\nUnsupported tensor size encountered\n"); + } + if (tensor->ne[0] % ggml_blck_size(new_type) != 0) { + new_type = GGML_TYPE_F16; + } + LLAMA_LOG_WARN(" - using fallback quantization %s\n", ggml_type_name(new_type)); + ++qs.n_fallback; + } + + return new_type; +} + +static size_t llama_tensor_quantize_impl(enum ggml_type new_type, const float * f32_data, void * new_data, const int64_t chunk_size, int64_t nrows, int64_t n_per_row, const float * imatrix, std::vector & workers, const int nthread) { + if (nthread < 2) { + // single-thread + size_t new_size = ggml_quantize_chunk(new_type, f32_data, new_data, 0, nrows, n_per_row, imatrix); + if (!ggml_validate_row_data(new_type, new_data, new_size)) { + throw std::runtime_error("quantized data validation failed"); + } + return new_size; + } + + std::mutex mutex; + int64_t counter = 0; + size_t new_size = 0; + bool valid = true; + auto compute = [&mutex, &counter, &new_size, &valid, new_type, f32_data, new_data, chunk_size, + nrows, n_per_row, imatrix]() { + const int64_t nrows_per_chunk = chunk_size / n_per_row; + size_t local_size = 0; + while (true) { + std::unique_lock lock(mutex); + int64_t first_row = counter; counter += nrows_per_chunk; + if (first_row >= nrows) { + if (local_size > 0) { + new_size += local_size; + } + break; + } + lock.unlock(); + const int64_t this_nrow = std::min(nrows - first_row, nrows_per_chunk); + size_t this_size = ggml_quantize_chunk(new_type, f32_data, new_data, first_row * n_per_row, this_nrow, n_per_row, imatrix); + local_size += this_size; + + // validate the quantized data + const size_t row_size = ggml_row_size(new_type, n_per_row); + void * this_data = (char *) new_data + first_row * row_size; + if (!ggml_validate_row_data(new_type, this_data, this_size)) { + std::unique_lock lock(mutex); + valid = false; + break; + } + } + }; + for (int it = 0; it < nthread - 1; ++it) { + workers.emplace_back(compute); + } + compute(); + for (auto & w : workers) { w.join(); } + workers.clear(); + if (!valid) { + throw std::runtime_error("quantized data validation failed"); + } + return new_size; +} + +static void llama_model_quantize_impl(const std::string & fname_inp, const std::string & fname_out, const llama_model_quantize_params * params) { + ggml_type default_type; + llama_ftype ftype = params->ftype; + + switch (params->ftype) { + case LLAMA_FTYPE_MOSTLY_Q4_0: default_type = GGML_TYPE_Q4_0; break; + case LLAMA_FTYPE_MOSTLY_Q4_1: default_type = GGML_TYPE_Q4_1; break; + case LLAMA_FTYPE_MOSTLY_Q5_0: default_type = GGML_TYPE_Q5_0; break; + case LLAMA_FTYPE_MOSTLY_Q5_1: default_type = GGML_TYPE_Q5_1; break; + case LLAMA_FTYPE_MOSTLY_Q8_0: default_type = GGML_TYPE_Q8_0; break; + case LLAMA_FTYPE_MOSTLY_F16: default_type = GGML_TYPE_F16; break; + case LLAMA_FTYPE_MOSTLY_BF16: default_type = GGML_TYPE_BF16; break; + case LLAMA_FTYPE_ALL_F32: default_type = GGML_TYPE_F32; break; + + // K-quants + case LLAMA_FTYPE_MOSTLY_Q2_K_S: + case LLAMA_FTYPE_MOSTLY_Q2_K: default_type = GGML_TYPE_Q2_K; break; + case LLAMA_FTYPE_MOSTLY_IQ3_XS: default_type = GGML_TYPE_IQ3_S; break; + case LLAMA_FTYPE_MOSTLY_Q3_K_S: + case LLAMA_FTYPE_MOSTLY_Q3_K_M: + case LLAMA_FTYPE_MOSTLY_Q3_K_L: default_type = GGML_TYPE_Q3_K; break; + case LLAMA_FTYPE_MOSTLY_Q4_K_S: + case LLAMA_FTYPE_MOSTLY_Q4_K_M: default_type = GGML_TYPE_Q4_K; break; + case LLAMA_FTYPE_MOSTLY_Q5_K_S: + case LLAMA_FTYPE_MOSTLY_Q5_K_M: default_type = GGML_TYPE_Q5_K; break; + case LLAMA_FTYPE_MOSTLY_Q6_K: default_type = GGML_TYPE_Q6_K; break; + case LLAMA_FTYPE_MOSTLY_TQ1_0: default_type = GGML_TYPE_TQ1_0; break; + case LLAMA_FTYPE_MOSTLY_TQ2_0: default_type = GGML_TYPE_TQ2_0; break; + case LLAMA_FTYPE_MOSTLY_IQ2_XXS: default_type = GGML_TYPE_IQ2_XXS; break; + case LLAMA_FTYPE_MOSTLY_IQ2_XS: default_type = GGML_TYPE_IQ2_XS; break; + case LLAMA_FTYPE_MOSTLY_IQ2_S: default_type = GGML_TYPE_IQ2_XS; break; + case LLAMA_FTYPE_MOSTLY_IQ2_M: default_type = GGML_TYPE_IQ2_S; break; + case LLAMA_FTYPE_MOSTLY_IQ3_XXS: default_type = GGML_TYPE_IQ3_XXS; break; + case LLAMA_FTYPE_MOSTLY_IQ1_S: default_type = GGML_TYPE_IQ1_S; break; + case LLAMA_FTYPE_MOSTLY_IQ1_M: default_type = GGML_TYPE_IQ1_M; break; + case LLAMA_FTYPE_MOSTLY_IQ4_NL: default_type = GGML_TYPE_IQ4_NL; break; + case LLAMA_FTYPE_MOSTLY_IQ4_XS: default_type = GGML_TYPE_IQ4_XS; break; + case LLAMA_FTYPE_MOSTLY_IQ3_S: default_type = GGML_TYPE_IQ3_S; break; + case LLAMA_FTYPE_MOSTLY_IQ3_M: default_type = GGML_TYPE_IQ3_S; break; + + default: throw std::runtime_error(format("invalid output file type %d\n", ftype)); + } + + int nthread = params->nthread; + + if (nthread <= 0) { + nthread = std::thread::hardware_concurrency(); + } + + // mmap consistently increases speed Linux, and also increases speed on Windows with + // hot cache. It may cause a slowdown on macOS, possibly related to free memory. +#if defined(__linux__) || defined(_WIN32) + constexpr bool use_mmap = true; +#else + constexpr bool use_mmap = false; +#endif + + llama_model_kv_override * kv_overrides = nullptr; + if (params->kv_overrides) { + auto v = (std::vector*)params->kv_overrides; + kv_overrides = v->data(); + } + + llama_model_loader ml(fname_inp, use_mmap, /*check_tensors*/ true, kv_overrides); + ml.init_mappings(false); // no prefetching + + llama_model model(llama_model_default_params()); + + model.load_arch (ml); + model.load_hparams(ml); + model.load_stats (ml); + + struct quantize_state_impl qs(model, params); + + if (params->only_copy) { + ftype = ml.ftype; + } + const std::unordered_map> * imatrix_data = nullptr; + if (params->imatrix) { + imatrix_data = static_cast>*>(params->imatrix); + if (imatrix_data) { + LLAMA_LOG_INFO("================================ Have weights data with %d entries\n",int(imatrix_data->size())); + qs.has_imatrix = true; + // check imatrix for nans or infs + for (const auto & kv : *imatrix_data) { + for (float f : kv.second) { + if (!std::isfinite(f)) { + throw std::runtime_error(format("imatrix contains non-finite value %f\n", f)); + } + } + } + } + } + + const size_t align = GGUF_DEFAULT_ALIGNMENT; + gguf_context_ptr ctx_out { gguf_init_empty() }; + + // copy the KV pairs from the input file + gguf_set_kv (ctx_out.get(), ml.meta.get()); + gguf_set_val_u32(ctx_out.get(), "general.quantization_version", GGML_QNT_VERSION); // TODO: use LLM_KV + gguf_set_val_u32(ctx_out.get(), "general.file_type", ftype); // TODO: use LLM_KV + + // Remove split metadata + gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_NO).c_str()); + gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_COUNT).c_str()); + gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_TENSORS_COUNT).c_str()); + + if (params->kv_overrides) { + const std::vector & overrides = *(const std::vector *)params->kv_overrides; + for (const auto & o : overrides) { + if (o.key[0] == 0) break; + if (o.tag == LLAMA_KV_OVERRIDE_TYPE_FLOAT) { + gguf_set_val_f32(ctx_out.get(), o.key, o.val_f64); + } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_INT) { + gguf_set_val_i32(ctx_out.get(), o.key, o.val_i64); + } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_BOOL) { + gguf_set_val_bool(ctx_out.get(), o.key, o.val_bool); + } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_STR) { + gguf_set_val_str(ctx_out.get(), o.key, o.val_str); + } else { + LLAMA_LOG_WARN("%s: unknown KV override type for key %s\n", __func__, o.key); + } + } + } + + // make a list of weights + std::vector tensors; + tensors.reserve(ml.weights_map.size()); + for (const auto & it : ml.weights_map) { + tensors.push_back(&it.second); + } + + // keep_split requires that the weights are sorted by split index + if (params->keep_split) { + std::sort(tensors.begin(), tensors.end(), [](const llama_model_loader::llama_tensor_weight * a, const llama_model_loader::llama_tensor_weight * b) { + if (a->idx == b->idx) { + return a->offs < b->offs; + } + return a->idx < b->idx; + }); + } + + for (const auto * it : tensors) { + const struct ggml_tensor * tensor = it->tensor; + + const std::string name = ggml_get_name(tensor); + + // TODO: avoid hardcoded tensor names - use the TN_* constants + if (name.find("attn_v.weight") != std::string::npos || + name.find("attn_qkv.weight") != std::string::npos || + name.find("attn_kv_b.weight")!= std::string::npos) { + ++qs.n_attention_wv; + } else if (name == LLM_TN(model.arch)(LLM_TENSOR_OUTPUT, "weight")) { + qs.has_output = true; + } + } + + qs.n_ffn_down = qs.n_ffn_gate = qs.n_ffn_up = (int)model.hparams.n_layer; + + // sanity checks for models that have attention layers + if (qs.n_attention_wv != 0) + { + const auto & n_head_kv_iter = model.hparams.n_head_kv_arr.begin(); + // attention layers have a non-zero number of kv heads + int32_t n_attn_layer = model.hparams.n_layer - std::count(n_head_kv_iter, n_head_kv_iter + model.hparams.n_layer, 0); + if (llama_model_has_encoder(&model)) { + n_attn_layer *= 3; + } + GGML_ASSERT((qs.n_attention_wv == n_attn_layer) && "n_attention_wv is unexpected"); + } + + size_t total_size_org = 0; + size_t total_size_new = 0; + + std::vector workers; + workers.reserve(nthread); + + int idx = 0; + + std::vector> read_data; + std::vector> work; + std::vector> f32_conv_buf; + + uint16_t n_split = 1; + + // Assume split index is continuous + if (params->keep_split) { + for (const auto * it : tensors) { + n_split = std::max(uint16_t(it->idx + 1), n_split); + } + } + std::vector ctx_outs(n_split); + ctx_outs[0] = std::move(ctx_out); + + // populate the original tensors so we get an initial meta data + for (const auto * it : tensors) { + uint16_t i_split = params->keep_split ? it->idx : 0; + struct ggml_tensor * tensor = it->tensor; + if (!ctx_outs[i_split]) { + ctx_outs[i_split].reset(gguf_init_empty()); + } + gguf_add_tensor(ctx_outs[i_split].get(), tensor); + } + + // Set split info if needed + if (n_split > 1) { + for (size_t i = 0; i < ctx_outs.size(); ++i) { + gguf_set_val_u16(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_NO).c_str(), i); + gguf_set_val_u16(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_COUNT).c_str(), n_split); + gguf_set_val_i32(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_TENSORS_COUNT).c_str(), ml.n_tensors); + } + } + + int cur_split = -1; + std::ofstream fout; + auto close_ofstream = [&]() { + // Write metadata and close file handler + if (fout.is_open()) { + fout.seekp(0); + std::vector data(gguf_get_meta_size(ctx_outs[cur_split].get())); + gguf_get_meta_data(ctx_outs[cur_split].get(), data.data()); + fout.write((const char *) data.data(), data.size()); + fout.close(); + } + }; + auto new_ofstream = [&](int index) { + cur_split = index; + GGML_ASSERT(ctx_outs[cur_split] && "Find uninitialized gguf_context"); + std::string fname = fname_out; + if (params->keep_split) { + std::vector split_path(llama_path_max(), 0); + llama_split_path(split_path.data(), split_path.size(), fname_out.c_str(), cur_split, n_split); + fname = std::string(split_path.data()); + } + + fout = std::ofstream(fname, std::ios::binary); + fout.exceptions(std::ofstream::failbit); // fail fast on write errors + const size_t meta_size = gguf_get_meta_size(ctx_outs[cur_split].get()); + // placeholder for the meta data + ::zeros(fout, meta_size); + }; + + const auto tn = LLM_TN(model.arch); + new_ofstream(0); + for (const auto * it : tensors) { + const auto & weight = *it; + struct ggml_tensor * tensor = weight.tensor; + if (weight.idx != cur_split && params->keep_split) { + close_ofstream(); + new_ofstream(weight.idx); + } + + const std::string name = ggml_get_name(tensor); + + if (!ml.use_mmap) { + if (read_data.size() < ggml_nbytes(tensor)) { + read_data.resize(ggml_nbytes(tensor)); + } + tensor->data = read_data.data(); + } + ml.load_data_for(tensor); + + LLAMA_LOG_INFO("[%4d/%4d] %36s - [%s], type = %6s, ", + ++idx, ml.n_tensors, + ggml_get_name(tensor), + llama_format_tensor_shape(tensor).c_str(), + ggml_type_name(tensor->type)); + + // This used to be a regex, but has an extreme cost to compile times. + bool quantize = name.rfind("weight") == name.size() - 6; // ends with 'weight'? + + // quantize only 2D and 3D tensors (experts) + quantize &= (ggml_n_dims(tensor) >= 2); + + // do not quantize norm tensors + quantize &= name.find("_norm.weight") == std::string::npos; + + quantize &= params->quantize_output_tensor || name != "output.weight"; + quantize &= !params->only_copy; + + // do not quantize expert gating tensors + // NOTE: can't use LLM_TN here because the layer number is not known + quantize &= name.find("ffn_gate_inp.weight") == std::string::npos; + + // do not quantize positional embeddings and token types (BERT) + quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_POS_EMBD, "weight"); + quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_TOKEN_TYPES, "weight"); + + // do not quantize Mamba's small yet 2D weights + // NOTE: can't use LLM_TN here because the layer number is not known + quantize &= name.find("ssm_conv1d.weight") == std::string::npos; + + // do not quantize RWKV's time_mix_first tensors + quantize &= name.find("time_mix_first.weight") == std::string::npos; + quantize &= name.find("time_mix_w1.weight") == std::string::npos; + quantize &= name.find("time_mix_w2.weight") == std::string::npos; + quantize &= name.find("time_mix_decay_w1.weight") == std::string::npos; + quantize &= name.find("time_mix_decay_w2.weight") == std::string::npos; + quantize &= name.find("time_mix_lerp_fused.weight") == std::string::npos; + + // do not quantize relative position bias (T5) + quantize &= name.find("attn_rel_b.weight") == std::string::npos; + + enum ggml_type new_type; + void * new_data; + size_t new_size; + + if (quantize) { + new_type = default_type; + + // get more optimal quantization type based on the tensor shape, layer, etc. + if (!params->pure && ggml_is_quantized(default_type)) { + new_type = llama_tensor_get_type(qs, new_type, tensor, ftype); + } + if (params->token_embedding_type < GGML_TYPE_COUNT && strcmp(tensor->name, "token_embd.weight") == 0) { + new_type = params->token_embedding_type; + } + if (params->output_tensor_type < GGML_TYPE_COUNT && strcmp(tensor->name, "output.weight") == 0) { + new_type = params->output_tensor_type; + } + + // If we've decided to quantize to the same type the tensor is already + // in then there's nothing to do. + quantize = tensor->type != new_type; + } + + if (!quantize) { + new_type = tensor->type; + new_data = tensor->data; + new_size = ggml_nbytes(tensor); + LLAMA_LOG_INFO("size = %8.3f MB\n", ggml_nbytes(tensor)/1024.0/1024.0); + } else { + const int64_t nelements = ggml_nelements(tensor); + + const float * imatrix = nullptr; + if (imatrix_data) { + auto it = imatrix_data->find(tensor->name); + if (it == imatrix_data->end()) { + LLAMA_LOG_INFO("\n====== %s: did not find weights for %s\n", __func__, tensor->name); + } else { + if (it->second.size() == (size_t)tensor->ne[0]*tensor->ne[2]) { + imatrix = it->second.data(); + } else { + LLAMA_LOG_INFO("\n====== %s: imatrix size %d is different from tensor size %d for %s\n", __func__, + int(it->second.size()), int(tensor->ne[0]*tensor->ne[2]), tensor->name); + + // this can happen when quantizing an old mixtral model with split tensors with a new incompatible imatrix + // this is a significant error and it may be good idea to abort the process if this happens, + // since many people will miss the error and not realize that most of the model is being quantized without an imatrix + // tok_embd should be ignored in this case, since it always causes this warning + if (name != tn(LLM_TENSOR_TOKEN_EMBD, "weight")) { + throw std::runtime_error(format("imatrix size %d is different from tensor size %d for %s", + int(it->second.size()), int(tensor->ne[0]*tensor->ne[2]), tensor->name)); + } + } + } + } + if ((new_type == GGML_TYPE_IQ2_XXS || + new_type == GGML_TYPE_IQ2_XS || + new_type == GGML_TYPE_IQ2_S || + new_type == GGML_TYPE_IQ1_S || + (new_type == GGML_TYPE_IQ1_M && strcmp(tensor->name, "token_embd.weight") && strcmp(tensor->name, "output.weight")) || + (new_type == GGML_TYPE_Q2_K && params->ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S && strcmp(tensor->name, "token_embd.weight") != 0)) && !imatrix) { + LLAMA_LOG_ERROR("\n\n============================================================\n"); + LLAMA_LOG_ERROR("Missing importance matrix for tensor %s in a very low-bit quantization\n", tensor->name); + LLAMA_LOG_ERROR("The result will be garbage, so bailing out\n"); + LLAMA_LOG_ERROR("============================================================\n\n"); + throw std::runtime_error(format("Missing importance matrix for tensor %s in a very low-bit quantization", tensor->name)); + } + + float * f32_data; + + if (tensor->type == GGML_TYPE_F32) { + f32_data = (float *) tensor->data; + } else if (ggml_is_quantized(tensor->type) && !params->allow_requantize) { + throw std::runtime_error(format("requantizing from type %s is disabled", ggml_type_name(tensor->type))); + } else { + llama_tensor_dequantize_impl(tensor, f32_conv_buf, workers, nelements, nthread); + f32_data = (float *) f32_conv_buf.data(); + } + + LLAMA_LOG_INFO("converting to %s .. ", ggml_type_name(new_type)); + fflush(stdout); + + if (work.size() < (size_t)nelements * 4) { + work.resize(nelements * 4); // upper bound on size + } + new_data = work.data(); + + const int64_t n_per_row = tensor->ne[0]; + const int64_t nrows = tensor->ne[1]; + + static const int64_t min_chunk_size = 32 * 512; + const int64_t chunk_size = (n_per_row >= min_chunk_size ? n_per_row : n_per_row * ((min_chunk_size + n_per_row - 1)/n_per_row)); + + const int64_t nelements_matrix = tensor->ne[0] * tensor->ne[1]; + const int64_t nchunk = (nelements_matrix + chunk_size - 1)/chunk_size; + const int64_t nthread_use = nthread > 1 ? std::max((int64_t)1, std::min((int64_t)nthread, nchunk)) : 1; + + // quantize each expert separately since they have different importance matrices + new_size = 0; + for (int64_t i03 = 0; i03 < tensor->ne[2]; ++i03) { + const float * f32_data_03 = f32_data + i03 * nelements_matrix; + void * new_data_03 = (char *)new_data + ggml_row_size(new_type, n_per_row) * i03 * nrows; + const float * imatrix_03 = imatrix ? imatrix + i03 * n_per_row : nullptr; + + new_size += llama_tensor_quantize_impl(new_type, f32_data_03, new_data_03, chunk_size, nrows, n_per_row, imatrix_03, workers, nthread_use); + } + LLAMA_LOG_INFO("size = %8.2f MiB -> %8.2f MiB\n", ggml_nbytes(tensor)/1024.0/1024.0, new_size/1024.0/1024.0); + } + total_size_org += ggml_nbytes(tensor); + total_size_new += new_size; + + // update the gguf meta data as we go + gguf_set_tensor_type(ctx_outs[cur_split].get(), name.c_str(), new_type); + GGML_ASSERT(gguf_get_tensor_size(ctx_outs[cur_split].get(), gguf_find_tensor(ctx_outs[cur_split].get(), name.c_str())) == new_size); + gguf_set_tensor_data(ctx_outs[cur_split].get(), name.c_str(), new_data); + + // write tensor data + padding + fout.write((const char *) new_data, new_size); + zeros(fout, GGML_PAD(new_size, align) - new_size); + } + close_ofstream(); + + LLAMA_LOG_INFO("%s: model size = %8.2f MB\n", __func__, total_size_org/1024.0/1024.0); + LLAMA_LOG_INFO("%s: quant size = %8.2f MB\n", __func__, total_size_new/1024.0/1024.0); + + if (qs.n_fallback > 0) { + LLAMA_LOG_WARN("%s: WARNING: %d of %d tensor(s) required fallback quantization\n", + __func__, qs.n_fallback, qs.n_k_quantized + qs.n_fallback); + } +} + +// +// interface implementation +// + +struct llama_model_quantize_params llama_model_quantize_default_params() { + struct llama_model_quantize_params result = { + /*.nthread =*/ 0, + /*.ftype =*/ LLAMA_FTYPE_MOSTLY_Q5_1, + /*.output_tensor_type =*/ GGML_TYPE_COUNT, + /*.token_embedding_type =*/ GGML_TYPE_COUNT, + /*.allow_requantize =*/ false, + /*.quantize_output_tensor =*/ true, + /*.only_copy =*/ false, + /*.pure =*/ false, + /*.keep_split =*/ false, + /*.imatrix =*/ nullptr, + /*.kv_overrides =*/ nullptr, + }; + + return result; +} + +uint32_t llama_model_quantize( + const char * fname_inp, + const char * fname_out, + const llama_model_quantize_params * params) { + try { + llama_model_quantize_impl(fname_inp, fname_out, params); + } catch (const std::exception & err) { + LLAMA_LOG_ERROR("%s: failed to quantize: %s\n", __func__, err.what()); + return 1; + } + + return 0; +} diff --git a/src/llama-quant.h b/src/llama-quant.h new file mode 100644 index 000000000..6f70f09be --- /dev/null +++ b/src/llama-quant.h @@ -0,0 +1 @@ +#pragma once diff --git a/src/llama-sampling.cpp b/src/llama-sampling.cpp index fd8ca8a9e..b3a12386e 100644 --- a/src/llama-sampling.cpp +++ b/src/llama-sampling.cpp @@ -1,5 +1,6 @@ #include "llama-sampling.h" +#include "llama-impl.h" #include "llama-vocab.h" #include "llama-grammar.h" @@ -14,6 +15,118 @@ #include #include #include +#include + +// the ring buffer works similarly to std::deque, but with a fixed capacity +template +struct ring_buffer { + ring_buffer(size_t cap) : capacity(cap), data(cap) {} + + T & front() { + if (sz == 0) { + throw std::runtime_error("ring buffer is empty"); + } + return data[first]; + } + + const T & front() const { + if (sz == 0) { + throw std::runtime_error("ring buffer is empty"); + } + return data[first]; + } + + T & back() { + if (sz == 0) { + throw std::runtime_error("ring buffer is empty"); + } + return data[pos]; + } + + const T & back() const { + if (sz == 0) { + throw std::runtime_error("ring buffer is empty"); + } + return data[pos]; + } + + void push_back(const T & value) { + if (capacity == 0) { + throw std::runtime_error("ring buffer: capacity is zero"); + } + + if (sz == capacity) { + // advance the start when buffer is full + first = (first + 1) % capacity; + } else { + sz++; + } + data[pos] = value; + pos = (pos + 1) % capacity; + } + + T pop_front() { + if (sz == 0) { + throw std::runtime_error("ring buffer is empty"); + } + T value = data[first]; + first = (first + 1) % capacity; + sz--; + return value; + } + + //T & operator[](size_t i) { + // if (i >= sz) { + // throw std::runtime_error("ring buffer: index out of bounds"); + // } + // return data[(first + i) % capacity]; + //} + + //const T & at(size_t i) const { + // if (i >= sz) { + // throw std::runtime_error("ring buffer: index out of bounds"); + // } + // return data[(first + i) % capacity]; + //} + + const T & rat(size_t i) const { + if (i >= sz) { + throw std::runtime_error("ring buffer: index out of bounds"); + } + return data[(first + sz - i - 1) % capacity]; + } + + std::vector to_vector() const { + std::vector result; + result.reserve(sz); + for (size_t i = 0; i < sz; i++) { + result.push_back(data[(first + i) % capacity]); + } + return result; + } + + void clear() { + // here only reset the status of the buffer + sz = 0; + first = 0; + pos = 0; + } + + bool empty() const { + return sz == 0; + } + + size_t size() const { + return sz; + } + + size_t capacity = 0; + size_t sz = 0; + size_t first = 0; + size_t pos = 0; + + std::vector data; +}; static int llama_sample_dist(llama_token_data_array * cur_p, std::mt19937 & rng) { // iterator for the probabilities @@ -144,7 +257,7 @@ static void llama_sampler_top_k_impl(llama_token_data_array * cur_p, int32_t k) for (int i = 0; i < (int)cur_p->size; ++i) { const float val = cur_p->data[i].logit; int ib = int(bucket_scale * val + bucket_inter); //nbuckets * (val - bucket_low) / (bucket_high - bucket_low); - ib = std::max(0, std::min(nbuckets-1, ib)); + ib = std::max(0, std::min(nbuckets - 1, ib)); bucket_idx[i] = ib; ++histo[ib]; } @@ -167,13 +280,13 @@ static void llama_sampler_top_k_impl(llama_token_data_array * cur_p, int32_t k) for (int i = 0; i < (int)cur_p->size; ++i) { int j = bucket_idx[i]; if (j >= ib) { - *bucket_ptrs[nbuckets-1-j]++ = cur_p->data[i]; + *bucket_ptrs[nbuckets - 1 - j]++ = cur_p->data[i]; } } ptr = tmp_tokens.data(); int ndone = 0; - for (int j = nbuckets-1; j > ib; --j) { + for (int j = nbuckets - 1; j > ib; --j) { std::sort(ptr, ptr + histo[j], comp); ptr += histo[j]; ndone += histo[j]; @@ -258,7 +371,10 @@ void llama_sampler_free(struct llama_sampler * smpl) { llama_token llama_sampler_sample(struct llama_sampler * smpl, struct llama_context * ctx, int32_t idx) { const auto * logits = llama_get_logits_ith(ctx, idx); - const int n_vocab = llama_n_vocab(llama_get_model(ctx)); + const llama_model * model = llama_get_model(ctx); + const llama_vocab * vocab = llama_model_get_vocab(model); + + const int n_vocab = llama_vocab_n_tokens(vocab); // TODO: do not allocate each time std::vector cur; @@ -1332,7 +1448,7 @@ static void llama_sampler_grammar_reset(struct llama_sampler * smpl) { static struct llama_sampler * llama_sampler_grammar_clone(const struct llama_sampler * smpl) { const auto * ctx = (const llama_sampler_grammar *) smpl->ctx; - auto * result = llama_sampler_init_grammar_impl(*ctx->vocab, nullptr, nullptr); + auto * result = llama_sampler_init_grammar(ctx->vocab, nullptr, nullptr); // copy the state { @@ -1368,19 +1484,19 @@ static struct llama_sampler_i llama_sampler_grammar_i = { /* .free = */ llama_sampler_grammar_free, }; -struct llama_sampler * llama_sampler_init_grammar_impl(const struct llama_vocab & vocab, const char * grammar_str, const char * grammar_root) { +struct llama_sampler * llama_sampler_init_grammar(const struct llama_vocab * vocab, const char * grammar_str, const char * grammar_root) { auto * ctx = new llama_sampler_grammar; if (grammar_str != nullptr && grammar_str[0] != '\0') { *ctx = { - /* .vocab = */ &vocab, + /* .vocab = */ vocab, /* .grammar_str = */ grammar_str, /* .grammar_root = */ grammar_root, - /* .grammar = */ llama_grammar_init_impl(&vocab, grammar_str, grammar_root), + /* .grammar = */ llama_grammar_init_impl(vocab, grammar_str, grammar_root), }; } else { *ctx = { - /* .vocab = */ &vocab, + /* .vocab = */ vocab, /* .grammar_str = */ {}, /* .grammar_root = */ {}, /* .grammar = */ nullptr, @@ -1396,19 +1512,15 @@ struct llama_sampler * llama_sampler_init_grammar_impl(const struct llama_vocab // penalties struct llama_sampler_penalties { - const int32_t n_vocab; - const llama_token special_eos_id; - const llama_token linefeed_id; - const int32_t penalty_last_n; const float penalty_repeat; const float penalty_freq; const float penalty_present; - const bool penalize_nl; - const bool ignore_eos; - ring_buffer prev; + + // a frequency map to count token occurrences + std::unordered_map token_count; }; static const char * llama_sampler_penalties_name(const struct llama_sampler * /*smpl*/) { @@ -1421,76 +1533,50 @@ static void llama_sampler_penalties_accept(struct llama_sampler * smpl, llama_to return; } + ctx->token_count[token]++; + + // if the ring buffer is full, remove the oldest token + if (ctx->prev.size() >= (size_t) ctx->penalty_last_n) { + const auto old = ctx->prev.front(); + + ctx->token_count[old]--; + if (ctx->token_count[old] == 0) { + ctx->token_count.erase(old); + } + } + ctx->prev.push_back(token); + +#if 0 + // sanity check + std::unordered_map tmp; + for (int i = 0; i < std::min(ctx->penalty_last_n, ctx->prev.size()); ++i) { + tmp[ctx->prev.rat(i)]++; + } + + assert(ctx->token_count == tmp); +#endif } static void llama_sampler_penalties_apply(struct llama_sampler * smpl, llama_token_data_array * cur_p) { auto * ctx = (llama_sampler_penalties *) smpl->ctx; - if (ctx->ignore_eos) { - assert(ctx->special_eos_id >= 0); - - // optimistically check if the candidates are not yet sorted/shuffled/truncated - if (cur_p->size > (size_t) ctx->special_eos_id && cur_p->data[ctx->special_eos_id].id == ctx->special_eos_id) { - cur_p->data[ctx->special_eos_id].logit = -INFINITY; - } else { - // else, search for the special EOS token - for (size_t i = 0; i < cur_p->size; ++i) { - if (cur_p->data[i].id == ctx->special_eos_id) { - cur_p->data[i].logit = -INFINITY; - break; - } - } - } - } - if ((ctx->penalty_last_n == 0) || (ctx->penalty_repeat == 1.0f && ctx->penalty_freq == 0.0f && ctx->penalty_present == 0.0f)) { return; } - bool nl_found = false; - size_t nl_idx = 0; - float nl_logit = -INFINITY; - if (!ctx->penalize_nl) { - assert(ctx->linefeed_id >= 0); - - // optimistically check if the candidates are not yet sorted/shuffled/truncated - if (cur_p->size > (size_t) ctx->linefeed_id && cur_p->data[ctx->linefeed_id].id == ctx->linefeed_id) { - nl_found = true; - nl_idx = ctx->linefeed_id; - nl_logit = cur_p->data[ctx->linefeed_id].logit; - } else { - // else, search for the linefeed token - for (size_t i = 0; i < cur_p->size; ++i) { - if (cur_p->data[i].id == ctx->linefeed_id) { - nl_found = true; - nl_idx = i; - nl_logit = cur_p->data[i].logit; - break; - } - } - } - } - - // Create a frequency map to count occurrences of each token in last_tokens - // TODO: optimize this by maintaining the token count in the sampler context - using llama_token_cnt = std::unordered_map; - llama_token_cnt token_count; - - for (int i = 0; i < std::min(ctx->penalty_last_n, ctx->prev.size()); ++i) { - token_count[ctx->prev.rat(i)]++; - } - // Apply frequency and presence penalties to the cur_p for (size_t i = 0; i < cur_p->size; ++i) { - const auto token_iter = token_count.find(cur_p->data[i].id); - if (token_iter == token_count.end()) { + const auto token_iter = ctx->token_count.find(cur_p->data[i].id); + if (token_iter == ctx->token_count.end()) { continue; } const int count = token_iter->second; + assert(count > 0 && count <= ctx->penalty_last_n); + // The academic publication that described this technique actually just only divided, but that would cause tokens with negative logits to become more likely, which is obviously wrong. // This is common fix for this problem, which is to multiply by the penalty instead of dividing. if (cur_p->data[i].logit <= 0) { @@ -1503,30 +1589,21 @@ static void llama_sampler_penalties_apply(struct llama_sampler * smpl, llama_tok } cur_p->sorted = false; - - if (!ctx->penalize_nl && nl_found) { - // restore the logit of the newline token if it was penalized - cur_p->data[nl_idx].logit = nl_logit; - } } static void llama_sampler_penalties_reset(struct llama_sampler * smpl) { auto * ctx = (llama_sampler_penalties *) smpl->ctx; ctx->prev.clear(); + ctx->token_count.clear(); } static struct llama_sampler * llama_sampler_penalties_clone(const struct llama_sampler * smpl) { const auto * ctx = (const llama_sampler_penalties *) smpl->ctx; auto * result = llama_sampler_init_penalties( - ctx->n_vocab, - ctx->special_eos_id, - ctx->linefeed_id, ctx->penalty_last_n, ctx->penalty_repeat, ctx->penalty_freq, - ctx->penalty_present, - ctx->penalize_nl, - ctx->ignore_eos); + ctx->penalty_present); // copy the state { @@ -1552,38 +1629,21 @@ static struct llama_sampler_i llama_sampler_penalties_i = { }; struct llama_sampler * llama_sampler_init_penalties( - int32_t n_vocab, - llama_token special_eos_id, - llama_token linefeed_id, int32_t penalty_last_n, float penalty_repeat, float penalty_freq, - float penalty_present, - bool penalize_nl, - bool ignore_eos) { - if (linefeed_id == LLAMA_TOKEN_NULL) { - penalize_nl = true; - } - - if (special_eos_id == LLAMA_TOKEN_NULL) { - ignore_eos = false; - } - + float penalty_present) { penalty_last_n = std::max(penalty_last_n, 0); return new llama_sampler { /* .iface = */ &llama_sampler_penalties_i, /* .ctx = */ new llama_sampler_penalties { - /* .n_vocab = */ n_vocab, - /* .special_eos_id = */ special_eos_id, - /* .linefeed_id = */ linefeed_id, /* .penalty_last_n = */ penalty_last_n, /* .penalty_repeat = */ penalty_repeat, /* .penalty_freq = */ penalty_freq, /* .penalty_present = */ penalty_present, - /* .penalize_nl = */ penalize_nl, - /* .ignore_eos = */ ignore_eos, /* .prev = */ ring_buffer(penalty_last_n), + /* .token_count = */ {}, }, }; } @@ -1606,12 +1666,13 @@ struct llama_sampler_dry { // Ported from Koboldcpp, original PR: https://github.com/LostRuins/koboldcpp/pull/982 (Original author: pi6am) static void get_overlapping_token_sequences(const llama_vocab & vocab, const std::string& str, std::unordered_multimap>& token_sequences, int max_tail_len = -1) { - for (llama_token token_id = 0; token_id < (llama_token)vocab.n_vocab; token_id++) { - std::string word = llama_detokenize(vocab, {token_id}, true); + for (llama_token token_id = 0; token_id < (llama_token) vocab.n_tokens(); token_id++) { + std::string word = vocab.detokenize({token_id}, true); if (word.find(str) != std::string::npos) { token_sequences.emplace(token_id, std::vector()); } else { - size_t word_len = word.size(), str_len = str.size(); + size_t word_len = word.size(); + size_t str_len = str.size(); size_t pos = -1; while ((pos = word.find(str[0], pos + 1)) != std::string::npos) { bool match = true; @@ -1623,7 +1684,7 @@ static void get_overlapping_token_sequences(const llama_vocab & vocab, const std } } if (match) { - std::vector tokenization = llama_tokenize_internal(vocab, str.substr(i), false, false); + std::vector tokenization = vocab.tokenize(str.substr(i), false, false); if (max_tail_len >= 0 && tokenization.size() > (size_t)max_tail_len) { tokenization.resize(max_tail_len); } @@ -1774,7 +1835,7 @@ static void llama_sampler_dry_apply(struct llama_sampler * smpl, llama_token_dat ctx->dry_repeat_count[last - k] = std::min(n, rep_limit); if (n > 0) { lt = k; - rt = k+n-1; + rt = k + n - 1; } } else { // If k is inside the current Z-box, consider two cases. @@ -1879,7 +1940,7 @@ static struct llama_sampler * llama_sampler_dry_clone(const struct llama_sampler llama_vocab dummy_vocab; // dummy vocab is passed because it is only needed for raw sequence breaker processing, which we have already done and will simply be copying - auto * result = llama_sampler_init_dry_impl(dummy_vocab, ctx->total_context_size, ctx->dry_multiplier, ctx->dry_base, ctx->dry_allowed_length, ctx->dry_penalty_last_n, NULL, 0); + auto * result = llama_sampler_init_dry(&dummy_vocab, ctx->total_context_size, ctx->dry_multiplier, ctx->dry_base, ctx->dry_allowed_length, ctx->dry_penalty_last_n, NULL, 0); // Copy the state, including the processed breakers { @@ -1906,7 +1967,7 @@ static struct llama_sampler_i llama_sampler_dry_i = { /* .free = */ llama_sampler_dry_free, }; -struct llama_sampler * llama_sampler_init_dry_impl(const struct llama_vocab & vocab, int32_t context_size, float dry_multiplier, float dry_base, int32_t dry_allowed_length, int32_t dry_penalty_last_n, const char** seq_breakers, size_t num_breakers) { +struct llama_sampler * llama_sampler_init_dry(const struct llama_vocab * vocab, int32_t context_size, float dry_multiplier, float dry_base, int32_t dry_allowed_length, int32_t dry_penalty_last_n, const char** seq_breakers, size_t num_breakers) { int32_t effective_dry_penalty_last_n = (dry_penalty_last_n == -1) ? context_size : std::max(dry_penalty_last_n, 0); std::unordered_multimap> processed_breakers; const int MAX_CHAR_LEN = 40; @@ -1933,7 +1994,7 @@ struct llama_sampler * llama_sampler_init_dry_impl(const struct llama_vocab & vo sequence_break.resize(MAX_CHAR_LEN); } - get_overlapping_token_sequences(vocab, sequence_break, processed_breakers, MAX_SEQ_LEN); + get_overlapping_token_sequences(*vocab, sequence_break, processed_breakers, MAX_SEQ_LEN); } } @@ -1956,7 +2017,7 @@ struct llama_sampler * llama_sampler_init_dry_impl(const struct llama_vocab & vo // wrapper for test-sampling.cpp struct llama_sampler * llama_sampler_init_dry_testing(int32_t context_size, float dry_multiplier, float dry_base, int32_t dry_allowed_length, int32_t dry_penalty_last_n, const std::vector>& seq_breakers) { llama_vocab dummy_vocab; - auto * result = llama_sampler_init_dry_impl(dummy_vocab, context_size, dry_multiplier, dry_base, dry_allowed_length, dry_penalty_last_n, NULL, 0); + auto * result = llama_sampler_init_dry(&dummy_vocab, context_size, dry_multiplier, dry_base, dry_allowed_length, dry_penalty_last_n, NULL, 0); auto * ctx = (llama_sampler_dry *) result->ctx; // Process the token-based sequence breakers @@ -2095,7 +2156,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ float p_eog_sum = 0.0f; for (size_t i = 0; i < cur_p->size; ++i) { - if (llama_token_is_eog_impl(*ctx->vocab, cur_p->data[i].id)) { + if (ctx->vocab->is_eog(cur_p->data[i].id)) { p_eog_sum += cur_p->data[i].p; } else { p_txt_sum += cur_p->data[i].p; @@ -2117,7 +2178,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ float p_sum = 0.0f; for (size_t i = 0; i < size_org; ++i) { - if (llama_token_is_eog_impl(*ctx->vocab, cur_p->data[i].id)) { + if (ctx->vocab->is_eog(cur_p->data[i].id)) { p_sum += cur_p->data[i].p; cur_p->data[cur_p->size++] = cur_p->data[i]; @@ -2145,17 +2206,17 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ continue; } - int len0 = llama_token_to_piece_impl(*ctx->vocab, cur_p->data[i0].id, ctx->buf0.data(), ctx->buf0.size(), 0, false); + int len0 = ctx->vocab->token_to_piece(cur_p->data[i0].id, ctx->buf0.data(), ctx->buf0.size(), 0, false); if (len0 < 0) { ctx->buf0.resize(len0); - len0 = llama_token_to_piece_impl(*ctx->vocab, cur_p->data[i0].id, ctx->buf0.data(), ctx->buf0.size(), 0, false); + len0 = ctx->vocab->token_to_piece(cur_p->data[i0].id, ctx->buf0.data(), ctx->buf0.size(), 0, false); assert(len0 > 0); } - int len1 = llama_token_to_piece_impl(*ctx->vocab, cur_p->data[i1].id, ctx->buf1.data(), ctx->buf1.size(), 0, false); + int len1 = ctx->vocab->token_to_piece(cur_p->data[i1].id, ctx->buf1.data(), ctx->buf1.size(), 0, false); if (len1 < 0) { ctx->buf1.resize(len1); - len1 = llama_token_to_piece_impl(*ctx->vocab, cur_p->data[i1].id, ctx->buf1.data(), ctx->buf1.size(), 0, false); + len1 = ctx->vocab->token_to_piece(cur_p->data[i1].id, ctx->buf1.data(), ctx->buf1.size(), 0, false); assert(len1 > 0); } @@ -2190,7 +2251,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ LOG_DBG_CUR("%s: n_combined = %zu, applying thold = %.3f\n", __func__, n_combined, thold); for (size_t i = 0; i < size_org; ++i) { - const bool is_eog = llama_token_is_eog_impl(*ctx->vocab, cur_p->data[i].id); + const bool is_eog = ctx->vocab->is_eog(cur_p->data[i].id); if (cur_p->data[i].p < thold && !is_eog) { continue; @@ -2211,7 +2272,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ // if no non-EOG tokens are left -> reduce cur_p to single EOT token if (n_non_eog == 0) { cur_p->size = 1; - cur_p->data[0].id = llama_token_eot_impl(*ctx->vocab); + cur_p->data[0].id = ctx->vocab->token_eot(); cur_p->data[0].logit = 1.0f; return; @@ -2233,7 +2294,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ LOG_DBG_CUR("%s: applying thold = %.3f\n", __func__, thold); for (size_t i = 0; i < size_org; ++i) { - const bool is_eog = llama_token_is_eog_impl(*ctx->vocab, cur_p->data[i].id); + const bool is_eog = ctx->vocab->is_eog(cur_p->data[i].id); if (cur_p->data[i].p < thold && !is_eog) { continue; @@ -2256,7 +2317,7 @@ static void llama_sampler_infill_apply(struct llama_sampler * smpl, llama_token_ static struct llama_sampler * llama_sampler_infill_clone(const struct llama_sampler * smpl) { const auto * ctx = (const llama_sampler_infill *) smpl->ctx; - return llama_sampler_init_infill_impl(*ctx->vocab); + return llama_sampler_init_infill(ctx->vocab); } static void llama_sampler_infill_free(struct llama_sampler * smpl) { @@ -2272,14 +2333,13 @@ static struct llama_sampler_i llama_sampler_infill_i = { /* .free = */ llama_sampler_infill_free, }; -struct llama_sampler * llama_sampler_init_infill_impl( - const struct llama_vocab & vocab) { +struct llama_sampler * llama_sampler_init_infill(const struct llama_vocab * vocab) { return new llama_sampler { /* .iface = */ &llama_sampler_infill_i, /* .ctx = */ new llama_sampler_infill { - /* .vocab = */ &vocab, - /* .buf0 = */ std::vector(512), - /* .buf1 = */ std::vector(512), + /* .vocab = */ vocab, + /* .buf0 = */ std::vector(512), + /* .buf1 = */ std::vector(512), }, }; } diff --git a/src/llama-sampling.h b/src/llama-sampling.h index 919f6fdfc..759dd7dcb 100644 --- a/src/llama-sampling.h +++ b/src/llama-sampling.h @@ -2,7 +2,9 @@ // TODO: rename llama-sampling.h/.cpp to llama-sampler.h/.cpp ? -#include "llama-grammar.h" +#include "llama.h" + +#include struct llama_vocab; struct llama_grammar; @@ -21,24 +23,6 @@ struct llama_sampler_chain { mutable int32_t n_sample; }; -struct llama_sampler * llama_sampler_init_grammar_impl( - const struct llama_vocab & vocab, - const char * grammar_str, - const char * grammar_root); - -struct llama_sampler * llama_sampler_init_infill_impl( - const struct llama_vocab & vocab); - -struct llama_sampler * llama_sampler_init_dry_impl( - const struct llama_vocab & vocab, - int32_t context_size, - float dry_multiplier, - float dry_base, - int32_t dry_allowed_length, - int32_t dry_penalty_last_n, - const char ** seq_breakers, - size_t num_breakers); - struct llama_sampler * llama_sampler_init_dry_testing( int32_t context_size, float dry_multiplier, diff --git a/src/llama-vocab.cpp b/src/llama-vocab.cpp index d1dc96276..4969d2628 100644 --- a/src/llama-vocab.cpp +++ b/src/llama-vocab.cpp @@ -1,5 +1,8 @@ #include "llama-vocab.h" +#include "llama-impl.h" +#include "llama-model-loader.h" + #include "unicode.h" #include @@ -9,29 +12,15 @@ #include #include #include +#include #include -#include +#include +#include // // helpers // -LLAMA_ATTRIBUTE_FORMAT(1, 2) -static std::string format(const char * fmt, ...) { - va_list ap; - va_list ap2; - va_start(ap, fmt); - va_copy(ap2, ap); - int size = vsnprintf(NULL, 0, fmt, ap); - GGML_ASSERT(size >= 0 && size < INT_MAX); // NOLINT - std::vector buf(size + 1); - int size2 = vsnprintf(buf.data(), size + 1, fmt, ap2); - GGML_ASSERT(size2 == size); - va_end(ap2); - va_end(ap); - return std::string(buf.data(), size); -} - struct naive_trie { naive_trie() : has_value(false), value(0) { } @@ -76,96 +65,14 @@ struct naive_trie { }; // -// impl +// tokenizers // struct llm_tokenizer { - llm_tokenizer() {} - virtual ~llm_tokenizer() = default; + llm_tokenizer() {} + virtual ~llm_tokenizer() = default; }; -llama_vocab::~llama_vocab() { - delete tokenizer; -} - -int llama_vocab::find_bpe_rank(const std::string & token_left, const std::string & token_right) const { - GGML_ASSERT(token_left.find(' ') == std::string::npos); - GGML_ASSERT(token_left.find('\n') == std::string::npos); - GGML_ASSERT(token_right.find(' ') == std::string::npos); - GGML_ASSERT(token_right.find('\n') == std::string::npos); - - auto it = bpe_ranks.find(std::make_pair(token_left, token_right)); - if (it == bpe_ranks.end()) { - return -1; - } - - return it->second; -} - -static enum llama_vocab_type llama_vocab_get_type(const llama_vocab & vocab) { - return vocab.type; -} - -static bool llama_is_normal_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_NORMAL; -} - -static bool llama_is_unknown_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_UNKNOWN; -} - -static bool llama_is_control_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_CONTROL; -} - -static bool llama_is_byte_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_BYTE; -} - -static bool llama_is_user_defined_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_USER_DEFINED; -} - -static bool llama_is_unused_token(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE); - return vocab.id_to_token[id].attr & LLAMA_TOKEN_ATTR_UNUSED; -} - -static uint8_t llama_token_to_byte(const llama_vocab & vocab, llama_token id) { - GGML_ASSERT(llama_vocab_get_type(vocab) != LLAMA_VOCAB_TYPE_NONE); - GGML_ASSERT(llama_is_byte_token(vocab, id)); - const auto & token_data = vocab.id_to_token.at(id); - switch (llama_vocab_get_type(vocab)) { - case LLAMA_VOCAB_TYPE_SPM: - case LLAMA_VOCAB_TYPE_UGM: { - auto buf = token_data.text.substr(3, 2); - return strtol(buf.c_str(), NULL, 16); - } - case LLAMA_VOCAB_TYPE_BPE: { - GGML_ABORT("fatal error"); - //return unicode_utf8_to_byte(token_data.text); // TODO: why is this here after GGML_ASSERT? - } - case LLAMA_VOCAB_TYPE_WPM: { - GGML_ABORT("fatal error"); - } - default: - GGML_ABORT("fatal error"); - } -} - -static void llama_escape_whitespace(std::string & text) { - replace_all(text, " ", "\xe2\x96\x81"); -} - -static void llama_unescape_whitespace(std::string & word) { - replace_all(word, "\xe2\x96\x81", " "); -} - struct llm_symbol { using index = int; index prev; @@ -197,14 +104,13 @@ struct llm_bigram_spm { }; struct llm_tokenizer_spm : llm_tokenizer { - llm_tokenizer_spm(const llama_vocab & /*vocab*/) : llm_tokenizer() {} + llm_tokenizer_spm(const llama_vocab & /*vocab*/) {} }; struct llm_tokenizer_spm_session { llm_tokenizer_spm_session(const llama_vocab & vocab) : vocab(vocab) {} - void tokenize(const std::string & text, std::vector & output) { - + void tokenize(const std::string & text, std::vector & output) { // split string into utf8 chars int index = 0; size_t offs = 0; @@ -263,13 +169,13 @@ struct llm_tokenizer_spm_session { } private: - void resegment(llm_symbol & symbol, std::vector & output) { + void resegment(llm_symbol & symbol, std::vector & output) { auto text = std::string(symbol.text, symbol.n); - auto token = vocab.token_to_id.find(text); + auto token = vocab.text_to_token(text); // Do we need to support is_unused? - if (token != vocab.token_to_id.end()) { - output.push_back((*token).second); + if (token != LLAMA_TOKEN_NULL) { + output.push_back(token); return; } @@ -279,8 +185,8 @@ private: // output any symbols that did not form tokens as bytes. output.reserve(output.size() + symbol.n); for (int j = 0; j < (int)symbol.n; ++j) { - llama_vocab::id token_id = llama_byte_to_token_impl(vocab, symbol.text[j]); - output.push_back(token_id); + llama_token id = vocab.byte_to_token(symbol.text[j]); + output.push_back(id); } return; } @@ -294,17 +200,17 @@ private: return; } const std::string text = std::string(symbols[left].text, symbols[left].n + symbols[right].n); - auto token = vocab.token_to_id.find(text); + auto token = vocab.text_to_token(text); - if (token == vocab.token_to_id.end()) { + if (token == LLAMA_TOKEN_NULL) { return; } - if (static_cast((*token).second) >= vocab.id_to_token.size()) { + if (static_cast(token) >= vocab.n_tokens()) { return; } - const auto & tok_data = vocab.id_to_token[(*token).second]; + const auto & tok_data = vocab.get_token_data(token); llm_bigram_spm bigram; bigram.left = left; @@ -367,9 +273,9 @@ struct llm_bigram_bpe { }; struct llm_tokenizer_bpe : llm_tokenizer { - llm_tokenizer_bpe(const llama_vocab & vocab) : llm_tokenizer() { - GGML_ASSERT(vocab.type == LLAMA_VOCAB_TYPE_BPE); - switch (vocab.type_pre) { + llm_tokenizer_bpe(const llama_vocab & vocab) { + GGML_ASSERT(vocab.get_type() == LLAMA_VOCAB_TYPE_BPE); + switch (vocab.get_pre_type()) { case LLAMA_VOCAB_PRE_TYPE_LLAMA3: regex_exprs = { // original regex from tokenizer.json @@ -396,6 +302,13 @@ struct llm_tokenizer_bpe : llm_tokenizer { "\\p{N}+", }; break; + case LLAMA_VOCAB_PRE_TYPE_DEEPSEEK3_LLM: + regex_exprs = { + "\\p{N}{1,3}", + "[一-龥぀-ゟ゠-ヿ]+", + "[!\"#$%&'()*+,\\-./:;<=>?@\\[\\\\\\]^_`{|}~][A-Za-z]+|[^\r\n\\p{L}\\p{P}\\p{S}]?[\\p{L}\\p{M}]+| ?[\\p{P}\\p{S}]+[\r\n]*|\\s*[\r\n]+|\\s+(?!\\S)|\\s+", + }; + break; case LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER: regex_exprs = { "[\r\n]", @@ -418,6 +331,7 @@ struct llm_tokenizer_bpe : llm_tokenizer { case LLAMA_VOCAB_PRE_TYPE_SMOLLM: case LLAMA_VOCAB_PRE_TYPE_CODESHELL: case LLAMA_VOCAB_PRE_TYPE_EXAONE: + case LLAMA_VOCAB_PRE_TYPE_MINERVA: regex_exprs = { "\\p{N}", "'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)", @@ -494,39 +408,38 @@ struct llm_tokenizer_bpe : llm_tokenizer { }; struct llm_tokenizer_bpe_session { - llm_tokenizer_bpe_session(const llama_vocab & vocab) : vocab(vocab), - bpe_tokenizer(static_cast(vocab.tokenizer)) {} + llm_tokenizer_bpe_session(const llama_vocab & vocab, const llm_tokenizer_bpe & tokenizer) : vocab(vocab), tokenizer(tokenizer) {} - static void append(const llama_vocab::id token_id, std::vector & output) { + static void append(const llama_token token_id, std::vector & output) { output.push_back(token_id); } - bool append_bos(std::vector & output) const { - if (vocab.tokenizer_add_bos) { - GGML_ASSERT(vocab.special_bos_id != -1); - output.push_back(vocab.special_bos_id); + bool append_bos(std::vector & output) const { + if (vocab.get_add_bos()) { + GGML_ASSERT(vocab.token_bos() != LLAMA_TOKEN_NULL); + output.push_back(vocab.token_bos()); return true; } return false; } - bool append_eos(std::vector & output) const { - if (vocab.tokenizer_add_eos) { - GGML_ASSERT(vocab.special_eos_id != -1); - output.push_back(vocab.special_eos_id); + bool append_eos(std::vector & output) const { + if (vocab.get_add_eos()) { + GGML_ASSERT(vocab.token_eos() != LLAMA_TOKEN_NULL); + output.push_back(vocab.token_eos()); return true; } return false; } - void check_double_bos_eos(const std::vector & output) const { - if (vocab.tokenizer_add_bos && output.size() >= 2 && output[1] == vocab.special_bos_id) { + void check_double_bos_eos(const std::vector & output) const { + if (vocab.get_add_bos() && output.size() >= 2 && output[1] == vocab.token_bos()) { LLAMA_LOG_WARN( "%s: Added a BOS token to the prompt as specified by the model but the prompt " "also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. " "Are you sure this is what you want?\n", __FUNCTION__); } - if (vocab.tokenizer_add_eos && output.size() >= 2 && *(output.end()-2) == vocab.special_eos_id) { + if (vocab.get_add_bos() && output.size() >= 2 && *(output.end()-2) == vocab.token_eos()) { LLAMA_LOG_WARN( "%s: Added a EOS token to the prompt as specified by the model but the prompt " "also ends with a EOS token. So now the final prompt ends with 2 EOS tokens. " @@ -534,9 +447,9 @@ struct llm_tokenizer_bpe_session { } } - void tokenize(const std::string & text, std::vector & output) { + void tokenize(const std::string & text, std::vector & output) { int final_prev_index = -1; - const auto word_collection = unicode_regex_split(text, bpe_tokenizer->regex_exprs); + const auto word_collection = unicode_regex_split(text, tokenizer.regex_exprs); symbols_final.clear(); @@ -547,7 +460,8 @@ struct llm_tokenizer_bpe_session { int index = 0; size_t offset = 0; - if (vocab.tokenizer_ignore_merges && vocab.token_to_id.find(word) != vocab.token_to_id.end()) { + //if (vocab.tokenizer_ignore_merges && vocab.token_to_id.find(word) != vocab.token_to_id.end()) { + if (vocab.get_ignore_merges() && vocab.text_to_token(word) != LLAMA_TOKEN_NULL) { symbols.emplace_back(llm_symbol{-1, -1, word.c_str(), word.size()}); offset = word.size(); } @@ -621,18 +535,18 @@ struct llm_tokenizer_bpe_session { } const std::string str = std::string(symbol.text, symbol.n); - const auto token = vocab.token_to_id.find(str); + const auto token = vocab.text_to_token(str); - if (token == vocab.token_to_id.end()) { + if (token == LLAMA_TOKEN_NULL) { for (auto j = str.begin(); j != str.end(); ++j) { std::string byte_str(1, *j); - auto token_multibyte = vocab.token_to_id.find(byte_str); - if (token_multibyte != vocab.token_to_id.end()) { - output.push_back(token_multibyte->second); + auto token_multibyte = vocab.text_to_token(byte_str); + if (token_multibyte != LLAMA_TOKEN_NULL) { + output.push_back(token_multibyte); } } } else { - output.push_back((*token).second); + output.push_back(token); } } } @@ -666,7 +580,7 @@ private: } const llama_vocab & vocab; - const llm_tokenizer_bpe * bpe_tokenizer; + const llm_tokenizer_bpe & tokenizer; std::vector symbols; std::vector symbols_final; @@ -678,14 +592,13 @@ private: // struct llm_tokenizer_wpm : llm_tokenizer { - llm_tokenizer_wpm(const llama_vocab & /*vocab*/) : llm_tokenizer() {} + llm_tokenizer_wpm(const llama_vocab & /*vocab*/) {} }; struct llm_tokenizer_wpm_session { llm_tokenizer_wpm_session(const llama_vocab & vocab) : vocab(vocab) {} - void tokenize(const std::string & text, std::vector & output) { - const auto & token_map = vocab.token_to_id; + void tokenize(const std::string & text, std::vector & output) { // normalize and split by whitespace std::vector words = preprocess(text); // bos token prepended already @@ -708,10 +621,10 @@ struct llm_tokenizer_wpm_session { for (int i = 0; i < n; ++i) { // loop through possible match length bool match = false; - for (int j = std::min(n, i + vocab.max_token_len + 1); j > i; j--) { - auto it = token_map.find(word1.substr(i, j - i)); - if (it != token_map.end()) { - output.push_back(it->second); + for (int j = std::min(n, i + vocab.max_token_len() + 1); j > i; j--) { + auto id = vocab.text_to_token(word1.substr(i, j - i)); + if (id != LLAMA_TOKEN_NULL) { + output.push_back(id); match = true; i = j - 1; break; @@ -726,7 +639,7 @@ struct llm_tokenizer_wpm_session { // we didn't find any matches for this word if (current_tokens == output.size()) { - output.push_back(vocab.special_unk_id); + output.push_back(vocab.token_unk()); } } } @@ -737,7 +650,7 @@ struct llm_tokenizer_wpm_session { std::vector words(1, ""); for (const uint32_t cpt : cpts_nfd) { - const auto flags = unicode_cpt_flags(cpt); + const auto flags = unicode_cpt_flags_from_cpt(cpt); if (flags.is_whitespace) { if (words.back().size()) { // finish previous word if any @@ -795,45 +708,45 @@ private: // struct llm_tokenizer_ugm : llm_tokenizer { - llm_tokenizer_ugm(const llama_vocab & vocab) : llm_tokenizer() { - if (vocab.precompiled_charsmap.size() > 0) { + llm_tokenizer_ugm(const llama_vocab & vocab, const std::vector & precompiled_charsmap) { + if (precompiled_charsmap.size() > 0) { size_t charsmap_offset = 0; // First four bytes of precompiled_charsmap contains length of binary // blob containing XOR-compressed compact double array (XCDA) entries - uint32_t xcda_blob_size = *(const uint32_t *) &vocab.precompiled_charsmap[0]; + uint32_t xcda_blob_size = *(const uint32_t *) &precompiled_charsmap[0]; charsmap_offset += sizeof(xcda_blob_size); - if (xcda_blob_size + charsmap_offset >= vocab.precompiled_charsmap.size()) { + if (xcda_blob_size + charsmap_offset >= precompiled_charsmap.size()) { throw std::runtime_error("Index out of array bounds in precompiled charsmap!"); } // Next xcda_blob_size bytes contain entries of XOR-compressed compact // double array (XCDA). Each entry is bit-packed into a 32-bit integer. - xcda_array = (const uint32_t *) &vocab.precompiled_charsmap[charsmap_offset]; + xcda_array = (const uint32_t *) &precompiled_charsmap[charsmap_offset]; xcda_array_size = xcda_blob_size / sizeof(uint32_t); charsmap_offset += xcda_blob_size; // Remaining bytes of precompiled charsmap contain null-terminated // replacement strings for prefixes matched by the XCDA. - prefix_replacements = &vocab.precompiled_charsmap[charsmap_offset]; - prefix_replacements_size = vocab.precompiled_charsmap.size() - charsmap_offset; + prefix_replacements = &precompiled_charsmap[charsmap_offset]; + prefix_replacements_size = precompiled_charsmap.size() - charsmap_offset; } - for (unsigned int id = 0; id < vocab.id_to_token.size(); ++id) { - const auto &token_data = vocab.id_to_token[id]; + for (uint32_t id = 0; id < vocab.n_tokens(); ++id) { + const auto & token_data = vocab.get_token_data(id); - if (llama_is_normal_token(vocab, id)) { + if (vocab.is_normal(id)) { min_score = std::min(min_score, token_data.score); max_score = std::max(max_score, token_data.score); } - if (llama_is_normal_token(vocab, id) || - llama_is_user_defined_token(vocab, id) || - llama_is_unused_token(vocab, id)) { + if (vocab.is_normal(id) || + vocab.is_user_defined(id) || + vocab.is_unused(id)) { token_matcher.insert(token_data.text.data(), token_data.text.size(), id); } - if (llama_is_user_defined_token(vocab, id)) { + if (vocab.is_user_defined(id)) { user_defined_token_matcher.insert(token_data.text.data(), token_data.text.size()); } } @@ -862,8 +775,7 @@ struct llm_tokenizer_ugm : llm_tokenizer { }; struct llm_tokenizer_ugm_session { - llm_tokenizer_ugm_session(const llama_vocab & vocab) : vocab(vocab), - ugm_tokenizer(static_cast(vocab.tokenizer)) {} + llm_tokenizer_ugm_session(const llama_vocab & vocab, const llm_tokenizer_ugm & tokenizer) : vocab(vocab), tokenizer(tokenizer) {} /* This implementation is based on SentencePiece optimized Viterbi algorithm for * unigram language models. The general idea is to: @@ -878,7 +790,7 @@ struct llm_tokenizer_ugm_session { * After processing the whole sequence we backtrack from the end to get * the best tokenization. */ - void tokenize(const std::string & text, std::vector & output) { + void tokenize(const std::string & text, std::vector & output) { // get current size of output (for reversal later) size_t output_size = output.size(); @@ -891,9 +803,9 @@ struct llm_tokenizer_ugm_session { } // initialize score_sum to -FLT_MAX so it will be always lower than sums of token scores - std::vector tokenization_results(input_len + 1, {vocab.special_unk_id, 0, -FLT_MAX}); + std::vector tokenization_results(input_len + 1, {vocab.token_unk(), 0, -FLT_MAX}); // at the beginning tokenization score is zero - tokenization_results[0] = { vocab.special_unk_id, 0, 0 }; + tokenization_results[0] = { vocab.token_unk(), 0, 0 }; for (size_t input_offset = 0; input_offset < input_len;) { size_t prefix_offset = input_offset; @@ -903,7 +815,7 @@ struct llm_tokenizer_ugm_session { // traverse the token matcher trie to find a matching token bool single_codepoint_token_found = false; const struct best_tokenization & current_best = tokenization_results[input_offset]; - const struct naive_trie * node = ugm_tokenizer->token_matcher.traverse(normalized[prefix_offset++]); + const struct naive_trie * node = tokenizer.token_matcher.traverse(normalized[prefix_offset++]); while (prefix_offset <= input_len && node != NULL) { // check if we found valid token in prefix @@ -913,13 +825,13 @@ struct llm_tokenizer_ugm_session { single_codepoint_token_found = true; } llama_token token_id = node->value; - const auto & token_data = vocab.id_to_token[token_id]; + const auto & token_data = vocab.get_token_data(token_id); // we set the user-defined token scores to 0 to make them more likely to be selected // (normal token scores are log probabilities, so they are negative) // score type is double here to make tokenization results exactly // the same as in the HF tokenizer using SentencePiece - const double token_score = llama_is_user_defined_token(vocab, token_id) ? 0.0 : token_data.score; + const double token_score = vocab.is_user_defined(token_id) ? 0.0 : token_data.score; const double challenger_score = current_best.score_sum + token_score; struct best_tokenization & current_champ = tokenization_results[prefix_offset]; if (challenger_score > current_champ.score_sum) { @@ -933,11 +845,11 @@ struct llm_tokenizer_ugm_session { // if we didn't find a valid token corresponding to the whole UTF code point // then use unknown token as the tokenization of this UTF code point if (!single_codepoint_token_found) { - const double challenger_score = current_best.score_sum + ugm_tokenizer->unknown_token_score; + const double challenger_score = current_best.score_sum + tokenizer.unknown_token_score; prefix_offset = input_offset + n_utf8_code_units; struct best_tokenization & current_champ = tokenization_results[prefix_offset]; if (challenger_score > current_champ.score_sum) { - struct best_tokenization challenger = { vocab.special_unk_id, input_offset, (float) challenger_score }; + struct best_tokenization challenger = { vocab.token_unk(), input_offset, (float) challenger_score }; current_champ = challenger; } } @@ -950,7 +862,7 @@ struct llm_tokenizer_ugm_session { // merge sequences of consecutive unknown tokens into single unknown tokens bool is_prev_unknown = false; for (struct best_tokenization & tokenization = tokenization_results[input_len]; ; tokenization = tokenization_results[tokenization.input_offset]) { - bool is_unknown = tokenization.token_id == vocab.special_unk_id; + bool is_unknown = tokenization.token_id == vocab.token_unk(); if (!(is_prev_unknown && is_unknown)) { output.push_back(tokenization.token_id); } @@ -977,11 +889,11 @@ private: normalized->clear(); normalized->reserve(input.size() * 3); - const std::string space = vocab.tokenizer_escape_whitespaces ? ugm_tokenizer->escaped_space : " "; + const std::string space = vocab.get_escape_whitespaces() ? tokenizer.escaped_space : " "; - bool shall_prepend_space = !vocab.tokenizer_treat_whitespace_as_suffix && vocab.tokenizer_add_space_prefix; - bool shall_append_space = vocab.tokenizer_treat_whitespace_as_suffix && vocab.tokenizer_add_space_prefix; - bool shall_merge_spaces = vocab.tokenizer_remove_extra_whitespaces; + const bool shall_prepend_space = !vocab.get_treat_whitespace_as_suffix() && vocab.get_add_space_prefix(); + const bool shall_append_space = vocab.get_treat_whitespace_as_suffix() && vocab.get_add_space_prefix(); + const bool shall_merge_spaces = vocab.get_remove_extra_whitespaces(); bool is_space_prepended = false; bool processing_non_ws = false; @@ -1073,7 +985,7 @@ private: // if input prefix matches some user-defined token return this token as normalization result auto user_defined_token_match = - ugm_tokenizer->user_defined_token_matcher.get_longest_prefix(&input[input_offset], input.size() - input_offset); + tokenizer.user_defined_token_matcher.get_longest_prefix(&input[input_offset], input.size() - input_offset); if (user_defined_token_match.second > 0) { return { &input[input_offset], user_defined_token_match.second, user_defined_token_match.second }; } @@ -1081,8 +993,8 @@ private: size_t longest_prefix_length = 0; size_t longest_prefix_offset = 0; - if (ugm_tokenizer->xcda_array_size > 0) { - struct xcda_array_view xcda_view(ugm_tokenizer->xcda_array, ugm_tokenizer->xcda_array_size); + if (tokenizer.xcda_array_size > 0) { + struct xcda_array_view xcda_view(tokenizer.xcda_array, tokenizer.xcda_array_size); // Find the longest normalized sequence matching the input prefix by walking // the XOR-compressed compact double array (XCDA) starting from the root node @@ -1118,10 +1030,10 @@ private: if (longest_prefix_length > 0) { // we have a match, so return the replacement sequence - if (longest_prefix_offset >= ugm_tokenizer->prefix_replacements_size) { + if (longest_prefix_offset >= tokenizer.prefix_replacements_size) { throw std::runtime_error("Index out of array bounds in precompiled charsmap!"); } - const char * prefix_replacement = &(ugm_tokenizer->prefix_replacements)[longest_prefix_offset]; + const char * prefix_replacement = &(tokenizer.prefix_replacements)[longest_prefix_offset]; return { prefix_replacement, strlen(prefix_replacement), longest_prefix_length }; } @@ -1138,7 +1050,7 @@ private: } const llama_vocab & vocab; - const llm_tokenizer_ugm * ugm_tokenizer; + const llm_tokenizer_ugm & tokenizer; }; // @@ -1200,15 +1112,15 @@ static std::vector llama_unescape_rwkv_token(const std::string & escape } struct llm_tokenizer_rwkv : llm_tokenizer { - llm_tokenizer_rwkv(const llama_vocab & vocab) : llm_tokenizer() { + llm_tokenizer_rwkv(const llama_vocab & vocab) { // RWKV supports arbitrary byte tokens, but the vocab struct only supports string tokens. // For now, we decode the vocab here into the lookup we'll use for tokenization. // build trie - for (unsigned int id = 0; id < vocab.id_to_token.size(); ++id) { - const auto & token = vocab.id_to_token[id]; - const auto data = llama_unescape_rwkv_token(token.text); - token_matcher.insert((const char *) data.data(), data.size(), id); + for (uint32_t id = 0; id < vocab.n_tokens(); ++id) { + const auto & data = vocab.get_token_data(id); + const auto text = llama_unescape_rwkv_token(data.text); + token_matcher.insert((const char *) text.data(), text.size(), id); } } @@ -1216,16 +1128,15 @@ struct llm_tokenizer_rwkv : llm_tokenizer { }; struct llm_tokenizer_rwkv_session { - llm_tokenizer_rwkv_session(const llama_vocab & vocab) : vocab(vocab), - rwkv_tokenizer(static_cast(*vocab.tokenizer)) {} + llm_tokenizer_rwkv_session(const llama_vocab & vocab, const llm_tokenizer_rwkv & tokenizer) : vocab(vocab), tokenizer(tokenizer) {} - void tokenize(const std::string & text, std::vector & output) { + void tokenize(const std::string & text, std::vector & output) { uint32_t position = 0; while (position < text.size()) { - const struct naive_trie * node = rwkv_tokenizer.token_matcher.traverse(text[position]); + const struct naive_trie * node = tokenizer.token_matcher.traverse(text[position]); if (node == NULL) { // no matching token found, add unknown token - output.push_back(vocab.special_unk_id); + output.push_back(vocab.token_unk()); position += 1; continue; } @@ -1249,33 +1160,11 @@ struct llm_tokenizer_rwkv_session { private: const llama_vocab & vocab; - const llm_tokenizer_rwkv & rwkv_tokenizer; + const llm_tokenizer_rwkv & tokenizer; }; -void llama_vocab::init_tokenizer() { - switch (type) { - case LLAMA_VOCAB_TYPE_SPM: - tokenizer = new llm_tokenizer_spm(*this); - break; - case LLAMA_VOCAB_TYPE_BPE: - tokenizer = new llm_tokenizer_bpe(*this); - break; - case LLAMA_VOCAB_TYPE_WPM: - tokenizer = new llm_tokenizer_wpm(*this); - break; - case LLAMA_VOCAB_TYPE_UGM: - tokenizer = new llm_tokenizer_ugm(*this); - break; - case LLAMA_VOCAB_TYPE_RWKV: - tokenizer = new llm_tokenizer_rwkv(*this); - break; - default: - GGML_ABORT("unsupported vocab type"); - } -} - // -// (de-) tokenize +// impl // typedef enum FRAGMENT_BUFFER_VARIANT_TYPE { @@ -1284,7 +1173,7 @@ typedef enum FRAGMENT_BUFFER_VARIANT_TYPE { } FRAGMENT_BUFFER_VARIANT_TYPE; struct fragment_buffer_variant { - fragment_buffer_variant(llama_vocab::id _token) + fragment_buffer_variant(llama_token _token) : type(FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN), token(_token), @@ -1295,7 +1184,7 @@ struct fragment_buffer_variant { fragment_buffer_variant(const std::string & _raw_text, int64_t _offset, int64_t _length) : type(FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT), - token((llama_vocab::id) - 1), + token((llama_token) - 1), raw_text(_raw_text), offset(_offset), length(_length){ @@ -1305,20 +1194,957 @@ struct fragment_buffer_variant { } const FRAGMENT_BUFFER_VARIANT_TYPE type; - const llama_vocab::id token; + const llama_token token; const std::string _dummy; const std::string & raw_text; const uint64_t offset; const uint64_t length; }; +struct llama_vocab::impl { + uint32_t n_token_types = 0; // for BERT-style token types + + enum llama_vocab_type type = LLAMA_VOCAB_TYPE_SPM; + enum llama_vocab_pre_type pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + + int max_token_len = 0; // used for optimizing longest token search + + // default LLaMA special tokens + // TODO: should we set all of these to LLAMA_TOKEN_NULL? + llama_token special_bos_id = 1; + llama_token special_eos_id = 2; + llama_token special_eot_id = LLAMA_TOKEN_NULL; + llama_token special_eom_id = LLAMA_TOKEN_NULL; + llama_token special_unk_id = 0; + llama_token special_sep_id = LLAMA_TOKEN_NULL; + llama_token special_pad_id = LLAMA_TOKEN_NULL; + llama_token special_mask_id = LLAMA_TOKEN_NULL; + + llama_token linefeed_id = 13; + + // fim tokens + llama_token special_fim_pre_id = LLAMA_TOKEN_NULL; + llama_token special_fim_suf_id = LLAMA_TOKEN_NULL; + llama_token special_fim_mid_id = LLAMA_TOKEN_NULL; + llama_token special_fim_pad_id = LLAMA_TOKEN_NULL; + llama_token special_fim_rep_id = LLAMA_TOKEN_NULL; // repo + llama_token special_fim_sep_id = LLAMA_TOKEN_NULL; // file separator + + // tokenizer flags + bool add_space_prefix = false; + bool add_bos = false; + bool add_eos = false; + bool ignore_merges = false; + bool clean_spaces = false; // clean_up_tokenization_spaces + bool remove_extra_whitespaces = false; + bool escape_whitespaces = true; + bool treat_whitespace_as_suffix = false; + + std::unordered_map token_to_id; + std::vector id_to_token; + + std::vector cache_special_tokens; + std::vector cache_token_to_piece; // llama_token_to_piece(special = true); + + std::map, int> bpe_ranks; + + // set of all tokens that cause "end of generation" + std::set special_eog_ids; + + std::unique_ptr tokenizer; + + std::vector precompiled_charsmap; + + impl(const llama_vocab & vocab) : vocab(vocab) { + } + + ~impl() = default; + + void load(llama_model_loader & ml, const LLM_KV & kv); + + enum llama_vocab_type get_type() const; + + std::string type_name() const; + + bool is_normal (llama_token id) const; + bool is_unknown (llama_token id) const; + bool is_control (llama_token id) const; + bool is_byte (llama_token id) const; + bool is_user_defined(llama_token id) const; + bool is_unused (llama_token id) const; + bool is_eog (llama_token id) const; + + uint8_t token_to_byte(llama_token id) const; + + llama_token_attr token_get_attr(llama_token id) const; + + void init_tokenizer(enum llama_vocab_type type); + + void tokenizer_st_partition(std::forward_list & buffer, bool parse_special) const; + + std::string token_to_piece_for_cache( + llama_token token, + bool special) const; + + + std::vector tokenize( + const std::string & raw_text, + bool add_special, + bool parse_special = false) const; + + int32_t tokenize( + const char * text, + int32_t text_len, + llama_token * tokens, + int32_t n_tokens_max, + bool add_special, + bool parse_special) const; + + // does not write null-terminator to buf + int32_t token_to_piece( + llama_token token, + char * buf, + int32_t length, + int32_t lstrip, + bool special) const; + + // use cached data + const std::string & token_to_piece(llama_token token) const; + + int32_t detokenize( + const llama_token * tokens, + int32_t n_tokens, + char * text, + int32_t text_len_max, + bool remove_special, + bool unparse_special) const; + + std::string detokenize( + const std::vector & tokens, + bool special) const; + + void print_info() const; + +private: + const llama_vocab & vocab; +}; + +void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) { + struct gguf_context * ctx = ml.meta.get(); + + // determine vocab type + { + std::string tokenizer_model; + std::string tokenizer_pre; + + ml.get_key(LLM_KV_TOKENIZER_MODEL, tokenizer_model); + ml.get_key(LLM_KV_TOKENIZER_PRE, tokenizer_pre, false); + + ml.get_key(LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, n_token_types, false); + + if (tokenizer_model == "no_vocab" || tokenizer_model == "none") { + type = LLAMA_VOCAB_TYPE_NONE; + + // default special tokens + special_bos_id = LLAMA_TOKEN_NULL; + special_eos_id = LLAMA_TOKEN_NULL; + special_unk_id = LLAMA_TOKEN_NULL; + special_sep_id = LLAMA_TOKEN_NULL; + special_pad_id = LLAMA_TOKEN_NULL; + special_mask_id = LLAMA_TOKEN_NULL; + linefeed_id = LLAMA_TOKEN_NULL; + + // read vocab size from metadata + uint32_t n_tokens = 0; + if (ml.get_key(LLM_KV_VOCAB_SIZE, n_tokens, false)) { + LLAMA_LOG_WARN("%s: adding %u dummy tokens\n", __func__, n_tokens); + id_to_token.resize(n_tokens); + } + + return; + } + + if (tokenizer_model == "llama") { + type = LLAMA_VOCAB_TYPE_SPM; + + // default special tokens + special_bos_id = 1; + special_eos_id = 2; + special_unk_id = 0; + special_sep_id = LLAMA_TOKEN_NULL; + special_pad_id = LLAMA_TOKEN_NULL; + special_mask_id = LLAMA_TOKEN_NULL; + } else if (tokenizer_model == "bert") { + type = LLAMA_VOCAB_TYPE_WPM; + + // default special tokens + special_bos_id = 101; + special_eos_id = LLAMA_TOKEN_NULL; + special_unk_id = 100; + special_sep_id = 102; + special_pad_id = 0; + special_mask_id = 103; + } else if (tokenizer_model == "gpt2") { + type = LLAMA_VOCAB_TYPE_BPE; + + // read bpe merges and populate bpe ranks + const int merges_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_MERGES).c_str()); + if (merges_keyidx == -1) { + throw std::runtime_error("cannot find tokenizer merges in model file\n"); + } + + const int n_merges = gguf_get_arr_n(ctx, merges_keyidx); + for (int i = 0; i < n_merges; i++) { + const std::string word = gguf_get_arr_str(ctx, merges_keyidx, i); + //GGML_ASSERT(unicode_cpts_from_utf8(word).size() > 0); + + std::string first; + std::string second; + + const size_t pos = word.find(' ', 1); + + if (pos != std::string::npos) { + first = word.substr(0, pos); + second = word.substr(pos + 1); + } + + bpe_ranks.emplace(std::make_pair(first, second), i); + } + + // default special tokens + special_bos_id = 11; + special_eos_id = 11; + special_unk_id = LLAMA_TOKEN_NULL; + special_sep_id = LLAMA_TOKEN_NULL; + special_pad_id = LLAMA_TOKEN_NULL; + special_mask_id = LLAMA_TOKEN_NULL; + } else if (tokenizer_model == "t5") { + type = LLAMA_VOCAB_TYPE_UGM; + + // default special tokens + special_bos_id = LLAMA_TOKEN_NULL; + special_eos_id = 1; + special_unk_id = 2; + special_sep_id = LLAMA_TOKEN_NULL; + special_pad_id = 0; + special_mask_id = LLAMA_TOKEN_NULL; + + const int precompiled_charsmap_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP).c_str()); + if (precompiled_charsmap_keyidx != -1) { + size_t n_precompiled_charsmap = gguf_get_arr_n(ctx, precompiled_charsmap_keyidx); + const char * pc = (const char *) gguf_get_arr_data(ctx, precompiled_charsmap_keyidx); + precompiled_charsmap.assign(pc, pc + n_precompiled_charsmap); +#ifdef IS_BIG_ENDIAN + // correct endiannes of data in precompiled_charsmap binary blob + uint32_t * xcda_blob_size = (uint32_t *) &precompiled_charsmap[0]; + *xcda_blob_size = __builtin_bswap32(*xcda_blob_size); + assert(*xcda_blob_size + sizeof(uint32_t) < n_precompiled_charsmap); + size_t xcda_array_size = *xcda_blob_size / sizeof(uint32_t); + uint32_t * xcda_array = (uint32_t *) &precompiled_charsmap[sizeof(uint32_t)]; + for (size_t i = 0; i < xcda_array_size; ++i) { + xcda_array[i] = __builtin_bswap32(xcda_array[i]); + } +#endif + } + } else if (tokenizer_model == "rwkv") { + type = LLAMA_VOCAB_TYPE_RWKV; + + // default special tokens + special_bos_id = LLAMA_TOKEN_NULL; + special_eos_id = LLAMA_TOKEN_NULL; + special_unk_id = LLAMA_TOKEN_NULL; + special_sep_id = LLAMA_TOKEN_NULL; + special_pad_id = LLAMA_TOKEN_NULL; + } else { + throw std::runtime_error(format("unknown tokenizer: '%s'", tokenizer_model.c_str())); + } + + // for now, only BPE models have pre-tokenizers + if (type == LLAMA_VOCAB_TYPE_BPE) { + add_space_prefix = false; + clean_spaces = true; + if (tokenizer_pre.empty()) { + LLAMA_LOG_WARN("%s: missing pre-tokenizer type, using: 'default'\n", __func__); + LLAMA_LOG_WARN("%s: \n", __func__); + LLAMA_LOG_WARN("%s: ************************************ \n", __func__); + LLAMA_LOG_WARN("%s: GENERATION QUALITY WILL BE DEGRADED! \n", __func__); + LLAMA_LOG_WARN("%s: CONSIDER REGENERATING THE MODEL \n", __func__); + LLAMA_LOG_WARN("%s: ************************************ \n", __func__); + LLAMA_LOG_WARN("%s: \n", __func__); + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + } else if (tokenizer_pre == "default") { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + } else if ( + tokenizer_pre == "llama3" || + tokenizer_pre == "llama-v3" || + tokenizer_pre == "llama-bpe"|| + tokenizer_pre == "falcon3") { + pre_type = LLAMA_VOCAB_PRE_TYPE_LLAMA3; + ignore_merges = true; + add_bos = true; + } else if ( + tokenizer_pre == "deepseek-llm") { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM; + clean_spaces = false; + } else if ( + tokenizer_pre == "deepseek-coder") { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER; + clean_spaces = false; + } else if ( + tokenizer_pre == "deepseek-v3") { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK3_LLM; + clean_spaces = false; + } else if ( + tokenizer_pre == "falcon") { + pre_type = LLAMA_VOCAB_PRE_TYPE_FALCON; + } else if ( + tokenizer_pre == "mpt") { + pre_type = LLAMA_VOCAB_PRE_TYPE_MPT; + } else if ( + tokenizer_pre == "starcoder") { + pre_type = LLAMA_VOCAB_PRE_TYPE_STARCODER; + } else if ( + tokenizer_pre == "gpt-2" || + tokenizer_pre == "phi-2" || + tokenizer_pre == "jina-es" || + tokenizer_pre == "jina-de" || + tokenizer_pre == "gigachat" || + tokenizer_pre == "jina-v1-en" || + tokenizer_pre == "jina-v2-es" || + tokenizer_pre == "jina-v2-de" || + tokenizer_pre == "jina-v2-code" || + tokenizer_pre == "roberta-bpe") { + pre_type = LLAMA_VOCAB_PRE_TYPE_GPT2; + } else if ( + tokenizer_pre == "refact") { + pre_type = LLAMA_VOCAB_PRE_TYPE_REFACT; + } else if ( + tokenizer_pre == "command-r") { + pre_type = LLAMA_VOCAB_PRE_TYPE_COMMAND_R; + clean_spaces = false; + } else if ( + tokenizer_pre == "qwen2") { + pre_type = LLAMA_VOCAB_PRE_TYPE_QWEN2; + clean_spaces = false; + } else if ( + tokenizer_pre == "stablelm2") { + pre_type = LLAMA_VOCAB_PRE_TYPE_STABLELM2; + } else if ( + tokenizer_pre == "olmo") { + pre_type = LLAMA_VOCAB_PRE_TYPE_OLMO; + } else if ( + tokenizer_pre == "dbrx") { + pre_type = LLAMA_VOCAB_PRE_TYPE_DBRX; + } else if ( + tokenizer_pre == "smaug-bpe") { + pre_type = LLAMA_VOCAB_PRE_TYPE_SMAUG; + } else if ( + tokenizer_pre == "poro-chat") { + pre_type = LLAMA_VOCAB_PRE_TYPE_PORO; + clean_spaces = false; + } else if ( + tokenizer_pre == "chatglm-bpe") { + pre_type = LLAMA_VOCAB_PRE_TYPE_CHATGLM4; + special_bos_id = LLAMA_TOKEN_NULL; + } else if ( + tokenizer_pre == "viking") { + pre_type = LLAMA_VOCAB_PRE_TYPE_VIKING; + clean_spaces = false; + } else if ( + tokenizer_pre == "jais") { + pre_type = LLAMA_VOCAB_PRE_TYPE_JAIS; + } else if ( + tokenizer_pre == "tekken") { + pre_type = LLAMA_VOCAB_PRE_TYPE_TEKKEN; + clean_spaces = false; + ignore_merges = true; + add_bos = true; + } else if ( + tokenizer_pre == "smollm") { + pre_type = LLAMA_VOCAB_PRE_TYPE_SMOLLM; + clean_spaces = false; + } else if ( + tokenizer_pre == "codeshell") { + pre_type = LLAMA_VOCAB_PRE_TYPE_CODESHELL; + } else if ( + tokenizer_pre == "bloom") { + pre_type = LLAMA_VOCAB_PRE_TYPE_BLOOM; + } else if ( + tokenizer_pre == "gpt3-finnish") { + pre_type = LLAMA_VOCAB_PRE_TYPE_GPT3_FINNISH; + } else if ( + tokenizer_pre == "exaone") { + pre_type = LLAMA_VOCAB_PRE_TYPE_EXAONE; + } else if ( + tokenizer_pre == "chameleon") { + pre_type = LLAMA_VOCAB_PRE_TYPE_CHAMELEON; + add_bos = true; + clean_spaces = false; + } else if ( + tokenizer_pre == "minerva-7b") { + pre_type = LLAMA_VOCAB_PRE_TYPE_MINERVA; + } else if ( + tokenizer_pre == "megrez") { + pre_type = LLAMA_VOCAB_PRE_TYPE_QWEN2; + } else { + throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str())); + } + } else if (type == LLAMA_VOCAB_TYPE_SPM) { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + add_space_prefix = true; + clean_spaces = false; + add_bos = true; + add_eos = false; + } else if (type == LLAMA_VOCAB_TYPE_WPM) { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + add_space_prefix = false; + clean_spaces = true; + add_bos = true; + add_eos = false; + } else if (type == LLAMA_VOCAB_TYPE_UGM) { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + add_bos = false; + add_eos = true; + } else if (type == LLAMA_VOCAB_TYPE_RWKV) { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + add_space_prefix = false; + clean_spaces = false; + add_bos = false; + add_eos = false; + } else { + pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT; + } + + ml.get_key(LLM_KV_TOKENIZER_ADD_PREFIX, add_space_prefix, false); + ml.get_key(LLM_KV_TOKENIZER_REMOVE_EXTRA_WS, remove_extra_whitespaces, false); + } + + const int token_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_LIST).c_str()); + if (token_idx == -1) { + throw std::runtime_error("cannot find tokenizer vocab in model file\n"); + } + + const float * scores = nullptr; + const int score_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_SCORES).c_str()); + if (score_idx != -1) { + scores = (const float * ) gguf_get_arr_data(ctx, score_idx); + } + + const int * toktypes = nullptr; + const int toktype_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_TOKEN_TYPE).c_str()); + if (toktype_idx != -1) { + toktypes = (const int * ) gguf_get_arr_data(ctx, toktype_idx); + } + + uint32_t n_tokens = gguf_get_arr_n(ctx, token_idx); + id_to_token.resize(n_tokens); + + for (uint32_t i = 0; i < n_tokens; i++) { + std::string word = gguf_get_arr_str(ctx, token_idx, i); + if (word.empty()) { + LLAMA_LOG_WARN("%s: empty token at index %u\n", __func__, i); + word = "[EMPTY_" + std::to_string(i) + "]"; + } + + token_to_id[word] = i; + max_token_len = std::max(max_token_len, (int) word.size()); + + auto & token_data = id_to_token[i]; + token_data.text = std::move(word); + token_data.score = scores ? scores[i] : 0.0f; + token_data.attr = LLAMA_TOKEN_ATTR_NORMAL; + + if (toktypes) { //TODO: remove, required until per token attributes are available from GGUF file + switch(toktypes[i]) { + case LLAMA_TOKEN_TYPE_UNKNOWN: token_data.attr = LLAMA_TOKEN_ATTR_UNKNOWN; break; + case LLAMA_TOKEN_TYPE_UNUSED: token_data.attr = LLAMA_TOKEN_ATTR_UNUSED; break; + case LLAMA_TOKEN_TYPE_NORMAL: token_data.attr = LLAMA_TOKEN_ATTR_NORMAL; break; + case LLAMA_TOKEN_TYPE_CONTROL: token_data.attr = LLAMA_TOKEN_ATTR_CONTROL; break; + case LLAMA_TOKEN_TYPE_USER_DEFINED: token_data.attr = LLAMA_TOKEN_ATTR_USER_DEFINED; break; + case LLAMA_TOKEN_TYPE_BYTE: token_data.attr = LLAMA_TOKEN_ATTR_BYTE; break; + case LLAMA_TOKEN_TYPE_UNDEFINED: token_data.attr = LLAMA_TOKEN_ATTR_UNDEFINED; break; + default: token_data.attr = LLAMA_TOKEN_ATTR_UNDEFINED; break; + } + } + } + GGML_ASSERT(id_to_token.size() == token_to_id.size()); + + init_tokenizer(type); + + // determine the newline token: LLaMA "<0x0A>" == 10 == '\n', Falcon 193 == '\n' + if (type == LLAMA_VOCAB_TYPE_SPM) { + try { + linefeed_id = vocab.byte_to_token('\n'); + } catch (const std::exception & e) { + LLAMA_LOG_WARN("%s: SPM vocabulary, but newline token not found: %s! Using special_pad_id instead.", __func__, e.what()); + linefeed_id = special_pad_id; + } + } else if (type == LLAMA_VOCAB_TYPE_WPM) { + linefeed_id = special_pad_id; + } else if (type == LLAMA_VOCAB_TYPE_RWKV) { + const std::vector ids = tokenize("\n", false); + GGML_ASSERT(!ids.empty() && "model vocab missing newline token"); + linefeed_id = ids[0]; + } else { + const std::vector ids = tokenize("\xC4\x8A", false); // U+010A + + //GGML_ASSERT(!ids.empty() && "model vocab missing newline token"); + if (ids.empty()) { + LLAMA_LOG_WARN("%s: model vocab missing newline token, using special_pad_id instead\n", __func__); + linefeed_id = special_pad_id; + } else { + linefeed_id = ids[0]; + } + } + + // special tokens + { + const std::vector> special_token_types = { + { LLM_KV_TOKENIZER_BOS_ID, special_bos_id }, + { LLM_KV_TOKENIZER_EOS_ID, special_eos_id }, + { LLM_KV_TOKENIZER_EOT_ID, special_eot_id }, + { LLM_KV_TOKENIZER_EOM_ID, special_eom_id }, + { LLM_KV_TOKENIZER_UNK_ID, special_unk_id }, + { LLM_KV_TOKENIZER_SEP_ID, special_sep_id }, + { LLM_KV_TOKENIZER_PAD_ID, special_pad_id }, + { LLM_KV_TOKENIZER_MASK_ID, special_mask_id }, + { LLM_KV_TOKENIZER_FIM_PRE_ID, special_fim_pre_id }, + { LLM_KV_TOKENIZER_FIM_SUF_ID, special_fim_suf_id }, + { LLM_KV_TOKENIZER_FIM_MID_ID, special_fim_mid_id }, + { LLM_KV_TOKENIZER_FIM_PAD_ID, special_fim_pad_id }, + { LLM_KV_TOKENIZER_FIM_REP_ID, special_fim_rep_id }, + { LLM_KV_TOKENIZER_FIM_SEP_ID, special_fim_sep_id }, + + // deprecated + { LLM_KV_TOKENIZER_PREFIX_ID, special_fim_pre_id }, + { LLM_KV_TOKENIZER_SUFFIX_ID, special_fim_suf_id }, + { LLM_KV_TOKENIZER_MIDDLE_ID, special_fim_mid_id }, + }; + + for (const auto & it : special_token_types) { + const std::string & key = kv(std::get<0>(it)); + int32_t & id = std::get<1>(it); + + uint32_t new_id; + if (!ml.get_key(std::get<0>(it), new_id, false)) { + continue; + } + if (new_id >= id_to_token.size()) { + LLAMA_LOG_WARN("%s: bad special token: '%s' = %u, using default id %d\n", + __func__, key.c_str(), new_id, id); + } else { + id = new_id; + } + } + + // Handle add_bos and add_eos + { + bool temp = true; + + if (ml.get_key(LLM_KV_TOKENIZER_ADD_BOS, temp, false)) { + add_bos = temp; + } + if (ml.get_key(LLM_KV_TOKENIZER_ADD_EOS, temp, false)) { + add_eos = temp; + } + } + + // auto-detect special tokens by text + // TODO: convert scripts should provide these tokens through the KV metadata LLM_KV_TOKENIZER_... + // for now, we apply this workaround to find the tokens based on their text + + for (const auto & t : token_to_id) { + // find EOT token: "<|eot_id|>", "<|im_end|>", "", etc. + if (special_eot_id == LLAMA_TOKEN_NULL) { + if (false + || t.first == "<|eot_id|>" + || t.first == "<|im_end|>" + || t.first == "<|end|>" + || t.first == "" + || t.first == "<|endoftext|>" + || t.first == "" + || t.first == "<｜end▁of▁sentence｜>" // DeepSeek + ) { + special_eot_id = t.second; + if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) { + LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n", + __func__, t.second, t.first.c_str()); + id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL; + } + } + } + + // find EOM token: "<|eom_id|>" + if (special_eom_id == LLAMA_TOKEN_NULL) { + if (false + || t.first == "<|eom_id|>" + ) { + special_eom_id = t.second; + if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) { + LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n", + __func__, t.second, t.first.c_str()); + id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL; + } + } + } + + // find FIM_PRE token: "<|fim_prefix|>", "", "
", etc.
+            if (special_fim_pre_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|fim_prefix|>"  // Qwen
+                        || t.first == ""
+                        || t.first == "<｜fim▁begin｜>" // DeepSeek
+                        || t.first == "
"
+                        ) {
+                    special_fim_pre_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+
+            // find FIM_SUF token: "<|fim_suffix|>", "", "", etc.
+            if (special_fim_suf_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|fim_suffix|>" // Qwen
+                        || t.first == ""
+                        || t.first == "<｜fim▁hole｜>" // DeepSeek
+                        || t.first == ""
+                        ) {
+                    special_fim_suf_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+
+            // find FIM_MID token: "<|fim_middle|>", "", "", etc.
+            if (special_fim_mid_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|fim_middle|>" // Qwen
+                        || t.first == ""
+                        || t.first == "<｜fim▁end｜>"  // DeepSeek
+                        || t.first == ""
+                        ) {
+                    special_fim_mid_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+
+            // find FIM_PAD token: "<|fim_pad|>", "", "", etc.
+            if (special_fim_pad_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|fim_pad|>" // Qwen
+                        || t.first == ""
+                        || t.first == ""
+                        ) {
+                    special_fim_pad_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+
+            // find FIM_REP token: "<|fim_repo|>", "", "", etc.
+            if (special_fim_rep_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|fim_repo|>"  // Qwen
+                        || t.first == "<|repo_name|>"
+                        || t.first == ""
+                        || t.first == ""
+                        ) {
+                    special_fim_rep_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+
+            // find FIM_SEP token: "<|file_sep|>"
+            if (special_fim_sep_id == LLAMA_TOKEN_NULL) {
+                if (false
+                        || t.first == "<|file_sep|>" // Qwen
+                        ) {
+                    special_fim_sep_id = t.second;
+                    if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                                __func__, t.second, t.first.c_str());
+                        id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                    }
+                }
+            }
+        }
+
+        // maintain a list of tokens that cause end-of-generation
+        // this is currently determined based on the token text, which is obviously not ideal
+        // ref: https://github.com/ggerganov/llama.cpp/issues/9606
+        special_eog_ids.clear();
+
+        if (special_fim_pad_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_fim_pad_id) == 0) {
+            special_eog_ids.insert(special_fim_pad_id);
+        }
+
+        if (special_fim_rep_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_fim_rep_id) == 0) {
+            special_eog_ids.insert(special_fim_rep_id);
+        }
+
+        if (special_fim_sep_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_fim_sep_id) == 0) {
+            special_eog_ids.insert(special_fim_sep_id);
+        }
+
+        for (const auto & t : token_to_id) {
+            if (false
+                    || t.first == "<|eot_id|>"
+                    || t.first == "<|im_end|>"
+                    || t.first == "<|end|>"
+                    || t.first == ""
+                    || t.first == "<|endoftext|>"
+                    || t.first == "<|eom_id|>"
+                    || t.first == ""
+               ) {
+                special_eog_ids.insert(t.second);
+                if ((id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
+                    LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
+                            __func__, t.second, t.first.c_str());
+                    id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
+                }
+            } else {
+                // token is control, but not marked as EOG -> print a debug log
+                if (id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL && special_eog_ids.count(t.second) == 0) {
+                    LLAMA_LOG_DEBUG("%s: control token: %6d '%s' is not marked as EOG\n",
+                            __func__, t.second, t.first.c_str());
+                }
+            }
+        }
+
+        // sanity checks
+        if (special_eos_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_eos_id) == 0) {
+            special_eog_ids.insert(special_eos_id);
+            LLAMA_LOG_WARN("%s: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
+        }
+
+        if (special_eot_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_eot_id) == 0) {
+            special_eog_ids.insert(special_eot_id);
+            LLAMA_LOG_WARN("%s: special_eot_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
+        }
+
+        if (special_eom_id != LLAMA_TOKEN_NULL && special_eog_ids.count(special_eom_id) == 0) {
+            special_eog_ids.insert(special_eom_id);
+            LLAMA_LOG_WARN("%s: special_eom_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
+        }
+    }
+
+    // build special tokens cache
+    {
+        for (llama_token id = 0; id < (llama_token) n_tokens; ++id) {
+            if (id_to_token[id].attr & (LLAMA_TOKEN_ATTR_CONTROL | LLAMA_TOKEN_ATTR_USER_DEFINED | LLAMA_TOKEN_ATTR_UNKNOWN)) {
+                cache_special_tokens.push_back(id);
+            }
+        }
+
+        std::sort(cache_special_tokens.begin(), cache_special_tokens.end(),
+            [&] (const llama_token a, const llama_token b) {
+                return id_to_token[a].text.size() > id_to_token[b].text.size();
+            }
+        );
+
+        LLAMA_LOG_INFO("%s: special tokens cache size = %u\n", __func__, (uint32_t) cache_special_tokens.size());
+    }
+
+    // build token to piece cache
+    {
+        size_t size_cache = 0;
+
+        std::vector cache(n_tokens);
+
+        for (uint32_t id = 0; id < n_tokens; ++id) {
+            cache[id] = token_to_piece_for_cache(id, true);
+
+            size_cache += cache[id].size();
+        }
+
+        std::swap(cache_token_to_piece, cache);
+
+        LLAMA_LOG_INFO("%s: token to piece cache size = %.4f MB\n", __func__, size_cache / 1024.0 / 1024.0);
+    }
+
+    // Handle per token attributes
+    //NOTE: Each model customizes per token attributes.
+    //NOTE: Per token attributes are missing from the GGUF file.
+    //TODO: Extract attributes from GGUF file.
+    {
+        auto _contains_any = [] (const std::string & str, const std::vector & substrs) -> bool {
+            for (const auto & substr : substrs) {
+                if (str.find(substr) < std::string::npos) {
+                    return true;
+                }
+            }
+            return false;
+        };
+
+        auto _set_tokenid_attr = [&] (const llama_token id, llama_token_attr attr, bool value) {
+            uint32_t current = id_to_token.at(id).attr;
+            current = value ? (current | attr) : (current & ~attr);
+            id_to_token[id].attr = (llama_token_attr) current;
+        };
+
+        auto _set_token_attr = [&] (const std::string & token, llama_token_attr attr, bool value) {
+            _set_tokenid_attr(token_to_id.at(token), attr, value);
+        };
+
+        std::string model_name;
+        std::string tokenizer_pre;
+
+        ml.get_key(LLM_KV_GENERAL_NAME,  model_name,    false);
+        ml.get_key(LLM_KV_TOKENIZER_PRE, tokenizer_pre, false);
+
+        // model name to lowercase
+        std::transform(model_name.begin(), model_name.end(), model_name.begin(),
+            [] (const std::string::value_type x) {
+                return std::tolower(x);
+            }
+        );
+
+        // set attributes by model/tokenizer name
+        if (_contains_any(tokenizer_pre, {"jina-v2-de", "jina-v2-es", "jina-v2-code"})) {
+            _set_token_attr("", LLAMA_TOKEN_ATTR_LSTRIP, true);
+        } else if (_contains_any(model_name, {"phi-3", "phi3"})) {
+            for (auto id : cache_special_tokens) {
+                _set_tokenid_attr(id, LLAMA_TOKEN_ATTR_RSTRIP, true);
+            }
+            for (const auto * token : {""}) {
+                _set_token_attr(token, LLAMA_TOKEN_ATTR_RSTRIP, true);
+            }
+            for (const auto * token : {"", "", "<|endoftext|>"}) {
+                _set_token_attr(token, LLAMA_TOKEN_ATTR_RSTRIP, false);
+            }
+        }
+    }
+}
+
+enum llama_vocab_type llama_vocab::impl::get_type() const {
+    return type;
+}
+
+std::string llama_vocab::impl::type_name() const{
+    switch (type) {
+        case LLAMA_VOCAB_TYPE_NONE: return "no vocab";
+        case LLAMA_VOCAB_TYPE_SPM:  return "SPM";
+        case LLAMA_VOCAB_TYPE_BPE:  return "BPE";
+        case LLAMA_VOCAB_TYPE_WPM:  return "WPM";
+        case LLAMA_VOCAB_TYPE_UGM:  return "UGM";
+        case LLAMA_VOCAB_TYPE_RWKV: return "RWKV";
+        default:                    return "unknown";
+    }
+}
+
+bool llama_vocab::impl::is_normal(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_NORMAL;
+}
+
+bool llama_vocab::impl::is_unknown(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_UNKNOWN;
+}
+
+bool llama_vocab::impl::is_control(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_CONTROL;
+}
+
+bool llama_vocab::impl::is_byte(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_BYTE;
+}
+
+bool llama_vocab::impl::is_user_defined(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_USER_DEFINED;
+}
+
+bool llama_vocab::impl::is_unused(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token[id].attr & LLAMA_TOKEN_ATTR_UNUSED;
+}
+
+bool llama_vocab::impl::is_eog(llama_token id) const {
+    return id != LLAMA_TOKEN_NULL && special_eog_ids.count(id) > 0;
+}
+
+uint8_t llama_vocab::impl::token_to_byte(llama_token id) const {
+    GGML_ASSERT(get_type() != LLAMA_VOCAB_TYPE_NONE);
+    GGML_ASSERT(is_byte(id));
+    const auto & token_data = id_to_token.at(id);
+    switch (get_type()) {
+        case LLAMA_VOCAB_TYPE_SPM:
+        case LLAMA_VOCAB_TYPE_UGM: {
+            auto buf = token_data.text.substr(3, 2);
+            return strtol(buf.c_str(), NULL, 16);
+        }
+        case LLAMA_VOCAB_TYPE_BPE: {
+            GGML_ABORT("fatal error");
+        }
+        case LLAMA_VOCAB_TYPE_WPM: {
+            GGML_ABORT("fatal error");
+        }
+        default:
+            GGML_ABORT("fatal error");
+    }
+}
+
+llama_token_attr llama_vocab::impl::token_get_attr(llama_token id) const {
+    GGML_ASSERT(type != LLAMA_VOCAB_TYPE_NONE);
+    return id_to_token.at(id).attr;
+}
+
+void llama_vocab::impl::init_tokenizer(enum llama_vocab_type type) {
+    LLAMA_LOG_DEBUG("%s: initializing tokenizer for type %d\n", __func__, type);
+
+    switch (type) {
+        case LLAMA_VOCAB_TYPE_SPM:
+            tokenizer = std::make_unique(vocab);
+            break;
+        case LLAMA_VOCAB_TYPE_BPE:
+            tokenizer = std::make_unique(vocab);
+            break;
+        case LLAMA_VOCAB_TYPE_WPM:
+            tokenizer = std::make_unique(vocab);
+            break;
+        case LLAMA_VOCAB_TYPE_UGM:
+            tokenizer = std::make_unique(vocab, precompiled_charsmap);
+            break;
+        case LLAMA_VOCAB_TYPE_RWKV:
+            tokenizer = std::make_unique(vocab);
+            break;
+        default:
+            GGML_ABORT("unsupported vocab type");
+    }
+}
+
+//
+// (de-) tokenize
+//
+
 // #define PRETOKENIZERDEBUG
 
-static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list & buffer, bool parse_special) {
+void llama_vocab::impl::tokenizer_st_partition(std::forward_list & buffer, bool parse_special) const {
     // for each special token
-    for (const llama_vocab::id special_id : vocab.cache_special_tokens) {
-        const auto & data = vocab.id_to_token[special_id];
-        const auto & special_token = data.text;
+    for (const llama_token special_id : cache_special_tokens) {
+        const auto & data = vocab.get_token_data(special_id);
+        const auto & text = data.text;
 
         if (!parse_special && (data.attr & (LLAMA_TOKEN_ATTR_CONTROL | LLAMA_TOKEN_ATTR_UNKNOWN))) {
             // Ignore control and unknown tokens when parse_special == false
@@ -1345,13 +2171,13 @@ static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list<
                     // find the first occurrence of a given special token in this fragment
                     //  passing offset argument only limit the "search area" but match coordinates
                     //  are still relative to the source full raw_text
-                    auto match = raw_text.find(special_token, raw_text_base_offset);
+                    auto match = raw_text.find(text, raw_text_base_offset);
 
                     // no occurrences found, stop processing this fragment for a given special token
                     if (match == std::string::npos) break;
 
                     // check if match is within bounds of offset <-> length
-                    if (match + special_token.length() > raw_text_base_offset + raw_text_base_length) break;
+                    if (match + text.length() > raw_text_base_offset + raw_text_base_length) break;
 
 #ifdef PRETOKENIZERDEBUG
                     LLAMA_LOG_WARN("FF: (%ld %ld %ld) '%s'\n", raw_text->length(), raw_text_base_offset, raw_text_base_length, raw_text->substr(raw_text_base_offset, raw_text_base_length).c_str());
@@ -1386,9 +2212,9 @@ static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list<
                     it++;
 
                     // right
-                    if (match + special_token.length() < raw_text_base_offset + raw_text_base_length) {
-                        int64_t right_reminder_offset = match + special_token.length();
-                        int64_t right_reminder_length = raw_text_base_length - ((match - raw_text_base_offset) + special_token.length());
+                    if (match + text.length() < raw_text_base_offset + raw_text_base_length) {
+                        int64_t right_reminder_offset = match + text.length();
+                        int64_t right_reminder_length = raw_text_base_length - ((match - raw_text_base_offset) + text.length());
 
                         if (data.attr & LLAMA_TOKEN_ATTR_RSTRIP) {
                             while (right_reminder_length > 0 && isspace(raw_text[right_reminder_offset])) {
@@ -1409,7 +2235,7 @@ static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list<
                         if (source == 0) {
                             buffer.erase_after(buffer.before_begin());
                         } else {
-                            buffer.erase_after(std::next(buffer.begin(), (source-1)));
+                            buffer.erase_after(std::next(buffer.begin(), (source - 1)));
                         }
 
                         // repeat for the right side
@@ -1423,7 +2249,7 @@ static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list<
                         if (source == 0) {
                             buffer.erase_after(buffer.before_begin());
                         } else {
-                            buffer.erase_after(std::next(buffer.begin(), (source-1)));
+                            buffer.erase_after(std::next(buffer.begin(), (source - 1)));
                         }
                         break;
                     }
@@ -1434,322 +2260,29 @@ static void tokenizer_st_partition(const llama_vocab & vocab, std::forward_list<
     }
 }
 
-std::vector llama_tokenize_internal(
-        const llama_vocab & vocab,
-        std::string raw_text,
-        bool add_special,
-        bool parse_special) {
-    GGML_ASSERT(vocab.tokenizer && "Tokenizer not initialized. Call llama_vocab::init_tokenizer() first.");
-
-    std::vector output;
-    std::forward_list fragment_buffer;
-
-    if (!raw_text.empty()) {
-        fragment_buffer.emplace_front(raw_text, 0, raw_text.length());
-        tokenizer_st_partition(vocab, fragment_buffer, parse_special);
+// NOTE: avoid ever using this except for building the token_to_piece caches
+std::string llama_vocab::impl::token_to_piece_for_cache(llama_token token, bool special) const {
+    std::string piece;
+    piece.resize(piece.capacity());  // using string internal cache
+    const int n_chars = vocab.token_to_piece(token, &piece[0], piece.size(), 0, special);
+    if (n_chars < 0) {
+        piece.resize(-n_chars);
+        int check = vocab.token_to_piece(token, &piece[0], piece.size(), 0, special);
+        GGML_ASSERT(check == -n_chars);
+    }
+    else {
+        piece.resize(n_chars);
     }
 
-    switch (vocab.type) {
-        case LLAMA_VOCAB_TYPE_SPM:
-            {
-                // OG tokenizer behavior:
-                //
-                // tokenizer.encode('', add_special_tokens=True)  returns [1]
-                // tokenizer.encode('', add_special_tokens=False) returns []
-
-                bool is_prev_special = true;  // prefix with space if first token
-
-                if (add_special && vocab.tokenizer_add_bos) {
-                    GGML_ASSERT(vocab.special_bos_id != -1);
-                    output.push_back(vocab.special_bos_id);
-                    is_prev_special = true;
-                }
-
-                for (const auto & fragment : fragment_buffer) {
-                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
-                        auto raw_text = fragment.raw_text.substr(fragment.offset, fragment.length);
-
-                        // prefix with space if previous is special
-                        if (vocab.tokenizer_add_space_prefix && is_prev_special) {
-                            raw_text = " " + raw_text;
-                        }
-
-#ifdef PRETOKENIZERDEBUG
-                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", raw_text.length(), fragment.offset, fragment.length, raw_text.c_str());
-#endif
-                        llama_escape_whitespace(raw_text);
-                        llm_tokenizer_spm_session session(vocab);
-                        session.tokenize(raw_text, output);
-                        is_prev_special = false;
-                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
-                        output.push_back(fragment.token);
-                        is_prev_special = true;
-                    }
-                }
-
-                if (add_special && vocab.tokenizer_add_bos && output.size() >= 2 && output[1] == vocab.special_bos_id) {
-                    LLAMA_LOG_WARN(
-                        "%s: Added a BOS token to the prompt as specified by the model but the prompt "
-                        "also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. "
-                        "Are you sure this is what you want?\n", __FUNCTION__);
-                }
-
-                if (add_special && vocab.tokenizer_add_eos) {
-                    GGML_ASSERT(vocab.special_eos_id != -1);
-                    output.push_back(vocab.special_eos_id);
-                }
-            } break;
-        case LLAMA_VOCAB_TYPE_BPE:
-            {
-                llm_tokenizer_bpe_session session(vocab);
-                // it calls some other methods that are not exist in llm_tokenizer,
-                // here just cast it to bpe tokenizer object
-                if (add_special) {
-                    session.append_bos(output);
-                }
-                for (const auto & fragment : fragment_buffer) {
-                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
-                        auto raw_text = fragment.raw_text.substr(fragment.offset, fragment.length);
-
-#ifdef PRETOKENIZERDEBUG
-                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", raw_text.length(), fragment.offset, fragment.length, raw_text.c_str());
-#endif
-                        session.tokenize(raw_text, output);
-                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
-                        session.append(fragment.token, output);
-                    }
-                }
-
-                if (add_special) {
-                    session.append_eos(output);
-                    session.check_double_bos_eos(output);
-                }
-            } break;
-        case LLAMA_VOCAB_TYPE_WPM:
-            {
-                if (add_special) {
-                    GGML_ASSERT(vocab.special_cls_id != -1);
-                    output.push_back(vocab.special_cls_id);
-                }
-
-                llm_tokenizer_wpm_session session(vocab);
-
-                for (const auto & fragment : fragment_buffer) {
-                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
-                        auto raw_text = fragment.raw_text.substr(fragment.offset, fragment.length);
-
-#ifdef PRETOKENIZERDEBUG
-                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", raw_text.length(), fragment.offset, fragment.length, raw_text.c_str());
-#endif
-                        session.tokenize(raw_text, output);
-                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
-                        output.push_back(fragment.token);
-                    }
-                }
-
-                if (add_special) {
-                    GGML_ASSERT(vocab.special_sep_id != -1);
-                    output.push_back(vocab.special_sep_id);
-                }
-            } break;
-        case LLAMA_VOCAB_TYPE_UGM:
-            {
-                if (add_special && vocab.tokenizer_add_bos) {
-                    GGML_ASSERT(vocab.special_bos_id != -1);
-                    output.push_back(vocab.special_bos_id);
-                }
-                llm_tokenizer_ugm_session session(vocab);
-
-                for (const auto & fragment : fragment_buffer) {
-                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
-                        auto raw_text = fragment.raw_text.substr(fragment.offset, fragment.length);
-#ifdef PRETOKENIZERDEBUG
-                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", raw_text.length(), fragment.offset, fragment.length, raw_text.c_str());
-#endif
-                        session.tokenize(raw_text, output);
-                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
-                        output.push_back(fragment.token);
-                    }
-                }
-
-                if (add_special && vocab.tokenizer_add_bos && output.size() >= 2 && output[1] == vocab.special_bos_id) {
-                    LLAMA_LOG_WARN(
-                        "%s: Added a BOS token to the prompt as specified by the model but the prompt "
-                        "also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. "
-                        "Are you sure this is what you want?\n", __FUNCTION__);
-                }
-
-                if (add_special && vocab.tokenizer_add_eos) {
-                    GGML_ASSERT(vocab.special_eos_id != -1);
-                    output.push_back(vocab.special_eos_id);
-                }
-            } break;
-        case LLAMA_VOCAB_TYPE_RWKV:
-            {
-                llm_tokenizer_rwkv_session session(vocab);
-                for (const auto & fragment : fragment_buffer) {
-                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
-                        auto raw_text = fragment.raw_text.substr(fragment.offset, fragment.length);
-
-#ifdef PRETOKENIZERDEBUG
-                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", raw_text.length(), fragment.offset, fragment.length, raw_text.c_str());
-#endif
-
-                        session.tokenize(raw_text, output);
-                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
-                        output.push_back(fragment.token);
-                    }
-                }
-            } break;
-        case LLAMA_VOCAB_TYPE_NONE:
-            GGML_ABORT("fatal error");
-    }
-
-    return output;
+    return piece;
 }
 
-llama_token llama_byte_to_token_impl(const llama_vocab & vocab, uint8_t ch) {
-    GGML_ASSERT(llama_vocab_get_type(vocab) != LLAMA_VOCAB_TYPE_NONE);
-    static const char * hex = "0123456789ABCDEF";
-    switch (llama_vocab_get_type(vocab)) {
-        case LLAMA_VOCAB_TYPE_SPM:
-        case LLAMA_VOCAB_TYPE_UGM: {
-            const char buf[7] = { '<', '0', 'x', hex[ch >> 4], hex[ch & 15], '>', 0 };
-            auto token = vocab.token_to_id.find(buf);
-            if (token != vocab.token_to_id.end()) {
-                return (*token).second;
-            }
-            // Try to fall back to just the byte as a string
-            const char buf2[2] = { (char)ch, 0 };
-            return vocab.token_to_id.at(buf2);
-        }
-        case LLAMA_VOCAB_TYPE_WPM:
-        case LLAMA_VOCAB_TYPE_BPE: {
-            return vocab.token_to_id.at(unicode_byte_to_utf8(ch));
-        }
-        default:
-            GGML_ABORT("fatal error");
-    }
+static void llama_escape_whitespace(std::string & text) {
+    replace_all(text, " ", "\xe2\x96\x81");
 }
 
-const char * llama_token_get_text_impl(const struct llama_vocab & vocab, llama_token token) {
-    GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE);
-    return vocab.id_to_token[token].text.c_str();
-}
-
-float llama_token_get_score_impl(const struct llama_vocab & vocab, llama_token token) {
-    GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE);
-    return vocab.id_to_token[token].score;
-}
-
-llama_token_attr llama_token_get_attr_impl(const struct llama_vocab & vocab, llama_token token) {
-    GGML_ASSERT(vocab.type != LLAMA_VOCAB_TYPE_NONE);
-    return vocab.id_to_token[token].attr;
-}
-
-bool llama_token_is_eog_impl(const struct llama_vocab & vocab, llama_token token) {
-    return token != -1 && vocab.special_eog_ids.count(token) > 0;
-}
-
-bool llama_token_is_control_impl(const struct llama_vocab & vocab, llama_token token) {
-    return llama_is_control_token(vocab, token);
-}
-
-llama_token llama_token_bos_impl(const struct llama_vocab & vocab) {
-    return vocab.special_bos_id;
-}
-
-llama_token llama_token_eos_impl(const struct llama_vocab & vocab) {
-    return vocab.special_eos_id;
-}
-
-llama_token llama_token_eot_impl(const struct llama_vocab & vocab) {
-    return vocab.special_eot_id;
-}
-
-llama_token llama_token_eom_impl(const struct llama_vocab & vocab) {
-    return vocab.special_eom_id;
-}
-
-llama_token llama_token_cls_impl(const struct llama_vocab & vocab) {
-    return vocab.special_cls_id;
-}
-
-llama_token llama_token_sep_impl(const struct llama_vocab & vocab) {
-    return vocab.special_sep_id;
-}
-
-llama_token llama_token_nl_impl(const struct llama_vocab & vocab) {
-    return vocab.linefeed_id;
-}
-
-llama_token llama_token_pad_impl(const struct llama_vocab & vocab) {
-    return vocab.special_pad_id;
-}
-
-bool llama_add_bos_token_impl(const struct llama_vocab & vocab) {
-    return vocab.tokenizer_add_bos;
-}
-
-bool llama_add_eos_token_impl(const struct llama_vocab & vocab) {
-    return vocab.tokenizer_add_eos;
-}
-
-llama_token llama_token_prefix_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_pre_id;
-}
-
-llama_token llama_token_middle_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_mid_id;
-}
-
-llama_token llama_token_suffix_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_suf_id;
-}
-
-llama_token llama_token_fim_pre_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_pre_id;
-}
-
-llama_token llama_token_fim_suf_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_suf_id;
-}
-
-llama_token llama_token_fim_mid_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_mid_id;
-}
-
-llama_token llama_token_fim_pad_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_pad_id;
-}
-
-llama_token llama_token_fim_rep_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_rep_id;
-}
-
-llama_token llama_token_fim_sep_impl(const struct llama_vocab & vocab) {
-    return vocab.special_fim_sep_id;
-}
-
-int32_t llama_tokenize_impl(
-        const struct llama_vocab & vocab,
-                      const char * text,
-                         int32_t   text_len,
-                     llama_token * tokens,
-                         int32_t   n_tokens_max,
-                            bool   add_special,
-                            bool   parse_special) {
-    auto res = llama_tokenize_internal(vocab, std::string(text, text_len), add_special, parse_special);
-    if (n_tokens_max < (int) res.size()) {
-        // LLAMA_LOG_ERROR("%s: too many tokens\n", __func__);
-        return -((int) res.size());
-    }
-
-    for (size_t i = 0; i < res.size(); i++) {
-        tokens[i] = res[i];
-    }
-
-    return res.size();
+static void llama_unescape_whitespace(std::string & word) {
+    replace_all(word, "\xe2\x96\x81", " ");
 }
 
 static std::string llama_decode_text(const std::string & text) {
@@ -1772,11 +2305,185 @@ static std::string llama_decode_text(const std::string & text) {
     return decoded_text;
 }
 
-// does not write null-terminator to buf
-int32_t llama_token_to_piece_impl(const struct llama_vocab & vocab, llama_token token, char * buf, int32_t length, int32_t lstrip, bool special) {
+std::vector llama_vocab::impl::tokenize(
+        const std::string & raw_text,
+        bool add_special,
+        bool parse_special) const {
+    GGML_ASSERT(tokenizer && "Tokenizer not initialized. Call llama_vocab::init_tokenizer() first.");
+
+    std::vector output;
+    std::forward_list fragment_buffer;
+
+    if (!raw_text.empty()) {
+        fragment_buffer.emplace_front(raw_text, 0, raw_text.length());
+        tokenizer_st_partition(fragment_buffer, parse_special);
+    }
+
+    switch (get_type()) {
+        case LLAMA_VOCAB_TYPE_SPM:
+            {
+                // OG tokenizer behavior:
+                //
+                // tokenizer.encode('', add_special_tokens=True)  returns [1]
+                // tokenizer.encode('', add_special_tokens=False) returns []
+
+                bool is_prev_special = true;  // prefix with space if first token
+
+                if (add_special && add_bos) {
+                    GGML_ASSERT(special_bos_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_bos_id);
+                    is_prev_special = true;
+                }
+
+                for (const auto & fragment : fragment_buffer) {
+                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
+                        std::string text;
+
+                        // prefix with space if previous is special
+                        if (add_space_prefix && is_prev_special) {
+                            text = ' ';
+                        }
+
+                        text += fragment.raw_text.substr(fragment.offset, fragment.length);
+
+#ifdef PRETOKENIZERDEBUG
+                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", text.length(), fragment.offset, fragment.length, text.c_str());
+#endif
+                        llama_escape_whitespace(text);
+                        llm_tokenizer_spm_session session(vocab);
+                        session.tokenize(text, output);
+                        is_prev_special = false;
+                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
+                        output.push_back(fragment.token);
+                        is_prev_special = true;
+                    }
+                }
+
+                if (add_special && add_bos && output.size() >= 2 && output[1] == special_bos_id) {
+                    LLAMA_LOG_WARN(
+                        "%s: Added a BOS token to the prompt as specified by the model but the prompt "
+                        "also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. "
+                        "Are you sure this is what you want?\n", __FUNCTION__);
+                }
+
+                if (add_special && add_eos) {
+                    GGML_ASSERT(special_eos_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_eos_id);
+                }
+            } break;
+        case LLAMA_VOCAB_TYPE_BPE:
+            {
+                llm_tokenizer_bpe_session session(vocab, *static_cast(tokenizer.get()));
+                // it calls some other methods that are not exist in llm_tokenizer,
+                // here just cast it to bpe tokenizer object
+                if (add_special) {
+                    session.append_bos(output);
+                }
+                for (const auto & fragment : fragment_buffer) {
+                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
+                        std::string text = fragment.raw_text.substr(fragment.offset, fragment.length);
+
+#ifdef PRETOKENIZERDEBUG
+                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", text.length(), fragment.offset, fragment.length, text.c_str());
+#endif
+                        session.tokenize(text, output);
+                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
+                        session.append(fragment.token, output);
+                    }
+                }
+
+                if (add_special) {
+                    session.append_eos(output);
+                    session.check_double_bos_eos(output);
+                }
+            } break;
+        case LLAMA_VOCAB_TYPE_WPM:
+            {
+                if (add_special) {
+                    GGML_ASSERT(special_bos_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_bos_id);
+                }
+
+                llm_tokenizer_wpm_session session(vocab);
+
+                for (const auto & fragment : fragment_buffer) {
+                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
+                        std::string text = fragment.raw_text.substr(fragment.offset, fragment.length);
+
+#ifdef PRETOKENIZERDEBUG
+                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", text.length(), fragment.offset, fragment.length, text.c_str());
+#endif
+                        session.tokenize(text, output);
+                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
+                        output.push_back(fragment.token);
+                    }
+                }
+
+                if (add_special) {
+                    GGML_ASSERT(special_sep_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_sep_id);
+                }
+            } break;
+        case LLAMA_VOCAB_TYPE_UGM:
+            {
+                if (add_special && add_bos) {
+                    GGML_ASSERT(special_bos_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_bos_id);
+                }
+                llm_tokenizer_ugm_session session(vocab, *static_cast(tokenizer.get()));
+
+                for (const auto & fragment : fragment_buffer) {
+                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
+                        std::string text = fragment.raw_text.substr(fragment.offset, fragment.length);
+#ifdef PRETOKENIZERDEBUG
+                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", text.length(), fragment.offset, fragment.length, text.c_str());
+#endif
+                        session.tokenize(text, output);
+                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
+                        output.push_back(fragment.token);
+                    }
+                }
+
+                if (add_special && add_bos && output.size() >= 2 && output[1] == special_bos_id) {
+                    LLAMA_LOG_WARN(
+                        "%s: Added a BOS token to the prompt as specified by the model but the prompt "
+                        "also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. "
+                        "Are you sure this is what you want?\n", __FUNCTION__);
+                }
+
+                if (add_special && add_eos) {
+                    GGML_ASSERT(special_eos_id != LLAMA_TOKEN_NULL);
+                    output.push_back(special_eos_id);
+                }
+            } break;
+        case LLAMA_VOCAB_TYPE_RWKV:
+            {
+                llm_tokenizer_rwkv_session session(vocab, *static_cast(tokenizer.get()));
+                for (const auto & fragment : fragment_buffer) {
+                    if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_RAW_TEXT) {
+                        std::string text = fragment.raw_text.substr(fragment.offset, fragment.length);
+
+#ifdef PRETOKENIZERDEBUG
+                        LLAMA_LOG_WARN("TT: (%ld %ld %ld) '%s'\n", text.length(), fragment.offset, fragment.length, text.c_str());
+#endif
+
+                        session.tokenize(text, output);
+                    } else { // if (fragment.type == FRAGMENT_BUFFER_VARIANT_TYPE_TOKEN)
+                        output.push_back(fragment.token);
+                    }
+                }
+            } break;
+        case LLAMA_VOCAB_TYPE_NONE:
+            GGML_ABORT("fatal error");
+    }
+
+    return output;
+}
+
+int32_t llama_vocab::impl::token_to_piece(llama_token token, char * buf, int32_t length, int32_t lstrip, bool special) const {
     // ref: https://github.com/ggerganov/llama.cpp/pull/7587#discussion_r1620983843
     static const int attr_special = LLAMA_TOKEN_ATTR_UNKNOWN | LLAMA_TOKEN_ATTR_CONTROL;
-    const llama_token_attr attr = llama_token_get_attr_impl(vocab, token);
+    const llama_token_attr attr = token_get_attr(token);
     if (!special && (attr & attr_special)) {
         return 0;
     }
@@ -1797,7 +2504,7 @@ int32_t llama_token_to_piece_impl(const struct llama_vocab & vocab, llama_token
 
     // if we have a cache - use it
     {
-        const auto & cache = vocab.cache_token_to_piece;
+        const auto & cache = cache_token_to_piece;
 
         if (!cache.empty()) {
             const auto & result = cache.at(token);
@@ -1805,9 +2512,9 @@ int32_t llama_token_to_piece_impl(const struct llama_vocab & vocab, llama_token
         }
     }
 
-    if (0 <= token && token < (int32_t) vocab.id_to_token.size()) {
-        const std::string & token_text = vocab.id_to_token[token].text;
-        switch (llama_vocab_get_type(vocab)) {
+    if (0 <= token && token < (int32_t) id_to_token.size()) {
+        const std::string & token_text = id_to_token[token].text;
+        switch (get_type()) {
             case LLAMA_VOCAB_TYPE_WPM:
             case LLAMA_VOCAB_TYPE_SPM:
             case LLAMA_VOCAB_TYPE_UGM: {
@@ -1822,7 +2529,7 @@ int32_t llama_token_to_piece_impl(const struct llama_vocab & vocab, llama_token
                     return _try_copy(result.data(), result.size());
                 }
                 if (attr & LLAMA_TOKEN_ATTR_BYTE) {
-                    char byte = (char) llama_token_to_byte(vocab, token);
+                    char byte = (char) token_to_byte(token);
                     return _try_copy((char*) &byte, 1);
                 }
                 break;
@@ -1858,39 +2565,46 @@ int32_t llama_token_to_piece_impl(const struct llama_vocab & vocab, llama_token
     return 0;
 }
 
-int32_t llama_detokenize_impl(
-        const struct llama_vocab & vocab,
+const std::string & llama_vocab::impl::token_to_piece(llama_token token) const {
+    return cache_token_to_piece.at(token);
+}
+
+int32_t llama_vocab::impl::detokenize(
                const llama_token * tokens,
                          int32_t   n_tokens,
                             char * text,
                          int32_t   text_len_max,
                             bool   remove_special,
-                            bool   unparse_special) {
-    GGML_ASSERT(vocab.tokenizer && "Tokenizer not initialized. Call llama_vocab::init_tokenizer() first.");
+                            bool   unparse_special) const {
+    if (type == LLAMA_VOCAB_TYPE_NONE) {
+        return 0;
+    }
+
+    GGML_ASSERT(tokenizer && "Tokenizer not initialized. Call llama_vocab::init_tokenizer() first.");
 
     int32_t avail = text_len_max;
     int32_t total = 0;
 
     // remove the leading space
-    bool remove_space = vocab.tokenizer_add_space_prefix;
+    bool remove_space = add_space_prefix;
 
-    if (remove_special && vocab.tokenizer_add_bos) {
-        if (n_tokens > 0 && tokens[0] == vocab.special_bos_id) {
+    if (remove_special && add_bos) {
+        if (n_tokens > 0 && tokens[0] == special_bos_id) {
             remove_space = false;
             n_tokens--;
             tokens++;
         }
     }
 
-    if (remove_special && vocab.tokenizer_add_eos) {
-        if (n_tokens > 0 && tokens[n_tokens-1] == vocab.special_eos_id) {
+    if (remove_special && add_eos) {
+        if (n_tokens > 0 && tokens[n_tokens - 1] == special_eos_id) {
             n_tokens--;
         }
     }
 
     for (int32_t i = 0; i < n_tokens; ++i) {
         GGML_ASSERT(avail >= 0);
-        int32_t n_chars = llama_token_to_piece_impl(vocab, tokens[i], text, avail, remove_space, unparse_special);
+        int32_t n_chars = token_to_piece(tokens[i], text, avail, remove_space, unparse_special);
         remove_space = false;
         if (n_chars < 0) {
             avail = 0;
@@ -1906,7 +2620,7 @@ int32_t llama_detokenize_impl(
         return -total;
     }
 
-    if (vocab.tokenizer_clean_spaces) {
+    if (clean_spaces) {
         text -= total;  // restart text
 
         // first pass: characters ?!.,  //TODO: where do these characters come from?
@@ -1967,13 +2681,321 @@ int32_t llama_detokenize_impl(
     return total <= text_len_max ? total : -total;
 }
 
-std::string llama_detokenize(const struct llama_vocab & vocab, const std::vector & tokens, bool special) {
+void llama_vocab::impl::print_info() const {
+    LLAMA_LOG_INFO("%s: vocab type       = %s\n",     __func__, type_name().c_str());
+    LLAMA_LOG_INFO("%s: n_vocab          = %u\n",     __func__, vocab.n_tokens());
+    LLAMA_LOG_INFO("%s: n_merges         = %u\n",     __func__, (uint32_t) bpe_ranks.size());
+
+    // special tokens
+    if (special_bos_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: BOS token        = %d '%s'\n", __func__, special_bos_id,     id_to_token[special_bos_id].text.c_str() );  }
+    if (special_eos_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: EOS token        = %d '%s'\n", __func__, special_eos_id,     id_to_token[special_eos_id].text.c_str() );  }
+    if (special_eot_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: EOT token        = %d '%s'\n", __func__, special_eot_id,     id_to_token[special_eot_id].text.c_str() );  }
+    if (special_eom_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: EOM token        = %d '%s'\n", __func__, special_eom_id,     id_to_token[special_eom_id].text.c_str() );  }
+    if (special_unk_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: UNK token        = %d '%s'\n", __func__, special_unk_id,     id_to_token[special_unk_id].text.c_str() );  }
+    if (special_sep_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: SEP token        = %d '%s'\n", __func__, special_sep_id,     id_to_token[special_sep_id].text.c_str() );  }
+    if (special_pad_id  != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: PAD token        = %d '%s'\n", __func__, special_pad_id,     id_to_token[special_pad_id].text.c_str() );  }
+    if (special_mask_id != LLAMA_TOKEN_NULL)    { LLAMA_LOG_INFO( "%s: MASK token       = %d '%s'\n", __func__, special_mask_id,    id_to_token[special_mask_id].text.c_str() ); }
+
+    if (linefeed_id != LLAMA_TOKEN_NULL)        { LLAMA_LOG_INFO( "%s: LF token         = %d '%s'\n", __func__, linefeed_id,        id_to_token[linefeed_id].text.c_str() ); }
+
+    if (special_fim_pre_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM PRE token    = %d '%s'\n", __func__, special_fim_pre_id, id_to_token[special_fim_pre_id].text.c_str() ); }
+    if (special_fim_suf_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM SUF token    = %d '%s'\n", __func__, special_fim_suf_id, id_to_token[special_fim_suf_id].text.c_str() ); }
+    if (special_fim_mid_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM MID token    = %d '%s'\n", __func__, special_fim_mid_id, id_to_token[special_fim_mid_id].text.c_str() ); }
+    if (special_fim_pad_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM PAD token    = %d '%s'\n", __func__, special_fim_pad_id, id_to_token[special_fim_pad_id].text.c_str() ); }
+    if (special_fim_rep_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM REP token    = %d '%s'\n", __func__, special_fim_rep_id, id_to_token[special_fim_rep_id].text.c_str() ); }
+    if (special_fim_sep_id != LLAMA_TOKEN_NULL) { LLAMA_LOG_INFO( "%s: FIM SEP token    = %d '%s'\n", __func__, special_fim_sep_id, id_to_token[special_fim_sep_id].text.c_str() ); }
+
+    for (const auto & id : special_eog_ids) {
+        LLAMA_LOG_INFO( "%s: EOG token        = %d '%s'\n", __func__, id, id_to_token[id].text.c_str() );
+    }
+
+    LLAMA_LOG_INFO("%s: max token length = %d\n", __func__, max_token_len);
+}
+
+llama_vocab::llama_vocab() : pimpl(new impl(*this)) {
+}
+
+llama_vocab::~llama_vocab() {
+}
+
+void llama_vocab::load(llama_model_loader & ml, const LLM_KV & kv) {
+    pimpl->load(ml, kv);
+}
+
+enum llama_vocab_type llama_vocab::get_type() const {
+    return pimpl->type;
+}
+
+enum llama_vocab_pre_type llama_vocab::get_pre_type() const {
+    return pimpl->pre_type;
+}
+
+uint32_t llama_vocab::n_tokens() const {
+    return (uint32_t) pimpl->id_to_token.size();
+}
+
+uint32_t llama_vocab::n_token_types() const {
+    return (uint32_t) pimpl->n_token_types;
+}
+
+std::string llama_vocab::type_name() const{
+    return pimpl->type_name();
+}
+
+bool llama_vocab::is_normal(llama_token id) const {
+    return pimpl->is_normal(id);
+}
+
+bool llama_vocab::is_unknown(llama_token id) const {
+    return pimpl->is_unknown(id);
+}
+
+bool llama_vocab::is_control(llama_token id) const {
+    return pimpl->is_control(id);
+}
+
+bool llama_vocab::is_byte(llama_token id) const {
+    return pimpl->is_byte(id);
+}
+
+bool llama_vocab::is_user_defined(llama_token id) const {
+    return pimpl->is_user_defined(id);
+}
+
+bool llama_vocab::is_unused(llama_token id) const {
+    return pimpl->is_unused(id);
+}
+
+bool llama_vocab::is_eog(llama_token id) const {
+    return pimpl->is_eog(id);
+}
+
+uint8_t llama_vocab::token_to_byte(llama_token id) const {
+    return pimpl->token_to_byte(id);
+}
+
+llama_token llama_vocab::byte_to_token(uint8_t ch) const {
+    GGML_ASSERT(get_type() != LLAMA_VOCAB_TYPE_NONE);
+    static const char * hex = "0123456789ABCDEF";
+    switch (get_type()) {
+        case LLAMA_VOCAB_TYPE_SPM:
+        case LLAMA_VOCAB_TYPE_UGM: {
+            const char buf[7] = { '<', '0', 'x', hex[ch >> 4], hex[ch & 15], '>', 0 };
+            auto token = pimpl->token_to_id.find(buf);
+            if (token != pimpl->token_to_id.end()) {
+                return (*token).second;
+            }
+            // Try to fall back to just the byte as a string
+            const char buf2[2] = { (char)ch, 0 };
+            return pimpl->token_to_id.at(buf2);
+        }
+        case LLAMA_VOCAB_TYPE_WPM:
+        case LLAMA_VOCAB_TYPE_BPE: {
+            return pimpl->token_to_id.at(unicode_byte_to_utf8(ch));
+        }
+        default:
+            GGML_ABORT("fatal error");
+    }
+}
+
+llama_token llama_vocab::text_to_token(const std::string & text) const {
+    GGML_ASSERT(pimpl->type != LLAMA_VOCAB_TYPE_NONE);
+    auto it = pimpl->token_to_id.find(text);
+    if (it != pimpl->token_to_id.end()) {
+        return (*it).second;
+    }
+    return LLAMA_TOKEN_NULL;
+}
+
+const llama_vocab::token_data & llama_vocab::get_token_data(llama_token id) const {
+    GGML_ASSERT(pimpl->type != LLAMA_VOCAB_TYPE_NONE);
+    return pimpl->id_to_token.at(id);
+}
+
+const char * llama_vocab::token_get_text(llama_token id) const {
+    GGML_ASSERT(pimpl->type != LLAMA_VOCAB_TYPE_NONE);
+    return pimpl->id_to_token.at(id).text.c_str();
+}
+
+float llama_vocab::token_get_score(llama_token id) const {
+    GGML_ASSERT(pimpl->type != LLAMA_VOCAB_TYPE_NONE);
+    return pimpl->id_to_token.at(id).score;
+}
+
+llama_token_attr llama_vocab::token_get_attr(llama_token id) const {
+    return pimpl->token_get_attr(id);
+}
+
+llama_token llama_vocab::token_bos() const {
+    return pimpl->special_bos_id;
+}
+
+llama_token llama_vocab::token_eos() const {
+    return pimpl->special_eos_id;
+}
+
+llama_token llama_vocab::token_eot() const {
+    return pimpl->special_eot_id;
+}
+
+llama_token llama_vocab::token_eom() const {
+    return pimpl->special_eom_id;
+}
+
+llama_token llama_vocab::token_unk() const {
+    return pimpl->special_unk_id;
+}
+
+llama_token llama_vocab::token_sep() const {
+    return pimpl->special_sep_id;
+}
+
+llama_token llama_vocab::token_nl() const {
+    return pimpl->linefeed_id;
+}
+
+llama_token llama_vocab::token_pad() const {
+    return pimpl->special_pad_id;
+}
+
+llama_token llama_vocab::token_prefix() const {
+    return pimpl->special_fim_pre_id;
+}
+
+llama_token llama_vocab::token_middle() const {
+    return pimpl->special_fim_mid_id;
+}
+
+llama_token llama_vocab::token_suffix() const {
+    return pimpl->special_fim_suf_id;
+}
+
+llama_token llama_vocab::token_fim_pre() const {
+    return pimpl->special_fim_pre_id;
+}
+
+llama_token llama_vocab::token_fim_suf() const {
+    return pimpl->special_fim_suf_id;
+}
+
+llama_token llama_vocab::token_fim_mid() const {
+    return pimpl->special_fim_mid_id;
+}
+
+llama_token llama_vocab::token_fim_pad() const {
+    return pimpl->special_fim_pad_id;
+}
+
+llama_token llama_vocab::token_fim_rep() const {
+    return pimpl->special_fim_rep_id;
+}
+
+llama_token llama_vocab::token_fim_sep() const {
+    return pimpl->special_fim_sep_id;
+}
+
+bool llama_vocab::get_add_space_prefix() const {
+    return pimpl->add_space_prefix;
+}
+
+bool llama_vocab::get_add_bos() const {
+    return pimpl->add_bos;
+}
+
+bool llama_vocab::get_add_eos() const {
+    return pimpl->add_eos;
+}
+
+bool llama_vocab::get_ignore_merges() const {
+    return pimpl->ignore_merges;
+}
+
+bool llama_vocab::get_clean_spaces() const {
+    return pimpl->clean_spaces;
+}
+
+bool llama_vocab::get_remove_extra_whitespaces() const {
+    return pimpl->remove_extra_whitespaces;
+}
+
+bool llama_vocab::get_escape_whitespaces() const {
+    return pimpl->escape_whitespaces;
+}
+
+bool llama_vocab::get_treat_whitespace_as_suffix() const {
+    return pimpl->treat_whitespace_as_suffix;
+}
+
+int llama_vocab::max_token_len() const {
+    return pimpl->max_token_len;
+}
+
+int llama_vocab::find_bpe_rank(const std::string & token_left, const std::string & token_right) const {
+    GGML_ASSERT(token_left.find(' ')   == std::string::npos);
+    GGML_ASSERT(token_left.find('\n')  == std::string::npos);
+    GGML_ASSERT(token_right.find(' ')  == std::string::npos);
+    GGML_ASSERT(token_right.find('\n') == std::string::npos);
+
+    auto it = pimpl->bpe_ranks.find(std::make_pair(token_left, token_right));
+    if (it == pimpl->bpe_ranks.end()) {
+        return -1;
+    }
+
+    return it->second;
+}
+
+int32_t llama_vocab::tokenize(
+                  const char * text,
+                     int32_t   text_len,
+                 llama_token * tokens,
+                     int32_t   n_tokens_max,
+                        bool   add_special,
+                        bool   parse_special) const {
+    auto res = tokenize(std::string(text, text_len), add_special, parse_special);
+    if (n_tokens_max < (int) res.size()) {
+        // LLAMA_LOG_ERROR("%s: too many tokens\n", __func__);
+        return -((int) res.size());
+    }
+
+    for (size_t i = 0; i < res.size(); i++) {
+        tokens[i] = res[i];
+    }
+
+    return res.size();
+}
+
+std::vector llama_vocab::tokenize(
+        const std::string & raw_text,
+        bool add_special,
+        bool parse_special) const {
+    return pimpl->tokenize(raw_text, add_special, parse_special);
+}
+
+const std::string & llama_vocab::token_to_piece(llama_token token) const {
+    return pimpl->token_to_piece(token);
+}
+
+int32_t llama_vocab::token_to_piece(llama_token token, char * buf, int32_t length, int32_t lstrip, bool special) const {
+    return pimpl->token_to_piece(token, buf, length, lstrip, special);
+}
+
+int32_t llama_vocab::detokenize(
+               const llama_token * tokens,
+                         int32_t   n_tokens,
+                            char * text,
+                         int32_t   text_len_max,
+                            bool   remove_special,
+                            bool   unparse_special) const {
+    return pimpl->detokenize(tokens, n_tokens, text, text_len_max, remove_special, unparse_special);
+}
+
+std::string llama_vocab::detokenize(const std::vector & tokens, bool special) const {
     std::string text;
     text.resize(std::max(text.capacity(), tokens.size()));
-    int32_t n_chars = llama_detokenize_impl(vocab, tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special);
+    int32_t n_chars = detokenize(tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special);
     if (n_chars < 0) {
         text.resize(-n_chars);
-        n_chars = llama_detokenize_impl(vocab, tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special);
+        n_chars = detokenize(tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special);
         GGML_ASSERT(n_chars <= (int32_t)text.size());  // whitespace trimming is performed after per-token detokenization
     }
 
@@ -1982,3 +3004,243 @@ std::string llama_detokenize(const struct llama_vocab & vocab, const std::vector
     // NOTE: the original tokenizer decodes bytes after collecting the pieces.
     return text;
 }
+
+void llama_vocab::print_info() const {
+    pimpl->print_info();
+}
+
+//
+// interface implementation
+//
+
+int32_t llama_vocab_n_tokens(const struct llama_vocab * vocab) {
+    return vocab->n_tokens();
+}
+
+// deprecated
+int32_t llama_n_vocab(const struct llama_vocab * vocab) {
+    return llama_vocab_n_tokens(vocab);
+}
+
+enum llama_vocab_type llama_vocab_type(const struct llama_vocab * vocab) {
+    return vocab->get_type();
+}
+
+const char * llama_vocab_get_text(const struct llama_vocab * vocab, llama_token token) {
+    return vocab->token_get_text(token);
+}
+
+float llama_vocab_get_score(const struct llama_vocab * vocab, llama_token token) {
+    return vocab->token_get_score(token);
+}
+
+enum llama_token_attr llama_vocab_get_attr(const struct llama_vocab * vocab, llama_token token) {
+    return vocab->token_get_attr(token);
+}
+
+bool llama_vocab_is_eog(const struct llama_vocab * vocab, llama_token token) {
+    return vocab->is_eog(token);
+}
+
+bool llama_vocab_is_control(const struct llama_vocab * vocab, llama_token token) {
+    return vocab->is_control(token);
+}
+
+llama_token llama_vocab_bos(const struct llama_vocab * vocab) {
+    return vocab->token_bos();
+}
+
+llama_token llama_vocab_eos(const struct llama_vocab * vocab) {
+    return vocab->token_eos();
+}
+
+llama_token llama_vocab_eot(const struct llama_vocab * vocab) {
+    return vocab->token_eot();
+}
+
+// deprecated
+llama_token llama_vocab_cls(const struct llama_vocab * vocab) {
+    return vocab->token_bos();
+}
+
+llama_token llama_vocab_sep(const struct llama_vocab * vocab) {
+    return vocab->token_sep();
+}
+
+llama_token llama_vocab_nl (const struct llama_vocab * vocab) {
+    return vocab->token_nl();
+}
+
+llama_token llama_vocab_pad(const struct llama_vocab * vocab) {
+    return vocab->token_pad();
+}
+
+bool llama_vocab_get_add_bos(const struct llama_vocab * vocab) {
+    return vocab->get_add_bos();
+}
+
+bool llama_vocab_get_add_eos(const struct llama_vocab * vocab) {
+    return vocab->get_add_eos();
+}
+
+llama_token llama_vocab_fim_pre(const struct llama_vocab * vocab) {
+    return vocab->token_fim_pre();
+}
+
+llama_token llama_vocab_fim_suf(const struct llama_vocab * vocab) {
+    return vocab->token_fim_suf();
+}
+
+llama_token llama_vocab_fim_mid(const struct llama_vocab * vocab) {
+    return vocab->token_fim_mid();
+}
+
+llama_token llama_vocab_fim_pad(const struct llama_vocab * vocab) {
+    return vocab->token_fim_pad();
+}
+
+llama_token llama_vocab_fim_rep(const struct llama_vocab * vocab) {
+    return vocab->token_fim_rep();
+}
+
+llama_token llama_vocab_fim_sep(const struct llama_vocab * vocab) {
+    return vocab->token_fim_sep();
+}
+
+// deprecated
+const char * llama_token_get_text(const struct llama_vocab * vocab, llama_token token) {
+    return llama_vocab_get_text(vocab, token);
+}
+
+// deprecated
+float llama_token_get_score(const struct llama_vocab * vocab, llama_token token) {
+    return llama_vocab_get_score(vocab, token);
+}
+
+// deprecated
+enum llama_token_attr llama_token_get_attr(const struct llama_vocab * vocab, llama_token token) {
+    return llama_vocab_get_attr(vocab, token);
+}
+
+// deprecated
+bool llama_token_is_eog(const struct llama_vocab * vocab, llama_token token) {
+    return llama_vocab_is_eog(vocab, token);
+}
+
+// deprecated
+bool llama_token_is_control(const struct llama_vocab * vocab, llama_token token) {
+    return llama_vocab_is_control(vocab, token);
+}
+
+// deprecated
+llama_token llama_token_bos(const struct llama_vocab * vocab) {
+    return llama_vocab_bos(vocab);
+}
+
+// deprecated
+llama_token llama_token_eos(const struct llama_vocab * vocab) {
+    return llama_vocab_eos(vocab);
+}
+
+// deprecated
+llama_token llama_token_eot(const struct llama_vocab * vocab) {
+    return llama_vocab_eot(vocab);
+}
+
+// deprecated
+llama_token llama_token_cls(const struct llama_vocab * vocab) {
+    //return llama_vocab_cls(vocab);
+    return llama_vocab_bos(vocab); // avoid deprecation warning
+}
+
+// deprecated
+llama_token llama_token_sep(const struct llama_vocab * vocab) {
+    return llama_vocab_sep(vocab);
+}
+
+// deprecated
+llama_token llama_token_nl (const struct llama_vocab * vocab) {
+    return llama_vocab_nl(vocab);
+}
+
+// deprecated
+llama_token llama_token_pad(const struct llama_vocab * vocab) {
+    return llama_vocab_pad(vocab);
+}
+
+// deprecated
+bool llama_add_bos_token(const struct llama_vocab * vocab) {
+    return llama_vocab_get_add_bos(vocab);
+}
+
+// deprecated
+bool llama_add_eos_token(const struct llama_vocab * vocab) {
+    return llama_vocab_get_add_eos(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_pre(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_pre(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_suf(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_suf(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_mid(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_mid(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_pad(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_pad(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_rep(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_rep(vocab);
+}
+
+// deprecated
+llama_token llama_token_fim_sep(const struct llama_vocab * vocab) {
+    return llama_vocab_fim_sep(vocab);
+}
+
+//
+// tokenization
+//
+
+int32_t llama_tokenize(
+    const struct llama_vocab * vocab,
+                  const char * text,
+                     int32_t   text_len,
+                 llama_token * tokens,
+                     int32_t   n_tokens_max,
+                        bool   add_special,
+                        bool   parse_special) {
+    return vocab->tokenize(text, text_len, tokens, n_tokens_max, add_special, parse_special);
+}
+
+int32_t llama_token_to_piece(
+    const struct llama_vocab * vocab,
+                 llama_token   token,
+                        char * buf,
+                     int32_t   length,
+                     int32_t   lstrip,
+                        bool   special) {
+    return vocab->token_to_piece(token, buf, length, lstrip, special);
+}
+
+int32_t llama_detokenize(
+    const struct llama_vocab * vocab,
+           const llama_token * tokens,
+                     int32_t   n_tokens,
+                        char * text,
+                     int32_t   text_len_max,
+                        bool   remove_special,
+                        bool   unparse_special) {
+    return vocab->detokenize(tokens, n_tokens, text, text_len_max, remove_special, unparse_special);
+}
+
diff --git a/src/llama-vocab.h b/src/llama-vocab.h
index 4bb16d2e4..5ce355214 100644
--- a/src/llama-vocab.h
+++ b/src/llama-vocab.h
@@ -1,170 +1,125 @@
 #pragma once
 
-#include "llama-impl.h"
+#include "llama.h"
 
 #include 
 #include 
-#include 
-#include 
-#include 
+#include 
 
-struct llm_tokenizer;
+struct LLM_KV;
+struct llama_model_loader;
 
 struct llama_vocab {
-    using id    = llama_token;
-    using token = std::string;
-    using tattr = llama_token_attr;
-
     struct token_data {
-        token text;
-        float score;
-        tattr attr;
+        std::string      text;
+        float            score;
+        llama_token_attr attr;
     };
 
-    uint32_t n_vocab = 0; // TODO: not great because has to keep in sync with hparams.n_vocab
-
-    enum llama_vocab_type     type     = LLAMA_VOCAB_TYPE_SPM;
-    enum llama_vocab_pre_type type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-
-    int max_token_len = 0; // used for optimizing longest token search
-
-    std::unordered_map token_to_id;
-    std::vector       id_to_token;
-
-    std::vector    cache_special_tokens;
-    std::vector cache_token_to_piece; // llama_token_to_piece(special = true);
-
-    std::map, int> bpe_ranks;
-
-    // default LLaMA special tokens
-    // TODO: should we set all of these to LLAMA_TOKEN_NULL?
-    id special_bos_id  = 1;
-    id special_eos_id  = 2;
-    id special_eot_id  = LLAMA_TOKEN_NULL;
-    id special_eom_id  = LLAMA_TOKEN_NULL;
-    id special_unk_id  = 0;
-    id special_sep_id  = LLAMA_TOKEN_NULL;
-    id special_pad_id  = LLAMA_TOKEN_NULL;
-    id special_cls_id  = LLAMA_TOKEN_NULL;
-    id special_mask_id = LLAMA_TOKEN_NULL;
-
-    id linefeed_id = 13;
-
-    // fim tokens
-    id special_fim_pre_id = LLAMA_TOKEN_NULL;
-    id special_fim_suf_id = LLAMA_TOKEN_NULL;
-    id special_fim_mid_id = LLAMA_TOKEN_NULL;
-    id special_fim_pad_id = LLAMA_TOKEN_NULL;
-    id special_fim_rep_id = LLAMA_TOKEN_NULL; // repo
-    id special_fim_sep_id = LLAMA_TOKEN_NULL; // file separator
-
-    // set of all tokens that cause "end of generation"
-    std::set special_eog_ids;
-
-    // tokenizer flags
-    bool tokenizer_add_space_prefix           = false;
-    bool tokenizer_add_bos                    = false;
-    bool tokenizer_add_eos                    = false;
-    bool tokenizer_ignore_merges              = false;
-    bool tokenizer_clean_spaces               = false;  // clean_up_tokenization_spaces
-    bool tokenizer_remove_extra_whitespaces   = false;
-    bool tokenizer_escape_whitespaces         = true;
-    bool tokenizer_treat_whitespace_as_suffix = false;
-
-    std::vector precompiled_charsmap;
-
-    llm_tokenizer * tokenizer = nullptr;
-
-    llama_vocab() = default;
+    llama_vocab();
     ~llama_vocab();
 
+    void load(llama_model_loader & ml, const LLM_KV & kv);
+
+    enum llama_vocab_type     get_type()     const;
+    enum llama_vocab_pre_type get_pre_type() const;
+
+    uint32_t n_tokens() const;
+    uint32_t n_token_types() const;
+
+    std::string type_name() const;
+
+    bool is_normal      (llama_token id) const;
+    bool is_unknown     (llama_token id) const;
+    bool is_control     (llama_token id) const;
+    bool is_byte        (llama_token id) const;
+    bool is_user_defined(llama_token id) const;
+    bool is_unused      (llama_token id) const;
+    bool is_eog         (llama_token id) const;
+
+    uint8_t     token_to_byte(llama_token id) const;
+    llama_token byte_to_token(uint8_t ch)     const;
+
+    llama_token text_to_token(const std::string & text) const;
+
+    const token_data & get_token_data(llama_token id) const;
+
+    const char *     token_get_text (llama_token id) const;
+    float            token_get_score(llama_token id) const;
+    llama_token_attr token_get_attr (llama_token id) const;
+
+    llama_token token_bos() const;
+    llama_token token_eos() const;
+    llama_token token_eot() const;
+    llama_token token_eom() const;
+    llama_token token_unk() const;
+    llama_token token_sep() const;
+    llama_token token_nl () const;
+    llama_token token_pad() const;
+
+    llama_token token_prefix() const;
+    llama_token token_middle() const;
+    llama_token token_suffix() const;
+
+    llama_token token_fim_pre() const;
+    llama_token token_fim_suf() const;
+    llama_token token_fim_mid() const;
+    llama_token token_fim_pad() const;
+    llama_token token_fim_rep() const;
+    llama_token token_fim_sep() const;
+
+    bool get_add_space_prefix          () const;
+    bool get_add_bos                   () const;
+    bool get_add_eos                   () const;
+    bool get_ignore_merges             () const;
+    bool get_clean_spaces              () const;
+    bool get_remove_extra_whitespaces  () const;
+    bool get_escape_whitespaces        () const;
+    bool get_treat_whitespace_as_suffix() const;
+
+    int max_token_len() const;
+
     int find_bpe_rank(const std::string & token_left, const std::string & token_right) const;
 
-    void init_tokenizer();
+    int32_t tokenize(
+                   const char * text,
+                      int32_t   text_len,
+                  llama_token * tokens,
+                      int32_t   n_tokens_max,
+                         bool   add_special,
+                         bool   parse_special) const;
+
+    std::vector tokenize(
+            const std::string & raw_text,
+                         bool   add_special,
+                         bool   parse_special = false) const;
+
+    // does not write null-terminator to buf
+    int32_t token_to_piece(
+                  llama_token   token,
+                         char * buf,
+                      int32_t   length,
+                      int32_t   lstrip,
+                         bool   special) const;
+
+    // use cached data
+    const std::string & token_to_piece(llama_token token) const;
+
+    int32_t detokenize(
+            const llama_token * tokens,
+                      int32_t   n_tokens,
+                         char * text,
+                      int32_t   text_len_max,
+                         bool   remove_special,
+                         bool   unparse_special) const;
+
+    std::string detokenize(
+            const std::vector & tokens,
+                                      bool   special) const;
+
+    void print_info() const;
+
+private:
+    struct impl;
+    std::unique_ptr pimpl;
 };
-
-//
-// internal API
-//
-
-// TODO: rename to llama_tokenize_impl
-// TODO: This should probably be in llama.h
-std::vector llama_tokenize_internal(
-        const llama_vocab & vocab,
-        std::string raw_text,
-        bool add_special,
-        bool parse_special = false);
-
-// TODO: move the API below as member functions of llama_vocab
-llama_token llama_byte_to_token_impl(const llama_vocab & vocab, uint8_t ch);
-
-const char * llama_token_get_text_impl(const struct llama_vocab & vocab, llama_token token);
-
-float llama_token_get_score_impl(const struct llama_vocab & vocab, llama_token token);
-
-llama_token_attr llama_token_get_attr_impl(const struct llama_vocab & vocab, llama_token token);
-
-bool llama_token_is_eog_impl(const struct llama_vocab & vocab, llama_token token);
-
-bool llama_token_is_control_impl(const struct llama_vocab & vocab, llama_token token);
-
-llama_token llama_token_bos_impl(const struct llama_vocab & vocab);
-llama_token llama_token_eos_impl(const struct llama_vocab & vocab);
-llama_token llama_token_eot_impl(const struct llama_vocab & vocab);
-llama_token llama_token_eom_impl(const struct llama_vocab & vocab);
-llama_token llama_token_cls_impl(const struct llama_vocab & vocab);
-llama_token llama_token_sep_impl(const struct llama_vocab & vocab);
-llama_token llama_token_nl_impl (const struct llama_vocab & vocab);
-llama_token llama_token_pad_impl(const struct llama_vocab & vocab);
-
-llama_token llama_token_prefix_impl(const struct llama_vocab & vocab);
-llama_token llama_token_middle_impl(const struct llama_vocab & vocab);
-llama_token llama_token_suffix_impl(const struct llama_vocab & vocab);
-
-llama_token llama_token_fim_pre_impl(const struct llama_vocab & vocab);
-llama_token llama_token_fim_suf_impl(const struct llama_vocab & vocab);
-llama_token llama_token_fim_mid_impl(const struct llama_vocab & vocab);
-llama_token llama_token_fim_pad_impl(const struct llama_vocab & vocab);
-llama_token llama_token_fim_rep_impl(const struct llama_vocab & vocab);
-llama_token llama_token_fim_sep_impl(const struct llama_vocab & vocab);
-
-bool llama_add_bos_token_impl(const struct llama_vocab & vocab);
-bool llama_add_eos_token_impl(const struct llama_vocab & vocab);
-
-int32_t llama_tokenize_impl(
-        const struct llama_vocab & vocab,
-                      const char * text,
-                         int32_t   text_len,
-                     llama_token * tokens,
-                         int32_t   n_tokens_max,
-                            bool   add_special,
-                            bool   parse_special);
-
-// does not write null-terminator to buf
-int32_t llama_token_to_piece_impl(
-        const struct llama_vocab & vocab,
-                     llama_token   token,
-                            char * buf,
-                         int32_t   length,
-                         int32_t   lstrip,
-                            bool   special);
-
-// check if token0 is contained as a prefix in token1
-bool llama_token_is_prefix_impl(
-        const struct llama_vocab & vocab,
-                     llama_token   token0,
-                     llama_token   token1);
-
-int32_t llama_detokenize_impl(
-        const struct llama_vocab & vocab,
-               const llama_token * tokens,
-                         int32_t   n_tokens,
-                            char * text,
-                         int32_t   text_len_max,
-                            bool   remove_special,
-                            bool   unparse_special);
-
-std::string llama_detokenize(
-        const struct llama_vocab & vocab,
-  const std::vector & tokens,
-                            bool   special);
diff --git a/src/llama.cpp b/src/llama.cpp
index 034441e1f..2e391b3b6 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -1,9208 +1,76 @@
 #include "llama-impl.h"
+
+#include "llama-chat.h"
+#include "llama-mmap.h"
+#include "llama-context.h"
 #include "llama-vocab.h"
 #include "llama-sampling.h"
-
-#include "unicode.h"
+#include "llama-kv-cache.h"
+#include "llama-model-loader.h"
+#include "llama-model.h"
 
 #include "ggml.h"
 #include "ggml-alloc.h"
 #include "ggml-backend.h"
 #include "ggml-cpp.h"
 
-// TODO: replace with ggml API call
-#define QK_K 256
-
-#ifdef __has_include
-    #if __has_include()
-        #include 
-        #if defined(_POSIX_MAPPED_FILES)
-            #include 
-            #include 
-        #endif
-        #if defined(_POSIX_MEMLOCK_RANGE)
-            #include 
-        #endif
-    #endif
-#endif
-
-#if defined(_WIN32)
-    #define WIN32_LEAN_AND_MEAN
-    #ifndef NOMINMAX
-        #define NOMINMAX
-    #endif
-    #include 
-    #ifndef PATH_MAX
-        #define PATH_MAX MAX_PATH
-    #endif
-    #include 
-#endif
-
-#if __cplusplus >= 202000L
-    #define LU8(x) (const char*)(u8##x)
-#else
-    #define LU8(x) u8##x
-#endif
-
 #include 
 #include 
 #include 
-#include 
 #include 
-#include 
-#include 
 #include 
-#include 
 #include 
 #include 
 #include 
 #include 
 #include 
-#include 
 #include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
 
 #if defined(_MSC_VER)
 #pragma warning(disable: 4244 4267) // possible loss of data
 #endif
 
-// bump if necessary
-#define LLAMA_MAX_LAYERS  512
-#define LLAMA_MAX_EXPERTS 160  // DeepSeekV2
-
-//
-// helpers
-//
-
-// trim whitespace from the beginning and end of a string
-static std::string trim(const std::string & str) {
-    size_t start = 0;
-    size_t end = str.size();
-    while (start < end && isspace(str[start])) {
-        start += 1;
-    }
-    while (end > start && isspace(str[end - 1])) {
-        end -= 1;
-    }
-    return str.substr(start, end - start);
-}
-
-static bool is_float_close(float a, float b, float abs_tol) {
-    // Check for non-negative tolerance
-    if (abs_tol < 0.0) {
-        throw std::invalid_argument("Tolerance must be non-negative");
-    }
-
-    // Exact equality check
-    if (a == b) {
-        return true;
-    }
-
-    // Check for infinities
-    if (std::isinf(a) || std::isinf(b)) {
-        return false;
-    }
-
-    // Regular comparison using the provided absolute tolerance
-    return std::fabs(b - a) <= abs_tol;
-}
-
-static void zeros(std::ofstream & file, size_t n) {
-    char zero = 0;
-    for (size_t i = 0; i < n; ++i) {
-        file.write(&zero, 1);
-    }
-}
-
-LLAMA_ATTRIBUTE_FORMAT(1, 2)
-static std::string format(const char * fmt, ...) {
-    va_list ap;
-    va_list ap2;
-    va_start(ap, fmt);
-    va_copy(ap2, ap);
-    int size = vsnprintf(NULL, 0, fmt, ap);
-    GGML_ASSERT(size >= 0 && size < INT_MAX); // NOLINT
-    std::vector buf(size + 1);
-    int size2 = vsnprintf(buf.data(), size + 1, fmt, ap2);
-    GGML_ASSERT(size2 == size);
-    va_end(ap2);
-    va_end(ap);
-    return std::string(buf.data(), size);
-}
-
-//
-// gguf constants (sync with gguf.py)
-//
-
-enum llm_arch {
-    LLM_ARCH_LLAMA,
-    LLM_ARCH_FALCON,
-    LLM_ARCH_BAICHUAN,
-    LLM_ARCH_GROK,
-    LLM_ARCH_GPT2,
-    LLM_ARCH_GPTJ,
-    LLM_ARCH_GPTNEOX,
-    LLM_ARCH_MPT,
-    LLM_ARCH_STARCODER,
-    LLM_ARCH_REFACT,
-    LLM_ARCH_BERT,
-    LLM_ARCH_NOMIC_BERT,
-    LLM_ARCH_JINA_BERT_V2,
-    LLM_ARCH_BLOOM,
-    LLM_ARCH_STABLELM,
-    LLM_ARCH_QWEN,
-    LLM_ARCH_QWEN2,
-    LLM_ARCH_QWEN2MOE,
-    LLM_ARCH_PHI2,
-    LLM_ARCH_PHI3,
-    LLM_ARCH_PLAMO,
-    LLM_ARCH_CODESHELL,
-    LLM_ARCH_ORION,
-    LLM_ARCH_INTERNLM2,
-    LLM_ARCH_MINICPM,
-    LLM_ARCH_MINICPM3,
-    LLM_ARCH_GEMMA,
-    LLM_ARCH_GEMMA2,
-    LLM_ARCH_STARCODER2,
-    LLM_ARCH_MAMBA,
-    LLM_ARCH_XVERSE,
-    LLM_ARCH_COMMAND_R,
-    LLM_ARCH_DBRX,
-    LLM_ARCH_OLMO,
-    LLM_ARCH_OLMOE,
-    LLM_ARCH_OPENELM,
-    LLM_ARCH_ARCTIC,
-    LLM_ARCH_DEEPSEEK2,
-    LLM_ARCH_CHATGLM,
-    LLM_ARCH_BITNET,
-    LLM_ARCH_T5,
-    LLM_ARCH_T5ENCODER,
-    LLM_ARCH_JAIS,
-    LLM_ARCH_NEMOTRON,
-    LLM_ARCH_EXAONE,
-    LLM_ARCH_RWKV6,
-    LLM_ARCH_GRANITE,
-    LLM_ARCH_GRANITE_MOE,
-    LLM_ARCH_CHAMELEON,
-    LLM_ARCH_UNKNOWN,
-};
-
-static const std::map LLM_ARCH_NAMES = {
-    { LLM_ARCH_LLAMA,           "llama"        },
-    { LLM_ARCH_FALCON,          "falcon"       },
-    { LLM_ARCH_GROK,            "grok"         },
-    { LLM_ARCH_GPT2,            "gpt2"         },
-    { LLM_ARCH_GPTJ,            "gptj"         },
-    { LLM_ARCH_GPTNEOX,         "gptneox"      },
-    { LLM_ARCH_MPT,             "mpt"          },
-    { LLM_ARCH_BAICHUAN,        "baichuan"     },
-    { LLM_ARCH_STARCODER,       "starcoder"    },
-    { LLM_ARCH_REFACT,          "refact"       },
-    { LLM_ARCH_BERT,            "bert"         },
-    { LLM_ARCH_NOMIC_BERT,      "nomic-bert"   },
-    { LLM_ARCH_JINA_BERT_V2,    "jina-bert-v2" },
-    { LLM_ARCH_BLOOM,           "bloom"        },
-    { LLM_ARCH_STABLELM,        "stablelm"     },
-    { LLM_ARCH_QWEN,            "qwen"         },
-    { LLM_ARCH_QWEN2,           "qwen2"        },
-    { LLM_ARCH_QWEN2MOE,        "qwen2moe"     },
-    { LLM_ARCH_PHI2,            "phi2"         },
-    { LLM_ARCH_PHI3,            "phi3"         },
-    { LLM_ARCH_PLAMO,           "plamo"        },
-    { LLM_ARCH_CODESHELL,       "codeshell"    },
-    { LLM_ARCH_ORION,           "orion"        },
-    { LLM_ARCH_INTERNLM2,       "internlm2"    },
-    { LLM_ARCH_MINICPM,         "minicpm"      },
-    { LLM_ARCH_MINICPM3,        "minicpm3"     },
-    { LLM_ARCH_GEMMA,           "gemma"        },
-    { LLM_ARCH_GEMMA2,          "gemma2"       },
-    { LLM_ARCH_STARCODER2,      "starcoder2"   },
-    { LLM_ARCH_MAMBA,           "mamba"        },
-    { LLM_ARCH_XVERSE,          "xverse"       },
-    { LLM_ARCH_COMMAND_R,       "command-r"    },
-    { LLM_ARCH_DBRX,            "dbrx"         },
-    { LLM_ARCH_OLMO,            "olmo"         },
-    { LLM_ARCH_OLMOE,           "olmoe"        },
-    { LLM_ARCH_OPENELM,         "openelm"      },
-    { LLM_ARCH_ARCTIC,          "arctic"       },
-    { LLM_ARCH_DEEPSEEK2,       "deepseek2"    },
-    { LLM_ARCH_CHATGLM,         "chatglm"      },
-    { LLM_ARCH_BITNET,          "bitnet"       },
-    { LLM_ARCH_T5,              "t5"           },
-    { LLM_ARCH_T5ENCODER,       "t5encoder"    },
-    { LLM_ARCH_JAIS,            "jais"         },
-    { LLM_ARCH_NEMOTRON,        "nemotron"     },
-    { LLM_ARCH_EXAONE,          "exaone"       },
-    { LLM_ARCH_RWKV6,           "rwkv6"        },
-    { LLM_ARCH_GRANITE,         "granite"      },
-    { LLM_ARCH_GRANITE_MOE,     "granitemoe"   },
-    { LLM_ARCH_CHAMELEON,       "chameleon"    },
-    { LLM_ARCH_UNKNOWN,         "(unknown)"    },
-};
-
-enum llm_kv {
-    LLM_KV_GENERAL_TYPE,
-    LLM_KV_GENERAL_ARCHITECTURE,
-    LLM_KV_GENERAL_QUANTIZATION_VERSION,
-    LLM_KV_GENERAL_ALIGNMENT,
-    LLM_KV_GENERAL_NAME,
-    LLM_KV_GENERAL_AUTHOR,
-    LLM_KV_GENERAL_VERSION,
-    LLM_KV_GENERAL_URL,
-    LLM_KV_GENERAL_DESCRIPTION,
-    LLM_KV_GENERAL_LICENSE,
-    LLM_KV_GENERAL_SOURCE_URL,
-    LLM_KV_GENERAL_SOURCE_HF_REPO,
-
-    LLM_KV_VOCAB_SIZE,
-    LLM_KV_CONTEXT_LENGTH,
-    LLM_KV_EMBEDDING_LENGTH,
-    LLM_KV_BLOCK_COUNT,
-    LLM_KV_LEADING_DENSE_BLOCK_COUNT,
-    LLM_KV_FEED_FORWARD_LENGTH,
-    LLM_KV_EXPERT_FEED_FORWARD_LENGTH,
-    LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH,
-    LLM_KV_USE_PARALLEL_RESIDUAL,
-    LLM_KV_TENSOR_DATA_LAYOUT,
-    LLM_KV_EXPERT_COUNT,
-    LLM_KV_EXPERT_USED_COUNT,
-    LLM_KV_EXPERT_SHARED_COUNT,
-    LLM_KV_EXPERT_WEIGHTS_SCALE,
-    LLM_KV_POOLING_TYPE,
-    LLM_KV_LOGIT_SCALE,
-    LLM_KV_DECODER_START_TOKEN_ID,
-    LLM_KV_ATTN_LOGIT_SOFTCAPPING,
-    LLM_KV_FINAL_LOGIT_SOFTCAPPING,
-    LLM_KV_SWIN_NORM,
-    LLM_KV_RESCALE_EVERY_N_LAYERS,
-    LLM_KV_TIME_MIX_EXTRA_DIM,
-    LLM_KV_TIME_DECAY_EXTRA_DIM,
-    LLM_KV_RESIDUAL_SCALE,
-    LLM_KV_EMBEDDING_SCALE,
-
-    LLM_KV_ATTENTION_HEAD_COUNT,
-    LLM_KV_ATTENTION_HEAD_COUNT_KV,
-    LLM_KV_ATTENTION_MAX_ALIBI_BIAS,
-    LLM_KV_ATTENTION_CLAMP_KQV,
-    LLM_KV_ATTENTION_KEY_LENGTH,
-    LLM_KV_ATTENTION_VALUE_LENGTH,
-    LLM_KV_ATTENTION_LAYERNORM_EPS,
-    LLM_KV_ATTENTION_LAYERNORM_RMS_EPS,
-    LLM_KV_ATTENTION_CAUSAL,
-    LLM_KV_ATTENTION_Q_LORA_RANK,
-    LLM_KV_ATTENTION_KV_LORA_RANK,
-    LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT,
-    LLM_KV_ATTENTION_SLIDING_WINDOW,
-    LLM_KV_ATTENTION_SCALE,
-
-    LLM_KV_ROPE_DIMENSION_COUNT,
-    LLM_KV_ROPE_FREQ_BASE,
-    LLM_KV_ROPE_SCALE_LINEAR,
-    LLM_KV_ROPE_SCALING_TYPE,
-    LLM_KV_ROPE_SCALING_FACTOR,
-    LLM_KV_ROPE_SCALING_ATTN_FACTOR,
-    LLM_KV_ROPE_SCALING_ORIG_CTX_LEN,
-    LLM_KV_ROPE_SCALING_FINETUNED,
-    LLM_KV_ROPE_SCALING_YARN_LOG_MUL,
-
-    LLM_KV_SPLIT_NO,
-    LLM_KV_SPLIT_COUNT,
-    LLM_KV_SPLIT_TENSORS_COUNT,
-
-    LLM_KV_SSM_INNER_SIZE,
-    LLM_KV_SSM_CONV_KERNEL,
-    LLM_KV_SSM_STATE_SIZE,
-    LLM_KV_SSM_TIME_STEP_RANK,
-    LLM_KV_SSM_DT_B_C_RMS,
-
-    LLM_KV_WKV_HEAD_SIZE,
-
-    LLM_KV_TOKENIZER_MODEL,
-    LLM_KV_TOKENIZER_PRE,
-    LLM_KV_TOKENIZER_LIST,
-    LLM_KV_TOKENIZER_TOKEN_TYPE,
-    LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT,
-    LLM_KV_TOKENIZER_SCORES,
-    LLM_KV_TOKENIZER_MERGES,
-    LLM_KV_TOKENIZER_BOS_ID,
-    LLM_KV_TOKENIZER_EOS_ID,
-    LLM_KV_TOKENIZER_EOT_ID,
-    LLM_KV_TOKENIZER_EOM_ID,
-    LLM_KV_TOKENIZER_UNK_ID,
-    LLM_KV_TOKENIZER_SEP_ID,
-    LLM_KV_TOKENIZER_PAD_ID,
-    LLM_KV_TOKENIZER_CLS_ID,
-    LLM_KV_TOKENIZER_MASK_ID,
-    LLM_KV_TOKENIZER_ADD_BOS,
-    LLM_KV_TOKENIZER_ADD_EOS,
-    LLM_KV_TOKENIZER_ADD_PREFIX,
-    LLM_KV_TOKENIZER_REMOVE_EXTRA_WS,
-    LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP,
-    LLM_KV_TOKENIZER_HF_JSON,
-    LLM_KV_TOKENIZER_RWKV,
-    LLM_KV_TOKENIZER_FIM_PRE_ID,
-    LLM_KV_TOKENIZER_FIM_SUF_ID,
-    LLM_KV_TOKENIZER_FIM_MID_ID,
-    LLM_KV_TOKENIZER_FIM_PAD_ID,
-    LLM_KV_TOKENIZER_FIM_REP_ID,
-    LLM_KV_TOKENIZER_FIM_SEP_ID,
-
-    LLM_KV_ADAPTER_TYPE,
-    LLM_KV_ADAPTER_LORA_ALPHA,
-
-    // deprecated:
-    LLM_KV_TOKENIZER_PREFIX_ID,
-    LLM_KV_TOKENIZER_SUFFIX_ID,
-    LLM_KV_TOKENIZER_MIDDLE_ID,
-};
-
-static const std::map LLM_KV_NAMES = {
-    { LLM_KV_GENERAL_TYPE,                  "general.type"                          },
-    { LLM_KV_GENERAL_ARCHITECTURE,          "general.architecture"                  },
-    { LLM_KV_GENERAL_QUANTIZATION_VERSION,  "general.quantization_version"          },
-    { LLM_KV_GENERAL_ALIGNMENT,             "general.alignment"                     },
-    { LLM_KV_GENERAL_NAME,                  "general.name"                          },
-    { LLM_KV_GENERAL_AUTHOR,                "general.author"                        },
-    { LLM_KV_GENERAL_VERSION,               "general.version"                       },
-    { LLM_KV_GENERAL_URL,                   "general.url"                           },
-    { LLM_KV_GENERAL_DESCRIPTION,           "general.description"                   },
-    { LLM_KV_GENERAL_LICENSE,               "general.license"                       },
-    { LLM_KV_GENERAL_SOURCE_URL,            "general.source.url"                    },
-    { LLM_KV_GENERAL_SOURCE_HF_REPO,        "general.source.huggingface.repository" },
-
-    { LLM_KV_VOCAB_SIZE,                        "%s.vocab_size"                        },
-    { LLM_KV_CONTEXT_LENGTH,                    "%s.context_length"                    },
-    { LLM_KV_EMBEDDING_LENGTH,                  "%s.embedding_length"                  },
-    { LLM_KV_BLOCK_COUNT,                       "%s.block_count"                       },
-    { LLM_KV_LEADING_DENSE_BLOCK_COUNT,         "%s.leading_dense_block_count"         },
-    { LLM_KV_FEED_FORWARD_LENGTH,               "%s.feed_forward_length"               },
-    { LLM_KV_EXPERT_FEED_FORWARD_LENGTH,        "%s.expert_feed_forward_length"        },
-    { LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, "%s.expert_shared_feed_forward_length" },
-    { LLM_KV_USE_PARALLEL_RESIDUAL,             "%s.use_parallel_residual"             },
-    { LLM_KV_TENSOR_DATA_LAYOUT,                "%s.tensor_data_layout"                },
-    { LLM_KV_EXPERT_COUNT,                      "%s.expert_count"                      },
-    { LLM_KV_EXPERT_USED_COUNT,                 "%s.expert_used_count"                 },
-    { LLM_KV_EXPERT_SHARED_COUNT,               "%s.expert_shared_count"               },
-    { LLM_KV_EXPERT_WEIGHTS_SCALE,              "%s.expert_weights_scale"              },
-    { LLM_KV_POOLING_TYPE,                      "%s.pooling_type"                      },
-    { LLM_KV_LOGIT_SCALE,                       "%s.logit_scale"                       },
-    { LLM_KV_DECODER_START_TOKEN_ID,            "%s.decoder_start_token_id"            },
-    { LLM_KV_ATTN_LOGIT_SOFTCAPPING,            "%s.attn_logit_softcapping"            },
-    { LLM_KV_FINAL_LOGIT_SOFTCAPPING,           "%s.final_logit_softcapping"           },
-    { LLM_KV_SWIN_NORM,                         "%s.swin_norm"                         },
-    { LLM_KV_RESCALE_EVERY_N_LAYERS,            "%s.rescale_every_n_layers"            },
-    { LLM_KV_TIME_MIX_EXTRA_DIM,                "%s.time_mix_extra_dim"                },
-    { LLM_KV_TIME_DECAY_EXTRA_DIM,              "%s.time_decay_extra_dim"              },
-    { LLM_KV_RESIDUAL_SCALE,                    "%s.residual_scale"                    },
-    { LLM_KV_EMBEDDING_SCALE,                   "%s.embedding_scale"                   },
-
-    { LLM_KV_ATTENTION_HEAD_COUNT,             "%s.attention.head_count"             },
-    { LLM_KV_ATTENTION_HEAD_COUNT_KV,          "%s.attention.head_count_kv"          },
-    { LLM_KV_ATTENTION_MAX_ALIBI_BIAS,         "%s.attention.max_alibi_bias"         },
-    { LLM_KV_ATTENTION_CLAMP_KQV,              "%s.attention.clamp_kqv"              },
-    { LLM_KV_ATTENTION_KEY_LENGTH,             "%s.attention.key_length"             },
-    { LLM_KV_ATTENTION_VALUE_LENGTH,           "%s.attention.value_length"           },
-    { LLM_KV_ATTENTION_LAYERNORM_EPS,          "%s.attention.layer_norm_epsilon"     },
-    { LLM_KV_ATTENTION_LAYERNORM_RMS_EPS,      "%s.attention.layer_norm_rms_epsilon" },
-    { LLM_KV_ATTENTION_CAUSAL,                 "%s.attention.causal"                 },
-    { LLM_KV_ATTENTION_Q_LORA_RANK,            "%s.attention.q_lora_rank"            },
-    { LLM_KV_ATTENTION_KV_LORA_RANK,           "%s.attention.kv_lora_rank"           },
-    { LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, "%s.attention.relative_buckets_count" },
-    { LLM_KV_ATTENTION_SLIDING_WINDOW,         "%s.attention.sliding_window"         },
-    { LLM_KV_ATTENTION_SCALE,                  "%s.attention.scale"                  },
-
-    { LLM_KV_ROPE_DIMENSION_COUNT,             "%s.rope.dimension_count"                 },
-    { LLM_KV_ROPE_FREQ_BASE,                   "%s.rope.freq_base"                       },
-    { LLM_KV_ROPE_SCALE_LINEAR,                "%s.rope.scale_linear"                    },
-    { LLM_KV_ROPE_SCALING_TYPE,                "%s.rope.scaling.type"                    },
-    { LLM_KV_ROPE_SCALING_FACTOR,              "%s.rope.scaling.factor"                  },
-    { LLM_KV_ROPE_SCALING_ATTN_FACTOR,         "%s.rope.scaling.attn_factor"             },
-    { LLM_KV_ROPE_SCALING_ORIG_CTX_LEN,        "%s.rope.scaling.original_context_length" },
-    { LLM_KV_ROPE_SCALING_FINETUNED,           "%s.rope.scaling.finetuned"               },
-    { LLM_KV_ROPE_SCALING_YARN_LOG_MUL,        "%s.rope.scaling.yarn_log_multiplier"     },
-
-    { LLM_KV_SPLIT_NO,                         "split.no"            },
-    { LLM_KV_SPLIT_COUNT,                      "split.count"         },
-    { LLM_KV_SPLIT_TENSORS_COUNT,              "split.tensors.count" },
-
-    { LLM_KV_SSM_CONV_KERNEL,                  "%s.ssm.conv_kernel"    },
-    { LLM_KV_SSM_INNER_SIZE,                   "%s.ssm.inner_size"     },
-    { LLM_KV_SSM_STATE_SIZE,                   "%s.ssm.state_size"     },
-    { LLM_KV_SSM_TIME_STEP_RANK,               "%s.ssm.time_step_rank" },
-    { LLM_KV_SSM_DT_B_C_RMS,                   "%s.ssm.dt_b_c_rms"     },
-
-    { LLM_KV_WKV_HEAD_SIZE,                    "%s.wkv.head_size" },
-
-    { LLM_KV_TOKENIZER_MODEL,                  "tokenizer.ggml.model"                    },
-    { LLM_KV_TOKENIZER_PRE,                    "tokenizer.ggml.pre"                      },
-    { LLM_KV_TOKENIZER_LIST,                   "tokenizer.ggml.tokens"                   },
-    { LLM_KV_TOKENIZER_TOKEN_TYPE,             "tokenizer.ggml.token_type"               },
-    { LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT,       "tokenizer.ggml.token_type_count"         },
-    { LLM_KV_TOKENIZER_SCORES,                 "tokenizer.ggml.scores"                   },
-    { LLM_KV_TOKENIZER_MERGES,                 "tokenizer.ggml.merges"                   },
-    { LLM_KV_TOKENIZER_BOS_ID,                 "tokenizer.ggml.bos_token_id"             },
-    { LLM_KV_TOKENIZER_EOS_ID,                 "tokenizer.ggml.eos_token_id"             },
-    { LLM_KV_TOKENIZER_EOT_ID,                 "tokenizer.ggml.eot_token_id"             },
-    { LLM_KV_TOKENIZER_EOM_ID,                 "tokenizer.ggml.eom_token_id"             },
-    { LLM_KV_TOKENIZER_UNK_ID,                 "tokenizer.ggml.unknown_token_id"         },
-    { LLM_KV_TOKENIZER_SEP_ID,                 "tokenizer.ggml.seperator_token_id"       },
-    { LLM_KV_TOKENIZER_PAD_ID,                 "tokenizer.ggml.padding_token_id"         },
-    { LLM_KV_TOKENIZER_CLS_ID,                 "tokenizer.ggml.cls_token_id"             },
-    { LLM_KV_TOKENIZER_MASK_ID,                "tokenizer.ggml.mask_token_id"            },
-    { LLM_KV_TOKENIZER_ADD_BOS,                "tokenizer.ggml.add_bos_token"            },
-    { LLM_KV_TOKENIZER_ADD_EOS,                "tokenizer.ggml.add_eos_token"            },
-    { LLM_KV_TOKENIZER_ADD_PREFIX,             "tokenizer.ggml.add_space_prefix"         },
-    { LLM_KV_TOKENIZER_REMOVE_EXTRA_WS,        "tokenizer.ggml.remove_extra_whitespaces" },
-    { LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP,   "tokenizer.ggml.precompiled_charsmap"     },
-    { LLM_KV_TOKENIZER_HF_JSON,                "tokenizer.huggingface.json"              },
-    { LLM_KV_TOKENIZER_RWKV,                   "tokenizer.rwkv.world"                    },
-    { LLM_KV_TOKENIZER_FIM_PRE_ID,             "tokenizer.ggml.fim_pre_token_id"         },
-    { LLM_KV_TOKENIZER_FIM_SUF_ID,             "tokenizer.ggml.fim_suf_token_id"         },
-    { LLM_KV_TOKENIZER_FIM_MID_ID,             "tokenizer.ggml.fim_mid_token_id"         },
-    { LLM_KV_TOKENIZER_FIM_PAD_ID,             "tokenizer.ggml.fim_pad_token_id"         },
-    { LLM_KV_TOKENIZER_FIM_REP_ID,             "tokenizer.ggml.fim_rep_token_id"         },
-    { LLM_KV_TOKENIZER_FIM_SEP_ID,             "tokenizer.ggml.fim_sep_token_id"         },
-
-    { LLM_KV_ADAPTER_TYPE,                     "adapter.type"       },
-    { LLM_KV_ADAPTER_LORA_ALPHA,               "adapter.lora.alpha" },
-
-    // deprecated
-    { LLM_KV_TOKENIZER_PREFIX_ID,              "tokenizer.ggml.prefix_token_id" },
-    { LLM_KV_TOKENIZER_SUFFIX_ID,              "tokenizer.ggml.suffix_token_id" },
-    { LLM_KV_TOKENIZER_MIDDLE_ID,              "tokenizer.ggml.middle_token_id" },
-};
-
-struct LLM_KV {
-    LLM_KV(llm_arch arch) : arch(arch) {}
-
-    llm_arch arch;
-
-    std::string operator()(llm_kv kv) const {
-        return ::format(LLM_KV_NAMES.at(kv), LLM_ARCH_NAMES.at(arch));
-    }
-};
-
-enum llm_tensor {
-    LLM_TENSOR_TOKEN_EMBD,
-    LLM_TENSOR_TOKEN_EMBD_NORM,
-    LLM_TENSOR_TOKEN_TYPES,
-    LLM_TENSOR_POS_EMBD,
-    LLM_TENSOR_OUTPUT,
-    LLM_TENSOR_OUTPUT_NORM,
-    LLM_TENSOR_ROPE_FREQS,
-    LLM_TENSOR_ROPE_FACTORS_LONG,
-    LLM_TENSOR_ROPE_FACTORS_SHORT,
-    LLM_TENSOR_ATTN_Q,
-    LLM_TENSOR_ATTN_K,
-    LLM_TENSOR_ATTN_V,
-    LLM_TENSOR_ATTN_QKV,
-    LLM_TENSOR_ATTN_OUT,
-    LLM_TENSOR_ATTN_NORM,
-    LLM_TENSOR_ATTN_NORM_2,
-    LLM_TENSOR_ATTN_OUT_NORM,
-    LLM_TENSOR_ATTN_POST_NORM,
-    LLM_TENSOR_ATTN_ROT_EMBD,
-    LLM_TENSOR_FFN_GATE_INP,
-    LLM_TENSOR_FFN_GATE_INP_SHEXP,
-    LLM_TENSOR_FFN_NORM,
-    LLM_TENSOR_FFN_POST_NORM,
-    LLM_TENSOR_FFN_GATE,
-    LLM_TENSOR_FFN_DOWN,
-    LLM_TENSOR_FFN_UP,
-    LLM_TENSOR_FFN_ACT,
-    LLM_TENSOR_FFN_DOWN_EXP,  // split experts for backward compatibility
-    LLM_TENSOR_FFN_GATE_EXP,
-    LLM_TENSOR_FFN_UP_EXP,
-    LLM_TENSOR_FFN_NORM_EXPS,
-    LLM_TENSOR_FFN_DOWN_EXPS, // merged experts
-    LLM_TENSOR_FFN_GATE_EXPS,
-    LLM_TENSOR_FFN_UP_EXPS,
-    LLM_TENSOR_FFN_DOWN_SHEXP,
-    LLM_TENSOR_FFN_GATE_SHEXP,
-    LLM_TENSOR_FFN_UP_SHEXP,
-    LLM_TENSOR_ATTN_Q_NORM,
-    LLM_TENSOR_ATTN_K_NORM,
-    LLM_TENSOR_LAYER_OUT_NORM,
-    LLM_TENSOR_SSM_IN,
-    LLM_TENSOR_SSM_CONV1D,
-    LLM_TENSOR_SSM_X,
-    LLM_TENSOR_SSM_DT,
-    LLM_TENSOR_SSM_A,
-    LLM_TENSOR_SSM_D,
-    LLM_TENSOR_SSM_OUT,
-    LLM_TENSOR_TIME_MIX_W1,
-    LLM_TENSOR_TIME_MIX_W2,
-    LLM_TENSOR_TIME_MIX_LERP_X,
-    LLM_TENSOR_TIME_MIX_LERP_W,
-    LLM_TENSOR_TIME_MIX_LERP_K,
-    LLM_TENSOR_TIME_MIX_LERP_V,
-    LLM_TENSOR_TIME_MIX_LERP_R,
-    LLM_TENSOR_TIME_MIX_LERP_G,
-    LLM_TENSOR_TIME_MIX_FIRST,
-    LLM_TENSOR_TIME_MIX_DECAY,
-    LLM_TENSOR_TIME_MIX_DECAY_W1,
-    LLM_TENSOR_TIME_MIX_DECAY_W2,
-    LLM_TENSOR_TIME_MIX_KEY,
-    LLM_TENSOR_TIME_MIX_VALUE,
-    LLM_TENSOR_TIME_MIX_RECEPTANCE,
-    LLM_TENSOR_TIME_MIX_GATE,
-    LLM_TENSOR_TIME_MIX_LN,
-    LLM_TENSOR_TIME_MIX_OUTPUT,
-    LLM_TENSOR_CHANNEL_MIX_LERP_K,
-    LLM_TENSOR_CHANNEL_MIX_LERP_R,
-    LLM_TENSOR_CHANNEL_MIX_KEY,
-    LLM_TENSOR_CHANNEL_MIX_RECEPTANCE,
-    LLM_TENSOR_CHANNEL_MIX_VALUE,
-    LLM_TENSOR_ATTN_Q_A,
-    LLM_TENSOR_ATTN_Q_B,
-    LLM_TENSOR_ATTN_KV_A_MQA,
-    LLM_TENSOR_ATTN_KV_B,
-    LLM_TENSOR_ATTN_Q_A_NORM,
-    LLM_TENSOR_ATTN_KV_A_NORM,
-    LLM_TENSOR_ATTN_SUB_NORM,
-    LLM_TENSOR_FFN_SUB_NORM,
-    LLM_TENSOR_DEC_ATTN_NORM,
-    LLM_TENSOR_DEC_ATTN_Q,
-    LLM_TENSOR_DEC_ATTN_K,
-    LLM_TENSOR_DEC_ATTN_V,
-    LLM_TENSOR_DEC_ATTN_OUT,
-    LLM_TENSOR_DEC_ATTN_REL_B,
-    LLM_TENSOR_DEC_CROSS_ATTN_NORM,
-    LLM_TENSOR_DEC_CROSS_ATTN_Q,
-    LLM_TENSOR_DEC_CROSS_ATTN_K,
-    LLM_TENSOR_DEC_CROSS_ATTN_V,
-    LLM_TENSOR_DEC_CROSS_ATTN_OUT,
-    LLM_TENSOR_DEC_CROSS_ATTN_REL_B,
-    LLM_TENSOR_DEC_FFN_NORM,
-    LLM_TENSOR_DEC_FFN_GATE,
-    LLM_TENSOR_DEC_FFN_DOWN,
-    LLM_TENSOR_DEC_FFN_UP,
-    LLM_TENSOR_DEC_OUTPUT_NORM,
-    LLM_TENSOR_ENC_ATTN_NORM,
-    LLM_TENSOR_ENC_ATTN_Q,
-    LLM_TENSOR_ENC_ATTN_K,
-    LLM_TENSOR_ENC_ATTN_V,
-    LLM_TENSOR_ENC_ATTN_OUT,
-    LLM_TENSOR_ENC_ATTN_REL_B,
-    LLM_TENSOR_ENC_FFN_NORM,
-    LLM_TENSOR_ENC_FFN_GATE,
-    LLM_TENSOR_ENC_FFN_DOWN,
-    LLM_TENSOR_ENC_FFN_UP,
-    LLM_TENSOR_ENC_OUTPUT_NORM,
-    LLM_TENSOR_CLS,
-    LLM_TENSOR_CLS_OUT,
-};
-
-static const std::map> LLM_TENSOR_NAMES = {
-    {
-        LLM_ARCH_LLAMA,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_GATE_EXP,    "blk.%d.ffn_gate.%d" },
-            { LLM_TENSOR_FFN_DOWN_EXP,    "blk.%d.ffn_down.%d" },
-            { LLM_TENSOR_FFN_UP_EXP,      "blk.%d.ffn_up.%d" },
-            { LLM_TENSOR_FFN_GATE_EXPS,   "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,   "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
-        },
-    },
-    {
-        LLM_ARCH_BAICHUAN,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_FALCON,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_NORM_2,     "blk.%d.attn_norm_2" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_GROK,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE_EXP,    "blk.%d.ffn_gate.%d" },
-            { LLM_TENSOR_FFN_DOWN_EXP,    "blk.%d.ffn_down.%d" },
-            { LLM_TENSOR_FFN_UP_EXP,      "blk.%d.ffn_up.%d" },
-            { LLM_TENSOR_FFN_GATE_EXPS,   "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,   "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
-            { LLM_TENSOR_LAYER_OUT_NORM,  "blk.%d.layer_output_norm" },
-            { LLM_TENSOR_ATTN_OUT_NORM,   "blk.%d.attn_output_norm" },
-        },
-    },
-    {
-        LLM_ARCH_GPT2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_POS_EMBD,        "position_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_GPTJ,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-        },
-    },
-    {
-        LLM_ARCH_GPTNEOX,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_MPT,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output"},
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_ACT,         "blk.%d.ffn.act" },
-            { LLM_TENSOR_POS_EMBD,        "position_embd" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm"},
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm"},
-        },
-    },
-    {
-        LLM_ARCH_STARCODER,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_POS_EMBD,        "position_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_REFACT,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_BERT,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" },
-            { LLM_TENSOR_TOKEN_TYPES,     "token_types" },
-            { LLM_TENSOR_POS_EMBD,        "position_embd" },
-            { LLM_TENSOR_ATTN_OUT_NORM,   "blk.%d.attn_output_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_LAYER_OUT_NORM,  "blk.%d.layer_output_norm" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_CLS,             "cls" },
-            { LLM_TENSOR_CLS_OUT,         "cls.output" },
-        },
-    },
-    {
-        LLM_ARCH_NOMIC_BERT,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" },
-            { LLM_TENSOR_TOKEN_TYPES,     "token_types" },
-            { LLM_TENSOR_ATTN_OUT_NORM,   "blk.%d.attn_output_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_LAYER_OUT_NORM,  "blk.%d.layer_output_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_JINA_BERT_V2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" },
-            { LLM_TENSOR_TOKEN_TYPES,     "token_types" },
-            { LLM_TENSOR_ATTN_NORM_2,     "blk.%d.attn_norm_2" },
-            { LLM_TENSOR_ATTN_OUT_NORM,   "blk.%d.attn_output_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_LAYER_OUT_NORM,  "blk.%d.layer_output_norm" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_CLS,             "cls" },
-        },
-    },
-    {
-        LLM_ARCH_BLOOM,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_TOKEN_EMBD_NORM, "token_embd_norm" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_STABLELM,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
-        },
-    },
-    {
-        LLM_ARCH_QWEN,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_QWEN2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_QWEN2MOE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_OUTPUT,             "output" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,             "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,             "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE_INP,       "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_GATE_EXPS,      "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,      "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,        "blk.%d.ffn_up_exps" },
-            { LLM_TENSOR_FFN_GATE_INP_SHEXP, "blk.%d.ffn_gate_inp_shexp" },
-            { LLM_TENSOR_FFN_GATE_SHEXP,     "blk.%d.ffn_gate_shexp" },
-            { LLM_TENSOR_FFN_DOWN_SHEXP,     "blk.%d.ffn_down_shexp" },
-            { LLM_TENSOR_FFN_UP_SHEXP,       "blk.%d.ffn_up_shexp" },
-        },
-    },
-    {
-        LLM_ARCH_PHI2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_PHI3,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_OUTPUT,             "output" },
-            { LLM_TENSOR_ROPE_FACTORS_LONG,  "rope_factors_long" },
-            { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,           "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,             "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,             "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_DOWN,           "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,             "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_PLAMO,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_CODESHELL,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_ORION,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_INTERNLM2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_MINICPM,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_GATE_EXP,    "blk.%d.ffn_gate.%d" },
-            { LLM_TENSOR_FFN_DOWN_EXP,    "blk.%d.ffn_down.%d" },
-            { LLM_TENSOR_FFN_UP_EXP,      "blk.%d.ffn_up.%d" },
-        },
-    },
-    {
-        LLM_ARCH_MINICPM3,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_OUTPUT,             "output" },
-            { LLM_TENSOR_ROPE_FACTORS_LONG,  "rope_factors_long" },
-            { LLM_TENSOR_ROPE_FACTORS_SHORT, "rope_factors_short" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q_A_NORM,      "blk.%d.attn_q_a_norm" },
-            { LLM_TENSOR_ATTN_KV_A_NORM,     "blk.%d.attn_kv_a_norm" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_Q_A,           "blk.%d.attn_q_a" },
-            { LLM_TENSOR_ATTN_Q_B,           "blk.%d.attn_q_b" },
-            { LLM_TENSOR_ATTN_KV_A_MQA,      "blk.%d.attn_kv_a_mqa" },
-            { LLM_TENSOR_ATTN_KV_B,          "blk.%d.attn_kv_b" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,           "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_UP,             "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,           "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_GEMMA,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_GEMMA2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_POST_NORM,  "blk.%d.post_attention_norm" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_POST_NORM,   "blk.%d.post_ffw_norm" },
-        },
-    },
-    {
-        LLM_ARCH_STARCODER2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_MAMBA,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_SSM_IN,          "blk.%d.ssm_in" },
-            { LLM_TENSOR_SSM_CONV1D,      "blk.%d.ssm_conv1d" },
-            { LLM_TENSOR_SSM_X,           "blk.%d.ssm_x" },
-            { LLM_TENSOR_SSM_DT,          "blk.%d.ssm_dt" },
-            { LLM_TENSOR_SSM_A,           "blk.%d.ssm_a" },
-            { LLM_TENSOR_SSM_D,           "blk.%d.ssm_d" },
-            { LLM_TENSOR_SSM_OUT,         "blk.%d.ssm_out" },
-        },
-    },
-    {
-        LLM_ARCH_XVERSE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_COMMAND_R,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
-        },
-    },
-    {
-        LLM_ARCH_DBRX,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_OUT_NORM,   "blk.%d.attn_output_norm" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_GATE_EXPS,   "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,   "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
-        },
-    },
-    {
-        LLM_ARCH_OLMO,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_OLMOE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_OUTPUT,             "output" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,             "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,             "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_Q_NORM,        "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K_NORM,        "blk.%d.attn_k_norm" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE_INP,       "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_GATE_EXPS,      "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,      "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,        "blk.%d.ffn_up_exps" },
-        },
-    },
-    {
-        LLM_ARCH_OPENELM,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_ARCTIC,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_NORM_EXPS,   "blk.%d.ffn_norm_exps" },
-            { LLM_TENSOR_FFN_GATE_EXPS,   "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,   "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
-        },
-    },
-    {
-        LLM_ARCH_DEEPSEEK2,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_OUTPUT,             "output" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q_A_NORM,      "blk.%d.attn_q_a_norm" },
-            { LLM_TENSOR_ATTN_KV_A_NORM,     "blk.%d.attn_kv_a_norm" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_Q_A,           "blk.%d.attn_q_a" },
-            { LLM_TENSOR_ATTN_Q_B,           "blk.%d.attn_q_b" },
-            { LLM_TENSOR_ATTN_KV_A_MQA,      "blk.%d.attn_kv_a_mqa" },
-            { LLM_TENSOR_ATTN_KV_B,          "blk.%d.attn_kv_b" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,           "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_UP,             "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,           "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_GATE_INP,       "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_GATE_EXPS,      "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,      "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,        "blk.%d.ffn_up_exps" },
-            { LLM_TENSOR_FFN_GATE_INP_SHEXP, "blk.%d.ffn_gate_inp_shexp" },
-            { LLM_TENSOR_FFN_GATE_SHEXP,     "blk.%d.ffn_gate_shexp" },
-            { LLM_TENSOR_FFN_DOWN_SHEXP,     "blk.%d.ffn_down_shexp" },
-            { LLM_TENSOR_FFN_UP_SHEXP,       "blk.%d.ffn_up_shexp" },
-        },
-    },
-    {
-        LLM_ARCH_CHATGLM,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_BITNET,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,         "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,        "output_norm" },
-            { LLM_TENSOR_ATTN_Q,             "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,             "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,             "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,           "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_NORM,          "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_SUB_NORM,      "blk.%d.attn_sub_norm" },
-            { LLM_TENSOR_FFN_GATE,           "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,           "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,             "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_NORM,           "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_SUB_NORM,       "blk.%d.ffn_sub_norm" },
-        },
-    },
-    {
-        LLM_ARCH_T5,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,           "token_embd" },
-            { LLM_TENSOR_OUTPUT,               "output" },
-            { LLM_TENSOR_DEC_OUTPUT_NORM,      "dec.output_norm" },
-            { LLM_TENSOR_DEC_ATTN_NORM,        "dec.blk.%d.attn_norm" },
-            { LLM_TENSOR_DEC_ATTN_Q,           "dec.blk.%d.attn_q" },
-            { LLM_TENSOR_DEC_ATTN_K,           "dec.blk.%d.attn_k" },
-            { LLM_TENSOR_DEC_ATTN_V,           "dec.blk.%d.attn_v" },
-            { LLM_TENSOR_DEC_ATTN_OUT,         "dec.blk.%d.attn_o" },
-            { LLM_TENSOR_DEC_ATTN_REL_B,       "dec.blk.%d.attn_rel_b" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_NORM,  "dec.blk.%d.cross_attn_norm" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_Q,     "dec.blk.%d.cross_attn_q" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_K,     "dec.blk.%d.cross_attn_k" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_V,     "dec.blk.%d.cross_attn_v" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_OUT,   "dec.blk.%d.cross_attn_o" },
-            { LLM_TENSOR_DEC_CROSS_ATTN_REL_B, "dec.blk.%d.cross_attn_rel_b" },
-            { LLM_TENSOR_DEC_FFN_NORM,         "dec.blk.%d.ffn_norm" },
-            { LLM_TENSOR_DEC_FFN_GATE,         "dec.blk.%d.ffn_gate" },
-            { LLM_TENSOR_DEC_FFN_DOWN,         "dec.blk.%d.ffn_down" },
-            { LLM_TENSOR_DEC_FFN_UP,           "dec.blk.%d.ffn_up" },
-            { LLM_TENSOR_ENC_OUTPUT_NORM,      "enc.output_norm" },
-            { LLM_TENSOR_ENC_ATTN_NORM,        "enc.blk.%d.attn_norm" },
-            { LLM_TENSOR_ENC_ATTN_Q,           "enc.blk.%d.attn_q" },
-            { LLM_TENSOR_ENC_ATTN_K,           "enc.blk.%d.attn_k" },
-            { LLM_TENSOR_ENC_ATTN_V,           "enc.blk.%d.attn_v" },
-            { LLM_TENSOR_ENC_ATTN_OUT,         "enc.blk.%d.attn_o" },
-            { LLM_TENSOR_ENC_ATTN_REL_B,       "enc.blk.%d.attn_rel_b" },
-            { LLM_TENSOR_ENC_FFN_NORM,         "enc.blk.%d.ffn_norm" },
-            { LLM_TENSOR_ENC_FFN_GATE,         "enc.blk.%d.ffn_gate" },
-            { LLM_TENSOR_ENC_FFN_DOWN,         "enc.blk.%d.ffn_down" },
-            { LLM_TENSOR_ENC_FFN_UP,           "enc.blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_T5ENCODER,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,           "token_embd" },
-            { LLM_TENSOR_OUTPUT,               "output" },
-            { LLM_TENSOR_ENC_OUTPUT_NORM,      "enc.output_norm" },
-            { LLM_TENSOR_ENC_ATTN_NORM,        "enc.blk.%d.attn_norm" },
-            { LLM_TENSOR_ENC_ATTN_Q,           "enc.blk.%d.attn_q" },
-            { LLM_TENSOR_ENC_ATTN_K,           "enc.blk.%d.attn_k" },
-            { LLM_TENSOR_ENC_ATTN_V,           "enc.blk.%d.attn_v" },
-            { LLM_TENSOR_ENC_ATTN_OUT,         "enc.blk.%d.attn_o" },
-            { LLM_TENSOR_ENC_ATTN_REL_B,       "enc.blk.%d.attn_rel_b" },
-            { LLM_TENSOR_ENC_FFN_NORM,         "enc.blk.%d.ffn_norm" },
-            { LLM_TENSOR_ENC_FFN_GATE,         "enc.blk.%d.ffn_gate" },
-            { LLM_TENSOR_ENC_FFN_DOWN,         "enc.blk.%d.ffn_down" },
-            { LLM_TENSOR_ENC_FFN_UP,           "enc.blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_JAIS,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-        },
-    },
-    {
-        LLM_ARCH_NEMOTRON,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_EXAONE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ROPE_FREQS,      "rope_freqs" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_ATTN_ROT_EMBD,   "blk.%d.attn_rot_embd" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_RWKV6,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,                "token_embd" },
-            { LLM_TENSOR_TOKEN_EMBD_NORM,           "token_embd_norm" },
-            { LLM_TENSOR_OUTPUT_NORM,               "output_norm" },
-            { LLM_TENSOR_OUTPUT,                    "output" },
-            { LLM_TENSOR_ATTN_NORM,                 "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_NORM_2,               "blk.%d.attn_norm_2" },
-            { LLM_TENSOR_TIME_MIX_W1,               "blk.%d.time_mix_w1" },
-            { LLM_TENSOR_TIME_MIX_W2,               "blk.%d.time_mix_w2" },
-            { LLM_TENSOR_TIME_MIX_LERP_X,           "blk.%d.time_mix_lerp_x" },
-            { LLM_TENSOR_TIME_MIX_LERP_W,           "blk.%d.time_mix_lerp_w" },
-            { LLM_TENSOR_TIME_MIX_LERP_K,           "blk.%d.time_mix_lerp_k" },
-            { LLM_TENSOR_TIME_MIX_LERP_V,           "blk.%d.time_mix_lerp_v" },
-            { LLM_TENSOR_TIME_MIX_LERP_R,           "blk.%d.time_mix_lerp_r" },
-            { LLM_TENSOR_TIME_MIX_LERP_G,           "blk.%d.time_mix_lerp_g" },
-            { LLM_TENSOR_TIME_MIX_FIRST,            "blk.%d.time_mix_first" },
-            { LLM_TENSOR_TIME_MIX_DECAY,            "blk.%d.time_mix_decay" },
-            { LLM_TENSOR_TIME_MIX_DECAY_W1,         "blk.%d.time_mix_decay_w1" },
-            { LLM_TENSOR_TIME_MIX_DECAY_W2,         "blk.%d.time_mix_decay_w2" },
-            { LLM_TENSOR_TIME_MIX_KEY,              "blk.%d.time_mix_key" },
-            { LLM_TENSOR_TIME_MIX_VALUE,            "blk.%d.time_mix_value" },
-            { LLM_TENSOR_TIME_MIX_RECEPTANCE,       "blk.%d.time_mix_receptance" },
-            { LLM_TENSOR_TIME_MIX_GATE,             "blk.%d.time_mix_gate" },
-            { LLM_TENSOR_TIME_MIX_LN,               "blk.%d.time_mix_ln" },
-            { LLM_TENSOR_TIME_MIX_OUTPUT,           "blk.%d.time_mix_output" },
-            { LLM_TENSOR_CHANNEL_MIX_LERP_K,        "blk.%d.channel_mix_lerp_k" },
-            { LLM_TENSOR_CHANNEL_MIX_LERP_R,        "blk.%d.channel_mix_lerp_r" },
-            { LLM_TENSOR_CHANNEL_MIX_KEY,           "blk.%d.channel_mix_key" },
-            { LLM_TENSOR_CHANNEL_MIX_VALUE,         "blk.%d.channel_mix_value" },
-            { LLM_TENSOR_CHANNEL_MIX_RECEPTANCE,    "blk.%d.channel_mix_receptance" },
-        },
-    },
-    {
-        LLM_ARCH_GRANITE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-        },
-    },
-    {
-        LLM_ARCH_GRANITE_MOE,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE_INP,    "blk.%d.ffn_gate_inp" },
-            { LLM_TENSOR_FFN_GATE_EXPS,   "blk.%d.ffn_gate_exps" },
-            { LLM_TENSOR_FFN_DOWN_EXPS,   "blk.%d.ffn_down_exps" },
-            { LLM_TENSOR_FFN_UP_EXPS,     "blk.%d.ffn_up_exps" },
-        },
-    },
-    {
-        LLM_ARCH_CHAMELEON,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-            { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
-            { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
-            { LLM_TENSOR_ATTN_Q,          "blk.%d.attn_q" },
-            { LLM_TENSOR_ATTN_K,          "blk.%d.attn_k" },
-            { LLM_TENSOR_ATTN_V,          "blk.%d.attn_v" },
-            { LLM_TENSOR_ATTN_OUT,        "blk.%d.attn_output" },
-            { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
-            { LLM_TENSOR_FFN_GATE,        "blk.%d.ffn_gate" },
-            { LLM_TENSOR_FFN_DOWN,        "blk.%d.ffn_down" },
-            { LLM_TENSOR_FFN_UP,          "blk.%d.ffn_up" },
-            { LLM_TENSOR_ATTN_Q_NORM,     "blk.%d.attn_q_norm" },
-            { LLM_TENSOR_ATTN_K_NORM,     "blk.%d.attn_k_norm" },
-        },
-    },
-    {
-        LLM_ARCH_UNKNOWN,
-        {
-            { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
-        },
-    },
-};
-
-static llm_arch llm_arch_from_string(const std::string & name) {
-    for (const auto & kv : LLM_ARCH_NAMES) { // NOLINT
-        if (kv.second == name) {
-            return kv.first;
-        }
-    }
-
-    return LLM_ARCH_UNKNOWN;
-}
-
-// helper to handle gguf constants
-// usage:
-//
-//   const auto tn = LLM_TN(LLM_ARCH_LLAMA);
-//
-//   std::string name = tn(LLM_TENSOR_OUTPUT);                     -> "output"
-//   std::string name = tn(LLM_TENSOR_TOKEN_EMBD, "bias");         -> "token_embd.bias"
-//   std::string name = tn(LLM_TENSOR_ATTN_NORM, "weight", 3);     -> "blk.3.attn_norm.weight"
-//
-struct LLM_TN_IMPL {
-    const llm_arch arch;
-    const llm_tensor tensor;
-    const char * const suffix;
-    const int bid;
-    const int xid;
-
-    std::string str() const {
-        if (LLM_TENSOR_NAMES.at(arch).find(tensor) == LLM_TENSOR_NAMES.at(arch).end()) {
-            return "__missing__";
-        }
-
-        std::string name = ::format(LLM_TENSOR_NAMES.at(arch).at(tensor), bid, xid);
-
-        if (suffix != nullptr) {
-            name += ".";
-            name += suffix;
-        }
-
-        return name;
-    }
-
-    operator std::string() const {
-        return str();
-    }
-
-    friend bool operator==(const std::string & str, const LLM_TN_IMPL & tn) {
-        return str == tn.str();
-    }
-
-    friend bool operator!=(const std::string & str, const LLM_TN_IMPL & tn) {
-        return str != tn.str();
-    }
-};
-
-struct LLM_TN {
-    LLM_TN(llm_arch arch) : arch(arch) {}
-
-    llm_arch arch;
-
-    LLM_TN_IMPL operator()(llm_tensor tensor, const char * suffix, int bid = -1, int xid = -1) const {
-        return { arch, tensor, suffix, bid, xid };
-    }
-
-    LLM_TN_IMPL operator()(llm_tensor tensor, int bid = -1, int xid = -1) const {
-        return { arch, tensor, nullptr, bid, xid };
-    }
-};
-
-//
-// gguf helpers
-//
-
-static const std::map LLAMA_ROPE_SCALING_TYPES = {
-    { LLAMA_ROPE_SCALING_TYPE_NONE,   "none"   },
-    { LLAMA_ROPE_SCALING_TYPE_LINEAR, "linear" },
-    { LLAMA_ROPE_SCALING_TYPE_YARN,   "yarn"   },
-};
-
-static llama_rope_scaling_type llama_rope_scaling_type_from_string(const std::string & name) {
-    for (const auto & kv : LLAMA_ROPE_SCALING_TYPES) {
-        if (kv.second == name) {
-            return (llama_rope_scaling_type) kv.first;
-        }
-    }
-
-    return LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED;
-}
-
-static std::string gguf_data_to_str(enum gguf_type type, const void * data, int i) {
-    switch (type) {
-        case GGUF_TYPE_UINT8:   return std::to_string(((const uint8_t  *)data)[i]);
-        case GGUF_TYPE_INT8:    return std::to_string(((const int8_t   *)data)[i]);
-        case GGUF_TYPE_UINT16:  return std::to_string(((const uint16_t *)data)[i]);
-        case GGUF_TYPE_INT16:   return std::to_string(((const int16_t  *)data)[i]);
-        case GGUF_TYPE_UINT32:  return std::to_string(((const uint32_t *)data)[i]);
-        case GGUF_TYPE_INT32:   return std::to_string(((const int32_t  *)data)[i]);
-        case GGUF_TYPE_UINT64:  return std::to_string(((const uint64_t *)data)[i]);
-        case GGUF_TYPE_INT64:   return std::to_string(((const int64_t  *)data)[i]);
-        case GGUF_TYPE_FLOAT32: return std::to_string(((const float    *)data)[i]);
-        case GGUF_TYPE_FLOAT64: return std::to_string(((const double   *)data)[i]);
-        case GGUF_TYPE_BOOL:    return ((const bool *)data)[i] ? "true" : "false";
-        default:                return format("unknown type %d", type);
-    }
-}
-
-static std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i) {
-    const enum gguf_type type = gguf_get_kv_type(ctx_gguf, i);
-
-    switch (type) {
-        case GGUF_TYPE_STRING:
-            return gguf_get_val_str(ctx_gguf, i);
-        case GGUF_TYPE_ARRAY:
-            {
-                const enum gguf_type arr_type = gguf_get_arr_type(ctx_gguf, i);
-                int arr_n = gguf_get_arr_n(ctx_gguf, i);
-                const void * data = gguf_get_arr_data(ctx_gguf, i);
-                std::stringstream ss;
-                ss << "[";
-                for (int j = 0; j < arr_n; j++) {
-                    if (arr_type == GGUF_TYPE_STRING) {
-                        std::string val = gguf_get_arr_str(ctx_gguf, i, j);
-                        // escape quotes
-                        replace_all(val, "\\", "\\\\");
-                        replace_all(val, "\"", "\\\"");
-                        ss << '"' << val << '"';
-                    } else if (arr_type == GGUF_TYPE_ARRAY) {
-                        ss << "???";
-                    } else {
-                        ss << gguf_data_to_str(arr_type, data, j);
-                    }
-                    if (j < arr_n - 1) {
-                        ss << ", ";
-                    }
-                }
-                ss << "]";
-                return ss.str();
-            }
-        default:
-            return gguf_data_to_str(type, gguf_get_val_data(ctx_gguf, i), 0);
-    }
-}
-
-//
-// llama helpers
-//
-
-#if defined(_WIN32)
-static std::string llama_format_win_err(DWORD err) {
-    LPSTR buf;
-    size_t size = FormatMessageA(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS,
-                                 NULL, err, MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), (LPSTR)&buf, 0, NULL);
-    if (!size) {
-        return "FormatMessageA failed";
-    }
-    std::string ret(buf, size);
-    LocalFree(buf);
-    return ret;
-}
-#endif
-
-template 
-struct no_init {
-    T value;
-    no_init() { /* do nothing */ }
-};
-
-struct llama_file {
-
-#if defined(_WIN32)
-    // use FILE * so we don't have to re-open the file to mmap
-    FILE * fp;
-    HANDLE fp_win32;
-    size_t size;
-
-private:
-    std::string GetErrorMessageWin32(DWORD error_code) const {
-        std::string ret;
-        LPSTR lpMsgBuf = NULL;
-        DWORD bufLen = FormatMessageA(FORMAT_MESSAGE_ALLOCATE_BUFFER | FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS,
-                                    NULL, error_code, MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), (LPSTR)&lpMsgBuf, 0, NULL);
-        if (!bufLen) {
-            ret = format("Win32 error code: %s", error_code);
-        } else {
-            ret = lpMsgBuf;
-            LocalFree(lpMsgBuf);
-        }
-
-        return ret;
-    }
-
-public:
-
-    llama_file(const char * fname, const char * mode) {
-        fp = ggml_fopen(fname, mode);
-        if (fp == NULL) {
-            throw std::runtime_error(format("failed to open %s: %s", fname, strerror(errno)));
-        }
-        fp_win32 = (HANDLE) _get_osfhandle(_fileno(fp));
-        seek(0, SEEK_END);
-        size = tell();
-        seek(0, SEEK_SET);
-    }
-
-    size_t tell() const {
-        // SetFilePointerEx returns the current position when seeking relative 0 bytes
-        LARGE_INTEGER li;
-        li.QuadPart = 0;
-        BOOL ret = SetFilePointerEx(fp_win32, li, &li, FILE_CURRENT);
-        if (!ret) {
-            throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str()));
-        }
-
-        return li.QuadPart;
-    }
-
-    void seek(size_t offset, int whence) const {
-        // no need to convert SEEK_* to FILE_*. The enums are the same.
-        // Still, keep static asserts to avoid failures in the future.
-        static_assert(SEEK_SET == FILE_BEGIN, "SEEK_SET != FILE_BEGIN");
-        static_assert(SEEK_CUR == FILE_CURRENT, "SEEK_CUR != FILE_CURRENT");
-        static_assert(SEEK_END == FILE_END, "SEEK_END != FILE_END");
-
-        LARGE_INTEGER li;
-        li.QuadPart = offset;
-        BOOL ret = SetFilePointerEx(fp_win32, li, NULL, whence);
-        if (!ret) {
-            throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str()));
-        }
-    }
-
-    void read_raw(void * ptr, size_t len) const {
-        // On Win32 ReadFile is significant faster than fread which is again significant faster than std::fstream. Thus
-        // use the Win32 API to do file io instead of the C/C++ library functions.
-
-        // There are conditions under which ReadFile cannot read chunks >64MB.
-        // Thus split the operation into smaller chunks if len exceeds this limit.
-        size_t bytes_read = 0;
-        while (bytes_read < len) {
-            size_t chunk_size = std::min(len - bytes_read, 64*1024*1024);
-            DWORD chunk_read = 0;
-            BOOL result = ReadFile(fp_win32, reinterpret_cast(ptr) + bytes_read, chunk_size, &chunk_read, NULL);
-            if (!result) {
-                throw std::runtime_error(format("read error: %s", GetErrorMessageWin32(GetLastError()).c_str()));
-            }
-            if (chunk_read < chunk_size || chunk_read == 0) {
-                throw std::runtime_error("unexpectedly reached end of file");
-            }
-
-            bytes_read += chunk_read;
-        } ;
-    }
-
-    uint32_t read_u32() const {
-        uint32_t val;
-        read_raw(&val, sizeof(val));
-        return val;
-    }
-
-    void write_raw(const void * ptr, size_t len) const {
-        // There are conditions under which WriteFile cannot write chunks >64MB.
-        // Thus split the operation into smaller chunks if len exceeds this limit.
-        size_t bytes_written = 0;
-        while (bytes_written < len) {
-            size_t chunk_size = std::min(len - bytes_written, 64*1024*1024);
-            DWORD chunk_written = 0;
-            BOOL result = WriteFile(fp_win32, reinterpret_cast(ptr) + bytes_written, chunk_size, &chunk_written, NULL);
-            if (!result) {
-                throw std::runtime_error(format("write error: %s", GetErrorMessageWin32(GetLastError()).c_str()));
-            }
-            if (chunk_written < chunk_size || chunk_written == 0) {
-                throw std::runtime_error("unexpectedly failed to write bytes");
-            }
-
-            bytes_written += chunk_written;
-        }
-    }
-
-    void write_u32(std::uint32_t val) const {
-        write_raw(&val, sizeof(val));
-    }
-
-    ~llama_file() {
-        if (fp) {
-            std::fclose(fp);
-        }
-    }
-#else
-    // use FILE * so we don't have to re-open the file to mmap
-    FILE * fp;
-    size_t size;
-
-    llama_file(const char * fname, const char * mode) {
-        fp = ggml_fopen(fname, mode);
-        if (fp == NULL) {
-            throw std::runtime_error(format("failed to open %s: %s", fname, strerror(errno)));
-        }
-        seek(0, SEEK_END);
-        size = tell();
-        seek(0, SEEK_SET);
-    }
-
-    size_t tell() const {
-#ifdef _WIN32
-        __int64 ret = _ftelli64(fp);
-#else
-        long ret = std::ftell(fp);
-#endif
-        if (ret == -1) {
-            throw std::runtime_error(format("ftell error: %s", strerror(errno)));
-        }
-
-        return (size_t) ret;
-    }
-
-    void seek(size_t offset, int whence) const {
-#ifdef _WIN32
-        int ret = _fseeki64(fp, (__int64) offset, whence);
-#else
-        int ret = std::fseek(fp, (long) offset, whence);
-#endif
-        if (ret != 0) {
-            throw std::runtime_error(format("seek error: %s", strerror(errno)));
-        }
-    }
-
-    void read_raw(void * ptr, size_t len) const {
-        if (len == 0) {
-            return;
-        }
-        errno = 0;
-        std::size_t ret = std::fread(ptr, len, 1, fp);
-        if (ferror(fp)) {
-            throw std::runtime_error(format("read error: %s", strerror(errno)));
-        }
-        if (ret != 1) {
-            throw std::runtime_error("unexpectedly reached end of file");
-        }
-    }
-
-    uint32_t read_u32() const {
-        uint32_t ret;
-        read_raw(&ret, sizeof(ret));
-        return ret;
-    }
-
-    void write_raw(const void * ptr, size_t len) const {
-        if (len == 0) {
-            return;
-        }
-        errno = 0;
-        size_t ret = std::fwrite(ptr, len, 1, fp);
-        if (ret != 1) {
-            throw std::runtime_error(format("write error: %s", strerror(errno)));
-        }
-    }
-
-    void write_u32(std::uint32_t val) const {
-        write_raw(&val, sizeof(val));
-    }
-
-    ~llama_file() {
-        if (fp) {
-            std::fclose(fp);
-        }
-    }
-#endif
-};
-using llama_files = std::vector>;
-
-struct llama_mmap {
-    void * addr;
-    size_t size;
-
-    llama_mmap(const llama_mmap &) = delete;
-
-#ifdef _POSIX_MAPPED_FILES
-    static constexpr bool SUPPORTED = true;
-
-    // list of mapped fragments (first_offset, last_offset)
-    std::vector> mapped_fragments;
-
-    llama_mmap(struct llama_file * file, size_t prefetch = (size_t) -1 /* -1 = max value */, bool numa = false) {
-        size = file->size;
-        int fd = fileno(file->fp);
-        int flags = MAP_SHARED;
-        // prefetch/readahead impairs performance on NUMA systems
-        if (numa)  { prefetch = 0; }
-#ifdef __linux__
-        // advise the kernel to read the file sequentially (increases readahead)
-        if (posix_fadvise(fd, 0, 0, POSIX_FADV_SEQUENTIAL)) {
-            LLAMA_LOG_WARN("warning: posix_fadvise(.., POSIX_FADV_SEQUENTIAL) failed: %s\n",
-                    strerror(errno));
-        }
-        if (prefetch) { flags |= MAP_POPULATE; }
-#endif
-        addr = mmap(NULL, file->size, PROT_READ, flags, fd, 0);
-        if (addr == MAP_FAILED) { // NOLINT
-            throw std::runtime_error(format("mmap failed: %s", strerror(errno)));
-        }
-
-        if (prefetch > 0) {
-            // advise the kernel to preload the mapped memory
-            if (posix_madvise(addr, std::min(file->size, prefetch), POSIX_MADV_WILLNEED)) {
-                LLAMA_LOG_WARN("warning: posix_madvise(.., POSIX_MADV_WILLNEED) failed: %s\n",
-                        strerror(errno));
-            }
-        }
-        if (numa) {
-            // advise the kernel not to use readahead
-            // (because the next page might not belong on the same node)
-            if (posix_madvise(addr, file->size, POSIX_MADV_RANDOM)) {
-                LLAMA_LOG_WARN("warning: posix_madvise(.., POSIX_MADV_RANDOM) failed: %s\n",
-                        strerror(errno));
-            }
-        }
-
-        // initialize list of mapped_fragments
-        mapped_fragments.emplace_back(0, file->size);
-    }
-
-    static void align_range(size_t * first, size_t * last, size_t page_size) {
-        // align first to the next page
-        size_t offset_in_page = *first & (page_size - 1);
-        size_t offset_to_page = offset_in_page == 0 ? 0 : page_size - offset_in_page;
-        *first += offset_to_page;
-
-        // align last to the previous page
-        *last = *last & ~(page_size - 1);
-
-        if (*last <= *first) {
-            *last = *first;
-        }
-    }
-
-    // partially unmap the file in the range [first, last)
-    void unmap_fragment(size_t first, size_t last) {
-        // note: this function must not be called multiple times with overlapping ranges
-        // otherwise, there is a risk of invalidating addresses that have been repurposed for other mappings
-        int page_size = sysconf(_SC_PAGESIZE);
-        align_range(&first, &last, page_size);
-        size_t len = last - first;
-
-        if (len == 0) {
-            return;
-        }
-
-        GGML_ASSERT(first % page_size == 0);
-        GGML_ASSERT(last % page_size == 0);
-        GGML_ASSERT(last > first);
-
-        void * next_page_start = (uint8_t *) addr + first;
-
-        // unmap the range
-        if (munmap(next_page_start, len)) {
-            LLAMA_LOG_WARN("warning: munmap failed: %s\n", strerror(errno));
-        }
-
-        // update the list of mapped fragments to avoid unmapping the same range again in the destructor
-        std::vector> new_mapped_fragments;
-        for (const auto & frag : mapped_fragments) {
-            if (frag.first < first && frag.second > last) {
-                // the range is in the middle of the fragment, split it
-                new_mapped_fragments.emplace_back(frag.first, first);
-                new_mapped_fragments.emplace_back(last, frag.second);
-            } else if (frag.first < first && frag.second > first) {
-                // the range starts in the middle of the fragment
-                new_mapped_fragments.emplace_back(frag.first, first);
-            } else if (frag.first < last && frag.second > last) {
-                // the range ends in the middle of the fragment
-                new_mapped_fragments.emplace_back(last, frag.second);
-            } else if (frag.first >= first && frag.second <= last) {
-                // the range covers the entire fragment
-            } else {
-                // the range is outside the fragment
-                new_mapped_fragments.push_back(frag);
-            }
-        }
-        mapped_fragments = std::move(new_mapped_fragments);
-    }
-
-    ~llama_mmap() {
-        for (const auto & frag : mapped_fragments) {
-            if (munmap((char *) addr + frag.first, frag.second - frag.first)) {
-                LLAMA_LOG_WARN("warning: munmap failed: %s\n", strerror(errno));
-            }
-        }
-    }
-#elif defined(_WIN32)
-    static constexpr bool SUPPORTED = true;
-
-    llama_mmap(struct llama_file * file, size_t prefetch = (size_t) -1, bool numa = false) {
-        GGML_UNUSED(numa);
-
-        size = file->size;
-
-        HANDLE hFile = (HANDLE) _get_osfhandle(_fileno(file->fp));
-
-        HANDLE hMapping = CreateFileMappingA(hFile, NULL, PAGE_READONLY, 0, 0, NULL);
-
-        if (hMapping == NULL) {
-            DWORD error = GetLastError();
-            throw std::runtime_error(format("CreateFileMappingA failed: %s", llama_format_win_err(error).c_str()));
-        }
-
-        addr = MapViewOfFile(hMapping, FILE_MAP_READ, 0, 0, 0);
-        DWORD error = GetLastError();
-        CloseHandle(hMapping);
-
-        if (addr == NULL) {
-            throw std::runtime_error(format("MapViewOfFile failed: %s", llama_format_win_err(error).c_str()));
-        }
-
-        if (prefetch > 0) {
-#if _WIN32_WINNT >= 0x602
-            // PrefetchVirtualMemory is only present on Windows 8 and above, so we dynamically load it
-            BOOL (WINAPI *pPrefetchVirtualMemory) (HANDLE, ULONG_PTR, PWIN32_MEMORY_RANGE_ENTRY, ULONG);
-            HMODULE hKernel32 = GetModuleHandleW(L"kernel32.dll");
-
-            // may fail on pre-Windows 8 systems
-            pPrefetchVirtualMemory = reinterpret_cast (GetProcAddress(hKernel32, "PrefetchVirtualMemory"));
-
-            if (pPrefetchVirtualMemory) {
-                // advise the kernel to preload the mapped memory
-                WIN32_MEMORY_RANGE_ENTRY range;
-                range.VirtualAddress = addr;
-                range.NumberOfBytes = (SIZE_T) std::min(size, prefetch);
-                if (!pPrefetchVirtualMemory(GetCurrentProcess(), 1, &range, 0)) {
-                    LLAMA_LOG_WARN("warning: PrefetchVirtualMemory failed: %s\n",
-                            llama_format_win_err(GetLastError()).c_str());
-                }
-            }
-#else
-            throw std::runtime_error("PrefetchVirtualMemory unavailable");
-#endif
-        }
-    }
-
-    void unmap_fragment(size_t first, size_t last) {
-        // not supported
-        GGML_UNUSED(first);
-        GGML_UNUSED(last);
-    }
-
-    ~llama_mmap() {
-        if (!UnmapViewOfFile(addr)) {
-            LLAMA_LOG_WARN("warning: UnmapViewOfFile failed: %s\n",
-                    llama_format_win_err(GetLastError()).c_str());
-        }
-    }
-#else
-    static constexpr bool SUPPORTED = false;
-
-    llama_mmap(struct llama_file * file, size_t prefetch = -1, bool numa = false) {
-        GGML_UNUSED(file);
-        GGML_UNUSED(prefetch);
-        GGML_UNUSED(numa);
-
-        throw std::runtime_error("mmap not supported");
-    }
-
-    void unmap_fragment(size_t first, size_t last) {
-        GGML_UNUSED(first);
-        GGML_UNUSED(last);
-
-        throw std::runtime_error("mmap not supported");
-    }
-#endif
-};
-using llama_mmaps = std::vector>;
-
-// Represents some region of memory being locked using mlock or VirtualLock;
-// will automatically unlock on destruction.
-struct llama_mlock {
-    void * addr = NULL;
-    size_t size = 0;
-
-    bool failed_already = false;
-
-    llama_mlock() {}
-    llama_mlock(const llama_mlock &) = delete;
-
-    ~llama_mlock() {
-        if (size) {
-            raw_unlock(addr, size);
-        }
-    }
-
-    void init(void * ptr) {
-        GGML_ASSERT(addr == NULL && size == 0); // NOLINT
-        addr = ptr;
-    }
-
-    void grow_to(size_t target_size) {
-        GGML_ASSERT(addr);
-        if (failed_already) {
-            return;
-        }
-        size_t granularity = lock_granularity();
-        target_size = (target_size + granularity - 1) & ~(granularity - 1);
-        if (target_size > size) {
-            if (raw_lock((uint8_t *) addr + size, target_size - size)) {
-                size = target_size;
-            } else {
-                failed_already = true;
-            }
-        }
-    }
-
-#ifdef _POSIX_MEMLOCK_RANGE
-    static constexpr bool SUPPORTED = true;
-
-    static size_t lock_granularity() {
-        return (size_t) sysconf(_SC_PAGESIZE);
-    }
-
-    #ifdef __APPLE__
-        #define MLOCK_SUGGESTION \
-            "Try increasing the sysctl values 'vm.user_wire_limit' and 'vm.global_user_wire_limit' and/or " \
-            "decreasing 'vm.global_no_user_wire_amount'.  Also try increasing RLIMIT_MEMLOCK (ulimit -l).\n"
-    #else
-        #define MLOCK_SUGGESTION \
-            "Try increasing RLIMIT_MEMLOCK ('ulimit -l' as root).\n"
-    #endif
-
-    bool raw_lock(const void * addr, size_t size) const {
-        if (!mlock(addr, size)) {
-            return true;
-        }
-
-        char* errmsg = std::strerror(errno);
-        bool suggest = (errno == ENOMEM);
-
-        // Check if the resource limit is fine after all
-        struct rlimit lock_limit;
-        if (suggest && getrlimit(RLIMIT_MEMLOCK, &lock_limit)) {
-            suggest = false;
-        }
-        if (suggest && (lock_limit.rlim_max > lock_limit.rlim_cur + size)) {
-            suggest = false;
-        }
-
-        LLAMA_LOG_WARN("warning: failed to mlock %zu-byte buffer (after previously locking %zu bytes): %s\n%s",
-                size, this->size, errmsg, suggest ? MLOCK_SUGGESTION : "");
-        return false;
-    }
-
-    #undef MLOCK_SUGGESTION
-
-    static void raw_unlock(void * addr, size_t size) {
-        if (munlock(addr, size)) {
-            LLAMA_LOG_WARN("warning: failed to munlock buffer: %s\n", std::strerror(errno));
-        }
-    }
-#elif defined(_WIN32)
-    static constexpr bool SUPPORTED = true;
-
-    static size_t lock_granularity() {
-        SYSTEM_INFO si;
-        GetSystemInfo(&si);
-        return (size_t) si.dwPageSize;
-    }
-
-    bool raw_lock(void * ptr, size_t len) const {
-        for (int tries = 1; ; tries++) {
-            if (VirtualLock(ptr, len)) {
-                return true;
-            }
-            if (tries == 2) {
-                LLAMA_LOG_WARN("warning: failed to VirtualLock %zu-byte buffer (after previously locking %zu bytes): %s\n",
-                    len, size, llama_format_win_err(GetLastError()).c_str());
-                return false;
-            }
-
-            // It failed but this was only the first try; increase the working
-            // set size and try again.
-            SIZE_T min_ws_size, max_ws_size;
-            if (!GetProcessWorkingSetSize(GetCurrentProcess(), &min_ws_size, &max_ws_size)) {
-                LLAMA_LOG_WARN("warning: GetProcessWorkingSetSize failed: %s\n",
-                        llama_format_win_err(GetLastError()).c_str());
-                return false;
-            }
-            // Per MSDN: "The maximum number of pages that a process can lock
-            // is equal to the number of pages in its minimum working set minus
-            // a small overhead."
-            // Hopefully a megabyte is enough overhead:
-            size_t increment = len + 1048576;
-            // The minimum must be <= the maximum, so we need to increase both:
-            min_ws_size += increment;
-            max_ws_size += increment;
-            if (!SetProcessWorkingSetSize(GetCurrentProcess(), min_ws_size, max_ws_size)) {
-                LLAMA_LOG_WARN("warning: SetProcessWorkingSetSize failed: %s\n",
-                        llama_format_win_err(GetLastError()).c_str());
-                return false;
-            }
-        }
-    }
-
-    static void raw_unlock(void * ptr, size_t len) {
-        if (!VirtualUnlock(ptr, len)) {
-            LLAMA_LOG_WARN("warning: failed to VirtualUnlock buffer: %s\n",
-                    llama_format_win_err(GetLastError()).c_str());
-        }
-    }
-#else
-    static constexpr bool SUPPORTED = false;
-
-    static size_t lock_granularity() {
-        return (size_t) 65536;
-    }
-
-    bool raw_lock(const void * addr, size_t len) const {
-        LLAMA_LOG_WARN("warning: mlock not supported on this system\n");
-        return false;
-    }
-
-    static void raw_unlock(const void * addr, size_t len) {}
-#endif
-};
-using llama_mlocks = std::vector>;
-
-// NOTE: avoid ever using this except for building the token_to_piece caches
-static std::string llama_token_to_piece(const struct llama_model * model, llama_token token, bool special) {
-    std::string piece;
-    piece.resize(piece.capacity());  // using string internal cache
-    const int n_chars = llama_token_to_piece(model, token, &piece[0], piece.size(), 0, special);
-    if (n_chars < 0) {
-        piece.resize(-n_chars);
-        int check = llama_token_to_piece(model, token, &piece[0], piece.size(), 0, special);
-        GGML_ASSERT(check == -n_chars);
-    }
-    else {
-        piece.resize(n_chars);
-    }
-
-    return piece;
-}
-
-//
-// globals
-//
-
-struct llama_logger_state {
-    ggml_log_callback log_callback = llama_log_callback_default;
-    void * log_callback_user_data = nullptr;
-};
-
-static llama_logger_state g_logger_state;
-
-// available llama models
-enum e_model {
-    MODEL_UNKNOWN,
-    MODEL_14M,
-    MODEL_17M,
-    MODEL_22M,
-    MODEL_33M,
-    MODEL_60M,
-    MODEL_70M,
-    MODEL_80M,
-    MODEL_109M,
-    MODEL_137M,
-    MODEL_160M,
-    MODEL_220M,
-    MODEL_250M,
-    MODEL_270M,
-    MODEL_335M,
-    MODEL_410M,
-    MODEL_450M,
-    MODEL_770M,
-    MODEL_780M,
-    MODEL_0_5B,
-    MODEL_1B,
-    MODEL_1_3B,
-    MODEL_1_4B,
-    MODEL_1_6B,
-    MODEL_2B,
-    MODEL_2_8B,
-    MODEL_3B,
-    MODEL_4B,
-    MODEL_6B,
-    MODEL_6_9B,
-    MODEL_7B,
-    MODEL_8B,
-    MODEL_9B,
-    MODEL_11B,
-    MODEL_12B,
-    MODEL_13B,
-    MODEL_14B,
-    MODEL_15B,
-    MODEL_16B,
-    MODEL_20B,
-    MODEL_30B,
-    MODEL_34B,
-    MODEL_35B,
-    MODEL_40B,
-    MODEL_65B,
-    MODEL_70B,
-    MODEL_236B,
-    MODEL_314B,
-    MODEL_SMALL,
-    MODEL_MEDIUM,
-    MODEL_LARGE,
-    MODEL_XL,
-    MODEL_A1_7B,
-    MODEL_A2_7B,
-    MODEL_8x7B,
-    MODEL_8x22B,
-    MODEL_16x12B,
-    MODEL_10B_128x3_66B,
-    MODEL_57B_A14B,
-    MODEL_27B,
-};
-
-static const size_t kiB = 1024;
-static const size_t MiB = 1024*kiB;
-static const size_t GiB = 1024*MiB;
-
-struct llama_hparams {
-    bool vocab_only;
-    bool rope_finetuned;
-    bool use_par_res;
-    bool swin_norm;
-
-    uint32_t n_vocab;
-    uint32_t n_ctx_train; // context size the model was trained on
-    uint32_t n_embd;
-    uint32_t n_layer;
-    uint32_t n_rot;
-    uint32_t n_swa = 0; // sliding window attention (SWA)
-    uint32_t n_embd_head_k; // dimension of keys (d_k). d_q is assumed to be the same, but there are n_head q heads, and only n_head_kv k-v heads
-    uint32_t n_embd_head_v; // dimension of values (d_v) aka n_embd_head
-    uint32_t n_expert = 0;
-    uint32_t n_expert_used = 0;
-    uint32_t n_vocab_type = 0; // for BERT-style token types
-    uint32_t n_rel_attn_bkts = 0;
-
-    std::array n_head_arr;
-    std::array n_head_kv_arr;
-    std::array n_ff_arr;
-
-    uint32_t n_layer_dense_lead = 0;
-    uint32_t n_lora_q = 0;
-    uint32_t n_lora_kv = 0;
-    uint32_t n_ff_exp = 0;
-    uint32_t n_ff_shexp = 0;
-    uint32_t n_expert_shared = 0;
-    float    expert_weights_scale = 0.0;
-
-    float f_norm_eps;
-    float f_norm_rms_eps;
-
-    float f_attn_logit_softcapping = 50.0f;
-    float f_final_logit_softcapping = 30.0f;
-
-    // for RWKV
-    uint32_t rescale_every_n_layers = 0;
-    uint32_t time_mix_extra_dim = 0;
-    uint32_t time_decay_extra_dim = 0;
-    uint32_t wkv_head_size = 0;
-
-    float    rope_attn_factor = 1.0f;
-    float    rope_freq_base_train;
-    float    rope_freq_scale_train;
-    uint32_t n_ctx_orig_yarn;
-    float    rope_yarn_log_mul;
-
-    // for State Space Models
-    uint32_t ssm_d_conv  = 0;
-    uint32_t ssm_d_inner = 0;
-    uint32_t ssm_d_state = 0;
-    uint32_t ssm_dt_rank = 0;
-    bool ssm_dt_b_c_rms = false;
-
-    float f_clamp_kqv      = 0.0f;
-    float f_max_alibi_bias = 0.0f;
-    float f_logit_scale    = 0.0f;
-
-    // Additional scale factors (Granite/Granite MoE)
-    float f_residual_scale  = 0.0f;
-    float f_embedding_scale = 0.0f;
-    float f_attention_scale = 0.0f;
-
-    bool causal_attn   = true;
-    bool use_alibi     = false;
-    bool attn_soft_cap = false;
-
-    // needed by encoder-decoder models (e.g. T5, FLAN-T5)
-    // ref: https://github.com/ggerganov/llama.cpp/pull/8141
-    llama_token dec_start_token_id = LLAMA_TOKEN_NULL;
-
-    enum llama_pooling_type      pooling_type            = LLAMA_POOLING_TYPE_NONE;
-    enum llama_rope_type         rope_type               = LLAMA_ROPE_TYPE_NONE;
-    enum llama_rope_scaling_type rope_scaling_type_train = LLAMA_ROPE_SCALING_TYPE_NONE;
-
-    bool operator!=(const llama_hparams & other) const {
-        if (this->vocab_only    != other.vocab_only)    return true;
-        if (this->n_vocab       != other.n_vocab)       return true;
-        if (this->n_ctx_train   != other.n_ctx_train)   return true;
-        if (this->n_embd        != other.n_embd)        return true;
-        if (this->n_layer       != other.n_layer)       return true;
-        if (this->n_rot         != other.n_rot)         return true;
-        if (this->n_swa         != other.n_swa)         return true;
-        if (this->n_embd_head_k != other.n_embd_head_k) return true;
-        if (this->n_embd_head_v != other.n_embd_head_v) return true;
-        if (this->n_expert      != other.n_expert)      return true;
-        if (this->n_expert_used != other.n_expert_used) return true;
-
-        if (this->n_head_arr    != other.n_head_arr)    return true;
-        if (this->n_head_kv_arr != other.n_head_kv_arr) return true;
-        if (this->n_ff_arr      != other.n_ff_arr)      return true;
-
-        if (this->n_rel_attn_bkts    != other.n_rel_attn_bkts)    return true;
-        if (this->n_layer_dense_lead != other.n_layer_dense_lead) return true;
-        if (this->n_lora_q           != other.n_lora_q)           return true;
-        if (this->n_lora_kv          != other.n_lora_kv)          return true;
-        if (this->n_ff_exp           != other.n_ff_exp)           return true;
-        if (this->n_ff_shexp         != other.n_ff_shexp)         return true;
-        if (this->n_expert_shared    != other.n_expert_shared)    return true;
-
-        if (this->rope_finetuned  != other.rope_finetuned)  return true;
-        if (this->n_ctx_orig_yarn != other.n_ctx_orig_yarn) return true;
-
-        if (this->ssm_d_conv  != other.ssm_d_conv)  return true;
-        if (this->ssm_d_inner != other.ssm_d_inner) return true;
-        if (this->ssm_d_state != other.ssm_d_state) return true;
-        if (this->ssm_dt_rank != other.ssm_dt_rank) return true;
-        if (this->ssm_dt_b_c_rms != other.ssm_dt_b_c_rms) return true;
-
-        if (this->rescale_every_n_layers != other.rescale_every_n_layers) return true;
-        if (this->time_mix_extra_dim     != other.time_mix_extra_dim)     return true;
-        if (this->time_decay_extra_dim   != other.time_decay_extra_dim)   return true;
-        if (this->wkv_head_size          != other.wkv_head_size)          return true;
-
-        if (this->dec_start_token_id != other.dec_start_token_id) return true;
-
-        const float EPSILON = 1e-9f;
-
-        if (!is_float_close(this->f_norm_eps,            other.f_norm_eps,            EPSILON)) return true;
-        if (!is_float_close(this->f_norm_rms_eps,        other.f_norm_rms_eps,        EPSILON)) return true;
-        if (!is_float_close(this->rope_attn_factor,      other.rope_attn_factor,      EPSILON)) return true;
-        if (!is_float_close(this->rope_freq_base_train,  other.rope_freq_base_train,  EPSILON)) return true;
-        if (!is_float_close(this->rope_freq_scale_train, other.rope_freq_scale_train, EPSILON)) return true;
-        if (!is_float_close(this->expert_weights_scale,  other.expert_weights_scale,  EPSILON)) return true;
-        if (!is_float_close(this->rope_yarn_log_mul,     other.rope_yarn_log_mul,     EPSILON)) return true;
-        if (!is_float_close(this->f_residual_scale,      other.f_residual_scale,      EPSILON)) return true;
-        if (!is_float_close(this->f_embedding_scale,     other.f_embedding_scale,     EPSILON)) return true;
-        if (!is_float_close(this->f_attention_scale,     other.f_attention_scale,     EPSILON)) return true;
-
-        return false;
-    }
-
-    uint32_t n_head(uint32_t il = 0) const {
-        if (il < n_layer) {
-            return n_head_arr[il];
-        }
-
-        GGML_ABORT("fatal error");
-    }
-
-    uint32_t n_head_kv(uint32_t il = 0) const {
-        if (il < n_layer) {
-            return n_head_kv_arr[il];
-        }
-
-        GGML_ABORT("fatal error");
-    }
-
-    uint32_t n_ff(uint32_t il = 0) const {
-        if (il < n_layer) {
-            return n_ff_arr[il];
-        }
-
-        GGML_ABORT("fatal error");
-    }
-
-    uint32_t n_gqa(uint32_t il = 0) const {
-        const uint32_t n_head    = this->n_head(il);
-        const uint32_t n_head_kv = this->n_head_kv(il);
-
-        if (n_head_kv == 0) {
-            return 0;
-        }
-
-        return n_head/n_head_kv;
-    }
-
-    uint32_t n_embd_k_gqa(uint32_t il = 0) const { // dimension of key embeddings across all k-v heads
-        const uint32_t n_head_kv = this->n_head_kv(il);
-
-        return n_embd_head_k * n_head_kv;
-    }
-
-    uint32_t n_embd_v_gqa(uint32_t il = 0) const { // dimension of value embeddings across all k-v heads
-        const uint32_t n_head_kv = this->n_head_kv(il);
-
-        return n_embd_head_v * n_head_kv;
-    }
-
-    uint32_t n_embd_k_s() const { // dimension of the rolling state embeddings
-        // corresponds to Mamba's conv_states size or RWKV's token_shift states size
-        if (wkv_head_size != 0) {
-            // for RWKV models
-            return 2 * n_embd;
-        } else {
-            // TODO: maybe support other convolution strides than 1
-            // NOTE: since the first column of the conv_state is shifted out each time, it's not actually needed
-            return (ssm_d_conv > 0 ? ssm_d_conv - 1 : 0) * ssm_d_inner;
-        }
-    }
-
-    uint32_t n_embd_v_s() const { // dimension of the recurrent state embeddings
-        if (wkv_head_size != 0) {
-            // corresponds to RWKV's wkv_states size
-            return n_embd * wkv_head_size;
-        } else {
-            // corresponds to Mamba's ssm_states size
-            return ssm_d_state * ssm_d_inner;
-        }
-    }
-};
-
-static_assert(std::is_trivially_copyable::value, "llama_hparams must be trivially copyable");
-
-struct llama_cparams {
-    uint32_t n_ctx;           // context size used during inference
-    uint32_t n_batch;
-    uint32_t n_ubatch;
-    uint32_t n_seq_max;
-    int      n_threads;       // number of threads to use for generation
-    int      n_threads_batch; // number of threads to use for batch processing
-
-    float rope_freq_base;
-    float rope_freq_scale;
-
-    uint32_t n_ctx_orig_yarn;
-    // These hyperparameters are not exposed in GGUF, because all
-    // existing YaRN models use the same values for them.
-    float yarn_ext_factor;
-    float yarn_attn_factor;
-    float yarn_beta_fast;
-    float yarn_beta_slow;
-    float defrag_thold;
-
-    bool embeddings;
-    bool causal_attn;
-    bool offload_kqv;
-    bool flash_attn;
-    bool no_perf;
-
-    enum llama_pooling_type pooling_type;
-
-    ggml_backend_sched_eval_callback cb_eval;
-    void * cb_eval_user_data;
-};
-
-// TODO: separate into "llama_layer_enc" and "llama_layer_dec"
-struct llama_layer {
-    llama_layer() {
-        // initialize all pointers to NULL
-        std::memset(this, 0, sizeof(*this));
-    }
-
-    // normalization
-    struct ggml_tensor * attn_norm;
-    struct ggml_tensor * attn_norm_b;
-    struct ggml_tensor * attn_norm_2;
-    struct ggml_tensor * attn_norm_2_b;
-    struct ggml_tensor * attn_q_norm;
-    struct ggml_tensor * attn_q_norm_b;
-    struct ggml_tensor * attn_k_norm;
-    struct ggml_tensor * attn_k_norm_b;
-    struct ggml_tensor * attn_out_norm;
-    struct ggml_tensor * attn_out_norm_b;
-    struct ggml_tensor * attn_q_a_norm;
-    struct ggml_tensor * attn_kv_a_norm;
-    struct ggml_tensor * attn_sub_norm;
-    struct ggml_tensor * attn_post_norm;
-    struct ggml_tensor * ffn_sub_norm;
-    struct ggml_tensor * attn_norm_cross;
-    struct ggml_tensor * attn_norm_enc;
-
-    // attention
-    struct ggml_tensor * wq;
-    struct ggml_tensor * wk;
-    struct ggml_tensor * wv;
-    struct ggml_tensor * wo;
-    struct ggml_tensor * wqkv;
-    struct ggml_tensor * wq_a;
-    struct ggml_tensor * wq_b;
-    struct ggml_tensor * wkv_a_mqa;
-    struct ggml_tensor * wkv_b;
-    struct ggml_tensor * wq_cross;
-    struct ggml_tensor * wk_cross;
-    struct ggml_tensor * wv_cross;
-    struct ggml_tensor * wo_cross;
-    struct ggml_tensor * wq_enc;
-    struct ggml_tensor * wk_enc;
-    struct ggml_tensor * wv_enc;
-    struct ggml_tensor * wo_enc;
-
-    // attention bias
-    struct ggml_tensor * bq;
-    struct ggml_tensor * bk;
-    struct ggml_tensor * bv;
-    struct ggml_tensor * bo;
-    struct ggml_tensor * bqkv;
-
-    // relative position bias
-    struct ggml_tensor * attn_rel_b;
-    struct ggml_tensor * attn_rel_b_enc;
-    struct ggml_tensor * attn_rel_b_cross;
-
-    // normalization
-    struct ggml_tensor * ffn_norm;
-    struct ggml_tensor * ffn_norm_b;
-    struct ggml_tensor * ffn_post_norm;
-    struct ggml_tensor * layer_out_norm;
-    struct ggml_tensor * layer_out_norm_b;
-    struct ggml_tensor * ffn_norm_exps;
-    struct ggml_tensor * ffn_norm_enc;
-
-    // ff
-    struct ggml_tensor * ffn_gate; // w1
-    struct ggml_tensor * ffn_down; // w2
-    struct ggml_tensor * ffn_up;   // w3
-    struct ggml_tensor * ffn_gate_enc;
-    struct ggml_tensor * ffn_down_enc;
-    struct ggml_tensor * ffn_up_enc;
-
-    // ff MoE
-    struct ggml_tensor * ffn_gate_inp;
-    struct ggml_tensor * ffn_gate_exps;
-    struct ggml_tensor * ffn_down_exps;
-    struct ggml_tensor * ffn_up_exps ;
-
-    // ff shared expert (shexp)
-    struct ggml_tensor * ffn_gate_inp_shexp;
-    struct ggml_tensor * ffn_gate_shexp;
-    struct ggml_tensor * ffn_down_shexp;
-    struct ggml_tensor * ffn_up_shexp;
-
-    // ff bias
-    struct ggml_tensor * ffn_gate_b;
-    struct ggml_tensor * ffn_down_b; // b2
-    struct ggml_tensor * ffn_up_b; // b3
-    struct ggml_tensor * ffn_act;
-
-    // mamba proj
-    struct ggml_tensor * ssm_in;
-    struct ggml_tensor * ssm_x;
-    struct ggml_tensor * ssm_dt;
-    struct ggml_tensor * ssm_out;
-
-    // mamba
-    struct ggml_tensor * ssm_conv1d;
-    struct ggml_tensor * ssm_a;
-    struct ggml_tensor * ssm_d;
-
-    // mamba bias
-    struct ggml_tensor * ssm_conv1d_b;
-    struct ggml_tensor * ssm_dt_b;
-
-    // rwkv
-    struct ggml_tensor * time_mix_w1;
-    struct ggml_tensor * time_mix_w2;
-    struct ggml_tensor * time_mix_lerp_x;
-    struct ggml_tensor * time_mix_lerp_w;
-    struct ggml_tensor * time_mix_lerp_k;
-    struct ggml_tensor * time_mix_lerp_v;
-    struct ggml_tensor * time_mix_lerp_r;
-    struct ggml_tensor * time_mix_lerp_g;
-
-    struct ggml_tensor * time_mix_first;
-    struct ggml_tensor * time_mix_decay;
-    struct ggml_tensor * time_mix_decay_w1;
-    struct ggml_tensor * time_mix_decay_w2;
-    struct ggml_tensor * time_mix_key;
-    struct ggml_tensor * time_mix_value;
-    struct ggml_tensor * time_mix_receptance;
-    struct ggml_tensor * time_mix_gate;
-
-    struct ggml_tensor * time_mix_ln;
-    struct ggml_tensor * time_mix_ln_b;
-    struct ggml_tensor * time_mix_output;
-
-    struct ggml_tensor * channel_mix_lerp_k;
-    struct ggml_tensor * channel_mix_lerp_r;
-
-    struct ggml_tensor * channel_mix_key;
-    struct ggml_tensor * channel_mix_receptance;
-    struct ggml_tensor * channel_mix_value;
-
-    // long rope factors
-    struct ggml_tensor * rope_long  = nullptr;
-    struct ggml_tensor * rope_short = nullptr;
-    struct ggml_tensor * rope_freqs = nullptr;
-
-    // bitnet scale
-    struct ggml_tensor * wq_scale;
-    struct ggml_tensor * wk_scale;
-    struct ggml_tensor * wv_scale;
-    struct ggml_tensor * wo_scale;
-    struct ggml_tensor * ffn_gate_scale;
-    struct ggml_tensor * ffn_up_scale;
-    struct ggml_tensor * ffn_down_scale;
-};
-
-// very similar to llama_batch,
-// but has more metadata about sequences
-struct llama_ubatch {
-    bool equal_seqs;
-    // TODO: whole_seqs for embeddings?
-
-    uint32_t n_tokens; // total tokens (n_seq_tokens * n_seqs)
-    uint32_t n_seq_tokens; // tokens per sequence
-    uint32_t n_seqs;
-
-    llama_token  *  token;    // [n_tokens]
-    float        *  embd;     // [n_embd, n_tokens]
-    llama_pos    *  pos;      // [n_tokens]
-    int32_t      *  n_seq_id; // [n_seqs]
-    llama_seq_id ** seq_id;   // [n_seqs]
-    int8_t       *  output;   // [n_tokens]
-};
-
-struct llama_kv_cell {
-    llama_pos pos   = -1;
-    llama_pos delta = 0;
-    int32_t   src   = -1; // used by recurrent state models to copy states
-    int32_t   tail  = -1;
-
-    std::set seq_id;
-
-    bool has_seq_id(const llama_seq_id & id) const {
-        return seq_id.find(id) != seq_id.end();
-    }
-
-    bool is_empty() const {
-        return seq_id.empty();
-    }
-
-    bool is_same_seq(const llama_kv_cell & other) const {
-        return seq_id == other.seq_id;
-    }
-};
-
-// ring-buffer of cached KV data
-struct llama_kv_cache {
-    bool has_shift = false;
-    bool do_defrag = false;
-    bool recurrent = false; // with recurrent state models, a cell can hold the state for more than one past token
-    bool v_trans   = true;  // the value tensor is transposed
-
-    // Note: The value of head isn't only used to optimize searching
-    // for a free KV slot. llama_decode_internal also uses it, so it
-    // cannot be freely changed after a slot has been allocated.
-    uint32_t head = 0;
-    uint32_t size = 0;
-    uint32_t used = 0; // used cells (i.e. at least one seq_id)
-
-    // computed before each graph build
-    uint32_t n = 0;
-
-    ggml_type type_k = GGML_TYPE_F16;
-    ggml_type type_v = GGML_TYPE_F16;
-
-    std::vector cells;
-
-    std::vector k_l; // per layer
-    std::vector v_l;
-
-    std::vector ctxs;
-    std::vector bufs;
-
-    size_t total_size() {
-        size_t size = 0;
-        for (auto & buf : bufs) {
-            size += ggml_backend_buffer_get_size(buf.get());
-        }
-        return size;
-    }
-};
-
-struct llama_control_vector {
-    std::vector tensors; // per layer
-    std::vector ctxs;
-    std::vector bufs;
-
-    int32_t layer_start = -1;
-    int32_t layer_end   = -1;
-
-    struct ggml_tensor * tensor_for(int il) const {
-        if (il < 0 || il < layer_start || il > layer_end || (size_t) il >= tensors.size()) {
-            return nullptr;
-        }
-        return tensors[il];
-    }
-
-    struct ggml_tensor * apply_to(struct ggml_context * ctx, struct ggml_tensor * cur, int  il) const {
-        ggml_tensor * layer_dir = tensor_for(il);
-        if (layer_dir != nullptr) {
-            cur = ggml_add(ctx, cur, layer_dir);
-        }
-        return cur;
-    }
-};
-
-struct llama_model {
-    e_model     type  = MODEL_UNKNOWN;
-    llm_arch    arch  = LLM_ARCH_UNKNOWN;
-    llama_ftype ftype = LLAMA_FTYPE_ALL_F32;
-
-    std::string name = "n/a";
-
-    llama_hparams hparams = {};
-    llama_vocab   vocab;
-
-    struct ggml_tensor * tok_embd = nullptr;
-    struct ggml_tensor * type_embd = nullptr;
-    struct ggml_tensor * pos_embd = nullptr;
-    struct ggml_tensor * tok_norm = nullptr;
-    struct ggml_tensor * tok_norm_b = nullptr;
-
-    struct ggml_tensor * output_norm = nullptr;
-    struct ggml_tensor * output_norm_b = nullptr;
-    struct ggml_tensor * output = nullptr;
-    struct ggml_tensor * output_b = nullptr;
-    struct ggml_tensor * output_norm_enc = nullptr;
-
-    // classifier
-    struct ggml_tensor * cls = nullptr;
-    struct ggml_tensor * cls_b = nullptr;
-    struct ggml_tensor * cls_out   = nullptr;
-    struct ggml_tensor * cls_out_b = nullptr;
-
-    std::vector layers;
-
-    // gguf metadata
-    std::unordered_map gguf_kv;
-
-    llama_split_mode split_mode;
-    int main_gpu;
-    int n_gpu_layers;
-
-    std::vector rpc_servers;
-
-    // list of devices used in this model
-    std::vector devices;
-
-
-    // lists of buffer types used for each layer
-    using buft_list_t = std::vector>;
-    buft_list_t cpu_buft_list;
-    std::map gpu_buft_list;
-
-    struct layer_dev {
-        ggml_backend_dev_t dev;
-        buft_list_t * buft_list;
-    };
-    layer_dev dev_input = {};
-    layer_dev dev_output = {};
-    std::vector dev_layer;
-
-    // contexts where the model tensors metadata is stored
-    std::vector ctxs;
-
-    // the model memory buffers for the tensor data
-    std::vector bufs;
-
-    // model memory mapped files
-    llama_mmaps mappings;
-
-    // objects representing data potentially being locked in memory
-    llama_mlocks mlock_bufs;
-    llama_mlocks mlock_mmaps;
-
-    // for quantize-stats only
-    std::vector> tensors_by_name;
-
-    int64_t t_load_us = 0;
-    int64_t t_start_us = 0;
-
-    // keep track of loaded lora adapters
-    std::set lora_adapters;
-
-    ~llama_model() {
-       while (!lora_adapters.empty()) {
-            llama_lora_adapter_free(*lora_adapters.begin());
-        }
-    }
-};
-
-struct llama_sbatch_seq {
-    int32_t n_seq_id;
-    llama_seq_id * seq_id;
-    size_t offset;
-    size_t length;
-};
-
-// sequence-length-aware batch splitting
-struct llama_sbatch {
-    // tokens left in this batch
-    size_t n_tokens;
-
-    size_t n_embd;
-
-    bool logits_all; // TODO: remove once lctx.logits_all is removed too
-
-    // sorted indices into the batch
-    std::vector ids;
-    // batch indices of the output
-    std::vector out_ids;
-    std::vector seq;
-    const llama_batch * batch = nullptr;
-
-    // buffers for the ubatch
-    std::vector    ubatch_token;
-    std::vector          ubatch_embd;
-    std::vector      ubatch_pos;
-    std::vector        ubatch_n_seq_id;
-    std::vector ubatch_seq_id;
-    std::vector         ubatch_output;
-
-    llama_ubatch reserve_ubatch(size_t n_ubatch, bool has_embd = false) {
-        // clear empty sequences
-        // the previous ubatch is assumed to be gone,
-        // so nothing should refer to values in these sequences anymore.
-        for (size_t i = seq.size(); i-- > 0;) {
-            if (seq[i].length == 0) {
-                seq.pop_back();
-            } else {
-                break;
-            }
-        }
-        ubatch_token.resize(!has_embd ? n_ubatch : 0);
-        ubatch_embd.resize(has_embd ? n_embd * n_ubatch : 0);
-        ubatch_pos.resize(n_ubatch);
-        ubatch_n_seq_id.resize(n_ubatch);
-        ubatch_seq_id.resize(n_ubatch);
-        ubatch_output.resize(n_ubatch);
-        llama_ubatch ubatch = {
-            /*equal_seqs   =*/ true,
-            /*n_tokens     =*/ 0,
-            /*n_seq_tokens =*/ 0,
-            /*n_seqs       =*/ 0,
-            /*token        =*/ !has_embd ? ubatch_token.data() : nullptr,
-            /*embd         =*/ has_embd  ? ubatch_embd.data()  : nullptr,
-            /*pos          =*/ ubatch_pos.data(),
-            /*n_seq_id     =*/ ubatch_n_seq_id.data(),
-            /*seq_id       =*/ ubatch_seq_id.data(),
-            /*output       =*/ ubatch_output.data(),
-        };
-        return ubatch;
-    }
-
-    void add_seq_to_ubatch(llama_ubatch & ubatch, llama_sbatch_seq & seq, size_t length) {
-        GGML_ASSERT(batch != nullptr);
-        GGML_ASSERT(length <= seq.length);
-        // Can only add sequences of equal lengths to a batch,
-        // otherwise it isn't clear to which sequence a token belongs
-        GGML_ASSERT(seq.n_seq_id == 0 || ubatch.n_seqs == 0 || length == (size_t) ubatch.n_tokens / ubatch.n_seqs);
-        GGML_ASSERT((seq.n_seq_id != 0) == ubatch.equal_seqs);
-        // NOTE: loops are separated for cache-friendliness
-        if (batch->token) {
-            if (ubatch.equal_seqs) {
-                for (size_t i = 0; i < length; ++i) {
-                    ubatch.token[ubatch.n_tokens + i] = batch->token[ids[seq.offset + i]];
-                }
-            } else {
-                // simple split
-                ubatch.token = batch->token + seq.offset;
-            }
-        } else {
-            ubatch.token = nullptr;
-        }
-        if (batch->embd) {
-            if (ubatch.equal_seqs) {
-                for (size_t i = 0; i < length; ++i) {
-                    memcpy(
-                        ubatch.embd + n_embd * (ubatch.n_tokens + i),
-                        batch->embd + n_embd * ids[seq.offset + i],
-                        n_embd * sizeof(float)
-                    );
-                }
-            } else {
-                // simple split
-                ubatch.embd = batch->embd + (n_embd * seq.offset);
-            }
-        } else {
-            ubatch.embd = nullptr;
-        }
-        if (ubatch.equal_seqs) {
-            for (size_t i = 0; i < length; ++i) {
-                ubatch.pos[ubatch.n_tokens + i] = batch->pos[ids[seq.offset + i]];
-            }
-        } else {
-            // simple split
-            ubatch.pos = batch->pos + seq.offset;
-        }
-        if (ubatch.equal_seqs) {
-            ubatch.n_seq_id[ubatch.n_seqs] = seq.n_seq_id;
-            if (seq.seq_id) {
-                ubatch.seq_id[ubatch.n_seqs] = seq.seq_id;
-            }
-        } else {
-            // simple split
-            if (batch->n_seq_id) {
-                ubatch.n_seq_id = batch->n_seq_id + seq.offset;
-            } else {
-                for (size_t i = 0; i < length; ++i) {
-                    ubatch.n_seq_id[ubatch.n_seqs + i] = 1;
-                }
-            }
-            if (batch->seq_id) {
-                ubatch.seq_id = batch->seq_id + seq.offset;
-            }
-        }
-        if (logits_all) {
-            for (size_t i = 0; i < length; ++i) {
-                ubatch.output[ubatch.n_tokens + i] = 1;
-                out_ids.push_back(ids[seq.offset + i]);
-            }
-        } else if (batch->logits) {
-            if (ubatch.equal_seqs) {
-                for (size_t i = 0; i < length; ++i) {
-                    size_t id = ids[seq.offset + i];
-                    int8_t is_output = batch->logits[id];
-                    ubatch.output[ubatch.n_tokens + i] = is_output;
-                    if (is_output) { out_ids.push_back(id); }
-                }
-            } else {
-                // simple split
-                ubatch.output = batch->logits + seq.offset;
-                for (size_t i = 0; i < length; ++i) {
-                    if (ubatch.output[i] != 0) { out_ids.push_back(seq.offset + i); }
-                }
-            }
-        } else {
-            // only get last output
-            for (size_t i = 0; i < length; ++i) {
-                size_t id = ids[seq.offset + i];
-                int8_t is_last = id == ids.size() - 1;
-                ubatch.output[ubatch.n_tokens + i] = is_last;
-                if (is_last) { out_ids.push_back(id); }
-            }
-        }
-        if (ubatch.n_tokens == 0 && ubatch.n_seqs == 0) {
-            ubatch.n_seq_tokens = ubatch.equal_seqs ? length : 1;
-        }
-        ubatch.n_tokens += length;
-        ubatch.n_seqs += ubatch.equal_seqs ? 1 : length; // virtual sequences for simple splits
-        seq.offset += length;
-        seq.length -= length;
-        n_tokens -= length;
-        GGML_ASSERT(ubatch.n_tokens == ubatch.n_seq_tokens * ubatch.n_seqs);
-    }
-
-    // simple split, unknown number of sequences of unequal lengths
-    llama_ubatch split_simple(size_t n_ubatch) {
-        n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch;
-        llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr);
-        ubatch.equal_seqs = false;
-        if (!seq.empty()) {
-            llama_sbatch_seq & s = seq[0];
-            size_t length = s.length < n_ubatch ? s.length : n_ubatch;
-            GGML_ASSERT(seq.size() == 1 && s.n_seq_id == 0); // don't mix with other splits
-            add_seq_to_ubatch(ubatch, s, length);
-        }
-        return ubatch;
-    }
-
-    // make batches of equal-length sequences
-    llama_ubatch split_equal(size_t n_ubatch) {
-        n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch;
-        llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr);
-        if (!seq.empty()) {
-            size_t length = 0;
-            size_t n_tokens_in_ubatch = 0;
-            GGML_ASSERT(seq[0].n_seq_id > 0); // should not be mixed with simple splits
-            // smallest first, because it's easier to split this way;
-            // starting from the end to pop in constant time.
-            for (size_t i = seq.size(); i-- > 0;) {
-                llama_sbatch_seq & s = seq[i];
-                GGML_ASSERT(s.length > 0);
-                if (length == 0) {
-                    length = s.length < n_ubatch ? s.length : n_ubatch;
-                }
-                add_seq_to_ubatch(ubatch, s, length);
-                n_tokens_in_ubatch += length;
-                // shared prompts can't be mixed with any of their sequences,
-                // so it's safer to compute them in their own ubatch
-                if (s.n_seq_id > 1) { break; }
-                // stop when there isn't enough space for another sequence
-                if (length + n_tokens_in_ubatch > n_ubatch) { break; }
-            }
-        }
-        return ubatch;
-    }
-
-    // sequence-wise split
-    llama_ubatch split_seq(size_t n_ubatch) {
-        n_ubatch = n_tokens < n_ubatch ? n_tokens : n_ubatch;
-        llama_ubatch ubatch = reserve_ubatch(n_ubatch, /* has_embd */ batch->embd != nullptr);
-        if (!seq.empty()) {
-            llama_sbatch_seq & s = seq[seq.size() - 1];
-            size_t length = s.length < n_ubatch ? s.length : n_ubatch;
-            GGML_ASSERT(s.n_seq_id > 0); // should not be mixed with simple splits
-            add_seq_to_ubatch(ubatch, s, length);
-        }
-        return ubatch;
-    }
-
-    void from_batch(const llama_batch & batch, const size_t n_embd, const bool simple_split = false, const bool logits_all = false) {
-        GGML_ASSERT(batch.n_tokens >= 0);
-        this->batch = &batch;
-        this->n_embd = n_embd;
-        this->logits_all = logits_all;
-
-        n_tokens = batch.n_tokens;
-        ids.resize(n_tokens);
-        out_ids.clear();
-        // TODO: reserve out_ids and seq
-
-        for (size_t i = 0; i < n_tokens; ++i) {
-            ids[i] = i;
-        }
-        if (simple_split) {
-            seq.resize(1);
-            llama_sbatch_seq & s = seq[0];
-            s.n_seq_id = 0;
-            s.seq_id = nullptr;
-            s.offset = 0;
-            s.length = n_tokens;
-            return;
-        }
-        std::sort(ids.begin(), ids.end(),
-            [&batch](size_t a, size_t b) {
-                int32_t n_seq_a = batch.n_seq_id ? batch.n_seq_id[a] : 1;
-                int32_t n_seq_b = batch.n_seq_id ? batch.n_seq_id[b] : 1;
-                // sort by seq_id, then by pos
-                if (n_seq_a == n_seq_b) {
-                    if (batch.seq_id) {
-                        for (int32_t i = 0; i < n_seq_a; ++i) {
-                            llama_seq_id seq_id_a = batch.seq_id[a][i];
-                            llama_seq_id seq_id_b = batch.seq_id[b][i];
-                            // smaller seq_ids go first
-                            if (seq_id_a != seq_id_b) {
-                                return seq_id_a < seq_id_b;
-                            }
-                        }
-                    }
-                    // when all else is equal, sort by pos
-                    if (batch.pos) {
-                        return batch.pos[a] < batch.pos[b];
-                    }
-                    // no pos, sort by id
-                    return a < b;
-                }
-                // shared prompts go first
-                return n_seq_a > n_seq_b;
-            }
-        );
-        // init seq
-        llama_sbatch_seq * last_seq = nullptr;
-
-        for (size_t i = 0; i < n_tokens; ++i) {
-            const size_t bi = ids[i];
-            const int32_t n_seqs = batch.n_seq_id[bi];
-            llama_seq_id * seq_ids = batch.seq_id[bi];
-            if (last_seq != nullptr) {
-                bool same = n_seqs == last_seq->n_seq_id;
-                for (int32_t j = 0; same && j < n_seqs; ++j) {
-                    if (seq_ids[j] != last_seq->seq_id[j]) {
-                        same = false;
-                    }
-                }
-                if (same) {
-                    last_seq->length += 1;
-                    continue;
-                }
-            }
-            llama_sbatch_seq new_seq = {n_seqs, seq_ids, i, 1};
-            seq.push_back(new_seq);
-            last_seq = &seq.back();
-        }
-        // keep shared prompts first at the end, then sort by length descending.
-        std::sort(seq.begin(), seq.end(),
-            [](llama_sbatch_seq & a, llama_sbatch_seq & b) {
-                if (a.n_seq_id == b.n_seq_id) {
-                    return a.length > b.length;
-                }
-                return a.n_seq_id < b.n_seq_id;
-            }
-        );
-    }
-};
-
-struct llama_context {
-    llama_context(const llama_model & model)
-        : model(model)
-        , t_start_us(model.t_start_us)
-        , t_load_us(model.t_load_us) {}
-
-    const struct llama_model & model;
-
-    struct llama_cparams        cparams;
-    struct llama_sbatch         sbatch;
-    struct llama_kv_cache       kv_self;
-    struct llama_control_vector cvec;
-
-    std::unordered_map lora_adapters;
-
-    std::vector backends;
-    std::vector> set_n_threads_fns;
-
-    ggml_backend_t backend_cpu = nullptr;
-
-    ggml_threadpool_t threadpool       = nullptr;
-    ggml_threadpool_t threadpool_batch = nullptr;
-
-    bool has_evaluated_once = false;
-
-    mutable int64_t t_start_us;
-    mutable int64_t t_load_us;
-    mutable int64_t t_p_eval_us = 0;
-    mutable int64_t t_eval_us   = 0;
-
-    mutable int64_t t_compute_start_us = 0;
-    mutable int64_t n_queued_tokens = 0;
-
-    mutable int32_t n_p_eval = 0; // number of tokens in eval calls for the prompt (with batch size > 1)
-    mutable int32_t n_eval   = 0; // number of eval calls
-
-    // host buffer for the model output (logits and embeddings)
-    ggml_backend_buffer_ptr buf_output;
-
-    // decode output (2-dimensional array: [n_outputs][n_vocab])
-    size_t  logits_size = 0; // capacity (of floats) for logits
-    float * logits      = nullptr;
-
-    std::vector output_ids; // map batch token positions to ids of the logits and embd buffers
-    size_t  output_size = 0; // capacity (of tokens positions) for the output buffers
-    int32_t n_outputs   = 0; // number of actually-used outputs in the current ubatch or last logical batch
-
-    bool logits_all = false;
-
-    // embeddings output (2-dimensional array: [n_outputs][n_embd])
-    // populated only when pooling_type == LLAMA_POOLING_TYPE_NONE
-    size_t  embd_size = 0; // capacity (of floats) for embeddings
-    float * embd      = nullptr;
-
-    // sequence embeddings output (map of [n_embd] vectors)
-    // populated only when pooling_type != LLAMA_POOLING_TYPE_NONE
-    std::map> embd_seq;
-
-    // whether we are computing encoder output or decoder output
-    bool is_encoding = false;
-
-    // output of the encoder part of the encoder-decoder models
-    std::vector embd_enc;
-    std::vector> seq_ids_enc;
-
-    // memory buffers used to evaluate the model
-    std::vector buf_compute_meta;
-    ggml_backend_sched_ptr sched;
-
-    ggml_abort_callback abort_callback      = nullptr;
-    void *              abort_callback_data = nullptr;
-
-    // input tensors
-    struct ggml_tensor * inp_tokens;      // I32 [n_batch]
-    struct ggml_tensor * inp_embd;        // F32 [n_embd, n_batch]
-    struct ggml_tensor * inp_pos;         // I32 [n_batch]
-    struct ggml_tensor * inp_out_ids;     // I32 [n_outputs]
-    struct ggml_tensor * inp_KQ_mask;     // F32 [kv_size, n_batch]
-    struct ggml_tensor * inp_KQ_mask_swa; // F32 [kv_size, n_batch]
-    struct ggml_tensor * inp_K_shift;     // I32 [kv_size]
-    struct ggml_tensor * inp_mean;        // F32 [n_batch, n_batch]
-    struct ggml_tensor * inp_cls;         // I32 [n_batch]
-    struct ggml_tensor * inp_s_copy;      // I32 [kv_size]
-    struct ggml_tensor * inp_s_mask;      // F32 [1, n_kv]
-    struct ggml_tensor * inp_s_seq;       // I32 [n_kv, n_batch]
-    struct ggml_tensor * inp_pos_bucket;    // I32 [n_batch|n_kv, n_batch]
-    struct ggml_tensor * inp_embd_enc;      // F32 [n_embd, n_outputs_enc]
-    struct ggml_tensor * inp_KQ_mask_cross; // F32 [n_outputs_enc, n_batch]
-};
-
-struct llama_lora_weight {
-    struct ggml_tensor * a = nullptr;
-    struct ggml_tensor * b = nullptr;
-    llama_lora_weight() = default;
-    llama_lora_weight(struct ggml_tensor * a, struct ggml_tensor * b): a(a), b(b) {}
-};
-
-struct llama_lora_adapter {
-    struct llama_model * base_model;
-    // map tensor name to lora_a_b
-    std::unordered_map ab_map;
-    std::vector ctxs;
-    std::vector bufs;
-
-    float alpha;
-
-    llama_lora_adapter(struct llama_model * base_model): base_model(base_model) {
-        base_model->lora_adapters.insert(this);
-    }
-
-    llama_lora_weight * get_weight(struct ggml_tensor * w) {
-        std::string name(w->name);
-        auto pos = ab_map.find(name);
-        if (ab_map.find(name) != ab_map.end()) {
-            return &pos->second;
-        }
-        return nullptr;
-    }
-
-    ~llama_lora_adapter() {
-        auto pos = base_model->lora_adapters.find(this);
-        if (pos != base_model->lora_adapters.end()) {
-            base_model->lora_adapters.erase(pos);
-        }
-    }
-};
-
-static int llama_get_device_count(const llama_model & model) {
-    return (int) model.devices.size();
-}
-
-template
-static bool buft_supported(ggml_backend_buffer_type_t buft, ggml_backend_dev_t dev, F & fn) {
-    ggml_init_params params = {
-        /*.mem_size   =*/ ggml_tensor_overhead()*8,
-        /*.mem_buffer =*/ NULL,
-        /*.no_alloc   =*/ true,
-    };
-    ggml_context_ptr ctx { ggml_init(params) };
-    if (!ctx) {
-        throw std::runtime_error(format("failed to create ggml context"));
-    }
-
-    ggml_backend_buffer_ptr buf { ggml_backend_buft_alloc_buffer(buft, 0) };
-    ggml_tensor * op_tensor = fn(ctx.get());
-    for (int i = 0; i < GGML_MAX_SRC; i++) {
-        if (op_tensor->src[i] != nullptr) {
-            assert(op_tensor->src[i]->buffer == nullptr);
-            op_tensor->src[i]->buffer = buf.get();
-        }
-    }
-    bool op_supported = ggml_backend_dev_supports_op(dev, op_tensor);
-
-    return op_supported;
-}
-
-template
-static ggml_backend_buffer_type_t select_buft(const llama_model::buft_list_t & buft_list, const F & fn) {
-    for (const auto & cur : buft_list) {
-        ggml_backend_dev_t cur_dev = cur.first;
-        ggml_backend_buffer_type_t cur_buft = cur.second;
-        if (buft_supported(cur_buft, cur_dev, fn)) {
-            return cur_buft;
-        }
-    }
-    throw std::runtime_error(format("no suitable buffer type found"));
-}
-
-//
-// kv cache helpers
-//
-
-static bool llama_kv_cache_init(
-             struct llama_kv_cache & cache,
-               const llama_context * ctx,
-                         ggml_type   type_k,
-                         ggml_type   type_v,
-                          uint32_t   kv_size,
-                              bool   offload) {
-    const llama_model & model = ctx->model;
-    const llama_cparams & cparams = ctx->cparams;
-
-    const struct llama_hparams & hparams = model.hparams;
-
-    const int64_t  n_layer = hparams.n_layer;
-
-    cache.has_shift = false;
-
-    cache.recurrent = llama_model_is_recurrent(&model);
-    cache.v_trans   = !cache.recurrent && !cparams.flash_attn;
-
-    cache.head = 0;
-    cache.size = kv_size;
-    cache.used = 0;
-
-    cache.type_k = type_k;
-    cache.type_v = type_v;
-
-    cache.cells.clear();
-    cache.cells.resize(kv_size);
-
-    // create a context for each buffer type
-    std::map ctx_map;
-    auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * {
-        auto it = ctx_map.find(buft);
-        if (it == ctx_map.end()) {
-            struct ggml_init_params params = {
-                /*.mem_size   =*/ size_t(2u*n_layer*ggml_tensor_overhead()),
-                /*.mem_buffer =*/ NULL,
-                /*.no_alloc   =*/ true,
-            };
-            ggml_context * ctx = ggml_init(params);
-            if (!ctx) {
-                return nullptr;
-            }
-            ctx_map[buft] = ctx;
-            cache.ctxs.emplace_back(ctx);
-            return ctx;
-        }
-        return it->second;
-    };
-
-    cache.k_l.reserve(n_layer);
-    cache.v_l.reserve(n_layer);
-
-    for (int i = 0; i < (int) n_layer; i++) {
-        const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(i) + hparams.n_embd_k_s();
-        const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(i) + hparams.n_embd_v_s();
-
-        const llama_model::buft_list_t * buft_list;
-        if (offload) {
-            buft_list = model.dev_layer.at(i).buft_list;
-        } else {
-            buft_list = &model.cpu_buft_list;
-        }
-        ggml_backend_buffer_type_t buft = select_buft(*buft_list,
-            [&](ggml_context * ctx) {
-                ggml_tensor * k = ggml_new_tensor_1d(ctx, type_k, n_embd_k_gqa*kv_size);
-                if (hparams.rope_type == LLAMA_ROPE_TYPE_NONE) {
-                    return k;
-                }
-                ggml_tensor * p = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 1);
-                return ggml_rope(ctx, k, p, hparams.n_rot, hparams.rope_type);
-            });
-        ggml_context * ctx = ctx_for_buft(buft);
-
-        if (!ctx) {
-            LLAMA_LOG_ERROR("%s: failed to create ggml context for kv cache\n", __func__);
-            return false;
-        }
-
-        ggml_tensor * k = ggml_new_tensor_1d(ctx, type_k, n_embd_k_gqa*kv_size);
-        ggml_tensor * v = ggml_new_tensor_1d(ctx, type_v, n_embd_v_gqa*kv_size);
-        ggml_format_name(k, "cache_k_l%d", i);
-        ggml_format_name(v, "cache_v_l%d", i);
-        cache.k_l.push_back(k);
-        cache.v_l.push_back(v);
-    }
-
-    // allocate tensors and initialize the buffers to avoid NaNs in the padding
-    for (auto it : ctx_map) {
-        auto * buft = it.first;
-        auto * ctx  = it.second;
-
-        ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft);
-        if (!buf) {
-            LLAMA_LOG_ERROR("%s: failed to allocate buffer for kv cache\n", __func__);
-            return false;
-        }
-        ggml_backend_buffer_clear(buf, 0);
-        LLAMA_LOG_INFO("%s: %10s KV buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf), ggml_backend_buffer_get_size(buf)/1024.0/1024.0);
-        cache.bufs.emplace_back(buf);
-    }
-
-    return true;
-}
-
-// find an empty slot of size "n_tokens" in the cache
-// updates the cache head
-// Note: On success, it's important that cache.head points
-// to the first cell of the slot.
-static bool llama_kv_cache_find_slot(
-           struct llama_kv_cache & cache,
-       const struct llama_ubatch & batch) {
-    const uint32_t n_tokens = batch.n_tokens;
-    const uint32_t n_seqs   = batch.n_seqs;
-    const uint32_t n_seq_tokens = batch.n_seq_tokens;
-
-    if (cache.recurrent) {
-        // For recurrent state architectures (like Mamba or RWKV),
-        // each cache cell can store the state for a whole sequence.
-        // A slot should be always be contiguous.
-
-        // can only process batches with an equal number of new tokens in each sequence
-        GGML_ASSERT(batch.equal_seqs);
-
-        int32_t min = cache.size - 1;
-        int32_t max = 0;
-
-        // everything should fit if all seq_ids are smaller than the max
-        for (uint32_t s = 0; s < n_seqs; ++s) {
-            const uint32_t n_seq_id = batch.n_seq_id[s];
-            for (uint32_t j = 0; j < n_seq_id; ++j) {
-                const llama_seq_id seq_id = batch.seq_id[s][j];
-
-                if (seq_id < 0 || (uint32_t) seq_id >= cache.size) {
-                    // too big seq_id
-                    // TODO: would it be possible to resize the cache instead?
-                    LLAMA_LOG_ERROR("%s: seq_id=%d >= n_seq_max=%d Try using a bigger --parallel value\n", __func__, seq_id, cache.size);
-                    return false;
-                }
-                if (j > 0) {
-                    llama_kv_cell & seq = cache.cells[seq_id];
-                    if (seq.tail >= 0) {
-                        llama_kv_cell & cell = cache.cells[seq.tail];
-                        // clear cells from seq_ids that become shared
-                        // (should not normally happen, but let's handle it anyway)
-                        cell.seq_id.erase(seq_id);
-                        seq.tail = -1;
-                        if (cell.seq_id.empty()) {
-                            cell.pos = -1;
-                            cell.src = -1;
-                            cache.used -= 1;
-                        }
-                    }
-                }
-            }
-        }
-
-#ifndef NDEBUG
-        {
-            std::vector tails_verif;
-            tails_verif.assign(cache.size, -1);
-            for (uint32_t i = 0; i < cache.size; ++i) {
-                llama_kv_cell & cell = cache.cells[i];
-                for (llama_seq_id seq_id : cell.seq_id) {
-                    if (tails_verif[seq_id] != -1) {
-                        LLAMA_LOG_ERROR("%s: duplicate tail for seq_id %d in cell %d and %d\n", __func__, seq_id, i, tails_verif[seq_id]);
-                    }
-                    tails_verif[seq_id] = i;
-                }
-            }
-            for (uint32_t i = 0; i < cache.size; ++i) {
-                if (tails_verif[i] != cache.cells[i].tail) {
-                    LLAMA_LOG_ERROR("%s: wrong tail for seq_id %d, (%d instead of %d)\n", __func__, i, cache.cells[i].tail, tails_verif[i]);
-                }
-            }
-        }
-#endif
-
-        // find next empty cell
-        uint32_t next_empty_cell = cache.head;
-
-        for (uint32_t i = 0; i < cache.size; ++i) {
-            if (next_empty_cell >= cache.size) { next_empty_cell -= cache.size; }
-            llama_kv_cell & cell = cache.cells[next_empty_cell];
-            if (cell.is_empty()) { break; }
-            next_empty_cell += 1;
-        }
-
-        // find usable cell range
-        for (uint32_t s = 0; s < n_seqs; ++s) {
-            const llama_seq_id seq_id = batch.seq_id[s][0];
-            llama_kv_cell & seq_meta = cache.cells[seq_id];
-            bool has_cell = false;
-            if (seq_meta.tail >= 0) {
-                llama_kv_cell & cell = cache.cells[seq_meta.tail];
-                GGML_ASSERT(cell.has_seq_id(seq_id));
-                // does this seq_id "own" the cell?
-                if (cell.seq_id.size() == 1) { has_cell = true; }
-            }
-            if (!has_cell) {
-                llama_kv_cell & empty_cell = cache.cells[next_empty_cell];
-                GGML_ASSERT(empty_cell.is_empty());
-                // copy old tail into the empty cell
-                if (seq_meta.tail >= 0) {
-                    llama_kv_cell & orig_cell = cache.cells[seq_meta.tail];
-                    empty_cell.pos = orig_cell.pos;
-                    empty_cell.src = orig_cell.src;
-                    orig_cell.seq_id.erase(seq_id);
-                    empty_cell.seq_id.insert(seq_id); // will be overwritten
-                }
-                seq_meta.tail = next_empty_cell;
-                // find next empty cell
-                if (s + 1 < n_seqs) {
-                    next_empty_cell += 1;
-                    for (uint32_t i = 0; i < cache.size; ++i) {
-                        if (next_empty_cell >= cache.size) { next_empty_cell -= cache.size; }
-                        llama_kv_cell & cell = cache.cells[next_empty_cell];
-                        if (cell.is_empty()) { break; }
-                        next_empty_cell += 1;
-                    }
-                }
-            }
-            if (min > seq_meta.tail) { min = seq_meta.tail; }
-            if (max < seq_meta.tail) { max = seq_meta.tail; }
-        }
-
-        // gather and re-order
-        for (uint32_t s = 0; s < n_seqs; ++s) {
-            int32_t dst_id = s + min;
-            int32_t src_id = cache.cells[batch.seq_id[s][0]].tail;
-            if (dst_id != src_id) {
-                llama_kv_cell & dst_cell = cache.cells[dst_id];
-                llama_kv_cell & src_cell = cache.cells[src_id];
-
-                std::swap(dst_cell.pos, src_cell.pos);
-                std::swap(dst_cell.src, src_cell.src);
-                std::swap(dst_cell.seq_id, src_cell.seq_id);
-
-                // swap tails (assuming they NEVER overlap)
-                for (const llama_seq_id seq_id : src_cell.seq_id) {
-                    cache.cells[seq_id].tail = src_id;
-                }
-                for (const llama_seq_id seq_id : dst_cell.seq_id) {
-                    cache.cells[seq_id].tail = dst_id;
-                }
-            }
-        }
-
-        // update the pos of the used seqs
-        for (uint32_t s = 0; s < n_seqs; ++s) {
-            const llama_pos last_pos = batch.pos[n_seq_tokens * s + n_seq_tokens - 1];
-            int32_t cell_id = s + min;
-            llama_kv_cell & cell = cache.cells[cell_id];
-
-            if (cell.pos >= 0 && last_pos != cell.pos + (llama_pos) n_seq_tokens) {
-                // What should happen when the pos backtracks or skips a value?
-                // Clearing the state mid-batch would require special-casing which isn't done.
-                LLAMA_LOG_WARN("%s: non-consecutive token position %d after %d for sequence %d with %u new tokens\n",
-                    __func__, last_pos, cell.pos, batch.seq_id[s][0], n_seq_tokens);
-            }
-            cell.pos = last_pos;
-            cell.seq_id.clear();
-            for (int32_t j = 0; j < batch.n_seq_id[s]; ++j) {
-                const llama_seq_id seq_id = batch.seq_id[s][j];
-                cell.seq_id.insert(seq_id);
-                cache.cells[seq_id].tail = cell_id;
-            }
-        }
-
-        // allow getting the range of used cells, from head to head + n
-        cache.head = min;
-        cache.n    = max - min + 1;
-
-        // sanity check
-        return cache.n >= n_seqs;
-    }
-    // otherwise, one cell per token.
-
-    if (n_tokens > cache.size) {
-        LLAMA_LOG_ERROR("%s: n_tokens=%d > cache.size=%d\n", __func__, n_tokens, cache.size);
-        return false;
-    }
-
-    uint32_t n_tested = 0;
-
-    while (true) {
-        if (cache.head + n_tokens > cache.size) {
-            n_tested += cache.size - cache.head;
-            cache.head = 0;
-            continue;
-        }
-
-        bool found = true;
-        for (uint32_t i = 0; i < n_tokens; i++) {
-            if (cache.cells[cache.head + i].pos >= 0) {
-                found = false;
-                cache.head += i + 1;
-                n_tested   += i + 1;
-                break;
-            }
-        }
-
-        if (found) {
-            break;
-        }
-
-        if (n_tested >= cache.size) {
-            //LLAMA_LOG_ERROR("%s: failed to find a slot for %d tokens\n", __func__, n_tokens);
-            return false;
-        }
-    }
-
-    for (uint32_t s = 0; s < n_seqs; s++) {
-        for (uint32_t i = 0; i < n_seq_tokens; ++i) {
-            uint32_t k = s*n_seq_tokens + i;
-            cache.cells[cache.head + k].pos = batch.pos[k];
-
-            for (int32_t j = 0; j < batch.n_seq_id[s]; j++) {
-                cache.cells[cache.head + k].seq_id.insert(batch.seq_id[s][j]);
-            }
-        }
-    }
-
-    cache.used += n_tokens;
-
-    return true;
-}
-
-// find how many cells are currently in use
-static uint32_t llama_kv_cache_cell_max(const struct llama_kv_cache & cache) {
-    for (uint32_t i = cache.size; i > 0; --i) {
-        const llama_kv_cell & cell = cache.cells[i - 1];
-
-        if (cell.pos >= 0 && !cell.is_empty()) {
-            return i;
-        }
-    }
-
-    return 0;
-}
-
-static void llama_kv_cache_clear(struct llama_kv_cache & cache) {
-    for (int32_t i = 0; i < (int32_t) cache.size; ++i) {
-        cache.cells[i].pos = -1;
-        cache.cells[i].seq_id.clear();
-        cache.cells[i].src = -1;
-        cache.cells[i].tail = -1;
-    }
-    cache.head = 0;
-    cache.used = 0;
-
-    for (auto & buf : cache.bufs) {
-        ggml_backend_buffer_clear(buf.get(), 0);
-    }
-}
-
-static bool llama_kv_cache_seq_rm(
-        struct llama_kv_cache & cache,
-                 llama_seq_id   seq_id,
-                    llama_pos   p0,
-                    llama_pos   p1) {
-    uint32_t new_head = cache.size;
-
-    if (p0 < 0) p0 = 0;
-    if (p1 < 0) p1 = std::numeric_limits::max();
-
-    // models like Mamba or RWKV can't have a state partially erased
-    if (cache.recurrent) {
-        if (seq_id >= (int64_t) cache.size) {
-            // could be fatal
-            return false;
-        }
-        if (0 <= seq_id) {
-            int32_t & tail_id = cache.cells[seq_id].tail;
-            if (tail_id >= 0) {
-                const llama_kv_cell & cell = cache.cells[tail_id];
-                // partial intersection is invalid
-                if ((0 < p0 && p0 <= cell.pos) || (0 < p1 && p1 <= cell.pos)) {
-                    return false;
-                }
-                // invalidate tails which will be cleared
-                if (p0 <= cell.pos && cell.pos < p1) {
-                    tail_id = -1;
-                }
-            }
-        } else {
-            // seq_id is negative, then the range should include everything or nothing
-            if (p0 != p1 && (p0 != 0 || p1 != std::numeric_limits::max())) {
-                return false;
-            }
-        }
-    }
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) {
-            if (seq_id < 0) {
-                cache.cells[i].seq_id.clear();
-            } else if (cache.cells[i].has_seq_id(seq_id)) {
-                cache.cells[i].seq_id.erase(seq_id);
-            } else {
-                continue;
-            }
-            if (cache.cells[i].is_empty()) {
-                // keep count of the number of used cells
-                if (cache.cells[i].pos >= 0) cache.used--;
-
-                cache.cells[i].pos = -1;
-                cache.cells[i].src = -1;
-                if (new_head == cache.size) new_head = i;
-            }
-        }
-    }
-
-    // If we freed up a slot, set head to it so searching can start there.
-    if (new_head != cache.size && new_head < cache.head) cache.head = new_head;
-
-    return true;
-}
-
-static void llama_kv_cache_seq_cp(
-        struct llama_kv_cache & cache,
-                 llama_seq_id   seq_id_src,
-                 llama_seq_id   seq_id_dst,
-                    llama_pos   p0,
-                    llama_pos   p1) {
-    if (p0 < 0) p0 = 0;
-    if (p1 < 0) p1 = std::numeric_limits::max();
-
-    if (cache.recurrent) {
-        if ((uint32_t) seq_id_dst < cache.size && (uint32_t) seq_id_src < cache.size) {
-            llama_kv_cell & tail_src = cache.cells[seq_id_src];
-            llama_kv_cell & tail_dst = cache.cells[seq_id_dst];
-            if (tail_dst.tail >= 0) {
-                // clear destination seq_id if it wasn't empty
-                llama_kv_cell & cell_dst = cache.cells[tail_dst.tail];
-
-                cell_dst.seq_id.erase(seq_id_dst);
-                tail_dst.tail = -1;
-                if (cell_dst.seq_id.empty()) {
-                    cell_dst.pos = -1;
-                    cell_dst.delta = -1;
-                    cell_dst.src = -1;
-                    cache.used -= 1;
-                }
-            }
-            if (tail_src.tail >= 0) {
-                llama_kv_cell & cell_src = cache.cells[tail_src.tail];
-
-                cell_src.seq_id.insert(seq_id_dst);
-                tail_dst.tail = tail_src.tail;
-            }
-        }
-
-        return;
-    }
-    // otherwise, this is the KV cache of a Transformer-like model
-
-    cache.head = 0;
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.cells[i].has_seq_id(seq_id_src) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) {
-            cache.cells[i].seq_id.insert(seq_id_dst);
-        }
-    }
-}
-
-static void llama_kv_cache_seq_keep(struct llama_kv_cache & cache, llama_seq_id seq_id) {
-    uint32_t new_head = cache.size;
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.recurrent && (llama_seq_id) i != seq_id) {
-            cache.cells[i].tail = -1;
-        }
-        if (!cache.cells[i].has_seq_id(seq_id)) {
-            if (cache.cells[i].pos >= 0) cache.used--;
-            cache.cells[i].pos = -1;
-            cache.cells[i].src = -1;
-            cache.cells[i].seq_id.clear();
-            if (new_head == cache.size) new_head = i;
-        } else {
-            cache.cells[i].seq_id.clear();
-            cache.cells[i].seq_id.insert(seq_id);
-        }
-    }
-
-    // If we freed up a slot, set head to it so searching can start there.
-    if (new_head != cache.size && new_head < cache.head) cache.head = new_head;
-}
-
-static void llama_kv_cache_seq_add(
-        struct llama_kv_cache & cache,
-                 llama_seq_id   seq_id,
-                    llama_pos   p0,
-                    llama_pos   p1,
-                    llama_pos   delta) {
-    uint32_t new_head = cache.size;
-
-    if (p0 < 0) p0 = 0;
-    if (p1 < 0) p1 = std::numeric_limits::max();
-    // If there is no range then return early to avoid looping over the cache.
-    if (p0 == p1) return;
-
-    if (cache.recurrent) {
-        // for Mamba-like or RWKV models, only the pos needs to be shifted
-        if (0 <= seq_id && seq_id < (int64_t) cache.size) {
-            const int32_t tail_id = cache.cells[seq_id].tail;
-            if (tail_id >= 0) {
-                llama_kv_cell & cell = cache.cells[tail_id];
-                if (cell.has_seq_id(seq_id) && p0 <= cell.pos && cell.pos < p1) {
-                    cell.pos += delta;
-                }
-            }
-        }
-        return;
-    }
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.cells[i].has_seq_id(seq_id) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) {
-            cache.has_shift = true;
-            cache.cells[i].pos   += delta;
-            cache.cells[i].delta += delta;
-
-            if (cache.cells[i].pos < 0) {
-                if (!cache.cells[i].is_empty()) {
-                    cache.used--;
-                }
-                cache.cells[i].pos = -1;
-                cache.cells[i].seq_id.clear();
-                if (new_head == cache.size) {
-                    new_head = i;
-                }
-            }
-        }
-    }
-
-    // If we freed up a slot, set head to it so searching can start there.
-    // Otherwise we just start the next search from the beginning.
-    cache.head = new_head != cache.size ? new_head : 0;
-}
-
-static void llama_kv_cache_seq_div(
-        struct llama_kv_cache & cache,
-                 llama_seq_id   seq_id,
-                    llama_pos   p0,
-                    llama_pos   p1,
-                          int   d) {
-    if (p0 < 0) p0 = 0;
-    if (p1 < 0) p1 = std::numeric_limits::max();
-    // If there is no range then return early to avoid looping over the cache.
-    if (p0 == p1) return;
-
-    if (cache.recurrent) {
-        // for Mamba-like or RWKV models, only the pos needs to be changed
-        if (0 <= seq_id && seq_id < (int64_t) cache.size) {
-            const int32_t tail_id = cache.cells[seq_id].tail;
-            if (tail_id >= 0) {
-                llama_kv_cell & cell = cache.cells[tail_id];
-                if (cell.has_seq_id(seq_id) && p0 <= cell.pos && cell.pos < p1) {
-                    cell.pos /= d;
-                }
-            }
-        }
-        return;
-    }
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.cells[i].has_seq_id(seq_id) && cache.cells[i].pos >= p0 && cache.cells[i].pos < p1) {
-            cache.has_shift = true;
-
-            {
-                llama_pos p_old = cache.cells[i].pos;
-                cache.cells[i].pos   /= d;
-                cache.cells[i].delta += cache.cells[i].pos - p_old;
-            }
-        }
-    }
-}
-
-static llama_pos llama_kv_cache_seq_pos_max(struct llama_kv_cache & cache, llama_seq_id seq_id) {
-    llama_pos result = 0;
-
-    for (uint32_t i = 0; i < cache.size; ++i) {
-        if (cache.cells[i].has_seq_id(seq_id)) {
-            result = std::max(result, cache.cells[i].pos);
-        }
-    }
-
-    return result;
-}
-
-static void llama_kv_cache_defrag(struct llama_kv_cache & cache) {
-    if (!cache.recurrent) {
-        cache.do_defrag = true;
-    }
-}
-
-static uint32_t llama_kv_cache_get_padding(const struct llama_cparams & cparams) {
-    // the FA kernels require padding to avoid extra runtime boundary checks
-    return cparams.flash_attn ? 256u : 32u;
-}
-
-//
-// model loading and saving
-//
-
-enum llama_fver {
-    GGUF_FILE_VERSION_V1 = 1,
-    GGUF_FILE_VERSION_V2 = 2,
-    GGUF_FILE_VERSION_V3 = 3,
-};
-
-static const char * llama_file_version_name(llama_fver version) {
-    switch (version) {
-        case GGUF_FILE_VERSION_V1: return "GGUF V1 (support until nov 2023)";
-        case GGUF_FILE_VERSION_V2: return "GGUF V2";
-        case GGUF_FILE_VERSION_V3: return "GGUF V3 (latest)";
-    }
-
-    return "unknown";
-}
-
-static std::string llama_format_tensor_shape(const std::vector & ne) {
-    char buf[256];
-    snprintf(buf, sizeof(buf), "%5" PRId64, ne.at(0));
-    for (size_t i = 1; i < ne.size(); i++) {
-        snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %5" PRId64, ne.at(i));
-    }
-    return buf;
-}
-
-static std::string llama_format_tensor_shape(const struct ggml_tensor * t) {
-    char buf[256];
-    snprintf(buf, sizeof(buf), "%5" PRId64, t->ne[0]);
-    for (int i = 1; i < GGML_MAX_DIMS; i++) {
-        snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %5" PRId64, t->ne[i]);
-    }
-    return buf;
-}
-
-namespace GGUFMeta {
-    template 
-    struct GKV_Base_Type {
-        static constexpr gguf_type gt = gt_;
-
-        static T getter(const gguf_context * ctx, const int kid) {
-            return gfun(ctx, kid);
-        }
-    };
-
-    template struct GKV_Base;
-
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-    template<> struct GKV_Base: GKV_Base_Type {};
-
-    template<> struct GKV_Base {
-        static constexpr gguf_type gt = GGUF_TYPE_STRING;
-
-        static std::string getter(const gguf_context * ctx, const int kid) {
-            return gguf_get_val_str(ctx, kid);
-        }
-    };
-
-    struct ArrayInfo {
-        const gguf_type gt;
-        const size_t length;
-        const void * data;
-    };
-
-    template<> struct GKV_Base {
-        public:
-        static constexpr gguf_type gt = GGUF_TYPE_ARRAY;
-        static ArrayInfo getter(const gguf_context *ctx, const int k) {
-            return ArrayInfo {
-                gguf_get_arr_type(ctx, k),
-                size_t(gguf_get_arr_n(ctx, k)),
-                gguf_get_arr_data(ctx, k),
-            };
-        }
-    };
-
-    template
-    class GKV : public GKV_Base {
-        GKV() = delete;
-
-        public:
-        static T get_kv(const gguf_context * ctx, const int k) {
-            const enum gguf_type kt = gguf_get_kv_type(ctx, k);
-
-            if (kt != GKV::gt) {
-                throw std::runtime_error(format("key %s has wrong type %s but expected type %s",
-                    gguf_get_key(ctx, k), gguf_type_name(kt), gguf_type_name(GKV::gt)));
-            }
-            return GKV::getter(ctx, k);
-        }
-
-        static const char * override_type_to_str(const llama_model_kv_override_type ty) {
-            switch (ty) {
-                case LLAMA_KV_OVERRIDE_TYPE_BOOL:  return "bool";
-                case LLAMA_KV_OVERRIDE_TYPE_INT:   return "int";
-                case LLAMA_KV_OVERRIDE_TYPE_FLOAT: return "float";
-                case LLAMA_KV_OVERRIDE_TYPE_STR:   return "str";
-            }
-            return "unknown";
-        }
-
-        static bool validate_override(const llama_model_kv_override_type expected_type, const struct llama_model_kv_override * ovrd) {
-            if (!ovrd) { return false; }
-            if (ovrd->tag == expected_type) {
-                LLAMA_LOG_INFO("%s: Using metadata override (%5s) '%s' = ",
-                    __func__, override_type_to_str(ovrd->tag), ovrd->key);
-                switch (ovrd->tag) {
-                    case LLAMA_KV_OVERRIDE_TYPE_BOOL:  {
-                        LLAMA_LOG_INFO("%s\n", ovrd->val_bool ? "true" : "false");
-                    } break;
-                    case LLAMA_KV_OVERRIDE_TYPE_INT:   {
-                        LLAMA_LOG_INFO("%" PRId64 "\n", ovrd->val_i64);
-                    } break;
-                    case LLAMA_KV_OVERRIDE_TYPE_FLOAT: {
-                        LLAMA_LOG_INFO("%.6f\n", ovrd->val_f64);
-                    } break;
-                    case LLAMA_KV_OVERRIDE_TYPE_STR: {
-                        LLAMA_LOG_INFO("%s\n", ovrd->val_str);
-                    } break;
-                    default:
-                        // Shouldn't be possible to end up here, but just in case...
-                        throw std::runtime_error(
-                            format("Unsupported attempt to override %s type for metadata key %s\n",
-                                override_type_to_str(ovrd->tag), ovrd->key));
-                }
-                return true;
-            }
-            LLAMA_LOG_WARN("%s: Warning: Bad metadata override type for key '%s', expected %s but got %s\n",
-                __func__, ovrd->key, override_type_to_str(expected_type), override_type_to_str(ovrd->tag));
-            return false;
-        }
-
-        template
-        static typename std::enable_if::value, bool>::type
-        try_override(OT & target, const struct llama_model_kv_override * ovrd) {
-            if (validate_override(LLAMA_KV_OVERRIDE_TYPE_BOOL, ovrd)) {
-                target = ovrd->val_bool;
-                return true;
-            }
-            return false;
-        }
-
-        template
-        static typename std::enable_if::value && std::is_integral::value, bool>::type
-        try_override(OT & target, const struct llama_model_kv_override * ovrd) {
-            if (validate_override(LLAMA_KV_OVERRIDE_TYPE_INT, ovrd)) {
-                target = ovrd->val_i64;
-                return true;
-            }
-            return false;
-        }
-
-        template
-        static typename std::enable_if::value, bool>::type
-        try_override(T & target, const struct llama_model_kv_override * ovrd) {
-            if (validate_override(LLAMA_KV_OVERRIDE_TYPE_FLOAT, ovrd)) {
-                target = ovrd->val_f64;
-                return true;
-            }
-            return false;
-        }
-
-        template
-        static typename std::enable_if::value, bool>::type
-        try_override(T & target, const struct llama_model_kv_override * ovrd) {
-            if (validate_override(LLAMA_KV_OVERRIDE_TYPE_STR, ovrd)) {
-                target = ovrd->val_str;
-                return true;
-            }
-            return false;
-        }
-
-        static bool set(const gguf_context * ctx, const int k, T & target, const struct llama_model_kv_override * ovrd = nullptr) {
-            if (try_override(target, ovrd)) {
-                return true;
-            }
-            if (k < 0) { return false; }
-            target = get_kv(ctx, k);
-            return true;
-        }
-
-        static bool set(const gguf_context * ctx, const char * key, T & target, const struct llama_model_kv_override * ovrd = nullptr) {
-            return set(ctx, gguf_find_key(ctx, key), target, ovrd);
-        }
-
-        static bool set(const gguf_context * ctx, const std::string & key, T & target, const struct llama_model_kv_override * ovrd = nullptr) {
-            return set(ctx, key.c_str(), target, ovrd);
-        }
-    };
-}
-
-using llama_buf_map = std::unordered_map;
-
-static size_t llama_model_max_nodes(const llama_model & model) {
-    return std::max(8192, model.tensors_by_name.size()*5);
-}
-
-struct llama_model_loader {
-    int n_kv      = 0;
-    int n_tensors = 0;
-    int n_created = 0;
-
-    int64_t n_elements = 0;
-    size_t  n_bytes    = 0;
-
-    bool use_mmap = false;
-    bool check_tensors;
-
-    llama_files files;
-    llama_ftype ftype;
-    llama_fver  fver;
-
-    llama_mmaps mappings;
-
-    // Holds information on a model weight
-    struct llama_tensor_weight {
-        uint16_t  idx; // source file index
-        size_t   offs; // tensor data offset in the original file
-
-        ggml_tensor * tensor;
-
-        llama_tensor_weight(const llama_file * file, uint16_t idx, const struct gguf_context * gguf_ctx, ggml_tensor * tensor) : idx(idx), tensor(tensor) {
-            const int tensor_idx = gguf_find_tensor(gguf_ctx,  ggml_get_name(tensor));
-            if (tensor_idx < 0) {
-                throw std::runtime_error(format("tensor '%s' not found in the model", ggml_get_name(tensor)));
-            }
-
-            offs = gguf_get_data_offset(gguf_ctx) + gguf_get_tensor_offset(gguf_ctx, tensor_idx);
-            if (offs + ggml_nbytes(tensor) < offs || offs + ggml_nbytes(tensor) > file->size) {
-                throw std::runtime_error(format("tensor '%s' data is not within the file bounds, model is corrupted or incomplete", ggml_get_name(tensor)));
-            }
-        }
-    };
-
-    // custom comparator to sort weights more nicely by layer
-    struct weight_name_comparer {
-        bool operator()(const std::string & a, const std::string & b) const {
-            int a_layer = -1;
-            int b_layer = -1;
-            sscanf(a.c_str(), "blk.%d.", &a_layer);
-            sscanf(b.c_str(), "blk.%d.", &b_layer);
-            if (a_layer != b_layer) {
-                return a_layer < b_layer;
-            }
-            return a < b;
-        }
-    };
-
-    std::map weights_map;
-    std::unordered_map kv_overrides;
-
-    gguf_context_ptr meta;
-    std::vector contexts;
-
-    std::string arch_name;
-    LLM_KV      llm_kv    = LLM_KV(LLM_ARCH_UNKNOWN);
-
-    llama_model_loader(const std::string & fname, bool use_mmap, bool check_tensors, const struct llama_model_kv_override * param_overrides_p) {
-        int trace = 0;
-        if (getenv("LLAMA_TRACE")) {
-            trace = atoi(getenv("LLAMA_TRACE"));
-        }
-
-        if (param_overrides_p != nullptr) {
-            for (const struct llama_model_kv_override * p = param_overrides_p; p->key[0] != 0; p++) {
-                kv_overrides.insert({std::string(p->key), *p});
-            }
-        }
-
-        struct ggml_context * ctx = NULL;
-        struct gguf_init_params params = {
-            /*.no_alloc = */ true,
-            /*.ctx      = */ &ctx,
-        };
-
-        meta.reset(gguf_init_from_file(fname.c_str(), params));
-        if (!meta) {
-            throw std::runtime_error(format("%s: failed to load model from %s\n", __func__, fname.c_str()));
-        }
-
-        get_key(llm_kv(LLM_KV_GENERAL_ARCHITECTURE), arch_name, false);
-        llm_kv = LLM_KV(llm_arch_from_string(arch_name));
-
-        files.emplace_back(new llama_file(fname.c_str(), "rb"));
-        contexts.emplace_back(ctx);
-
-        // Save tensors data offset of the main file.
-        // For subsidiary files, `meta` tensor data offset must not be used,
-        // so we build a unified tensors index for weights.
-        for (ggml_tensor * cur = ggml_get_first_tensor(ctx); cur; cur = ggml_get_next_tensor(ctx, cur)) {
-            std::string tensor_name = std::string(cur->name);
-            // make sure there is no duplicated tensor names
-            if (weights_map.find(tensor_name) != weights_map.end()) {
-                throw std::runtime_error(format("invalid model: tensor '%s' is duplicated", ggml_get_name(cur)));
-            }
-            n_elements += ggml_nelements(cur);
-            n_bytes    += ggml_nbytes(cur);
-            weights_map.emplace(tensor_name, llama_tensor_weight(files.back().get(), 0, meta.get(), cur));
-        }
-        uint16_t n_split = 0;
-        get_key(llm_kv(LLM_KV_SPLIT_COUNT), n_split, false);
-
-        // Load additional GGML contexts
-        if (n_split > 1) {
-            uint16_t idx = 0;
-            get_key(llm_kv(LLM_KV_SPLIT_NO), idx);
-            if (idx != 0) {
-                throw std::runtime_error(format("illegal split file: %d, model must be loaded with the first split", idx));
-            }
-
-            char split_prefix[PATH_MAX] = {0};
-            if (!llama_split_prefix(split_prefix, sizeof(split_prefix), fname.c_str(), idx, n_split)) {
-                throw std::runtime_error(format("invalid split file: %s", fname.c_str()));
-            }
-
-            if (trace > 0) {
-                LLAMA_LOG_INFO("%s: loading additional %d GGUFs\n", __func__, n_split);
-            }
-
-            char split_path[PATH_MAX] = {0};
-            for (idx = 1; idx < n_split; idx++) {
-                llama_split_path(split_path, sizeof(split_path), split_prefix, idx, n_split);
-
-                struct gguf_init_params split_params = {
-                    /*.no_alloc = */ true,
-                    /*.ctx      = */ &ctx,
-                };
-                gguf_context_ptr ctx_gguf { gguf_init_from_file(split_path, split_params) };
-                if (!ctx_gguf) {
-                    throw std::runtime_error(format("%s: failed to load GGUF split from %s\n", __func__, split_path));
-                }
-
-                files.emplace_back(new llama_file(split_path, "rb"));
-                contexts.emplace_back(ctx);
-
-                // Save tensors data offset info of the shard.
-                for (ggml_tensor * cur = ggml_get_first_tensor(ctx); cur; cur = ggml_get_next_tensor(ctx, cur)) {
-                    std::string tensor_name = std::string(cur->name);
-                    // make sure there is no duplicated tensor names
-                    if (weights_map.find(tensor_name) != weights_map.end()) {
-                        throw std::runtime_error(format("invalid model: tensor '%s' is duplicated", ggml_get_name(cur)));
-                    }
-                    n_elements += ggml_nelements(cur);
-                    n_bytes    += ggml_nbytes(cur);
-                    weights_map.emplace(tensor_name, llama_tensor_weight(files.back().get(), idx, ctx_gguf.get(), cur));
-                }
-            }
-
-            get_key(llm_kv(LLM_KV_SPLIT_TENSORS_COUNT), n_tensors);
-
-            // sanity check
-            {
-                const int n_tensors_loaded = (int) weights_map.size();
-                if (n_tensors != n_tensors_loaded) {
-                    throw std::runtime_error(format("corrupted model: %d tensors expected but %d found", n_tensors, n_tensors_loaded));
-                }
-            }
-
-            LLAMA_LOG_INFO("%s: additional %d GGUFs metadata loaded.\n",  __func__, n_split - 1);
-        }
-
-        n_kv      = gguf_get_n_kv(meta.get());
-        n_tensors = weights_map.size();
-
-        fver = (enum llama_fver) gguf_get_version(meta.get());
-
-        LLAMA_LOG_INFO("%s: loaded meta data with %d key-value pairs and %d tensors from %s (version %s)\n",
-                __func__, n_kv, n_tensors, fname.c_str(), llama_file_version_name(fver));
-
-        // determine file type based on the number of tensors for each quantization and print meta data
-        // TODO: make optional
-        {
-            std::map n_type;
-
-            uint32_t n_type_max = 0;
-            enum ggml_type type_max = GGML_TYPE_F32;
-
-            for (const auto & it : weights_map) {
-                const llama_tensor_weight & w = it.second;
-                const ggml_tensor * tensor = w.tensor;
-
-                enum ggml_type type = tensor->type;
-
-                n_type[type]++;
-
-                if (n_type_max < n_type[type]) {
-                    n_type_max = n_type[type];
-                    type_max   = type;
-                }
-
-                if (trace > 0) {
-                    const uint16_t sid = w.idx;
-                    LLAMA_LOG_INFO("%s: - tensor split %2d: %32s %-8s [ %s ]\n", __func__, sid, ggml_get_name(tensor), ggml_type_name(type), llama_format_tensor_shape(tensor).c_str());
-                }
-            }
-
-            switch (type_max) {
-                case GGML_TYPE_F32:     ftype = LLAMA_FTYPE_ALL_F32;        break;
-                case GGML_TYPE_F16:     ftype = LLAMA_FTYPE_MOSTLY_F16;     break;
-                case GGML_TYPE_BF16:    ftype = LLAMA_FTYPE_MOSTLY_BF16;    break;
-                case GGML_TYPE_Q4_0:    ftype = LLAMA_FTYPE_MOSTLY_Q4_0;    break;
-                case GGML_TYPE_Q4_1:    ftype = LLAMA_FTYPE_MOSTLY_Q4_1;    break;
-                case GGML_TYPE_Q5_0:    ftype = LLAMA_FTYPE_MOSTLY_Q5_0;    break;
-                case GGML_TYPE_Q5_1:    ftype = LLAMA_FTYPE_MOSTLY_Q5_1;    break;
-                case GGML_TYPE_Q8_0:    ftype = LLAMA_FTYPE_MOSTLY_Q8_0;    break;
-                case GGML_TYPE_Q2_K:    ftype = LLAMA_FTYPE_MOSTLY_Q2_K;    break;
-                case GGML_TYPE_Q3_K:    ftype = LLAMA_FTYPE_MOSTLY_Q3_K_M;  break;
-                case GGML_TYPE_Q4_K:    ftype = LLAMA_FTYPE_MOSTLY_Q4_K_M;  break;
-                case GGML_TYPE_Q5_K:    ftype = LLAMA_FTYPE_MOSTLY_Q5_K_M;  break;
-                case GGML_TYPE_Q6_K:    ftype = LLAMA_FTYPE_MOSTLY_Q6_K;    break;
-                case GGML_TYPE_TQ1_0:   ftype = LLAMA_FTYPE_MOSTLY_TQ1_0;   break;
-                case GGML_TYPE_TQ2_0:   ftype = LLAMA_FTYPE_MOSTLY_TQ2_0;   break;
-                case GGML_TYPE_IQ2_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ2_XXS; break;
-                case GGML_TYPE_IQ2_XS:  ftype = LLAMA_FTYPE_MOSTLY_IQ2_XS;  break;
-                case GGML_TYPE_IQ2_S:   ftype = LLAMA_FTYPE_MOSTLY_IQ2_S;   break;
-                case GGML_TYPE_IQ3_XXS: ftype = LLAMA_FTYPE_MOSTLY_IQ3_XXS; break;
-                case GGML_TYPE_IQ1_S:   ftype = LLAMA_FTYPE_MOSTLY_IQ1_S;   break;
-                case GGML_TYPE_IQ1_M:   ftype = LLAMA_FTYPE_MOSTLY_IQ1_M;   break;
-                case GGML_TYPE_IQ4_NL:  ftype = LLAMA_FTYPE_MOSTLY_IQ4_NL;  break;
-                case GGML_TYPE_IQ4_XS:  ftype = LLAMA_FTYPE_MOSTLY_IQ4_XS;  break;
-                case GGML_TYPE_IQ3_S:   ftype = LLAMA_FTYPE_MOSTLY_IQ3_S;   break;
-                case GGML_TYPE_Q4_0_4_4: ftype = LLAMA_FTYPE_MOSTLY_Q4_0_4_4; break;
-                case GGML_TYPE_Q4_0_4_8: ftype = LLAMA_FTYPE_MOSTLY_Q4_0_4_8; break;
-                case GGML_TYPE_Q4_0_8_8: ftype = LLAMA_FTYPE_MOSTLY_Q4_0_8_8; break;
-                default:
-                    {
-                        LLAMA_LOG_WARN("%s: unknown type %s\n", __func__, ggml_type_name(type_max));
-                        ftype = LLAMA_FTYPE_ALL_F32;
-                    } break;
-            }
-
-            // this is a way to mark that we have "guessed" the file type
-            ftype = (llama_ftype) (ftype | LLAMA_FTYPE_GUESSED);
-
-            {
-                const int kid = gguf_find_key(meta.get(), "general.file_type"); // TODO: use LLM_KV
-                if (kid >= 0) {
-                    ftype = (llama_ftype) gguf_get_val_u32(meta.get(), kid);
-                }
-            }
-
-            LLAMA_LOG_INFO("%s: Dumping metadata keys/values. Note: KV overrides do not apply in this output.\n", __func__);
-
-            for (int i = 0; i < n_kv; i++) {
-                const char * name           = gguf_get_key(meta.get(), i);
-                const enum gguf_type type   = gguf_get_kv_type(meta.get(), i);
-                const std::string type_name =
-                    type == GGUF_TYPE_ARRAY
-                    ? format("%s[%s,%d]", gguf_type_name(type), gguf_type_name(gguf_get_arr_type(meta.get(), i)), gguf_get_arr_n(meta.get(), i))
-                    : gguf_type_name(type);
-
-                std::string value          = gguf_kv_to_str(meta.get(), i);
-                const size_t MAX_VALUE_LEN = 40;
-                if (value.size() > MAX_VALUE_LEN) {
-                    value = format("%s...", value.substr(0, MAX_VALUE_LEN - 3).c_str());
-                }
-                replace_all(value, "\n", "\\n");
-
-                LLAMA_LOG_INFO("%s: - kv %3d: %42s %-16s = %s\n", __func__, i, name, type_name.c_str(), value.c_str());
-            }
-
-            // print type counts
-            for (auto & kv : n_type) {
-                if (kv.second == 0) {
-                    continue;
-                }
-
-                LLAMA_LOG_INFO("%s: - type %4s: %4d tensors\n", __func__, ggml_type_name(kv.first), kv.second);
-            }
-        }
-
-        if (!llama_mmap::SUPPORTED) {
-            LLAMA_LOG_WARN("%s: mmap is not supported on this platform\n", __func__);
-            use_mmap = false;
-        }
-
-        this->use_mmap = use_mmap;
-        this->check_tensors = check_tensors;
-    }
-
-    template
-    typename std::enable_if::value, bool>::type
-    get_arr_n(const std::string & key, T & result, const bool required = true) {
-        const int kid = gguf_find_key(meta.get(), key.c_str());
-
-        if (kid < 0) {
-            if (required) {
-                throw std::runtime_error(format("key not found in model: %s", key.c_str()));
-            }
-            return false;
-        }
-
-        struct GGUFMeta::ArrayInfo arr_info =
-            GGUFMeta::GKV::get_kv(meta.get(), kid);
-
-
-        result = arr_info.length;
-        return true;
-    }
-
-    template
-    typename std::enable_if::value, bool>::type
-    get_arr_n(const enum llm_kv kid, T & result, const bool required = true) {
-        return get_arr_n(llm_kv(kid), result, required);
-    }
-
-    template
-    bool get_arr(const std::string & key, std::vector & result, const bool required = true) {
-        const int kid = gguf_find_key(meta.get(), key.c_str());
-
-        if (kid < 0 || gguf_get_kv_type(meta.get(), kid) != GGUF_TYPE_ARRAY) {
-            if (required) {
-                throw std::runtime_error(format("array key not found in model: %s", key.c_str()));
-            }
-            return false;
-        }
-
-        struct GGUFMeta::ArrayInfo arr_info =
-            GGUFMeta::GKV::get_kv(meta.get(), kid);
-
-        switch (arr_info.gt) {
-            case GGUF_TYPE_FLOAT32: GGML_ASSERT((std::is_same::value)); break;
-            case GGUF_TYPE_INT32:   GGML_ASSERT(
-                                            (std::is_same::value) ||
-                                            (std::is_same::value));  break;
-            default:
-                throw std::runtime_error(format("%s is not a float32, int32 array", key.c_str()));
-        }
-
-        result.resize(arr_info.length);
-        result.assign((const T*)arr_info.data, (const T *)arr_info.data + arr_info.length);
-
-        return true;
-    }
-
-    template
-    bool get_arr(const std::string & key, std::array & result, const bool required = true) {
-        const int kid = gguf_find_key(meta.get(), key.c_str());
-
-        if (kid < 0 || gguf_get_kv_type(meta.get(), kid) != GGUF_TYPE_ARRAY) {
-            if (required) {
-                throw std::runtime_error(format("array key not found in model: %s", key.c_str()));
-            }
-            return false;
-        }
-
-        struct GGUFMeta::ArrayInfo arr_info =
-            GGUFMeta::GKV::get_kv(meta.get(), kid);
-
-        switch (arr_info.gt) {
-            case GGUF_TYPE_FLOAT32: GGML_ASSERT((std::is_same::value)); break;
-            case GGUF_TYPE_INT32:   GGML_ASSERT(
-                                            (std::is_same::value) ||
-                                            (std::is_same::value));  break;
-            default:
-                throw std::runtime_error(format("%s is not a float32, int32 array", key.c_str()));
-        }
-
-        if (arr_info.length > N_MAX) {
-            throw std::runtime_error(format("array length %u for key %s exceeds max %u", (uint32_t) arr_info.length, key.c_str(), (uint32_t) N_MAX));
-        }
-
-        std::copy((const T*)arr_info.data, (const T *)arr_info.data + arr_info.length, result.begin());
-
-        return true;
-    }
-
-    template
-    bool get_arr(const enum llm_kv kid, T & result, const bool required = true) {
-        return get_arr(llm_kv(kid), result, required);
-    }
-
-    template
-    bool get_key(const std::string & key, T & result, const bool required = true) {
-        auto it = kv_overrides.find(key);
-
-        const struct llama_model_kv_override * override =
-            it != kv_overrides.end() ? &it->second : nullptr;
-
-        const bool found = GGUFMeta::GKV::set(meta.get(), key, result, override);
-
-        if (required && !found) {
-            throw std::runtime_error(format("key not found in model: %s", key.c_str()));
-        }
-
-        return found;
-    }
-
-    template
-    bool get_key(const enum llm_kv kid, T & result, const bool required = true) {
-        return get_key(llm_kv(kid), result, required);
-    }
-
-    // get array of n <= N_MAX elements, or a single element repeated n times
-    template
-    bool get_key_or_arr(const std::string & key, std::array & result, uint32_t n, const bool required = true) {
-        const int kid = gguf_find_key(meta.get(), key.c_str());
-
-        if (kid < 0) {
-            if (required) {
-                throw std::runtime_error(format("key not found in model: %s", key.c_str()));
-            }
-            return false;
-        }
-
-        if (n > N_MAX) {
-            throw std::runtime_error(format("n > N_MAX: %u > %u for key %s", (uint32_t) n, (uint32_t) N_MAX, key.c_str()));
-        }
-
-        if (gguf_get_kv_type(meta.get(), kid) == GGUF_TYPE_ARRAY) {
-            struct GGUFMeta::ArrayInfo arr_info =
-                GGUFMeta::GKV::get_kv(meta.get(), kid);
-
-            if (n != arr_info.length) {
-                throw std::runtime_error(format("key %s has wrong array length; expected %u, got %u", key.c_str(), n, (uint32_t) arr_info.length));
-            }
-
-            return get_arr(key, result, required);
-        } else {
-            T value;
-
-            bool ok = get_key(key, value, required);
-            if (!ok) {
-                return false;
-            }
-
-            for (uint32_t i = 0; i < n; i++) {
-                result[i] = value;
-            }
-
-            return true;
-        }
-    }
-
-    template
-    bool get_key_or_arr(const enum llm_kv kid, T & result, uint32_t n, const bool required = true) {
-        return get_key_or_arr(llm_kv(kid), result, n, required);
-    }
-
-    std::string get_arch_name() const {
-        return arch_name;
-    }
-
-    enum llm_arch get_arch() const {
-        return llm_kv.arch;
-    }
-
-    const llama_tensor_weight * get_weight(const char * name) const {
-        auto pos = weights_map.find(name);
-        if (pos != weights_map.end()) {
-            return &pos->second;
-        }
-
-        return nullptr;
-    }
-
-    const llama_tensor_weight & require_weight(const char * name) const {
-        const llama_tensor_weight * weight = get_weight(name);
-        if (!weight) {
-            throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name));
-        }
-        return *weight;
-    }
-
-    struct ggml_tensor * get_tensor_meta(const char * name) const {
-        const auto * weight = get_weight(name);
-        if (!weight) {
-            return nullptr;
-        }
-        return weight->tensor;
-    }
-
-    struct ggml_tensor * require_tensor_meta(const std::string & name) const {
-        struct ggml_tensor * tensor = get_tensor_meta(name.c_str());
-        if (!tensor) {
-            throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name.c_str()));
-        }
-        return tensor;
-    }
-
-    const struct ggml_tensor * check_tensor_dims(const std::string & name, const std::vector & ne, bool required) const {
-        const struct ggml_tensor * cur = get_tensor_meta(name.c_str());
-
-        if (cur == NULL) {
-            if (!required) {
-                return NULL;
-            }
-            throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name.c_str()));
-        }
-
-        {
-            bool is_ok = true;
-            for (size_t i = 0; i < GGML_MAX_DIMS; ++i) {
-                if ((i < ne.size() && ne[i] != cur->ne[i]) || (i >= ne.size() && cur->ne[i] != 1)) {
-                    is_ok = false;
-                    break;
-                }
-            }
-            if (!is_ok) {
-                throw std::runtime_error(
-                        format("%s: tensor '%s' has wrong shape; expected %s, got %s",
-                            __func__, name.c_str(),
-                            llama_format_tensor_shape(ne).c_str(),
-                            llama_format_tensor_shape(cur).c_str()));
-            }
-        }
-
-        return cur;
-    }
-
-    static const int TENSOR_NOT_REQUIRED = 1;
-    static const int TENSOR_DUPLICATED   = 2;
-
-    struct ggml_tensor * create_tensor(struct ggml_context * ctx, const std::string & name, const std::initializer_list & ne, int flags = 0) {
-        const struct ggml_tensor * cur = check_tensor_dims(name, ne, !(flags & TENSOR_NOT_REQUIRED));
-
-        if (cur == NULL) {
-            return NULL;
-        }
-
-        bool duplicated = flags & TENSOR_DUPLICATED;
-
-        struct ggml_tensor * tensor = ggml_dup_tensor(ctx, cur);
-        ggml_set_name(tensor, ggml_get_name(cur));
-
-        if (duplicated) {
-            size_data += ggml_nbytes(cur);
-        } else {
-            n_created++;
-        }
-
-        return tensor;
-
-    }
-
-    struct ggml_tensor * create_tensor_as_view(struct ggml_context * ctx, struct ggml_tensor * base, const std::string & name, const std::initializer_list & ne, size_t offset, bool required = true) {
-        const struct ggml_tensor * cur = check_tensor_dims(name, ne, required);
-
-        if (cur == NULL) {
-            return NULL;
-        }
-
-        if (cur->type != base->type) {
-            throw std::runtime_error(format("%s: tensor '%s' has wrong type; expected %s, got %s", __func__, name.c_str(), ggml_type_name(base->type), ggml_type_name(cur->type)));
-        }
-
-        std::array dims;
-        for (size_t i = 0; i < GGML_MAX_DIMS; ++i) {
-            dims[i] = i < ne.size() ? ne.begin()[i] : 1;
-        }
-
-        struct ggml_tensor * tensor = ggml_view_4d(ctx, base,
-                                        dims[0], dims[1], dims[2], dims[3],
-                                        cur->nb[1], cur->nb[2], cur->nb[3],
-                                        offset);
-
-        ggml_set_name(tensor, name.c_str());
-
-        n_created++;
-
-        return tensor;
-    }
-
-    void done_getting_tensors() const {
-        if (n_created != n_tensors) {
-            throw std::runtime_error(format("%s: wrong number of tensors; expected %d, got %d", __func__, n_tensors, n_created));
-        }
-    }
-
-    void init_mappings(bool prefetch = true, llama_mlocks * mlock_mmaps = nullptr) {
-        if (use_mmap) {
-            mappings.reserve(files.size());
-            mmaps_used.reserve(files.size());
-            for (const auto & file : files) {
-                std::unique_ptr mapping(new llama_mmap(file.get(), prefetch ? -1 : 0, ggml_is_numa()));
-                mmaps_used.emplace_back(mapping->size, 0);
-                if (mlock_mmaps) {
-                    std::unique_ptr mlock_mmap(new llama_mlock());
-                    mlock_mmap->init(mapping->addr);
-                    mlock_mmaps->emplace_back(std::move(mlock_mmap));
-                }
-                mappings.emplace_back(std::move(mapping));
-            }
-        }
-
-        // compute the total size of all tensors for progress reporting
-        for (const auto & it : weights_map) {
-            size_data += ggml_nbytes(it.second.tensor);
-        }
-    }
-
-    void get_mapping_range(size_t * first, size_t * last, void ** addr, int idx, ggml_context * ctx) const {
-        GGML_ASSERT(!mappings.empty());
-        const auto & mapping = mappings.at(idx);
-
-        *first = mapping->size;
-        *last  = 0;
-        *addr = mapping->addr;
-        for (ggml_tensor * tensor = ggml_get_first_tensor(ctx); tensor; tensor = ggml_get_next_tensor(ctx, tensor)) {
-            const auto * weight = get_weight(ggml_get_name(tensor));
-            if (!weight || weight->idx != idx) {
-                continue;
-            }
-            *first = std::min(*first, weight->offs);
-            *last  = std::max(*last,  weight->offs + ggml_nbytes(tensor));
-        }
-    }
-
-    // for backwards compatibility, does not support ggml-backend
-    void load_data_for(struct ggml_tensor * cur) const {
-        const auto & w = require_weight(ggml_get_name(cur));
-
-        if (use_mmap) {
-            const auto & mapping = mappings.at(w.idx);
-            if (cur->data == nullptr) {
-                cur->data = (uint8_t *)mapping->addr + w.offs;
-            } else {
-                memcpy(cur->data, (uint8_t *)mapping->addr + w.offs, ggml_nbytes(cur));
-            }
-        } else {
-            GGML_ASSERT(cur->data != nullptr);
-            GGML_ASSERT(w.idx < files.size());
-            const auto & file = files.at(w.idx);
-            file->seek(w.offs, SEEK_SET);
-            file->read_raw(cur->data, ggml_nbytes(cur));
-        }
-
-        if (check_tensors && !ggml_validate_row_data(cur->type, cur->data, ggml_nbytes(cur))) {
-            throw std::runtime_error(format("tensor '%s' has invalid data", ggml_get_name(cur)));
-        }
-    }
-
-    size_t size_done = 0;
-    size_t size_data = 0;
-    std::vector> mmaps_used;
-
-    // Returns false if cancelled by progress_callback
-    bool load_all_data(
-            struct ggml_context * ctx,
-            llama_buf_map & bufs,
-            llama_mlocks * lmlocks,
-            llama_progress_callback progress_callback,
-            void * progress_callback_user_data) {
-        GGML_ASSERT(size_data != 0 && "call init_mappings() first");
-
-        std::vector> read_buf;
-        std::vector>> validation_result;
-
-        // 4 staging buffers for async uploads, each sized 1MB seems to be a good default for single NVMe drives.
-        // NVMe raid configurations might require more / larger buffers.
-        constexpr size_t n_buffers = 4;
-        constexpr size_t buffer_size = 1 * 1024 * 1024; // 1MB
-
-        std::vector host_buffers;
-        std::vector events;
-        std::vector host_ptrs;
-        size_t buffer_idx = 0; // buffer to use for async loads
-        ggml_backend_t upload_backend = [&](const char * func) -> ggml_backend_t {
-            if (use_mmap || check_tensors) {
-                return nullptr;
-            }
-            // When not using mmaped io use async uploads from pinned memory to GPU memory.
-            // First determine if the backend supports the necessary features for async uploads.
-            auto * buf = bufs.count(0) ? bufs.at(0) : nullptr;
-            if (!buf) {
-                LLAMA_LOG_DEBUG("%s: no buffer found for async uploads\n", func);
-                return nullptr;
-            }
-
-            auto * buft = ggml_backend_buffer_get_type(buf);
-            auto * dev = ggml_backend_buft_get_device(buft);
-            if (!dev) {
-                LLAMA_LOG_DEBUG("%s: no device found for buffer type %s for async uploads\n", func,
-                    ggml_backend_buft_name(buft));
-                return nullptr;
-            }
-
-            if (buft != ggml_backend_dev_buffer_type(dev)) {
-                LLAMA_LOG_DEBUG("%s: buffer type %s is not the default buffer type for device %s for async uploads\n", func,
-                    ggml_backend_buft_name(buft), ggml_backend_dev_name(dev));
-                return nullptr;
-            }
-
-            ggml_backend_dev_props props;
-            ggml_backend_dev_get_props(dev, &props);
-            if (!props.caps.async || !props.caps.host_buffer || !props.caps.events) {
-                LLAMA_LOG_DEBUG("%s: device %s does not support async, host buffers or events\n", func,
-                    ggml_backend_dev_name(dev));
-                return nullptr;
-            }
-
-            auto * host_buft = ggml_backend_dev_host_buffer_type(dev);
-            if (!host_buft) {
-                LLAMA_LOG_DEBUG("%s: no host buffer type found for device %s\n", func,
-                    ggml_backend_dev_name(dev));
-                return nullptr;
-            }
-
-            // If the backend is supported, create pinned memory buffers and events for synchronisation.
-            for (size_t idx = 0; idx < n_buffers; ++idx) {
-                auto * buf = ggml_backend_buft_alloc_buffer(host_buft, buffer_size);
-                if (!buf) {
-                    LLAMA_LOG_DEBUG("%s: failed to allocate host buffer for async uploads for device %s\n", func,
-                        ggml_backend_dev_name(dev));
-                    return nullptr;
-                }
-
-                host_buffers.emplace_back(buf);
-                host_ptrs.emplace_back(ggml_backend_buffer_get_base(buf));
-
-                auto * event = ggml_backend_event_new(dev);
-                if (!event) {
-                    LLAMA_LOG_DEBUG("%s: failed to create event for async uploads for device %s\n", func,
-                        ggml_backend_dev_name(dev));
-                    return nullptr;
-                }
-
-                events.emplace_back(event);
-            }
-
-            ggml_backend_t backend = ggml_backend_dev_init(dev, nullptr);
-            if (!backend) {
-                LLAMA_LOG_DEBUG("%s: failed to initialize backend for device %s for async uploads\n", func,
-                    ggml_backend_dev_name(dev));
-                return nullptr;
-            }
-
-            return backend;
-        }(__func__);
-
-        if (upload_backend) {
-            LLAMA_LOG_DEBUG("%s: using async uploads for device %s, buffer type %s, backend %s\n", __func__,
-                ggml_backend_dev_name(ggml_backend_get_device(upload_backend)),
-                ggml_backend_buft_name(ggml_backend_buffer_get_type(bufs.at(0))),
-                ggml_backend_name(upload_backend));
-        }
-
-        for (struct ggml_tensor * cur = ggml_get_first_tensor(ctx); cur != NULL; cur = ggml_get_next_tensor(ctx, cur)) {
-            const auto * weight = get_weight(ggml_get_name(cur));
-            if (weight == nullptr) {
-                // this can happen with split experts models
-                continue;
-            }
-
-            if (progress_callback) {
-                if (!progress_callback((float) size_done / size_data, progress_callback_user_data)) {
-                    return false;
-                }
-            }
-
-            size_t n_size = ggml_nbytes(cur);
-
-            if (use_mmap) {
-                const auto & mapping = mappings.at(weight->idx);
-                ggml_backend_buffer_t buf_mmap = nullptr;
-                if (bufs.count(weight->idx)) {
-                    buf_mmap = bufs.at(weight->idx);
-                }
-                uint8_t * data = (uint8_t *) mapping->addr + weight->offs;
-
-                if (check_tensors) {
-                    validation_result.emplace_back(std::async(std::launch::async, [cur, data, n_size] {
-                        return std::make_pair(cur, ggml_validate_row_data(cur->type, data, n_size));
-                    }));
-                }
-
-                GGML_ASSERT(buf_mmap || cur->data); // either we have a buffer to allocate the tensor in, or it is already allocated
-                if (buf_mmap && cur->data == nullptr) {
-                    ggml_backend_tensor_alloc(buf_mmap, cur, data);
-                    if (lmlocks) {
-                        const auto & lmlock = lmlocks->at(weight->idx);
-                        lmlock->grow_to(weight->offs + n_size);
-                    }
-
-                    auto & mmap_used = mmaps_used[weight->idx];
-                    mmap_used.first  = std::min(mmap_used.first,  weight->offs);
-                    mmap_used.second = std::max(mmap_used.second, weight->offs + n_size);
-                } else {
-                    ggml_backend_tensor_set(cur, data, 0, n_size);
-                }
-            } else {
-                const auto & file = files.at(weight->idx);
-                if (ggml_backend_buffer_is_host(cur->buffer)) {
-                    file->seek(weight->offs, SEEK_SET);
-                    file->read_raw(cur->data, n_size);
-                    if (check_tensors) {
-                        validation_result.emplace_back(std::async(std::launch::async, [cur, n_size] {
-                            return std::make_pair(cur, ggml_validate_row_data(cur->type, cur->data, n_size));
-                        }));
-                    }
-                } else {
-                    // If upload_backend is valid load the tensor in chunks to pinned memory and upload the buffers asynchronously to the GPU.
-                    if (upload_backend) {
-                        file->seek(weight->offs, SEEK_SET);
-
-                        size_t bytes_read = 0;
-
-                        while (bytes_read < n_size) {
-                            size_t read_iteration = std::min(buffer_size, n_size - bytes_read);
-
-                            ggml_backend_event_synchronize(events[buffer_idx]);
-                            file->read_raw(host_ptrs[buffer_idx], read_iteration);
-                            ggml_backend_tensor_set_async(upload_backend, cur, host_ptrs[buffer_idx], bytes_read, read_iteration);
-                            ggml_backend_event_record(events[buffer_idx], upload_backend);
-
-                            bytes_read += read_iteration;
-                            ++buffer_idx;
-                            buffer_idx %= n_buffers;
-                        }
-                    } else {
-                        read_buf.resize(n_size);
-                        file->seek(weight->offs, SEEK_SET);
-                        file->read_raw(read_buf.data(), n_size);
-                        ggml_backend_tensor_set(cur, read_buf.data(), 0, n_size);
-                        if (check_tensors && !ggml_validate_row_data(cur->type, read_buf.data(), n_size)) {
-                            throw std::runtime_error(format("tensor '%s' has invalid data", ggml_get_name(cur)));
-                        }
-                    }
-                }
-            }
-
-            size_done += n_size;
-        }
-
-        // free temporary resources used for async uploads
-        for (auto * event : events) {
-            ggml_backend_event_synchronize(event);
-            ggml_backend_event_free(event);
-        }
-        for (auto * buf : host_buffers) {
-            ggml_backend_buffer_free(buf);
-        }
-        ggml_backend_free(upload_backend);
-
-        // check validation results
-        bool validation_failed = false;
-        for (auto & future : validation_result) {
-            auto result = future.get();
-            if (!result.second) {
-                LLAMA_LOG_ERROR("%s: tensor '%s' has invalid data\n", __func__, ggml_get_name(result.first));
-                validation_failed = true;
-            }
-        }
-        if (validation_failed) {
-            throw std::runtime_error("found tensors with invalid data");
-        }
-
-        // check if this is the last call and do final cleanup
-        if (size_done >= size_data) {
-            // unmap offloaded tensors and metadata
-            if (use_mmap) {
-                for (uint32_t idx = 0; idx < mappings.size(); idx++) {
-                    const auto & mmap_used = mmaps_used.at(idx);
-                    auto & mapping = mappings.at(idx);
-                    mapping->unmap_fragment(0, mmap_used.first);
-                    if (mmap_used.second != 0) {
-                        mapping->unmap_fragment(mmap_used.second, mapping->size);
-                    }
-                }
-            }
-            if (progress_callback) {
-                // Even though the model is done loading, we still honor
-                // cancellation since we need to free allocations.
-                return progress_callback(1.0f, progress_callback_user_data);
-            }
-        }
-
-        return true;
-    }
-};
-
-// temporary allocate memory for the input batch if needed
-static const llama_seq_id batch_default_seq_id = 0;
-struct llama_batch_allocr {
-    std::array seq_id_0 = {batch_default_seq_id};
-    std::vector      pos;
-    std::vector        n_seq_id;
-    std::vector seq_id;
-    std::vector         logits;
-    struct llama_batch          batch;
-    // optionally fulfill the batch returned by llama_batch_get_one
-    llama_batch_allocr(llama_context & ctx, struct llama_batch in_batch) {
-        batch = in_batch;
-        GGML_ASSERT(batch.n_tokens > 0);
-        if (!batch.pos) {
-            // determine the last position in KV cache
-            llama_pos last_pos = -1;
-            for (const auto & cell : ctx.kv_self.cells) {
-                if (cell.has_seq_id(batch_default_seq_id)) {
-                    last_pos = std::max(last_pos, cell.pos);
-                }
-            }
-            last_pos++; // next position
-            pos.resize(batch.n_tokens);
-            for (int32_t i = 0; i < batch.n_tokens; i++) {
-                pos[i] = i+last_pos;
-            }
-            batch.pos = pos.data();
-        }
-        if (!batch.n_seq_id) {
-            n_seq_id.resize(batch.n_tokens);
-            for (int32_t i = 0; i < batch.n_tokens; i++) {
-                n_seq_id[i] = seq_id_0.size();
-            }
-            batch.n_seq_id = n_seq_id.data();
-        }
-        if (!batch.seq_id) {
-            seq_id.resize(batch.n_tokens + 1);
-            seq_id[batch.n_tokens] = NULL;
-            for (int32_t i = 0; i < batch.n_tokens; i++) {
-                seq_id[i] = seq_id_0.data();
-            }
-            batch.seq_id = seq_id.data();
-        }
-        if (!batch.logits) {
-            logits.resize(batch.n_tokens);
-            logits[logits.size() - 1] = true;
-            batch.logits = logits.data();
-        }
-    }
-};
-
-template<>
-bool llama_model_loader::get_key(const enum llm_kv kid, enum llama_pooling_type & result, const bool required) {
-    uint32_t tmp;
-    const bool found = get_key(kid, tmp, required);
-    if (found) {
-        result = (enum llama_pooling_type) tmp;
-    } else {
-        result = LLAMA_POOLING_TYPE_UNSPECIFIED;
-    }
-    return found;
-}
-
-
-//
-// load LLaMA models
-//
-
-static const char * llama_model_arch_name(llm_arch arch) {
-    auto it = LLM_ARCH_NAMES.find(arch);
-    if (it == LLM_ARCH_NAMES.end()) {
-        return "unknown";
-    }
-    return it->second;
-}
-
-static std::string llama_model_ftype_name(llama_ftype ftype) {
-    if (ftype & LLAMA_FTYPE_GUESSED) {
-        return llama_model_ftype_name((enum llama_ftype) (ftype & ~LLAMA_FTYPE_GUESSED)) + " (guessed)";
-    }
-
-    switch (ftype) {
-        case LLAMA_FTYPE_ALL_F32:         return "all F32";
-        case LLAMA_FTYPE_MOSTLY_F16:      return "F16";
-        case LLAMA_FTYPE_MOSTLY_BF16:     return "BF16";
-        case LLAMA_FTYPE_MOSTLY_Q4_0:     return "Q4_0";
-        case LLAMA_FTYPE_MOSTLY_Q4_1:     return "Q4_1";
-        case LLAMA_FTYPE_MOSTLY_Q5_0:     return "Q5_0";
-        case LLAMA_FTYPE_MOSTLY_Q5_1:     return "Q5_1";
-        case LLAMA_FTYPE_MOSTLY_Q8_0:     return "Q8_0";
-        case LLAMA_FTYPE_MOSTLY_Q2_K:     return "Q2_K - Medium";
-        case LLAMA_FTYPE_MOSTLY_Q2_K_S:   return "Q2_K - Small";
-        case LLAMA_FTYPE_MOSTLY_Q3_K_S:   return "Q3_K - Small";
-        case LLAMA_FTYPE_MOSTLY_Q3_K_M:   return "Q3_K - Medium";
-        case LLAMA_FTYPE_MOSTLY_Q3_K_L:   return "Q3_K - Large";
-        case LLAMA_FTYPE_MOSTLY_Q4_K_S:   return "Q4_K - Small";
-        case LLAMA_FTYPE_MOSTLY_Q4_K_M:   return "Q4_K - Medium";
-        case LLAMA_FTYPE_MOSTLY_Q5_K_S:   return "Q5_K - Small";
-        case LLAMA_FTYPE_MOSTLY_Q5_K_M:   return "Q5_K - Medium";
-        case LLAMA_FTYPE_MOSTLY_Q6_K:     return "Q6_K";
-        case LLAMA_FTYPE_MOSTLY_TQ1_0:    return "TQ1_0 - 1.69 bpw ternary";
-        case LLAMA_FTYPE_MOSTLY_TQ2_0:    return "TQ2_0 - 2.06 bpw ternary";
-        case LLAMA_FTYPE_MOSTLY_IQ2_XXS:  return "IQ2_XXS - 2.0625 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ2_XS:   return "IQ2_XS - 2.3125 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ2_S:    return "IQ2_S - 2.5 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ2_M:    return "IQ2_M - 2.7 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ3_XS:   return "IQ3_XS - 3.3 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ3_XXS:  return "IQ3_XXS - 3.0625 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ1_S:    return "IQ1_S - 1.5625 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ1_M:    return "IQ1_M - 1.75 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ4_NL:   return "IQ4_NL - 4.5 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ4_XS:   return "IQ4_XS - 4.25 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ3_S:    return "IQ3_S - 3.4375 bpw";
-        case LLAMA_FTYPE_MOSTLY_IQ3_M:    return "IQ3_S mix - 3.66 bpw";
-        case LLAMA_FTYPE_MOSTLY_Q4_0_4_4: return "Q4_0_4_4";
-        case LLAMA_FTYPE_MOSTLY_Q4_0_4_8: return "Q4_0_4_8";
-        case LLAMA_FTYPE_MOSTLY_Q4_0_8_8: return "Q4_0_8_8";
-
-        default: return "unknown, may not work";
-    }
-}
-
-static const char * llama_model_type_name(e_model type) {
-    switch (type) {
-        case MODEL_14M:           return "14M";
-        case MODEL_17M:           return "17M";
-        case MODEL_22M:           return "22M";
-        case MODEL_33M:           return "33M";
-        case MODEL_60M:           return "60M";
-        case MODEL_70M:           return "70M";
-        case MODEL_80M:           return "80M";
-        case MODEL_109M:          return "109M";
-        case MODEL_137M:          return "137M";
-        case MODEL_160M:          return "160M";
-        case MODEL_220M:          return "220M";
-        case MODEL_250M:          return "250M";
-        case MODEL_270M:          return "270M";
-        case MODEL_335M:          return "335M";
-        case MODEL_410M:          return "410M";
-        case MODEL_450M:          return "450M";
-        case MODEL_770M:          return "770M";
-        case MODEL_780M:          return "780M";
-        case MODEL_0_5B:          return "0.5B";
-        case MODEL_1B:            return "1B";
-        case MODEL_1_3B:          return "1.3B";
-        case MODEL_1_4B:          return "1.4B";
-        case MODEL_1_6B:          return "1.6B";
-        case MODEL_2B:            return "2B";
-        case MODEL_2_8B:          return "2.8B";
-        case MODEL_3B:            return "3B";
-        case MODEL_4B:            return "4B";
-        case MODEL_6B:            return "6B";
-        case MODEL_6_9B:          return "6.9B";
-        case MODEL_7B:            return "7B";
-        case MODEL_8B:            return "8B";
-        case MODEL_9B:            return "9B";
-        case MODEL_11B:           return "11B";
-        case MODEL_12B:           return "12B";
-        case MODEL_13B:           return "13B";
-        case MODEL_14B:           return "14B";
-        case MODEL_15B:           return "15B";
-        case MODEL_16B:           return "16B";
-        case MODEL_20B:           return "20B";
-        case MODEL_30B:           return "30B";
-        case MODEL_34B:           return "34B";
-        case MODEL_35B:           return "35B";
-        case MODEL_40B:           return "40B";
-        case MODEL_65B:           return "65B";
-        case MODEL_70B:           return "70B";
-        case MODEL_236B:          return "236B";
-        case MODEL_314B:          return "314B";
-        case MODEL_SMALL:         return "0.1B";
-        case MODEL_MEDIUM:        return "0.4B";
-        case MODEL_LARGE:         return "0.8B";
-        case MODEL_XL:            return "1.5B";
-        case MODEL_A1_7B:         return "A1.7B";
-        case MODEL_A2_7B:         return "A2.7B";
-        case MODEL_8x7B:          return "8x7B";
-        case MODEL_8x22B:         return "8x22B";
-        case MODEL_16x12B:        return "16x12B";
-        case MODEL_10B_128x3_66B: return "10B+128x3.66B";
-        case MODEL_57B_A14B:      return "57B.A14B";
-        case MODEL_27B:           return "27B";
-        default:                  return "?B";
-    }
-}
-
-static const char * llama_model_vocab_type_name(enum llama_vocab_type type){
-    switch (type) {
-        case LLAMA_VOCAB_TYPE_NONE: return "no vocab";
-        case LLAMA_VOCAB_TYPE_SPM:  return "SPM";
-        case LLAMA_VOCAB_TYPE_BPE:  return "BPE";
-        case LLAMA_VOCAB_TYPE_WPM:  return "WPM";
-        case LLAMA_VOCAB_TYPE_UGM:  return "UGM";
-        case LLAMA_VOCAB_TYPE_RWKV: return "RWKV";
-        default:                    return "unknown";
-    }
-}
-
-static void llm_load_arch(llama_model_loader & ml, llama_model & model) {
-    model.arch = ml.get_arch();
-    if (model.arch == LLM_ARCH_UNKNOWN) {
-        throw std::runtime_error("unknown model architecture: '" + ml.get_arch_name() + "'");
-    }
-}
-
-static void llm_load_hparams(
-        llama_model_loader & ml,
-        llama_model & model) {
-    auto & hparams = model.hparams;
-    const gguf_context * ctx = ml.meta.get();
-
-    // get metadata as string
-    for (int i = 0; i < gguf_get_n_kv(ctx); i++) {
-        enum gguf_type type = gguf_get_kv_type(ctx, i);
-        if (type == GGUF_TYPE_ARRAY) {
-            continue;
-        }
-        const char * name = gguf_get_key(ctx, i);
-        const std::string value = gguf_kv_to_str(ctx, i);
-        model.gguf_kv.emplace(name, value);
-    }
-
-    // get general kv
-    ml.get_key(LLM_KV_GENERAL_NAME, model.name, false);
-
-    // get hparams kv
-    ml.get_key(LLM_KV_VOCAB_SIZE, hparams.n_vocab, false) || ml.get_arr_n(LLM_KV_TOKENIZER_LIST, hparams.n_vocab);
-
-    // everything past this point is not vocab-related
-    if (hparams.vocab_only) {
-        return;
-    }
-
-    ml.get_key(LLM_KV_CONTEXT_LENGTH,    hparams.n_ctx_train);
-    ml.get_key(LLM_KV_EMBEDDING_LENGTH,  hparams.n_embd);
-    ml.get_key(LLM_KV_BLOCK_COUNT,       hparams.n_layer);
-    ml.get_key(LLM_KV_EXPERT_COUNT,      hparams.n_expert,      false);
-    ml.get_key(LLM_KV_EXPERT_USED_COUNT, hparams.n_expert_used, false);
-
-    GGML_ASSERT(hparams.n_expert <= LLAMA_MAX_EXPERTS);
-    GGML_ASSERT(hparams.n_expert_used <= hparams.n_expert);
-    if (hparams.n_expert > 0) {
-        GGML_ASSERT(hparams.n_expert_used > 0);
-    } else {
-        GGML_ASSERT(hparams.n_expert_used == 0);
-    }
-
-    // zero-out the per-layer hparams
-    std::fill(hparams.n_head_arr.begin(),    hparams.n_head_arr.end(),    0);
-    std::fill(hparams.n_head_kv_arr.begin(), hparams.n_head_kv_arr.end(), 0);
-    std::fill(hparams.n_ff_arr.begin(),      hparams.n_ff_arr.end(),      0);
-
-    ml.get_key_or_arr(LLM_KV_FEED_FORWARD_LENGTH,  hparams.n_ff_arr,   hparams.n_layer);
-    ml.get_key_or_arr(LLM_KV_ATTENTION_HEAD_COUNT, hparams.n_head_arr, hparams.n_layer);
-
-    // n_head_kv is optional, default to n_head
-    hparams.n_head_kv_arr = hparams.n_head_arr;
-
-    ml.get_key_or_arr(LLM_KV_ATTENTION_HEAD_COUNT_KV, hparams.n_head_kv_arr, hparams.n_layer, false);
-
-    bool rope_finetuned = false;
-    ml.get_key(LLM_KV_ROPE_SCALING_FINETUNED, rope_finetuned, false);
-    hparams.rope_finetuned = rope_finetuned;
-
-    hparams.n_ctx_orig_yarn = hparams.n_ctx_train;
-    ml.get_key(LLM_KV_ROPE_SCALING_ORIG_CTX_LEN, hparams.n_ctx_orig_yarn, false);
-
-    // rope_freq_base (optional)
-    hparams.rope_freq_base_train = 10000.0f;
-    ml.get_key(LLM_KV_ROPE_FREQ_BASE, hparams.rope_freq_base_train, false);
-
-    std::string rope_scaling("linear");
-    ml.get_key(LLM_KV_ROPE_SCALING_TYPE, rope_scaling, false);
-    hparams.rope_scaling_type_train = llama_rope_scaling_type_from_string(rope_scaling);
-    GGML_ASSERT(hparams.rope_scaling_type_train != LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED);
-
-    // rope_freq_scale (inverse of the kv) is optional
-    float ropescale = 0.0f;
-    if (!ml.get_key(LLM_KV_ROPE_SCALING_FACTOR, ropescale, false)) {
-        // try the old key name
-        ml.get_key(LLM_KV_ROPE_SCALE_LINEAR, ropescale, false);
-    }
-    hparams.rope_freq_scale_train = ropescale == 0.0f ? 1.0f : 1.0f/ropescale;
-
-    ml.get_key(LLM_KV_ROPE_SCALING_ATTN_FACTOR, hparams.rope_attn_factor, false);
-
-    // non-transformer models do not have attention heads
-    if (hparams.n_head() > 0) {
-        // gpt-neox n_rot = rotary_pct * (n_embd / n_head)
-        // gpt-j n_rot = rotary_dim
-
-        hparams.n_embd_head_k = hparams.n_embd / hparams.n_head();
-        ml.get_key(LLM_KV_ATTENTION_KEY_LENGTH, hparams.n_embd_head_k, false);
-
-        hparams.n_embd_head_v = hparams.n_embd / hparams.n_head();
-        ml.get_key(LLM_KV_ATTENTION_VALUE_LENGTH, hparams.n_embd_head_v, false);
-
-        // sanity check for n_rot (optional)
-        hparams.n_rot = hparams.n_embd_head_k;
-
-        ml.get_key(LLM_KV_ROPE_DIMENSION_COUNT, hparams.n_rot, false);
-
-        if (model.arch == LLM_ARCH_LLAMA || model.arch == LLM_ARCH_FALCON) {
-            if (hparams.n_rot != hparams.n_embd_head_k) {
-                throw std::runtime_error(format("invalid n_rot: %u, expected %u", hparams.n_rot, hparams.n_embd_head_k));
-            }
-        }
-    } else {
-        hparams.n_rot = 0;
-        hparams.n_embd_head_k = 0;
-        hparams.n_embd_head_v = 0;
-    }
-
-    // arch-specific KVs
-    switch (model.arch) {
-        case LLM_ARCH_LLAMA:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                if (hparams.n_expert == 8) {
-                    switch (hparams.n_layer) {
-                        case 32: model.type = e_model::MODEL_8x7B; break;
-                        case 56: model.type = e_model::MODEL_8x22B; break;
-                        default: model.type = e_model::MODEL_UNKNOWN;
-                    }
-                } else {
-                    switch (hparams.n_layer) {
-                        case 16: model.type = e_model::MODEL_1B; break; // Llama 3.2 1B
-                        case 22: model.type = e_model::MODEL_1B; break;
-                        case 26: model.type = e_model::MODEL_3B; break;
-                        case 28: model.type = e_model::MODEL_3B; break; // Llama 3.2 3B
-                        // granite uses a vocab with len 49152
-                        case 32: model.type = hparams.n_vocab == 49152 ? e_model::MODEL_3B : (hparams.n_vocab < 40000 ? e_model::MODEL_7B : e_model::MODEL_8B); break;
-                        case 36: model.type = e_model::MODEL_8B; break; // granite
-                        case 40: model.type = e_model::MODEL_13B; break;
-                        case 48: model.type = e_model::MODEL_34B; break;
-                        case 60: model.type = e_model::MODEL_30B; break;
-                        case 80: model.type = hparams.n_head() == hparams.n_head_kv() ? e_model::MODEL_65B : e_model::MODEL_70B; break;
-                        default: model.type = e_model::MODEL_UNKNOWN;
-                    }
-                }
-            } break;
-        case LLM_ARCH_MINICPM:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 40: model.type = e_model::MODEL_2B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_MINICPM3:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q);
-                ml.get_key(LLM_KV_ATTENTION_KV_LORA_RANK, hparams.n_lora_kv);
-
-                switch (hparams.n_layer) {
-                    case 62: model.type = e_model::MODEL_4B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_GROK:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 64: model.type = e_model::MODEL_314B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_FALCON:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 60: model.type = e_model::MODEL_40B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_BAICHUAN:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = e_model::MODEL_13B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-
-                if (model.type == e_model::MODEL_13B) {
-                    // TODO: become GGUF KV parameter
-                    hparams.f_max_alibi_bias = 8.0f;
-                }
-            } break;
-        case LLM_ARCH_STARCODER:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1B; break;
-                    case 36: model.type = e_model::MODEL_3B; break;
-                    case 42: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = e_model::MODEL_15B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_REFACT:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_1B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-
-                // TODO: become GGUF KV parameter
-                hparams.f_max_alibi_bias = 8.0f;
-            } break;
-        case LLM_ARCH_BERT:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,    hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_CAUSAL,           hparams.causal_attn);
-                ml.get_key(LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, hparams.n_vocab_type);
-                ml.get_key(LLM_KV_POOLING_TYPE,               hparams.pooling_type, false);
-
-                switch (hparams.n_layer) {
-                    case 3:
-                        model.type = e_model::MODEL_17M; break; // bge-micro
-                    case 6:
-                        model.type = e_model::MODEL_22M; break; // MiniLM-L6
-                    case 12:
-                        switch (hparams.n_embd) {
-                            case 384: model.type = e_model::MODEL_33M; break; // MiniLM-L12, bge-small
-                            case 768: model.type = e_model::MODEL_109M; break; // bge-base
-                        } break;
-                    case 24:
-                        model.type = e_model::MODEL_335M; break; // bge-large
-                }
-            } break;
-        case LLM_ARCH_JINA_BERT_V2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,    hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_CAUSAL,           hparams.causal_attn);
-                ml.get_key(LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, hparams.n_vocab_type);
-                ml.get_key(LLM_KV_POOLING_TYPE,               hparams.pooling_type, false);
-                hparams.f_max_alibi_bias = 8.0f;
-
-                switch (hparams.n_layer) {
-                    case 4:  model.type = e_model::MODEL_33M;  break; // jina-embeddings-small
-                    case 12: model.type = e_model::MODEL_137M; break; // jina-embeddings-base
-                }
-            } break;
-        case LLM_ARCH_NOMIC_BERT:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,    hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_CAUSAL,           hparams.causal_attn);
-                ml.get_key(LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, hparams.n_vocab_type);
-                ml.get_key(LLM_KV_POOLING_TYPE,               hparams.pooling_type);
-
-                if (hparams.n_layer == 12 && hparams.n_embd == 768) {
-                    model.type = e_model::MODEL_137M;
-                }
-            } break;
-        case LLM_ARCH_BLOOM:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1B; break;
-                    case 30:
-                        switch (hparams.n_embd) {
-                            case 2560: model.type = e_model::MODEL_3B; break;
-                            case 4096: model.type = e_model::MODEL_7B; break;
-                        } break;
-                }
-
-                // TODO: become GGUF KV parameter
-                hparams.f_max_alibi_bias = 8.0f;
-            } break;
-        case LLM_ARCH_MPT:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,  hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV,      hparams.f_clamp_kqv, false);
-                ml.get_key(LLM_KV_ATTENTION_MAX_ALIBI_BIAS, hparams.f_max_alibi_bias);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 48: model.type = e_model::MODEL_30B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_STABLELM:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1B; break;
-                    case 32: model.type = e_model::MODEL_3B; break;
-                    case 40: model.type = e_model::MODEL_12B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        case LLM_ARCH_QWEN:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = e_model::MODEL_13B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_QWEN2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 24: model.type = hparams.n_embd == 1024 ? e_model::MODEL_0_5B : e_model::MODEL_1B; break;
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = hparams.n_head() == 20 ? e_model::MODEL_4B : e_model::MODEL_13B; break;
-                    case 80: model.type = e_model::MODEL_70B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_QWEN2MOE:
-            {
-                ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp, false);
-                ml.get_key(LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, hparams.n_ff_shexp, false);
-
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_A2_7B; break;
-                    case 28: model.type = e_model::MODEL_57B_A14B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_PHI2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1B; break;
-                    case 32: model.type = e_model::MODEL_3B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_PHI3:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1B; break;
-                    case 32: model.type = e_model::MODEL_3B; break;
-                    case 40: model.type = e_model::MODEL_14B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-
-                // for backward compatibility ; see: https://github.com/ggerganov/llama.cpp/pull/8931
-                if ((hparams.n_layer == 32 || hparams.n_layer == 40) && hparams.n_ctx_train == 4096) {
-                    // default value for Phi-3-mini-4k-instruct and Phi-3-medium-4k-instruct
-                    hparams.n_swa = 2047;
-                } else if (hparams.n_layer == 32 && hparams.n_head_kv(0) == 32 && hparams.n_ctx_train == 131072) {
-                    // default value for Phi-3-mini-128k-instruct
-                    hparams.n_swa = 262144;
-                } else if (hparams.n_layer == 40 && hparams.n_ctx_train == 131072) {
-                    // default value for Phi-3-medium-128k-instruct
-                    hparams.n_swa = 131072;
-                }
-                bool found_swa = ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false);
-                if (!found_swa && hparams.n_swa == 0) {
-                    throw std::runtime_error("invalid value for sliding_window");
-                }
-            } break;
-        case LLM_ARCH_PLAMO:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 40: model.type = e_model::MODEL_13B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        case LLM_ARCH_GPT2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 12: model.type = e_model::MODEL_SMALL; break;
-                    case 24: model.type = e_model::MODEL_MEDIUM; break;
-                    case 36: model.type = e_model::MODEL_LARGE; break;
-                    case 48: model.type = e_model::MODEL_XL; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_CODESHELL:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 42: model.type = e_model::MODEL_7B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_ORION:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-
-                switch (hparams.n_layer) {
-                    case 40: model.type = e_model::MODEL_14B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_INTERNLM2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 48: model.type = e_model::MODEL_20B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_GEMMA:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 18: model.type = e_model::MODEL_2B; break;
-                    case 28: model.type = e_model::MODEL_7B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        case LLM_ARCH_GEMMA2:
-            {
-                hparams.n_swa = 4096; // default value of gemma 2
-                ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false);
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_ATTN_LOGIT_SOFTCAPPING, hparams.f_attn_logit_softcapping, false);
-                ml.get_key(LLM_KV_FINAL_LOGIT_SOFTCAPPING, hparams.f_final_logit_softcapping, false);
-                hparams.attn_soft_cap = true;
-
-                switch (hparams.n_layer) {
-                    case 26: model.type = e_model::MODEL_2B; break;
-                    case 42: model.type = e_model::MODEL_9B; break;
-                    case 46: model.type = e_model::MODEL_27B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        case LLM_ARCH_STARCODER2:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 30: model.type = e_model::MODEL_3B; break;
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = e_model::MODEL_15B; break;
-                    case 52: model.type = e_model::MODEL_20B; break; // granite
-                    case 88: model.type = e_model::MODEL_34B; break; // granite
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_MAMBA:
-            {
-                ml.get_key(LLM_KV_SSM_CONV_KERNEL,    hparams.ssm_d_conv);
-                ml.get_key(LLM_KV_SSM_INNER_SIZE,     hparams.ssm_d_inner);
-                ml.get_key(LLM_KV_SSM_STATE_SIZE,     hparams.ssm_d_state);
-                ml.get_key(LLM_KV_SSM_TIME_STEP_RANK, hparams.ssm_dt_rank);
-                ml.get_key(LLM_KV_SSM_DT_B_C_RMS, hparams.ssm_dt_b_c_rms, false);
-
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 24:
-                        switch (hparams.n_embd) {
-                            case 768: model.type = e_model::MODEL_SMALL; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 48:
-                        switch (hparams.n_embd) {
-                            case 1024: model.type = e_model::MODEL_MEDIUM; break;
-                            case 1536: model.type = e_model::MODEL_LARGE; break;
-                            case 2048: model.type = e_model::MODEL_XL; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 64:
-                        switch (hparams.n_embd) {
-                            case 2560: model.type = e_model::MODEL_3B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_XVERSE:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 40: model.type = e_model::MODEL_13B; break;
-                    case 80: model.type = e_model::MODEL_65B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_COMMAND_R:
-            {
-                ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale);
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 40: model.type = e_model::MODEL_35B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_DBRX:
-        {
-            ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS,  hparams.f_norm_eps);
-            ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV,      hparams.f_clamp_kqv);
-
-            switch (hparams.n_layer) {
-                case 40: model.type = e_model::MODEL_16x12B; break;
-                default: model.type = e_model::MODEL_UNKNOWN;
-            }
-        } break;
-        case LLM_ARCH_OLMO:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_CLAMP_KQV,     hparams.f_clamp_kqv, false);
-
-                switch (hparams.n_layer) {
-                    case 22: model.type = e_model::MODEL_1B; break;
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 80: model.type = e_model::MODEL_70B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_OLMOE:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 16: model.type = e_model::MODEL_A1_7B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_OPENELM:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                case 16: model.type = e_model::MODEL_270M; break;
-                case 20: model.type = e_model::MODEL_450M; break;
-                case 28: model.type = e_model::MODEL_1B; break;
-                case 36: model.type = e_model::MODEL_3B; break;
-                default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_GPTNEOX:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                ml.get_key(LLM_KV_USE_PARALLEL_RESIDUAL, hparams.use_par_res);
-                switch (hparams.n_layer) {
-                    case 6:
-                        switch (hparams.n_ff()) {
-                            case 512: model.type = e_model::MODEL_14M; break;
-                            case 2048: model.type = e_model::MODEL_70M; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 12:
-                        switch (hparams.n_ff()) {
-                            case 3072: model.type = e_model::MODEL_160M; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 16:
-                        switch (hparams.n_ff()) {
-                            case 8192: model.type = e_model::MODEL_1B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 24:
-                        switch (hparams.n_ff()) {
-                            case 4096: model.type = e_model::MODEL_410M; break;
-                            case 8192: model.type = e_model::MODEL_1_4B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 32:
-                        switch (hparams.n_ff()) {
-                            case 10240: model.type = e_model::MODEL_2_8B; break;
-                            case 16384: model.type = e_model::MODEL_6_9B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 36:
-                        switch (hparams.n_ff()) {
-                            case 20480: model.type = e_model::MODEL_12B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 44:
-                        switch (hparams.n_ff()) {
-                            case 24576: model.type = e_model::MODEL_20B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_ARCTIC:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                if (hparams.n_expert == 128) {
-                    switch (hparams.n_layer) {
-                        case 35: model.type = e_model::MODEL_10B_128x3_66B; break;
-                        default: model.type = e_model::MODEL_UNKNOWN;
-                    }
-                } else {
-                    model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_DEEPSEEK2:
-            {
-                bool is_lite = (hparams.n_layer == 27);
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT, hparams.n_layer_dense_lead);
-                if (!is_lite) {
-                    ml.get_key(LLM_KV_ATTENTION_Q_LORA_RANK, hparams.n_lora_q);
-                }
-                ml.get_key(LLM_KV_ATTENTION_KV_LORA_RANK, hparams.n_lora_kv);
-                ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp);
-                ml.get_key(LLM_KV_EXPERT_SHARED_COUNT, hparams.n_expert_shared);
-                ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE, hparams.expert_weights_scale);
-                ml.get_key(LLM_KV_ROPE_SCALING_YARN_LOG_MUL, hparams.rope_yarn_log_mul);
-
-                switch (hparams.n_layer) {
-                    case 27: model.type = e_model::MODEL_16B; break;
-                    case 60: model.type = e_model::MODEL_236B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_CHATGLM:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                switch (hparams.n_layer) {
-                    case 28: model.type = e_model::MODEL_6B; break;
-                    case 40: model.type = e_model::MODEL_9B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_BITNET:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 26: model.type = e_model::MODEL_3B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_T5:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, hparams.n_rel_attn_bkts);
-
-                uint32_t dec_start_token_id;
-                if (ml.get_key(LLM_KV_DECODER_START_TOKEN_ID, dec_start_token_id, false)) {
-                    hparams.dec_start_token_id = dec_start_token_id;
-                }
-
-                switch (hparams.n_layer) {
-                    case 6:  model.type = e_model::MODEL_60M;  break; // t5-small
-                    case 8:  model.type = e_model::MODEL_80M;  break; // flan-t5-small
-                    case 12:
-                        switch (hparams.n_ff()) {
-                            case 3072: model.type = e_model::MODEL_220M; break; // t5-base
-                            case 2048: model.type = e_model::MODEL_250M; break; // flan-t5-base
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 24:
-                        switch (hparams.n_ff()) {
-                            case 4096:  model.type = e_model::MODEL_770M; break; // t5-large
-                            case 2816:  model.type = e_model::MODEL_780M; break; // flan-t5-large
-                            case 16384: model.type = e_model::MODEL_3B;   break; // t5-3b
-                            case 5120:  model.type = e_model::MODEL_3B;   break; // flan-t5-xl
-                            case 65536: model.type = e_model::MODEL_11B;  break; // t5-11b
-                            case 10240: model.type = e_model::MODEL_11B;  break; // flan-t5-xxl
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        case LLM_ARCH_T5ENCODER:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_ATTENTION_RELATIVE_BUCKETS_COUNT, hparams.n_rel_attn_bkts);
-                model.type = e_model::MODEL_UNKNOWN;
-            } break;
-        case LLM_ARCH_JAIS:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                ml.get_key(LLM_KV_ATTENTION_MAX_ALIBI_BIAS, hparams.f_max_alibi_bias);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1_3B; break;
-                    case 40: model.type = e_model::MODEL_13B; break;
-                    /* TODO: add variants */
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_NEMOTRON:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_4B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_EXAONE:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_8B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_RWKV6:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
-                ml.get_key(LLM_KV_WKV_HEAD_SIZE, hparams.wkv_head_size);
-                ml.get_key(LLM_KV_TIME_MIX_EXTRA_DIM, hparams.time_mix_extra_dim);
-                ml.get_key(LLM_KV_TIME_DECAY_EXTRA_DIM, hparams.time_decay_extra_dim);
-                ml.get_key(LLM_KV_RESCALE_EVERY_N_LAYERS, hparams.rescale_every_n_layers, false);
-
-                switch (hparams.n_layer) {
-                    case 24: model.type = e_model::MODEL_1_6B; break;
-                    case 32:
-                        switch (hparams.n_embd) {
-                            case 2560: model.type = e_model::MODEL_3B; break;
-                            case 4096: model.type = e_model::MODEL_7B; break;
-                            default: model.type = e_model::MODEL_UNKNOWN;
-                        } break;
-                    case 61: model.type = e_model::MODEL_14B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_GRANITE:
-        case LLM_ARCH_GRANITE_MOE:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                ml.get_key(LLM_KV_LOGIT_SCALE, hparams.f_logit_scale);
-                ml.get_key(LLM_KV_RESIDUAL_SCALE, hparams.f_residual_scale);
-                ml.get_key(LLM_KV_EMBEDDING_SCALE, hparams.f_embedding_scale);
-                ml.get_key(LLM_KV_ATTENTION_SCALE, hparams.f_attention_scale);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_3B; break;
-                    case 40: model.type = e_model::MODEL_3B; break;
-                    // Add additional layer/vocab/etc checks here for other model sizes
-                    default: model.type = e_model::MODEL_UNKNOWN;
-                }
-            } break;
-        case LLM_ARCH_CHAMELEON:
-            {
-                ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
-                hparams.f_norm_eps = 1e-5;  // eps for qk-norm, torch default
-                ml.get_key(LLM_KV_SWIN_NORM, hparams.swin_norm);
-
-                switch (hparams.n_layer) {
-                    case 32: model.type = e_model::MODEL_7B; break;
-                    case 48: model.type = e_model::MODEL_34B; break;
-                    default: model.type = e_model::MODEL_UNKNOWN;
-               }
-            } break;
-        default: (void)0;
-    }
-
-    model.ftype = ml.ftype;
-
-    if (hparams.f_max_alibi_bias > 0.0f) {
-        hparams.use_alibi = true;
-    }
-
-    hparams.rope_type = llama_rope_type(&model);
-}
-
-static void llm_load_vocab(
-        llama_model_loader & ml,
-        llama_model & model) {
-    auto & vocab = model.vocab;
-
-    struct gguf_context * ctx = ml.meta.get();
-
-    const auto kv = LLM_KV(model.arch);
-
-    // determine vocab type
-    {
-        std::string tokenizer_model;
-        std::string tokenizer_pre;
-
-        ml.get_key(LLM_KV_TOKENIZER_MODEL, tokenizer_model);
-        ml.get_key(LLM_KV_TOKENIZER_PRE,   tokenizer_pre, false);
-
-        if (tokenizer_model == "no_vocab") {
-            vocab.type = LLAMA_VOCAB_TYPE_NONE;
-
-            // default special tokens
-            vocab.special_bos_id  = LLAMA_TOKEN_NULL;
-            vocab.special_eos_id  = LLAMA_TOKEN_NULL;
-            vocab.special_unk_id  = LLAMA_TOKEN_NULL;
-            vocab.special_sep_id  = LLAMA_TOKEN_NULL;
-            vocab.special_pad_id  = LLAMA_TOKEN_NULL;
-            vocab.special_cls_id  = LLAMA_TOKEN_NULL;
-            vocab.special_mask_id = LLAMA_TOKEN_NULL;
-            vocab.linefeed_id     = LLAMA_TOKEN_NULL;
-
-            // read vocab size from metadata
-            if (!ml.get_key(LLM_KV_VOCAB_SIZE, vocab.n_vocab, false)) {
-                vocab.n_vocab = 0;
-                LLAMA_LOG_WARN("%s: there is no vocab_size in metadata, vocab.n_vocab will be set to %u\n", __func__, vocab.n_vocab);
-            }
-            return;
-        }
-
-        if (tokenizer_model == "llama") {
-            vocab.type = LLAMA_VOCAB_TYPE_SPM;
-
-            // default special tokens
-            vocab.special_bos_id  = 1;
-            vocab.special_eos_id  = 2;
-            vocab.special_unk_id  = 0;
-            vocab.special_sep_id  = LLAMA_TOKEN_NULL;
-            vocab.special_pad_id  = LLAMA_TOKEN_NULL;
-            vocab.special_cls_id  = LLAMA_TOKEN_NULL;
-            vocab.special_mask_id = LLAMA_TOKEN_NULL;
-        } else if (tokenizer_model == "bert") {
-            vocab.type = LLAMA_VOCAB_TYPE_WPM;
-
-            // default special tokens
-            vocab.special_bos_id  = LLAMA_TOKEN_NULL;
-            vocab.special_eos_id  = LLAMA_TOKEN_NULL;
-            vocab.special_unk_id  = 100;
-            vocab.special_sep_id  = 102;
-            vocab.special_pad_id  = 0;
-            vocab.special_cls_id  = 101;
-            vocab.special_mask_id = 103;
-        } else if (tokenizer_model == "gpt2") {
-            vocab.type = LLAMA_VOCAB_TYPE_BPE;
-
-            // read bpe merges and populate bpe ranks
-            const int merges_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_MERGES).c_str());
-            if (merges_keyidx == -1) {
-                throw std::runtime_error("cannot find tokenizer merges in model file\n");
-            }
-
-            const int n_merges = gguf_get_arr_n(ctx, merges_keyidx);
-            for (int i = 0; i < n_merges; i++) {
-                const std::string word = gguf_get_arr_str(ctx, merges_keyidx, i);
-                GGML_ASSERT(unicode_cpts_from_utf8(word).size() > 0);
-
-                std::string first;
-                std::string second;
-
-                const size_t pos = word.find(' ', 1);
-
-                if (pos != std::string::npos) {
-                    first  = word.substr(0, pos);
-                    second = word.substr(pos + 1);
-                }
-
-                vocab.bpe_ranks.emplace(std::make_pair(first, second), i);
-            }
-
-            // default special tokens
-            vocab.special_bos_id  = 11;
-            vocab.special_eos_id  = 11;
-            vocab.special_unk_id  = LLAMA_TOKEN_NULL;
-            vocab.special_sep_id  = LLAMA_TOKEN_NULL;
-            vocab.special_pad_id  = LLAMA_TOKEN_NULL;
-            vocab.special_cls_id  = LLAMA_TOKEN_NULL;
-            vocab.special_mask_id = LLAMA_TOKEN_NULL;
-        } else if (tokenizer_model == "t5") {
-            vocab.type = LLAMA_VOCAB_TYPE_UGM;
-
-            // default special tokens
-            vocab.special_bos_id  = LLAMA_TOKEN_NULL;
-            vocab.special_eos_id  = 1;
-            vocab.special_unk_id  = 2;
-            vocab.special_sep_id  = LLAMA_TOKEN_NULL;
-            vocab.special_pad_id  = 0;
-            vocab.special_cls_id  = LLAMA_TOKEN_NULL;
-            vocab.special_mask_id = LLAMA_TOKEN_NULL;
-
-            const int precompiled_charsmap_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_PRECOMPILED_CHARSMAP).c_str());
-            if (precompiled_charsmap_keyidx != -1) {
-                size_t n_precompiled_charsmap = gguf_get_arr_n(ctx, precompiled_charsmap_keyidx);
-                const char * precompiled_charsmap = (const char *) gguf_get_arr_data(ctx, precompiled_charsmap_keyidx);
-                vocab.precompiled_charsmap.assign(precompiled_charsmap, precompiled_charsmap + n_precompiled_charsmap);
-#ifdef IS_BIG_ENDIAN
-                // correct endiannes of data in precompiled_charsmap binary blob
-                uint32_t * xcda_blob_size = (uint32_t *) &vocab.precompiled_charsmap[0];
-                *xcda_blob_size = __builtin_bswap32(*xcda_blob_size);
-                assert(*xcda_blob_size + sizeof(uint32_t) < n_precompiled_charsmap);
-                size_t xcda_array_size = *xcda_blob_size / sizeof(uint32_t);
-                uint32_t * xcda_array = (uint32_t *) &vocab.precompiled_charsmap[sizeof(uint32_t)];
-                for (size_t i = 0; i < xcda_array_size; ++i) {
-                    xcda_array[i] = __builtin_bswap32(xcda_array[i]);
-                }
-#endif
-            }
-        } else if (tokenizer_model == "rwkv") {
-            vocab.type = LLAMA_VOCAB_TYPE_RWKV;
-
-            // default special tokens
-            vocab.special_bos_id = LLAMA_TOKEN_NULL;
-            vocab.special_eos_id = LLAMA_TOKEN_NULL;
-            vocab.special_unk_id = LLAMA_TOKEN_NULL;
-            vocab.special_sep_id = LLAMA_TOKEN_NULL;
-            vocab.special_pad_id = LLAMA_TOKEN_NULL;
-        } else {
-            throw std::runtime_error(format("unknown tokenizer: '%s'", tokenizer_model.c_str()));
-        }
-
-        // for now, only BPE models have pre-tokenizers
-        if (vocab.type == LLAMA_VOCAB_TYPE_BPE) {
-            vocab.tokenizer_add_space_prefix = false;
-            vocab.tokenizer_clean_spaces = true;
-            if (tokenizer_pre.empty()) {
-                LLAMA_LOG_WARN("%s: missing pre-tokenizer type, using: 'default'\n", __func__);
-                LLAMA_LOG_WARN("%s:                                             \n", __func__);
-                LLAMA_LOG_WARN("%s: ************************************        \n", __func__);
-                LLAMA_LOG_WARN("%s: GENERATION QUALITY WILL BE DEGRADED!        \n", __func__);
-                LLAMA_LOG_WARN("%s: CONSIDER REGENERATING THE MODEL             \n", __func__);
-                LLAMA_LOG_WARN("%s: ************************************        \n", __func__);
-                LLAMA_LOG_WARN("%s:                                             \n", __func__);
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            } else if (tokenizer_pre == "default") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            } else if (
-                    tokenizer_pre == "llama3"   ||
-                    tokenizer_pre == "llama-v3" ||
-                    tokenizer_pre == "llama-bpe") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_LLAMA3;
-                vocab.tokenizer_ignore_merges = true;
-                vocab.tokenizer_add_bos = true;
-            } else if (
-                    tokenizer_pre == "deepseek-llm") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                    tokenizer_pre == "deepseek-coder") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                    tokenizer_pre == "falcon") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_FALCON;
-            } else if (
-                    tokenizer_pre == "mpt") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_MPT;
-            } else if (
-                    tokenizer_pre == "starcoder") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_STARCODER;
-            } else if (
-                    tokenizer_pre == "gpt-2"   ||
-                    tokenizer_pre == "phi-2"   ||
-                    tokenizer_pre == "jina-es" ||
-                    tokenizer_pre == "jina-de" ||
-                    tokenizer_pre == "jina-v1-en" ||
-                    tokenizer_pre == "jina-v2-es" ||
-                    tokenizer_pre == "jina-v2-de" ||
-                    tokenizer_pre == "jina-v2-code") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_GPT2;
-            } else if (
-                    tokenizer_pre == "refact") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_REFACT;
-            } else if (
-                tokenizer_pre == "command-r") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_COMMAND_R;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                tokenizer_pre == "qwen2") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_QWEN2;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                tokenizer_pre == "stablelm2") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_STABLELM2;
-            } else if (
-                tokenizer_pre == "olmo") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_OLMO;
-            } else if (
-                tokenizer_pre == "dbrx") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DBRX;
-            } else if (
-                tokenizer_pre == "smaug-bpe") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_SMAUG;
-            } else if (
-                tokenizer_pre == "poro-chat") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_PORO;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                tokenizer_pre == "chatglm-bpe") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_CHATGLM4;
-                vocab.special_bos_id = LLAMA_TOKEN_NULL;
-            } else if (
-                tokenizer_pre == "viking") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_VIKING;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                tokenizer_pre == "jais") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_JAIS;
-            } else if (
-                tokenizer_pre == "tekken") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_TEKKEN;
-                vocab.tokenizer_clean_spaces = false;
-                vocab.tokenizer_ignore_merges = true;
-                vocab.tokenizer_add_bos = true;
-            } else if (
-                tokenizer_pre == "smollm") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_SMOLLM;
-                vocab.tokenizer_clean_spaces = false;
-            } else if (
-                tokenizer_pre == "codeshell") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_CODESHELL;
-            } else if (
-                tokenizer_pre == "bloom") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_BLOOM;
-            } else if (
-                tokenizer_pre == "gpt3-finnish") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_GPT3_FINNISH;
-            } else if (
-                tokenizer_pre == "exaone") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_EXAONE;
-            } else if (
-                tokenizer_pre == "chameleon") {
-                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_CHAMELEON;
-                vocab.tokenizer_add_bos = true;
-                vocab.tokenizer_clean_spaces = false;
-            } else {
-                throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
-            }
-        } else if (vocab.type == LLAMA_VOCAB_TYPE_SPM) {
-            vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            vocab.tokenizer_add_space_prefix = true;
-            vocab.tokenizer_clean_spaces = false;
-            vocab.tokenizer_add_bos = true;
-            vocab.tokenizer_add_eos = false;
-        } else if (vocab.type == LLAMA_VOCAB_TYPE_WPM) {
-            vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            vocab.tokenizer_add_space_prefix = false;
-            vocab.tokenizer_clean_spaces = true;
-            vocab.tokenizer_add_bos = true;
-            vocab.tokenizer_add_eos = false;
-        } else if (vocab.type == LLAMA_VOCAB_TYPE_UGM) {
-            vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            vocab.tokenizer_add_bos = false;
-            vocab.tokenizer_add_eos = true;
-        } else if (vocab.type == LLAMA_VOCAB_TYPE_RWKV) {
-            vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-            vocab.tokenizer_add_space_prefix = false;
-            vocab.tokenizer_clean_spaces = false;
-            vocab.tokenizer_add_bos = false;
-            vocab.tokenizer_add_eos = false;
-        } else {
-            vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
-        }
-
-        ml.get_key(LLM_KV_TOKENIZER_ADD_PREFIX,      vocab.tokenizer_add_space_prefix,         false);
-        ml.get_key(LLM_KV_TOKENIZER_REMOVE_EXTRA_WS, vocab.tokenizer_remove_extra_whitespaces, false);
-    }
-
-    const int token_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_LIST).c_str());
-    if (token_idx == -1) {
-        throw std::runtime_error("cannot find tokenizer vocab in model file\n");
-    }
-
-    const float * scores = nullptr;
-    const int score_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_SCORES).c_str());
-    if (score_idx != -1) {
-        scores = (const float * ) gguf_get_arr_data(ctx, score_idx);
-    }
-
-    const int * toktypes = nullptr;
-    const int toktype_idx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_TOKEN_TYPE).c_str());
-    if (toktype_idx != -1) {
-        toktypes = (const int * ) gguf_get_arr_data(ctx, toktype_idx);
-    }
-
-    const uint32_t n_vocab = gguf_get_arr_n(ctx, token_idx);
-
-    vocab.n_vocab = n_vocab;
-    vocab.id_to_token.resize(n_vocab);
-
-    for (uint32_t i = 0; i < n_vocab; i++) {
-        std::string word = gguf_get_arr_str(ctx, token_idx, i);
-
-        //GGML_ASSERT(unicode_cpts_from_utf8(word).size() > 0);
-        if (word.empty()) {
-            LLAMA_LOG_WARN("%s: empty token at index %u\n", __func__, i);
-            word = "[EMPTY_" + std::to_string(i) + "]";
-        }
-
-        vocab.token_to_id[word] = i;
-        vocab.max_token_len = std::max(vocab.max_token_len, (int) word.size());
-
-        auto & token_data = vocab.id_to_token[i];
-        token_data.text  = std::move(word);
-        token_data.score = scores ? scores[i] : 0.0f;
-        token_data.attr  = LLAMA_TOKEN_ATTR_NORMAL;
-
-        if (toktypes) {  //TODO: remove, required until per token attributes are available from GGUF file
-            switch(toktypes[i]) {
-                case LLAMA_TOKEN_TYPE_UNKNOWN:      token_data.attr = LLAMA_TOKEN_ATTR_UNKNOWN;      break;
-                case LLAMA_TOKEN_TYPE_UNUSED:       token_data.attr = LLAMA_TOKEN_ATTR_UNUSED;       break;
-                case LLAMA_TOKEN_TYPE_NORMAL:       token_data.attr = LLAMA_TOKEN_ATTR_NORMAL;       break;
-                case LLAMA_TOKEN_TYPE_CONTROL:      token_data.attr = LLAMA_TOKEN_ATTR_CONTROL;      break;
-                case LLAMA_TOKEN_TYPE_USER_DEFINED: token_data.attr = LLAMA_TOKEN_ATTR_USER_DEFINED; break;
-                case LLAMA_TOKEN_TYPE_BYTE:         token_data.attr = LLAMA_TOKEN_ATTR_BYTE;         break;
-                case LLAMA_TOKEN_TYPE_UNDEFINED:    token_data.attr = LLAMA_TOKEN_ATTR_UNDEFINED;    break;
-                default:                            token_data.attr = LLAMA_TOKEN_ATTR_UNDEFINED;    break;
-            }
-        }
-    }
-    GGML_ASSERT(vocab.id_to_token.size() == vocab.token_to_id.size());
-
-    vocab.init_tokenizer();
-
-    // determine the newline token: LLaMA "<0x0A>" == 10 == '\n', Falcon 193 == '\n'
-    if (vocab.type == LLAMA_VOCAB_TYPE_SPM) {
-        try {
-            vocab.linefeed_id = llama_byte_to_token_impl(vocab, '\n');
-        } catch (const std::exception & e) {
-            LLAMA_LOG_WARN("%s: SPM vocabulary, but newline token not found: %s! Using special_pad_id instead.", __func__, e.what());
-            vocab.linefeed_id = vocab.special_pad_id;
-        }
-    } else if (vocab.type == LLAMA_VOCAB_TYPE_WPM) {
-        vocab.linefeed_id = vocab.special_pad_id;
-    } else if (vocab.type == LLAMA_VOCAB_TYPE_RWKV) {
-        const std::vector ids = llama_tokenize_internal(vocab, "\n", false);
-        GGML_ASSERT(!ids.empty() && "model vocab missing newline token");
-        vocab.linefeed_id = ids[0];
-    } else {
-        const std::vector ids = llama_tokenize_internal(vocab, "\xC4\x8A", false); // U+010A
-
-        //GGML_ASSERT(!ids.empty() && "model vocab missing newline token");
-        if (ids.empty()) {
-            LLAMA_LOG_WARN("%s: model vocab missing newline token, using special_pad_id instead\n", __func__);
-            vocab.linefeed_id = vocab.special_pad_id;
-        } else {
-            vocab.linefeed_id = ids[0];
-        }
-    }
-
-    // special tokens
-    {
-        const std::vector> special_token_types = {
-            { LLM_KV_TOKENIZER_BOS_ID,     vocab.special_bos_id     },
-            { LLM_KV_TOKENIZER_EOS_ID,     vocab.special_eos_id     },
-            { LLM_KV_TOKENIZER_EOT_ID,     vocab.special_eot_id     },
-            { LLM_KV_TOKENIZER_EOM_ID,     vocab.special_eom_id     },
-            { LLM_KV_TOKENIZER_UNK_ID,     vocab.special_unk_id     },
-            { LLM_KV_TOKENIZER_SEP_ID,     vocab.special_sep_id     },
-            { LLM_KV_TOKENIZER_PAD_ID,     vocab.special_pad_id     },
-            { LLM_KV_TOKENIZER_CLS_ID,     vocab.special_cls_id     },
-            { LLM_KV_TOKENIZER_MASK_ID,    vocab.special_mask_id    },
-            { LLM_KV_TOKENIZER_FIM_PRE_ID, vocab.special_fim_pre_id },
-            { LLM_KV_TOKENIZER_FIM_SUF_ID, vocab.special_fim_suf_id },
-            { LLM_KV_TOKENIZER_FIM_MID_ID, vocab.special_fim_mid_id },
-            { LLM_KV_TOKENIZER_FIM_PAD_ID, vocab.special_fim_pad_id },
-            { LLM_KV_TOKENIZER_FIM_REP_ID, vocab.special_fim_rep_id },
-            { LLM_KV_TOKENIZER_FIM_SEP_ID, vocab.special_fim_sep_id },
-
-            // deprecated
-            { LLM_KV_TOKENIZER_PREFIX_ID, vocab.special_fim_pre_id },
-            { LLM_KV_TOKENIZER_SUFFIX_ID, vocab.special_fim_suf_id },
-            { LLM_KV_TOKENIZER_MIDDLE_ID, vocab.special_fim_mid_id },
-        };
-
-        for (const auto & it : special_token_types) {
-            const std::string & key = kv(std::get<0>(it));
-            int32_t & id = std::get<1>(it);
-
-            uint32_t new_id;
-            if (!ml.get_key(std::get<0>(it), new_id, false)) {
-                continue;
-            }
-            if (new_id >= vocab.id_to_token.size()) {
-                LLAMA_LOG_WARN("%s: bad special token: '%s' = %ud, using default id %d\n",
-                    __func__, key.c_str(), new_id, id);
-            } else {
-                id = new_id;
-            }
-        }
-
-        // Handle add_bos_token and add_eos_token
-        {
-            bool temp = true;
-
-            if (ml.get_key(LLM_KV_TOKENIZER_ADD_BOS, temp, false)) {
-                vocab.tokenizer_add_bos = temp;
-            }
-            if (ml.get_key(LLM_KV_TOKENIZER_ADD_EOS, temp, false)) {
-                vocab.tokenizer_add_eos = temp;
-            }
-        }
-
-        // auto-detect special tokens by text
-        // TODO: convert scripts should provide these tokens through the KV metadata LLM_KV_TOKENIZER_...
-        //       for now, we apply this workaround to find the tokens based on their text
-
-        for (const auto & t : vocab.token_to_id) {
-            // find EOT token: "<|eot_id|>", "<|im_end|>", "", etc.
-            if (vocab.special_eot_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|eot_id|>"
-                        || t.first == "<|im_end|>"
-                        || t.first == "<|end|>"
-                        || t.first == ""
-                        || t.first == "<|endoftext|>"
-                        || t.first == ""
-                        || t.first == "<｜end▁of▁sentence｜>" // DeepSeek
-                   ) {
-                    vocab.special_eot_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find EOM token: "<|eom_id|>"
-            if (vocab.special_eom_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|eom_id|>"
-                        ) {
-                    vocab.special_eom_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_PRE token: "<|fim_prefix|>", "", "
", etc.
-            if (vocab.special_fim_pre_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|fim_prefix|>"  // Qwen
-                        || t.first == ""
-                        || t.first == "<｜fim▁begin｜>" // DeepSeek
-                        || t.first == "
"
-                        ) {
-                    vocab.special_fim_pre_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_SUF token: "<|fim_suffix|>", "", "", etc.
-            if (vocab.special_fim_suf_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|fim_suffix|>" // Qwen
-                        || t.first == ""
-                        || t.first == "<｜fim▁hole｜>" // DeepSeek
-                        || t.first == ""
-                        ) {
-                    vocab.special_fim_suf_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_MID token: "<|fim_middle|>", "", "", etc.
-            if (vocab.special_fim_mid_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|fim_middle|>" // Qwen
-                        || t.first == ""
-                        || t.first == "<｜fim▁end｜>"  // DeepSeek
-                        || t.first == ""
-                        ) {
-                    vocab.special_fim_mid_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_PAD token: "<|fim_pad|>", "", "", etc.
-            if (vocab.special_fim_pad_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|fim_pad|>" // Qwen
-                        || t.first == ""
-                        || t.first == ""
-                        ) {
-                    vocab.special_fim_pad_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_REP token: "<|fim_repo|>", "", "", etc.
-            if (vocab.special_fim_rep_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|fim_repo|>"  // Qwen
-                        || t.first == "<|repo_name|>"
-                        || t.first == ""
-                        || t.first == ""
-                        ) {
-                    vocab.special_fim_rep_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-
-            // find FIM_SEP token: "<|file_sep|>"
-            if (vocab.special_fim_sep_id == LLAMA_TOKEN_NULL) {
-                if (false
-                        || t.first == "<|file_sep|>" // Qwen
-                        ) {
-                    vocab.special_fim_sep_id = t.second;
-                    if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                        LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                                __func__, t.second, t.first.c_str());
-                        vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                    }
-                }
-            }
-        }
-
-        // maintain a list of tokens that cause end-of-generation
-        // this is currently determined based on the token text, which is obviously not ideal
-        // ref: https://github.com/ggerganov/llama.cpp/issues/9606
-        vocab.special_eog_ids.clear();
-
-        if (vocab.special_fim_pad_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_fim_pad_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_fim_pad_id);
-        }
-
-        if (vocab.special_fim_rep_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_fim_rep_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_fim_rep_id);
-        }
-
-        if (vocab.special_fim_sep_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_fim_sep_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_fim_sep_id);
-        }
-
-        for (const auto & t : vocab.token_to_id) {
-            if (false
-                    || t.first == "<|eot_id|>"
-                    || t.first == "<|im_end|>"
-                    || t.first == "<|end|>"
-                    || t.first == ""
-                    || t.first == "<|endoftext|>"
-                    || t.first == "<|eom_id|>"
-                    || t.first == ""
-               ) {
-                vocab.special_eog_ids.insert(t.second);
-                if ((vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL) == 0) {
-                    LLAMA_LOG_WARN("%s: control-looking token: %6d '%s' was not control-type; this is probably a bug in the model. its type will be overridden\n",
-                            __func__, t.second, t.first.c_str());
-                    vocab.id_to_token[t.second].attr = LLAMA_TOKEN_ATTR_CONTROL;
-                }
-            } else {
-                // token is control, but not marked as EOG -> print a debug log
-                if (vocab.id_to_token[t.second].attr & LLAMA_TOKEN_ATTR_CONTROL && vocab.special_eog_ids.count(t.second) == 0) {
-                    LLAMA_LOG_DEBUG("%s: control token: %6d '%s' is not marked as EOG\n",
-                            __func__, t.second, t.first.c_str());
-                }
-            }
-        }
-
-        // sanity checks
-        if (vocab.special_eos_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_eos_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_eos_id);
-            LLAMA_LOG_WARN("%s: special_eos_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
-        }
-
-        if (vocab.special_eot_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_eot_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_eot_id);
-            LLAMA_LOG_WARN("%s: special_eot_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
-        }
-
-        if (vocab.special_eom_id != LLAMA_TOKEN_NULL && vocab.special_eog_ids.count(vocab.special_eom_id) == 0) {
-            vocab.special_eog_ids.insert(vocab.special_eom_id);
-            LLAMA_LOG_WARN("%s: special_eom_id is not in special_eog_ids - the tokenizer config may be incorrect\n", __func__);
-        }
-    }
-
-    // build special tokens cache
-    {
-        for (llama_vocab::id id = 0; id < (llama_vocab::id)n_vocab; ++id) {
-            if (vocab.id_to_token[id].attr & (LLAMA_TOKEN_ATTR_CONTROL | LLAMA_TOKEN_ATTR_USER_DEFINED | LLAMA_TOKEN_ATTR_UNKNOWN)) {
-                vocab.cache_special_tokens.push_back(id);
-            }
-        }
-
-        std::sort(vocab.cache_special_tokens.begin(), vocab.cache_special_tokens.end(),
-            [&] (const llama_vocab::id a, const llama_vocab::id b) {
-                return vocab.id_to_token[a].text.size() > vocab.id_to_token[b].text.size();
-            }
-        );
-
-        LLAMA_LOG_INFO("%s: special tokens cache size = %u\n", __func__, (uint32_t)vocab.cache_special_tokens.size());
-    }
-
-    // build token to piece cache
-    {
-        size_t size_cache = 0;
-
-        std::vector cache_token_to_piece(n_vocab);
-
-        for (uint32_t id = 0; id < n_vocab; ++id) {
-            cache_token_to_piece[id] = llama_token_to_piece(&model, id, true);
-
-            size_cache += cache_token_to_piece[id].size();
-        }
-
-        std::swap(vocab.cache_token_to_piece, cache_token_to_piece);
-
-        LLAMA_LOG_INFO("%s: token to piece cache size = %.4f MB\n", __func__, size_cache / 1024.0 / 1024.0);
-    }
-
-    // Handle per token attributes
-    //NOTE: Each model customizes per token attributes.
-    //NOTE: Per token attributes are missing from the GGUF file.
-    //TODO: Extract attributes from GGUF file.
-    {
-        auto _contains_any = [] (const std::string &str, const std::vector &substrs) -> bool {
-            for (auto substr : substrs) {
-                if (str.find(substr) < std::string::npos) {
-                    return true;
-                }
-            }
-            return false;
-        };
-
-        auto _set_tokenid_attr = [&] (const llama_vocab::id id, llama_token_attr attr, bool value) {
-            uint32_t current = vocab.id_to_token.at(id).attr;
-            current = value ? (current | attr) : (current & ~attr);
-            vocab.id_to_token[id].attr = (llama_token_attr) current;
-        };
-
-        auto _set_token_attr = [&] (const std::string & token, llama_token_attr attr, bool value) {
-            _set_tokenid_attr(vocab.token_to_id.at(token), attr, value);
-        };
-
-        std::string model_name;
-        std::string tokenizer_pre;
-
-        ml.get_key(LLM_KV_GENERAL_NAME, model_name, false);
-        ml.get_key(LLM_KV_TOKENIZER_PRE, tokenizer_pre, false);
-
-        // model name to lowercase
-        std::transform(model_name.begin(), model_name.end(), model_name.begin(),
-            [] (const std::string::value_type x) {
-                return std::tolower(x);
-            }
-        );
-
-        // set attributes by model/tokenizer name
-        if (_contains_any(tokenizer_pre, {"jina-v2-de", "jina-v2-es", "jina-v2-code"})) {
-            _set_token_attr("", LLAMA_TOKEN_ATTR_LSTRIP, true);
-        } else if (_contains_any(model_name, {"phi-3", "phi3"})) {
-            for (auto id : vocab.cache_special_tokens) {
-                _set_tokenid_attr(id, LLAMA_TOKEN_ATTR_RSTRIP, true);
-            }
-            for (auto token : {""}) {
-                _set_token_attr(token, LLAMA_TOKEN_ATTR_RSTRIP, true);
-            }
-            for (auto token : {"", "", "<|endoftext|>"}) {
-                _set_token_attr(token, LLAMA_TOKEN_ATTR_RSTRIP, false);
-            }
-        }
-    }
-}
-
-static void llm_load_print_meta(llama_model_loader & ml, llama_model & model) {
-    const auto & hparams = model.hparams;
-    const auto & vocab   = model.vocab;
-
-    const char * rope_scaling_type = LLAMA_ROPE_SCALING_TYPES.at(hparams.rope_scaling_type_train);
-
-    auto print_f = [](const std::function & f, uint32_t n) {
-        bool is_var = false;
-
-        std::vector v;
-        for (uint32_t i = 0; i < n; ++i) {
-            v.push_back(f(i));
-            if (v[i] != v[0]) {
-                is_var = true;
-            }
-        }
-
-        std::stringstream ss;
-
-        if (is_var) {
-            ss << "[";
-            for (uint32_t i = 0; i < n; ++i) {
-                ss << v[i];
-                if (i < n - 1) {
-                    ss << ", ";
-                }
-            }
-            ss << "]";
-        } else {
-            ss << v[0];
-        }
-
-        return ss.str();
-    };
-
-    // hparams
-    LLAMA_LOG_INFO("%s: format           = %s\n",     __func__, llama_file_version_name(ml.fver));
-    LLAMA_LOG_INFO("%s: arch             = %s\n",     __func__, LLM_ARCH_NAMES.at(model.arch));
-    LLAMA_LOG_INFO("%s: vocab type       = %s\n",     __func__, llama_model_vocab_type_name(vocab.type));
-    LLAMA_LOG_INFO("%s: n_vocab          = %u\n",     __func__, hparams.n_vocab);
-    LLAMA_LOG_INFO("%s: n_merges         = %u\n",     __func__, (int) vocab.bpe_ranks.size());
-    LLAMA_LOG_INFO("%s: vocab_only       = %d\n",     __func__, hparams.vocab_only);
-
-    if (!hparams.vocab_only) {
-        LLAMA_LOG_INFO("%s: n_ctx_train      = %u\n",     __func__, hparams.n_ctx_train);
-        LLAMA_LOG_INFO("%s: n_embd           = %u\n",     __func__, hparams.n_embd);
-        LLAMA_LOG_INFO("%s: n_layer          = %u\n",     __func__, hparams.n_layer);
-        LLAMA_LOG_INFO("%s: n_head           = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_head(il);    }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: n_head_kv        = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_head_kv(il); }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: n_rot            = %u\n",     __func__, hparams.n_rot);
-        LLAMA_LOG_INFO("%s: n_swa            = %u\n",     __func__, hparams.n_swa);
-        LLAMA_LOG_INFO("%s: n_embd_head_k    = %u\n",     __func__, hparams.n_embd_head_k);
-        LLAMA_LOG_INFO("%s: n_embd_head_v    = %u\n",     __func__, hparams.n_embd_head_v);
-        LLAMA_LOG_INFO("%s: n_gqa            = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_gqa(il);        }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: n_embd_k_gqa     = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_embd_k_gqa(il); }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: n_embd_v_gqa     = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_embd_v_gqa(il); }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: f_norm_eps       = %.1e\n",   __func__, hparams.f_norm_eps);
-        LLAMA_LOG_INFO("%s: f_norm_rms_eps   = %.1e\n",   __func__, hparams.f_norm_rms_eps);
-        LLAMA_LOG_INFO("%s: f_clamp_kqv      = %.1e\n",   __func__, hparams.f_clamp_kqv);
-        LLAMA_LOG_INFO("%s: f_max_alibi_bias = %.1e\n",   __func__, hparams.f_max_alibi_bias);
-        LLAMA_LOG_INFO("%s: f_logit_scale    = %.1e\n",   __func__, hparams.f_logit_scale);
-        LLAMA_LOG_INFO("%s: n_ff             = %s\n",     __func__, print_f([&](uint32_t il) { return hparams.n_ff(il); }, hparams.n_layer).c_str());
-        LLAMA_LOG_INFO("%s: n_expert         = %u\n",     __func__, hparams.n_expert);
-        LLAMA_LOG_INFO("%s: n_expert_used    = %u\n",     __func__, hparams.n_expert_used);
-        LLAMA_LOG_INFO("%s: causal attn      = %d\n",     __func__, hparams.causal_attn);
-        LLAMA_LOG_INFO("%s: pooling type     = %d\n",     __func__, hparams.pooling_type);
-        LLAMA_LOG_INFO("%s: rope type        = %d\n",     __func__, hparams.rope_type);
-        LLAMA_LOG_INFO("%s: rope scaling     = %s\n",     __func__, rope_scaling_type);
-        LLAMA_LOG_INFO("%s: freq_base_train  = %.1f\n",   __func__, hparams.rope_freq_base_train);
-        LLAMA_LOG_INFO("%s: freq_scale_train = %g\n",     __func__, hparams.rope_freq_scale_train);
-        LLAMA_LOG_INFO("%s: n_ctx_orig_yarn  = %u\n",     __func__, hparams.n_ctx_orig_yarn);
-        LLAMA_LOG_INFO("%s: rope_finetuned   = %s\n",     __func__, hparams.rope_finetuned ? "yes" : "unknown");
-        LLAMA_LOG_INFO("%s: ssm_d_conv       = %u\n",     __func__, hparams.ssm_d_conv);
-        LLAMA_LOG_INFO("%s: ssm_d_inner      = %u\n",     __func__, hparams.ssm_d_inner);
-        LLAMA_LOG_INFO("%s: ssm_d_state      = %u\n",     __func__, hparams.ssm_d_state);
-        LLAMA_LOG_INFO("%s: ssm_dt_rank      = %u\n",     __func__, hparams.ssm_dt_rank);
-        LLAMA_LOG_INFO("%s: ssm_dt_b_c_rms   = %d\n",     __func__, hparams.ssm_dt_b_c_rms);
-    }
-
-    LLAMA_LOG_INFO("%s: model type       = %s\n",     __func__, llama_model_type_name(model.type));
-    LLAMA_LOG_INFO("%s: model ftype      = %s\n",     __func__, llama_model_ftype_name(model.ftype).c_str());
-    if (ml.n_elements >= 1e12) {
-        LLAMA_LOG_INFO("%s: model params     = %.2f T\n", __func__, ml.n_elements*1e-12);
-    } else if (ml.n_elements >= 1e9) {
-        LLAMA_LOG_INFO("%s: model params     = %.2f B\n", __func__, ml.n_elements*1e-9);
-    } else if (ml.n_elements >= 1e6) {
-        LLAMA_LOG_INFO("%s: model params     = %.2f M\n", __func__, ml.n_elements*1e-6);
-    } else {
-        LLAMA_LOG_INFO("%s: model params     = %.2f K\n", __func__, ml.n_elements*1e-3);
-    }
-    if (ml.n_bytes < GiB) {
-        LLAMA_LOG_INFO("%s: model size       = %.2f MiB (%.2f BPW) \n", __func__, ml.n_bytes/1024.0/1024.0,        ml.n_bytes*8.0/ml.n_elements);
-    } else {
-        LLAMA_LOG_INFO("%s: model size       = %.2f GiB (%.2f BPW) \n", __func__, ml.n_bytes/1024.0/1024.0/1024.0, ml.n_bytes*8.0/ml.n_elements);
-    }
-
-    // general kv
-    LLAMA_LOG_INFO("%s: general.name     = %s\n",    __func__, model.name.c_str());
-
-    // special tokens
-    if (vocab.special_bos_id  != -1)    { LLAMA_LOG_INFO( "%s: BOS token        = %d '%s'\n", __func__, vocab.special_bos_id,     vocab.id_to_token[vocab.special_bos_id].text.c_str() );  }
-    if (vocab.special_eos_id  != -1)    { LLAMA_LOG_INFO( "%s: EOS token        = %d '%s'\n", __func__, vocab.special_eos_id,     vocab.id_to_token[vocab.special_eos_id].text.c_str() );  }
-    if (vocab.special_eot_id  != -1)    { LLAMA_LOG_INFO( "%s: EOT token        = %d '%s'\n", __func__, vocab.special_eot_id,     vocab.id_to_token[vocab.special_eot_id].text.c_str() );  }
-    if (vocab.special_eom_id  != -1)    { LLAMA_LOG_INFO( "%s: EOM token        = %d '%s'\n", __func__, vocab.special_eom_id,     vocab.id_to_token[vocab.special_eom_id].text.c_str() );  }
-    if (vocab.special_unk_id  != -1)    { LLAMA_LOG_INFO( "%s: UNK token        = %d '%s'\n", __func__, vocab.special_unk_id,     vocab.id_to_token[vocab.special_unk_id].text.c_str() );  }
-    if (vocab.special_sep_id  != -1)    { LLAMA_LOG_INFO( "%s: SEP token        = %d '%s'\n", __func__, vocab.special_sep_id,     vocab.id_to_token[vocab.special_sep_id].text.c_str() );  }
-    if (vocab.special_pad_id  != -1)    { LLAMA_LOG_INFO( "%s: PAD token        = %d '%s'\n", __func__, vocab.special_pad_id,     vocab.id_to_token[vocab.special_pad_id].text.c_str() );  }
-    if (vocab.special_cls_id  != -1)    { LLAMA_LOG_INFO( "%s: CLS token        = %d '%s'\n", __func__, vocab.special_cls_id,     vocab.id_to_token[vocab.special_cls_id].text.c_str() );  }
-    if (vocab.special_mask_id != -1)    { LLAMA_LOG_INFO( "%s: MASK token       = %d '%s'\n", __func__, vocab.special_mask_id,    vocab.id_to_token[vocab.special_mask_id].text.c_str() ); }
-
-    if (vocab.linefeed_id != -1)        { LLAMA_LOG_INFO( "%s: LF token         = %d '%s'\n", __func__, vocab.linefeed_id,        vocab.id_to_token[vocab.linefeed_id].text.c_str() ); }
-
-    if (vocab.special_fim_pre_id != -1) { LLAMA_LOG_INFO( "%s: FIM PRE token    = %d '%s'\n", __func__, vocab.special_fim_pre_id, vocab.id_to_token[vocab.special_fim_pre_id].text.c_str() ); }
-    if (vocab.special_fim_suf_id != -1) { LLAMA_LOG_INFO( "%s: FIM SUF token    = %d '%s'\n", __func__, vocab.special_fim_suf_id, vocab.id_to_token[vocab.special_fim_suf_id].text.c_str() ); }
-    if (vocab.special_fim_mid_id != -1) { LLAMA_LOG_INFO( "%s: FIM MID token    = %d '%s'\n", __func__, vocab.special_fim_mid_id, vocab.id_to_token[vocab.special_fim_mid_id].text.c_str() ); }
-    if (vocab.special_fim_pad_id != -1) { LLAMA_LOG_INFO( "%s: FIM PAD token    = %d '%s'\n", __func__, vocab.special_fim_pad_id, vocab.id_to_token[vocab.special_fim_pad_id].text.c_str() ); }
-    if (vocab.special_fim_rep_id != -1) { LLAMA_LOG_INFO( "%s: FIM REP token    = %d '%s'\n", __func__, vocab.special_fim_rep_id, vocab.id_to_token[vocab.special_fim_rep_id].text.c_str() ); }
-    if (vocab.special_fim_sep_id != -1) { LLAMA_LOG_INFO( "%s: FIM SEP token    = %d '%s'\n", __func__, vocab.special_fim_sep_id, vocab.id_to_token[vocab.special_fim_sep_id].text.c_str() ); }
-
-    for (const auto & id : vocab.special_eog_ids) {
-        LLAMA_LOG_INFO( "%s: EOG token        = %d '%s'\n", __func__, id, vocab.id_to_token[id].text.c_str() );
-    }
-
-    LLAMA_LOG_INFO("%s: max token length = %d\n", __func__, vocab.max_token_len);
-
-    if (model.arch == LLM_ARCH_DEEPSEEK2) {
-        LLAMA_LOG_INFO("%s: n_layer_dense_lead   = %d\n",     __func__, hparams.n_layer_dense_lead);
-        LLAMA_LOG_INFO("%s: n_lora_q             = %d\n",     __func__, hparams.n_lora_q);
-        LLAMA_LOG_INFO("%s: n_lora_kv            = %d\n",     __func__, hparams.n_lora_kv);
-        LLAMA_LOG_INFO("%s: n_ff_exp             = %d\n",     __func__, hparams.n_ff_exp);
-        LLAMA_LOG_INFO("%s: n_expert_shared      = %d\n",     __func__, hparams.n_expert_shared);
-        LLAMA_LOG_INFO("%s: expert_weights_scale = %.1f\n",   __func__, hparams.expert_weights_scale);
-        LLAMA_LOG_INFO("%s: rope_yarn_log_mul    = %.4f\n",   __func__, hparams.rope_yarn_log_mul);
-    }
-
-    if (model.arch == LLM_ARCH_QWEN2MOE) {
-        LLAMA_LOG_INFO("%s: n_ff_exp         = %d\n",     __func__, hparams.n_ff_exp);
-        LLAMA_LOG_INFO("%s: n_ff_shexp       = %d\n",     __func__, hparams.n_ff_shexp);
-    }
-
-    if (model.arch == LLM_ARCH_GRANITE || model.arch == LLM_ARCH_GRANITE_MOE) {
-        LLAMA_LOG_INFO("%s: f_embedding_scale = %f\n", __func__, hparams.f_embedding_scale);
-        LLAMA_LOG_INFO("%s: f_residual_scale  = %f\n", __func__, hparams.f_residual_scale);
-        LLAMA_LOG_INFO("%s: f_attention_scale = %f\n", __func__, hparams.f_attention_scale);
-    }
-}
-
-enum llm_tensor_layer {
-    LLM_TENSOR_LAYER_INPUT,
-    LLM_TENSOR_LAYER_REPEATING,
-    LLM_TENSOR_LAYER_OUTPUT,
-};
-
-struct llm_tensor_info {
-    llm_tensor_layer layer;
-    ggml_op op;
-};
-
-static const std::map llm_tensor_info_mapping = {
-    {LLM_TENSOR_TOKEN_EMBD,                 {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_POS_EMBD,                   {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_TOKEN_EMBD_NORM,            {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_TOKEN_TYPES,                {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_OUTPUT,                     {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_CLS,                        {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_CLS_OUT,                    {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_OUTPUT_NORM,                {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}},
-    {LLM_TENSOR_DEC_OUTPUT_NORM,            {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}},
-    {LLM_TENSOR_ENC_OUTPUT_NORM,            {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL}},
-    {LLM_TENSOR_ROPE_FREQS,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}},
-    {LLM_TENSOR_ROPE_FACTORS_LONG,          {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}},
-    {LLM_TENSOR_ROPE_FACTORS_SHORT,         {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ROPE}},
-    {LLM_TENSOR_ATTN_Q,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_K,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_V,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_QKV,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_OUT,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_DOWN,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_UP,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_DOWN_SHEXP,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE_SHEXP,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_UP_SHEXP,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_Q_A,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_Q_B,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_KV_A_MQA,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_KV_B,                  {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_Q,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_K,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_Q,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_K,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_V,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_QKV,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_OUT,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_DOWN,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_UP,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_DOWN_SHEXP,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE_SHEXP,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_UP_SHEXP,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_Q_A,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_Q_B,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_KV_A_MQA,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ATTN_KV_B,                  {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_Q,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_K,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_V,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_ATTN_OUT,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_CROSS_ATTN_Q,           {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_CROSS_ATTN_K,           {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_CROSS_ATTN_V,           {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_CROSS_ATTN_OUT,         {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_FFN_GATE,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_FFN_DOWN,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_DEC_FFN_UP,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_ATTN_Q,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_ATTN_K,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_ATTN_V,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_ATTN_OUT,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_FFN_GATE,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_FFN_DOWN,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_ENC_FFN_UP,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE_INP_SHEXP,         {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_GATE_INP,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_SSM_IN,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_SSM_X,                      {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_SSM_DT,                     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_SSM_OUT,                    {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_W1,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_W2,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_DECAY_W1,          {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_DECAY_W2,          {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_KEY,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_VALUE,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_RECEPTANCE,        {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_GATE,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_TIME_MIX_OUTPUT,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_CHANNEL_MIX_KEY,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_CHANNEL_MIX_RECEPTANCE,     {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_CHANNEL_MIX_VALUE,          {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT}},
-    {LLM_TENSOR_FFN_ACT,                    {LLM_TENSOR_LAYER_REPEATING, GGML_OP_DIV}},
-    {LLM_TENSOR_SSM_CONV1D,                 {LLM_TENSOR_LAYER_REPEATING, GGML_OP_SSM_CONV}},
-    {LLM_TENSOR_SSM_A,                      {LLM_TENSOR_LAYER_REPEATING, GGML_OP_SSM_SCAN}},
-    {LLM_TENSOR_SSM_D,                      {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_TIME_MIX_LERP_X,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_TIME_MIX_LN,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_CHANNEL_MIX_LERP_K,         {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_CHANNEL_MIX_LERP_R,         {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_TIME_MIX_LERP_W,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_LERP_K,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_LERP_V,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_LERP_R,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_LERP_G,            {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_DECAY,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_ADD}},
-    {LLM_TENSOR_TIME_MIX_FIRST,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_RWKV_WKV6}},
-    {LLM_TENSOR_ATTN_NORM,                  {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_NORM_2,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_OUT_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_POST_NORM,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_FFN_NORM,                   {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_FFN_POST_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_FFN_NORM_EXPS,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_Q_NORM,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_K_NORM,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_LAYER_OUT_NORM,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_Q_A_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_KV_A_NORM,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ATTN_SUB_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_FFN_SUB_NORM,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_DEC_ATTN_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_DEC_CROSS_ATTN_NORM,        {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_DEC_FFN_NORM,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ENC_ATTN_NORM,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_ENC_FFN_NORM,               {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
-    {LLM_TENSOR_DEC_ATTN_REL_B,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_ENC_ATTN_REL_B,             {LLM_TENSOR_LAYER_REPEATING, GGML_OP_GET_ROWS}},
-    {LLM_TENSOR_FFN_DOWN_EXPS,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}},
-    {LLM_TENSOR_FFN_GATE_EXPS,              {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}},
-    {LLM_TENSOR_FFN_UP_EXPS,                {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL_MAT_ID}},
-    // this tensor is loaded for T5, but never used
-    {LLM_TENSOR_DEC_CROSS_ATTN_REL_B,       {LLM_TENSOR_LAYER_REPEATING, GGML_OP_NONE}},
-};
-
-// checks if the weight tensor can be used with the specified buffer type and device
-static bool weight_buft_supported(const llama_hparams & hparams, ggml_tensor * w, ggml_op op, ggml_backend_buffer_type_t buft, ggml_backend_dev_t dev) {
-    GGML_ASSERT(w != nullptr);
-
-    if (op == GGML_OP_NONE) {
-        return true;
-    }
-
-    ggml_init_params params = {
-        /*.mem_size   =*/ ggml_tensor_overhead()*8,
-        /*.mem_buffer =*/ NULL,
-        /*.no_alloc   =*/ true,
-    };
-    ggml_context_ptr ctx_ptr { ggml_init(params) };
-    if (!ctx_ptr) {
-        throw std::runtime_error(format("failed to create ggml context"));
-    }
-    ggml_context * ctx = ctx_ptr.get();
-
-    ggml_tensor * op_tensor = nullptr;
-
-    switch (op) {
-        case GGML_OP_GET_ROWS:
-            {
-                ggml_tensor * b = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 512);
-                op_tensor = ggml_get_rows(ctx, w, b);
-            } break;
-        case GGML_OP_MUL_MAT:
-            {
-                ggml_tensor * b = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], 512, w->ne[2], w->ne[3]);
-                op_tensor = ggml_mul_mat(ctx, w, b);
-            } break;
-        case GGML_OP_MUL_MAT_ID:
-            {
-                int n_expert_used = hparams.n_expert_used;
-                ggml_tensor * b = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, w->ne[0], n_expert_used, 512);
-                ggml_tensor * ids = ggml_new_tensor_2d(ctx, GGML_TYPE_I32, n_expert_used, 512);
-                op_tensor = ggml_mul_mat_id(ctx, w, b, ids);
-            } break;
-        case GGML_OP_ADD:
-            {
-                ggml_tensor * a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, w->ne[0], 512);
-                op_tensor = ggml_add(ctx, a, w);
-            } break;
-        case GGML_OP_MUL:
-            {
-                ggml_tensor * a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, w->ne[0], 512);
-                op_tensor = ggml_mul(ctx, a, w);
-            } break;
-        case GGML_OP_DIV:
-            {
-                ggml_tensor * a = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, w->ne[0]);
-                op_tensor = ggml_div(ctx, a, w);
-            } break;
-        case GGML_OP_ROPE:
-            {
-                int n_embd_head = hparams.n_embd_head_v;
-                int n_head = hparams.n_head();
-                ggml_tensor * a = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, n_embd_head, n_head, 512);
-                ggml_tensor * b = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, 512);
-                op_tensor = ggml_rope_ext(
-                    ctx, a, b, w,
-                    0, 0, 0, 0, 0,
-                    0, 0, 0, 0
-                );
-
-            } break;
-        case GGML_OP_SSM_CONV:
-            {
-                // FIXME
-                ggml_tensor * conv_x = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, 12345, w->ne[1], 6789);
-                op_tensor = ggml_ssm_conv(ctx, conv_x, w);
-            } break;
-        case GGML_OP_SSM_SCAN:
-            {
-                // FIXME
-                const int64_t d_state      = w->ne[0];
-                const int64_t d_inner      = w->ne[1];
-                const int64_t n_seq_tokens = 512;
-                const int64_t n_seqs       = 1;
-                ggml_tensor * s  = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, d_inner, n_seqs);
-                ggml_tensor * x = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_inner, n_seq_tokens, n_seqs);
-                ggml_tensor * dt = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_inner, n_seq_tokens, n_seqs);
-                ggml_tensor * B = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, n_seq_tokens, n_seqs);
-                ggml_tensor * C = ggml_new_tensor_3d(ctx, GGML_TYPE_F32, d_state, n_seq_tokens, n_seqs);
-                op_tensor = ggml_ssm_scan(ctx, s, x, dt, w, B, C);
-            } break;
-        case GGML_OP_RWKV_WKV6:
-            {
-                // FIXME
-                const int64_t S = 123;
-                const int64_t H = 123;
-                const int64_t n_tokens = 123;
-                const int64_t n_seqs = 123;
-                ggml_tensor  * k = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, S, 1, H, n_tokens);
-                ggml_tensor  * v = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, 1, S, H, n_tokens);
-                ggml_tensor  * r = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, 1, S, H, n_tokens);
-                ggml_tensor  * tf = w;
-                ggml_tensor  * td = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, 1, S, H, n_tokens);
-                ggml_tensor  * state = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, S, n_seqs, S, H);
-                op_tensor = ggml_rwkv_wkv6(ctx, k, v, r, tf, td, state);
-            } break;
-        default:
-            GGML_ABORT("%s: missing test for op %s for tensor %s", __func__, ggml_op_name(op), w->name);
-    }
-
-    // create a temporary dummy buffer for the weight so that supports_op can check the buffer type
-    GGML_ASSERT(w->buffer == nullptr);
-    w->buffer = ggml_backend_buft_alloc_buffer(buft, 0);
-    bool op_supported = ggml_backend_dev_supports_op(dev, op_tensor);
-    ggml_backend_buffer_free(w->buffer);
-    w->buffer = nullptr;
-
-    return op_supported;
-}
-
-// find the first buffer type in the list that can use the tensor
-static ggml_backend_buffer_type_t select_weight_buft(const llama_model & model, ggml_tensor * tensor, ggml_op op, const llama_model::buft_list_t & buft_list) {
-    GGML_ASSERT(!buft_list.empty());
-    for (const auto & cur : buft_list) {
-        ggml_backend_dev_t cur_dev = cur.first;
-        ggml_backend_buffer_type_t cur_buft = cur.second;
-        if (weight_buft_supported(model.hparams, tensor, op, cur_buft, cur_dev)) {
-            return cur_buft;
-        }
-    }
-    return nullptr;
-}
-
-// CPU: ACCEL -> CPU extra -> GPU host -> CPU
-static llama_model::buft_list_t make_cpu_buft_list(llama_model & model) {
-    llama_model::buft_list_t buft_list;
-
-    // add ACCEL buffer types
-    for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
-        ggml_backend_dev_t dev = ggml_backend_dev_get(i);
-        if (ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_ACCEL) {
-            auto * buft = ggml_backend_dev_buffer_type(dev);
-            // skip
-            if (buft != ggml_backend_cpu_buffer_type()) {
-                buft_list.emplace_back(dev, buft);
-            }
-        }
-    }
-
-    // add extra buffer types
-    auto * cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
-    auto * cpu_reg = ggml_backend_dev_backend_reg(cpu_dev);
-    auto ggml_backend_dev_get_extra_bufts_fn = (ggml_backend_dev_get_extra_bufts_t)
-        ggml_backend_reg_get_proc_address(cpu_reg, "ggml_backend_cpu_get_extra_bufts");
-    if (ggml_backend_dev_get_extra_bufts_fn) {
-        ggml_backend_buffer_type_t * extra_bufts = ggml_backend_dev_get_extra_bufts_fn(cpu_dev);
-        while (extra_bufts && *extra_bufts) {
-            buft_list.emplace_back(cpu_dev, *extra_bufts);
-            ++extra_bufts;
-        }
-    }
-
-    // add a host buffer type
-    // storing the tensors in a host buffer is useful when the processing of large batches
-    // is offloaded to a GPU device, since it reduces the time spent on data transfers
-    // generally, this will be done using the first device in the list
-    // a better approach would be to handle this on a weight-by-weight basis using the offload_op
-    // function of the device to determine if it would benefit from being stored in a host buffer
-    for (auto * dev : model.devices) {
-        ggml_backend_buffer_type_t buft = ggml_backend_dev_host_buffer_type(dev);
-        if (buft) {
-            buft_list.emplace_back(dev, buft);
-            break;
-        }
-    }
-
-    // add the CPU buffer type
-    for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
-        ggml_backend_dev_t dev = ggml_backend_dev_get(i);
-        if (ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_CPU) {
-            buft_list.emplace_back(dev, ggml_backend_dev_buffer_type(dev));
-        }
-    }
-
-    return buft_list;
-}
-
-// GPU: split if LLAMA_SPLIT_MODE_ROW -> GPU
-static llama_model::buft_list_t make_gpu_buft_list(ggml_backend_dev_t dev, enum llama_split_mode split_mode, const float * tensor_split) {
-    llama_model::buft_list_t buft_list;
-
-    // add the device split buffer type if requested and available
-    if (split_mode == LLAMA_SPLIT_MODE_ROW) {
-        ggml_backend_reg_t reg = ggml_backend_dev_backend_reg(dev);
-        auto ggml_backend_split_buffer_type_fn = (ggml_backend_split_buffer_type_t)
-            ggml_backend_reg_get_proc_address(reg, "ggml_backend_split_buffer_type");
-        if (ggml_backend_split_buffer_type_fn) {
-            size_t dev_index = [&]() {
-                auto * reg = ggml_backend_dev_backend_reg(dev);
-                for (size_t i = 0; i < ggml_backend_reg_dev_count(reg); ++i) {
-                    if (ggml_backend_reg_dev_get(reg, i) == dev) {
-                        return i;
-                    }
-                }
-                throw std::runtime_error(format("device %s not found in its backend reg", ggml_backend_dev_name(dev)));
-            }();
-            auto * buft = ggml_backend_split_buffer_type_fn(dev_index, tensor_split);
-            if (buft != nullptr) {
-                buft_list.emplace_back(dev, buft);
-            }
-        }
-    }
-
-    // add the device default buffer type
-    buft_list.emplace_back(dev, ggml_backend_dev_buffer_type(dev));
-
-    return buft_list;
-}
-
-// Returns false if cancelled by progress_callback
-static bool llm_load_tensors(
-        llama_model_loader & ml,
-        llama_model & model,
-        int n_gpu_layers,
-        enum llama_split_mode split_mode,
-        int main_gpu,
-        const float * tensor_split,
-        bool use_mlock,
-        llama_progress_callback progress_callback,
-        void * progress_callback_user_data) {
-    auto & hparams = model.hparams;
-
-    model.split_mode   = split_mode;
-    model.main_gpu     = main_gpu;
-    model.n_gpu_layers = n_gpu_layers;
-
-    const int n_layer     = hparams.n_layer;
-    bool use_mmap_buffer = true;
-
-    // build a list of buffer types for the CPU and GPU devices
-    model.cpu_buft_list = make_cpu_buft_list(model);
-    for (auto * dev : model.devices) {
-        llama_model::buft_list_t buft_list = make_gpu_buft_list(dev, split_mode, tensor_split);
-        // add CPU buffer types as a fallback
-        buft_list.insert(buft_list.end(), model.cpu_buft_list.begin(), model.cpu_buft_list.end());
-        model.gpu_buft_list.emplace(dev, std::move(buft_list));
-    }
-
-    // calculate the split points
-    int device_count = llama_get_device_count(model);
-    bool all_zero = tensor_split == nullptr || std::all_of(tensor_split, tensor_split + device_count, [](float x) { return x == 0.0f; });
-    std::vector splits(device_count);
-    if (all_zero) {
-        // default split, by free memory
-        for (int i = 0; i < device_count; ++i) {
-            ggml_backend_dev_t dev = model.devices[i];
-            size_t total;
-            size_t free;
-            ggml_backend_dev_memory(dev, &free, &total);
-            splits[i] = free;
-        }
-    } else {
-        std::copy(tensor_split, tensor_split + device_count, splits.begin());
-    }
-
-    // sum and normalize the splits to get the split points
-    float split_sum = 0.0f;
-    for (int i = 0; i < device_count; ++i) {
-        split_sum += splits[i];
-        splits[i] = split_sum;
-    }
-    for (int i = 0; i < device_count; ++i) {
-        splits[i] /= split_sum;
-    }
-
-    ggml_backend_dev_t cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
-    const int i_gpu_start = std::max((int) hparams.n_layer - n_gpu_layers, (int) 0);
-    const int act_gpu_layers = model.devices.empty() ? 0 : std::min(n_gpu_layers, (int)n_layer + 1);
-    auto get_layer_buft_list = [&](int il) -> llama_model::layer_dev {
-        if (il < i_gpu_start || (il - i_gpu_start) >= act_gpu_layers) {
-            return {cpu_dev, &model.cpu_buft_list};
-        }
-        int layer_gpu = std::upper_bound(splits.begin(), splits.begin() + device_count, float(il - i_gpu_start)/act_gpu_layers) - splits.begin();
-        auto * dev = model.devices.at(layer_gpu);
-        return {dev, &model.gpu_buft_list.at(dev)};
-    };
-
-    // assign the input layer
-    // there is very little benefit to offloading the input layer, so always keep it on the CPU
-    model.dev_input = { cpu_dev, &model.cpu_buft_list };
-
-    // assign the repeating layers to the devices according to the splits
-    model.dev_layer.resize(n_layer);
-    for (int il = 0; il < n_layer; ++il) {
-        model.dev_layer[il] = get_layer_buft_list(il);
-    }
-    // assign the output layer
-    model.dev_output = get_layer_buft_list(n_layer);
-
-    // one ggml context per buffer type
-    int max_n_tensors = ml.n_tensors;
-    max_n_tensors += 1;         // duplicated output tensor
-    max_n_tensors += n_layer*2; // duplicated rope freq tensors
-    const size_t ctx_size = ggml_tensor_overhead()*max_n_tensors;
-
-    std::map ctx_map;
-    auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * {
-        auto it = ctx_map.find(buft);
-        if (it == ctx_map.end()) {
-            ggml_init_params params = {
-                /*.mem_size   =*/ ctx_size,
-                /*.mem_buffer =*/ NULL,
-                /*.no_alloc   =*/ true,
-            };
-            ggml_context * ctx = ggml_init(params);
-            if (!ctx) {
-                throw std::runtime_error(format("failed to create ggml context"));
-            }
-            ctx_map[buft] = ctx;
-            model.ctxs.emplace_back(ctx);
-            return ctx;
-        }
-        return it->second;
-    };
-
-    // create tensors for the weights
-    {
-        // note: cast to int64_t since we will use these for the tensor dimensions
-        const int64_t n_head        = hparams.n_head();
-        const int64_t n_head_kv     = hparams.n_head_kv();
-        const int64_t n_embd        = hparams.n_embd;
-        const int64_t n_embd_k_gqa  = hparams.n_embd_k_gqa();
-        const int64_t n_embd_v_gqa  = hparams.n_embd_v_gqa();
-        const int64_t n_embd_head_k = hparams.n_embd_head_k;
-        const int64_t n_embd_head_v = hparams.n_embd_head_v;
-        const int64_t n_ff          = hparams.n_ff();
-        const int64_t n_embd_gqa    = n_embd_v_gqa;
-        const int64_t n_vocab       = hparams.n_vocab;
-        const int64_t n_vocab_type  = hparams.n_vocab_type;
-        const int64_t n_rot         = hparams.n_rot;
-        const int64_t n_expert      = hparams.n_expert;
-        const int64_t n_expert_used = hparams.n_expert_used;
-        const int64_t n_ctx_train   = hparams.n_ctx_train;
-
-        if (n_expert > 0 && hparams.n_expert_used == 0) {
-            throw std::runtime_error("model has expert layers but no expert layers are used");
-        }
-
-        int n_moved_tensors = 0;
-        ggml_tensor * first_moved_tensor = nullptr;
-        ggml_backend_buffer_type_t first_moved_from_buft = nullptr;
-        ggml_backend_buffer_type_t first_moved_to_buft = nullptr;
-
-        auto create_tensor = [&](const LLM_TN_IMPL & tn, const std::initializer_list & ne, int flags) -> ggml_tensor * {
-            ggml_tensor * t_meta = ml.get_tensor_meta(tn.str().c_str());
-
-            if (!t_meta) {
-                if (flags & llama_model_loader::TENSOR_NOT_REQUIRED) {
-                    return nullptr;
-                }
-                throw std::runtime_error(format("missing tensor '%s'", tn.str().c_str()));
-            }
-
-            // some models use the token embedding tensor as the output, but since these are used in different layers and with different ops
-            // the tensor is duplicated
-            // to handle this, we check if the tensor is duplicated, and if so, we assume that it is being loaded as the output tensor
-            llm_tensor tn_tensor = tn.tensor;
-            if (tn.tensor == LLM_TENSOR_TOKEN_EMBD && flags & llama_model_loader::TENSOR_DUPLICATED) {
-                tn_tensor = LLM_TENSOR_OUTPUT;
-            }
-
-            auto it = llm_tensor_info_mapping.find(tn_tensor);
-            if (it == llm_tensor_info_mapping.end()) {
-                throw std::runtime_error(format("missing tensor info mapping for %s", tn.str().c_str()));
-            }
-            const auto & info = it->second;
-
-            // tensors with "bias" suffix are always used with GGML_OP_ADD
-            ggml_op op;
-            bool bias = tn.suffix != nullptr && strcmp(tn.suffix, "bias") == 0;
-            if (bias) {
-                op = GGML_OP_ADD;
-            } else {
-                op = info.op;
-            }
-
-            // sanity checks
-            if (info.layer == LLM_TENSOR_LAYER_INPUT || info.layer == LLM_TENSOR_LAYER_OUTPUT) {
-                if (tn.bid != -1) {
-                    GGML_ABORT("input/output layer tensor %s used with a layer number", tn.str().c_str());
-                }
-            } else {
-                if (tn.bid == -1) {
-                    GGML_ABORT("repeating layer tensor %s used without a layer number", tn.str().c_str());
-                }
-            }
-
-            // select the buffer type for this tensor
-            llama_model::buft_list_t * buft_list;
-            switch (info.layer) {
-                case LLM_TENSOR_LAYER_INPUT:
-                    buft_list = model.dev_input.buft_list;
-                    break;
-                case LLM_TENSOR_LAYER_OUTPUT:
-                    buft_list = model.dev_output.buft_list;
-                    break;
-                case LLM_TENSOR_LAYER_REPEATING:
-                    buft_list = model.dev_layer.at(tn.bid).buft_list;
-                    break;
-                default:
-                    GGML_ABORT("invalid layer %d for tensor %s", info.layer, tn.str().c_str());
-            }
-
-            ggml_backend_buffer_type_t buft = select_weight_buft(model, t_meta, op, *buft_list);
-            if (!buft) {
-                throw std::runtime_error(format("failed to find a compatible buffer type for tensor %s", tn.str().c_str()));
-            }
-
-            // avoid using a host buffer when using mmap
-            auto * buft_dev = ggml_backend_buft_get_device(buft);
-            if (ml.use_mmap && buft == ggml_backend_dev_host_buffer_type(buft_dev)) {
-                auto * cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
-                buft = ggml_backend_dev_buffer_type(cpu_dev);
-            }
-
-            if (buft != buft_list->front().second) {
-                n_moved_tensors++;
-                if (!first_moved_tensor) {
-                    first_moved_tensor = t_meta;
-                    first_moved_from_buft = buft_list->front().second;
-                    first_moved_to_buft   = buft;
-                }
-            }
-
-            ggml_context * ctx = ctx_for_buft(buft);
-
-            // if duplicated, check if the original tensor was allocated in the same buffer type context and avoid creating a new one
-            if (flags & llama_model_loader::TENSOR_DUPLICATED) {
-                ggml_tensor * t = ggml_get_tensor(ctx, tn.str().c_str());
-                if (t) {
-                    return t;
-                }
-            }
-            return ml.create_tensor(ctx, tn, ne, flags);
-        };
-
-        model.layers.resize(n_layer);
-
-        // TODO: move to a separate function
-        const auto tn = LLM_TN(model.arch);
-        switch (model.arch) {
-            case LLM_ARCH_LLAMA:
-            case LLM_ARCH_REFACT:
-            case LLM_ARCH_MINICPM:
-            case LLM_ARCH_GRANITE:
-            case LLM_ARCH_GRANITE_MOE:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd_head_k * n_head}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
-
-                        // optional bias tensors
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd},     llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd},     llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-
-                        if (n_expert == 0) {
-                            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                            layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-
-                            // optional MLP bias
-                            layer.ffn_gate_b = create_tensor(tn(LLM_TENSOR_FFN_GATE, "bias", i), {n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                            layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                            layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i), {n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        } else {
-                            layer.ffn_gate_inp  = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP,  "weight", i), {n_embd, n_expert}, 0);
-                            layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd,   n_ff, n_expert}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                            layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {  n_ff, n_embd, n_expert}, 0);
-                            layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd,   n_ff, n_expert}, 0);
-                        }
-                    }
-                } break;
-            case LLM_ARCH_MINICPM3:
-                {
-                    const int64_t n_embd_head_qk_rope = hparams.n_rot;
-                    const int64_t n_embd_head_qk_nope = hparams.n_embd_head_k - hparams.n_rot;
-
-                    const int64_t q_lora_rank  = hparams.n_lora_q;
-                    const int64_t kv_lora_rank = hparams.n_lora_kv;
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0);
-
-                        layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, 0);
-
-                        layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0);
-                        layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k}, 0);
-
-                        layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + (n_embd_head_qk_rope)}, 0);
-                        layer.wkv_b     = create_tensor(tn(LLM_TENSOR_ATTN_KV_B,     "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)}, 0);
-                        layer.wo        = create_tensor(tn(LLM_TENSOR_ATTN_OUT,      "weight", i), {              n_head * (                      n_embd_head_v), n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-
-                        layer.rope_long  = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG,  "weight", i), { n_embd_head_qk_rope/2 }, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-                        layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), { n_embd_head_qk_rope/2 }, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-                    }
-                } break;
-            case LLM_ARCH_GROK:
-                {
-                    if (n_expert == 0) {
-                        throw std::runtime_error("Grok model cannot have zero experts");
-                    }
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.attn_out_norm   = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate_inp  = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP,  "weight", i), {n_embd, n_expert}, 0);
-                        layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {  n_ff, n_embd, n_expert}, 0);
-                        layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd,   n_ff, n_expert}, 0);
-
-                        layer.layer_out_norm   = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_DBRX:
-                {
-                    if (n_expert == 0) {
-                        throw std::runtime_error("DBRX model cannot have zero experts");
-                    }
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.attn_out_norm = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate_inp  = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP,  "weight", i), {n_embd, n_expert}, 0);
-                        layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff,   n_expert}, 0);
-                        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff,   n_embd, n_expert}, 0);
-                        layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd, n_ff,   n_expert}, 0);
-                    }
-                } break;
-            case LLM_ARCH_BAICHUAN:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-                    {
-                        model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                        model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_FALCON:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    {
-                        model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                        model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-
-                        model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        if (!model.output) {
-                            model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED); // needs to be on GPU
-                        }
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.attn_norm_2   = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_STARCODER:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-                    model.pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD,   "weight"), {n_embd, n_ctx_train}, 0);
-
-                    // output
-                    {
-                        model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                        model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                        model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        if (!model.output) {
-                            // needs to be on GPU
-                            model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                        }
-
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i),   {n_embd, n_ff}, 0);
-                        layer.ffn_up_b = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i),     {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_BERT:
-            case LLM_ARCH_NOMIC_BERT:
-                {
-                    model.tok_embd     = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,  "weight"), {n_embd, n_vocab}, 0);
-                    model.type_embd    = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_vocab_type}, 0);
-
-                    if (model.arch == LLM_ARCH_BERT) {
-                        model.pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD,    "weight"), {n_embd, n_ctx_train}, 0);
-
-                        model.cls   = create_tensor(tn(LLM_TENSOR_CLS, "weight"), {n_embd, n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        model.cls_b = create_tensor(tn(LLM_TENSOR_CLS, "bias"),   {n_embd},         llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        model.cls_out   = create_tensor(tn(LLM_TENSOR_CLS_OUT, "weight"), {n_embd, 1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        model.cls_out_b = create_tensor(tn(LLM_TENSOR_CLS_OUT, "bias"),   {1},         llama_model_loader::TENSOR_NOT_REQUIRED);
-                    }
-
-                    model.tok_norm   = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0);
-                    model.tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"),   {n_embd}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        if (model.arch == LLM_ARCH_BERT) {
-                            layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                            layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i),   {n_embd}, 0);
-
-                            layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                            layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i),   {n_embd_gqa}, 0);
-
-                            layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                            layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i),   {n_embd_gqa}, 0);
-                        } else {
-                            layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        }
-
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT,      "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.attn_out_norm   = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_out_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,        "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN,      "weight", i), {n_ff, n_embd}, 0);
-
-                        if (model.arch == LLM_ARCH_BERT) {
-                            layer.bo         = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0);
-                            layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i), {n_ff}, 0);
-                            layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0);
-                        } else {
-                            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
-                        }
-
-                        layer.layer_out_norm   = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0);
-                        layer.layer_out_norm_b = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "bias", i),   {n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_JINA_BERT_V2:
-                {
-                    model.tok_embd  = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,  "weight"), {n_embd, n_vocab}, 0); // word_embeddings
-                    model.type_embd = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_vocab_type}, 0); // token_type_embeddings
-
-                    model.tok_norm   = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0); // LayerNorm
-                    model.tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"),   {n_embd}, 0); //LayerNorm bias
-
-                    model.cls   = create_tensor(tn(LLM_TENSOR_CLS, "weight"), {n_embd, 1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    model.cls_b = create_tensor(tn(LLM_TENSOR_CLS, "bias"),   {1},         llama_model_loader::TENSOR_NOT_REQUIRED);
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i]; // JinaBertLayer
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i),   {n_embd}, 0);
-
-                        layer.attn_q_norm   = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias",   i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias",   i), {n_embd_gqa}, 0);
-
-                        layer.attn_k_norm   = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias",   i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias",   i), {n_embd_gqa}, 0);
-
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0); //output_dens
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias",   i), {n_embd}, 0); //output_dens
-
-                        layer.attn_out_norm   = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "weight", i), {n_embd}, 0); //output_norm
-                        layer.attn_out_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_OUT_NORM, "bias",   i), {n_embd}, 0);
-
-                        layer.attn_norm_2   = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias",   i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias",   i), {n_embd}, 0);
-
-                        layer.layer_out_norm   = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "weight", i), {n_embd}, 0);
-                        layer.layer_out_norm_b = create_tensor(tn(LLM_TENSOR_LAYER_OUT_NORM, "bias",   i), {n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_BLOOM:
-                {
-                    model.tok_embd   = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,      "weight"), {n_embd, n_vocab}, 0);
-                    model.tok_norm   = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0);
-                    model.tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"),   {n_embd}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias",   i), {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias",   i), {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias",   i), {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias",   i), {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias",   i), {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias",   i), {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_MPT:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-                    model.pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD,   "weight"), {n_embd, n_ctx_train}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    if (!model.output) {
-                        model.output    = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED); // needs to be on GPU
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i),   {n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.attn_q_norm   = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias",   i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.attn_k_norm   = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias",   i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        // AWQ ScaleActivation layer
-                        layer.ffn_act = create_tensor(tn(LLM_TENSOR_FFN_ACT, "scales", i), {n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    }
-                } break;
-            case LLM_ARCH_STABLELM:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm =   create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        // optional bias tensors, present in Stable LM 2 1.6B
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd},     llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        // optional q and k layernorms, present in StableLM 2 12B
-                        layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head},    llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        // optional FFN norm, not present in StableLM 2 12B which uses parallel residual
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_QWEN:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd*3}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd*3}, 0);
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff/2}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff/2, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff/2}, 0);
-                    }
-                } break;
-            case LLM_ARCH_QWEN2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        // optional bias tensors
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd}, 0);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, 0);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_QWEN2MOE:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        // optional bias tensors
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd}, 0);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, 0);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
-
-                        if (n_expert == 0) {
-                            throw std::runtime_error("n_expert must be > 0 for QWEN2MOE");
-                        }
-                        if (n_expert_used == 0) {
-                            throw std::runtime_error("n_expert_used must be > 0 for QWEN2MOE");
-                        }
-
-                        // MoE branch
-                        const int64_t n_ff_exp = hparams.n_ff_exp ? hparams.n_ff_exp : n_ff / n_expert_used;
-
-                        layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {  n_embd, n_ff_exp, n_expert}, 0);
-                        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp,   n_embd, n_expert}, 0);
-                        layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {  n_embd, n_ff_exp, n_expert}, 0);
-
-                        // Shared expert branch
-                        const int64_t n_ff_shexp = hparams.n_ff_shexp ? hparams.n_ff_shexp : n_ff;
-
-                        layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {    n_embd, n_ff_shexp}, 0);
-                        layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {n_ff_shexp,     n_embd}, 0);
-                        layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {    n_embd, n_ff_shexp}, 0);
-                    }
-                } break;
-            case LLM_ARCH_PHI2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-                    model.output_b      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "bias"),   {n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        if (layer.wqkv == nullptr) {
-                            layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd}, 0);
-                            layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i),   {n_embd}, 0);
-
-                            layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0);
-                            layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i),   {n_embd_gqa}, 0);
-
-                            layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0);
-                            layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i),   {n_embd_gqa}, 0);
-                        }
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i),   {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_PHI3:
-                {
-                    const int64_t n_embd_head = n_embd / n_head;
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), { n_embd }, 0);
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), { n_embd, n_vocab }, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), { n_embd }, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), { n_embd, n_embd + 2 * n_embd_gqa }, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), { n_embd, n_embd }, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), { n_embd }, 0);
-
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd }, 0);
-                        layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), { n_embd, 2 * n_ff }, 0);
-
-                        layer.rope_long  = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_LONG,  "weight", i), { n_embd_head/2 }, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-                        layer.rope_short = create_tensor(tn(LLM_TENSOR_ROPE_FACTORS_SHORT, "weight", i), { n_embd_head/2 }, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-                    }
-                } break;
-            case LLM_ARCH_PLAMO:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_GPT2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-                    model.pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD,   "weight"), {n_embd, n_ctx_train}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM,   "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM,   "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i),   {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_CODESHELL:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i),   {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP, "bias", i),     {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_ORION:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_INTERNLM2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        // layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_GEMMA:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,  "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED); // same as tok_embd, duplicated to allow offloading
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd_head_k * n_head}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_GEMMA2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,  "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED); // same as tok_embd, duplicated to allow offloading
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd_head_k * n_head}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
-                        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_post_norm = create_tensor(tn(LLM_TENSOR_FFN_POST_NORM, "weight", i), {n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_STARCODER2:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        // optional bias tensors
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd}, 0);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, 0);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, 0);
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-
-                        // optional bias tensors
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP ,  "bias", i), {  n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_MAMBA:
-                {
-                    const int64_t d_conv  = hparams.ssm_d_conv;
-                    const int64_t d_inner = hparams.ssm_d_inner;
-                    const int64_t d_state = hparams.ssm_d_state;
-                    const int64_t dt_rank = hparams.ssm_dt_rank;
-
-                    // only an expansion factor of 2 is supported for now
-                    if (2 * n_embd != d_inner) {
-                        throw std::runtime_error("only an expansion factor of 2 is supported for now");
-                    }
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed, duplicated to allow offloading
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        // norm
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ssm_in = create_tensor(tn(LLM_TENSOR_SSM_IN, "weight", i), {n_embd, 2*d_inner}, 0);
-
-                        layer.ssm_conv1d = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "weight", i), {d_conv, d_inner}, 0);
-                        layer.ssm_conv1d_b = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "bias", i), {d_inner}, 0);
-
-                        layer.ssm_x = create_tensor(tn(LLM_TENSOR_SSM_X, "weight", i), {d_inner, dt_rank + 2*d_state}, 0);
-
-                        layer.ssm_dt = create_tensor(tn(LLM_TENSOR_SSM_DT, "weight", i), {dt_rank, d_inner}, 0);
-                        layer.ssm_dt_b = create_tensor(tn(LLM_TENSOR_SSM_DT, "bias", i), {d_inner}, 0);
-
-                        // no "weight" suffix for these
-                        layer.ssm_a = create_tensor(tn(LLM_TENSOR_SSM_A, i), {d_state, d_inner}, 0);
-                        layer.ssm_d = create_tensor(tn(LLM_TENSOR_SSM_D, i), {d_inner}, 0);
-
-                        // out_proj
-                        layer.ssm_out = create_tensor(tn(LLM_TENSOR_SSM_OUT, "weight", i), {d_inner, n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_XVERSE:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_COMMAND_R:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    // init output from the input tok embed
-                    model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        if (n_layer >= 64){
-                            layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head}, 0);
-                            layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, 0);
-                        }
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_OLMO:  // adapted from LLM_ARCH_LLAMA with norm params removed
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_OLMOE:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
-
-                        if (n_expert == 0) {
-                            throw std::runtime_error("n_expert must be > 0");
-                        }
-                        if (n_expert_used == 0) {
-                            throw std::runtime_error("n_expert_used must be > 0");
-                        }
-
-                        // MoE branch
-                        layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff,   n_expert}, 0);
-                        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff,   n_embd, n_expert}, 0);
-                        layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd, n_ff,   n_expert}, 0);
-                    }
-                } break;
-            case LLM_ARCH_OPENELM:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    // init output from the input tok embed
-                    model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        const int64_t n_head      =   hparams.n_head(i);
-                        const int64_t n_head_qkv  = 2*hparams.n_head_kv(i) + n_head;
-                        const int64_t n_ff        =   hparams.n_ff(i);
-
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_head_qkv*n_embd_head_k}, 0);
-                        layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k}, 0);
-                        layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_head*n_embd_head_k, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_GPTNEOX:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i),   {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_ARCTIC:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_embd}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_embd, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
-                        layer.ffn_norm_exps = create_tensor(tn(LLM_TENSOR_FFN_NORM_EXPS, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd,   n_ff, n_expert}, false);
-                        layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {  n_ff, n_embd, n_expert}, 0);
-                        layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd,   n_ff, n_expert}, 0);
-                    }
-                } break;
-            case LLM_ARCH_DEEPSEEK2:
-                {
-                    const bool is_lite = (hparams.n_layer == 27);
-
-                    const int64_t n_embd_head_qk_rope = hparams.n_rot;
-                    const int64_t n_embd_head_qk_nope = hparams.n_embd_head_k - hparams.n_rot;
-
-                    const int64_t q_lora_rank  = hparams.n_lora_q;
-                    const int64_t kv_lora_rank = hparams.n_lora_kv;
-
-                    const int64_t n_ff_exp        = hparams.n_ff_exp;
-                    const int64_t n_expert_shared = hparams.n_expert_shared;
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        if (!is_lite) {
-                            layer.attn_q_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_A_NORM, "weight", i), {q_lora_rank}, 0);
-                        }
-
-                        layer.attn_kv_a_norm = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_NORM, "weight", i), {kv_lora_rank}, 0);
-
-                        if (!is_lite) {
-                            layer.wq_a = create_tensor(tn(LLM_TENSOR_ATTN_Q_A, "weight", i), {n_embd, q_lora_rank}, 0);
-                            layer.wq_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_B, "weight", i), {q_lora_rank, n_head * n_embd_head_k}, 0);
-                        } else {
-                            layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        }
-
-                        layer.wkv_a_mqa = create_tensor(tn(LLM_TENSOR_ATTN_KV_A_MQA, "weight", i), {n_embd, kv_lora_rank + (n_embd_head_qk_rope)}, 0);
-                        layer.wkv_b     = create_tensor(tn(LLM_TENSOR_ATTN_KV_B,     "weight", i), {kv_lora_rank, n_head * (n_embd_head_qk_nope + n_embd_head_v)}, 0);
-                        layer.wo        = create_tensor(tn(LLM_TENSOR_ATTN_OUT,      "weight", i), {              n_head * (                      n_embd_head_v), n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        if (i < (int) hparams.n_layer_dense_lead) {
-                            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                            layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                        } else {
-                            layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);
-
-                            if (n_expert == 0) {
-                                throw std::runtime_error("n_expert must be > 0");
-                            }
-                            if (n_expert_used == 0) {
-                                throw std::runtime_error("n_expert_used must be > 0");
-                            }
-
-                            // MoE branch
-                            layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {  n_embd, n_ff_exp, n_expert}, 0);
-                            layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp,   n_embd, n_expert}, 0);
-                            layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {  n_embd, n_ff_exp, n_expert}, 0);
-
-                            // Shared expert branch
-                            layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0);
-                            layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {        n_ff_exp * n_expert_shared, n_embd}, 0);
-                            layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {n_embd, n_ff_exp * n_expert_shared}, 0);
-                        }
-                    }
-                } break;
-            case LLM_ARCH_BITNET:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm     = create_tensor(tn(LLM_TENSOR_ATTN_NORM,     "weight", i), {n_embd}, 0);
-                        layer.attn_sub_norm = create_tensor(tn(LLM_TENSOR_ATTN_SUB_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq       = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wq_scale = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.wk       = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wk_scale = create_tensor(tn(LLM_TENSOR_ATTN_K,   "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.wv       = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv_scale = create_tensor(tn(LLM_TENSOR_ATTN_V,   "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.wo       = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.wo_scale = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_norm     = create_tensor(tn(LLM_TENSOR_FFN_NORM,     "weight", i), {n_embd}, 0);
-                        layer.ffn_sub_norm = create_tensor(tn(LLM_TENSOR_FFN_SUB_NORM, "weight", i), {n_ff}, 0);
-
-                        layer.ffn_gate       = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_gate_scale = create_tensor(tn(LLM_TENSOR_FFN_GATE, "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_down       = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_scale = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_up         = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_scale   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "scale",  i), {1}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    }
-                } break;
-            case LLM_ARCH_T5:
-                {
-                    const auto n_rel_attn_bkts = hparams.n_rel_attn_bkts;
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm     = create_tensor(tn(LLM_TENSOR_DEC_OUTPUT_NORM, "weight"), {n_embd}, 0);
-
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm_enc  = create_tensor(tn(LLM_TENSOR_ENC_ATTN_NORM,  "weight", i), {n_embd}, 0);
-                        layer.attn_rel_b_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wq_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_Q,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wk_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0);
-
-                        layer.ffn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_GATE, "weight", i), {n_embd,   n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_down_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up_enc   = create_tensor(tn(LLM_TENSOR_ENC_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-
-                        layer.attn_norm  = create_tensor(tn(LLM_TENSOR_DEC_ATTN_NORM,  "weight", i), {n_embd}, 0);
-                        layer.attn_rel_b = create_tensor(tn(LLM_TENSOR_DEC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_DEC_ATTN_Q,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_DEC_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_DEC_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_DEC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0);
-
-                        layer.attn_norm_cross  = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_NORM,  "weight", i), {n_embd}, 0);
-                        // this tensor seems to be unused in HF transformers implementation
-                        layer.attn_rel_b_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wq_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_Q,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wk_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo_cross = create_tensor(tn(LLM_TENSOR_DEC_CROSS_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_DEC_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_DEC_FFN_GATE, "weight", i), {n_embd,   n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_DEC_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_DEC_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_T5ENCODER:
-                {
-                    const auto n_rel_attn_bkts = hparams.n_rel_attn_bkts;
-
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm_enc  = create_tensor(tn(LLM_TENSOR_ENC_ATTN_NORM,  "weight", i), {n_embd}, 0);
-                        layer.attn_rel_b_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_REL_B, "weight", i), {n_head, n_rel_attn_bkts}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wq_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_Q,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wk_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo_enc = create_tensor(tn(LLM_TENSOR_ENC_ATTN_OUT, "weight", i), {n_embd_v_gqa, n_embd}, 0);
-
-                        layer.ffn_norm_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_gate_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_GATE, "weight", i), {n_embd,   n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_down_enc = create_tensor(tn(LLM_TENSOR_ENC_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up_enc   = create_tensor(tn(LLM_TENSOR_ENC_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_JAIS:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM,   "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM,   "bias", i),   {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i),   {n_embd}, 0);
-
-                        layer.ffn_gate   = create_tensor(tn(LLM_TENSOR_FFN_GATE,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_gate_b = create_tensor(tn(LLM_TENSOR_FFN_GATE,   "bias", i),   {n_ff}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i),   {n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_CHATGLM:
-                {
-                    model.tok_embd   = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,      "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), {n_embd, n_embd + 2*n_embd_gqa}, 0);
-                        layer.bqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "bias", i),   {n_embd + 2*n_embd_gqa}, 0);
-
-                        layer.wo   = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff * 2}, 0);
-
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);
-                    }
-                } break;
-            case LLM_ARCH_NEMOTRON:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm   = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0);
-                    model.output        = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        // optional bias tensors
-                        layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "bias", i), {n_embd},     llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "bias", i), {n_embd_gqa}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd},     llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-                        layer.ffn_norm_b = create_tensor(tn(LLM_TENSOR_FFN_NORM, "bias", i), {n_embd}, 0);
-
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-
-                        // optional MLP bias
-                        layer.ffn_down_b = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "bias", i), {n_embd}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.ffn_up_b   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "bias", i), {n_ff}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    }
-                } break;
-            case LLM_ARCH_EXAONE:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd_head_k * n_head}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_k_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_v_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
-
-                        layer.ffn_norm   = create_tensor(tn(LLM_TENSOR_FFN_NORM,   "weight", i), {n_embd}, 0);
-                        layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, llama_model_loader::TENSOR_NOT_REQUIRED | (i != 0 ? llama_model_loader::TENSOR_DUPLICATED : 0));
-                        layer.ffn_gate   = create_tensor(tn(LLM_TENSOR_FFN_GATE,   "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down   = create_tensor(tn(LLM_TENSOR_FFN_DOWN,   "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up     = create_tensor(tn(LLM_TENSOR_FFN_UP,     "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            case LLM_ARCH_RWKV6:
-                {
-                    model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                    // Block 0, LN0
-                    model.tok_norm = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "weight"), {n_embd}, 0);
-                    model.tok_norm_b = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD_NORM, "bias"), {n_embd}, 0);
-
-                    // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output_norm_b = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "bias"), {n_embd}, 0);
-                    model.output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, 0);
-
-                    const int time_mix_extra_dim = hparams.time_mix_extra_dim;
-                    const int time_decay_extra_dim = hparams.time_decay_extra_dim;
-                    const int head_size = hparams.wkv_head_size;
-                    const int attn_hidden_size = n_embd;
-                    const int ffn_size = hparams.n_ff_arr[0];
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm   = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "bias", i),   {n_embd}, 0);
-
-                        layer.attn_norm_2   = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "weight", i), {n_embd}, 0);
-                        layer.attn_norm_2_b = create_tensor(tn(LLM_TENSOR_ATTN_NORM_2, "bias", i),   {n_embd}, 0);
-
-                        layer.time_mix_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W1, "weight", i), {n_embd, time_mix_extra_dim * 5}, 0);
-                        layer.time_mix_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_W2, "weight", i), {time_mix_extra_dim, n_embd, 5}, 0);
-
-                        layer.time_mix_lerp_x = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_X, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.time_mix_lerp_w = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_W, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.time_mix_lerp_k = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_K, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.time_mix_lerp_v = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_V, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.time_mix_lerp_r = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_R, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.time_mix_lerp_g = create_tensor(tn(LLM_TENSOR_TIME_MIX_LERP_G, "weight", i), {n_embd, 1, 1}, 0);
-
-                        layer.time_mix_first = create_tensor(tn(LLM_TENSOR_TIME_MIX_FIRST, "weight", i), {head_size, n_embd / head_size}, 0);
-                        layer.time_mix_decay = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY, "weight", i), {n_embd}, 0);
-                        layer.time_mix_decay_w1 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W1, "weight", i), {n_embd, time_decay_extra_dim}, 0);
-                        layer.time_mix_decay_w2 = create_tensor(tn(LLM_TENSOR_TIME_MIX_DECAY_W2, "weight", i), {time_decay_extra_dim, attn_hidden_size}, 0);
-                        layer.time_mix_key = create_tensor(tn(LLM_TENSOR_TIME_MIX_KEY, "weight", i), {attn_hidden_size, n_embd}, 0);
-                        layer.time_mix_value = create_tensor(tn(LLM_TENSOR_TIME_MIX_VALUE, "weight", i), {attn_hidden_size, n_embd}, 0);
-                        layer.time_mix_receptance = create_tensor(tn(LLM_TENSOR_TIME_MIX_RECEPTANCE, "weight", i), {attn_hidden_size, n_embd}, 0);
-                        layer.time_mix_gate = create_tensor(tn(LLM_TENSOR_TIME_MIX_GATE, "weight", i), {attn_hidden_size, n_embd}, 0);
-
-                        layer.time_mix_ln = create_tensor(tn(LLM_TENSOR_TIME_MIX_LN, "weight", i), {n_embd}, 0);
-                        layer.time_mix_ln_b = create_tensor(tn(LLM_TENSOR_TIME_MIX_LN, "bias", i), {n_embd}, 0);
-                        layer.time_mix_output = create_tensor(tn(LLM_TENSOR_TIME_MIX_OUTPUT, "weight", i), {n_embd, attn_hidden_size}, 0);
-
-                        layer.channel_mix_lerp_k = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_LERP_K, "weight", i), {n_embd, 1, 1}, 0);
-                        layer.channel_mix_lerp_r = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_LERP_R, "weight", i), {n_embd, 1, 1}, 0);
-
-                        layer.channel_mix_key = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_KEY, "weight", i), {n_embd, ffn_size}, 0);
-                        layer.channel_mix_value = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_VALUE, "weight", i), {ffn_size, n_embd}, 0);
-                        layer.channel_mix_receptance = create_tensor(tn(LLM_TENSOR_CHANNEL_MIX_RECEPTANCE, "weight", i), {n_embd, n_embd}, 0);
-                    }
-
-                } break;
-            case LLM_ARCH_CHAMELEON:
-                {
-                 model.tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
-
-                 // output
-                    model.output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    model.output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                    // if output is NULL, init from the input tok embed
-                    if (model.output == NULL) {
-                        model.output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
-                    }
-
-                    for (int i = 0; i < n_layer; ++i) {
-                        auto & layer = model.layers[i];
-
-                        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
-                        layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head}, 0);
-                        layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv}, 0);
-                        layer.attn_q_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i),  {n_embd_head_k, n_head}, llama_model_loader::TENSOR_NOT_REQUIRED);
-                        layer.attn_k_norm_b = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "bias", i),  {n_embd_head_k, n_head_kv}, llama_model_loader::TENSOR_NOT_REQUIRED);
-
-                        layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q,   "weight", i), {n_embd, n_embd}, 0);
-                        layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V,   "weight", i), {n_embd, n_embd_gqa}, 0);
-                        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd}, 0);
-
-                        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
-
-                        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
-                        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
-                        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
-                    }
-                } break;
-            default:
-                throw std::runtime_error("unknown architecture");
-        }
-
-        if (n_moved_tensors > 0) {
-            LLAMA_LOG_DEBUG("%s: tensor '%s' (%s) (and %d others) cannot be used with preferred buffer type %s, using %s instead\n",
-                __func__, first_moved_tensor->name, ggml_type_name(first_moved_tensor->type), n_moved_tensors - 1,
-                ggml_backend_buft_name(first_moved_from_buft), ggml_backend_buft_name(first_moved_to_buft));
-        }
-    }
-
-    ml.done_getting_tensors();
-
-    ml.init_mappings(true, use_mlock ? &model.mlock_mmaps : nullptr);
-    model.mappings.reserve(ml.mappings.size());
-
-    // create the backend buffers
-    std::vector> ctx_bufs;
-    ctx_bufs.reserve(ctx_map.size());
-
-    // Ensure we have enough capacity for the maximum backend buffer we will potentially create
-    const size_t n_max_backend_buffer = ctx_map.size() * ml.files.size();
-    model.bufs.reserve(n_max_backend_buffer);
-
-    for (auto & it : ctx_map) {
-        ggml_backend_buffer_type_t buft = it.first;
-        ggml_context * ctx              = it.second;
-
-        // skip contexts without tensors
-        if (ggml_get_first_tensor(ctx) == nullptr) {
-            continue;
-        }
-
-        llama_buf_map bufs;
-        bufs.reserve(n_max_backend_buffer);
-
-        // check if it is possible to use buffer_from_host_ptr with this buffer type
-        ggml_backend_dev_t dev = ggml_backend_buft_get_device(buft);
-        ggml_backend_dev_props props;
-        ggml_backend_dev_get_props(dev, &props);
-        bool buffer_from_host_ptr_supported = props.caps.buffer_from_host_ptr;
-        bool is_default_buft = buft == ggml_backend_dev_buffer_type(dev);
-
-        if (ml.use_mmap && use_mmap_buffer && buffer_from_host_ptr_supported && is_default_buft) {
-            for (uint32_t idx = 0; idx < ml.files.size(); idx++) {
-                // only the mmap region containing the tensors in the model is mapped to the backend buffer
-                // this is important for metal with apple silicon: if the entire model could be mapped to a metal buffer, then we could just use metal for all layers
-                // this allows using partial offloading when the model size exceeds the metal buffer size, but not the RAM size
-                void * addr = nullptr;
-                size_t first, last; // NOLINT
-                ml.get_mapping_range(&first, &last, &addr, idx, ctx);
-                if (first >= last) {
-                    continue;
-                }
-                const size_t max_size = ggml_get_max_tensor_size(ctx);
-                ggml_backend_buffer_t buf = ggml_backend_dev_buffer_from_host_ptr(dev, (char *) addr + first, last - first, max_size);
-                if (buf == nullptr) {
-                    throw std::runtime_error(format("unable to allocate %s buffer", ggml_backend_buft_name(buft)));
-                }
-                model.bufs.emplace_back(buf);
-                bufs.emplace(idx, buf);
-            }
-        }
-        else {
-            ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft);
-            if (buf == nullptr) {
-                throw std::runtime_error(format("unable to allocate %s buffer", ggml_backend_buft_name(buft)));
-            }
-            model.bufs.emplace_back(buf);
-            if (use_mlock && ggml_backend_buffer_is_host(buf)) {
-                model.mlock_bufs.emplace_back(new llama_mlock);
-                auto & mlock_buf = model.mlock_bufs.back();
-                mlock_buf->init   (ggml_backend_buffer_get_base(buf));
-                mlock_buf->grow_to(ggml_backend_buffer_get_size(buf));
-            }
-            for (uint32_t idx = 0; idx < ml.files.size(); idx++) {
-                bufs.emplace(idx, buf);
-            }
-        }
-
-        if (bufs.empty()) {
-            throw std::runtime_error("failed to allocate buffer");
-        }
-
-        for (auto & buf : bufs) {
-            // indicate that this buffer contains weights
-            // this is used by ggml_backend_sched to improve op scheduling: ops that use a weight are preferably scheduled to the backend that contains the weight
-            ggml_backend_buffer_set_usage(buf.second, GGML_BACKEND_BUFFER_USAGE_WEIGHTS);
-        }
-
-        ctx_bufs.emplace_back(ctx, bufs);
-    }
-
-    if (llama_supports_gpu_offload()) {
-        const int n_gpu = std::min(n_gpu_layers, int(hparams.n_layer));
-
-        LLAMA_LOG_INFO("%s: offloading %d repeating layers to GPU\n", __func__, n_gpu);
-        if (n_gpu_layers > (int) hparams.n_layer) {
-            LLAMA_LOG_INFO("%s: offloading output layer to GPU\n", __func__);
-        }
-
-        const int max_backend_supported_layers = hparams.n_layer + 1;
-        const int max_offloadable_layers       = hparams.n_layer + 1;
-
-        LLAMA_LOG_INFO("%s: offloaded %d/%d layers to GPU\n", __func__, std::min(n_gpu_layers, max_offloadable_layers), max_backend_supported_layers);
-    }
-
-    // print memory requirements per buffer type
-    for (auto & buf : model.bufs) {
-        LLAMA_LOG_INFO("%s: %12s model buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get()) / 1024.0 / 1024.0);
-    }
-
-    // populate tensors_by_name
-    for (auto & ctx : model.ctxs) {
-        for (auto * cur = ggml_get_first_tensor(ctx.get()); cur != NULL; cur = ggml_get_next_tensor(ctx.get(), cur)) {
-            model.tensors_by_name.emplace_back(ggml_get_name(cur), cur);
-        }
-    }
-
-    // load tensor data
-    for (auto & it : ctx_bufs) {
-        ggml_context * ctx = it.first;
-        auto & bufs = it.second;
-        if (!ml.load_all_data(ctx, bufs, use_mlock ? &model.mlock_mmaps : NULL, progress_callback, progress_callback_user_data)) {
-            return false;
-        }
-    }
-
-    if (use_mmap_buffer) {
-        for (auto & mapping : ml.mappings) {
-            model.mappings.emplace_back(std::move(mapping));
-        }
-    }
-
-    return true;
-}
-
 // Returns 0 on success, -1 on error, and -2 on cancellation via llama_progress_callback
 static int llama_model_load(const std::string & fname, llama_model & model, llama_model_params & params) {
-    model.t_start_us = ggml_time_us();
+    // loading time will be recalculated after the first eval, so
+    // we take page faults deferred by mmap() into consideration
+    model.t_load_us = 0;
+    time_meas tm(model.t_load_us);
+
+    model.t_start_us = tm.t_start_us;
 
     try {
         llama_model_loader ml(fname, params.use_mmap, params.check_tensors, params.kv_overrides);
 
+        ml.print_info();
+
         model.hparams.vocab_only = params.vocab_only;
 
         try {
-            llm_load_arch(ml, model);
+            model.load_arch(ml);
         } catch(const std::exception & e) {
             throw std::runtime_error("error loading model architecture: " + std::string(e.what()));
         }
         try {
-            llm_load_hparams(ml, model);
+            model.load_hparams(ml);
         } catch(const std::exception & e) {
             throw std::runtime_error("error loading model hyperparameters: " + std::string(e.what()));
         }
         try {
-            llm_load_vocab(ml, model);
+            model.load_vocab(ml);
         } catch(const std::exception & e) {
             throw std::runtime_error("error loading model vocabulary: " + std::string(e.what()));
         }
 
-        llm_load_print_meta(ml, model);
-
-        if (model.vocab.type != LLAMA_VOCAB_TYPE_NONE &&
-            model.hparams.n_vocab != model.vocab.id_to_token.size()) {
-            throw std::runtime_error("vocab size mismatch");
-        }
+        model.load_stats(ml);
+        model.print_info();
 
         if (params.vocab_only) {
             LLAMA_LOG_INFO("%s: vocab only - skipping tensors\n", __func__);
             return 0;
         }
 
-        if (!llm_load_tensors(
-            ml, model, params.n_gpu_layers, params.split_mode,  params.main_gpu, params.tensor_split, params.use_mlock,
-            params.progress_callback, params.progress_callback_user_data
-        )) {
+        if (!model.load_tensors(ml)) {
             return -2;
         }
     } catch (const std::exception & err) {
@@ -9210,10 +78,6 @@ static int llama_model_load(const std::string & fname, llama_model & model, llam
         return -1;
     }
 
-    // loading time will be recalculate after the first eval, so
-    // we take page faults deferred by mmap() into consideration
-    model.t_load_us = ggml_time_us() - model.t_start_us;
-
     return 0;
 }
 
@@ -9239,27 +103,43 @@ enum llm_ffn_gate_type {
 enum llm_norm_type {
     LLM_NORM,
     LLM_NORM_RMS,
+    LLM_NORM_GROUP,
 };
 
 static struct ggml_tensor * llm_build_inp_embd(
         struct ggml_context * ctx,
        struct llama_context & lctx,
         const llama_hparams & hparams,
-         const llama_ubatch & batch,
+         const llama_ubatch & ubatch,
          struct ggml_tensor * tok_embd,
          const llm_build_cb & cb) {
     const int64_t n_embd = hparams.n_embd;
 
     struct ggml_tensor * inpL;
 
-    if (batch.token) {
-        lctx.inp_tokens = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, batch.n_tokens);
+    if (ubatch.token) {
+        lctx.inp_tokens = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ubatch.n_tokens);
         cb(lctx.inp_tokens, "inp_tokens", -1);
         ggml_set_input(lctx.inp_tokens);
 
         inpL = ggml_get_rows(ctx, tok_embd, lctx.inp_tokens);
+
+        // apply lora for embedding tokens if needed
+        for (auto & it : lctx.lora) {
+            struct llama_adapter_lora_weight * lw = it.first->get_weight(tok_embd);
+            if (lw == nullptr) {
+                continue;
+            }
+            const float adapter_scale = it.second;
+            const float scale = lw->get_scale(it.first->alpha, adapter_scale);
+            struct ggml_tensor * inpL_delta = ggml_scale(ctx, ggml_mul_mat(
+                ctx, lw->b, // non-transposed lora_b
+                ggml_get_rows(ctx, lw->a, lctx.inp_tokens)
+            ), scale);
+            inpL = ggml_add(ctx, inpL, inpL_delta);
+        }
     } else {
-       lctx.inp_embd = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, n_embd, batch.n_tokens);
+        lctx.inp_embd = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, n_embd, ubatch.n_tokens);
         inpL = lctx.inp_embd;
         ggml_set_input(lctx.inp_embd);
     }
@@ -9325,17 +205,16 @@ static struct ggml_tensor * llm_build_lora_mm(
           struct ggml_tensor * w,
           struct ggml_tensor * cur) {
     struct ggml_tensor * res = ggml_mul_mat(ctx0, w, cur);
-    for (auto & it : lctx.lora_adapters) {
-        struct llama_lora_weight * lora = it.first->get_weight(w);
-        if (lora == nullptr) {
+    for (auto & it : lctx.lora) {
+        struct llama_adapter_lora_weight * lw = it.first->get_weight(w);
+        if (lw == nullptr) {
             continue;
         }
-        const float alpha = it.first->alpha;
-        const float rank  = (float) lora->b->ne[0];
-        const float scale = alpha ? it.second * alpha / rank : it.second;
+        const float adapter_scale = it.second;
+        const float scale = lw->get_scale(it.first->alpha, adapter_scale);
         struct ggml_tensor * ab_cur = ggml_mul_mat(
-            ctx0, lora->b,
-            ggml_mul_mat(ctx0, lora->a, cur)
+            ctx0, lw->b,
+            ggml_mul_mat(ctx0, lw->a, cur)
         );
         ab_cur = ggml_scale(ctx0, ab_cur, scale);
         res = ggml_add(ctx0, res, ab_cur);
@@ -9351,17 +230,17 @@ static struct ggml_tensor * llm_build_lora_mm_id(
           struct ggml_tensor * cur, // struct ggml_tensor * b
           struct ggml_tensor * ids) {
     struct ggml_tensor * res = ggml_mul_mat_id(ctx0, w, cur, ids);
-    for (auto & it : lctx.lora_adapters) {
-        struct llama_lora_weight * lora = it.first->get_weight(w);
-        if (lora == nullptr) {
+    for (auto & it : lctx.lora) {
+        struct llama_adapter_lora_weight * lw = it.first->get_weight(w);
+        if (lw == nullptr) {
             continue;
         }
         const float alpha = it.first->alpha;
-        const float rank  = (float) lora->b->ne[0];
+        const float rank  = (float) lw->b->ne[0];
         const float scale = alpha ? it.second * alpha / rank : it.second;
         struct ggml_tensor * ab_cur = ggml_mul_mat_id(
-            ctx0, lora->b,
-            ggml_mul_mat_id(ctx0, lora->a, cur, ids),
+            ctx0, lw->b,
+            ggml_mul_mat_id(ctx0, lw->a, cur, ids),
             ids
         );
         ab_cur = ggml_scale(ctx0, ab_cur, scale);
@@ -9380,8 +259,14 @@ static struct ggml_tensor * llm_build_norm(
          const llm_build_cb & cb,
                         int   il) {
     switch (type) {
-        case LLM_NORM:     cur = ggml_norm    (ctx, cur, hparams.f_norm_eps);     break;
-        case LLM_NORM_RMS: cur = ggml_rms_norm(ctx, cur, hparams.f_norm_rms_eps); break;
+        case LLM_NORM:       cur = ggml_norm      (ctx, cur, hparams.f_norm_eps);     break;
+        case LLM_NORM_RMS:   cur = ggml_rms_norm  (ctx, cur, hparams.f_norm_rms_eps); break;
+        case LLM_NORM_GROUP:
+            {
+                cur = ggml_reshape_3d(ctx, cur, cur->ne[0], 1, cur->ne[1]);
+                cur = ggml_group_norm(ctx, cur, hparams.n_norm_groups, hparams.f_norm_group_eps);
+                cur = ggml_reshape_2d(ctx, cur, cur->ne[0],    cur->ne[2]);
+            } break;
     }
 
     if (mw || mb) {
@@ -9537,12 +422,14 @@ static struct ggml_tensor * llm_build_moe_ffn(
          struct ggml_tensor * up_exps,
          struct ggml_tensor * gate_exps,
          struct ggml_tensor * down_exps,
+         struct ggml_tensor * exp_probs_b,
                     int64_t   n_expert,
                     int64_t   n_expert_used,
             llm_ffn_op_type   type_op,
                        bool   norm_w,
                        bool   scale_w,
                       float   w_scale,
+llama_expert_gating_func_type gating_op,
          const llm_build_cb & cb,
                         int   il) {
     int64_t n_embd = cur->ne[0];
@@ -9551,11 +438,31 @@ static struct ggml_tensor * llm_build_moe_ffn(
     ggml_tensor * logits = llm_build_lora_mm(lctx, ctx, gate_inp, cur); // [n_expert, n_tokens]
     cb(logits, "ffn_moe_logits", il);
 
-    ggml_tensor * probs = ggml_soft_max(ctx, logits); // [n_expert, n_tokens]
+    ggml_tensor * probs = nullptr;
+    switch (gating_op) {
+        case LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX:
+            {
+                probs = ggml_soft_max(ctx, logits); // [n_expert, n_tokens]
+            } break;
+        case LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID:
+            {
+                probs = ggml_sigmoid(ctx, logits); // [n_expert, n_tokens]
+            } break;
+        default:
+            GGML_ABORT("fatal error");
+    }
     cb(probs, "ffn_moe_probs", il);
 
+    // add experts selection bias - introduced in DeepSeek V3
+    // leave probs unbiased as it's later used to get expert weights
+    ggml_tensor * selection_probs = probs;
+    if (exp_probs_b != nullptr) {
+        selection_probs = ggml_add(ctx, probs, exp_probs_b);
+        cb(selection_probs, "ffn_moe_probs_biased", il);
+    }
+
     // select experts
-    ggml_tensor * selected_experts = ggml_top_k(ctx, probs, n_expert_used); // [n_expert_used, n_tokens]
+    ggml_tensor * selected_experts = ggml_top_k(ctx, selection_probs, n_expert_used); // [n_expert_used, n_tokens]
     cb(selected_experts->src[0], "ffn_moe_argsort", il);
     cb(selected_experts, "ffn_moe_topk", il);
 
@@ -9826,7 +733,7 @@ static struct ggml_tensor * llm_build_copy_mask_state(
 static struct ggml_tensor * llm_build_mamba(
         struct ggml_context * ctx,
        struct llama_context & lctx,
-         const llama_ubatch & batch,
+         const llama_ubatch & ubatch,
          struct ggml_cgraph * graph,
          struct ggml_tensor * cur,
          struct ggml_tensor * state_copy,
@@ -9842,17 +749,17 @@ static struct ggml_tensor * llm_build_mamba(
     const int64_t d_inner = hparams.ssm_d_inner;
     const int64_t d_state = hparams.ssm_d_state;
     const int64_t dt_rank = hparams.ssm_dt_rank;
-    const int64_t n_seqs  = batch.n_seqs;
+    const int64_t n_seqs  = ubatch.n_seqs;
     // Some variants of Mamba arch (e.g. FalconMamba do apply layer norm on B and Dt layers)
     const bool ssm_dt_b_c_rms = hparams.ssm_dt_b_c_rms;
     // Use the same RMS norm as the final layer norm
     const float norm_rms_eps = hparams.f_norm_rms_eps;
 
-    const int64_t n_seq_tokens = batch.n_seq_tokens;
+    const int64_t n_seq_tokens = ubatch.n_seq_tokens;
 
     GGML_ASSERT(n_seqs != 0);
-    GGML_ASSERT(batch.equal_seqs);
-    GGML_ASSERT(batch.n_tokens == n_seq_tokens * n_seqs);
+    GGML_ASSERT(ubatch.equal_seqs);
+    GGML_ASSERT(ubatch.n_tokens == n_seq_tokens * n_seqs);
 
     struct ggml_tensor * conv_states_all = kv.k_l[il];
     struct ggml_tensor * ssm_states_all  = kv.v_l[il];
@@ -9964,16 +871,20 @@ static struct ggml_tensor * llm_build_rwkv6_time_mix(
         const struct llama_layer * layer,
         struct ggml_tensor * cur,
         struct ggml_tensor * x_prev,
-        struct ggml_tensor ** wkv_state) {
+        struct ggml_tensor ** wkv_state,
+        size_t wkv_head_size,
+        size_t head_count_kv) {
     size_t n_embd       = cur->ne[0];
     size_t n_seq_tokens = cur->ne[1];
     size_t n_seqs       = cur->ne[2];
 
-    size_t head_size  = layer->time_mix_first->ne[0];
-    size_t head_count = layer->time_mix_first->ne[1];
+    size_t head_size  = wkv_head_size;
+    size_t head_count = n_embd / head_size;
 
     size_t n_tokens = n_seqs * n_seq_tokens;
 
+    bool is_qrwkv = layer->time_mix_first == nullptr;
+
     struct ggml_tensor * sx = ggml_sub(ctx, x_prev, cur);
 
     sx  = ggml_reshape_2d(ctx, sx,  n_embd, n_tokens);
@@ -10002,69 +913,64 @@ static struct ggml_tensor * llm_build_rwkv6_time_mix(
         xxx
     );
 
-    struct ggml_tensor *mw = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], 0);
-    struct ggml_tensor *mk = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * sizeof(float));
-    struct ggml_tensor *mv = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 2 * sizeof(float));
-    struct ggml_tensor *mr = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 3 * sizeof(float));
-    struct ggml_tensor *mg = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 4 * sizeof(float));
+    struct ggml_tensor *xw, *xk, *xv, *xr, *xg;
+    if (layer->time_mix_lerp_fused) {
+        // fusing these weights makes some performance improvement
+        sx  = ggml_reshape_3d(ctx, sx,  n_embd, 1, n_tokens);
+        cur = ggml_reshape_3d(ctx, cur, n_embd, 1, n_tokens);
+        xxx = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xxx, layer->time_mix_lerp_fused), sx), cur);
+        xw = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], 0);
+        xk = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * sizeof(float));
+        xv = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 2 * sizeof(float));
+        xr = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 3 * sizeof(float));
+        xg = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 4 * sizeof(float));
+    } else {
+        // for backward compatibility
+        xw = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], 0);
+        xk = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * sizeof(float));
+        xv = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 2 * sizeof(float));
+        xr = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 3 * sizeof(float));
+        xg = ggml_view_2d(ctx, xxx, n_embd, n_tokens, xxx->nb[1], n_embd * n_tokens * 4 * sizeof(float));
 
-    struct ggml_tensor * xw = ggml_add(
-        ctx,
-        ggml_mul(
-            ctx,
-            ggml_add(ctx, mw, layer->time_mix_lerp_w),
-            sx
-        ),
-        cur
-    );
+        xw = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xw, layer->time_mix_lerp_w), sx), cur);
+        xk = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xk, layer->time_mix_lerp_k), sx), cur);
+        xv = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xv, layer->time_mix_lerp_v), sx), cur);
+        xr = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xr, layer->time_mix_lerp_r), sx), cur);
+        xg = ggml_add(ctx, ggml_mul(ctx, ggml_add(ctx, xg, layer->time_mix_lerp_g), sx), cur);
+    }
 
-    struct ggml_tensor * xk = ggml_add(
-        ctx,
-        ggml_mul(
-            ctx,
-            ggml_add(ctx, mk, layer->time_mix_lerp_k),
-            sx
-        ),
-        cur
-    );
+    struct ggml_tensor * r = llm_build_lora_mm(lctx, ctx, layer->time_mix_receptance, xr);
+    struct ggml_tensor * k = llm_build_lora_mm(lctx, ctx, layer->time_mix_key,        xk);
+    struct ggml_tensor * v = llm_build_lora_mm(lctx, ctx, layer->time_mix_value,      xv);
+    if (layer->time_mix_receptance_b) {
+        r = ggml_add(ctx, r, layer->time_mix_receptance_b);
+    }
+    if (layer->time_mix_key_b) {
+        k = ggml_add(ctx, k, layer->time_mix_key_b);
+    }
+    if (layer->time_mix_value_b) {
+        v = ggml_add(ctx, v, layer->time_mix_value_b);
+    }
 
-    struct ggml_tensor * xv = ggml_add(
-        ctx,
-        ggml_mul(
-            ctx,
-            ggml_add(ctx, mv, layer->time_mix_lerp_v),
-            sx
-        ),
-        cur
-    );
+    struct ggml_tensor * g = llm_build_lora_mm(lctx, ctx, layer->time_mix_gate, xg);
+    if (is_qrwkv) {
+        g = ggml_sigmoid(ctx, g);
+    } else {
+        g = ggml_silu(ctx, g);
+    }
 
-    struct ggml_tensor * xr = ggml_add(
-        ctx,
-        ggml_mul(
-            ctx,
-            ggml_add(ctx, mr, layer->time_mix_lerp_r),
-            sx
-        ),
-        cur
-    );
+    if (head_count_kv != head_count) {
+        GGML_ASSERT(head_count % head_count_kv == 0);
+        k = ggml_reshape_4d(ctx, k, head_size, 1, head_count_kv, n_tokens);
+        v = ggml_reshape_4d(ctx, v, head_size, 1, head_count_kv, n_tokens);
+        struct ggml_tensor * tmp = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, head_size, head_count / head_count_kv, head_count_kv, n_tokens);
+        k = ggml_repeat(ctx, k, tmp);
+        v = ggml_repeat(ctx, v, tmp);
+    }
 
-    struct ggml_tensor * xg = ggml_add(
-        ctx,
-        ggml_mul(
-            ctx,
-            ggml_add(ctx, mg, layer->time_mix_lerp_g),
-            sx
-        ),
-        cur
-    );
-
-    struct ggml_tensor * r = ggml_reshape_4d(ctx, llm_build_lora_mm(lctx, ctx, layer->time_mix_receptance, xr), head_size, 1,         head_count, n_tokens);
-    struct ggml_tensor * k = ggml_reshape_4d(ctx, llm_build_lora_mm(lctx, ctx, layer->time_mix_key,        xk), 1,         head_size, head_count, n_tokens);
-    struct ggml_tensor * v = ggml_reshape_4d(ctx, llm_build_lora_mm(lctx, ctx, layer->time_mix_value,      xv), head_size, 1,         head_count, n_tokens);
-    struct ggml_tensor * g = ggml_silu(
-        ctx,
-        llm_build_lora_mm(lctx, ctx, layer->time_mix_gate, xg)
-    );
+    k = ggml_reshape_3d(ctx, k, head_size, head_count, n_tokens);
+    v = ggml_reshape_3d(ctx, v, head_size, head_count, n_tokens);
+    r = ggml_reshape_3d(ctx, r, head_size, head_count, n_tokens);
 
     struct ggml_tensor * w = ggml_mul_mat(
         ctx,
@@ -10075,25 +981,35 @@ static struct ggml_tensor * llm_build_rwkv6_time_mix(
         )
     );
 
-    w = ggml_add(ctx, w, ggml_reshape_1d(ctx, layer->time_mix_decay, n_embd));
+    w = ggml_add(ctx, w, layer->time_mix_decay);
     w = ggml_exp(ctx, ggml_neg(ctx, ggml_exp(ctx, w)));
-    w = ggml_reshape_4d(ctx, w, 1, head_size, head_count, n_tokens);
+    w = ggml_reshape_3d(ctx, w, head_size, head_count, n_tokens);
 
-    k = ggml_transpose(ctx, k);
-    v = ggml_transpose(ctx, v);
-    r = ggml_transpose(ctx, r);
+    if (is_qrwkv) {
+        // k = k * (1 - w)
+        k = ggml_sub(ctx, k, ggml_mul(ctx, k, w));
+    }
 
-    struct ggml_tensor * wkv_output = ggml_rwkv_wkv6(ctx, k, v, r, layer->time_mix_first, w, *wkv_state);
+    struct ggml_tensor * wkv_output;
+    if (!layer->time_mix_first) {
+        wkv_output = ggml_gated_linear_attn(ctx, k, v, r, w, *wkv_state, pow(head_size, -0.5f));
+    } else {
+        wkv_output = ggml_rwkv_wkv6(ctx, k, v, r, layer->time_mix_first, w, *wkv_state);
+    }
     cur = ggml_view_1d(ctx, wkv_output, n_embd * n_tokens, 0);
     *wkv_state = ggml_view_1d(ctx, wkv_output, n_embd * head_size * n_seqs, n_embd * n_tokens * sizeof(float));
 
-    // group norm with head_count groups
-    cur = ggml_reshape_3d(ctx, cur, n_embd / head_count, head_count, n_tokens);
-    cur = ggml_norm(ctx, cur, 64e-5f);
+    if (!is_qrwkv) {
+        // group norm with head_count groups
+        cur = ggml_reshape_3d(ctx, cur, n_embd / head_count, head_count, n_tokens);
+        cur = ggml_norm(ctx, cur, 64e-5f);
 
-    // Convert back to regular vectors.
-    cur = ggml_reshape_2d(ctx, cur, n_embd, n_tokens);
-    cur = ggml_add(ctx, ggml_mul(ctx, cur, layer->time_mix_ln), layer->time_mix_ln_b);
+        // Convert back to regular vectors.
+        cur = ggml_reshape_2d(ctx, cur, n_embd, n_tokens);
+        cur = ggml_add(ctx, ggml_mul(ctx, cur, layer->time_mix_ln), layer->time_mix_ln_b);
+    } else {
+        cur = ggml_reshape_2d(ctx, cur, n_embd, n_tokens);
+    }
 
     cur = ggml_mul(ctx, cur, g);
     cur = llm_build_lora_mm(lctx, ctx, layer->time_mix_output, cur);
@@ -10249,7 +1165,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_k_shift() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         GGML_ASSERT(kv_self.size == n_ctx);
 
@@ -10299,7 +1215,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_defrag(const std::vector & ids) {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         for (uint32_t i = 0; i < ids.size(); ++i) {
             const uint32_t id = ids[i];
@@ -10558,7 +1474,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_llama() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -10652,6 +1568,7 @@ struct llm_build_context {
 
             // feed-forward network
             if (model.layers[il].ffn_gate_inp == nullptr) {
+
                 cur = llm_build_norm(ctx0, ffn_inp, hparams,
                         model.layers[il].ffn_norm, NULL,
                         LLM_NORM_RMS, cb, il);
@@ -10676,9 +1593,11 @@ struct llm_build_context {
                         model.layers[il].ffn_up_exps,
                         model.layers[il].ffn_gate_exps,
                         model.layers[il].ffn_down_exps,
+                        nullptr,
                         n_expert, n_expert_used,
                         LLM_FFN_SILU, true,
                         false, 0.0,
+                        LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                         cb, il);
                 cb(cur, "ffn_moe_out", il);
             }
@@ -10720,8 +1639,11 @@ struct llm_build_context {
         return gf;
     }
 
-    struct ggml_cgraph * build_baichuan() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+    struct ggml_cgraph * build_deci() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        // mutable variable, needed during the last layer of the computation to skip unused tokens
+        int32_t n_tokens = this->n_tokens;
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -10733,7 +1655,165 @@ struct llm_build_context {
         inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
 
         // inp_pos - contains the positions
-        struct ggml_tensor * inp_pos = model.type == MODEL_7B ? build_inp_pos() : nullptr;
+        struct ggml_tensor * inp_pos = build_inp_pos();
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
+
+        const float kq_scale = hparams.f_attention_scale == 0.0f ? 1.0f/sqrtf(float(n_embd_head)) : hparams.f_attention_scale;
+        for (int il = 0; il < n_layer; ++il) {
+            struct ggml_tensor * inpSA = inpL;
+            const int64_t n_head_kv = hparams.n_head_kv(il);
+            const int64_t n_head    = hparams.n_head(il);
+
+            if (n_head == 0) {
+                // attention-free layer of Llama-3_1-Nemotron-51B
+                cur = inpL;
+            } else {
+                // norm
+                cur = llm_build_norm(ctx0, inpL, hparams,
+                        model.layers[il].attn_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(cur, "attn_norm", il);
+            }
+
+            if (n_head > 0 && n_head_kv == 0) {
+                // "linear attention" of Llama-3_1-Nemotron-51B
+                cur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wo, cur);
+                cb(cur, "wo", il);
+            } else if (n_head > 0) {
+                // self-attention
+                // rope freq factors for llama3; may return nullptr for llama2 and other models
+                struct ggml_tensor * rope_factors = build_rope_factors(il);
+
+                // compute Q and K and RoPE them
+                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+                if (model.layers[il].bq) {
+                    Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
+                    cb(Qcur, "Qcur", il);
+                }
+
+                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+                if (model.layers[il].bk) {
+                    Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
+                    cb(Kcur, "Kcur", il);
+                }
+
+                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+                if (model.layers[il].bv) {
+                    Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
+                    cb(Vcur, "Vcur", il);
+                }
+
+                Qcur = ggml_rope_ext(
+                    ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, rope_factors,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Qcur, "Qcur", il);
+
+                Kcur = ggml_rope_ext(
+                    ctx0, ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, rope_factors,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Kcur, "Kcur", il);
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
+                        model.layers[il].wo, model.layers[il].bo,
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, kq_scale, cb, il);
+            }
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+                n_tokens = n_outputs;
+                cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);
+                inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+            }
+
+            // For Granite architecture
+            if (hparams.f_residual_scale) {
+                cur = ggml_scale(ctx0, cur, hparams.f_residual_scale);
+            }
+
+            // modified to support attention-free layer of Llama-3_1-Nemotron-51B
+            struct ggml_tensor * ffn_inp = cur;
+            if (n_head > 0) {
+                ffn_inp = ggml_add(ctx0, cur, inpSA);
+                cb(ffn_inp, "ffn_inp", il);
+            }
+
+            // feed-forward network
+            if (model.layers[il].ffn_gate_inp == nullptr) {
+                cur = llm_build_norm(ctx0, ffn_inp, hparams,
+                        model.layers[il].ffn_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(cur, "ffn_norm", il);
+
+                cur = llm_build_ffn(ctx0, lctx, cur,
+                        model.layers[il].ffn_up,   model.layers[il].ffn_up_b,   NULL,
+                        model.layers[il].ffn_gate, model.layers[il].ffn_gate_b, NULL,
+                        model.layers[il].ffn_down, model.layers[il].ffn_down_b, NULL,
+                        NULL,
+                        LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+                cb(cur, "ffn_out", il);
+            }
+
+            // For Granite architecture
+            if (hparams.f_residual_scale) {
+                cur = ggml_scale(ctx0, cur, hparams.f_residual_scale);
+            }
+
+            cur = ggml_add(ctx0, cur, ffn_inp);
+            cb(cur, "ffn_out", il);
+
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+
+        // For Granite architecture
+        if (hparams.f_logit_scale) {
+            cur = ggml_scale(ctx0, cur, 1.0f / hparams.f_logit_scale);
+        }
+
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
+    struct ggml_cgraph * build_baichuan() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        struct ggml_tensor * inp_pos = model.type == LLM_TYPE_7B ? build_inp_pos() : nullptr;
 
         // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
         struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
@@ -10758,7 +1838,7 @@ struct llm_build_context {
                 cb(Vcur, "Vcur", il);
 
                 switch (model.type) {
-                    case MODEL_7B:
+                    case LLM_TYPE_7B:
                         Qcur = ggml_rope_ext(
                             ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, nullptr,
                             n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
@@ -10770,7 +1850,7 @@ struct llm_build_context {
                             ext_factor, attn_factor, beta_fast, beta_slow
                         );
                         break;
-                    case MODEL_13B:
+                    case LLM_TYPE_13B:
                         Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd/n_head, n_head, n_tokens);
                         Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd/n_head, n_head, n_tokens);
                         break;
@@ -10836,7 +1916,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_xverse() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -10939,7 +2019,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_falcon() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -11059,7 +2139,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_grok() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -11167,9 +2247,11 @@ struct llm_build_context {
                     model.layers[il].ffn_up_exps,
                     model.layers[il].ffn_gate_exps,
                     model.layers[il].ffn_down_exps,
+                    nullptr,
                     n_expert, n_expert_used,
                     LLM_FFN_GELU, true,
                     false, 0.0,
+                    LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                     cb, il);
             cb(cur, "ffn_moe_out", il);
 
@@ -11216,7 +2298,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_dbrx() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -11308,9 +2390,11 @@ struct llm_build_context {
                     model.layers[il].ffn_up_exps,
                     model.layers[il].ffn_gate_exps,
                     model.layers[il].ffn_down_exps,
+                    nullptr,
                     n_expert, n_expert_used,
                     LLM_FFN_SILU, true,
                     false, 0.0,
+                    LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                     cb, il);
             cb(cur, "ffn_moe_out", il);
 
@@ -11342,7 +2426,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_starcoder() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -11446,7 +2530,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_refact() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -11540,7 +2624,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_bert() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -11734,7 +2818,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_bloom() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -11835,7 +2919,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_mpt() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -12125,7 +3209,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_qwen() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -12237,7 +3321,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_qwen2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -12348,8 +3432,126 @@ struct llm_build_context {
         return gf;
     }
 
+    struct ggml_cgraph * build_qwen2vl() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        lctx.inp_pos = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, n_tokens * 4);
+        cb(lctx.inp_pos, "inp_pos", -1);
+        ggml_set_input(lctx.inp_pos);
+        struct ggml_tensor * inp_pos = lctx.inp_pos;
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
+        int sections[4];
+        std::copy(std::begin(hparams.rope_sections), std::begin(hparams.rope_sections) + 4, sections);
+
+        for (int il = 0; il < n_layer; ++il) {
+            struct ggml_tensor * inpSA = inpL;
+
+            // norm
+            cur = llm_build_norm(ctx0, inpL, hparams,
+                    model.layers[il].attn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "attn_norm", il);
+
+            // self-attention
+            {
+                // compute Q and K and RoPE them
+                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+                Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
+                cb(Qcur, "Qcur", il);
+
+                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+                Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
+                cb(Kcur, "Kcur", il);
+
+                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+                Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
+                cb(Vcur, "Vcur", il);
+
+                Qcur = ggml_rope_multi(
+                    ctx0,
+                    ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, nullptr,
+                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Qcur, "Qcur", il);
+
+                Kcur = ggml_rope_multi(
+                    ctx0,
+                    ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, nullptr,
+                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Kcur, "Kcur", il);
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
+                        model.layers[il].wo, model.layers[il].bo,
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
+            }
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+                cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);
+                inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+            }
+
+            struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+            cb(ffn_inp, "ffn_inp", il);
+
+            // feed-forward network
+            cur = llm_build_norm(ctx0, ffn_inp, hparams,
+                    model.layers[il].ffn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "ffn_norm", il);
+
+            cur = llm_build_ffn(ctx0, lctx, cur,
+                    model.layers[il].ffn_up,   NULL, NULL,
+                    model.layers[il].ffn_gate, NULL, NULL,
+                    model.layers[il].ffn_down, NULL, NULL,
+                    NULL,
+                    LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+            cb(cur, "ffn_out", il);
+
+            cur = ggml_add(ctx0, cur, ffn_inp);
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
     struct ggml_cgraph * build_qwen2moe() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -12438,9 +3640,11 @@ struct llm_build_context {
                         model.layers[il].ffn_up_exps,
                         model.layers[il].ffn_gate_exps,
                         model.layers[il].ffn_down_exps,
+                        nullptr,
                         n_expert, n_expert_used,
                         LLM_FFN_SILU, false,
                         false, 0.0,
+                        LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                         cb, il);
             cb(cur, "ffn_moe_out", il);
 
@@ -12495,7 +3699,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_phi2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -12616,7 +3820,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_phi3() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa = hparams.n_embd_v_gqa();
@@ -12631,7 +3835,13 @@ struct llm_build_context {
         struct ggml_tensor * inp_pos = build_inp_pos();
 
         // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
-        struct ggml_tensor * KQ_mask_swa = build_inp_KQ_mask_swa();
+        struct ggml_tensor * KQ_mask = nullptr;
+        if (hparams.n_swa == 0) {
+            // Phi-4 doesn't use sliding window attention
+            KQ_mask = build_inp_KQ_mask();
+        } else {
+            KQ_mask = build_inp_KQ_mask_swa();
+        }
 
         for (int il = 0; il < n_layer; ++il) {
             auto residual = inpL;
@@ -12643,7 +3853,7 @@ struct llm_build_context {
 
                 struct ggml_tensor* attn_norm_output = llm_build_norm(ctx0, inpL, hparams,
                     model.layers[il].attn_norm,
-                    NULL,
+                    model.layers[il].attn_norm_b,
                     LLM_NORM_RMS, cb, il);
                 cb(attn_norm_output, "attn_norm", il);
 
@@ -12658,8 +3868,7 @@ struct llm_build_context {
                     Qcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd,     n_tokens, cur->nb[1], 0 * sizeof(float) * (n_embd)));
                     Kcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1 * sizeof(float) * (n_embd)));
                     Vcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1 * sizeof(float) * (n_embd + n_embd_gqa)));
-                }
-                else {
+                } else {
                     Qcur = ggml_add(ctx0, llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, attn_norm_output), model.layers[il].bq);
                     Kcur = ggml_add(ctx0, llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, attn_norm_output), model.layers[il].bk);
                     Vcur = ggml_add(ctx0, llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, attn_norm_output), model.layers[il].bv);
@@ -12689,7 +3898,7 @@ struct llm_build_context {
 
                 cur = llm_build_kv(ctx0, lctx, kv_self, gf,
                         model.layers[il].wo, model.layers[il].bo,
-                        Kcur, Vcur, Qcur, KQ_mask_swa, n_tokens, kv_head, n_kv, 1.0f, cb, il);
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f, cb, il);
             }
 
             if (il == n_layer - 1) {
@@ -12703,14 +3912,12 @@ struct llm_build_context {
             residual = cur;
 
             cur = llm_build_norm(ctx0, cur, hparams,
-                model.layers[il].ffn_norm, NULL,
+                model.layers[il].ffn_norm, model.layers[il].ffn_norm_b,
                 LLM_NORM_RMS, cb, il);
             cb(cur, "ffn_norm", il);
 
-            // FF
-            // special-case: the up and gate tensors are merged into a single tensor
-            // TOOD: support into llm_build_ffn
-            {
+            // feed-forward network
+            if (model.layers[il].ffn_gate_inp == nullptr) {
                 cur = llm_build_ffn(ctx0, lctx, cur,
                         model.layers[il].ffn_up,   NULL, NULL,
                         NULL,                      NULL, NULL,
@@ -12718,6 +3925,20 @@ struct llm_build_context {
                         NULL,
                         LLM_FFN_SWIGLU, LLM_FFN_SEQ, cb, il);
                 cb(cur, "ffn_out", il);
+            } else {
+                // MoE branch
+                cur = llm_build_moe_ffn(ctx0, lctx, cur,
+                        model.layers[il].ffn_gate_inp,
+                        model.layers[il].ffn_up_exps,
+                        model.layers[il].ffn_gate_exps,
+                        model.layers[il].ffn_down_exps,
+                        nullptr,
+                        n_expert, n_expert_used,
+                        LLM_FFN_SILU, true,
+                        false, 0.0,
+                        LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
+                        cb, il);
+                cb(cur, "ffn_moe_out", il);
             }
 
             cur = ggml_add(ctx0, residual, cur);
@@ -12730,11 +3951,16 @@ struct llm_build_context {
 
         cur = llm_build_norm(ctx0, inpL, hparams,
             model.output_norm,
-            NULL,
+            model.output_norm_b,
             LLM_NORM_RMS, cb, -1);
         cb(cur, "result_norm", -1);
 
         cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+
+        if (model.output_b != nullptr) {
+            cb(cur, "result_output_no_bias", -1);
+            cur = ggml_add(ctx0, cur, model.output_b);
+        }
         cb(cur, "result_output", -1);
 
         ggml_build_forward_expand(gf, cur);
@@ -12848,7 +4074,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_gpt2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -12953,7 +4179,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_codeshell() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -13064,7 +4290,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_orion() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -13182,7 +4408,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_internlm2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -13299,155 +4525,8 @@ struct llm_build_context {
         return gf;
     }
 
-    // ref: https://arxiv.org/abs/2203.03466
-    //      https://github.com/ggerganov/llama.cpp/issues/5276#issuecomment-1925774738
-    // based on the original build_llama() function
-    struct ggml_cgraph * build_minicpm() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
-
-        const int64_t n_embd_head = hparams.n_embd_head_v;
-        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
-        GGML_ASSERT(n_embd_head == hparams.n_rot);
-
-        const int64_t n_embd = hparams.n_embd;
-        //TODO: if the model varies, these parameters need to be read from the model
-        const int64_t n_embd_base = 256;
-        const float scale_embd  = 12.0f;
-        const float scale_depth = 1.4f;
-
-        struct ggml_tensor * cur;
-        struct ggml_tensor * inpL;
-
-        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
-
-        // scale the input embeddings
-        inpL = ggml_scale(ctx0, inpL, scale_embd);
-        cb(inpL, "inp_scaled", -1);
-
-        // inp_pos - contains the positions
-        struct ggml_tensor * inp_pos = build_inp_pos();
-
-        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
-        struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
-
-        for (int il = 0; il < n_layer; ++il) {
-            struct ggml_tensor * inpSA = inpL;
-
-            // norm
-            cur = llm_build_norm(ctx0, inpL, hparams,
-                    model.layers[il].attn_norm, NULL,
-                    LLM_NORM_RMS, cb, il);
-            cb(cur, "attn_norm", il);
-
-            // self-attention
-            {
-                // compute Q and K and RoPE them
-                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
-                cb(Qcur, "Qcur", il);
-                if (model.layers[il].bq) {
-                    Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
-                    cb(Qcur, "Qcur", il);
-                }
-
-                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
-                cb(Kcur, "Kcur", il);
-                if (model.layers[il].bk) {
-                    Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
-                    cb(Kcur, "Kcur", il);
-                }
-
-                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
-                cb(Vcur, "Vcur", il);
-                if (model.layers[il].bv) {
-                    Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
-                    cb(Vcur, "Vcur", il);
-                }
-
-                Qcur = ggml_rope_ext(
-                    ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head,    n_tokens), inp_pos, nullptr,
-                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
-                    ext_factor, attn_factor, beta_fast, beta_slow
-                );
-                cb(Qcur, "Qcur", il);
-
-                Kcur = ggml_rope_ext(
-                    ctx0, ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, nullptr,
-                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
-                    ext_factor, attn_factor, beta_fast, beta_slow
-                );
-                cb(Kcur, "Kcur", il);
-
-                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
-                        model.layers[il].wo, model.layers[il].bo,
-                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
-            }
-
-            if (il == n_layer - 1) {
-                // skip computing output for unused tokens
-                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
-                cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);
-                inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
-            }
-
-            // scale_res - scale the hidden states for residual connection
-            const float scale_res = scale_depth/sqrtf(float(n_layer));
-            cur = ggml_scale(ctx0, cur, scale_res);
-            cb(cur, "hidden_scaled", -1);
-
-            struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
-            cb(ffn_inp, "ffn_inp", il);
-
-            // feed-forward network
-            {
-                cur = llm_build_norm(ctx0, ffn_inp, hparams,
-                        model.layers[il].ffn_norm, NULL,
-                        LLM_NORM_RMS, cb, il);
-                cb(cur, "ffn_norm", il);
-
-                cur = llm_build_ffn(ctx0, lctx, cur,
-                        model.layers[il].ffn_up,   NULL, NULL,
-                        model.layers[il].ffn_gate, NULL, NULL,
-                        model.layers[il].ffn_down, NULL, NULL,
-                        NULL,
-                        LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
-                cb(cur, "ffn_out", il);
-            }
-
-            // scale the hidden states for residual connection
-            cur = ggml_scale(ctx0, cur, scale_res);
-            cb(cur, "hidden_scaled_ffn", -1);
-
-            cur = ggml_add(ctx0, cur, ffn_inp);
-            cur = lctx.cvec.apply_to(ctx0, cur, il);
-            cb(cur, "l_out", il);
-
-            // input for next layer
-            inpL = cur;
-        }
-
-        cur = inpL;
-
-        cur = llm_build_norm(ctx0, cur, hparams,
-                model.output_norm, NULL,
-                LLM_NORM_RMS, cb, -1);
-        cb(cur, "result_norm", -1);
-
-        // lm_head scaling
-        const float scale_lmhead = float(n_embd_base)/float(n_embd);
-        cur = ggml_scale(ctx0, cur, scale_lmhead);
-        cb(cur, "lmhead_scaling", -1);
-
-        // lm_head
-        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
-        cb(cur, "result_output", -1);
-
-        ggml_build_forward_expand(gf, cur);
-
-        return gf;
-    }
-
     struct ggml_cgraph * build_minicpm3() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         //TODO: if the model varies, these parameters need to be read from the model
         const int64_t n_embd_base = 256;
@@ -13563,7 +4642,7 @@ struct llm_build_context {
                     0);
                 cb(v_states, "v_states", il);
 
-                q_pe = ggml_cont(ctx0, q_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
+                q_pe = ggml_cont(ctx0, q_pe); // TODO: the CUDA backend used to not support non-cont. RoPE, investigate removing this
                 q_pe = ggml_rope_ext(
                     ctx0, q_pe, inp_pos, rope_factors,
                     n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
@@ -13572,7 +4651,7 @@ struct llm_build_context {
                 cb(q_pe, "q_pe", il);
 
                 // shared RoPE key
-                k_pe = ggml_cont(ctx0, k_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
+                k_pe = ggml_cont(ctx0, k_pe); // TODO: the CUDA backend used to not support non-cont. RoPE, investigate removing this
                 k_pe = ggml_rope_ext(
                     ctx0, k_pe, inp_pos, rope_factors,
                     n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
@@ -13656,7 +4735,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_gemma() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head_k = hparams.n_embd_head_k;
 
@@ -13764,7 +4843,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_gemma2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head_k = hparams.n_embd_head_k;
 
@@ -13814,9 +4893,9 @@ struct llm_build_context {
 
                 // ref: https://github.com/google/gemma_pytorch/commit/03e657582d17cb5a8617ebf333c1c16f3694670e
                 switch (model.type) {
-                    case e_model::MODEL_2B:
-                    case e_model::MODEL_9B:  Qcur = ggml_scale(ctx0, Qcur, 1.0f / sqrtf(float(n_embd_head_k)));   break;
-                    case e_model::MODEL_27B: Qcur = ggml_scale(ctx0, Qcur, 1.0f / sqrtf(float(n_embd / n_head))); break;
+                    case LLM_TYPE_2B:
+                    case LLM_TYPE_9B:  Qcur = ggml_scale(ctx0, Qcur, 1.0f / sqrtf(float(n_embd_head_k)));   break;
+                    case LLM_TYPE_27B: Qcur = ggml_scale(ctx0, Qcur, 1.0f / sqrtf(float(n_embd / n_head))); break;
                     default: GGML_ABORT("fatal error");
                 };
                 cb(Qcur, "Qcur_scaled", il);
@@ -13900,7 +4979,7 @@ struct llm_build_context {
 
 
     struct ggml_cgraph * build_starcoder2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -14019,7 +5098,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_mamba() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         struct ggml_tensor * cur;
         struct ggml_tensor * inpL;
@@ -14074,7 +5153,7 @@ struct llm_build_context {
 
     struct ggml_cgraph * build_command_r() {
 
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -14221,6 +5300,137 @@ struct llm_build_context {
 
     }
 
+    struct ggml_cgraph * build_cohere2() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        const float f_logit_scale = hparams.f_logit_scale;
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        struct ggml_tensor * inp_pos = build_inp_pos();
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        // cohere2 requires different mask for layers using sliding window (SWA)
+        struct ggml_tensor * KQ_mask     = build_inp_KQ_mask();
+        struct ggml_tensor * KQ_mask_swa = build_inp_KQ_mask_swa();
+
+        // sliding window switch pattern
+        const int32_t sliding_window_pattern = 4;
+
+        for (int il = 0; il < n_layer; ++il) {
+            // three layers sliding window attention (window size 4096) and ROPE
+            // fourth layer uses global attention without positional embeddings
+            const bool           is_sliding = il % sliding_window_pattern < (sliding_window_pattern - 1);
+            struct ggml_tensor * KQ_mask_l = is_sliding ? KQ_mask_swa : KQ_mask;
+
+            // norm
+            cur = llm_build_norm(ctx0, inpL, hparams, model.layers[il].attn_norm, NULL, LLM_NORM, cb, il);
+            cb(cur, "attn_norm", il);
+            struct ggml_tensor * ffn_inp = cur;
+
+            // self-attention
+            {
+                // rope freq factors for 128k context
+                struct ggml_tensor * rope_factors = build_rope_factors(il);
+
+                // compute Q and K and RoPE them
+                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+                if (model.layers[il].bq) {
+                    Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
+                    cb(Qcur, "Qcur", il);
+                }
+
+                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+                if (model.layers[il].bk) {
+                    Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
+                    cb(Kcur, "Kcur", il);
+                }
+
+                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+                if (model.layers[il].bv) {
+                    Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
+                    cb(Vcur, "Vcur", il);
+                }
+
+                if (is_sliding) {
+                    Qcur = ggml_rope_ext(ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, rope_factors,
+                                        n_rot, rope_type, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor,
+                                        beta_fast, beta_slow);
+                    cb(Qcur, "Qcur", il);
+
+                    Kcur = ggml_rope_ext(ctx0, ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos,
+                                        rope_factors, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale, ext_factor,
+                                        attn_factor, beta_fast, beta_slow);
+                    cb(Kcur, "Kcur", il);
+                } else {
+                    // For non-sliding layers, just reshape without applying RoPE
+                    Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
+                    cb(Qcur, "Qcur", il);
+
+                    Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
+                    cb(Kcur, "Kcur", il);
+                }
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf, model.layers[il].wo, model.layers[il].bo, Kcur, Vcur, Qcur,
+                                   KQ_mask_l, n_tokens, kv_head, n_kv, 1.0f / sqrtf(float(n_embd_head)), cb, il);
+            }
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+                cur                              = ggml_get_rows(ctx0, cur, inp_out_ids);
+                inpL                             = ggml_get_rows(ctx0, inpL, inp_out_ids);
+                ffn_inp                          = ggml_get_rows(ctx0, ffn_inp, inp_out_ids);
+            }
+
+            struct ggml_tensor * attn_out = cur;
+
+            // feed-forward network
+            {
+                cur = llm_build_ffn(ctx0, lctx, ffn_inp, model.layers[il].ffn_up, NULL, NULL, model.layers[il].ffn_gate,
+                                    NULL, NULL, model.layers[il].ffn_down, NULL, NULL, NULL, LLM_FFN_SILU, LLM_FFN_PAR,
+                                    cb, il);
+                cb(cur, "ffn_out", il);
+            }
+
+            // add together residual + FFN + self-attention
+            cur = ggml_add(ctx0, cur, inpL);
+            cur = ggml_add(ctx0, cur, attn_out);
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams, model.output_norm, NULL, LLM_NORM, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+
+        if (f_logit_scale) {
+            cur = ggml_scale(ctx0, cur, f_logit_scale);
+        }
+
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
     // ref: https://allenai.org/olmo
     // based on the original build_llama() function, changes:
     //   * non-parametric layer norm
@@ -14228,7 +5438,7 @@ struct llm_build_context {
     //   * removed bias
     //   * removed MoE
     struct ggml_cgraph * build_olmo() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -14351,12 +5561,136 @@ struct llm_build_context {
         return gf;
     }
 
+    struct ggml_cgraph * build_olmo2() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        // mutable variable, needed during the last layer of the computation to skip unused tokens
+        int32_t n_tokens = this->n_tokens;
+
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        struct ggml_tensor * inp_pos = build_inp_pos();
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
+
+        for (int il = 0; il < n_layer; ++il) {
+            struct ggml_tensor * inpSA = inpL;
+
+            cur = inpL;
+
+            // self_attention
+            {
+                // compute Q and K and RoPE them
+                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+
+                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+
+                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+
+                Qcur = llm_build_norm(ctx0, Qcur, hparams, model.layers[il].attn_q_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(Qcur, "Qcur_normed", il);
+
+                Kcur = llm_build_norm(ctx0, Kcur, hparams, model.layers[il].attn_k_norm, NULL,
+                        LLM_NORM_RMS, cb, il);
+                cb(Kcur, "Kcur_normed", il);
+
+                Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
+                Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
+
+                Qcur = ggml_rope_ext(
+                    ctx0, Qcur, inp_pos, nullptr,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Qcur, "Qcur_rope", il);
+
+                Kcur = ggml_rope_ext(
+                    ctx0, Kcur, inp_pos, nullptr,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Kcur, "Kcur_rope", il);
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
+                        model.layers[il].wo, NULL,
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
+            }
+
+            cur = llm_build_norm(ctx0, cur, hparams,
+                    model.layers[il].attn_post_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "attn_post_norm", il);
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+                n_tokens = n_outputs;
+                cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);
+                inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+            }
+
+            struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+            cb(ffn_inp, "ffn_inp", il);
+
+            // feed-forward network
+            cur = llm_build_ffn(ctx0, lctx, ffn_inp,
+                    model.layers[il].ffn_up,   NULL, NULL,
+                    model.layers[il].ffn_gate, NULL, NULL,
+                    model.layers[il].ffn_down, NULL, NULL,
+                    NULL,
+                    LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+            cb(cur, "ffn_out", il);
+
+            cur = llm_build_norm(ctx0, cur, hparams,
+                model.layers[il].ffn_post_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+            cb(cur, "ffn_post_norm", -1);
+
+            cur = ggml_add(ctx0, cur, ffn_inp);
+            cb(cur, "ffn_out", il);
+
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
     // based on the build_qwen2moe() function, changes:
     //   * removed shared experts
     //   * removed bias
     //   * added q, k norm
     struct ggml_cgraph * build_olmoe() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -14449,9 +5783,11 @@ struct llm_build_context {
                     model.layers[il].ffn_up_exps,
                     model.layers[il].ffn_gate_exps,
                     model.layers[il].ffn_down_exps,
+                    nullptr,
                     n_expert, n_expert_used,
                     LLM_FFN_SILU, false,
                     false, 0.0,
+                    LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                     cb, il);
             cb(cur, "ffn_moe_out", il);
 
@@ -14480,7 +5816,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_openelm() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -14605,7 +5941,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_gptneox() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -14747,7 +6083,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_arctic() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -14846,9 +6182,11 @@ struct llm_build_context {
                     model.layers[il].ffn_up_exps,
                     model.layers[il].ffn_gate_exps,
                     model.layers[il].ffn_down_exps,
+                    nullptr,
                     n_expert, n_expert_used,
                     LLM_FFN_SILU, true,
                     false, 0.0,
+                    LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
                     cb, il);
             cb(cur, "ffn_moe_out", il);
 
@@ -14878,8 +6216,165 @@ struct llm_build_context {
         return gf;
     }
 
+    struct ggml_cgraph * build_deepseek() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        // mutable variable, needed during the last layer of the computation to skip unused tokens
+        int32_t n_tokens = this->n_tokens;
+
+        const int64_t n_embd_head = hparams.n_embd_head_v;
+        GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+        GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        // inp_pos - contains the positions
+        struct ggml_tensor * inp_pos = build_inp_pos();
+
+        // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
+        struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
+        const float kq_scale = hparams.f_attention_scale == 0.0f ? 1.0f/sqrtf(float(n_embd_head)) : hparams.f_attention_scale;
+        for (int il = 0; il < n_layer; ++il) {
+            struct ggml_tensor * inpSA = inpL;
+
+            // norm
+            cur = llm_build_norm(ctx0, inpL, hparams,
+                    model.layers[il].attn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "attn_norm", il);
+
+            // self-attention
+            {
+                // rope freq factors for llama3; may return nullptr for llama2 and other models
+                struct ggml_tensor * rope_factors = build_rope_factors(il);
+
+                // compute Q and K and RoPE them
+                struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
+                cb(Qcur, "Qcur", il);
+                if (model.layers[il].bq) {
+                    Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
+                    cb(Qcur, "Qcur", il);
+                }
+
+                struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
+                cb(Kcur, "Kcur", il);
+                if (model.layers[il].bk) {
+                    Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
+                    cb(Kcur, "Kcur", il);
+                }
+
+                struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
+                cb(Vcur, "Vcur", il);
+                if (model.layers[il].bv) {
+                    Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
+                    cb(Vcur, "Vcur", il);
+                }
+
+                Qcur = ggml_rope_ext(
+                    ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, rope_factors,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Qcur, "Qcur", il);
+
+                Kcur = ggml_rope_ext(
+                    ctx0, ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, rope_factors,
+                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+                    ext_factor, attn_factor, beta_fast, beta_slow
+                );
+                cb(Kcur, "Kcur", il);
+
+                cur = llm_build_kv(ctx0, lctx, kv_self, gf,
+                        model.layers[il].wo, model.layers[il].bo,
+                        Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, kq_scale, cb, il);
+            }
+
+            if (il == n_layer - 1) {
+                // skip computing output for unused tokens
+                struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+                n_tokens = n_outputs;
+                cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);
+                inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+            }
+
+
+            struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+            cb(ffn_inp, "ffn_inp", il);
+
+            cur = llm_build_norm(ctx0, ffn_inp, hparams,
+                    model.layers[il].ffn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "ffn_norm", il);
+
+            if ((uint32_t) il < hparams.n_layer_dense_lead) {
+                cur = llm_build_ffn(ctx0, lctx, cur,
+                        model.layers[il].ffn_up,   NULL, NULL,
+                        model.layers[il].ffn_gate, NULL, NULL,
+                        model.layers[il].ffn_down, NULL, NULL,
+                        NULL,
+                        LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+                cb(cur, "ffn_out", il);
+            } else {
+                // MoE branch
+                ggml_tensor * moe_out =
+                        llm_build_moe_ffn(ctx0, lctx, cur,
+                            model.layers[il].ffn_gate_inp,
+                            model.layers[il].ffn_up_exps,
+                            model.layers[il].ffn_gate_exps,
+                            model.layers[il].ffn_down_exps,
+                            nullptr,
+                            n_expert, n_expert_used,
+                            LLM_FFN_SILU, false,
+                            false, hparams.expert_weights_scale,
+                            LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
+                            cb, il);
+                cb(moe_out, "ffn_moe_out", il);
+
+                // FFN shared expert
+                {
+                    ggml_tensor * ffn_shexp = llm_build_ffn(ctx0, lctx, cur,
+                            model.layers[il].ffn_up_shexp,   NULL, NULL,
+                            model.layers[il].ffn_gate_shexp, NULL, NULL,
+                            model.layers[il].ffn_down_shexp, NULL, NULL,
+                            NULL,
+                            LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+                    cb(ffn_shexp, "ffn_shexp", il);
+
+                    cur = ggml_add(ctx0, moe_out, ffn_shexp);
+                    cb(cur, "ffn_out", il);
+                }
+            }
+
+            cur = ggml_add(ctx0, cur, ffn_inp);
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm, NULL,
+                LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
     struct ggml_cgraph * build_deepseek2() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -15001,7 +6496,7 @@ struct llm_build_context {
                     0);
                 cb(v_states, "v_states", il);
 
-                q_pe = ggml_cont(ctx0, q_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
+                q_pe = ggml_cont(ctx0, q_pe); // TODO: the CUDA backend used to not support non-cont. RoPE, investigate removing this
                 q_pe = ggml_rope_ext(
                     ctx0, q_pe, inp_pos, nullptr,
                     n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
@@ -15010,7 +6505,7 @@ struct llm_build_context {
                 cb(q_pe, "q_pe", il);
 
                 // shared RoPE key
-                k_pe = ggml_cont(ctx0, k_pe); // TODO: the CUDA backend does not support non-contiguous RoPE
+                k_pe = ggml_cont(ctx0, k_pe); // TODO: the CUDA backend used to not support non-cont. RoPE, investigate removing this
                 k_pe = ggml_rope_ext(
                     ctx0, k_pe, inp_pos, nullptr,
                     n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
@@ -15061,9 +6556,11 @@ struct llm_build_context {
                             model.layers[il].ffn_up_exps,
                             model.layers[il].ffn_gate_exps,
                             model.layers[il].ffn_down_exps,
+                            model.layers[il].ffn_exp_probs_b,
                             n_expert, n_expert_used,
-                            LLM_FFN_SILU, false,
+                            LLM_FFN_SILU, hparams.expert_weights_norm,
                             true, hparams.expert_weights_scale,
+                            (enum llama_expert_gating_func_type) hparams.expert_gating_func,
                             cb, il);
                 cb(moe_out, "ffn_moe_out", il);
 
@@ -15107,7 +6604,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_bitnet() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -15257,8 +6754,8 @@ struct llm_build_context {
         return gf;
     }
 
-    struct ggml_cgraph * build_t5_encoder() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+    struct ggml_cgraph * build_t5_enc() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -15389,8 +6886,8 @@ struct llm_build_context {
         return gf;
     }
 
-    struct ggml_cgraph * build_t5_decoder() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+    struct ggml_cgraph * build_t5_dec() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -15595,7 +7092,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_jais() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -15687,7 +7184,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_chatglm() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         const int64_t n_embd_gqa  = hparams.n_embd_v_gqa();
@@ -15801,7 +7298,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_nemotron() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         const int64_t n_embd_head = hparams.n_embd_head_v;
         GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
@@ -15922,7 +7419,7 @@ struct llm_build_context {
     }
 
     struct ggml_cgraph * build_exaone() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -16049,7 +7546,7 @@ struct llm_build_context {
     }
 
     ggml_cgraph * build_rwkv6() {
-        ggml_cgraph *gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // Token shift state dimensions should be 2 * n_emb
         GGML_ASSERT(n_embd == hparams.n_embd_k_s() / 2);
@@ -16094,7 +7591,7 @@ struct llm_build_context {
                 1
             );
 
-            cur = ggml_add(ctx0, cur, llm_build_rwkv6_time_mix(lctx, ctx0, layer, x_norm_att, x_prev, &wkv_states));
+            cur = ggml_add(ctx0, cur, llm_build_rwkv6_time_mix(lctx, ctx0, layer, x_norm_att, x_prev, &wkv_states, hparams.wkv_head_size, n_embd / hparams.wkv_head_size));
             ggml_build_forward_expand(gf, cur);
             ggml_build_forward_expand(
                 gf,
@@ -16161,6 +7658,118 @@ struct llm_build_context {
         return gf;
     }
 
+    // ref: https://huggingface.co/recursal/QRWKV6-32B-Instruct-Preview-v0.1/blob/main/modeling_rwkv6qwen2.py
+    ggml_cgraph * build_rwkv6qwen2() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        GGML_ASSERT(n_embd == hparams.n_embd_k_s());
+
+        const int64_t n_seqs = ubatch.n_seqs;
+        const int64_t n_seq_tokens = ubatch.n_seq_tokens;
+        const int64_t n_tokens = ubatch.n_tokens;
+        GGML_ASSERT(n_seqs != 0);
+        GGML_ASSERT(ubatch.equal_seqs);
+        GGML_ASSERT(n_tokens == n_seq_tokens * n_seqs);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+        struct ggml_tensor * state_copy = build_inp_s_copy();
+        struct ggml_tensor * state_mask = build_inp_s_mask();
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        for (int il = 0; il < n_layer; ++il) {
+            const llama_layer * layer = &model.layers[il];
+
+            // (ab)using the KV cache to store the states
+            struct ggml_tensor * token_shift = llm_build_copy_mask_state(ctx0,
+                    gf, kv_self.k_l[il], state_copy, state_mask,
+                    hparams.n_embd_k_s(), kv_self.size, kv_head, n_kv, n_seqs);
+            struct ggml_tensor * wkv_states = llm_build_copy_mask_state(ctx0,
+                    gf, kv_self.v_l[il], state_copy, state_mask,
+                    hparams.n_embd_v_s(), kv_self.size, kv_head, n_kv, n_seqs);
+
+            cur = ggml_reshape_3d(ctx0, inpL, n_embd, n_seq_tokens, n_seqs);
+            token_shift = ggml_reshape_3d(ctx0, token_shift, n_embd, 1, n_seqs);
+
+            struct ggml_tensor * x_norm_att = llm_build_norm(ctx0, cur, hparams, layer->attn_norm, layer->attn_norm_b, LLM_NORM_RMS, cb, il);
+            struct ggml_tensor * x_prev = ggml_concat(
+                ctx0,
+                token_shift,
+                ggml_view_3d(ctx0, x_norm_att, n_embd, n_seq_tokens - 1, n_seqs, x_norm_att->nb[1], x_norm_att->nb[2], 0),
+                1
+            );
+
+            ggml_build_forward_expand(
+                gf,
+                ggml_cpy(
+                    ctx0,
+                    wkv_states,
+                    ggml_view_1d(
+                        ctx0,
+                        kv_self.v_l[il],
+                        hparams.n_embd_v_s() * n_seqs,
+                        hparams.n_embd_v_s() * kv_head * ggml_element_size(kv_self.v_l[il])
+                    )
+                )
+            );
+
+            struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, llm_build_rwkv6_time_mix(lctx, ctx0, layer, x_norm_att, x_prev, &wkv_states, hparams.wkv_head_size, hparams.n_head_kv()));
+            ggml_build_forward_expand(gf, ffn_inp);
+            ggml_build_forward_expand(
+                gf,
+                ggml_cpy(
+                    ctx0,
+                    wkv_states,
+                    ggml_view_1d(
+                        ctx0,
+                        kv_self.v_l[il],
+                        hparams.n_embd_v_s() * n_seqs,
+                        hparams.n_embd_v_s() * kv_head * ggml_element_size(kv_self.v_l[il])
+                    )
+                )
+            );
+
+            cb(ffn_inp, "ffn_inp", il);
+
+            // feed-forward network
+            cur = llm_build_norm(ctx0, ffn_inp, hparams,
+                    model.layers[il].ffn_norm, NULL,
+                    LLM_NORM_RMS, cb, il);
+            cb(cur, "ffn_norm", il);
+
+            cur = llm_build_ffn(ctx0, lctx, cur,
+                    model.layers[il].ffn_up,   NULL, NULL,
+                    model.layers[il].ffn_gate, NULL, NULL,
+                    model.layers[il].ffn_down, NULL, NULL,
+                    NULL,
+                    LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
+            cb(cur, "ffn_out", il);
+
+            cur = ggml_add(ctx0, cur, ffn_inp);
+            cur = lctx.cvec.apply_to(ctx0, cur, il);
+            cb(cur, "l_out", il);
+
+            // input for next layer
+            inpL = cur;
+        }
+
+        cur = inpL;
+        struct ggml_tensor * inp_out_ids = build_inp_out_ids();
+        cur = ggml_reshape_2d(ctx0, cur, n_embd, n_tokens);
+        cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+
+        cur = llm_build_norm(ctx0, cur, hparams, model.output_norm, model.output_norm_b, LLM_NORM_RMS, cb, -1);
+        cb(cur, "result_norm", -1);
+
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+        cb(cur, "result_output", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
+
     // ref: https://github.com/facebookresearch/chameleon
     // based on the original build_llama() function, changes:
     //   * qk-norm
@@ -16168,7 +7777,7 @@ struct llm_build_context {
     //   * removed bias
     //   * removed MoE
     struct ggml_cgraph * build_chameleon() {
-        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
 
         // mutable variable, needed during the last layer of the computation to skip unused tokens
         int32_t n_tokens = this->n_tokens;
@@ -16338,6 +7947,158 @@ struct llm_build_context {
 
         return gf;
     }
+
+    struct ggml_cgraph * build_wavtokenizer_dec() {
+        struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, model.max_nodes(), false);
+
+        struct ggml_tensor * cur;
+        struct ggml_tensor * inpL;
+
+        inpL = llm_build_inp_embd(ctx0, lctx, hparams, ubatch, model.tok_embd, cb);
+
+        cur = ggml_cont(ctx0, ggml_transpose(ctx0, inpL));
+
+        cur = ggml_conv_1d_ph(ctx0, model.conv1d, cur, 1, 1);
+        cur = ggml_add(ctx0, cur, model.conv1d_b);
+
+        // posnet
+        for (uint32_t il = 0; il < hparams.posnet.n_layer; ++il) {
+            const auto & layer = model.layers[il].posnet;
+
+            inpL = cur;
+
+            switch (il) {
+                case 0:
+                case 1:
+                case 3:
+                case 4:
+                    {
+                        cur = llm_build_norm(ctx0, cur, hparams,
+                                layer.norm1,
+                                layer.norm1_b,
+                                LLM_NORM_GROUP, cb, 0);
+
+                        cur = ggml_mul(ctx0, ggml_sigmoid(ctx0, cur), cur);
+
+                        cur = ggml_conv_1d_ph(ctx0, layer.conv1, cur, 1, 1);
+                        cur = ggml_add(ctx0, cur, layer.conv1_b);
+
+                        cur = llm_build_norm(ctx0, cur, hparams,
+                                layer.norm2,
+                                layer.norm2_b,
+                                LLM_NORM_GROUP, cb, 0);
+
+                        cur = ggml_mul(ctx0, ggml_sigmoid(ctx0, cur), cur);
+
+                        cur = ggml_conv_1d_ph(ctx0, layer.conv2, cur, 1, 1);
+                        cur = ggml_add(ctx0, cur, layer.conv2_b);
+
+                        cur = ggml_add(ctx0, cur, inpL);
+                    } break;
+                case 2:
+                    {
+                        cur = llm_build_norm(ctx0, cur, hparams,
+                                layer.attn_norm,
+                                layer.attn_norm_b,
+                                LLM_NORM_GROUP, cb, 0);
+
+                        struct ggml_tensor * q;
+                        struct ggml_tensor * k;
+                        struct ggml_tensor * v;
+
+                        q = ggml_conv_1d_ph(ctx0, layer.attn_q, cur, 1, 1);
+                        k = ggml_conv_1d_ph(ctx0, layer.attn_k, cur, 1, 1);
+                        v = ggml_conv_1d_ph(ctx0, layer.attn_v, cur, 1, 1);
+
+                        q = ggml_add(ctx0, q, layer.attn_q_b);
+                        k = ggml_add(ctx0, k, layer.attn_k_b);
+                        v = ggml_add(ctx0, v, layer.attn_v_b);
+
+                        q = ggml_cont(ctx0, ggml_transpose(ctx0, q));
+                        k = ggml_cont(ctx0, ggml_transpose(ctx0, k));
+
+                        struct ggml_tensor * kq = ggml_mul_mat(ctx0, k, q);
+
+                        kq = ggml_soft_max_ext(ctx0, kq, nullptr, 1.0f/sqrtf(float(hparams.posnet.n_embd)), 0.0f);
+
+                        cur = ggml_mul_mat(ctx0, kq, v);
+
+                        cur = ggml_conv_1d_ph(ctx0, layer.attn_o, cur, 1, 1);
+                        cur = ggml_add(ctx0, cur, layer.attn_o_b);
+
+                        cur = ggml_add(ctx0, cur, inpL);
+                    } break;
+                case 5:
+                    {
+                        cur = llm_build_norm(ctx0, cur, hparams,
+                                layer.norm,
+                                layer.norm_b,
+                                LLM_NORM_GROUP, cb, 0);
+                    } break;
+                default: GGML_ABORT("unknown posnet layer");
+            };
+        }
+
+        cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.tok_norm,
+                model.tok_norm_b,
+                LLM_NORM, cb, -1);
+
+        cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));
+
+        inpL = cur;
+
+        // convnext
+        for (uint32_t il = 0; il < hparams.convnext.n_layer; ++il) {
+            const auto & layer = model.layers[il].convnext;
+
+            cur = inpL;
+
+            cur = ggml_conv_1d_dw_ph(ctx0, layer.dw, cur, 1, 1);
+            cur = ggml_add(ctx0, cur, layer.dw_b);
+
+            cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));
+
+            cur = llm_build_norm(ctx0, cur, hparams,
+                    layer.norm,
+                    layer.norm_b,
+                    LLM_NORM, cb, -1);
+
+            cur = llm_build_ffn(ctx0, lctx, cur,
+                    layer.pw1, layer.pw1_b, NULL,
+                    NULL,      NULL,        NULL,
+                    layer.pw2, layer.pw2_b, NULL,
+                    NULL,
+                    LLM_FFN_GELU, LLM_FFN_SEQ, cb, il);
+
+            cur = ggml_mul(ctx0, cur, layer.gamma);
+
+            cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));
+
+            inpL = ggml_add(ctx0, cur, inpL);
+        }
+
+        cur = inpL;
+
+        cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));
+
+        cur = llm_build_norm(ctx0, cur, hparams,
+                model.output_norm,
+                model.output_norm_b,
+                LLM_NORM, cb, -1);
+
+        // lm_head
+        cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
+
+        cur = ggml_add(ctx0, cur, model.output_b);
+        cb(cur, "result_embd", -1);
+
+        ggml_build_forward_expand(gf, cur);
+
+        return gf;
+    }
 };
 
 static struct ggml_cgraph * llama_build_graph_defrag(llama_context & lctx, const std::vector & ids) {
@@ -16397,12 +8158,12 @@ static struct ggml_cgraph * llama_build_graph(
 
         // norm may be automatically assigned to the backend of the previous layer, increasing data transfer between backends
         // FIXME: fix in ggml_backend_sched
-        const bool full_offload = lctx.model.n_gpu_layers > (int)lctx.model.hparams.n_layer;
+        const bool full_offload = lctx.model.params.n_gpu_layers > (int) lctx.model.hparams.n_layer;
         if (ubatch.n_tokens < 32 || full_offload) {
             if (il != -1 && strcmp(name, "norm") == 0) {
-                const auto & dev_layer = lctx.model.dev_layer.at(il);
+                const auto & dev_layer = lctx.model.dev_layer(il);
                 for (auto & backend : lctx.backends) {
-                    if (ggml_backend_get_device(backend.get()) == dev_layer.dev) {
+                    if (ggml_backend_get_device(backend.get()) == dev_layer) {
                         if (ggml_backend_supports_op(backend.get(), cur)) {
                             ggml_backend_sched_set_tensor_backend(lctx.sched.get(), cur, backend.get());
                         }
@@ -16420,11 +8181,16 @@ static struct ggml_cgraph * llama_build_graph(
 
     switch (model.arch) {
         case LLM_ARCH_LLAMA:
+        case LLM_ARCH_MINICPM:
         case LLM_ARCH_GRANITE:
         case LLM_ARCH_GRANITE_MOE:
             {
                 result = llm.build_llama();
             } break;
+        case LLM_ARCH_DECI:
+            {
+                result = llm.build_deci();
+            } break;
         case LLM_ARCH_BAICHUAN:
             {
                 result = llm.build_baichuan();
@@ -16471,6 +8237,11 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_qwen2();
             } break;
+        case LLM_ARCH_QWEN2VL:
+            {
+                lctx.n_pos_per_token = 4;
+                result = llm.build_qwen2vl();
+            } break;
         case LLM_ARCH_QWEN2MOE:
             {
                 result = llm.build_qwen2moe();
@@ -16480,6 +8251,7 @@ static struct ggml_cgraph * llama_build_graph(
                 result = llm.build_phi2();
             } break;
         case LLM_ARCH_PHI3:
+        case LLM_ARCH_PHIMOE:
             {
                 result = llm.build_phi3();
             } break;
@@ -16503,10 +8275,6 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_internlm2();
             } break;
-        case LLM_ARCH_MINICPM:
-            {
-                result = llm.build_minicpm();
-            } break;
         case LLM_ARCH_MINICPM3:
             {
                 result = llm.build_minicpm3();
@@ -16535,6 +8303,10 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_command_r();
             } break;
+        case LLM_ARCH_COHERE2:
+            {
+                result = llm.build_cohere2();
+            } break;
         case LLM_ARCH_DBRX:
             {
                 result = llm.build_dbrx();
@@ -16543,6 +8315,10 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_olmo();
             } break;
+        case LLM_ARCH_OLMO2:
+            {
+                result = llm.build_olmo2();
+            } break;
         case LLM_ARCH_OLMOE:
             {
                 result = llm.build_olmoe();
@@ -16559,6 +8335,10 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_arctic();
             } break;
+        case LLM_ARCH_DEEPSEEK:
+            {
+                result = llm.build_deepseek();
+            } break;
         case LLM_ARCH_DEEPSEEK2:
             {
                 result = llm.build_deepseek2();
@@ -16574,14 +8354,14 @@ static struct ggml_cgraph * llama_build_graph(
         case LLM_ARCH_T5:
             {
                 if (lctx.is_encoding) {
-                    result = llm.build_t5_encoder();
+                    result = llm.build_t5_enc();
                 } else {
-                    result = llm.build_t5_decoder();
+                    result = llm.build_t5_dec();
                 }
             } break;
         case LLM_ARCH_T5ENCODER:
             {
-                result = llm.build_t5_encoder();
+                result = llm.build_t5_enc();
             } break;
         case LLM_ARCH_JAIS:
             {
@@ -16599,10 +8379,18 @@ static struct ggml_cgraph * llama_build_graph(
             {
                 result = llm.build_rwkv6();
             } break;
+        case LLM_ARCH_RWKV6QWEN2:
+            {
+                result = llm.build_rwkv6qwen2();
+            } break;
         case LLM_ARCH_CHAMELEON:
             {
                 result = llm.build_chameleon();
             } break;
+        case LLM_ARCH_WAVTOKENIZER_DEC:
+            {
+                result = llm.build_wavtokenizer_dec();
+            } break;
         default:
             GGML_ABORT("fatal error");
     }
@@ -16617,575 +8405,16 @@ static struct ggml_cgraph * llama_build_graph(
     return result;
 }
 
-static void llama_set_k_shift(llama_context & lctx) {
-    const int64_t kv_size = lctx.kv_self.size;
-
-    assert(ggml_backend_buffer_is_host(lctx.inp_K_shift->buffer));
-
-    int32_t * data = (int32_t *) lctx.inp_K_shift->data;
-
-    for (int i = 0; i < kv_size; ++i) {
-        data[i] = lctx.kv_self.cells[i].delta;
-    }
-}
-
-static void llama_set_s_copy(llama_context & lctx) {
-    const int64_t kv_size = lctx.kv_self.size;
-
-    assert(ggml_backend_buffer_is_host(lctx.inp_s_copy->buffer));
-
-    int32_t * data = (int32_t *) lctx.inp_s_copy->data;
-
-    for (int i = 0; i < kv_size; ++i) {
-        data[i] = lctx.kv_self.cells[i].src;
-    }
-}
-
-static int32_t llama_relative_position_bucket(llama_pos x, llama_pos y, uint64_t n_buckets, bool bidirectional) {
-    // TODO move to hparams if a T5 variant appears that uses a different value
-    const int64_t max_distance = 128;
-
-    if (bidirectional) {
-        n_buckets >>= 1;
-    }
-
-    const int64_t max_exact = n_buckets >> 1;
-
-    int32_t relative_position = x - y;
-    int32_t relative_bucket = 0;
-    if (bidirectional) {
-        relative_bucket += (relative_position > 0) * n_buckets;
-        relative_position = abs(relative_position);
-    } else {
-        relative_position = -std::min(relative_position, 0);
-    }
-    int32_t relative_position_if_large = floorf(max_exact + logf(1.0 * relative_position / max_exact) * (n_buckets - max_exact) / log(1.0 * max_distance / max_exact));
-    relative_position_if_large = std::min(relative_position_if_large, n_buckets - 1);
-    relative_bucket += (relative_position < max_exact ? relative_position : relative_position_if_large);
-    return relative_bucket;
-}
-
-static void llama_set_inputs(llama_context & lctx, const llama_ubatch & ubatch) {
-    //
-    // set input data
-    //
-
-    const auto & hparams = lctx.model.hparams;
-    const auto & cparams = lctx.cparams;
-    const auto & kv_self = lctx.kv_self;
-
-    if (ubatch.token) {
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        ggml_backend_tensor_set(lctx.inp_tokens, ubatch.token, 0, n_tokens*ggml_element_size(lctx.inp_tokens));
-    }
-
-    if (ubatch.embd) {
-        const int64_t n_embd   = hparams.n_embd;
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        ggml_backend_tensor_set(lctx.inp_embd, ubatch.embd, 0, n_tokens*n_embd*ggml_element_size(lctx.inp_embd));
-    }
-
-    if (ubatch.pos && lctx.inp_pos) {
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        ggml_backend_tensor_set(lctx.inp_pos, ubatch.pos, 0, n_tokens*ggml_element_size(lctx.inp_pos));
-    }
-
-    if (hparams.causal_attn || cparams.pooling_type == LLAMA_POOLING_TYPE_NONE) {
-        GGML_ASSERT(lctx.inp_out_ids && "every model that can must skip unused outputs");
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_out_ids->buffer));
-        int32_t * data = (int32_t *) lctx.inp_out_ids->data;
-
-        if (lctx.n_outputs == n_tokens) {
-            for (int i = 0; i < n_tokens; ++i) {
-                data[i] = i;
-            }
-        } else if (ubatch.output) {
-            int32_t n_outputs = 0;
-            for (int i = 0; i < n_tokens; ++i) {
-                if (ubatch.output[i]) {
-                    data[n_outputs++] = i;
-                }
-            }
-            // the graph needs to have been passed the correct number of outputs
-            GGML_ASSERT(lctx.n_outputs == n_outputs);
-        } else if (lctx.n_outputs == 1) {
-            // only keep last output
-            data[0] = n_tokens - 1;
-        } else {
-            GGML_ASSERT(lctx.n_outputs == 0);
-        }
-    }
-
-    GGML_ASSERT(
-        // (!a || b) is a logical implication (a -> b)
-        // !hparams.causal_attn -> !cparams.causal_attn
-        (hparams.causal_attn || !cparams.causal_attn) &&
-        "causal attention is not supported by this model"
-    );
-
-    if (lctx.inp_KQ_mask || lctx.inp_KQ_mask_swa) {
-        // NOTE: hparams.causal_attn indicates the model is capable of generation and uses the kv cache.
-        if (cparams.causal_attn && !lctx.is_encoding) {
-            const int64_t n_kv         = kv_self.n;
-            const int64_t n_tokens     = ubatch.n_tokens;
-            const int64_t n_seq_tokens = ubatch.n_seq_tokens;
-            const int64_t n_seqs       = ubatch.n_seqs;
-
-
-            float * data     = nullptr;
-            float * data_swa = nullptr;
-
-            if (lctx.inp_KQ_mask) {
-                GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask->buffer));
-                data = (float *) lctx.inp_KQ_mask->data;
-            }
-
-            if (lctx.inp_KQ_mask_swa) {
-                GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask_swa->buffer));
-                data_swa = (float *) lctx.inp_KQ_mask_swa->data;
-            }
-
-            // For causal attention, use only the previous KV cells
-            // of the correct sequence for each token of the ubatch.
-            // It's assumed that if a token in the batch has multiple sequences, they are equivalent.
-            for (int h = 0; h < 1; ++h) {
-                for (int s = 0; s < n_seqs; ++s) {
-                    const llama_seq_id seq_id = ubatch.seq_id[s][0];
-
-                    for (int j = 0; j < n_seq_tokens; ++j) {
-                        const llama_pos pos = ubatch.pos[s*n_seq_tokens + j];
-
-                        for (int i = 0; i < n_kv; ++i) {
-                            float f;
-                            if (!kv_self.cells[i].has_seq_id(seq_id) || kv_self.cells[i].pos > pos) {
-                                f = -INFINITY;
-                            } else {
-                                if (hparams.use_alibi) {
-                                    f = -std::abs(kv_self.cells[i].pos - pos);
-                                } else {
-                                    f = 0.0f;
-                                }
-                            }
-
-                            if (data) {
-                                data[h*(n_kv*n_tokens) + s*(n_kv*n_seq_tokens) + j*n_kv + i] = f;
-                            }
-
-                            // may need to cut off old tokens for sliding window
-                            if (data_swa) {
-                                if (pos - kv_self.cells[i].pos >= (int32_t)hparams.n_swa) {
-                                    f = -INFINITY;
-                                }
-                                data_swa[h*(n_kv*n_tokens) + s*(n_kv*n_seq_tokens) + j*n_kv + i] = f;
-                            }
-                        }
-                    }
-                }
-
-                if (data) {
-                    for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) {
-                        for (int j = 0; j < n_kv; ++j) {
-                            data[h*(n_kv*n_tokens) + i*n_kv + j] = -INFINITY;
-                        }
-                    }
-                }
-
-                if (data_swa) {
-                    for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) {
-                        for (int j = 0; j < n_kv; ++j) {
-                            data_swa[h*(n_kv*n_tokens) + i*n_kv + j] = -INFINITY;
-                        }
-                    }
-                }
-            }
-        } else {
-            const int64_t n_tokens     = ubatch.n_tokens;
-            const int64_t n_seq_tokens = ubatch.n_seq_tokens;
-            const int64_t n_seqs       = ubatch.n_seqs;
-            // when using kv cache, the mask needs to match the kv cache size
-            const int64_t n_stride = hparams.causal_attn && !lctx.is_encoding ? kv_self.n : n_tokens;
-
-            GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask->buffer));
-
-            float * data = (float *) lctx.inp_KQ_mask->data;
-
-            for (int h = 0; h < 1; ++h) {
-                for (int s1 = 0; s1 < n_seqs; ++s1) {
-                    const llama_seq_id seq_id = ubatch.seq_id[s1][0];
-
-                    for (int j = 0; j < n_seq_tokens; ++j) {
-                        const int32_t tj = s1*n_seq_tokens + j;
-
-                        for (int s0 = 0; s0 < n_seqs; ++s0) {
-                            for (int i = 0; i < n_seq_tokens; ++i) {
-                                const int32_t ti = s0*n_seq_tokens + i;
-                                float f = -INFINITY;
-
-                                for (int s = 0; s < ubatch.n_seq_id[s0]; ++s) {
-                                    if (ubatch.seq_id[s0][s] == seq_id) {
-                                        if (hparams.use_alibi) {
-                                            f = -std::abs(ubatch.pos[ti] - ubatch.pos[tj]);
-                                        } else {
-                                            f = 0.0f;
-                                        }
-                                        break;
-                                    }
-                                }
-
-                                data[h*(n_tokens*n_tokens) + tj*n_stride + ti] = f;
-                            }
-                        }
-
-                        for (int i = n_tokens; i < n_stride; ++i) {
-                            data[h*(n_tokens*n_tokens) + tj*n_stride + i] = -INFINITY;
-                        }
-                    }
-                }
-            }
-        }
-    }
-
-    if (cparams.embeddings && cparams.pooling_type == LLAMA_POOLING_TYPE_MEAN) {
-        const int64_t n_tokens     = ubatch.n_tokens;
-        const int64_t n_seq_tokens = ubatch.n_seq_tokens;
-        const int64_t n_seqs       = ubatch.n_seqs;
-
-        GGML_ASSERT(lctx.inp_mean);
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_mean->buffer));
-
-        float * data = (float *) lctx.inp_mean->data;
-        memset(lctx.inp_mean->data, 0, n_tokens * n_tokens * ggml_element_size(lctx.inp_mean));
-
-        std::vector sum(n_tokens, 0);
-
-        for (int s = 0; s < n_seqs; ++s) {
-            const llama_seq_id seq_id = ubatch.seq_id[s][0];
-
-            // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true
-            GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == MEAN");
-
-            sum[seq_id] += ubatch.n_seq_tokens;
-        }
-
-        std::vector div(n_tokens, 0.0f);
-        for (int i = 0; i < n_tokens; ++i) {
-            const uint64_t s = sum[i];
-            if (s > 0) {
-                div[i] = 1.0f/float(s);
-            }
-        }
-
-        for (int s = 0; s < n_seqs; ++s) {
-            const llama_seq_id seq_id = ubatch.seq_id[s][0];
-
-            for (int i = 0; i < n_seq_tokens; ++i) {
-                data[seq_id*n_tokens + s*n_seq_tokens + i] = div[seq_id];
-            }
-        }
-    }
-
-    if (cparams.embeddings && (
-                cparams.pooling_type == LLAMA_POOLING_TYPE_CLS ||
-                cparams.pooling_type == LLAMA_POOLING_TYPE_RANK)) {
-        const int64_t n_tokens     = ubatch.n_tokens;
-        const int64_t n_seq_tokens = ubatch.n_seq_tokens;
-        const int64_t n_seqs       = ubatch.n_seqs;
-
-        GGML_ASSERT(lctx.inp_cls);
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_cls->buffer));
-
-        uint32_t * data = (uint32_t *) lctx.inp_cls->data;
-        memset(lctx.inp_cls->data, 0, n_tokens * ggml_element_size(lctx.inp_cls));
-
-        for (int s = 0; s < n_seqs; ++s) {
-            const llama_seq_id seq_id = ubatch.seq_id[s][0];
-
-            // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true
-            GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == CLS or RANK");
-
-            for (int i = 0; i < n_seq_tokens; ++i) {
-                const llama_pos pos = ubatch.pos[s*n_seq_tokens + i];
-
-                if (pos == 0) {
-                    data[seq_id] = s*n_seq_tokens + i;
-                }
-            }
-        }
-    }
-
-    if (cparams.embeddings && cparams.pooling_type == LLAMA_POOLING_TYPE_LAST) {
-        const int64_t n_tokens     = ubatch.n_tokens;
-        const int64_t n_seq_tokens = ubatch.n_seq_tokens;
-        const int64_t n_seqs       = ubatch.n_seqs;
-
-        GGML_ASSERT(lctx.inp_cls);
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_cls->buffer));
-
-        uint32_t * data = (uint32_t *) lctx.inp_cls->data;
-        memset(lctx.inp_cls->data, 0, n_tokens * ggml_element_size(lctx.inp_cls));
-
-        std::vector last_pos(n_tokens, -1);
-        std::vector last_row(n_tokens, -1);
-
-        for (int s = 0; s < n_seqs; ++s) {
-            const llama_seq_id seq_id = ubatch.seq_id[s][0];
-
-            // TODO: adapt limits to n_seqs when ubatch.equal_seqs is true
-            GGML_ASSERT(seq_id < n_tokens && "seq_id cannot be larger than n_tokens with pooling_type == LAST");
-
-            for (int i = 0; i < n_seq_tokens; ++i) {
-                const llama_pos pos = ubatch.pos[s*n_seq_tokens + i];
-
-                if (pos >= last_pos[seq_id]) {
-                    last_pos[seq_id] = pos;
-                    last_row[seq_id] = s*n_seq_tokens + i;
-                }
-            }
-        }
-
-        for (int i = 0; i < n_tokens; ++i) {
-            if (last_row[i] >= 0) {
-                data[i] = last_row[i];
-            }
-        }
-    }
-
-    if (kv_self.recurrent) {
-        const int64_t n_kv = kv_self.n;
-
-        if (lctx.inp_s_mask) {
-            GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_s_mask->buffer));
-            float * data = (float *) lctx.inp_s_mask->data;
-
-            // clear unused states
-            for (int i = 0; i < n_kv; ++i) {
-                const uint32_t  cell_id = i + kv_self.head;
-                llama_kv_cell & kv_cell = lctx.kv_self.cells[cell_id];
-
-                data[i] = (float) (kv_cell.src >= 0);
-
-                // only clear once
-                if (kv_cell.src < 0) {
-                    kv_cell.src = cell_id;
-                }
-            }
-        }
-
-        if (lctx.inp_s_copy) {
-            GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_s_copy->buffer));
-            int32_t * data = (int32_t *) lctx.inp_s_copy->data;
-
-            // assuming copy destinations ALWAYS happen ONLY on the cells between head and head+n
-            for (uint32_t i = 0; i < n_kv; ++i) {
-                const uint32_t  cell_id = i + kv_self.head;
-                llama_kv_cell & kv_cell = lctx.kv_self.cells[cell_id];
-
-                // prevent out-of-bound sources
-                if (kv_cell.src < 0 || (uint32_t) kv_cell.src >= kv_self.size) {
-                    kv_cell.src = cell_id;
-                }
-
-                data[i] = kv_cell.src;
-
-                // ensure copy only happens once
-                if (kv_cell.src != (int32_t) cell_id) {
-                    kv_cell.src = cell_id;
-                }
-            }
-        }
-    }
-
-    if (lctx.inp_pos_bucket) {
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_pos_bucket->buffer));
-        GGML_ASSERT(!ubatch.equal_seqs); // TODO: use ubatch.n_seqs instead of failing
-
-        int32_t * data = (int32_t *) lctx.inp_pos_bucket->data;
-
-        if (!lctx.is_encoding) {
-            const int64_t n_kv = kv_self.n;
-            for (int h = 0; h < 1; ++h) {
-                for (int j = 0; j < n_tokens; ++j) {
-                    for (int i = 0; i < n_kv; ++i) {
-                        data[h*(n_kv*n_tokens) + j*n_kv + i] = llama_relative_position_bucket(lctx.kv_self.cells[i].pos, ubatch.pos[j], hparams.n_rel_attn_bkts, lctx.is_encoding);
-                    }
-                }
-            }
-        } else {
-            for (int h = 0; h < 1; ++h) {
-                for (int j = 0; j < n_tokens; ++j) {
-                    for (int i = 0; i < n_tokens; ++i) {
-                        data[h*(n_tokens*n_tokens) + j*n_tokens + i] = llama_relative_position_bucket(ubatch.pos[i], ubatch.pos[j], hparams.n_rel_attn_bkts, lctx.is_encoding);
-                    }
-                }
-            }
-        }
-    }
-
-    if (!lctx.is_encoding && lctx.inp_embd_enc) {
-        assert(lctx.inp_embd_enc->type == GGML_TYPE_F32);
-        assert((size_t) ggml_nelements(lctx.inp_embd_enc) == lctx.embd_enc.size());
-
-        ggml_backend_tensor_set(lctx.inp_embd_enc, lctx.embd_enc.data(), 0, ggml_nbytes(lctx.inp_embd_enc));
-    }
-
-    if (!lctx.is_encoding && lctx.inp_KQ_mask_cross) {
-        const int64_t n_output_enc = lctx.embd_enc.size() / hparams.n_embd;
-        const int64_t n_tokens = ubatch.n_tokens;
-
-        GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_KQ_mask_cross->buffer));
-        GGML_ASSERT(!ubatch.equal_seqs); // TODO: use ubatch.n_seqs instead of failing
-
-        float * data = (float *) lctx.inp_KQ_mask_cross->data;
-
-        for (int h = 0; h < 1; ++h) {
-            for (int j = 0; j < n_tokens; ++j) {
-                for (int i = 0; i < n_output_enc; ++i) {
-                    float f = -INFINITY;
-                    for (int s = 0; s < ubatch.n_seq_id[j]; ++s) {
-                        const llama_seq_id seq_id = ubatch.seq_id[j][s];
-                        if (lctx.seq_ids_enc[i].find(seq_id) != lctx.seq_ids_enc[i].end()) {
-                            f = 0.0f;
-                        }
-                    }
-                    data[h*(n_output_enc*n_tokens) + j*n_output_enc + i] = f;
-                }
-            }
-
-            for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) {
-                for (int j = 0; j < n_output_enc; ++j) {
-                    data[h*(n_output_enc*n_tokens) + i*n_output_enc + j] = -INFINITY;
-                }
-            }
-        }
-    }
-}
-
-// Make sure enough space is available for outputs.
-// Returns max number of outputs for which space was reserved.
-static size_t llama_output_reserve(llama_context & lctx, size_t n_outputs) {
-    const auto & cparams = lctx.cparams;
-    const auto & hparams = lctx.model.hparams;
-
-    const size_t n_outputs_max = std::max(n_outputs, (size_t) cparams.n_seq_max);
-
-    const auto n_batch = cparams.n_batch;
-    const auto n_vocab = hparams.n_vocab;
-    const auto n_embd  = hparams.n_embd;
-
-    // TODO: use a per-batch flag for logits presence instead
-    const bool has_logits = !cparams.embeddings;
-    const bool has_embd   =  cparams.embeddings && (cparams.pooling_type == LLAMA_POOLING_TYPE_NONE);
-
-    const size_t logits_size = has_logits ? n_vocab*n_outputs_max : 0;
-    const size_t embd_size   = has_embd   ?  n_embd*n_outputs_max : 0;
-
-    if (lctx.output_ids.empty()) {
-        // init, never resized afterwards
-        lctx.output_ids.resize(n_batch);
-    }
-
-    const size_t prev_size = lctx.buf_output ? ggml_backend_buffer_get_size(lctx.buf_output.get()) : 0;
-    const size_t new_size  = (logits_size + embd_size) * sizeof(float);
-
-    // alloc only when more than the current capacity is required
-    // TODO: also consider shrinking the buffer
-    if (!lctx.buf_output || prev_size < new_size) {
-        if (lctx.buf_output) {
-#ifndef NDEBUG
-            // This doesn't happen often, but may be annoying in some cases (like the HellaSwag benchmark)
-            LLAMA_LOG_INFO("%s: reallocating output buffer from size %.02f MiB to %.02f MiB\n", __func__, prev_size / 1024.0 / 1024.0, new_size / 1024.0 / 1024.0);
-#endif
-            lctx.buf_output = nullptr;
-            lctx.logits = nullptr;
-            lctx.embd = nullptr;
-        }
-
-        auto * buft = ggml_backend_cpu_buffer_type();
-        // try to use the host buffer of the device where the output tensor is allocated for faster transfer to system memory
-        auto * output_dev = lctx.model.dev_output.dev;
-        auto * output_dev_host_buft = output_dev ? ggml_backend_dev_host_buffer_type(output_dev) : nullptr;
-        if (output_dev_host_buft) {
-            buft = output_dev_host_buft;
-        }
-        lctx.buf_output.reset(ggml_backend_buft_alloc_buffer(buft, new_size));
-        if (lctx.buf_output == nullptr) {
-            LLAMA_LOG_ERROR("%s: failed to allocate output buffer of size %.2f MiB\n", __func__, new_size / (1024.0 * 1024.0));
-            return 0;
-        }
-    }
-
-    float * output_base = (float *) ggml_backend_buffer_get_base(lctx.buf_output.get());
-
-    lctx.logits = has_logits ? output_base               : nullptr;
-    lctx.embd   = has_embd   ? output_base + logits_size : nullptr;
-
-    lctx.output_size = n_outputs_max;
-    lctx.logits_size = logits_size;
-    lctx.embd_size   = embd_size;
-
-    // set all ids as invalid (negative)
-    std::fill(lctx.output_ids.begin(), lctx.output_ids.end(), -1);
-
-    ggml_backend_buffer_clear(lctx.buf_output.get(), 0);
-
-    lctx.n_outputs = 0;
-
-    return n_outputs_max;
-}
-
-// make the outputs have the same order they had in the user-provided batch
-static void llama_output_reorder(struct llama_context * ctx) {
-    std::vector & out_ids = ctx->sbatch.out_ids;
-    if (!out_ids.empty()) {
-        uint32_t n_vocab = ctx->model.hparams.n_vocab;
-        uint32_t n_embd  = ctx->model.hparams.n_embd;
-        int32_t n_outputs = ctx->n_outputs;
-        GGML_ASSERT((size_t) n_outputs == out_ids.size());
-        // TODO: is there something more efficient which also minimizes swaps?
-        // selection sort, to minimize swaps (from https://en.wikipedia.org/wiki/Selection_sort)
-        for (int32_t i = 0; i < n_outputs - 1; ++i) {
-            int32_t j_min = i;
-            for (int32_t j = i + 1; j < n_outputs; ++j) {
-                if (out_ids[j] < out_ids[j_min]) {
-                    j_min = j;
-                }
-            }
-            if (j_min == i) { continue; }
-            std::swap(out_ids[i], out_ids[j_min]);
-            if (ctx->logits_size > 0) {
-                for (uint32_t k = 0; k < n_vocab; k++) {
-                    std::swap(ctx->logits[i*n_vocab + k], ctx->logits[j_min*n_vocab + k]);
-                }
-            }
-            if (ctx->embd_size > 0) {
-                for (uint32_t k = 0; k < n_embd; k++) {
-                    std::swap(ctx->embd[i*n_embd + k], ctx->embd[j_min*n_embd + k]);
-                }
-            }
-        }
-        std::fill(ctx->output_ids.begin(), ctx->output_ids.end(), -1);
-        for (int32_t i = 0; i < n_outputs; ++i) {
-            ctx->output_ids[out_ids[i]] = i;
-        }
-        out_ids.clear();
-    }
-}
-
-static void llama_graph_compute(
+// returns the result of ggml_backend_sched_graph_compute_async execution
+static enum ggml_status llama_graph_compute(
           llama_context & lctx,
             ggml_cgraph * gf,
                     int   n_threads,
         ggml_threadpool * threadpool) {
     if (lctx.backend_cpu != nullptr) {
-        ggml_backend_cpu_set_threadpool(lctx.backend_cpu, threadpool);
-        ggml_backend_cpu_set_abort_callback(lctx.backend_cpu, lctx.abort_callback, lctx.abort_callback_data);
+        auto * reg = ggml_backend_dev_backend_reg(ggml_backend_get_device(lctx.backend_cpu));
+        auto * set_threadpool_fn = (decltype(ggml_backend_cpu_set_threadpool) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_set_threadpool");
+        set_threadpool_fn(lctx.backend_cpu, threadpool);
     }
 
     // set the number of threads for all the backends
@@ -17193,15 +8422,20 @@ static void llama_graph_compute(
         set_n_threads_fn.second(set_n_threads_fn.first, n_threads);
     }
 
-    auto err = ggml_backend_sched_graph_compute_async(lctx.sched.get(), gf);
-    if (err != GGML_STATUS_SUCCESS) {
-        LLAMA_LOG_ERROR("%s: ggml_backend_sched_graph_compute_async failed with error %d\n", __func__, err);
+    auto status = ggml_backend_sched_graph_compute_async(lctx.sched.get(), gf);
+    if (status != GGML_STATUS_SUCCESS) {
+        LLAMA_LOG_ERROR("%s: ggml_backend_sched_graph_compute_async failed with error %d\n", __func__, status);
     }
 
     // fprintf(stderr, "splits: %d\n", ggml_backend_sched_get_n_splits(lctx.sched));
+
+    return status;
 }
 
 // decode a batch of tokens by evaluating the transformer
+// in case of unsuccessful decoding (error or warning),
+// the kv_cache state will be returned to its original state
+// (for non-recurrent models) or cleaned (for recurrent models)
 //
 //   - lctx:      llama context
 //   - batch:     batch to evaluate
@@ -17210,7 +8444,7 @@ static void llama_graph_compute(
 // return positive int on warning
 // return negative int on error
 //
-static int llama_decode_internal(
+static int llama_decode_impl(
          llama_context & lctx,
            llama_batch   inp_batch) {
 
@@ -17222,11 +8456,13 @@ static int llama_decode_internal(
     }
 
     // temporary allocate memory for the input batch if needed
-    llama_batch_allocr batch_allocr(lctx, inp_batch);
+    llama_batch_allocr batch_allocr(inp_batch, inp_batch.pos ? -1 : lctx.kv_self.max_pos() + 1);
+
     const llama_batch & batch = batch_allocr.batch;
     const uint32_t n_tokens_all = batch.n_tokens;
 
     const auto & model   = lctx.model;
+    const auto & vocab   = model.vocab;
     const auto & hparams = model.hparams;
     const auto & cparams = lctx.cparams;
 
@@ -17234,7 +8470,7 @@ static int llama_decode_internal(
 
     if (batch.token) {
         for (uint32_t i = 0; i < n_tokens_all; ++i) {
-            if (batch.token[i] < 0 || (uint32_t)batch.token[i] >= model.vocab.n_vocab) {
+            if (batch.token[i] < 0 || (uint32_t) batch.token[i] >= model.vocab.n_tokens()) {
                 LLAMA_LOG_ERROR("%s: invalid token[%d] = %d\n", __func__, i, batch.token[i]);
                 return -1;
             }
@@ -17251,9 +8487,10 @@ static int llama_decode_internal(
     lctx.n_queued_tokens += n_tokens_all;
 
     auto & kv_self = lctx.kv_self;
+    llama_kv_slot_restorer kv_slot_restorer(kv_self);
 
     const int64_t n_embd  = hparams.n_embd;
-    const int64_t n_vocab = hparams.n_vocab;
+    const int64_t n_vocab = vocab.n_tokens();
 
     uint32_t n_outputs = 0;
     uint32_t n_outputs_prev = 0;
@@ -17335,9 +8572,11 @@ static int llama_decode_internal(
                 kv_self.head = 0;
             }
 
-            if (!llama_kv_cache_find_slot(kv_self, ubatch)) {
+            const auto slot = llama_kv_cache_find_slot(kv_self, ubatch);
+            if (!slot) {
                 return 1;
             }
+            kv_slot_restorer.save(slot);
 
             if (!kv_self.recurrent) {
                 // a heuristic, to avoid attending the full cache if it is not yet utilized
@@ -17378,13 +8617,26 @@ static int llama_decode_internal(
             embd = nullptr; // do not extract embeddings when not needed
             GGML_ASSERT(strcmp(res->name, "result_output") == 0 && "missing result_output tensor");
         }
+
         // LLAMA_LOG_INFO("graph build time: %.3f ms (%d nodes, %d leafs)\n", (ggml_time_us() - t_start_us)/1000.0, gf->n_nodes, gf->n_leafs);
 
         ggml_backend_sched_alloc_graph(lctx.sched.get(), gf);
 
         llama_set_inputs(lctx, ubatch);
 
-        llama_graph_compute(lctx, gf, n_threads, threadpool);
+        const auto compute_status = llama_graph_compute(lctx, gf, n_threads, threadpool);
+        if (compute_status != GGML_STATUS_SUCCESS) {
+            kv_slot_restorer.restore(kv_self);
+            switch (compute_status) {
+                case GGML_STATUS_ABORTED:
+                    return 2;
+                case GGML_STATUS_ALLOC_FAILED:
+                    return -2;
+                case GGML_STATUS_FAILED:
+                default:
+                    return -3;
+            }
+        }
 
         // update the kv ring buffer
         {
@@ -17528,7 +8780,7 @@ static int llama_decode_internal(
 // return positive int on warning
 // return negative int on error
 //
-static int llama_encode_internal(
+static int llama_encode_impl(
          llama_context & lctx,
            llama_batch   inp_batch) {
 
@@ -17540,7 +8792,8 @@ static int llama_encode_internal(
     }
 
     // temporary allocate memory for the input batch if needed
-    llama_batch_allocr batch_allocr(lctx, inp_batch);
+    llama_batch_allocr batch_allocr(inp_batch, inp_batch.pos ? -1 : lctx.kv_self.max_pos() + 1);
+
     const llama_batch & batch = batch_allocr.batch;
     const uint32_t n_tokens = batch.n_tokens;
 
@@ -17552,7 +8805,7 @@ static int llama_encode_internal(
 
     if (batch.token) {
         for (uint32_t i = 0; i < n_tokens; ++i) {
-            if (batch.token[i] < 0 || (uint32_t)batch.token[i] >= model.vocab.n_vocab) {
+            if (batch.token[i] < 0 || (uint32_t) batch.token[i] >= model.vocab.n_tokens()) {
                 LLAMA_LOG_ERROR("%s: invalid token[%d] = %d\n", __func__, i, batch.token[i]);
                 return -1;
             }
@@ -17621,7 +8874,18 @@ static int llama_encode_internal(
 
     llama_set_inputs(lctx, ubatch);
 
-    llama_graph_compute(lctx, gf, n_threads, threadpool);
+    const auto compute_status = llama_graph_compute(lctx, gf, n_threads, threadpool);
+    switch (compute_status) {
+        case GGML_STATUS_SUCCESS:
+            break;
+        case GGML_STATUS_ABORTED:
+            return 2;
+        case GGML_STATUS_ALLOC_FAILED:
+            return -2;
+        case GGML_STATUS_FAILED:
+        default:
+            return -3;
+    }
 
     // extract embeddings
     if (embd) {
@@ -17698,7 +8962,7 @@ static int llama_encode_internal(
 }
 
 // find holes from the beginning of the KV cache and fill them by moving data from the end of the cache
-static void llama_kv_cache_defrag_internal(struct llama_context & lctx) {
+static void llama_kv_cache_defrag_impl(struct llama_context & lctx) {
     auto & kv_self = lctx.kv_self;
 
     const auto & hparams = lctx.model.hparams;
@@ -17718,9 +8982,9 @@ static void llama_kv_cache_defrag_internal(struct llama_context & lctx) {
     // each move requires 6*n_layer tensors (see build_defrag)
     //   - source view, destination view, copy operation
     //   - x2 for keys and values
-    //const uint32_t max_moves = llama_model_max_nodes(model)/(6*n_layer);
+    //const uint32_t max_moves = model.max_nodes()/(6*n_layer);
     // TODO: tmp fix https://github.com/ggerganov/llama.cpp/issues/6685#issuecomment-2057579516
-    const uint32_t max_moves = (llama_model_max_nodes(lctx.model) - 2*n_layer)/(6*n_layer);
+    const uint32_t max_moves = (lctx.model.max_nodes() - 2*n_layer)/(6*n_layer);
 
     // determine which KV cells to move where
     //
@@ -17918,16 +9182,16 @@ static void llama_kv_cache_defrag_internal(struct llama_context & lctx) {
     //LLAMA_LOG_INFO("(tmp log) KV defrag time: %.3f ms\n", (t_end - t_start)/1000.0);
 }
 
-static void llama_kv_cache_update_internal(struct llama_context & lctx) {
+static void llama_kv_cache_update_impl(struct llama_context & lctx) {
     bool need_reserve = false;
 
-    // apply K-shift if needed
-    if (lctx.model.hparams.rope_type != LLAMA_ROPE_TYPE_NONE && lctx.kv_self.has_shift) {
-        if (lctx.model.arch == LLM_ARCH_DEEPSEEK2) { // not supported due to MLA
-            GGML_ABORT("Deepseek2 does not support K-shift");
+    if (lctx.kv_self.has_shift) {
+        if (!llama_kv_cache_can_shift(&lctx)) {
+            GGML_ABORT("The current context does not support K-shift");
         }
 
-        {
+        // apply K-shift if needed
+        if (lctx.model.hparams.rope_type != LLAMA_ROPE_TYPE_NONE) {
             ggml_backend_sched_reset(lctx.sched.get());
 
             ggml_cgraph * gf = llama_build_graph_k_shift(lctx);
@@ -17954,7 +9218,7 @@ static void llama_kv_cache_update_internal(struct llama_context & lctx) {
 
     // defragment the KV cache if needed
     if (lctx.kv_self.do_defrag) {
-        llama_kv_cache_defrag_internal(lctx);
+        llama_kv_cache_defrag_impl(lctx);
 
         need_reserve = true;
 
@@ -17967,7 +9231,7 @@ static void llama_kv_cache_update_internal(struct llama_context & lctx) {
         // build worst-case graph
         uint32_t n_seqs = 1; // TODO: worst-case number of sequences
         uint32_t n_tokens = std::min(lctx.cparams.n_ctx, lctx.cparams.n_ubatch);
-        llama_token token = llama_token_bos(&lctx.model); // not actually used by llama_build_graph, but required to choose between token and embedding inputs graph
+        llama_token token = lctx.model.vocab.token_bos(); // not actually used by llama_build_graph, but required to choose between token and embedding inputs graph
         llama_ubatch ubatch = { true, n_tokens, n_tokens / n_seqs, n_seqs, &token, nullptr, nullptr, nullptr, nullptr, nullptr};
         ggml_cgraph * gf = llama_build_graph(lctx, ubatch, true);
 
@@ -17979,1118 +9243,43 @@ static void llama_kv_cache_update_internal(struct llama_context & lctx) {
     }
 }
 
-//
-// quantization
-//
-
-struct quantize_state_internal {
-    const llama_model                 & model;
-    const llama_model_quantize_params * params;
-
-    int n_attention_wv    = 0;
-    int n_ffn_down        = 0;
-    int n_ffn_gate        = 0;
-    int n_ffn_up          = 0;
-    int i_attention_wv    = 0;
-    int i_ffn_down        = 0;
-    int i_ffn_gate        = 0;
-    int i_ffn_up          = 0;
-
-    int n_k_quantized     = 0;
-    int n_fallback        = 0;
-
-    bool has_imatrix      = false;
-
-    // used to figure out if a model shares tok_embd with the output weight
-    bool has_output       = false;
-
-    quantize_state_internal(const llama_model & model, const llama_model_quantize_params * params)
-        : model(model)
-        , params(params)
-        {}
-};
-
-static void llama_tensor_dequantize_internal(
-    struct ggml_tensor * tensor, std::vector> & output, std::vector & workers,
-    const size_t nelements, const int nthread
-) {
-    if (output.size() < nelements) {
-        output.resize(nelements);
-    }
-    float * f32_output = (float *) output.data();
-
-    const ggml_type_traits * qtype = ggml_get_type_traits(tensor->type);
-    if (ggml_is_quantized(tensor->type)) {
-        if (qtype->to_float == NULL) {
-            throw std::runtime_error(format("type %s unsupported for integer quantization: no dequantization available", ggml_type_name(tensor->type)));
-        }
-    } else if (tensor->type != GGML_TYPE_F16 &&
-               tensor->type != GGML_TYPE_BF16) {
-        throw std::runtime_error(format("cannot dequantize/convert tensor type %s", ggml_type_name(tensor->type)));
-    }
-
-    if (nthread < 2) {
-        if (tensor->type == GGML_TYPE_F16) {
-            ggml_fp16_to_fp32_row((ggml_fp16_t *)tensor->data, f32_output, nelements);
-        } else if (tensor->type == GGML_TYPE_BF16) {
-            ggml_bf16_to_fp32_row((ggml_bf16_t *)tensor->data, f32_output, nelements);
-        } else if (ggml_is_quantized(tensor->type)) {
-            qtype->to_float(tensor->data, f32_output, nelements);
-        } else {
-            GGML_ABORT("fatal error"); // unreachable
-        }
-        return;
-    }
-
-    size_t block_size;
-    if (tensor->type == GGML_TYPE_F16 ||
-        tensor->type == GGML_TYPE_BF16) {
-        block_size = 1;
-    } else {
-        block_size = (size_t)ggml_blck_size(tensor->type);
-    }
-
-    size_t block_size_bytes = ggml_type_size(tensor->type);
-
-    GGML_ASSERT(nelements % block_size == 0);
-    size_t nblocks = nelements / block_size;
-    size_t blocks_per_thread = nblocks / nthread;
-    size_t spare_blocks = nblocks - (blocks_per_thread * nthread); // if blocks aren't divisible by thread count
-
-    size_t in_buff_offs = 0;
-    size_t out_buff_offs = 0;
-
-    for (int tnum = 0; tnum < nthread; tnum++) {
-        size_t thr_blocks = blocks_per_thread + (tnum == nthread - 1 ? spare_blocks : 0); // num blocks for this thread
-        size_t thr_elems = thr_blocks * block_size; // number of elements for this thread
-        size_t thr_block_bytes = thr_blocks * block_size_bytes; // number of input bytes for this thread
-
-        auto compute = [qtype] (ggml_type typ, uint8_t * inbuf, float * outbuf, int nels) {
-            if (typ == GGML_TYPE_F16) {
-                ggml_fp16_to_fp32_row((ggml_fp16_t *)inbuf, outbuf, nels);
-            } else if (typ == GGML_TYPE_BF16) {
-                ggml_bf16_to_fp32_row((ggml_bf16_t *)inbuf, outbuf, nels);
-            } else {
-                qtype->to_float(inbuf, outbuf, nels);
-            }
-        };
-        workers.emplace_back(compute, tensor->type, (uint8_t *) tensor->data + in_buff_offs, f32_output + out_buff_offs, thr_elems);
-        in_buff_offs += thr_block_bytes;
-        out_buff_offs += thr_elems;
-    }
-    for (auto & w : workers) { w.join(); }
-    workers.clear();
-}
-
-static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type new_type, const ggml_tensor * tensor, llama_ftype ftype) {
-    const std::string name = ggml_get_name(tensor);
-
-    // TODO: avoid hardcoded tensor names - use the TN_* constants
-    const llm_arch arch = qs.model.arch;
-    const auto       tn = LLM_TN(arch);
-
-    auto use_more_bits = [](int i_layer, int n_layers) -> bool {
-        return i_layer < n_layers/8 || i_layer >= 7*n_layers/8 || (i_layer - n_layers/8)%3 == 2;
-    };
-    const int n_expert = std::max(1, (int)qs.model.hparams.n_expert);
-    auto layer_info = [n_expert] (int i_layer, int n_layer, const char * name) {
-        if (n_expert > 1) {
-            // Believe it or not, "experts" in the FFN of Mixtral-8x7B are not consecutive, but occasionally randomly
-            // sprinkled in the model. Hence, simply dividing i_ffn_down by n_expert does not work
-            // for getting the current layer as I initially thought, and we need to resort to parsing the
-            // tensor name.
-            if (sscanf(name, "blk.%d.", &i_layer) != 1) {
-                throw std::runtime_error(format("Failed to determine layer for tensor %s", name));
-            }
-            if (i_layer < 0 || i_layer >= n_layer) {
-                throw std::runtime_error(format("Bad layer %d for tensor %s. Must be in [0, %d)", i_layer, name, n_layer));
-            }
-        }
-        return std::make_pair(i_layer, n_layer);
-    };
-
-    // for arches that share the same tensor between the token embeddings and the output, we quantize the token embeddings
-    // with the quantization of the output tensor
-    if (name == tn(LLM_TENSOR_OUTPUT, "weight") || (!qs.has_output && name == tn(LLM_TENSOR_TOKEN_EMBD, "weight"))) {
-        if (qs.params->output_tensor_type < GGML_TYPE_COUNT) {
-            new_type = qs.params->output_tensor_type;
-        } else {
-            int nx = tensor->ne[0];
-            if (arch == LLM_ARCH_FALCON || nx % QK_K != 0) {
-                new_type = GGML_TYPE_Q8_0;
-            }
-            else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS ||
-                     ftype == LLAMA_FTYPE_MOSTLY_IQ1_S   || ftype == LLAMA_FTYPE_MOSTLY_IQ2_S  || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M   ||
-                     ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) {
-                new_type = GGML_TYPE_Q5_K;
-            }
-            else if (new_type != GGML_TYPE_Q8_0) {
-                new_type = GGML_TYPE_Q6_K;
-            }
-        }
-    } else if (name == "token_embd.weight") {
-        if (qs.params->token_embedding_type < GGML_TYPE_COUNT) {
-            new_type = qs.params->token_embedding_type;
-        } else {
-            if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS ||
-                ftype == LLAMA_FTYPE_MOSTLY_IQ1_S   || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) {
-                new_type = GGML_TYPE_Q2_K;
-            }
-            else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M) {
-                new_type = GGML_TYPE_IQ3_S;
-            }
-            else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) {
-                new_type = GGML_TYPE_IQ3_S;
-            }
-            else if (new_type == GGML_TYPE_Q4_0_4_4 || new_type == GGML_TYPE_Q4_0_4_8 ||
-                     new_type == GGML_TYPE_Q4_0_8_8) {
-                new_type = GGML_TYPE_Q4_0;
-            }
-            else if (ftype == LLAMA_FTYPE_MOSTLY_TQ1_0 || ftype == LLAMA_FTYPE_MOSTLY_TQ2_0) {
-                new_type = GGML_TYPE_Q4_K;
-            }
-        }
-    } else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ1_S ||
-               ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M    || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) {
-        if (name.find("attn_v.weight") != std::string::npos) {
-            if (qs.model.hparams.n_gqa() >= 4 || qs.model.hparams.n_expert >= 4) new_type = GGML_TYPE_Q4_K;
-            else new_type = ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M ? GGML_TYPE_IQ3_S : GGML_TYPE_Q2_K;
-            ++qs.i_attention_wv;
-        }
-        else if (qs.model.hparams.n_expert == 8 && name.find("attn_k.weight") != std::string::npos) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (name.find("ffn_down") != std::string::npos) {
-            if (qs.i_ffn_down < qs.n_ffn_down/8) {
-                new_type = ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M ? GGML_TYPE_IQ3_S : GGML_TYPE_Q2_K;
-            }
-            ++qs.i_ffn_down;
-        }
-        else if (name.find("attn_output.weight") != std::string::npos) {
-            if (qs.model.hparams.n_expert == 8) {
-                new_type = GGML_TYPE_Q5_K;
-            } else {
-                if (ftype == LLAMA_FTYPE_MOSTLY_IQ1_S || ftype == LLAMA_FTYPE_MOSTLY_IQ1_M) new_type = GGML_TYPE_IQ2_XXS;
-                else if (ftype == LLAMA_FTYPE_MOSTLY_IQ2_S || ftype == LLAMA_FTYPE_MOSTLY_IQ2_M) new_type = GGML_TYPE_IQ3_S;
-            }
-        }
-    } else if (name.find("attn_v.weight") != std::string::npos) {
-        if      (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) {
-            new_type = qs.model.hparams.n_gqa() >= 4 ? GGML_TYPE_Q4_K : GGML_TYPE_Q3_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S && qs.model.hparams.n_gqa() >= 4) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) {
-            new_type = qs.model.hparams.n_gqa() >= 4 ? GGML_TYPE_Q4_K : !qs.has_imatrix ? GGML_TYPE_IQ3_S : GGML_TYPE_IQ3_XXS;
-        }
-        else if ((ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_S) && qs.model.hparams.n_gqa() >= 4) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) {
-            new_type = qs.i_attention_wv < 2 ? GGML_TYPE_Q5_K : GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q5_K;
-        else if ((ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) && qs.model.hparams.n_gqa() >= 4) {
-            new_type = GGML_TYPE_Q5_K;
-        }
-        else if ((ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M || ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M) &&
-                use_more_bits(qs.i_attention_wv, qs.n_attention_wv)) new_type = GGML_TYPE_Q6_K;
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S && qs.i_attention_wv < 4) new_type = GGML_TYPE_Q5_K;
-        if (qs.model.type == MODEL_70B) {
-            // In the 70B model we have 8 heads sharing the same attn_v weights. As a result, the attn_v.weight tensor is
-            // 8x smaller compared to attn_q.weight. Hence, we can get a nice boost in quantization accuracy with
-            // nearly negligible increase in model size by quantizing this tensor with more bits:
-            if (new_type == GGML_TYPE_Q3_K || new_type == GGML_TYPE_Q4_K) new_type = GGML_TYPE_Q5_K;
-        }
-        if (qs.model.hparams.n_expert == 8) {
-            // for the 8-expert model, bumping this to Q8_0 trades just ~128MB
-            // TODO: explore better strategies
-            new_type = GGML_TYPE_Q8_0;
-        }
-        ++qs.i_attention_wv;
-    } else if (name.find("attn_k.weight") != std::string::npos) {
-        if (qs.model.hparams.n_expert == 8) {
-            // for the 8-expert model, bumping this to Q8_0 trades just ~128MB
-            // TODO: explore better strategies
-            new_type = GGML_TYPE_Q8_0;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS) {
-            new_type = GGML_TYPE_IQ3_XXS;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) {
-            new_type = GGML_TYPE_IQ2_S;
-        }
-    } else if (name.find("attn_q.weight") != std::string::npos) {
-        if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS) {
-            new_type = GGML_TYPE_IQ3_XXS;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) {
-            new_type = GGML_TYPE_IQ2_S;
-        }
-    } else if (name.find("ffn_down") != std::string::npos) {
-        auto info = layer_info(qs.i_ffn_down, qs.n_ffn_down, name.c_str());
-        int i_layer = info.first, n_layer = info.second;
-        if      (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K;
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S) {
-            if (i_layer < n_layer/8) new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS && !qs.has_imatrix) {
-            new_type = i_layer < n_layer/8 ? GGML_TYPE_Q4_K : GGML_TYPE_Q3_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M) {
-            new_type = i_layer < n_layer/16 ? GGML_TYPE_Q5_K
-                     : arch != LLM_ARCH_FALCON || use_more_bits(i_layer, n_layer) ? GGML_TYPE_Q4_K
-                     : GGML_TYPE_Q3_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M && (i_layer < n_layer/8 ||
-                    (qs.model.hparams.n_expert == 8 && use_more_bits(i_layer, n_layer)))) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) {
-            new_type = arch == LLM_ARCH_FALCON ? GGML_TYPE_Q4_K : GGML_TYPE_Q5_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M) {
-            if (arch == LLM_ARCH_FALCON) {
-                new_type = i_layer < n_layer/16 ? GGML_TYPE_Q6_K :
-                           use_more_bits(i_layer, n_layer) ? GGML_TYPE_Q5_K : GGML_TYPE_Q4_K;
-            } else {
-                if (use_more_bits(i_layer, n_layer)) new_type = GGML_TYPE_Q6_K;
-            }
-        }
-        else if (i_layer < n_layer/8 && (ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) && !qs.has_imatrix) {
-            new_type = GGML_TYPE_Q5_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M && use_more_bits(i_layer, n_layer)) new_type = GGML_TYPE_Q6_K;
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S && arch != LLM_ARCH_FALCON && i_layer < n_layer/8) {
-            new_type = GGML_TYPE_Q5_K;
-        }
-        else if ((ftype == LLAMA_FTYPE_MOSTLY_Q4_0 || ftype == LLAMA_FTYPE_MOSTLY_Q5_0)
-                && qs.has_imatrix && i_layer < n_layer/8) {
-            // Guard against craziness in the first few ffn_down layers that can happen even with imatrix for Q4_0/Q5_0.
-            // We only do it when an imatrix is provided because a) we want to make sure that one can always get the
-            // same quantization as before imatrix stuff, and b) Q4_1/Q5_1 do go crazy on ffn_down without an imatrix.
-            new_type = ftype == LLAMA_FTYPE_MOSTLY_Q4_0 ? GGML_TYPE_Q4_1 : GGML_TYPE_Q5_1;
-        }
-        ++qs.i_ffn_down;
-    } else if (name.find("attn_output.weight") != std::string::npos) {
-        if (arch != LLM_ARCH_FALCON) {
-            if (qs.model.hparams.n_expert == 8) {
-                if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K   || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS || ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS ||
-                    ftype == LLAMA_FTYPE_MOSTLY_Q3_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M  || ftype == LLAMA_FTYPE_MOSTLY_IQ4_NL  ||
-                    ftype == LLAMA_FTYPE_MOSTLY_Q4_K_S || ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M  || ftype == LLAMA_FTYPE_MOSTLY_IQ3_S  ||
-                    ftype == LLAMA_FTYPE_MOSTLY_IQ3_M  || ftype == LLAMA_FTYPE_MOSTLY_IQ4_XS) {
-                    new_type = GGML_TYPE_Q5_K;
-                }
-            } else {
-                if      (ftype == LLAMA_FTYPE_MOSTLY_Q2_K   ) new_type = GGML_TYPE_Q3_K;
-                else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) new_type = GGML_TYPE_IQ3_S;
-                else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M ) new_type = GGML_TYPE_Q4_K;
-                else if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L ) new_type = GGML_TYPE_Q5_K;
-                else if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_M  ) new_type = GGML_TYPE_Q4_K;
-            }
-        } else {
-            if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L) new_type = GGML_TYPE_Q4_K;
-        }
-    }
-    else if (name.find("attn_qkv.weight") != std::string::npos) {
-        if (ftype == LLAMA_FTYPE_MOSTLY_Q3_K_M || ftype == LLAMA_FTYPE_MOSTLY_Q3_K_L || ftype == LLAMA_FTYPE_MOSTLY_IQ3_M) {
-            new_type = GGML_TYPE_Q4_K;
-        }
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q4_K_M) new_type = GGML_TYPE_Q5_K;
-        else if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_M) new_type = GGML_TYPE_Q6_K;
-    }
-    else if (name.find("ffn_gate") != std::string::npos) {
-        auto info = layer_info(qs.i_ffn_gate, qs.n_ffn_gate, name.c_str());
-        int i_layer = info.first, n_layer = info.second;
-        if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS && (i_layer >= n_layer/8 && i_layer < 7*n_layer/8)) {
-            new_type = GGML_TYPE_IQ3_XXS;
-        }
-        ++qs.i_ffn_gate;
-    }
-    else if (name.find("ffn_up") != std::string::npos) {
-        auto info = layer_info(qs.i_ffn_up, qs.n_ffn_up, name.c_str());
-        int i_layer = info.first, n_layer = info.second;
-        if (ftype == LLAMA_FTYPE_MOSTLY_IQ3_XS && (i_layer >= n_layer/8 && i_layer < 7*n_layer/8)) {
-            new_type = GGML_TYPE_IQ3_XXS;
-        }
-        ++qs.i_ffn_up;
-    }
-
-    //    if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K;
-    //}
-    // IK: let's remove this, else Q2_K is almost the same as Q3_K_S
-    //else if (name.find("ffn_gate") != std::string::npos || name.find("ffn_up") != std::string::npos) {
-    //    if (ftype == LLAMA_FTYPE_MOSTLY_Q2_K) new_type = GGML_TYPE_Q3_K;
-    //}
-    // This can be used to reduce the size of the Q5_K_S model.
-    // The associated PPL increase is fully in line with the size reduction
-    //else {
-    //    if (ftype == LLAMA_FTYPE_MOSTLY_Q5_K_S) new_type = GGML_TYPE_Q4_K;
-    //}
-    bool convert_incompatible_tensor = false;
-    if (new_type == GGML_TYPE_Q2_K    || new_type == GGML_TYPE_Q3_K    || new_type == GGML_TYPE_Q4_K   ||
-        new_type == GGML_TYPE_Q5_K    || new_type == GGML_TYPE_Q6_K    || new_type == GGML_TYPE_IQ4_XS ||
-        new_type == GGML_TYPE_IQ2_XS  || new_type == GGML_TYPE_IQ2_XXS || new_type == GGML_TYPE_IQ2_S  ||
-        new_type == GGML_TYPE_IQ3_XXS || new_type == GGML_TYPE_IQ1_S   || new_type == GGML_TYPE_IQ3_S  ||
-        new_type == GGML_TYPE_IQ1_M) {
-        int nx = tensor->ne[0];
-        int ny = tensor->ne[1];
-        if (nx % QK_K != 0) {
-            LLAMA_LOG_WARN("\n\n%s : tensor cols %d x %d are not divisible by %d, required for %s", __func__, nx, ny, QK_K, ggml_type_name(new_type));
-            convert_incompatible_tensor = true;
-        } else {
-            ++qs.n_k_quantized;
-        }
-    }
-    if (convert_incompatible_tensor) {
-        switch (new_type) {
-            case GGML_TYPE_TQ1_0:
-            case GGML_TYPE_TQ2_0:  new_type = GGML_TYPE_Q4_0; break;  // TODO: use a symmetric type instead
-            case GGML_TYPE_IQ2_XXS:
-            case GGML_TYPE_IQ2_XS:
-            case GGML_TYPE_IQ2_S:
-            case GGML_TYPE_IQ3_XXS:
-            case GGML_TYPE_IQ3_S:
-            case GGML_TYPE_IQ1_S:
-            case GGML_TYPE_IQ1_M:
-            case GGML_TYPE_Q2_K:
-            case GGML_TYPE_Q3_K:
-            case GGML_TYPE_IQ4_XS: new_type = GGML_TYPE_IQ4_NL; break;
-            case GGML_TYPE_Q4_K:   new_type = GGML_TYPE_Q5_0;   break;
-            case GGML_TYPE_Q5_K:   new_type = GGML_TYPE_Q5_1;   break;
-            case GGML_TYPE_Q6_K:   new_type = GGML_TYPE_Q8_0;   break;
-            default: throw std::runtime_error("\nUnsupported tensor size encountered\n");
-        }
-        if (tensor->ne[0] % ggml_blck_size(new_type) != 0) {
-            new_type = GGML_TYPE_F16;
-        }
-        LLAMA_LOG_WARN(" - using fallback quantization %s\n", ggml_type_name(new_type));
-        ++qs.n_fallback;
-    }
-
-    return new_type;
-}
-
-static size_t llama_tensor_quantize_internal(enum ggml_type new_type, const float * f32_data, void * new_data, const int64_t chunk_size, int64_t nrows, int64_t n_per_row, const float * imatrix, std::vector & workers, const int nthread) {
-    if (nthread < 2) {
-        // single-thread
-        size_t new_size = ggml_quantize_chunk(new_type, f32_data, new_data, 0, nrows, n_per_row, imatrix);
-        if (!ggml_validate_row_data(new_type, new_data, new_size)) {
-            throw std::runtime_error("quantized data validation failed");
-        }
-        return new_size;
-    }
-
-    std::mutex mutex;
-    int64_t counter = 0;
-    size_t new_size = 0;
-    bool valid = true;
-    auto compute = [&mutex, &counter, &new_size, &valid, new_type, f32_data, new_data, chunk_size,
-            nrows, n_per_row, imatrix]() {
-        const int64_t nrows_per_chunk = chunk_size / n_per_row;
-        size_t local_size = 0;
-        while (true) {
-            std::unique_lock lock(mutex);
-            int64_t first_row = counter; counter += nrows_per_chunk;
-            if (first_row >= nrows) {
-                if (local_size > 0) {
-                    new_size += local_size;
-                }
-                break;
-            }
-            lock.unlock();
-            const int64_t this_nrow = std::min(nrows - first_row, nrows_per_chunk);
-            size_t this_size = ggml_quantize_chunk(new_type, f32_data, new_data, first_row * n_per_row, this_nrow, n_per_row, imatrix);
-            local_size += this_size;
-
-            // validate the quantized data
-            const size_t row_size  = ggml_row_size(new_type, n_per_row);
-            void * this_data = (char *) new_data + first_row * row_size;
-            if (!ggml_validate_row_data(new_type, this_data, this_size)) {
-                std::unique_lock lock(mutex);
-                valid = false;
-                break;
-            }
-        }
-    };
-    for (int it = 0; it < nthread - 1; ++it) {
-        workers.emplace_back(compute);
-    }
-    compute();
-    for (auto & w : workers) { w.join(); }
-    workers.clear();
-    if (!valid) {
-        throw std::runtime_error("quantized data validation failed");
-    }
-    return new_size;
-}
-
-static void llama_model_quantize_internal(const std::string & fname_inp, const std::string & fname_out, const llama_model_quantize_params * params) {
-    ggml_type default_type;
-    llama_ftype ftype = params->ftype;
-
-    switch (params->ftype) {
-        case LLAMA_FTYPE_MOSTLY_Q4_0: default_type = GGML_TYPE_Q4_0; break;
-        case LLAMA_FTYPE_MOSTLY_Q4_1: default_type = GGML_TYPE_Q4_1; break;
-        case LLAMA_FTYPE_MOSTLY_Q5_0: default_type = GGML_TYPE_Q5_0; break;
-        case LLAMA_FTYPE_MOSTLY_Q5_1: default_type = GGML_TYPE_Q5_1; break;
-        case LLAMA_FTYPE_MOSTLY_Q8_0: default_type = GGML_TYPE_Q8_0; break;
-        case LLAMA_FTYPE_MOSTLY_F16:  default_type = GGML_TYPE_F16;  break;
-        case LLAMA_FTYPE_MOSTLY_BF16: default_type = GGML_TYPE_BF16; break;
-        case LLAMA_FTYPE_ALL_F32:     default_type = GGML_TYPE_F32;  break;
-
-        // K-quants
-        case LLAMA_FTYPE_MOSTLY_Q2_K_S:
-        case LLAMA_FTYPE_MOSTLY_Q2_K:    default_type = GGML_TYPE_Q2_K;    break;
-        case LLAMA_FTYPE_MOSTLY_IQ3_XS:  default_type = GGML_TYPE_IQ3_S;   break;
-        case LLAMA_FTYPE_MOSTLY_Q3_K_S:
-        case LLAMA_FTYPE_MOSTLY_Q3_K_M:
-        case LLAMA_FTYPE_MOSTLY_Q3_K_L:  default_type = GGML_TYPE_Q3_K;    break;
-        case LLAMA_FTYPE_MOSTLY_Q4_K_S:
-        case LLAMA_FTYPE_MOSTLY_Q4_K_M:  default_type = GGML_TYPE_Q4_K;    break;
-        case LLAMA_FTYPE_MOSTLY_Q5_K_S:
-        case LLAMA_FTYPE_MOSTLY_Q5_K_M:  default_type = GGML_TYPE_Q5_K;    break;
-        case LLAMA_FTYPE_MOSTLY_Q6_K:    default_type = GGML_TYPE_Q6_K;    break;
-        case LLAMA_FTYPE_MOSTLY_TQ1_0:   default_type = GGML_TYPE_TQ1_0;   break;
-        case LLAMA_FTYPE_MOSTLY_TQ2_0:   default_type = GGML_TYPE_TQ2_0;   break;
-        case LLAMA_FTYPE_MOSTLY_IQ2_XXS: default_type = GGML_TYPE_IQ2_XXS; break;
-        case LLAMA_FTYPE_MOSTLY_IQ2_XS:  default_type = GGML_TYPE_IQ2_XS;  break;
-        case LLAMA_FTYPE_MOSTLY_IQ2_S:   default_type = GGML_TYPE_IQ2_XS;  break;
-        case LLAMA_FTYPE_MOSTLY_IQ2_M:   default_type = GGML_TYPE_IQ2_S;   break;
-        case LLAMA_FTYPE_MOSTLY_IQ3_XXS: default_type = GGML_TYPE_IQ3_XXS; break;
-        case LLAMA_FTYPE_MOSTLY_IQ1_S:   default_type = GGML_TYPE_IQ1_S;   break;
-        case LLAMA_FTYPE_MOSTLY_IQ1_M:   default_type = GGML_TYPE_IQ1_M;   break;
-        case LLAMA_FTYPE_MOSTLY_IQ4_NL:  default_type = GGML_TYPE_IQ4_NL;  break;
-        case LLAMA_FTYPE_MOSTLY_IQ4_XS:  default_type = GGML_TYPE_IQ4_XS;  break;
-        case LLAMA_FTYPE_MOSTLY_IQ3_S:   default_type = GGML_TYPE_IQ3_S;   break;
-        case LLAMA_FTYPE_MOSTLY_IQ3_M:   default_type = GGML_TYPE_IQ3_S;   break;
-        case LLAMA_FTYPE_MOSTLY_Q4_0_4_4: default_type = GGML_TYPE_Q4_0_4_4; break;
-        case LLAMA_FTYPE_MOSTLY_Q4_0_4_8: default_type = GGML_TYPE_Q4_0_4_8; break;
-        case LLAMA_FTYPE_MOSTLY_Q4_0_8_8: default_type = GGML_TYPE_Q4_0_8_8; break;
-
-        default: throw std::runtime_error(format("invalid output file type %d\n", ftype));
-    }
-
-    int nthread = params->nthread;
-
-    if (nthread <= 0) {
-        nthread = std::thread::hardware_concurrency();
-    }
-
-    // mmap consistently increases speed Linux, and also increases speed on Windows with
-    // hot cache. It may cause a slowdown on macOS, possibly related to free memory.
-#if defined(__linux__) || defined(_WIN32)
-    constexpr bool use_mmap = true;
-#else
-    constexpr bool use_mmap = false;
-#endif
-
-    llama_model_kv_override * kv_overrides = nullptr;
-    if (params->kv_overrides) {
-        auto v = (std::vector*)params->kv_overrides;
-        kv_overrides = v->data();
-    }
-    llama_model_loader ml(fname_inp, use_mmap, /*check_tensors*/ true, kv_overrides);
-    ml.init_mappings(false); // no prefetching
-
-    llama_model model;
-    llm_load_arch(ml, model);
-    llm_load_hparams(ml, model);
-
-    struct quantize_state_internal qs(model, params);
-
-    if (params->only_copy) {
-        ftype = model.ftype;
-    }
-    const std::unordered_map> * imatrix_data = nullptr;
-    if (params->imatrix) {
-        imatrix_data = static_cast>*>(params->imatrix);
-        if (imatrix_data) {
-            LLAMA_LOG_INFO("================================ Have weights data with %d entries\n",int(imatrix_data->size()));
-            qs.has_imatrix = true;
-            // check imatrix for nans or infs
-            for (const auto & kv : *imatrix_data) {
-                for (float f : kv.second) {
-                    if (!std::isfinite(f)) {
-                        throw std::runtime_error(format("imatrix contains non-finite value %f\n", f));
-                    }
-                }
-            }
-        }
-    }
-
-    const size_t align = GGUF_DEFAULT_ALIGNMENT;
-    gguf_context_ptr ctx_out { gguf_init_empty() };
-
-    // copy the KV pairs from the input file
-    gguf_set_kv     (ctx_out.get(), ml.meta.get());
-    gguf_set_val_u32(ctx_out.get(), "general.quantization_version", GGML_QNT_VERSION); // TODO: use LLM_KV
-    gguf_set_val_u32(ctx_out.get(), "general.file_type", ftype); // TODO: use LLM_KV
-
-    // Remove split metadata
-    gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_NO).c_str());
-    gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_COUNT).c_str());
-    gguf_remove_key(ctx_out.get(), ml.llm_kv(LLM_KV_SPLIT_TENSORS_COUNT).c_str());
-
-    if (params->kv_overrides) {
-        const std::vector & overrides = *(const std::vector *)params->kv_overrides;
-        for (const auto & o : overrides) {
-            if (o.key[0] == 0) break;
-            if (o.tag == LLAMA_KV_OVERRIDE_TYPE_FLOAT) {
-                gguf_set_val_f32(ctx_out.get(), o.key, o.val_f64);
-            } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_INT) {
-                gguf_set_val_i32(ctx_out.get(), o.key, o.val_i64);
-            } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_BOOL) {
-                gguf_set_val_bool(ctx_out.get(), o.key, o.val_bool);
-            } else if (o.tag == LLAMA_KV_OVERRIDE_TYPE_STR) {
-                gguf_set_val_str(ctx_out.get(), o.key, o.val_str);
-            } else {
-                LLAMA_LOG_WARN("%s: unknown KV override type for key %s\n", __func__, o.key);
-            }
-        }
-    }
-
-    // make a list of weights
-    std::vector tensors;
-    tensors.reserve(ml.weights_map.size());
-    for (const auto & it : ml.weights_map) {
-        tensors.push_back(&it.second);
-    }
-
-    // keep_split requires that the weights are sorted by split index
-    if (params->keep_split) {
-        std::sort(tensors.begin(), tensors.end(), [](const llama_model_loader::llama_tensor_weight * a, const llama_model_loader::llama_tensor_weight * b) {
-            if (a->idx == b->idx) {
-                return a->offs < b->offs;
-            }
-            return a->idx < b->idx;
-        });
-    }
-
-    for (const auto * it : tensors) {
-        const struct ggml_tensor * tensor = it->tensor;
-
-        const std::string name = ggml_get_name(tensor);
-
-        // TODO: avoid hardcoded tensor names - use the TN_* constants
-        if (name.find("attn_v.weight")   != std::string::npos ||
-            name.find("attn_qkv.weight") != std::string::npos ||
-            name.find("attn_kv_b.weight")!= std::string::npos) {
-            ++qs.n_attention_wv;
-        } else if (name == LLM_TN(model.arch)(LLM_TENSOR_OUTPUT, "weight")) {
-            qs.has_output = true;
-        }
-    }
-
-    qs.n_ffn_down = qs.n_ffn_gate = qs.n_ffn_up = (int)model.hparams.n_layer;
-
-    // sanity checks
-    {
-        const auto & n_head_kv_iter = model.hparams.n_head_kv_arr.begin();
-        // attention layers have a non-zero number of kv heads
-        int32_t n_attn_layer = model.hparams.n_layer - std::count(n_head_kv_iter, n_head_kv_iter + model.hparams.n_layer, 0);
-        if (llama_model_has_encoder(&model)) {
-            n_attn_layer *= 3;
-        }
-        GGML_ASSERT((qs.n_attention_wv == n_attn_layer) && "n_attention_wv is unexpected");
-    }
-
-    size_t total_size_org = 0;
-    size_t total_size_new = 0;
-
-    std::vector workers;
-    workers.reserve(nthread);
-
-    int idx = 0;
-
-    std::vector> read_data;
-    std::vector> work;
-    std::vector> f32_conv_buf;
-
-    uint16_t n_split = 1;
-
-    // Assume split index is continuous
-    if (params->keep_split) {
-        for (const auto * it : tensors) {
-            n_split = std::max(uint16_t(it->idx + 1), n_split);
-        }
-    }
-    std::vector ctx_outs(n_split);
-    ctx_outs[0] = std::move(ctx_out);
-
-    // populate the original tensors so we get an initial meta data
-    for (const auto * it : tensors) {
-        uint16_t i_split = params->keep_split ? it->idx : 0;
-        struct ggml_tensor * tensor = it->tensor;
-        if (!ctx_outs[i_split]) {
-            ctx_outs[i_split].reset(gguf_init_empty());
-        }
-        gguf_add_tensor(ctx_outs[i_split].get(), tensor);
-    }
-
-    // Set split info if needed
-    if (n_split > 1) {
-        for (size_t i = 0; i < ctx_outs.size(); ++i) {
-            gguf_set_val_u16(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_NO).c_str(), i);
-            gguf_set_val_u16(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_COUNT).c_str(), n_split);
-            gguf_set_val_i32(ctx_outs[i].get(), ml.llm_kv(LLM_KV_SPLIT_TENSORS_COUNT).c_str(), ml.n_tensors);
-        }
-    }
-
-    int cur_split = -1;
-    std::ofstream fout;
-    auto close_ofstream = [&]() {
-        // Write metadata and close file handler
-        if (fout.is_open()) {
-            fout.seekp(0);
-            std::vector data(gguf_get_meta_size(ctx_outs[cur_split].get()));
-            gguf_get_meta_data(ctx_outs[cur_split].get(), data.data());
-            fout.write((const char *) data.data(), data.size());
-            fout.close();
-        }
-    };
-    auto new_ofstream = [&](int index) {
-        cur_split = index;
-        GGML_ASSERT(ctx_outs[cur_split] && "Find uninitialized gguf_context");
-        std::string fname = fname_out;
-        if (params->keep_split) {
-            char split_path[PATH_MAX] = {0};
-            llama_split_path(split_path, sizeof(split_path), fname_out.c_str(), cur_split, n_split);
-            fname = std::string(split_path);
-        }
-
-        fout = std::ofstream(fname, std::ios::binary);
-        fout.exceptions(std::ofstream::failbit); // fail fast on write errors
-        const size_t meta_size = gguf_get_meta_size(ctx_outs[cur_split].get());
-        // placeholder for the meta data
-        ::zeros(fout, meta_size);
-    };
-
-    const auto tn = LLM_TN(model.arch);
-    new_ofstream(0);
-    for (const auto * it : tensors) {
-        const auto & weight = *it;
-        struct ggml_tensor * tensor = weight.tensor;
-        if (weight.idx != cur_split && params->keep_split) {
-            close_ofstream();
-            new_ofstream(weight.idx);
-        }
-
-        const std::string name = ggml_get_name(tensor);
-
-        if (!ml.use_mmap) {
-            if (read_data.size() < ggml_nbytes(tensor)) {
-                read_data.resize(ggml_nbytes(tensor));
-            }
-            tensor->data = read_data.data();
-        }
-        ml.load_data_for(tensor);
-
-        LLAMA_LOG_INFO("[%4d/%4d] %36s - [%s], type = %6s, ",
-               ++idx, ml.n_tensors,
-               ggml_get_name(tensor),
-               llama_format_tensor_shape(tensor).c_str(),
-               ggml_type_name(tensor->type));
-
-        // This used to be a regex, but  has an extreme cost to compile times.
-        bool quantize = name.rfind("weight") == name.size() - 6; // ends with 'weight'?
-
-        // quantize only 2D and 3D tensors (experts)
-        quantize &= (ggml_n_dims(tensor) >= 2);
-
-        // do not quantize norm tensors
-        quantize &= name.find("_norm.weight") == std::string::npos;
-
-        quantize &= params->quantize_output_tensor || name != "output.weight";
-        quantize &= !params->only_copy;
-
-        // do not quantize expert gating tensors
-        // NOTE: can't use LLM_TN here because the layer number is not known
-        quantize &= name.find("ffn_gate_inp.weight") == std::string::npos;
-
-        // do not quantize positional embeddings and token types (BERT)
-        quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_POS_EMBD,    "weight");
-        quantize &= name != LLM_TN(model.arch)(LLM_TENSOR_TOKEN_TYPES, "weight");
-
-        // do not quantize Mamba's small yet 2D weights
-        // NOTE: can't use LLM_TN here because the layer number is not known
-        quantize &= name.find("ssm_conv1d.weight") == std::string::npos;
-
-        // do not quantize RWKV's time_mix_first tensors
-        quantize &= name.find("time_mix_first.weight") == std::string::npos;
-        quantize &= name.find("time_mix_w1.weight") == std::string::npos;
-        quantize &= name.find("time_mix_w2.weight") == std::string::npos;
-        quantize &= name.find("time_mix_decay_w1.weight") == std::string::npos;
-        quantize &= name.find("time_mix_decay_w2.weight") == std::string::npos;
-
-        // do not quantize relative position bias (T5)
-        quantize &= name.find("attn_rel_b.weight") == std::string::npos;
-
-        enum ggml_type new_type;
-        void * new_data;
-        size_t new_size;
-
-        if (quantize) {
-            new_type = default_type;
-
-            // get more optimal quantization type based on the tensor shape, layer, etc.
-            if (!params->pure && ggml_is_quantized(default_type)) {
-                new_type = llama_tensor_get_type(qs, new_type, tensor, ftype);
-            }
-            if (params->token_embedding_type < GGML_TYPE_COUNT && strcmp(tensor->name, "token_embd.weight") == 0) {
-                new_type = params->token_embedding_type;
-            }
-            if (params->output_tensor_type < GGML_TYPE_COUNT && strcmp(tensor->name, "output.weight") == 0) {
-                new_type = params->output_tensor_type;
-            }
-
-            // If we've decided to quantize to the same type the tensor is already
-            // in then there's nothing to do.
-            quantize = tensor->type != new_type;
-        }
-
-        if (!quantize) {
-            new_type = tensor->type;
-            new_data = tensor->data;
-            new_size = ggml_nbytes(tensor);
-            LLAMA_LOG_INFO("size = %8.3f MB\n", ggml_nbytes(tensor)/1024.0/1024.0);
-        } else {
-            const int64_t nelements = ggml_nelements(tensor);
-
-            const float * imatrix = nullptr;
-            if (imatrix_data) {
-                auto it = imatrix_data->find(tensor->name);
-                if (it == imatrix_data->end()) {
-                    LLAMA_LOG_INFO("\n====== %s: did not find weights for %s\n", __func__, tensor->name);
-                } else {
-                    if (it->second.size() == (size_t)tensor->ne[0]*tensor->ne[2]) {
-                        imatrix = it->second.data();
-                    } else {
-                        LLAMA_LOG_INFO("\n====== %s: imatrix size %d is different from tensor size %d for %s\n", __func__,
-                                int(it->second.size()), int(tensor->ne[0]*tensor->ne[2]), tensor->name);
-
-                        // this can happen when quantizing an old mixtral model with split tensors with a new incompatible imatrix
-                        // this is a significant error and it may be good idea to abort the process if this happens,
-                        // since many people will miss the error and not realize that most of the model is being quantized without an imatrix
-                        // tok_embd should be ignored in this case, since it always causes this warning
-                        if (name != tn(LLM_TENSOR_TOKEN_EMBD, "weight")) {
-                            throw std::runtime_error(format("imatrix size %d is different from tensor size %d for %s",
-                                    int(it->second.size()), int(tensor->ne[0]*tensor->ne[2]), tensor->name));
-                        }
-                    }
-                }
-            }
-            if ((new_type == GGML_TYPE_IQ2_XXS ||
-                 new_type == GGML_TYPE_IQ2_XS  ||
-                 new_type == GGML_TYPE_IQ2_S   ||
-                 new_type == GGML_TYPE_IQ1_S   ||
-                (new_type == GGML_TYPE_IQ1_M && strcmp(tensor->name, "token_embd.weight") && strcmp(tensor->name, "output.weight"))  ||
-                (new_type == GGML_TYPE_Q2_K && params->ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S && strcmp(tensor->name, "token_embd.weight") != 0)) && !imatrix) {
-                LLAMA_LOG_ERROR("\n\n============================================================\n");
-                LLAMA_LOG_ERROR("Missing importance matrix for tensor %s in a very low-bit quantization\n", tensor->name);
-                LLAMA_LOG_ERROR("The result will be garbage, so bailing out\n");
-                LLAMA_LOG_ERROR("============================================================\n\n");
-                throw std::runtime_error(format("Missing importance matrix for tensor %s in a very low-bit quantization", tensor->name));
-            }
-
-            float * f32_data;
-
-            if (tensor->type == GGML_TYPE_F32) {
-                f32_data = (float *) tensor->data;
-            } else if (ggml_is_quantized(tensor->type) && !params->allow_requantize) {
-                throw std::runtime_error(format("requantizing from type %s is disabled", ggml_type_name(tensor->type)));
-            } else {
-                llama_tensor_dequantize_internal(tensor, f32_conv_buf, workers, nelements, nthread);
-                f32_data = (float *) f32_conv_buf.data();
-            }
-
-            int chunk_size_multiplier = 1;
-            if (new_type == GGML_TYPE_Q4_0_4_4 || new_type == GGML_TYPE_Q4_0_4_8 || new_type == GGML_TYPE_Q4_0_8_8) {
-                if ((new_type == GGML_TYPE_Q4_0_8_8) && (tensor->ne[1] % 8 != 0)) new_type = GGML_TYPE_Q4_0;
-                else if (tensor->ne[1] % 4 != 0) new_type = GGML_TYPE_Q4_0;
-                if (new_type == GGML_TYPE_Q4_0_8_8) chunk_size_multiplier = 8;
-                else if (new_type == GGML_TYPE_Q4_0_4_4 || new_type == GGML_TYPE_Q4_0_4_8) chunk_size_multiplier = 4;
-            }
-
-            LLAMA_LOG_INFO("converting to %s .. ", ggml_type_name(new_type));
-            fflush(stdout);
-
-            if (work.size() < (size_t)nelements * 4) {
-                work.resize(nelements * 4); // upper bound on size
-            }
-            new_data = work.data();
-
-            const int64_t n_per_row = tensor->ne[0];
-            const int64_t nrows = tensor->ne[1];
-
-            static const int64_t min_chunk_size = 32 * 512;
-            const int64_t chunk_size = (n_per_row >= min_chunk_size ? n_per_row : n_per_row * ((min_chunk_size + n_per_row - 1)/n_per_row)) *
-                                       chunk_size_multiplier;
-
-            const int64_t nelements_matrix = tensor->ne[0] * tensor->ne[1];
-            const int64_t nchunk = (nelements_matrix + chunk_size - 1)/chunk_size;
-            const int64_t nthread_use = nthread > 1 ? std::max((int64_t)1, std::min((int64_t)nthread, nchunk)) : 1;
-
-            // quantize each expert separately since they have different importance matrices
-            new_size = 0;
-            for (int64_t i03 = 0; i03 < tensor->ne[2]; ++i03) {
-                const float * f32_data_03 = f32_data + i03 * nelements_matrix;
-                void * new_data_03 = (char *)new_data + ggml_row_size(new_type, n_per_row) * i03 * nrows;
-                const float * imatrix_03 = imatrix ? imatrix + i03 * n_per_row : nullptr;
-
-                new_size += llama_tensor_quantize_internal(new_type, f32_data_03, new_data_03, chunk_size, nrows, n_per_row, imatrix_03, workers, nthread_use);
-            }
-            LLAMA_LOG_INFO("size = %8.2f MiB -> %8.2f MiB\n", ggml_nbytes(tensor)/1024.0/1024.0, new_size/1024.0/1024.0);
-        }
-        total_size_org += ggml_nbytes(tensor);
-        total_size_new += new_size;
-
-        // update the gguf meta data as we go
-        gguf_set_tensor_type(ctx_outs[cur_split].get(), name.c_str(), new_type);
-        gguf_set_tensor_data(ctx_outs[cur_split].get(), name.c_str(), new_data, new_size);
-
-        // write tensor data + padding
-        fout.write((const char *) new_data, new_size);
-        zeros(fout, GGML_PAD(new_size, align) - new_size);
-    }
-    close_ofstream();
-
-    LLAMA_LOG_INFO("%s: model size  = %8.2f MB\n", __func__, total_size_org/1024.0/1024.0);
-    LLAMA_LOG_INFO("%s: quant size  = %8.2f MB\n", __func__, total_size_new/1024.0/1024.0);
-
-    if (qs.n_fallback > 0) {
-        LLAMA_LOG_WARN("%s: WARNING: %d of %d tensor(s) required fallback quantization\n",
-                __func__, qs.n_fallback, qs.n_k_quantized + qs.n_fallback);
-    }
-}
-
-static void llama_lora_adapter_init_internal(struct llama_model * model, const char * path_lora, struct llama_lora_adapter & adapter) {
-    LLAMA_LOG_INFO("%s: loading lora adapter from '%s' ...\n", __func__, path_lora);
-
-    ggml_context * ctx_init;
-    struct gguf_init_params meta_gguf_params = {
-        /* .no_alloc = */ true,
-        /* .ctx      = */ &ctx_init,
-    };
-
-    gguf_context_ptr ctx_gguf { gguf_init_from_file(path_lora, meta_gguf_params) };
-    if (!ctx_gguf) {
-        throw std::runtime_error("failed to load lora adapter file from " + std::string(path_lora));
-    }
-
-    ggml_context_ptr ctx { ctx_init };
-
-    // check metadata
-    {
-        auto get_kv_str = [&](const std::string & key) -> std::string {
-            int id = gguf_find_key(ctx_gguf.get(), key.c_str());
-            return id < 0 ? "" : std::string(gguf_get_val_str(ctx_gguf.get(), id));
-        };
-        auto get_kv_f32 = [&](const std::string & key) -> float {
-            int id = gguf_find_key(ctx_gguf.get(), key.c_str());
-            return id < 0 ? 0.0f : gguf_get_val_f32(ctx_gguf.get(), id);
-        };
-        LLM_KV llm_kv = LLM_KV(LLM_ARCH_UNKNOWN);
-
-        auto general_type = get_kv_str(llm_kv(LLM_KV_GENERAL_TYPE));
-        if (general_type != "adapter") {
-            throw std::runtime_error("expect general.type to be 'adapter', but got: " + general_type);
-        }
-
-        auto general_arch_str = get_kv_str(llm_kv(LLM_KV_GENERAL_ARCHITECTURE));
-        auto general_arch = llm_arch_from_string(general_arch_str);
-        if (general_arch != model->arch) {
-            throw std::runtime_error("model arch and LoRA arch mismatch");
-        }
-
-        auto adapter_type = get_kv_str(llm_kv(LLM_KV_ADAPTER_TYPE));
-        if (adapter_type != "lora") {
-            throw std::runtime_error("expect adapter.type to be 'lora', but got: " + adapter_type);
-        }
-
-        adapter.alpha = get_kv_f32(llm_kv(LLM_KV_ADAPTER_LORA_ALPHA));
-    }
-
-    int n_tensors = gguf_get_n_tensors(ctx_gguf.get());
-
-    // contexts for each buffer type
-    std::map ctx_map;
-    auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * {
-        auto it = ctx_map.find(buft);
-        if (it == ctx_map.end()) {
-            // add a new context
-            struct ggml_init_params params = {
-                /*.mem_size   =*/ n_tensors*ggml_tensor_overhead(),
-                /*.mem_buffer =*/ NULL,
-                /*.no_alloc   =*/ true,
-            };
-            ggml_context * buft_ctx = ggml_init(params);
-            if (!buft_ctx) {
-                return nullptr;
-            }
-            ctx_map[buft] = buft_ctx;
-            adapter.ctxs.emplace_back(buft_ctx);
-            return buft_ctx;
-        };
-        return it->second;
-    };
-
-    // bundle lora_a and lora_b into pairs
-    std::map ab_map;
-    auto str_endswith = [](const std::string & str, const std::string & suffix) {
-        return str.size() >= suffix.size() && str.compare(str.size()-suffix.size(), suffix.size(), suffix) == 0;
-    };
-    for (ggml_tensor * cur = ggml_get_first_tensor(ctx.get()); cur; cur = ggml_get_next_tensor(ctx.get(), cur)) {
-        std::string name(cur->name);
-        if (str_endswith(name, ".lora_a")) {
-            replace_all(name, ".lora_a", "");
-            if (ab_map.find(name) == ab_map.end()) {
-                ab_map[name] = llama_lora_weight(cur, nullptr);
-            } else {
-                ab_map[name].a = cur;
-            }
-        } else if (str_endswith(name, ".lora_b")) {
-            replace_all(name, ".lora_b", "");
-            if (ab_map.find(name) == ab_map.end()) {
-                ab_map[name] = llama_lora_weight(nullptr, cur);
-            } else {
-                ab_map[name].b = cur;
-            }
-        } else {
-            throw std::runtime_error("LoRA tensor '" + name + "' has unexpected suffix");
-        }
-    }
-
-    // add tensors
-    for (auto & it : ab_map) {
-        const std::string & name = it.first;
-        llama_lora_weight & w = it.second;
-
-        if (!w.a || !w.b) {
-            throw std::runtime_error("LoRA tensor pair for '" + name + "' is missing one component");
-        }
-
-        // device buft and device ctx
-        auto * model_tensor = llama_get_model_tensor(model, name.c_str());
-        if (!model_tensor) {
-            throw std::runtime_error("LoRA tensor '" + name + "' does not exist in base model");
-        }
-        struct ggml_context * dev_ctx = ctx_for_buft(ggml_backend_buffer_get_type(model_tensor->buffer));
-        // validate tensor shape
-        if (model_tensor->ne[0] != w.a->ne[0] || model_tensor->ne[1] != w.b->ne[1]) {
-            throw std::runtime_error("tensor '" + name + "' has incorrect shape");
-        }
-        if (w.a->ne[1] != w.b->ne[0]) {
-            throw std::runtime_error("lora_a tensor is not transposed (hint: adapter from \"finetune\" example is no longer supported)");
-        }
-        // save tensor to adapter
-        struct ggml_tensor * tensor_a = ggml_dup_tensor(dev_ctx, w.a);
-        struct ggml_tensor * tensor_b = ggml_dup_tensor(dev_ctx, w.b);
-        ggml_set_name(tensor_a, w.a->name);
-        ggml_set_name(tensor_b, w.b->name);
-        adapter.ab_map[name] = llama_lora_weight(tensor_a, tensor_b);
-    }
-
-    // allocate tensors / buffers and zero
-    {
-        adapter.ctxs.reserve(ctx_map.size());
-        adapter.bufs.reserve(ctx_map.size());
-        for (auto & it : ctx_map) {
-            ggml_backend_buffer_type_t buft = it.first;
-            ggml_context * ctx_dev = it.second;
-            ggml_backend_buffer_ptr buf { ggml_backend_alloc_ctx_tensors_from_buft(ctx_dev, buft) };
-            if (!buf) {
-                throw std::runtime_error("failed to allocate buffer for lora adapter\n");
-            }
-            LLAMA_LOG_INFO("%s: %10s LoRA buffer size = %8.2f MiB\n", __func__, ggml_backend_buffer_name(buf.get()), ggml_backend_buffer_get_size(buf.get())/1024.0/1024.0);
-            adapter.bufs.emplace_back(std::move(buf));
-        }
-    }
-
-    // set tensor data
-    {
-        llama_file gguf_file(path_lora, "rb");
-        std::vector read_buf;
-        auto set_tensor = [&](struct ggml_tensor * orig, struct ggml_tensor * dev) {
-            size_t offs = gguf_get_data_offset(ctx_gguf.get()) + gguf_get_tensor_offset(ctx_gguf.get(), gguf_find_tensor(ctx_gguf.get(), orig->name));
-            size_t size = ggml_nbytes(orig);
-            read_buf.resize(size);
-            gguf_file.seek(offs, SEEK_SET);
-            gguf_file.read_raw(read_buf.data(), size);
-            ggml_backend_tensor_set(dev, read_buf.data(), 0, size);
-        };
-        for (auto & it : adapter.ab_map) {
-            auto orig = ab_map[it.first];
-            auto dev  = it.second;
-            set_tensor(orig.a, dev.a);
-            set_tensor(orig.b, dev.b);
-        }
-    }
-
-    LLAMA_LOG_INFO("%s: loaded %zu tensors from lora file\n", __func__, adapter.ab_map.size()*2);
-}
-
-int32_t llama_lora_adapter_set(
+int32_t llama_set_adapter_lora(
             struct llama_context * ctx,
-            struct llama_lora_adapter * adapter,
+            struct llama_adapter_lora * adapter,
             float scale) {
-    if (ctx->cparams.flash_attn) {
-        LLAMA_LOG_ERROR("%s: flash_attn is not compatible with LoRA\n", __func__);
-        return -1;
-    }
-    ctx->lora_adapters[adapter] = scale;
+    ctx->lora[adapter] = scale;
     return 0;
 }
 
-int32_t llama_lora_adapter_remove(
+int32_t llama_rm_adapter_lora(
             struct llama_context * ctx,
-            struct llama_lora_adapter * adapter) {
-    auto pos = ctx->lora_adapters.find(adapter);
-    if (pos != ctx->lora_adapters.end()) {
-        ctx->lora_adapters.erase(pos);
+            struct llama_adapter_lora * adapter) {
+    auto pos = ctx->lora.find(adapter);
+    if (pos != ctx->lora.end()) {
+        ctx->lora.erase(pos);
         return 0;
     }
+
     return -1;
 }
 
-void llama_lora_adapter_clear(struct llama_context * ctx) {
-    ctx->lora_adapters.clear();
+void llama_clear_adapter_lora(struct llama_context * ctx) {
+    ctx->lora.clear();
 }
 
-void llama_lora_adapter_free(struct llama_lora_adapter * adapter) {
-    delete adapter;
+int32_t llama_apply_adapter_cvec(
+        struct llama_context * ctx,
+                 const float * data,
+                      size_t   len,
+                     int32_t   n_embd,
+                     int32_t   il_start,
+                     int32_t   il_end) {
+    return ctx->cvec.apply(ctx->model, data, len, n_embd, il_start, il_end);
 }
 
 //
 // interface implementation
 //
-struct llama_model_params llama_model_default_params() {
-    struct llama_model_params result = {
-        /*.n_gpu_layers                =*/ 0,
-        /*.split_mode                  =*/ LLAMA_SPLIT_MODE_LAYER,
-        /*.main_gpu                    =*/ 0,
-        /*.tensor_split                =*/ nullptr,
-        /*.rpc_servers                 =*/ nullptr,
-        /*.progress_callback           =*/ nullptr,
-        /*.progress_callback_user_data =*/ nullptr,
-        /*.kv_overrides                =*/ nullptr,
-        /*.vocab_only                  =*/ false,
-        /*.use_mmap                    =*/ true,
-        /*.use_mlock                   =*/ false,
-        /*.check_tensors               =*/ false,
-    };
-
-#ifdef GGML_USE_METAL
-    // note: we usually have plenty of VRAM, so by default offload all layers to the GPU
-    result.n_gpu_layers = 999;
-#endif
-
-    return result;
-}
 
 struct llama_context_params llama_context_default_params() {
     struct llama_context_params result = {
@@ -19135,24 +9324,6 @@ struct llama_sampler_chain_params llama_sampler_chain_default_params() {
     return result;
 }
 
-struct llama_model_quantize_params llama_model_quantize_default_params() {
-    struct llama_model_quantize_params result = {
-        /*.nthread                     =*/ 0,
-        /*.ftype                       =*/ LLAMA_FTYPE_MOSTLY_Q5_1,
-        /*.output_tensor_type          =*/ GGML_TYPE_COUNT,
-        /*.token_embedding_type        =*/ GGML_TYPE_COUNT,
-        /*.allow_requantize            =*/ false,
-        /*.quantize_output_tensor      =*/ true,
-        /*.only_copy                   =*/ false,
-        /*.pure                        =*/ false,
-        /*.keep_split                  =*/ false,
-        /*.imatrix                     =*/ nullptr,
-        /*.kv_overrides                =*/ nullptr,
-    };
-
-    return result;
-}
-
 size_t llama_max_devices(void) {
     return 16;
 }
@@ -19187,23 +9358,14 @@ void llama_backend_init(void) {
 
 void llama_numa_init(enum ggml_numa_strategy numa) {
     if (numa != GGML_NUMA_STRATEGY_DISABLED) {
-        ggml_numa_init(numa);
+        auto * dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
+        GGML_ASSERT(dev && "CPU backend is not loaded");
+        auto * reg = ggml_backend_dev_backend_reg(dev);
+        auto * numa_init_fn = (decltype(ggml_numa_init) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_numa_init");
+        numa_init_fn(numa);
     }
 }
 
-void llama_attach_threadpool(
-             struct llama_context * ctx,
-        ggml_threadpool_t   threadpool,
-        ggml_threadpool_t   threadpool_batch) {
-    ctx->threadpool       = threadpool;
-    ctx->threadpool_batch = threadpool_batch ? threadpool_batch : threadpool;
-}
-
-void llama_detach_threadpool(struct llama_context * ctx) {
-    ctx->threadpool       = nullptr;
-    ctx->threadpool_batch = nullptr;
-}
-
 void llama_backend_free(void) {
     ggml_quantize_free();
 }
@@ -19214,10 +9376,16 @@ int64_t llama_time_us(void) {
 
 struct llama_model * llama_load_model_from_file(
         const char * path_model,
-        struct llama_model_params   params) {
+        struct llama_model_params params) {
+    return llama_model_load_from_file(path_model, params);
+}
+
+struct llama_model * llama_model_load_from_file(
+        const char * path_model,
+        struct llama_model_params params) {
     ggml_time_init();
 
-    llama_model * model = new llama_model;
+    llama_model * model = new llama_model(params);
 
     unsigned cur_percentage = 0;
     if (params.progress_callback == NULL) {
@@ -19253,7 +9421,7 @@ struct llama_model * llama_load_model_from_file(
         ggml_backend_reg_t rpc_reg = ggml_backend_reg_by_name("RPC");
         if (!rpc_reg) {
             LLAMA_LOG_ERROR("%s: failed to find RPC backend\n", __func__);
-            llama_free_model(model);
+            llama_model_free(model);
             return nullptr;
         }
 
@@ -19261,7 +9429,7 @@ struct llama_model * llama_load_model_from_file(
         ggml_backend_rpc_add_device_t ggml_backend_rpc_add_device_fn = (ggml_backend_rpc_add_device_t) ggml_backend_reg_get_proc_address(rpc_reg, "ggml_backend_rpc_add_device");
         if (!ggml_backend_rpc_add_device_fn) {
             LLAMA_LOG_ERROR("%s: failed to find RPC device add function\n", __func__);
-            llama_free_model(model);
+            llama_model_free(model);
             return nullptr;
         }
 
@@ -19271,26 +9439,31 @@ struct llama_model * llama_load_model_from_file(
                 model->devices.push_back(dev);
             } else {
                 LLAMA_LOG_ERROR("%s: failed to add RPC device for server '%s'\n", __func__, server.c_str());
-                llama_free_model(model);
+                llama_model_free(model);
                 return nullptr;
             }
         }
     }
 
     // create list of devices to use with this model
-    // currently, we use all available devices
-    // TODO: rework API to give user more control over device selection
-    for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
-        ggml_backend_dev_t dev = ggml_backend_dev_get(i);
-        switch (ggml_backend_dev_type(dev)) {
-            case GGML_BACKEND_DEVICE_TYPE_CPU:
-            case GGML_BACKEND_DEVICE_TYPE_ACCEL:
-                // skip CPU backends since they are handled separately
-                break;
+    if (params.devices) {
+        for (ggml_backend_dev_t * dev = params.devices; *dev; ++dev) {
+            model->devices.push_back(*dev);
+        }
+    } else {
+        // use all available devices
+        for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
+            ggml_backend_dev_t dev = ggml_backend_dev_get(i);
+            switch (ggml_backend_dev_type(dev)) {
+                case GGML_BACKEND_DEVICE_TYPE_CPU:
+                case GGML_BACKEND_DEVICE_TYPE_ACCEL:
+                    // skip CPU backends since they are handled separately
+                    break;
 
-            case GGML_BACKEND_DEVICE_TYPE_GPU:
-                model->devices.push_back(dev);
-                break;
+                case GGML_BACKEND_DEVICE_TYPE_GPU:
+                    model->devices.push_back(dev);
+                    break;
+            }
         }
     }
 
@@ -19298,7 +9471,7 @@ struct llama_model * llama_load_model_from_file(
     if (params.split_mode == LLAMA_SPLIT_MODE_NONE) {
         if (params.main_gpu < 0 || params.main_gpu >= (int)model->devices.size()) {
             LLAMA_LOG_ERROR("%s: invalid value for main_gpu: %d (available devices: %d)\n", __func__, params.main_gpu, (int)model->devices.size());
-            llama_free_model(model);
+            llama_model_free(model);
             return nullptr;
         }
         ggml_backend_dev_t main_gpu = model->devices[params.main_gpu];
@@ -19312,7 +9485,7 @@ struct llama_model * llama_load_model_from_file(
         LLAMA_LOG_INFO("%s: using device %s (%s) - %zu MiB free\n", __func__, ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), free/1024/1024);
     }
 
-    int status = llama_model_load(path_model, *model, params);
+    const int status = llama_model_load(path_model, *model, params);
     GGML_ASSERT(status <= 0);
     if (status < 0) {
         if (status == -1) {
@@ -19320,18 +9493,15 @@ struct llama_model * llama_load_model_from_file(
         } else if (status == -2) {
             LLAMA_LOG_INFO("%s: cancelled model load\n", __func__);
         }
-        llama_free_model(model);
+
+        llama_model_free(model);
         return nullptr;
     }
 
     return model;
 }
 
-void llama_free_model(struct llama_model * model) {
-    delete model;
-}
-
-struct llama_context * llama_new_context_with_model(
+struct llama_context * llama_init_from_model(
                  struct llama_model * model,
         struct llama_context_params   params) {
 
@@ -19461,9 +9631,6 @@ struct llama_context * llama_new_context_with_model(
                 __func__, n_ctx_per_seq, hparams.n_ctx_train);
     }
 
-    ctx->abort_callback      = params.abort_callback;
-    ctx->abort_callback_data = params.abort_callback_data;
-
     ctx->logits_all = params.logits_all;
 
     // build worst-case graph for encoder if a model contains encoder
@@ -19512,7 +9679,7 @@ struct llama_context * llama_new_context_with_model(
         }
 
         // add CPU backend
-        ctx->backend_cpu = ggml_backend_cpu_init();
+        ctx->backend_cpu = ggml_backend_init_by_type(GGML_BACKEND_DEVICE_TYPE_CPU, nullptr);
         if (ctx->backend_cpu == nullptr) {
             LLAMA_LOG_ERROR("%s: failed to initialize CPU backend\n", __func__);
             llama_free(ctx);
@@ -19532,7 +9699,9 @@ struct llama_context * llama_new_context_with_model(
             }
         }
 
-        if (!llama_kv_cache_init(ctx->kv_self, ctx, type_k, type_v, kv_size, cparams.offload_kqv)) {
+        llama_set_abort_callback(ctx, params.abort_callback, params.abort_callback_data);
+
+        if (!llama_kv_cache_init(ctx->kv_self, ctx->model, ctx->cparams, type_k, type_v, kv_size, cparams.offload_kqv)) {
             LLAMA_LOG_ERROR("%s: llama_kv_cache_init() failed for self-attention cache\n", __func__);
             llama_free(ctx);
             return nullptr;
@@ -19577,7 +9746,8 @@ struct llama_context * llama_new_context_with_model(
             std::vector backend_ptrs;
             for (auto & backend : ctx->backends) {
                 auto * buft = ggml_backend_get_default_buffer_type(backend.get());
-                if (ggml_backend_is_cpu(backend.get()) && !model->devices.empty()) {
+                auto backend_type = ggml_backend_dev_type(ggml_backend_get_device(backend.get()));
+                if (backend_type == GGML_BACKEND_DEVICE_TYPE_CPU && !model->devices.empty()) {
                     // use the host buffer of the first device CPU for faster transfer of the intermediate state
                     auto * dev = model->devices[0];
                     auto * host_buft = ggml_backend_dev_host_buffer_type(dev);
@@ -19589,7 +9759,7 @@ struct llama_context * llama_new_context_with_model(
                 backend_ptrs.push_back(backend.get());
             }
 
-            const size_t max_nodes = llama_model_max_nodes(*model);
+            const size_t max_nodes = model->max_nodes();
 
             // buffer used to store the computation graph and the tensor meta data
             ctx->buf_compute_meta.resize(ggml_tensor_overhead()*max_nodes + ggml_graph_overhead_custom(max_nodes, false));
@@ -19597,15 +9767,16 @@ struct llama_context * llama_new_context_with_model(
             // TODO: move these checks to ggml_backend_sched
             // enabling pipeline parallelism in the scheduler increases memory usage, so it is only done when necessary
             bool pipeline_parallel =
-                llama_get_device_count(*model) > 1 &&
-                model->n_gpu_layers > (int)model->hparams.n_layer &&
-                model->split_mode == LLAMA_SPLIT_MODE_LAYER &&
+                model->n_devices() > 1 &&
+                model->params.n_gpu_layers > (int)model->hparams.n_layer &&
+                model->params.split_mode == LLAMA_SPLIT_MODE_LAYER &&
                 params.offload_kqv;
 
             // pipeline parallelism requires support for async compute and events in all devices
             if (pipeline_parallel) {
                 for (auto & backend : ctx->backends) {
-                    if (ggml_backend_is_cpu(backend.get())) {
+                    auto dev_type = ggml_backend_dev_type(ggml_backend_get_device(backend.get()));
+                    if (dev_type == GGML_BACKEND_DEVICE_TYPE_CPU) {
                         // ignore CPU backend
                         continue;
                     }
@@ -19629,7 +9800,7 @@ struct llama_context * llama_new_context_with_model(
             // initialize scheduler with the worst-case graph
             uint32_t n_seqs = 1; // TODO: worst-case number of sequences
             uint32_t n_tokens = std::min(cparams.n_ctx, cparams.n_ubatch);
-            llama_token token = llama_token_bos(&ctx->model); // not actually used by llama_build_graph, but required to choose between token and embedding inputs graph
+            llama_token token = ctx->model.vocab.token_bos(); // not actually used by llama_build_graph, but required to choose between token and embedding inputs graph
 
             llama_ubatch ubatch_pp = { true, n_tokens, n_tokens / n_seqs, n_seqs, &token, nullptr, nullptr, nullptr, nullptr, nullptr};
             ggml_cgraph * gf_pp = llama_build_graph(*ctx, ubatch_pp, true);
@@ -19681,454 +9852,32 @@ struct llama_context * llama_new_context_with_model(
     return ctx;
 }
 
-void llama_free(struct llama_context * ctx) {
-    delete ctx;
+struct llama_context * llama_new_context_with_model(
+                 struct llama_model * model,
+        struct llama_context_params   params) {
+    return llama_init_from_model(model, params);
 }
 
-uint32_t llama_n_ctx(const struct llama_context * ctx) {
-    return ctx->cparams.n_ctx;
-}
+//
+// kv cache
+//
 
-uint32_t llama_n_batch(const struct llama_context * ctx) {
-    return ctx->cparams.n_batch;
-}
-
-uint32_t llama_n_ubatch(const struct llama_context * ctx) {
-    return ctx->cparams.n_ubatch;
-}
-
-uint32_t llama_n_seq_max(const struct llama_context * ctx) {
-    return ctx->kv_self.size;
-}
-
-enum llama_vocab_type llama_vocab_type(const struct llama_model * model) {
-    return model->vocab.type;
-}
-
-int32_t llama_n_vocab(const struct llama_model * model) {
-    return model->hparams.n_vocab;
-}
-
-int32_t llama_n_ctx_train(const struct llama_model * model) {
-    return model->hparams.n_ctx_train;
-}
-
-int32_t llama_n_embd(const struct llama_model * model) {
-    return model->hparams.n_embd;
-}
-
-int32_t llama_n_layer(const struct llama_model * model) {
-    return model->hparams.n_layer;
-}
-
-int32_t llama_n_head(const struct llama_model * model) {
-    return model->hparams.n_head();
-}
-
-const struct llama_model * llama_get_model(const struct llama_context * ctx) {
-    return &ctx->model;
-}
-
-enum llama_pooling_type llama_pooling_type(const struct llama_context * ctx) {
-    return ctx->cparams.pooling_type;
-}
-
-enum llama_rope_type llama_rope_type(const struct llama_model * model) {
-    switch (model->arch) {
-        // these models do not use RoPE
-        case LLM_ARCH_GPT2:
-        case LLM_ARCH_GPTJ:
-        case LLM_ARCH_MPT:
-        case LLM_ARCH_REFACT:
-        case LLM_ARCH_BLOOM:
-        case LLM_ARCH_MAMBA:
-        case LLM_ARCH_JINA_BERT_V2:
-        case LLM_ARCH_T5:
-        case LLM_ARCH_T5ENCODER:
-        case LLM_ARCH_JAIS:
-        case LLM_ARCH_RWKV6:
-            return LLAMA_ROPE_TYPE_NONE;
-
-        // use what we call a normal RoPE, operating on pairs of consecutive head values
-        case LLM_ARCH_LLAMA:
-        case LLM_ARCH_BAICHUAN:
-        case LLM_ARCH_STARCODER:
-        case LLM_ARCH_PLAMO:
-        case LLM_ARCH_ORION:
-        case LLM_ARCH_INTERNLM2:
-        case LLM_ARCH_MINICPM:
-        case LLM_ARCH_XVERSE:
-        case LLM_ARCH_COMMAND_R:
-        case LLM_ARCH_OLMO:
-        case LLM_ARCH_ARCTIC:
-        case LLM_ARCH_DEEPSEEK2:
-        case LLM_ARCH_CHATGLM:
-        case LLM_ARCH_GRANITE:
-        case LLM_ARCH_GRANITE_MOE:
-        case LLM_ARCH_CHAMELEON:
-            return LLAMA_ROPE_TYPE_NORM;
-
-        // the pairs of head values are offset by n_rot/2
-        case LLM_ARCH_FALCON:
-        case LLM_ARCH_GROK:
-        case LLM_ARCH_DBRX:
-        case LLM_ARCH_BERT:
-        case LLM_ARCH_NOMIC_BERT:
-        case LLM_ARCH_STABLELM:
-        case LLM_ARCH_BITNET:
-        case LLM_ARCH_QWEN:
-        case LLM_ARCH_QWEN2:
-        case LLM_ARCH_QWEN2MOE:
-        case LLM_ARCH_OLMOE:
-        case LLM_ARCH_PHI2:
-        case LLM_ARCH_PHI3:
-        case LLM_ARCH_GEMMA:
-        case LLM_ARCH_GEMMA2:
-        case LLM_ARCH_STARCODER2:
-        case LLM_ARCH_OPENELM:
-        case LLM_ARCH_GPTNEOX:
-        case LLM_ARCH_CODESHELL:
-        case LLM_ARCH_NEMOTRON:
-        case LLM_ARCH_EXAONE:
-        case LLM_ARCH_MINICPM3:
-            return LLAMA_ROPE_TYPE_NEOX;
-
-        // all model arches should be listed explicitly here
-        case LLM_ARCH_UNKNOWN:
-            GGML_ABORT("unknown architecture");
-    }
-
-    return LLAMA_ROPE_TYPE_NONE;
-}
-
-float llama_rope_freq_scale_train(const struct llama_model * model) {
-    return model->hparams.rope_freq_scale_train;
-}
-
-int32_t llama_model_meta_val_str(const struct llama_model * model, const char * key, char * buf, size_t buf_size) {
-    const auto & it = model->gguf_kv.find(key);
-    if (it == model->gguf_kv.end()) {
-        if (buf_size > 0) {
-            buf[0] = '\0';
-        }
-        return -1;
-    }
-    return snprintf(buf, buf_size, "%s", it->second.c_str());
-}
-
-int32_t llama_model_meta_count(const struct llama_model * model) {
-    return (int)model->gguf_kv.size();
-}
-
-int32_t llama_model_meta_key_by_index(const struct llama_model * model, int i, char * buf, size_t buf_size) {
-    if (i < 0 || i >= (int)model->gguf_kv.size()) {
-        if (buf_size > 0) {
-            buf[0] = '\0';
-        }
-        return -1;
-    }
-    auto it = model->gguf_kv.begin();
-    std::advance(it, i);
-    return snprintf(buf, buf_size, "%s", it->first.c_str());
-}
-
-int32_t llama_model_meta_val_str_by_index(const struct llama_model * model, int32_t i, char * buf, size_t buf_size) {
-    if (i < 0 || i >= (int)model->gguf_kv.size()) {
-        if (buf_size > 0) {
-            buf[0] = '\0';
-        }
-        return -1;
-    }
-    auto it = model->gguf_kv.begin();
-    std::advance(it, i);
-    return snprintf(buf, buf_size, "%s", it->second.c_str());
-}
-
-int32_t llama_model_desc(const struct llama_model * model, char * buf, size_t buf_size) {
-    return snprintf(buf, buf_size, "%s %s %s",
-            llama_model_arch_name(model->arch),
-            llama_model_type_name(model->type),
-            llama_model_ftype_name(model->ftype).c_str());
-}
-
-uint64_t llama_model_size(const struct llama_model * model) {
-    uint64_t size = 0;
-    for (const auto & it : model->tensors_by_name) {
-        size += ggml_nbytes(it.second);
-    }
-    return size;
-}
-
-uint64_t llama_model_n_params(const struct llama_model * model) {
-    uint64_t nparams = 0;
-    for (const auto & it : model->tensors_by_name) {
-        nparams += ggml_nelements(it.second);
-    }
-    return nparams;
-}
-
-struct ggml_tensor * llama_get_model_tensor(struct llama_model * model, const char * name) {
-    auto it = std::find_if(model->tensors_by_name.begin(), model->tensors_by_name.end(),
-            [name](const std::pair & it) {
-                return it.first == name;
-            });
-    if (it == model->tensors_by_name.end()) {
-        return nullptr;
-    }
-    return it->second;
-}
-
-bool llama_model_has_encoder(const struct llama_model * model) {
-    switch (model->arch) {
-        case LLM_ARCH_T5:        return true;
-        case LLM_ARCH_T5ENCODER: return true;
-        default:                 return false;
-    }
-}
-
-bool llama_model_has_decoder(const struct llama_model * model) {
-    switch (model->arch) {
-        case LLM_ARCH_T5ENCODER: return false;
-        default:                 return true;
-    }
-}
-
-llama_token llama_model_decoder_start_token(const struct llama_model * model) {
-    return model->hparams.dec_start_token_id;
-}
-
-bool llama_model_is_recurrent(const struct llama_model * model) {
-    switch (model->arch) {
-        case LLM_ARCH_MAMBA:  return true;
-        case LLM_ARCH_RWKV6:  return true;
-        default:              return false;
-    }
-}
-
-uint32_t llama_model_quantize(
-        const char * fname_inp,
-        const char * fname_out,
-        const llama_model_quantize_params * params) {
-    try {
-        llama_model_quantize_internal(fname_inp, fname_out, params);
-        return 0;
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: failed to quantize: %s\n", __func__, err.what());
-        return 1;
-    }
-}
-
-struct llama_lora_adapter * llama_lora_adapter_init(struct llama_model * model, const char * path_lora) {
-    try {
-        struct llama_lora_adapter * adapter = new llama_lora_adapter(model);
-        llama_lora_adapter_init_internal(model, path_lora, *adapter);
-        return adapter;
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: failed to apply lora adapter: %s\n", __func__, err.what());
-        return nullptr;
-    }
-}
-
-static bool llama_control_vector_init(struct llama_control_vector & cvec, const llama_model & model) {
-    GGML_ASSERT(cvec.tensors.empty());
-    GGML_ASSERT(cvec.ctxs.empty());
-    GGML_ASSERT(cvec.bufs.empty());
-
-    // create a context for each buffer type
-    std::map ctx_map;
-    auto ctx_for_buft = [&](ggml_backend_buffer_type_t buft) -> ggml_context * {
-        auto it = ctx_map.find(buft);
-        if (it == ctx_map.end()) {
-            struct ggml_init_params params = {
-                /*.mem_size   =*/ model.hparams.n_layer*ggml_tensor_overhead(),
-                /*.mem_buffer =*/ NULL,
-                /*.no_alloc   =*/ true,
-            };
-            ggml_context * ctx = ggml_init(params);
-            if (!ctx) {
-                return nullptr;
-            }
-            ctx_map[buft] = ctx;
-            cvec.ctxs.emplace_back(ctx);
-            return ctx;
-        }
-        return it->second;
-    };
-
-    // make tensors
-    cvec.tensors.reserve(model.hparams.n_layer);
-    cvec.tensors.push_back(nullptr); // there's never a tensor for layer 0
-    for (size_t il = 1; il < model.hparams.n_layer; il++) {
-        ggml_backend_buffer_type_t buft = select_buft(*model.dev_layer.at(il).buft_list,
-            [&](ggml_context * ctx) {
-                ggml_tensor * cur = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, model.hparams.n_embd);
-                ggml_tensor * layer_dir = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, model.hparams.n_embd);
-                return ggml_add(ctx, cur, layer_dir);
-            });
-        ggml_context * ctx = ctx_for_buft(buft);
-        if (!ctx) {
-            LLAMA_LOG_ERROR("%s: failed to allocate context for control vector\n", __func__);
-            return false;
-        }
-        ggml_tensor * tensor = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, model.hparams.n_embd);
-        cvec.tensors.push_back(tensor);
-    }
-
-    // allocate tensors / buffers and zero
-    cvec.bufs.reserve(ctx_map.size());
-    for (auto it : ctx_map) {
-        ggml_backend_buffer_type_t buft = it.first;
-        ggml_context * ctx = it.second;
-        ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors_from_buft(ctx, buft);
-        if (!buf) {
-            LLAMA_LOG_ERROR("%s: failed to allocate buffer for control vector\n", __func__);
-            return false;
-        }
-        ggml_backend_buffer_clear(buf, 0);
-        cvec.bufs.emplace_back(buf);
-    }
-
-    return true;
-}
-
-int32_t llama_control_vector_apply(struct llama_context * lctx, const float * data, size_t len, int32_t n_embd, int32_t il_start, int32_t il_end) {
-    const llama_model & model = lctx->model;
-    llama_control_vector & cvec = lctx->cvec;
-
-    if (data == nullptr) {
-        // disable the current control vector (but leave allocated for later)
-        cvec.layer_start = -1;
-        cvec.layer_end   = -1;
-        return 0;
-    }
-
-    if (n_embd != (int) model.hparams.n_embd) {
-        LLAMA_LOG_ERROR("%s: control vector n_embd does not match model\n", __func__);
-        return 1;
-    }
-
-    if (cvec.tensors.empty()) {
-        if (!llama_control_vector_init(cvec, model)) {
-            return 1;
-        }
-    }
-
-    cvec.layer_start = il_start;
-    cvec.layer_end   = il_end;
-
-    for (size_t il = 1; il < model.hparams.n_layer; il++) {
-        assert(cvec.tensors[il] != nullptr);
-
-        const size_t off = n_embd * (il - 1); // buffer doesn't have data for layer 0, since it's never present
-        if (off + n_embd <= len) {
-            ggml_backend_tensor_set(cvec.tensors[il], data + off, 0, n_embd * ggml_element_size(cvec.tensors[il]));
-        }
-    }
-
-    return 0;
-}
+// TODO: tmp bridges below until `struct llama_kv_cache` is exposed through the public API
 
 struct llama_kv_cache_view llama_kv_cache_view_init(const struct llama_context * ctx, int32_t n_seq_max) {
-    struct llama_kv_cache_view result = {
-        /*.n_cells            = */ 0,
-        /*.n_seq_max          = */ n_seq_max,
-        /*.token_count        = */ 0,
-        /*.used_cells         = */ llama_get_kv_cache_used_cells(ctx),
-        /*.max_contiguous     = */ 0,
-        /*.max_contiguous_idx = */ -1,
-        /*.cells              = */ nullptr,
-        /*.cells_sequences    = */ nullptr,
-    };
-    return result;
-}
-
-void llama_kv_cache_view_free(struct llama_kv_cache_view * view) {
-    if (view->cells != nullptr) {
-        free(view->cells);
-        view->cells = nullptr;
-    }
-    if (view->cells_sequences != nullptr) {
-        free(view->cells_sequences);
-        view->cells_sequences = nullptr;
-    }
+    return llama_kv_cache_view_init(ctx->kv_self, n_seq_max);
 }
 
 void llama_kv_cache_view_update(const struct llama_context * ctx, struct llama_kv_cache_view * view) {
-    if (uint32_t(view->n_cells) < ctx->kv_self.size || view->cells == nullptr) {
-        view->n_cells = int32_t(ctx->kv_self.size);
-        void * p = realloc(view->cells, sizeof(struct llama_kv_cache_view_cell) * view->n_cells);
-        GGML_ASSERT(p != nullptr && "Failed to alloc kv_cache_view cells");
-        view->cells = (struct llama_kv_cache_view_cell *)p;
-        p = realloc(view->cells_sequences, sizeof(llama_seq_id) * view->n_seq_max * view->n_cells);
-        GGML_ASSERT(p != nullptr && "Failed to alloc kv_cache_view cells sequences");
-        view->cells_sequences = (llama_seq_id *)p;
-    }
-
-    const std::vector & kv_cells = ctx->kv_self.cells;
-    llama_kv_cache_view_cell * c_curr = view->cells;
-    llama_seq_id * cs_curr = view->cells_sequences;
-    int32_t used_cells = 0;
-    int32_t token_count = 0;
-    int32_t curr_contig_idx = -1;
-    uint32_t max_contig = 0;
-    int32_t max_contig_idx = -1;
-
-    for (int32_t i = 0; i < int32_t(ctx->kv_self.size); i++, c_curr++, cs_curr += view->n_seq_max) {
-        const size_t curr_size = kv_cells[i].seq_id.size();
-        token_count += curr_size;
-        c_curr->pos = kv_cells[i].pos + kv_cells[i].delta;
-
-        if (curr_size > 0) {
-            if (curr_contig_idx >= 0 && uint32_t(i - curr_contig_idx) > max_contig) {
-                max_contig = i - curr_contig_idx;
-                max_contig_idx = curr_contig_idx;
-            }
-            curr_contig_idx = -1;
-        } else if (curr_contig_idx < 0) {
-            curr_contig_idx = i;
-        }
-
-        int seq_idx = 0;
-        for (const llama_seq_id it : kv_cells[i].seq_id) {
-            if (seq_idx >= view->n_seq_max) {
-                break;
-            }
-            cs_curr[seq_idx] = it;
-            seq_idx++;
-        }
-        if (seq_idx != 0) {
-            used_cells++;
-        }
-        for (; seq_idx < view->n_seq_max; seq_idx++) {
-            cs_curr[seq_idx] = -1;
-        }
-    }
-    if (curr_contig_idx >= 0 && kv_cells.size() - curr_contig_idx > max_contig) {
-        max_contig_idx = curr_contig_idx;
-        max_contig = kv_cells.size() - curr_contig_idx;
-    }
-    view->max_contiguous = max_contig;
-    view->max_contiguous_idx = max_contig_idx;
-    view->token_count = token_count;
-    view->used_cells = used_cells;
-    if (uint32_t(used_cells) != ctx->kv_self.used) {
-        LLAMA_LOG_ERROR("%s: used cells mismatch. kv_cache says %d but we calculated %d\n",
-            __func__, ctx->kv_self.used, used_cells);
-    }
+    llama_kv_cache_view_update(view, ctx->kv_self);
 }
 
 int32_t llama_get_kv_cache_token_count(const struct llama_context * ctx) {
-    int result = 0;
-
-    for (uint32_t i = 0; i < ctx->kv_self.size; i++) {
-        result += ctx->kv_self.cells[i].seq_id.size();
-    }
-
-    return result;
+    return llama_get_kv_cache_token_count(ctx->kv_self);
 }
 
 int32_t llama_get_kv_cache_used_cells(const struct llama_context * ctx) {
-    return ctx->kv_self.used;
+    return llama_get_kv_cache_used_cells(ctx->kv_self);
 }
 
 void llama_kv_cache_clear(struct llama_context * ctx) {
@@ -20175,1065 +9924,19 @@ void llama_kv_cache_defrag(struct llama_context * ctx) {
 }
 
 void llama_kv_cache_update(struct llama_context * ctx) {
-    llama_kv_cache_update_internal(*ctx);
+    llama_kv_cache_update_impl(*ctx);
 }
 
-// deprecated
-size_t llama_get_state_size(struct llama_context * ctx) {
-    return llama_state_get_size(ctx);
+bool llama_kv_cache_can_shift(struct llama_context * ctx) {
+    return llama_kv_cache_can_shift(ctx->kv_self);
 }
 
-// deprecated
-size_t llama_copy_state_data(struct llama_context * ctx, uint8_t * dst) {
-    return llama_state_get_data(ctx, dst, -1);
-}
-
-// deprecated
-size_t llama_set_state_data(struct llama_context * ctx, const uint8_t * src) {
-    return llama_state_set_data(ctx, src, -1);
-}
-
-// deprecated
-bool llama_load_session_file(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) {
-    return llama_state_load_file(ctx, path_session, tokens_out, n_token_capacity, n_token_count_out);
-}
-
-// deprecated
-bool llama_save_session_file(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) {
-    return llama_state_save_file(ctx, path_session, tokens, n_token_count);
-}
-
-// TODO: replace all non-fatal assertions with returned errors or exceptions
-struct llama_data_write {
-    virtual void write(const void * src, size_t size) = 0;
-    virtual void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) = 0;
-    virtual size_t get_size_written() = 0;
-    virtual ~llama_data_write() = default;
-
-    void write_string(const std::string & str) {
-        uint32_t str_size = str.size();
-
-        write(&str_size,  sizeof(str_size));
-        write(str.data(), str_size);
-    }
-
-    void write_model_info(const struct llama_context * ctx) {
-        std::string arch_str = LLM_ARCH_NAMES.at(ctx->model.arch);
-        write_string(arch_str);
-        // TODO: add more model-specific info which should prevent loading the session file if not identical
-    }
-
-    //void write_rng(const std::mt19937 & rng) {
-    //    std::ostringstream rng_ss;
-    //    rng_ss << rng;
-
-    //    const std::string & rng_str = rng_ss.str();
-
-    //    write_string(rng_str);
-    //}
-
-    void write_output_ids(struct llama_context * ctx) {
-        llama_output_reorder(ctx);
-
-        const uint32_t n_outputs = ctx->n_outputs;
-
-        std::vector output_pos;
-
-        const size_t    n_batch = ctx->cparams.n_batch;
-        const auto & output_ids = ctx->output_ids;
-
-        GGML_ASSERT(n_outputs <= ctx->output_size);
-
-        output_pos.resize(n_outputs);
-
-        // build a more compact representation of the output ids
-        for (size_t i = 0; i < n_batch; ++i) {
-            // map an output id to a position in the batch
-            int32_t pos = output_ids[i];
-            if (pos >= 0) {
-                GGML_ASSERT((uint32_t) pos < n_outputs);
-                output_pos[pos] = i;
-            }
-        }
-
-        write(&n_outputs, sizeof(n_outputs));
-
-        if (n_outputs) {
-            write(output_pos.data(), n_outputs * sizeof(int32_t));
-        }
-    }
-
-    void write_logits(const struct llama_context * ctx) {
-        const uint64_t logits_size = std::min((uint64_t) ctx->logits_size, (uint64_t) ctx->n_outputs * ctx->model.hparams.n_vocab);
-
-        write(&logits_size, sizeof(logits_size));
-
-        if (logits_size) {
-            write(ctx->logits, logits_size * sizeof(float));
-        }
-    }
-
-    void write_embeddings(const struct llama_context * ctx) {
-        const uint64_t embeddings_size = std::min((uint64_t) ctx->embd_size, (uint64_t) ctx->n_outputs * ctx->model.hparams.n_embd);
-
-        write(&embeddings_size, sizeof(embeddings_size));
-
-        if (embeddings_size) {
-            write(ctx->embd, embeddings_size * sizeof(float));
-        }
-    }
-
-    void write_kv_cache_meta(const llama_kv_cache & kv_self, const std::vector> & cell_ranges, llama_seq_id seq_id = -1) {
-
-        for (const auto & range : cell_ranges) {
-            for (uint32_t i = range.first; i < range.second; ++i) {
-                const auto & cell = kv_self.cells[i];
-                const llama_pos pos      = cell.pos;
-                const uint32_t  n_seq_id = seq_id == -1 ? cell.seq_id.size() : 0;
-
-                write(&pos,      sizeof(pos));
-                write(&n_seq_id, sizeof(n_seq_id));
-
-                if (n_seq_id) {
-                    for (auto seq_id : cell.seq_id) {
-                        write(&seq_id, sizeof(seq_id));
-                    }
-                }
-            }
-        }
-    }
-
-    void write_kv_cache_data(const struct llama_context * ctx, const std::vector> & cell_ranges) {
-        const struct llama_kv_cache & kv_self = ctx->kv_self;
-        const struct llama_hparams & hparams = ctx->model.hparams;
-
-        const uint32_t v_trans = kv_self.v_trans ? 1 : 0;
-        const uint32_t n_layer = hparams.n_layer;
-
-        write(&v_trans, sizeof(v_trans));
-        write(&n_layer, sizeof(n_layer));
-
-        std::vector tmp_buf;
-
-        // Iterate and write all the keys first, each row is a cell
-        // Get whole range at a time
-        for (uint32_t il = 0; il < n_layer; ++il) {
-            const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(il) + hparams.n_embd_k_s();
-
-            // Write key type
-            const int32_t k_type_i = (int32_t)kv_self.k_l[il]->type;
-            write(&k_type_i, sizeof(k_type_i));
-
-            // Write row size of key
-            const uint64_t k_size_row = ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa);
-            write(&k_size_row, sizeof(k_size_row));
-
-            // Read each range of cells of k_size length each into tmp_buf and write out
-            for (const auto & range : cell_ranges) {
-                const size_t range_size = range.second - range.first;
-                const size_t buf_size = range_size * k_size_row;
-                write_tensor_data(kv_self.k_l[il], range.first * k_size_row, buf_size);
-            }
-        }
-
-        if (!kv_self.v_trans) {
-            for (uint32_t il = 0; il < n_layer; ++il) {
-                const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s();
-
-                // Write value type
-                const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type;
-                write(&v_type_i, sizeof(v_type_i));
-
-                // Write row size of value
-                const uint64_t v_size_row = ggml_row_size(kv_self.v_l[il]->type, n_embd_v_gqa);
-                write(&v_size_row, sizeof(v_size_row));
-
-                // Read each range of cells of v_size length each into tmp_buf and write out
-                for (const auto & range : cell_ranges) {
-                    const size_t range_size = range.second - range.first;
-                    const size_t buf_size = range_size * v_size_row;
-                    write_tensor_data(kv_self.v_l[il], range.first * v_size_row, buf_size);
-                }
-            }
-        } else {
-            // When v is transposed, we also need the element size and get the element ranges from each row
-            const uint32_t kv_size = kv_self.size;
-            for (uint32_t il = 0; il < n_layer; ++il) {
-                const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s();
-
-                // Write value type
-                const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type;
-                write(&v_type_i, sizeof(v_type_i));
-
-                // Write element size
-                const uint32_t v_size_el = ggml_type_size(kv_self.v_l[il]->type);
-                write(&v_size_el, sizeof(v_size_el));
-
-                // Write GQA embedding size
-                write(&n_embd_v_gqa, sizeof(n_embd_v_gqa));
-
-                // For each row, we get the element values of each cell
-                for (uint32_t j = 0; j < n_embd_v_gqa; ++j) {
-                    // Read each range of cells of v_size_el length each into tmp_buf and write out
-                    for (const auto & range : cell_ranges) {
-                        const size_t range_size = range.second - range.first;
-                        const size_t src_offset = (range.first + j * kv_size) * v_size_el;
-                        const size_t buf_size = range_size * v_size_el;
-                        write_tensor_data(kv_self.v_l[il], src_offset, buf_size);
-                    }
-                }
-            }
-        }
-    }
-
-    void write_kv_cache(const struct llama_context * ctx, llama_seq_id seq_id = -1) {
-        const struct llama_kv_cache & kv_self = ctx->kv_self;
-        std::vector> cell_ranges; // ranges, from inclusive, to exclusive
-        uint32_t cell_count = 0;
-
-        // Count the number of cells with the specified seq_id
-        // Find all the ranges of cells with this seq id (or all, when -1)
-        uint32_t cell_range_begin = kv_self.size;
-        for (uint32_t i = 0; i < kv_self.size; ++i) {
-            const auto & cell = kv_self.cells[i];
-            if ((seq_id == -1 && !cell.is_empty()) || cell.has_seq_id(seq_id)) {
-                ++cell_count;
-                if (cell_range_begin == kv_self.size) {
-                    cell_range_begin = i;
-                }
-            } else {
-                if (cell_range_begin != kv_self.size) {
-                    cell_ranges.emplace_back(cell_range_begin, i);
-                    cell_range_begin = kv_self.size;
-                }
-            }
-        }
-        if (cell_range_begin != kv_self.size) {
-            cell_ranges.emplace_back(cell_range_begin, kv_self.size);
-        }
-
-        // DEBUG CHECK: Sum of cell counts in ranges should equal the total cell count
-        uint32_t cell_count_check = 0;
-        for (const auto & range : cell_ranges) {
-            cell_count_check += range.second - range.first;
-        }
-        GGML_ASSERT(cell_count == cell_count_check);
-
-        write(&cell_count, sizeof(cell_count));
-
-        write_kv_cache_meta(kv_self, cell_ranges, seq_id);
-        write_kv_cache_data(ctx, cell_ranges);
-    }
-};
-
-struct llama_data_read {
-    virtual const uint8_t * read(size_t size) = 0;
-    virtual void read_to(void * dst, size_t size) = 0;
-    virtual size_t get_size_read() = 0;
-    virtual ~llama_data_read() = default;
-
-    void read_string(std::string & str) {
-        uint32_t str_size;
-        read_to(&str_size, sizeof(str_size));
-
-        str.assign((const char *) read(str_size), str_size);
-    }
-
-    // validate model information
-    void read_model_info(const struct llama_context * ctx) {
-        std::string cur_arch_str = LLM_ARCH_NAMES.at(ctx->model.arch);
-        std::string arch_str;
-        read_string(arch_str);
-        if (cur_arch_str != arch_str) {
-            throw std::runtime_error(format("wrong model arch: '%s' instead of '%s'", arch_str.c_str(), cur_arch_str.c_str()));
-        }
-        // TODO: add more info which needs to be identical but which is not verified otherwise
-    }
-
-    //void read_rng(std::mt19937 & rng) {
-    //    std::string rng_str;
-    //    read_string(rng_str);
-
-    //    std::istringstream rng_ss(rng_str);
-    //    rng_ss >> rng;
-
-    //    if (rng_ss.fail()) {
-    //        throw std::runtime_error("failed to load RNG state");
-    //    }
-    //}
-
-    void read_output_ids(struct llama_context * ctx) {
-        std::vector output_pos;
-
-        uint32_t n_outputs;
-        read_to(&n_outputs, sizeof(n_outputs));
-
-        if (n_outputs > llama_output_reserve(*ctx, n_outputs)) {
-            throw std::runtime_error("could not reserve outputs");
-        }
-
-        if (n_outputs) {
-            output_pos.resize(n_outputs);
-            read_to(output_pos.data(), n_outputs * sizeof(int32_t));
-
-            for (int32_t i = 0; i < (int32_t) output_pos.size(); ++i) {
-                int32_t id = output_pos[i];
-                if ((uint32_t) id >= ctx->cparams.n_batch) {
-                    throw std::runtime_error(format("invalid output id, %d does not fit in batch size of %u", id, ctx->cparams.n_batch));
-                }
-                ctx->output_ids[id] = i;
-            }
-
-            ctx->n_outputs = n_outputs;
-        }
-    }
-
-    void read_logits(struct llama_context * ctx) {
-        uint64_t logits_size;
-        read_to(&logits_size, sizeof(logits_size));
-
-        if (ctx->logits_size < logits_size) {
-            throw std::runtime_error("logits buffer too small");
-        }
-
-        if (logits_size) {
-            read_to(ctx->logits, logits_size * sizeof(float));
-        }
-    }
-
-    void read_embeddings(struct llama_context * ctx) {
-        uint64_t embeddings_size;
-        read_to(&embeddings_size, sizeof(embeddings_size));
-
-        if (ctx->embd_size < embeddings_size) {
-            throw std::runtime_error("embeddings buffer too small");
-        }
-
-        if (embeddings_size) {
-            read_to(ctx->embd, embeddings_size * sizeof(float));
-        }
-    }
-
-    bool read_kv_cache_meta(struct llama_context * ctx, uint32_t cell_count, llama_seq_id dest_seq_id = -1) {
-        struct llama_kv_cache & kv_self = ctx->kv_self;
-
-        if (dest_seq_id != -1) {
-            // single sequence
-
-            llama_kv_cache_seq_rm(kv_self, dest_seq_id, -1, -1);
-
-            llama_ubatch batch = ctx->sbatch.reserve_ubatch(cell_count, /* has_embd */ false);
-            batch.n_tokens = cell_count;
-            batch.n_seq_tokens = cell_count;
-            batch.n_seqs = 1;
-
-            for (uint32_t i = 0; i < cell_count; ++i) {
-                llama_pos pos;
-                uint32_t n_seq_id;
-
-                read_to(&pos, sizeof(pos));
-                read_to(&n_seq_id, sizeof(n_seq_id));
-
-                if (n_seq_id != 0) {
-                    LLAMA_LOG_ERROR("%s: invalid seq_id-agnostic kv cell\n", __func__);
-                    return false;
-                }
-
-                batch.pos[i] = pos;
-            }
-            batch.n_seq_id[0] = 1;
-            batch.seq_id[0] = &dest_seq_id;
-            if (!llama_kv_cache_find_slot(kv_self, batch)) {
-                LLAMA_LOG_ERROR("%s: failed to find available cells in kv cache\n", __func__);
-                return false;
-            }
-
-            // DEBUG CHECK: kv_self.head should be our first cell, kv_self.head + cell_count - 1 should be our last cell (verify seq_id and pos values)
-            // Assume that this is one contiguous block of cells
-            GGML_ASSERT(kv_self.head + cell_count <= kv_self.size);
-            GGML_ASSERT(kv_self.cells[kv_self.head].pos == batch.pos[0]);
-            GGML_ASSERT(kv_self.cells[kv_self.head + cell_count - 1].pos == batch.pos[cell_count - 1]);
-            GGML_ASSERT(kv_self.cells[kv_self.head].has_seq_id(dest_seq_id));
-            GGML_ASSERT(kv_self.cells[kv_self.head + cell_count - 1].has_seq_id(dest_seq_id));
-        } else {
-            // whole KV cache restore
-
-            if (cell_count > kv_self.size) {
-                LLAMA_LOG_ERROR("%s: not enough cells in kv cache\n", __func__);
-                return false;
-            }
-
-            llama_kv_cache_clear(kv_self);
-
-            for (uint32_t i = 0; i < cell_count; ++i) {
-                llama_kv_cell & cell = kv_self.cells[i];
-
-                llama_pos pos;
-                uint32_t  n_seq_id;
-
-                read_to(&pos,      sizeof(pos));
-                read_to(&n_seq_id, sizeof(n_seq_id));
-
-                cell.pos = pos;
-
-                for (uint32_t j = 0; j < n_seq_id; ++j) {
-                    llama_seq_id seq_id;
-                    read_to(&seq_id, sizeof(seq_id));
-
-                    if (seq_id < 0 || (uint32_t) seq_id >= llama_n_seq_max(ctx)) {
-                        LLAMA_LOG_ERROR("%s: invalid seq_id, %d is out of range [0, %u)\n", __func__, seq_id, llama_n_seq_max(ctx));
-                        return false;
-                    }
-
-                    cell.seq_id.insert(seq_id);
-
-                    if (kv_self.recurrent) {
-                        int32_t & tail = kv_self.cells[seq_id].tail;
-                        if (tail != -1) {
-                            LLAMA_LOG_ERROR("%s: duplicate tail for seq_id %d in cell %d and %d\n", __func__, seq_id, i, tail);
-                            return false;
-                        }
-                        tail = i;
-                    }
-                }
-            }
-
-            kv_self.head = 0;
-            kv_self.used = cell_count;
-        }
-
-        if (kv_self.recurrent) {
-            for (uint32_t i = 0; i < cell_count; ++i) {
-                uint32_t cell_id = kv_self.head + i;
-                // make sure the recurrent states will keep their restored state
-                kv_self.cells[cell_id].src = cell_id;
-            }
-        }
-
-        return true;
-    }
-
-    bool read_kv_cache_data(struct llama_context * ctx, uint32_t cell_count) {
-        const struct llama_hparams & hparams = ctx->model.hparams;
-        struct llama_kv_cache & kv_self = ctx->kv_self;
-        uint32_t v_trans;
-        uint32_t n_layer;
-        read_to(&v_trans, sizeof(v_trans));
-        read_to(&n_layer, sizeof(n_layer));
-
-        if (n_layer != hparams.n_layer) {
-            LLAMA_LOG_ERROR("%s: mismatched layer count (%u instead of %u)\n", __func__, n_layer, hparams.n_layer);
-            return false;
-        }
-        if (cell_count > kv_self.size) {
-            LLAMA_LOG_ERROR("%s: not enough cells in kv cache to restore state (%u > %u)\n", __func__, cell_count, kv_self.size);
-            return false;
-        }
-        if (kv_self.v_trans != (bool) v_trans) {
-            LLAMA_LOG_ERROR("%s: incompatible V transposition\n", __func__);
-            return false;
-        }
-
-        // For each layer, read the keys for each cell, one row is one cell, read as one contiguous block
-        for (uint32_t il = 0; il < n_layer; ++il) {
-            const uint32_t n_embd_k_gqa = hparams.n_embd_k_gqa(il) + hparams.n_embd_k_s();
-
-            // Read type of key
-            int32_t k_type_i_ref;
-            read_to(&k_type_i_ref, sizeof(k_type_i_ref));
-            const int32_t k_type_i = (int32_t)kv_self.k_l[il]->type;
-            if (k_type_i != k_type_i_ref) {
-                LLAMA_LOG_ERROR("%s: mismatched key type (%d != %d, layer %d)\n", __func__, k_type_i, k_type_i_ref, il);
-                return false;
-            }
-
-            // Read row size of key
-            uint64_t k_size_row_ref;
-            read_to(&k_size_row_ref, sizeof(k_size_row_ref));
-            const size_t k_size_row = ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa);
-            if (k_size_row != k_size_row_ref) {
-                LLAMA_LOG_ERROR("%s: mismatched key row size (%zu != %zu, layer %d)\n", __func__, k_size_row, (size_t) k_size_row_ref, il);
-                return false;
-            }
-
-            if (cell_count) {
-                // Read and set the keys for the whole cell range
-                ggml_backend_tensor_set(kv_self.k_l[il], read(cell_count * k_size_row), kv_self.head * k_size_row, cell_count * k_size_row);
-            }
-        }
-
-        if (!kv_self.v_trans) {
-            for (uint32_t il = 0; il < n_layer; ++il) {
-                const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s();
-
-                // Read type of value
-                int32_t v_type_i_ref;
-                read_to(&v_type_i_ref, sizeof(v_type_i_ref));
-                const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type;
-                if (v_type_i != v_type_i_ref) {
-                    LLAMA_LOG_ERROR("%s: mismatched value type (%d != %d, layer %d)\n", __func__, v_type_i, v_type_i_ref, il);
-                    return false;
-                }
-
-                // Read row size of value
-                uint64_t v_size_row_ref;
-                read_to(&v_size_row_ref, sizeof(v_size_row_ref));
-                const size_t v_size_row = ggml_row_size(kv_self.v_l[il]->type, n_embd_v_gqa);
-                if (v_size_row != v_size_row_ref) {
-                    LLAMA_LOG_ERROR("%s: mismatched value row size (%zu != %zu, layer %d)\n", __func__, v_size_row, (size_t) v_size_row_ref, il);
-                    return false;
-                }
-
-                if (cell_count) {
-                    // Read and set the values for the whole cell range
-                    ggml_backend_tensor_set(kv_self.v_l[il], read(cell_count * v_size_row), kv_self.head * v_size_row, cell_count * v_size_row);
-                }
-            }
-        } else {
-            // For each layer, read the values for each cell (transposed)
-            for (uint32_t il = 0; il < n_layer; ++il) {
-                const uint32_t n_embd_v_gqa = hparams.n_embd_v_gqa(il) + hparams.n_embd_v_s();
-
-                // Read type of value
-                int32_t v_type_i_ref;
-                read_to(&v_type_i_ref, sizeof(v_type_i_ref));
-                const int32_t v_type_i = (int32_t)kv_self.v_l[il]->type;
-                if (v_type_i != v_type_i_ref) {
-                    LLAMA_LOG_ERROR("%s: mismatched value type (%d != %d, layer %d)\n", __func__, v_type_i, v_type_i_ref, il);
-                    return false;
-                }
-
-                // Read element size of value
-                uint32_t v_size_el_ref;
-                read_to(&v_size_el_ref, sizeof(v_size_el_ref));
-                const size_t v_size_el = ggml_type_size(kv_self.v_l[il]->type);
-                if (v_size_el != v_size_el_ref) {
-                    LLAMA_LOG_ERROR("%s: mismatched value element size (%zu != %zu, layer %d)\n", __func__, v_size_el, (size_t) v_size_el_ref, il);
-                    return false;
-                }
-
-                // Read GQA embedding size
-                uint32_t n_embd_v_gqa_ref;
-                read_to(&n_embd_v_gqa_ref, sizeof(n_embd_v_gqa_ref));
-                if (n_embd_v_gqa != n_embd_v_gqa_ref) {
-                    LLAMA_LOG_ERROR("%s: mismatched GQA embedding size (%u != %u, layer %d)\n", __func__, n_embd_v_gqa, n_embd_v_gqa_ref, il);
-                    return false;
-                }
-
-                if (cell_count) {
-                    // For each row in the transposed matrix, read the values for the whole cell range
-                    for (uint32_t j = 0; j < n_embd_v_gqa; ++j) {
-                        const size_t dst_offset = (kv_self.head + j * kv_self.size) * v_size_el;
-                        ggml_backend_tensor_set(kv_self.v_l[il], read(cell_count * v_size_el), dst_offset, cell_count * v_size_el);
-                    }
-                }
-            }
-        }
-        return true;
-    }
-
-    void read_kv_cache(struct llama_context * ctx, llama_seq_id seq_id = -1) {
-        uint32_t cell_count;
-        read_to(&cell_count, sizeof(cell_count));
-
-        bool res = read_kv_cache_meta(ctx, cell_count, seq_id) && read_kv_cache_data(ctx, cell_count);
-
-        if (!res) {
-            if (seq_id == -1) {
-                llama_kv_cache_clear(ctx);
-            } else {
-                llama_kv_cache_seq_rm(ctx, seq_id, -1, -1);
-            }
-            throw std::runtime_error("failed to restore kv cache");
-        }
-    }
-};
-
-struct llama_data_write_dummy : llama_data_write {
-    size_t size_written = 0;
-
-    llama_data_write_dummy() {}
-
-    void write(const void * /* src */, size_t size) override {
-        size_written += size;
-    }
-
-    void write_tensor_data(const struct ggml_tensor * /* tensor */, size_t /* offset */, size_t size) override {
-        size_written += size;
-    }
-
-    size_t get_size_written() override {
-        return size_written;
-    }
-};
-
-struct llama_data_write_buffer : llama_data_write {
-    uint8_t * ptr;
-    size_t buf_size = 0;
-    size_t size_written = 0;
-
-    llama_data_write_buffer(uint8_t * p, size_t len) : ptr(p), buf_size(len) {}
-
-    void write(const void * src, size_t size) override {
-        if (size > buf_size) {
-            throw std::runtime_error("unexpectedly reached end of buffer");
-        }
-        memcpy(ptr, src, size);
-        ptr += size;
-        size_written += size;
-        buf_size -= size;
-    }
-
-    void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) override {
-        if (size > buf_size) {
-            throw std::runtime_error("unexpectedly reached end of buffer");
-        }
-        ggml_backend_tensor_get(tensor, ptr, offset, size);
-        ptr += size;
-        size_written += size;
-        buf_size -= size;
-    }
-
-    size_t get_size_written() override {
-        return size_written;
-    }
-};
-
-struct llama_data_read_buffer : llama_data_read {
-    const uint8_t * ptr;
-    size_t buf_size = 0;
-    size_t size_read = 0;
-
-    llama_data_read_buffer(const uint8_t * p, size_t len) : ptr(p), buf_size(len) {}
-
-    const uint8_t * read(size_t size) override {
-        const uint8_t * base_ptr = ptr;
-        if (size > buf_size) {
-            throw std::runtime_error("unexpectedly reached end of buffer");
-        }
-        ptr += size;
-        size_read += size;
-        buf_size -= size;
-        return base_ptr;
-    }
-
-    void read_to(void * dst, size_t size) override {
-        memcpy(dst, read(size), size);
-    }
-
-    size_t get_size_read() override {
-        return size_read;
-    }
-};
-
-struct llama_data_write_file : llama_data_write {
-    llama_file * file;
-    size_t size_written = 0;
-    std::vector temp_buffer;
-
-    llama_data_write_file(llama_file * f) : file(f) {}
-
-    void write(const void * src, size_t size) override {
-        file->write_raw(src, size);
-        size_written += size;
-    }
-
-    void write_tensor_data(const struct ggml_tensor * tensor, size_t offset, size_t size) override {
-        temp_buffer.resize(size);
-        ggml_backend_tensor_get(tensor, temp_buffer.data(), offset, size);
-        write(temp_buffer.data(), temp_buffer.size());
-    }
-
-    size_t get_size_written() override {
-        return size_written;
-    }
-};
-
-struct llama_data_read_file : llama_data_read {
-    llama_file * file;
-    size_t size_read = 0;
-    std::vector temp_buffer;
-
-    llama_data_read_file(llama_file * f) : file(f) {}
-
-    void read_to(void * dst, size_t size) override {
-        file->read_raw(dst, size);
-        size_read += size;
-    }
-
-    const uint8_t * read(size_t size) override {
-        temp_buffer.resize(size);
-        read_to(temp_buffer.data(), size);
-        return temp_buffer.data();
-    }
-
-    size_t get_size_read() override {
-        return size_read;
-    }
-};
-
-/** copy state data into either a buffer or file depending on the passed in context
- *
- * file context:
- * llama_file file("/path", "wb");
- * llama_data_write_file data_ctx(&file);
- * llama_state_get_data_internal(ctx, data_ctx);
- *
- * buffer context:
- * std::vector buf(max_size, 0);
- * llama_data_write_buffer data_ctx(buf.data(), max_size);
- * llama_state_get_data_internal(ctx, data_ctx);
- *
-*/
-static size_t llama_state_get_data_internal(struct llama_context * ctx, llama_data_write & data_ctx) {
-    llama_synchronize(ctx);
-
-    data_ctx.write_model_info(ctx);
-
-    // copy outputs
-    data_ctx.write_output_ids(ctx);
-    data_ctx.write_logits(ctx);
-    data_ctx.write_embeddings(ctx);
-
-    data_ctx.write_kv_cache(ctx);
-
-    return data_ctx.get_size_written();
-}
-
-size_t llama_state_get_data(struct llama_context * ctx, uint8_t * dst, size_t size) {
-    llama_data_write_buffer data_ctx(dst, size);
-    try {
-        return llama_state_get_data_internal(ctx, data_ctx);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error saving state: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-// Returns the *actual* size of the state.
-// Intended to be used when saving to state to a buffer.
-size_t llama_state_get_size(struct llama_context * ctx) {
-    llama_data_write_dummy data_ctx;
-    try {
-        return llama_state_get_data_internal(ctx, data_ctx);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error getting state size: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-static size_t llama_state_set_data_internal(struct llama_context * ctx, llama_data_read & data_ctx) {
-    llama_synchronize(ctx);
-
-    data_ctx.read_model_info(ctx);
-
-    // set outputs
-    data_ctx.read_output_ids(ctx);
-    data_ctx.read_logits(ctx);
-    data_ctx.read_embeddings(ctx);
-
-    data_ctx.read_kv_cache(ctx);
-
-    return data_ctx.get_size_read();
-}
-
-// Sets the state reading from the specified source address
-size_t llama_state_set_data(struct llama_context * ctx, const uint8_t * src, size_t size) {
-    llama_data_read_buffer data_ctx(src, size);
-    try {
-        return llama_state_set_data_internal(ctx, data_ctx);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error loading state: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-static bool llama_state_load_file_internal(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) {
-    llama_file file(path_session, "rb");
-
-    // sanity checks
-    {
-        const uint32_t magic   = file.read_u32();
-        const uint32_t version = file.read_u32();
-
-        if (magic != LLAMA_SESSION_MAGIC || version != LLAMA_SESSION_VERSION) {
-            LLAMA_LOG_ERROR("%s: unknown (magic, version) for session file: %08x, %08x\n", __func__, magic, version);
-            return false;
-        }
-    }
-
-    // load the prompt
-    {
-        const uint32_t n_token_count = file.read_u32();
-
-        if (n_token_count > n_token_capacity) {
-            LLAMA_LOG_ERROR("%s: token count in session file exceeded capacity! %u > %zu\n", __func__, n_token_count, n_token_capacity);
-            return false;
-        }
-
-        file.read_raw(tokens_out, sizeof(llama_token) * n_token_count);
-        *n_token_count_out = n_token_count;
-    }
-
-    // restore the context state
-    {
-        const size_t n_state_size_cur = file.size - file.tell();
-
-        llama_data_read_file data_ctx(&file);
-        const size_t n_read = llama_state_set_data_internal(ctx, data_ctx);
-
-        if (n_read != n_state_size_cur) {
-            LLAMA_LOG_ERROR("%s: did not read all of the session file data! size %zu, got %zu\n", __func__, n_state_size_cur, n_read);
-            return false;
-        }
-    }
-    return true;
-}
-
-bool llama_state_load_file(struct llama_context * ctx, const char * path_session, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) {
-    try {
-        return llama_state_load_file_internal(ctx, path_session, tokens_out, n_token_capacity, n_token_count_out);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error loading session file: %s\n", __func__, err.what());
-        return false;
-    }
-}
-
-static bool llama_state_save_file_internal(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) {
-    llama_file file(path_session, "wb");
-
-    file.write_u32(LLAMA_SESSION_MAGIC);
-    file.write_u32(LLAMA_SESSION_VERSION);
-
-    // save the prompt
-    file.write_u32((uint32_t) n_token_count);
-    file.write_raw(tokens, sizeof(llama_token) * n_token_count);
-
-    // save the context state using stream saving
-    llama_data_write_file data_ctx(&file);
-    llama_state_get_data_internal(ctx, data_ctx);
-
-    return true;
-}
-
-bool llama_state_save_file(struct llama_context * ctx, const char * path_session, const llama_token * tokens, size_t n_token_count) {
-    try {
-        return llama_state_save_file_internal(ctx, path_session, tokens, n_token_count);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error saving session file: %s\n", __func__, err.what());
-        return false;
-    }
-}
-
-static size_t llama_state_seq_get_data_internal(struct llama_context * ctx, llama_data_write & data_ctx, llama_seq_id seq_id) {
-    llama_synchronize(ctx);
-
-    data_ctx.write_kv_cache(ctx, seq_id);
-
-    return data_ctx.get_size_written();
-}
-
-size_t llama_state_seq_get_size(struct llama_context * ctx, llama_seq_id seq_id) {
-    llama_data_write_dummy data_ctx;
-    return llama_state_seq_get_data_internal(ctx, data_ctx, seq_id);
-}
-
-size_t llama_state_seq_get_data(struct llama_context * ctx, uint8_t * dst, size_t size, llama_seq_id seq_id) {
-    llama_data_write_buffer data_ctx(dst, size);
-    try {
-        return llama_state_seq_get_data_internal(ctx, data_ctx, seq_id);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error saving sequence state: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-static size_t llama_state_seq_set_data_internal(struct llama_context * ctx, llama_data_read & data_ctx, llama_seq_id dest_seq_id) {
-    llama_synchronize(ctx);
-
-    data_ctx.read_kv_cache(ctx, dest_seq_id);
-
-    return data_ctx.get_size_read();
-}
-
-size_t llama_state_seq_set_data(struct llama_context * ctx, const uint8_t * src, size_t size, llama_seq_id dest_seq_id) {
-    llama_data_read_buffer data_ctx(src, size);
-    try {
-        return llama_state_seq_set_data_internal(ctx, data_ctx, dest_seq_id);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error loading sequence state: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-static size_t llama_state_seq_save_file_internal(struct llama_context * ctx, const char * filepath, llama_seq_id seq_id, const llama_token * tokens, size_t n_token_count) {
-    llama_file file(filepath, "wb");
-
-    file.write_u32(LLAMA_STATE_SEQ_MAGIC);
-    file.write_u32(LLAMA_STATE_SEQ_VERSION);
-
-    // save the prompt
-    file.write_u32((uint32_t) n_token_count);
-    file.write_raw(tokens, sizeof(llama_token) * n_token_count);
-
-    // save the context state using stream saving
-    llama_data_write_file data_ctx(&file);
-    llama_state_seq_get_data_internal(ctx, data_ctx, seq_id);
-
-    const size_t res = file.tell();
-    GGML_ASSERT(res == sizeof(uint32_t) * 3 + sizeof(llama_token) * n_token_count + data_ctx.get_size_written());
-    return res;
-}
-
-static size_t llama_state_seq_load_file_internal(struct llama_context * ctx, const char * filepath, llama_seq_id dest_seq_id, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) {
-    llama_file file(filepath, "rb");
-
-    // version checks
-    {
-        const uint32_t magic   = file.read_u32();
-        const uint32_t version = file.read_u32();
-
-        if (magic != LLAMA_STATE_SEQ_MAGIC || version != LLAMA_STATE_SEQ_VERSION) {
-            LLAMA_LOG_ERROR("%s: unknown (magic, version) for sequence state file: %08x, %08x\n", __func__, magic, version);
-            return 0;
-        }
-    }
-
-    // load the prompt
-    {
-        const uint32_t n_token_count = file.read_u32();
-
-        if (n_token_count > n_token_capacity) {
-            LLAMA_LOG_ERROR("%s: token count in sequence state file exceeded capacity! %u > %zu\n", __func__, n_token_count, n_token_capacity);
-            return 0;
-        }
-
-        file.read_raw(tokens_out, sizeof(llama_token) * n_token_count);
-        *n_token_count_out = n_token_count;
-    }
-
-    // restore the context state
-    {
-        const size_t state_size = file.size - file.tell();
-        llama_data_read_file data_ctx(&file);
-        const size_t nread = llama_state_seq_set_data_internal(ctx, data_ctx, dest_seq_id);
-        if (!nread) {
-            LLAMA_LOG_ERROR("%s: failed to restore sequence state\n", __func__);
-            return 0;
-        }
-        GGML_ASSERT(nread <= state_size);
-        GGML_ASSERT(nread + sizeof(uint32_t) * 3 + sizeof(llama_token) * *n_token_count_out == file.tell());
-    }
-
-    return file.tell();
-}
-
-size_t llama_state_seq_save_file(struct llama_context * ctx, const char * filepath, llama_seq_id seq_id, const llama_token * tokens, size_t n_token_count) {
-    try {
-        return llama_state_seq_save_file_internal(ctx, filepath, seq_id, tokens, n_token_count);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error saving sequence state file: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-size_t llama_state_seq_load_file(struct llama_context * ctx, const char * filepath, llama_seq_id dest_seq_id, llama_token * tokens_out, size_t n_token_capacity, size_t * n_token_count_out) {
-    try {
-        return llama_state_seq_load_file_internal(ctx, filepath, dest_seq_id, tokens_out, n_token_capacity, n_token_count_out);
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: error loading sequence state file: %s\n", __func__, err.what());
-        return 0;
-    }
-}
-
-void llama_set_n_threads(struct llama_context * ctx, int32_t n_threads, int32_t n_threads_batch) {
-    ctx->cparams.n_threads       = n_threads;
-    ctx->cparams.n_threads_batch = n_threads_batch;
-}
-
-int32_t llama_n_threads(struct llama_context * ctx) {
-    return ctx->cparams.n_threads;
-}
-
-int32_t llama_n_threads_batch(struct llama_context * ctx) {
-    return ctx->cparams.n_threads_batch;
-}
-
-void llama_set_abort_callback(struct llama_context * ctx, bool (*abort_callback)(void * data), void * abort_callback_data) {
-    ctx->abort_callback      = abort_callback;
-    ctx->abort_callback_data = abort_callback_data;
-}
-
-void llama_set_embeddings(struct llama_context * ctx, bool embeddings) {
-    ctx->cparams.embeddings = embeddings;
-}
-
-void llama_set_causal_attn(struct llama_context * ctx, bool causal_attn) {
-    ctx->cparams.causal_attn = causal_attn;
-}
-
-struct llama_batch llama_batch_get_one(
-             llama_token * tokens,
-                 int32_t   n_tokens) {
-    return {
-        /*n_tokens       =*/ n_tokens,
-        /*tokens         =*/ tokens,
-        /*embd           =*/ nullptr,
-        /*pos            =*/ nullptr,
-        /*n_seq_id       =*/ nullptr,
-        /*seq_id         =*/ nullptr,
-        /*logits         =*/ nullptr,
-    };
-}
-
-struct llama_batch llama_batch_init(int32_t n_tokens_alloc, int32_t embd, int32_t n_seq_max) {
-    llama_batch batch = {
-        /*n_tokens       =*/ 0,
-        /*tokens         =*/ nullptr,
-        /*embd           =*/ nullptr,
-        /*pos            =*/ nullptr,
-        /*n_seq_id       =*/ nullptr,
-        /*seq_id         =*/ nullptr,
-        /*logits         =*/ nullptr,
-    };
-
-    if (embd) {
-        batch.embd = (float *) malloc(sizeof(float) * n_tokens_alloc * embd);
-    } else {
-        batch.token = (llama_token *) malloc(sizeof(llama_token) * n_tokens_alloc);
-    }
-
-    batch.pos      = (llama_pos *)     malloc(sizeof(llama_pos)      * n_tokens_alloc);
-    batch.n_seq_id = (int32_t *)       malloc(sizeof(int32_t)        * n_tokens_alloc);
-    batch.seq_id   = (llama_seq_id **) malloc(sizeof(llama_seq_id *) * (n_tokens_alloc + 1));
-    for (int i = 0; i < n_tokens_alloc; ++i) {
-        batch.seq_id[i] = (llama_seq_id *) malloc(sizeof(llama_seq_id) * n_seq_max);
-    }
-    batch.seq_id[n_tokens_alloc] = nullptr;
-
-    batch.logits   = (int8_t *)        malloc(sizeof(int8_t)         * n_tokens_alloc);
-
-    return batch;
-}
-
-void llama_batch_free(struct llama_batch batch) {
-    if (batch.token)    free(batch.token);
-    if (batch.embd)     free(batch.embd);
-    if (batch.pos)      free(batch.pos);
-    if (batch.n_seq_id) free(batch.n_seq_id);
-    if (batch.seq_id) {
-        for (int i = 0; batch.seq_id[i] != nullptr; ++i) {
-            free(batch.seq_id[i]);
-        }
-        free(batch.seq_id);
-    }
-    if (batch.logits)   free(batch.logits);
-}
+///
 
 int32_t llama_encode(
         struct llama_context * ctx,
           struct llama_batch   batch) {
-    const int ret = llama_encode_internal(*ctx, batch);
+    const int ret = llama_encode_impl(*ctx, batch);
     if (ret != 0) {
         LLAMA_LOG_ERROR("%s: failed to encode, ret = %d\n", __func__, ret);
     }
@@ -21244,7 +9947,7 @@ int32_t llama_encode(
 int32_t llama_decode(
         struct llama_context * ctx,
           struct llama_batch   batch) {
-    const int ret = llama_decode_internal(*ctx, batch);
+    const int ret = llama_decode_impl(*ctx, batch);
     if (ret != 0) {
         LLAMA_LOG_ERROR("%s: failed to decode, ret = %d\n", __func__, ret);
     }
@@ -21252,592 +9955,18 @@ int32_t llama_decode(
     return ret;
 }
 
-void llama_synchronize(struct llama_context * ctx) {
-    ggml_backend_sched_synchronize(ctx->sched.get());
-
-    // FIXME: if multiple single tokens are evaluated without a synchronization,
-    // the stats will be added to the prompt evaluation stats
-    // this should only happen when using batch size 1 to evaluate a batch
-
-    // add the evaluation to the stats
-    if (ctx->n_queued_tokens == 1) {
-        if (!ctx->cparams.no_perf) {
-            ctx->t_eval_us += ggml_time_us() - ctx->t_compute_start_us;
-        }
-        ctx->n_eval++;
-    } else if (ctx->n_queued_tokens > 1) {
-        if (!ctx->cparams.no_perf) {
-            ctx->t_p_eval_us += ggml_time_us() - ctx->t_compute_start_us;
-        }
-        ctx->n_p_eval += ctx->n_queued_tokens;
-    }
-
-    // get a more accurate load time, upon first eval
-    if (ctx->n_queued_tokens > 0 && !ctx->has_evaluated_once) {
-        ctx->t_load_us = ggml_time_us() - ctx->t_start_us;
-        ctx->has_evaluated_once = true;
-    }
-
-    ctx->n_queued_tokens = 0;
-    ctx->t_compute_start_us = 0;
-}
-
-float * llama_get_logits(struct llama_context * ctx) {
-    llama_synchronize(ctx);
-
-    // reorder logits for backward compatibility
-    // TODO: maybe deprecate this
-    llama_output_reorder(ctx);
-
-    return ctx->logits;
-}
-
-float * llama_get_logits_ith(struct llama_context * ctx, int32_t i) {
-    int32_t j = -1;
-    llama_synchronize(ctx);
-
-    try {
-        if (ctx->logits == nullptr) {
-            throw std::runtime_error("no logits");
-        }
-
-        if (i < 0) {
-            j = ctx->n_outputs + i;
-            if (j < 0) {
-                throw std::runtime_error(format("negative index out of range [0, %d)", ctx->n_outputs));
-            }
-        } else if ((size_t) i >= ctx->output_ids.size()) {
-            throw std::runtime_error(format("out of range [0, %zu)", ctx->output_ids.size()));
-        } else {
-            j = ctx->output_ids[i];
-        }
-
-        if (j < 0) {
-            throw std::runtime_error(format("batch.logits[%d] != true", i));
-        }
-        if (j >= ctx->n_outputs) {
-            // This should not happen
-            throw std::runtime_error(format("corrupt output buffer (j=%d, n_outputs=%d)", j, ctx->n_outputs));
-        }
-
-        return ctx->logits + j*ctx->model.hparams.n_vocab;
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: invalid logits id %d, reason: %s\n", __func__, i, err.what());
-#ifndef NDEBUG
-        GGML_ABORT("fatal error");
-#else
-        return nullptr;
-#endif
-    }
-}
-
-float * llama_get_embeddings(struct llama_context * ctx) {
-    llama_synchronize(ctx);
-
-    // reorder embeddings for backward compatibility
-    // TODO: maybe deprecate this
-    llama_output_reorder(ctx);
-
-    return ctx->embd;
-}
-
-float * llama_get_embeddings_ith(struct llama_context * ctx, int32_t i) {
-    int32_t j = -1;
-
-    llama_synchronize(ctx);
-
-    try {
-        if (ctx->embd == nullptr) {
-            throw std::runtime_error("no embeddings");
-        }
-
-        if (i < 0) {
-            j = ctx->n_outputs + i;
-            if (j < 0) {
-                throw std::runtime_error(format("negative index out of range [0, %d)", ctx->n_outputs));
-            }
-        } else if ((size_t) i >= ctx->output_ids.size()) {
-            throw std::runtime_error(format("out of range [0, %lu)", ctx->output_ids.size()));
-        } else {
-            j = ctx->output_ids[i];
-        }
-
-        if (j < 0) {
-            throw std::runtime_error(format("batch.logits[%d] != true", i));
-        }
-        if (j >= ctx->n_outputs) {
-            // This should not happen
-            throw std::runtime_error(format("corrupt output buffer (j=%d, n_outputs=%d)", j, ctx->n_outputs));
-        }
-
-        return ctx->embd + j*ctx->model.hparams.n_embd;
-    } catch (const std::exception & err) {
-        LLAMA_LOG_ERROR("%s: invalid embeddings id %d, reason: %s\n", __func__, i, err.what());
-#ifndef NDEBUG
-        GGML_ABORT("fatal error");
-#else
-        return nullptr;
-#endif
-    }
-}
-
-float * llama_get_embeddings_seq(struct llama_context * ctx, llama_seq_id seq_id) {
-    llama_synchronize(ctx);
-
-    auto it = ctx->embd_seq.find(seq_id);
-    if (it == ctx->embd_seq.end()) {
-        return nullptr;
-    }
-
-    return it->second.data();
-}
-
-//
-// vocab
-//
-
-const char * llama_token_get_text(const struct llama_model * model, llama_token token) {
-    return llama_token_get_text_impl(model->vocab, token);
-}
-
-float llama_token_get_score(const struct llama_model * model, llama_token token) {
-    return llama_token_get_score_impl(model->vocab, token);
-}
-
-enum llama_token_attr llama_token_get_attr(const struct llama_model * model, llama_token token) {
-    return llama_token_get_attr_impl(model->vocab, token);
-}
-
-bool llama_token_is_eog(const struct llama_model * model, llama_token token) {
-    return llama_token_is_eog_impl(model->vocab, token);
-}
-
-bool llama_token_is_control(const struct llama_model * model, llama_token token) {
-    return llama_token_is_control_impl(model->vocab, token);
-}
-
-llama_token llama_token_bos(const struct llama_model * model) {
-    return llama_token_bos_impl(model->vocab);
-}
-
-llama_token llama_token_eos(const struct llama_model * model) {
-    return llama_token_eos_impl(model->vocab);
-}
-
-llama_token llama_token_eot(const struct llama_model * model) {
-    return llama_token_eot_impl(model->vocab);
-}
-
-llama_token llama_token_cls(const struct llama_model * model) {
-    return llama_token_cls_impl(model->vocab);
-}
-
-llama_token llama_token_sep(const struct llama_model * model) {
-    return llama_token_sep_impl(model->vocab);
-}
-
-llama_token llama_token_nl (const struct llama_model * model) {
-    return llama_token_nl_impl(model->vocab);
-}
-
-llama_token llama_token_pad(const struct llama_model * model) {
-    return llama_token_pad_impl(model->vocab);
-}
-
-bool llama_add_bos_token(const struct llama_model * model) {
-    return llama_add_bos_token_impl(model->vocab);
-}
-
-bool llama_add_eos_token(const struct llama_model * model) {
-    return llama_add_eos_token_impl(model->vocab);
-}
-
-llama_token llama_token_prefix(const struct llama_model * model) {
-    return llama_token_prefix_impl(model->vocab);
-}
-
-llama_token llama_token_middle(const struct llama_model * model) {
-    return llama_token_middle_impl(model->vocab);
-}
-
-llama_token llama_token_suffix(const struct llama_model * model) {
-    return llama_token_suffix_impl(model->vocab);
-}
-
-llama_token llama_token_fim_pre(const struct llama_model * model) {
-    return llama_token_fim_pre_impl(model->vocab);
-}
-
-llama_token llama_token_fim_suf(const struct llama_model * model) {
-    return llama_token_fim_suf_impl(model->vocab);
-}
-
-llama_token llama_token_fim_mid(const struct llama_model * model) {
-    return llama_token_fim_mid_impl(model->vocab);
-}
-
-llama_token llama_token_fim_pad(const struct llama_model * model) {
-    return llama_token_fim_pad_impl(model->vocab);
-}
-
-llama_token llama_token_fim_rep(const struct llama_model * model) {
-    return llama_token_fim_rep_impl(model->vocab);
-}
-
-llama_token llama_token_fim_sep(const struct llama_model * model) {
-    return llama_token_fim_sep_impl(model->vocab);
-}
-
-//
-// tokenization
-//
-
-int32_t llama_tokenize(
-    const struct llama_model * model,
-                  const char * text,
-                     int32_t   text_len,
-                 llama_token * tokens,
-                     int32_t   n_tokens_max,
-                        bool   add_special,
-                        bool   parse_special) {
-    return llama_tokenize_impl(model->vocab, text, text_len, tokens, n_tokens_max, add_special, parse_special);
-}
-
-int32_t llama_token_to_piece(
-    const struct llama_model * model,
-                 llama_token   token,
-                        char * buf,
-                     int32_t   length,
-                     int32_t   lstrip,
-                        bool   special) {
-    return llama_token_to_piece_impl(model->vocab, token, buf, length, lstrip, special);
-}
-
-int32_t llama_detokenize(
-    const struct llama_model * model,
-           const llama_token * tokens,
-                     int32_t   n_tokens,
-                        char * text,
-                     int32_t   text_len_max,
-                        bool   remove_special,
-                        bool   unparse_special) {
-    return llama_detokenize_impl(model->vocab, tokens, n_tokens, text, text_len_max, remove_special, unparse_special);
-}
-
 //
 // chat templates
 //
 
-// Simple version of "llama_apply_chat_template" that only works with strings
-// This function uses heuristic checks to determine commonly used template. It is not a jinja parser.
-static int32_t llama_chat_apply_template_internal(
-    const std::string & tmpl,
-    const std::vector & chat,
-    std::string & dest, bool add_ass) {
-    // Taken from the research: https://github.com/ggerganov/llama.cpp/issues/5527
-    std::stringstream ss;
-    auto tmpl_contains = [&tmpl](std::string haystack) -> bool {
-        return tmpl.find(haystack) != std::string::npos;
-    };
-    if (tmpl == "chatml" || tmpl_contains("<|im_start|>")) {
-        // chatml template
-        for (auto message : chat) {
-            ss << "<|im_start|>" << message->role << "\n" << message->content << "<|im_end|>\n";
-        }
-        if (add_ass) {
-            ss << "<|im_start|>assistant\n";
-        }
-    } else if (tmpl == "llama2" || tmpl == "mistral" || tmpl_contains("[INST]")) {
-        // llama2 template and its variants
-        // [variant] support system message
-        bool support_system_message = tmpl_contains("<>") || tmpl == "mistral";
-        // [variant] space before + after response
-        bool space_around_response = tmpl_contains("' ' + eos_token");
-        // [variant] add BOS inside history
-        bool add_bos_inside_history = tmpl_contains("bos_token + '[INST]");
-        // [variant] trim spaces from the input message
-        bool strip_message = tmpl_contains("content.strip()");
-        // construct the prompt
-        bool is_inside_turn = true; // skip BOS at the beginning
-        ss << "[INST] ";
-        for (auto message : chat) {
-            std::string content = strip_message ? trim(message->content) : message->content;
-            std::string role(message->role);
-            if (!is_inside_turn) {
-                is_inside_turn = true;
-                ss << (add_bos_inside_history ? "[INST] " : "[INST] ");
-            }
-            if (role == "system") {
-                if (support_system_message) {
-                    ss << "<>\n" << content << "\n<>\n\n";
-                } else {
-                    // if the model does not support system message, we still include it in the first message, but without <>
-                    ss << content << "\n";
-                }
-            } else if (role == "user") {
-                ss << content << " [/INST]";
-            } else {
-                ss << (space_around_response ? " " : "") << content << (space_around_response ? " " : "") << "";
-                is_inside_turn = false;
-            }
-        }
-        // llama2 templates seem to not care about "add_generation_prompt"
-    } else if (tmpl == "phi3" || (tmpl_contains("<|assistant|>") && tmpl_contains("<|end|>"))) {
-        // Phi 3
-        for (auto message : chat) {
-            std::string role(message->role);
-            ss << "<|" << role << "|>\n" << message->content << "<|end|>\n";
-        }
-        if (add_ass) {
-            ss << "<|assistant|>\n";
-        }
-    } else if (tmpl == "zephyr" || tmpl_contains("<|user|>")) {
-        // zephyr template
-        for (auto message : chat) {
-            ss << "<|" << message->role << "|>" << "\n" << message->content << "<|endoftext|>\n";
-        }
-        if (add_ass) {
-            ss << "<|assistant|>\n";
-        }
-    } else if (tmpl == "monarch" || tmpl_contains("bos_token + message['role']")) {
-        // mlabonne/AlphaMonarch-7B template (the  is included inside history)
-        for (auto message : chat) {
-            std::string bos = (message == chat.front()) ? "" : ""; // skip BOS for first message
-            ss << bos << message->role << "\n" << message->content << "\n";
-        }
-        if (add_ass) {
-            ss << "assistant\n";
-        }
-    } else if (tmpl == "gemma" || tmpl == "gemma2" || tmpl_contains("")) {
-        // google/gemma-7b-it
-        std::string system_prompt = "";
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                // there is no system message for gemma, but we will merge it with user prompt, so nothing is broken
-                system_prompt = trim(message->content);
-                continue;
-            }
-            // in gemma, "assistant" is "model"
-            role = role == "assistant" ? "model" : message->role;
-            ss << "" << role << "\n";
-            if (!system_prompt.empty() && role != "model") {
-                ss << system_prompt << "\n\n";
-                system_prompt = "";
-            }
-            ss << trim(message->content) << "\n";
-        }
-        if (add_ass) {
-            ss << "model\n";
-        }
-    } else if (tmpl == "orion" || tmpl_contains("'\\n\\nAssistant: ' + eos_token")) {
-        // OrionStarAI/Orion-14B-Chat
-        std::string system_prompt = "";
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                // there is no system message support, we will merge it with user prompt
-                system_prompt = message->content;
-                continue;
-            } else if (role == "user") {
-                ss << "Human: ";
-                if (!system_prompt.empty()) {
-                    ss << system_prompt << "\n\n";
-                    system_prompt = "";
-                }
-                ss << message->content << "\n\nAssistant: ";
-            } else {
-                ss << message->content << "";
-            }
-        }
-    } else if (tmpl == "openchat" || tmpl_contains("GPT4 Correct ")) {
-        // openchat/openchat-3.5-0106,
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                ss << message->content << "<|end_of_turn|>";
-            } else {
-                role[0] = toupper(role[0]);
-                ss << "GPT4 Correct " << role << ": " << message->content << "<|end_of_turn|>";
-            }
-        }
-        if (add_ass) {
-            ss << "GPT4 Correct Assistant:";
-        }
-    } else if (tmpl == "vicuna" || tmpl == "vicuna-orca" || (tmpl_contains("USER: ") && tmpl_contains("ASSISTANT: "))) {
-        // eachadea/vicuna-13b-1.1 (and Orca variant)
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                // Orca-Vicuna variant uses a system prefix
-                if (tmpl == "vicuna-orca" || tmpl_contains("SYSTEM: ")) {
-                    ss << "SYSTEM: " << message->content << "\n";
-                } else {
-                    ss << message->content << "\n\n";
-                }
-            } else if (role == "user") {
-                ss << "USER: " << message->content << "\n";
-            } else if (role == "assistant") {
-                ss << "ASSISTANT: " << message->content << "\n";
-            }
-        }
-        if (add_ass) {
-            ss << "ASSISTANT:";
-        }
-    } else if (tmpl == "deepseek" || (tmpl_contains("### Instruction:") && tmpl_contains("<|EOT|>"))) {
-        // deepseek-ai/deepseek-coder-33b-instruct
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                ss << message->content;
-            } else if (role == "user") {
-                ss << "### Instruction:\n" << message->content << "\n";
-            } else if (role == "assistant") {
-                ss << "### Response:\n" << message->content << "\n<|EOT|>\n";
-            }
-        }
-        if (add_ass) {
-            ss << "### Response:\n";
-        }
-    } else if (tmpl == "command-r" || (tmpl_contains("<|START_OF_TURN_TOKEN|>") && tmpl_contains("<|USER_TOKEN|>"))) {
-        // CohereForAI/c4ai-command-r-plus
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                ss << "<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>";
-            } else if (role == "user") {
-                ss << "<|START_OF_TURN_TOKEN|><|USER_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>";
-            } else if (role == "assistant") {
-                ss << "<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>" << trim(message->content) << "<|END_OF_TURN_TOKEN|>";
-            }
-        }
-        if (add_ass) {
-            ss << "<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>";
-        }
-    } else if (tmpl == "llama3" || (tmpl_contains("<|start_header_id|>") && tmpl_contains("<|end_header_id|>"))) {
-        // Llama 3
-        for (auto message : chat) {
-            std::string role(message->role);
-            ss << "<|start_header_id|>" << role << "<|end_header_id|>\n\n" << trim(message->content) << "<|eot_id|>";
-        }
-        if (add_ass) {
-            ss << "<|start_header_id|>assistant<|end_header_id|>\n\n";
-        }
-    } else if (tmpl == "chatglm3" || tmpl_contains("[gMASK]sop")) {
-        // chatglm3-6b
-        ss << "[gMASK]" << "sop";
-        for (auto message : chat) {
-            std::string role(message->role);
-            ss << "<|" << role << "|>" << "\n " << message->content;
-        }
-        if (add_ass) {
-            ss << "<|assistant|>";
-        }
-    } else if (tmpl == "chatglm4" || tmpl_contains("[gMASK]")) {
-        ss << "[gMASK]" << "";
-        for (auto message : chat) {
-            std::string role(message->role);
-            ss << "<|" << role << "|>" << "\n" << message->content;
-        }
-        if (add_ass) {
-            ss << "<|assistant|>";
-        }
-    } else if (tmpl == "minicpm" || tmpl_contains(LU8("<用户>"))) {
-        // MiniCPM-3B-OpenHermes-2.5-v2-GGUF
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "user") {
-                ss << LU8("<用户>");
-                ss << trim(message->content);
-                ss << "";
-            } else {
-                ss << trim(message->content);
-            }
-        }
-    } else if (tmpl == "deepseek2" || tmpl_contains("'Assistant: ' + message['content'] + eos_token")) {
-        // DeepSeek-V2
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                ss << message->content << "\n\n";
-            } else if (role == "user") {
-                ss << "User: " << message->content << "\n\n";
-            } else if (role == "assistant") {
-                ss << "Assistant: " << message->content << LU8("<｜end▁of▁sentence｜>");
-            }
-        }
-        if (add_ass) {
-            ss << "Assistant:";
-        }
-    } else if (tmpl == "exaone3" || (tmpl_contains("[|system|]") && tmpl_contains("[|assistant|]") && tmpl_contains("[|endofturn|]"))) {
-        // ref: https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct/discussions/8#66bae61b1893d14ee8ed85bb
-        // EXAONE-3.0-7.8B-Instruct
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "system") {
-                ss << "[|system|]" << trim(message->content) << "[|endofturn|]\n";
-            } else if (role == "user") {
-                ss << "[|user|]" << trim(message->content) << "\n";
-            } else if (role == "assistant") {
-                ss << "[|assistant|]" << trim(message->content) << "[|endofturn|]\n";
-            }
-        }
-        if (add_ass) {
-            ss << "[|assistant|]";
-        }
-    } else if (tmpl == "rwkv-world" || tmpl_contains("rwkv-world")) {
-        // this template requires the model to have "\n\n" as EOT token
-        for (auto message : chat) {
-            std::string role(message->role);
-            if (role == "user") {
-                ss << "User: " << message->content << "\n\nAssistant:";
-            } else {
-                ss << message->content << "\n\n";
-            }
-        }
-    } else if (tmpl == "granite" || tmpl_contains("<|start_of_role|>")) {
-        // IBM Granite template
-        for (const auto & message : chat) {
-            std::string role(message->role);
-            ss << "<|start_of_role|>" << role << "<|end_of_role|>";
-            if (role == "assistant_tool_call") {
-                ss << "<|tool_call|>";
-            }
-            ss << message->content << "<|end_of_text|>\n";
-        }
-        if (add_ass) {
-            ss << "<|start_of_role|>assistant<|end_of_role|>\n";
-        }
-    } else {
-        // template not supported
-        return -1;
-    }
-    dest = ss.str();
-    return dest.size();
-}
-
 int32_t llama_chat_apply_template(
-                const struct llama_model * model,
                               const char * tmpl,
          const struct llama_chat_message * chat,
                                   size_t   n_msg,
                                     bool   add_ass,
                                     char * buf,
                                  int32_t   length) {
-    std::string curr_tmpl(tmpl == nullptr ? "" : tmpl);
-    if (tmpl == nullptr) {
-        GGML_ASSERT(model != nullptr);
-        // load template from model
-        std::vector model_template(2048, 0); // longest known template is about 1200 bytes
-        std::string template_key = "tokenizer.chat_template";
-        int32_t res = llama_model_meta_val_str(model, template_key.c_str(), model_template.data(), model_template.size());
-        if (res < 0) {
-            // worst case: there is no information about template, we will use chatml by default
-            curr_tmpl = "chatml"; // see llama_chat_apply_template_internal
-        } else {
-            curr_tmpl = std::string(model_template.data(), model_template.size());
-        }
-    }
+    const std::string curr_tmpl(tmpl == nullptr ? "chatml" : tmpl);
 
     // format the chat to string
     std::vector chat_vec;
@@ -21847,7 +9976,11 @@ int32_t llama_chat_apply_template(
     }
 
     std::string formatted_chat;
-    int32_t res = llama_chat_apply_template_internal(curr_tmpl, chat_vec, formatted_chat, add_ass);
+    llm_chat_template detected_tmpl = llm_chat_detect_template(curr_tmpl);
+    if (detected_tmpl == LLM_CHAT_TEMPLATE_UNKNOWN) {
+        return -1;
+    }
+    int32_t res = llm_chat_apply_template(detected_tmpl, chat_vec, formatted_chat, add_ass);
     if (res < 0) {
         return res;
     }
@@ -21857,23 +9990,6 @@ int32_t llama_chat_apply_template(
     return res;
 }
 
-//
-// sampling
-//
-
-// TODO: remove indirection when vocab becomes accesible in llama-sampling.cpp
-struct llama_sampler * llama_sampler_init_grammar(const struct llama_model * model, const char * grammar_str, const char * grammar_root) {
-    return llama_sampler_init_grammar_impl(model->vocab, grammar_str, grammar_root);
-}
-
-struct llama_sampler * llama_sampler_init_infill(const struct llama_model * model) {
-    return llama_sampler_init_infill_impl(model->vocab);
-}
-
-struct llama_sampler * llama_sampler_init_dry(const struct llama_model * model, float dry_multiplier, float dry_base, int32_t dry_allowed_length, int32_t dry_penalty_last_n, const char** seq_breakers, size_t num_breakers) {
-    return llama_sampler_init_dry_impl(model->vocab, llama_n_ctx_train(model), dry_multiplier, dry_base, dry_allowed_length, dry_penalty_last_n, seq_breakers, num_breakers);
-}
-
 //
 // model split
 //
@@ -21886,16 +10002,16 @@ int llama_split_path(char * split_path, size_t maxlen, const char * path_prefix,
     return 0;
 }
 
-int llama_split_prefix(char * dest, size_t maxlen, const char * split_path, int split_no, int split_count) {
+int llama_split_prefix(char * split_prefix, size_t maxlen, const char * split_path, int split_no, int split_count) {
     std::string str_split_path(split_path);
     char postfix[32];
     snprintf(postfix, 32, "-%05d-of-%05d.gguf", split_no + 1, split_count);
     std::string str_postfix(postfix);
 
-    // check if dest ends with postfix
+    // check if split_prefix ends with postfix
     int size_prefix = str_split_path.size() - str_postfix.size();
     if (size_prefix > 0 && str_split_path.find(str_postfix, size_prefix) != std::string::npos) {
-        snprintf(dest, std::min((size_t) size_prefix + 1, maxlen), "%s", split_path);
+        snprintf(split_prefix, std::min((size_t) size_prefix + 1, maxlen), "%s", split_path);
         return size_prefix;
     }
 
@@ -21903,37 +10019,33 @@ int llama_split_prefix(char * dest, size_t maxlen, const char * split_path, int
 }
 
 const char * llama_print_system_info(void) {
-    ggml_cpu_init(); // some ARM features are detected at runtime
-
     static std::string s;
+    s.clear(); // Clear the string, since it's static, otherwise it will accumulate data from previous calls.
 
-    s  = "";
-    s += "AVX = "         + std::to_string(ggml_cpu_has_avx())         + " | ";
-    s += "AVX_VNNI = "    + std::to_string(ggml_cpu_has_avx_vnni())    + " | ";
-    s += "AVX2 = "        + std::to_string(ggml_cpu_has_avx2())        + " | ";
-    s += "AVX512 = "      + std::to_string(ggml_cpu_has_avx512())      + " | ";
-    s += "AVX512_VBMI = " + std::to_string(ggml_cpu_has_avx512_vbmi()) + " | ";
-    s += "AVX512_VNNI = " + std::to_string(ggml_cpu_has_avx512_vnni()) + " | ";
-    s += "AVX512_BF16 = " + std::to_string(ggml_cpu_has_avx512_bf16()) + " | ";
-    s += "AMX_INT8 = "    + std::to_string(ggml_cpu_has_amx_int8())    + " | ";
-    s += "FMA = "         + std::to_string(ggml_cpu_has_fma())         + " | ";
-    s += "NEON = "        + std::to_string(ggml_cpu_has_neon())        + " | ";
-    s += "SVE = "         + std::to_string(ggml_cpu_has_sve())         + " | ";
-    s += "ARM_FMA = "     + std::to_string(ggml_cpu_has_arm_fma())     + " | ";
-    s += "F16C = "        + std::to_string(ggml_cpu_has_f16c())        + " | ";
-    s += "FP16_VA = "     + std::to_string(ggml_cpu_has_fp16_va())     + " | ";
-    s += "RISCV_VECT = "  + std::to_string(ggml_cpu_has_riscv_v())     + " | ";
-    s += "WASM_SIMD = "   + std::to_string(ggml_cpu_has_wasm_simd())   + " | ";
-    s += "BLAS = "        + std::to_string(ggml_cpu_has_blas())        + " | ";
-    s += "SSE3 = "        + std::to_string(ggml_cpu_has_sse3())        + " | ";
-    s += "SSSE3 = "       + std::to_string(ggml_cpu_has_ssse3())       + " | ";
-    s += "VSX = "         + std::to_string(ggml_cpu_has_vsx())         + " | ";
-    s += "MATMUL_INT8 = " + std::to_string(ggml_cpu_has_matmul_int8()) + " | ";
-    s += "LLAMAFILE = "   + std::to_string(ggml_cpu_has_llamafile())   + " | ";
+
+    for (size_t i = 0; i < ggml_backend_reg_count(); i++) {
+        auto * reg = ggml_backend_reg_get(i);
+        auto * get_features_fn = (ggml_backend_get_features_t) ggml_backend_reg_get_proc_address(reg, "ggml_backend_get_features");
+        if (get_features_fn) {
+            ggml_backend_feature * features = get_features_fn(reg);
+            s += ggml_backend_reg_name(reg);
+            s += " : ";
+            for (; features->name; features++) {
+                s += features->name;
+                s += " = ";
+                s += features->value;
+                s += " | ";
+            }
+        }
+    }
 
     return s.c_str();
 }
 
+//
+// perf
+//
+
 struct llama_perf_context_data llama_perf_context(const struct llama_context * ctx) {
     struct llama_perf_context_data data = {};
 
@@ -21969,69 +10081,3 @@ void llama_perf_context_reset(struct llama_context * ctx) {
     ctx->t_eval_us   = ctx->n_eval = 0;
     ctx->t_p_eval_us = ctx->n_p_eval = 0;
 }
-
-void llama_perf_dump_yaml(FILE * stream, const llama_context * ctx) {
-    fprintf(stream, "\n");
-    fprintf(stream, "###########\n");
-    fprintf(stream, "# Timings #\n");
-    fprintf(stream, "###########\n");
-    fprintf(stream, "\n");
-
-    fprintf(stream, "mst_eval: %.2f  # ms / token during generation\n",
-            1.0e-3 * ctx->t_eval_us / ctx->n_eval);
-    fprintf(stream, "mst_p_eval: %.2f  # ms / token during prompt processing\n",
-            1.0e-3 * ctx->t_p_eval_us / ctx->n_p_eval);
-    fprintf(stream, "n_eval: %d  # number of tokens generated (excluding the first one)\n", ctx->n_eval);
-    fprintf(stream, "n_p_eval: %d  # number of tokens processed in batches at the beginning\n", ctx->n_p_eval);
-    fprintf(stream, "t_eval_us: %" PRId64 "  # total microseconds spent generating tokens\n", ctx->t_eval_us);
-    fprintf(stream, "t_load_us: %" PRId64 "  # total microseconds spent loading the model\n", ctx->t_load_us);
-    fprintf(stream, "t_p_eval_us: %" PRId64 "  # total microseconds spent prompt processing\n", ctx->t_p_eval_us);
-    fprintf(stream, "ts_eval: %.2f  # tokens / second during generation\n",
-            1.0e6 * ctx->n_eval / ctx->t_eval_us);
-    fprintf(stream, "ts_p_eval: %.2f  # tokens / second during prompt processing\n",
-            1.0e6 * ctx->n_p_eval / ctx->t_p_eval_us);
-}
-
-// For internal test use
-const std::vector> & llama_internal_get_tensor_map(
-    struct llama_context * ctx
-) {
-    return ctx->model.tensors_by_name;
-}
-
-void llama_log_set(ggml_log_callback log_callback, void * user_data) {
-    ggml_log_set(log_callback, user_data);
-    g_logger_state.log_callback = log_callback ? log_callback : llama_log_callback_default;
-    g_logger_state.log_callback_user_data = user_data;
-}
-
-static void llama_log_internal_v(ggml_log_level level, const char * format, va_list args) {
-    va_list args_copy;
-    va_copy(args_copy, args);
-    char buffer[128];
-    int len = vsnprintf(buffer, 128, format, args);
-    if (len < 128) {
-        g_logger_state.log_callback(level, buffer, g_logger_state.log_callback_user_data);
-    } else {
-        char * buffer2 = new char[len + 1];
-        vsnprintf(buffer2, len + 1, format, args_copy);
-        buffer2[len] = 0;
-        g_logger_state.log_callback(level, buffer2, g_logger_state.log_callback_user_data);
-        delete[] buffer2;
-    }
-    va_end(args_copy);
-}
-
-void llama_log_internal(ggml_log_level level, const char * format, ...) {
-    va_list args;
-    va_start(args, format);
-    llama_log_internal_v(level, format, args);
-    va_end(args);
-}
-
-void llama_log_callback_default(ggml_log_level level, const char * text, void * user_data) {
-    (void) level;
-    (void) user_data;
-    fputs(text, stderr);
-    fflush(stderr);
-}
diff --git a/src/unicode.cpp b/src/unicode.cpp
index 50b35bbbc..7aca6544b 100644
--- a/src/unicode.cpp
+++ b/src/unicode.cpp
@@ -71,15 +71,15 @@ uint32_t unicode_cpt_from_utf8(const std::string & utf8, size_t & offset) {
     throw std::invalid_argument("failed to convert utf8 to codepoint");
 }
 
-//static std::vector unicode_cpt_to_utf16(uint32_t cp) {
+//static std::vector unicode_cpt_to_utf16(uint32_t cpt) {
 //    std::vector result;
-//    if (/* 0x0000 <= cp && */ cp <= 0xffff) {
-//        result.emplace_back(cp);
+//    if (/* 0x0000 <= cpt && */ cpt <= 0xffff) {
+//        result.emplace_back(cpt);
 //        return result;
 //    }
-//    if (0x10000 <= cp && cp <= 0x10ffff) {
-//        result.emplace_back(0xd800 | ((cp - 0x10000) >> 10));
-//        result.emplace_back(0xdc00 | ((cp - 0x10000) & 0x03ff));
+//    if (0x10000 <= cpt && cpt <= 0x10ffff) {
+//        result.emplace_back(0xd800 | ((cpt - 0x10000) >> 10));
+//        result.emplace_back(0xdc00 | ((cpt - 0x10000) & 0x03ff));
 //        return result;
 //    }
 //    throw std::invalid_argument("failed to convert codepoint to utf16");
@@ -120,8 +120,8 @@ uint32_t unicode_cpt_from_utf8(const std::string & utf8, size_t & offset) {
 //    return result;
 //}
 
-static std::vector unicode_cpt_flags_array() {
-    std::vector cpt_flags(MAX_CODEPOINTS, codepoint_flags::UNDEFINED);
+static std::vector unicode_cpt_flags_array() {
+    std::vector cpt_flags(MAX_CODEPOINTS, unicode_cpt_flags::UNDEFINED);
 
     assert (unicode_ranges_flags.begin()[0].first == 0);
     assert (unicode_ranges_flags.begin()[unicode_ranges_flags.size()-1].first == MAX_CODEPOINTS);
@@ -201,7 +201,18 @@ static std::unordered_map unicode_utf8_to_byte_map() {
 }
 
 static inline std::wstring unicode_wstring_from_utf8(const std::string & s) {
+#if defined(__clang__)
+    // disable C++17 deprecation warning for std::codecvt_utf8
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wdeprecated-declarations"
+#endif
+
     std::wstring_convert> conv;
+
+#if defined(__clang__)
+#    pragma clang diagnostic pop
+#endif
+
     return conv.from_bytes(s);
 }
 
@@ -242,8 +253,8 @@ static std::vector unicode_regex_split_custom_gpt2(const std::string & t
             return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE;
         };
 
-        auto _get_flags = [&] (const size_t pos) -> codepoint_flags {
-            return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{};
+        auto _get_flags = [&] (const size_t pos) -> unicode_cpt_flags {
+            return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags_from_cpt(cpts[pos]) : unicode_cpt_flags{};
         };
 
         size_t _prev_end = offset_ini;
@@ -360,8 +371,8 @@ static std::vector unicode_regex_split_custom_llama3(const std::string &
             return (offset_ini <= pos && pos < offset_end) ? cpts[pos] : OUT_OF_RANGE;
         };
 
-        auto _get_flags = [&] (const size_t pos) -> codepoint_flags {
-            return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags(cpts[pos]) : codepoint_flags{};
+        auto _get_flags = [&] (const size_t pos) -> unicode_cpt_flags {
+            return (offset_ini <= pos && pos < offset_end) ? unicode_cpt_flags_from_cpt(cpts[pos]) : unicode_cpt_flags{};
         };
 
         size_t _prev_end = offset_ini;
@@ -561,29 +572,29 @@ static std::vector unicode_regex_split_custom(const std::string & text,
 // interface
 //
 
-std::string unicode_cpt_to_utf8(uint32_t cp) {
+std::string unicode_cpt_to_utf8(uint32_t cpt) {
     std::string result;
 
-    if (/* 0x00 <= cp && */ cp <= 0x7f) {
-        result.push_back(cp);
+    if (/* 0x00 <= cpt && */ cpt <= 0x7f) {
+        result.push_back(cpt);
         return result;
     }
-    if (0x80 <= cp && cp <= 0x7ff) {
-        result.push_back(0xc0 | ((cp >> 6) & 0x1f));
-        result.push_back(0x80 | (cp & 0x3f));
+    if (0x80 <= cpt && cpt <= 0x7ff) {
+        result.push_back(0xc0 | ((cpt >> 6) & 0x1f));
+        result.push_back(0x80 | (cpt & 0x3f));
         return result;
     }
-    if (0x800 <= cp && cp <= 0xffff) {
-        result.push_back(0xe0 | ((cp >> 12) & 0x0f));
-        result.push_back(0x80 | ((cp >> 6) & 0x3f));
-        result.push_back(0x80 | (cp & 0x3f));
+    if (0x800 <= cpt && cpt <= 0xffff) {
+        result.push_back(0xe0 | ((cpt >> 12) & 0x0f));
+        result.push_back(0x80 | ((cpt >> 6) & 0x3f));
+        result.push_back(0x80 | (cpt & 0x3f));
         return result;
     }
-    if (0x10000 <= cp && cp <= 0x10ffff) {
-        result.push_back(0xf0 | ((cp >> 18) & 0x07));
-        result.push_back(0x80 | ((cp >> 12) & 0x3f));
-        result.push_back(0x80 | ((cp >> 6) & 0x3f));
-        result.push_back(0x80 | (cp & 0x3f));
+    if (0x10000 <= cpt && cpt <= 0x10ffff) {
+        result.push_back(0xf0 | ((cpt >> 18) & 0x07));
+        result.push_back(0x80 | ((cpt >> 12) & 0x3f));
+        result.push_back(0x80 | ((cpt >> 6) & 0x3f));
+        result.push_back(0x80 | (cpt & 0x3f));
         return result;
     }
 
@@ -613,19 +624,19 @@ std::vector unicode_cpts_from_utf8(const std::string & utf8) {
     return result;
 }
 
-codepoint_flags unicode_cpt_flags(const uint32_t cp) {
-    static const codepoint_flags undef(codepoint_flags::UNDEFINED);
+unicode_cpt_flags unicode_cpt_flags_from_cpt(const uint32_t cpt) {
+    static const unicode_cpt_flags undef(unicode_cpt_flags::UNDEFINED);
     static const auto cpt_flags = unicode_cpt_flags_array();
-    return cp < cpt_flags.size() ? cpt_flags[cp] : undef;
+    return cpt < cpt_flags.size() ? cpt_flags[cpt] : undef;
 }
 
-codepoint_flags unicode_cpt_flags(const std::string & utf8) {
-    static const codepoint_flags undef(codepoint_flags::UNDEFINED);
+unicode_cpt_flags unicode_cpt_flags_from_utf8(const std::string & utf8) {
+    static const unicode_cpt_flags undef(unicode_cpt_flags::UNDEFINED);
     if (utf8.empty()) {
         return undef;  // undefined
     }
     size_t offset = 0;
-    return unicode_cpt_flags(unicode_cpt_from_utf8(utf8, offset));
+    return unicode_cpt_flags_from_cpt(unicode_cpt_from_utf8(utf8, offset));
 }
 
 std::string unicode_byte_to_utf8(uint8_t byte) {
@@ -638,41 +649,47 @@ uint8_t unicode_utf8_to_byte(const std::string & utf8) {
     return map.at(utf8);
 }
 
-uint32_t unicode_tolower(uint32_t cp) {
+uint32_t unicode_tolower(uint32_t cpt) {
     // binary search
-    auto it = std::lower_bound(unicode_map_lowercase.begin(), unicode_map_lowercase.end(), cp,
+    auto it = std::lower_bound(unicode_map_lowercase.begin(), unicode_map_lowercase.end(), cpt,
         [](const std::pair & pair, uint32_t value) {
             return pair.first < value;
         });
-    if (it != unicode_map_lowercase.end() && it->first == cp) {
+    if (it != unicode_map_lowercase.end() && it->first == cpt) {
         return it->second;
     }
-    return cp;  // Return the original code point if no lowercase mapping is found
+    return cpt;  // Return the original code point if no lowercase mapping is found
 }
 
 std::vector unicode_regex_split(const std::string & text, const std::vector & regex_exprs) {
     // unicode categories
     static const std::map k_ucat_enum = {
-        { "\\p{N}", codepoint_flags::NUMBER },
-        { "\\p{L}", codepoint_flags::LETTER },
-        { "\\p{P}", codepoint_flags::PUNCTUATION },
+        { "\\p{N}", unicode_cpt_flags::NUMBER },
+        { "\\p{L}", unicode_cpt_flags::LETTER },
+        { "\\p{P}", unicode_cpt_flags::PUNCTUATION },
+        { "\\p{M}", unicode_cpt_flags::ACCENT_MARK },
+        { "\\p{S}", unicode_cpt_flags::SYMBOL },
     };
 
     static const std::map k_ucat_cpt = {
-        { codepoint_flags::NUMBER,        0xD1 },
-        { codepoint_flags::LETTER,        0xD2 },
-        { codepoint_flags::PUNCTUATION,   0xD3 },
+        { unicode_cpt_flags::NUMBER,      0xD1 },
+        { unicode_cpt_flags::LETTER,      0xD2 },
+        { unicode_cpt_flags::PUNCTUATION, 0xD3 },
+        { unicode_cpt_flags::ACCENT_MARK, 0xD4 },
+        { unicode_cpt_flags::SYMBOL,      0xD5 },
     };
 
     static const std::map k_ucat_map = {
-        { codepoint_flags::NUMBER,        "\x30-\x39" }, // 0-9
-        { codepoint_flags::LETTER,        "\x41-\x5A\x61-\x7A" }, // A-Za-z
-        { codepoint_flags::PUNCTUATION,   "\x21-\x23\x25-\x2A\x2C-\x2F\x3A-\x3B\x3F-\x40\\\x5B-\\\x5D\x5F\\\x7B\\\x7D" }, // !-#%-*,-/:-;?-@\[-\]_\{\}
+        { unicode_cpt_flags::NUMBER,      "\x30-\x39" }, // 0-9
+        { unicode_cpt_flags::LETTER,      "\x41-\x5A\x61-\x7A" }, // A-Za-z
+        { unicode_cpt_flags::PUNCTUATION, "\x21-\x23\x25-\x2A\x2C-\x2F\x3A-\x3B\x3F-\x40\\\x5B-\\\x5D\x5F\\\x7B\\\x7D" }, // !-#%-*,-/:-;?-@\[-\]_\{\}
+        { unicode_cpt_flags::ACCENT_MARK, "" }, // no sub-128 codepoints
+        { unicode_cpt_flags::SYMBOL,      "\\\x24\\\x2B\x3C-\x3E\x5E\x60\\\x7C" }, // $+<=>^`|
     };
 
     // compute collapsed codepoints only if needed by at least one regex
     bool need_collapse = false;
-    for (auto & regex_expr : regex_exprs) {
+    for (const auto & regex_expr : regex_exprs) {
         // search for unicode categories
         for (const auto & ucat : k_ucat_enum) {
             if (std::string::npos != regex_expr.find(ucat.first)) {
@@ -698,7 +715,7 @@ std::vector unicode_regex_split(const std::string & text, const std
                 continue;
             }
 
-            const auto flags = unicode_cpt_flags(cpts[i]);
+            const auto flags = unicode_cpt_flags_from_cpt(cpts[i]);
 
             if (flags.is_whitespace) {
                 //NOTE: C++ std::regex \s does not mach 0x85, Rust and Python regex does.
@@ -714,7 +731,7 @@ std::vector unicode_regex_split(const std::string & text, const std
 
     std::vector bpe_offsets = { cpts.size() };
 
-    for (auto & regex_expr : regex_exprs) {
+    for (const auto & regex_expr : regex_exprs) {
         // first, see if we have an efficient custom regex implementation
         auto tmp = unicode_regex_split_custom(text, regex_expr, bpe_offsets);
 
@@ -728,7 +745,7 @@ std::vector unicode_regex_split(const std::string & text, const std
             // if a unicode category is used in the regex, we use the collapsed text and replace the unicode category
             // with the corresponding collapsed representation
             bool use_collapsed = false;
-            for (auto & ucat : k_ucat_enum) {
+            for (const auto & ucat : k_ucat_enum) {
                 if (std::string::npos != regex_expr.find(ucat.first)) {
                     use_collapsed = true;
                     break;
@@ -794,7 +811,7 @@ std::vector unicode_regex_split(const std::string & text, const std
                 // std::wregex \s does not mach non-ASCII whitespaces, using 0x0B as fallback
                 std::wstring wtext(cpts.begin(), cpts.end());
                 for (size_t i = 0; i < wtext.size(); ++i) {
-                    if (wtext[i] > 0x7F && unicode_cpt_flags(wtext[i]).is_whitespace) {
+                    if (wtext[i] > 0x7F && unicode_cpt_flags_from_cpt(wtext[i]).is_whitespace) {
                         wtext[i] = 0x0B;
                     }
                 }
diff --git a/src/unicode.h b/src/unicode.h
index 008532a24..c27098df7 100644
--- a/src/unicode.h
+++ b/src/unicode.h
@@ -4,9 +4,7 @@
 #include 
 #include 
 
-// TODO: prefix all symbols with "llama_"
-
-struct codepoint_flags {
+struct unicode_cpt_flags {
     enum {
         UNDEFINED       = 0x0001,
         NUMBER          = 0x0002,  // regex: \p{N}
@@ -35,7 +33,7 @@ struct codepoint_flags {
     uint16_t is_nfd         : 1;
 
     // decode from uint16
-    inline codepoint_flags(const uint16_t flags=0) {
+    inline unicode_cpt_flags(const uint16_t flags = 0) {
         *reinterpret_cast(this) = flags;
     }
 
@@ -50,18 +48,19 @@ struct codepoint_flags {
 
 size_t unicode_len_utf8(char src);
 
-std::string unicode_cpt_to_utf8(uint32_t cp);
-uint32_t unicode_cpt_from_utf8(const std::string & utf8, size_t & offset);
+std::string unicode_cpt_to_utf8  (uint32_t cpt);
+uint32_t    unicode_cpt_from_utf8(const std::string & utf8, size_t & offset);
+
 std::vector unicode_cpts_from_utf8(const std::string & utf8);
 
 std::vector unicode_cpts_normalize_nfd(const std::vector & cpts);
 
-codepoint_flags unicode_cpt_flags(const uint32_t cp);
-codepoint_flags unicode_cpt_flags(const std::string & utf8);
+unicode_cpt_flags unicode_cpt_flags_from_cpt (uint32_t cpt);
+unicode_cpt_flags unicode_cpt_flags_from_utf8(const std::string & utf8);
 
 std::string unicode_byte_to_utf8(uint8_t byte);
-uint8_t unicode_utf8_to_byte(const std::string & utf8);
+uint8_t     unicode_utf8_to_byte(const std::string & utf8);
 
-uint32_t unicode_tolower(uint32_t cp);
+uint32_t unicode_tolower(uint32_t cpt);
 
 std::vector unicode_regex_split(const std::string & text, const std::vector & regex_exprs);
diff --git a/tests/CMakeLists.txt b/tests/CMakeLists.txt
index 08ad66b49..2b5e5fd4a 100644
--- a/tests/CMakeLists.txt
+++ b/tests/CMakeLists.txt
@@ -84,56 +84,66 @@ llama_test(test-tokenizer-0 NAME test-tokenizer-0-qwen2             ARGS ${CMAKE
 llama_test(test-tokenizer-0 NAME test-tokenizer-0-refact            ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-refact.gguf)
 llama_test(test-tokenizer-0 NAME test-tokenizer-0-starcoder         ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
 
-# build test-tokenizer-1-bpe target once and add many tests
-add_executable(test-tokenizer-1-bpe test-tokenizer-1-bpe.cpp)
-target_link_libraries(test-tokenizer-1-bpe PRIVATE common)
-install(TARGETS test-tokenizer-1-bpe RUNTIME)
 
-# TODO: disabled due to slowness
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-aquila    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-aquila.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-falcon    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt-2     ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-2.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt-neox  ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-neox.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-llama-bpe ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-bpe.gguf --ignore-merges)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-mpt       ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-mpt.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-refact    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-refact.gguf)
-#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-starcoder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
+if (NOT WIN32)
+    # these tests are disabled on Windows because they use internal functions not exported with LLAMA_API
+    llama_target_and_test(test-sampling.cpp)
+    llama_target_and_test(test-grammar-parser.cpp)
+    llama_target_and_test(test-grammar-integration.cpp)
+    llama_target_and_test(test-llama-grammar.cpp)
+    # TODO: disabled on loongarch64 because the ggml-ci node lacks Python 3.8
+    if (NOT ${CMAKE_SYSTEM_PROCESSOR} MATCHES "loongarch64")
+        llama_target_and_test(test-json-schema-to-grammar.cpp   WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/..)
+        target_include_directories(test-json-schema-to-grammar PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/../examples/server)
+    endif()
 
-# build test-tokenizer-1-spm target once and add many tests
-add_executable(test-tokenizer-1-spm test-tokenizer-1-spm.cpp)
-target_link_libraries(test-tokenizer-1-spm PRIVATE common)
-install(TARGETS test-tokenizer-1-spm RUNTIME)
 
-llama_test(test-tokenizer-1-spm  NAME test-tokenizer-1-llama-spm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-spm.gguf)
-#llama_test(test-tokenizer-1-spm  NAME test-tokenizer-1-baichuan  ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-baichuan.gguf)
+    # build test-tokenizer-1-bpe target once and add many tests
+    add_executable(test-tokenizer-1-bpe test-tokenizer-1-bpe.cpp)
+    target_link_libraries(test-tokenizer-1-bpe PRIVATE common)
+    install(TARGETS test-tokenizer-1-bpe RUNTIME)
+
+    # TODO: disabled due to slowness
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-aquila    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-aquila.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-falcon    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt-2     ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-2.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt-neox  ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-neox.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-llama-bpe ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-bpe.gguf --ignore-merges)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-mpt       ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-mpt.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-refact    ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-refact.gguf)
+    #llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-starcoder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
+
+    # build test-tokenizer-1-spm target once and add many tests
+    add_executable(test-tokenizer-1-spm test-tokenizer-1-spm.cpp)
+    target_link_libraries(test-tokenizer-1-spm PRIVATE common)
+    install(TARGETS test-tokenizer-1-spm RUNTIME)
+
+    llama_test(test-tokenizer-1-spm  NAME test-tokenizer-1-llama-spm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-spm.gguf)
+    #llama_test(test-tokenizer-1-spm  NAME test-tokenizer-1-baichuan  ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-baichuan.gguf)
+
+    # llama_target_and_test(test-double-float.cpp) # SLOW
+endif()
 
-# llama_target_and_test(test-double-float.cpp) # SLOW
 llama_target_and_test(test-log.cpp)
 llama_target_and_test(test-arg-parser.cpp)
-llama_target_and_test(test-quantize-fns.cpp)
-llama_target_and_test(test-quantize-perf.cpp)
-llama_target_and_test(test-sampling.cpp)
 llama_target_and_test(test-chat-template.cpp)
 
-llama_target_and_test(test-grammar-parser.cpp)
-llama_target_and_test(test-llama-grammar.cpp)
-llama_target_and_test(test-grammar-integration.cpp)
-llama_target_and_test(test-grad0.cpp)
-llama_target_and_test(test-barrier.cpp)
 # llama_target_and_test(test-opt.cpp) # SLOW
+llama_target_and_test(test-gguf.cpp)
 llama_target_and_test(test-backend-ops.cpp)
 
-llama_target_and_test(test-rope.cpp)
-
 llama_target_and_test(test-model-load-cancel.cpp  LABEL "model")
 llama_target_and_test(test-autorelease.cpp        LABEL "model")
 
-# TODO: disabled on loongarch64 because the ggml-ci node lacks Python 3.8
-if (NOT ${CMAKE_SYSTEM_PROCESSOR} MATCHES "loongarch64")
-    llama_target_and_test(test-json-schema-to-grammar.cpp   WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/..)
-    target_include_directories(test-json-schema-to-grammar PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/../examples/server)
+if (NOT GGML_BACKEND_DL)
+    # these tests use the backends directly and cannot be built with dynamic loading
+    llama_target_and_test(test-barrier.cpp)
+    llama_target_and_test(test-quantize-fns.cpp)
+    llama_target_and_test(test-quantize-perf.cpp)
+    llama_target_and_test(test-rope.cpp)
 endif()
 
+
 # dummy executable - not installed
 get_filename_component(TEST_TARGET test-c.c NAME_WE)
 add_executable(${TEST_TARGET} test-c.c)
diff --git a/tests/test-arg-parser.cpp b/tests/test-arg-parser.cpp
index 3665238b5..69604b87c 100644
--- a/tests/test-arg-parser.cpp
+++ b/tests/test-arg-parser.cpp
@@ -70,7 +70,7 @@ int main(void) {
 
     // non-existence arg in specific example (--draft cannot be used outside llama-speculative)
     argv = {"binary_name", "--draft", "123"};
-    assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), params, LLAMA_EXAMPLE_SERVER));
+    assert(false == common_params_parse(argv.size(), list_str_to_char(argv).data(), params, LLAMA_EXAMPLE_EMBEDDING));
 
 
     printf("test-arg-parser: test valid usage\n\n");
@@ -96,7 +96,7 @@ int main(void) {
     // --draft cannot be used outside llama-speculative
     argv = {"binary_name", "--draft", "123"};
     assert(true == common_params_parse(argv.size(), list_str_to_char(argv).data(), params, LLAMA_EXAMPLE_SPECULATIVE));
-    assert(params.n_draft == 123);
+    assert(params.speculative.n_max == 123);
 
 // skip this part on windows, because setenv is not supported
 #ifdef _WIN32
diff --git a/tests/test-autorelease.cpp b/tests/test-autorelease.cpp
index 57fa00011..35b09aaea 100644
--- a/tests/test-autorelease.cpp
+++ b/tests/test-autorelease.cpp
@@ -13,10 +13,10 @@ int main(int argc, char ** argv) {
 
     std::thread([&model_path]() {
         llama_backend_init();
-        auto * model = llama_load_model_from_file(model_path, llama_model_default_params());
-        auto * ctx = llama_new_context_with_model(model, llama_context_default_params());
+        auto * model = llama_model_load_from_file(model_path, llama_model_default_params());
+        auto * ctx = llama_init_from_model(model, llama_context_default_params());
         llama_free(ctx);
-        llama_free_model(model);
+        llama_model_free(model);
         llama_backend_free();
     }).join();
 
diff --git a/tests/test-backend-ops.cpp b/tests/test-backend-ops.cpp
index 9d48a2717..4c8464d8b 100644
--- a/tests/test-backend-ops.cpp
+++ b/tests/test-backend-ops.cpp
@@ -16,7 +16,6 @@
 
 
 #include 
-#include 
 #include 
 #include 
 
@@ -26,7 +25,6 @@
 #include 
 #include 
 #include 
-#include 
 #include 
 #include 
 #include 
@@ -639,19 +637,20 @@ struct test_case {
 
         // determine number of runs
         int n_runs;
+        bool is_cpu = ggml_backend_dev_type(ggml_backend_get_device(backend)) == GGML_BACKEND_DEVICE_TYPE_CPU;
         if (op_flops(out) > 0) {
             // based on flops
             const uint64_t GFLOP = 1000 * 1000 * 1000;
             const uint64_t target_flops_cpu =   8ULL * GFLOP;
             const uint64_t target_flops_gpu = 100ULL * GFLOP;
-            uint64_t target_flops = ggml_backend_is_cpu(backend) ? target_flops_cpu : target_flops_gpu;
+            uint64_t target_flops = is_cpu ? target_flops_cpu : target_flops_gpu;
             n_runs = std::min(ggml_graph_size(gf) - ggml_graph_n_nodes(gf), target_flops / op_flops(out)) + 1;
         } else {
             // based on memory size
             const size_t GB = 1ULL << 30;
             const size_t target_size_cpu =  8 * GB;
             const size_t target_size_gpu = 32 * GB;
-            size_t target_size = ggml_backend_is_cpu(backend) ? target_size_cpu : target_size_gpu;
+            size_t target_size = is_cpu ? target_size_cpu : target_size_gpu;
             n_runs = std::min(ggml_graph_size(gf) - ggml_graph_n_nodes(gf), target_size / op_size(out)) + 1;
         }
 
@@ -681,6 +680,7 @@ struct test_case {
 
         // run
         int64_t total_time_us = 0;
+        int64_t total_mem = 0;
         int total_runs = 0;
         do {
             int64_t start_time = ggml_time_us();
@@ -688,6 +688,7 @@ struct test_case {
             int64_t end_time = ggml_time_us();
 
             total_time_us += end_time - start_time;
+            total_mem += mem;
             total_runs += n_runs;
         } while (total_time_us < 1000*1000); // run for at least 1 second
 
@@ -717,7 +718,7 @@ struct test_case {
         } else {
             printf("%8zu kB/run - \033[1;34m%7.2f GB/s\033[0m",
                 op_size(out) / 1024,
-                mem / (total_time_us / 1e6) / 1024.0 / 1024.0 / 1024.0);
+                total_mem / (total_time_us / 1e6) / 1024.0 / 1024.0 / 1024.0);
         }
         printf("\n");
 
@@ -809,15 +810,14 @@ struct test_case {
 
         ggml_build_forward_expand(gf, out);
         ggml_graph_cpy(gf, gb);
-        ggml_build_backward_expand(ctx, gf, gb, false);
+        ggml_build_backward_expand(ctx, ctx, gb, false);
         if (expect.size() != 1 || expect[0] != 0.0f) {
             GGML_ASSERT(ggml_graph_n_nodes(gb) > ggml_graph_n_nodes(gf));
             for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
-                GGML_ASSERT(!(t->flags & GGML_TENSOR_FLAG_PARAM) || t->grad->op != GGML_OP_NONE);
+                GGML_ASSERT(!(t->flags & GGML_TENSOR_FLAG_PARAM) || ggml_graph_get_grad(gb, t)->op != GGML_OP_NONE);
             }
         }
 
-        // TODO: refactor so that this check is only needed once
         for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
             if (!ggml_backend_supports_op(backend, t)) {
                 printf("not supported [%s] ", ggml_backend_name(backend));
@@ -860,7 +860,13 @@ struct test_case {
             const char * bn = ggml_backend_name(backend);
             const int64_t ne = ggml_nelements(t);
 
-            std::vector ga = tensor_to_float(t->grad);
+            std::vector ga;
+            struct ggml_tensor * grad = ggml_graph_get_grad(gb, t);
+            if (grad) {
+                ga = tensor_to_float(grad);
+            } else {
+                ga.resize(ne); // default value is 0.0f
+            }
 
             for (int64_t i = 0; i < ne; ++i) { // gradient algebraic
                 // check for nans
@@ -1147,6 +1153,26 @@ struct test_argmax : public test_case {
         return out;
     }
 
+    void initialize_tensors(ggml_context * ctx) override {
+        std::random_device rd;
+        std::default_random_engine rng(rd());
+        for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
+            if (t->type == GGML_TYPE_F32) {
+                // initialize with unique values to avoid ties
+                for (int64_t r = 0; r < ggml_nrows(t); r++) {
+                    std::vector data(t->ne[0]);
+                    for (int i = 0; i < t->ne[0]; i++) {
+                        data[i] = i;
+                    }
+                    std::shuffle(data.begin(), data.end(), rng);
+                    ggml_backend_tensor_set(t, data.data(), r * t->nb[1], t->ne[0] * sizeof(float));
+                }
+            } else {
+                init_tensor_uniform(t);
+            }
+        }
+    }
+
     double max_nmse_err() override {
         return 0.0;
     }
@@ -1633,17 +1659,46 @@ struct test_rwkv_wkv6 : public test_case {
 
     ggml_tensor * build_graph(ggml_context * ctx) override {
         const int64_t n_tokens = n_seq_tokens * n_seqs;
-        ggml_tensor * r   = ggml_new_tensor(ctx, type, 4, std::vector{ 1, head_size, head_count, n_tokens }.data());
-        ggml_tensor * k   = ggml_new_tensor(ctx, type, 4, std::vector{ head_size, 1, head_count, n_tokens }.data());
-        ggml_tensor * v   = ggml_new_tensor(ctx, type, 4, std::vector{ 1, head_size, head_count, n_tokens }.data());
+        ggml_tensor * r   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * k   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * v   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
         ggml_tensor * tf  = ggml_new_tensor(ctx, type, 2, std::vector{ head_size, head_count }.data());
-        ggml_tensor * td  = ggml_new_tensor(ctx, type, 4, std::vector{ 1, head_size, head_count, n_tokens }.data());
+        ggml_tensor * td  = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
         ggml_tensor * s   = ggml_new_tensor(ctx, type, 2, std::vector{ head_size * head_size * head_count, n_seqs }.data());
         ggml_tensor * out = ggml_rwkv_wkv6(ctx, k, v, r, tf, td, s);
         return out;
     }
 };
 
+// GGML_OP_GATED_LINEAR_ATTN
+struct test_gla : public test_case {
+    const ggml_type type;
+
+    const int64_t head_count;
+    const int64_t head_size;
+    const int64_t n_seq_tokens;
+    const int64_t n_seqs;
+
+    std::string vars() override {
+        return VARS_TO_STR5(type, head_count, head_size, n_seq_tokens, n_seqs);
+    }
+
+    test_gla(ggml_type type = GGML_TYPE_F32,
+            int64_t head_count = 32, int64_t head_size = 64, int64_t n_seq_tokens = 32, int64_t n_seqs = 32)
+        : type(type), head_count(head_count), head_size(head_size), n_seq_tokens(n_seq_tokens), n_seqs(n_seqs) {}
+
+    ggml_tensor * build_graph(ggml_context * ctx) override {
+        const int64_t n_tokens = n_seq_tokens * n_seqs;
+        ggml_tensor * q   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * k   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * v   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * g   = ggml_new_tensor(ctx, type, 3, std::vector{ head_size, head_count, n_tokens }.data());
+        ggml_tensor * s   = ggml_new_tensor(ctx, type, 2, std::vector{ head_size * head_size * head_count, n_seqs }.data());
+        ggml_tensor * out = ggml_gated_linear_attn(ctx, k, v, q, g, s, pow(head_size, -0.5));
+        return out;
+    }
+};
+
 // GGML_OP_MUL_MAT
 struct test_mul_mat : public test_case {
     const ggml_type type_a;
@@ -2137,7 +2192,7 @@ struct test_soft_max : public test_case {
 };
 
 
-// GGML_OP_ROPE
+// GGML_OP_ROPE + GGML_OP_ROPE_BACK
 struct test_rope : public test_case {
     const ggml_type type;
     const std::array ne_a;
@@ -2149,33 +2204,48 @@ struct test_rope : public test_case {
     float af; // attn_factor
     bool ff;
     int v; // view (1 : non-contiguous a)
+    bool forward;
 
     std::string vars() override {
+        // forward can be inferred from the op, does not need to be printed
         return VARS_TO_STR10(type, ne_a, n_dims, mode, n_ctx, fs, ef, af, ff, v);
     }
 
     test_rope(ggml_type type = GGML_TYPE_F32,
             std::array ne_a = {10, 5, 3, 1},
-            int n_dims = 10, int mode = 0, int n_ctx = 512, float fs = 1.0f, float ef = 0.0f, float af = 0.0f, bool ff = false, int v = 0)
-        : type(type), ne_a(ne_a), n_dims(n_dims), mode(mode), n_ctx(n_ctx), fs(fs), ef(ef), af(af), ff(ff), v(v) {}
+            int n_dims = 10, int mode = 0, int n_ctx = 512, float fs = 1.0f,
+            float ef = 0.0f, float af = 0.0f, bool ff = false, int v = 0, bool forward = true)
+        : type(type), ne_a(ne_a), n_dims(n_dims), mode(mode), n_ctx(n_ctx), fs(fs), ef(ef), af(af), ff(ff), v(v), forward(forward) {}
 
     ggml_tensor * build_graph(ggml_context * ctx) override {
         ggml_tensor * a;
         if (v & 1) {
             auto ne = ne_a; ne[0] *= 2; ne[1] *= 4; ne[2] *= 3;
             a = ggml_new_tensor(ctx, type, 4, ne.data());
-            ggml_set_param(ctx, a);
+            if (forward) {
+                ggml_set_param(ctx, a);
+            }
             ggml_set_name(a, "a");
 
             a = ggml_view_4d(ctx, a, ne_a[0], ne_a[1], ne_a[2], ne_a[3], a->nb[1], a->nb[2], a->nb[3], 0);
             ggml_set_name(a, "view_of_a");
         } else {
             a = ggml_new_tensor(ctx, type, 4, ne_a.data());
-            ggml_set_param(ctx, a);
+            if (forward) {
+                ggml_set_param(ctx, a);
+            }
             ggml_set_name(a, "a");
         }
 
-        ggml_tensor * pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ne_a[2]);
+        const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE;
+        const bool is_vision = mode == GGML_ROPE_TYPE_VISION;
+
+        ggml_tensor * pos;
+        if (is_mrope || is_vision) {
+            pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ne_a[2] * 4);
+        } else {
+            pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, ne_a[2]);
+        }
         ggml_set_name(pos, "pos");
 
         ggml_tensor * freq = nullptr;
@@ -2184,7 +2254,32 @@ struct test_rope : public test_case {
             ggml_set_name(freq, "freq");
         }
 
-        ggml_tensor * out = ggml_rope_ext(ctx, a, pos, freq, n_dims, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+        ggml_tensor * out;
+        if (is_mrope) {
+            if (is_vision) {
+                GGML_ASSERT(n_dims/4 > 0);
+                int rope_sections[4] = {n_dims/4, n_dims/4, 0, 0}; // Vision-RoPE only use first two dimension for image (x, y) coordinate
+                if (forward) {
+                    out = ggml_rope_multi     (ctx, a, pos, freq, n_dims/2, rope_sections, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+                } else {
+                    out = ggml_rope_multi_back(ctx, a, pos, freq, n_dims/2, rope_sections, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+                }
+            } else {
+                GGML_ASSERT(n_dims/3 > 0);
+                int rope_sections[4] = {n_dims/3, n_dims/3, n_dims/3, 0};
+                if (forward) {
+                    out = ggml_rope_multi     (ctx, a, pos, freq, n_dims, rope_sections, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+                } else {
+                    out = ggml_rope_multi_back(ctx, a, pos, freq, n_dims, rope_sections, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+                }
+            }
+        } else {
+            if (forward) {
+                out = ggml_rope_ext     (ctx, a, pos, freq, n_dims, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+            } else {
+                out = ggml_rope_ext_back(ctx, a, pos, freq, n_dims, mode, 0, 10000.0f, fs, ef, af, 1.0f, 1.0f);
+            }
+        }
         ggml_set_name(out, "out");
 
         return out;
@@ -2194,11 +2289,12 @@ struct test_rope : public test_case {
         for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
             if (t->type == GGML_TYPE_I32) {
                 // pos
-                std::vector data(ne_a[2]);
-                for (int i = 0; i < ne_a[2]; i++) {
+                const int num_pos_ids = (mode & GGML_ROPE_TYPE_MROPE) ? ne_a[2] * 4 : ne_a[2];
+                std::vector data(num_pos_ids);
+                for (int i = 0; i < num_pos_ids; i++) {
                     data[i] = rand() % n_ctx;
                 }
-                ggml_backend_tensor_set(t, data.data(), 0, ne_a[2] * sizeof(int));
+                ggml_backend_tensor_set(t, data.data(), 0, num_pos_ids * sizeof(int));
             } else {
                 if (t->ne[0] == n_dims/2) {
                     // frequency factors in the range [0.9f, 1.1f]
@@ -2498,6 +2594,35 @@ struct test_sum_rows : public test_case {
     }
 };
 
+// GGML_OP_MEAN
+struct test_mean : public test_case {
+    const ggml_type type;
+    const std::array ne;
+
+    std::string vars() override {
+        return VARS_TO_STR2(type, ne);
+    }
+
+    test_mean(ggml_type type = GGML_TYPE_F32,
+            std::array ne = {10, 5, 4, 3})
+        : type(type), ne(ne) {}
+
+    ggml_tensor * build_graph(ggml_context * ctx) override {
+        ggml_tensor * a = ggml_new_tensor(ctx, type, 4, ne.data());
+        ggml_set_param(ctx, a);
+        ggml_set_name(a, "a");
+
+        ggml_tensor * out = ggml_mean(ctx, a);
+        ggml_set_name(out, "out");
+
+        return out;
+    }
+
+    float grad_eps() override {
+        return 0.1f * ne[0]*ne[1]*ne[2]*ne[3];
+    }
+};
+
 // GGML_OP_UPSCALE
 struct test_upscale : public test_case {
     const ggml_type type;
@@ -2642,6 +2767,33 @@ struct test_pad : public test_case {
     }
 };
 
+// GGML_OP_PAD_REFLECT_1D
+struct test_pad_reflect_1d : public test_case {
+    const ggml_type type;
+    const std::array ne_a;
+    const int pad_0;
+    const int pad_1;
+
+    std::string vars() override {
+        return VARS_TO_STR4(type, ne_a, pad_0, pad_1);
+    }
+
+    test_pad_reflect_1d(ggml_type type = GGML_TYPE_F32,
+            std::array ne_a = {512, 34, 2, 1},
+            int pad_0 = 10, int pad_1 = 9)
+        : type(type), ne_a(ne_a), pad_0(pad_0), pad_1(pad_1)  {}
+
+    ggml_tensor * build_graph(ggml_context * ctx) override {
+        ggml_tensor * a = ggml_new_tensor(ctx, type, 2, ne_a.data());
+        ggml_set_name(a, "a");
+
+        ggml_tensor * out = ggml_pad_reflect_1d(ctx, a, pad_0, pad_1);
+        ggml_set_name(out, "out");
+
+        return out;
+    }
+};
+
 // GGML_OP_ARANGE
 struct test_arange : public test_case {
     const ggml_type type;
@@ -2740,6 +2892,13 @@ struct test_flash_attn_ext : public test_case {
         return 5e-4;
     }
 
+    uint64_t op_flops(ggml_tensor * t) override {
+        GGML_UNUSED(t);
+        // Just counting matmul costs:
+        // Q*K^T is nb x hs x kv, P*V is nb x kv x hs, per head
+        return 2 * 2 * nh * nb * hs * kv;
+    }
+
     test_flash_attn_ext(int64_t hs = 128, int64_t nh = 32, int64_t kv = 96, int64_t nb = 8,
                         bool mask = true, float max_bias = 0.0f, float logit_softcap = 0.0f, ggml_type type_KV = GGML_TYPE_F16)
         : hs(hs), nh(nh), kv(kv), nb(nb), mask(mask), max_bias(max_bias), logit_softcap(logit_softcap), type_KV(type_KV) {}
@@ -2825,24 +2984,14 @@ struct test_cross_entropy_loss : public test_case {
 struct test_opt_step_adamw : public test_case {
     const ggml_type type;
     const std::array ne;
-    const float alpha;
-    const float beta1;
-    const float beta2;
-    const float eps;
-    const float wd;
 
     std::string vars() override {
-        return VARS_TO_STR7(type, ne, alpha, beta1, beta2, eps, wd);
+        return VARS_TO_STR2(type, ne);
     }
 
     test_opt_step_adamw(ggml_type type = GGML_TYPE_F32,
-            std::array ne = {10, 5, 4, 3},
-            float alpha = 1e-3f,
-            float beta1 = 0.9f,
-            float beta2 = 0.999f,
-            float eps = 1e-8f,
-            float wd = 0.0f)
-        : type(type), ne(ne), alpha(alpha), beta1(beta1), beta2(beta2), eps(eps), wd(wd) {}
+            std::array ne = {10, 5, 4, 3})
+        : type(type), ne(ne) {}
 
     ggml_tensor * build_graph(ggml_context * ctx) override {
         ggml_tensor * a = ggml_new_tensor_4d(ctx, type, ne[0], ne[1], ne[2], ne[3]);
@@ -2852,7 +3001,16 @@ struct test_opt_step_adamw : public test_case {
         ggml_tensor * grad = ggml_new_tensor_4d(ctx, type, ne[0], ne[1], ne[2], ne[3]);
         ggml_set_name(grad, "grad");
 
-        ggml_tensor * out = ggml_opt_step_adamw(ctx, a, grad, alpha, beta1, beta2, eps, wd);
+        ggml_tensor * grad_m = ggml_new_tensor_4d(ctx, type, ne[0], ne[1], ne[2], ne[3]);
+        ggml_set_name(grad_m, "grad_m");
+
+        ggml_tensor * grad_v = ggml_new_tensor_4d(ctx, type, ne[0], ne[1], ne[2], ne[3]);
+        ggml_set_name(grad_v, "grad_v");
+
+        ggml_tensor * adamw_params = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, 7);
+        ggml_set_name(adamw_params, "adamw_params");
+
+        ggml_tensor * out = ggml_opt_step_adamw(ctx, a, grad, grad_m, grad_v, adamw_params);
         ggml_set_name(out, "out");
 
         return out;
@@ -2860,7 +3018,7 @@ struct test_opt_step_adamw : public test_case {
 
     void initialize_tensors(ggml_context * ctx) override {
         for (ggml_tensor * t = ggml_get_first_tensor(ctx); t != NULL; t = ggml_get_next_tensor(ctx, t)) {
-            init_tensor_uniform(t, 0.0f, 1.0f); // grad_v needs non-negative values.
+            init_tensor_uniform(t, 0.0f, 1.0f); // grad_v and adamw_params need non-negative values.
         }
     }
 
@@ -3273,7 +3431,9 @@ static const ggml_type all_types[] = {
 
 static const ggml_type base_types[] = {
     GGML_TYPE_F32, GGML_TYPE_F16,
+    GGML_TYPE_Q8_0, // for I8MM tests
     GGML_TYPE_Q4_0,
+    GGML_TYPE_Q4_1, // for I8MM tests
     GGML_TYPE_Q4_K,
     GGML_TYPE_IQ2_XXS
 };
@@ -3397,9 +3557,15 @@ static std::vector> make_test_cases_eval() {
     test_cases.emplace_back(new test_conv_transpose_1d({3,2,1,1}, {3,1,2,1}, 1, 0, 1));
     test_cases.emplace_back(new test_conv_transpose_1d({2,1,1,1}, {3,1,1,1}, 1, 0, 1));
 
-    test_cases.emplace_back(new test_argmax());
     test_cases.emplace_back(new test_count_equal());
 
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {32,    1, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {100,  10, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {1024, 10, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {1024, 12, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {2000, 10, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {5438,  3, 1, 1}));
+
     for (int ne3 : {1, 3}) { // CUDA backward pass only supports ne3 == 1
         test_cases.emplace_back(new test_repeat(GGML_TYPE_F32, {10, 5, 4, ne3}, {1, 1, 1, 1}));
         test_cases.emplace_back(new test_repeat(GGML_TYPE_F32, {10, 5, 4, ne3}, {2, 1, 1, 1}));
@@ -3425,10 +3591,14 @@ static std::vector> make_test_cases_eval() {
         test_cases.emplace_back(new test_set(GGML_TYPE_F32, GGML_TYPE_F32, {6, 5, 4, 3}, dim));
     }
 
+    for (int dim = 1; dim < GGML_MAX_DIMS; ++dim) {
+        test_cases.emplace_back(new test_set(GGML_TYPE_I32, GGML_TYPE_I32, {6, 5, 4, 3}, dim));
+    }
+
     for (ggml_type type_src : {GGML_TYPE_F16, GGML_TYPE_F32}) {
         for (ggml_type type_dst : all_types) {
-           test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 4, 4, 4}));
-           test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 2, 3, 4}, {0, 2, 1, 3})); // cpy by rows
+            test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 4, 4, 4}));
+            test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 2, 3, 4}, {0, 2, 1, 3})); // cpy by rows
         }
     }
     for (ggml_type type_src : {GGML_TYPE_F16, GGML_TYPE_F32}) {
@@ -3504,6 +3674,24 @@ static std::vector> make_test_cases_eval() {
     test_cases.emplace_back(new test_rwkv_wkv6(GGML_TYPE_F32, 32, 64, 32, 4));
     test_cases.emplace_back(new test_rwkv_wkv6(GGML_TYPE_F32, 32, 64, 128, 4));
 
+    test_cases.emplace_back(new test_gla(GGML_TYPE_F32, 32, 64, 1, 1));
+    test_cases.emplace_back(new test_gla(GGML_TYPE_F32, 32, 64, 32, 1));
+    test_cases.emplace_back(new test_gla(GGML_TYPE_F32, 32, 64, 32, 4));
+    test_cases.emplace_back(new test_gla(GGML_TYPE_F32, 32, 64, 128, 4));
+
+    for (int i = 1; i < 9; ++i) {
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16,    GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q4_0,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q4_1,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q5_0,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q5_1,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q8_0,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q4_K,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q5_K,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_Q6_K,   GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+        test_cases.emplace_back(new test_mul_mat(GGML_TYPE_IQ4_NL, GGML_TYPE_F32, 16,  i, 256, { 1,  1}, {1, 1}));
+    }
+
 #if 1
     for (ggml_type type_a : base_types) {
         for (ggml_type type_b : {GGML_TYPE_F32, GGML_TYPE_F16}) {
@@ -3675,7 +3863,7 @@ static std::vector> make_test_cases_eval() {
     test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true,  0.1f, 0.0f));
     test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {32, 2, 32, 1}, true,  0.1f, 8.0f));
 
-    {
+    for (bool fw : {true, false}) { // fw == forward
         bool all = true;
 
         for (float v : { 0, 1 }) {
@@ -3684,23 +3872,29 @@ static std::vector> make_test_cases_eval() {
                     for (float af : { 1.0f, 1.4245f }) {
                         for (ggml_type type : {GGML_TYPE_F32, GGML_TYPE_F16}) {
                             for (bool ff : {false, true}) { // freq_factors
-                                test_cases.emplace_back(new test_rope(type, {128,  32, 2, 1}, 128, 0, 512, fs, ef, af, ff, v)); // llama 7B
+                                test_cases.emplace_back(new test_rope(type, {128,  32, 2, 1}, 128, 0, 512, fs, ef, af, ff, v, fw)); // llama 7B
 
                                 if (all) {
-                                    test_cases.emplace_back(new test_rope(type, {128,  40, 2, 1}, 128, 0, 512, fs, ef, af, ff, v)); // llama 13B
-                                    test_cases.emplace_back(new test_rope(type, {128,  52, 2, 1}, 128, 0, 512, fs, ef, af, ff, v)); // llama 30B
-                                    test_cases.emplace_back(new test_rope(type, {128,  64, 2, 1}, 128, 0, 512, fs, ef, af, ff, v)); // llama 65B
+                                    test_cases.emplace_back(new test_rope(type, {128,  40, 2, 1}, 128, 0, 512, fs, ef, af, ff, v, fw)); // llama 13B
+                                    test_cases.emplace_back(new test_rope(type, {128,  52, 2, 1}, 128, 0, 512, fs, ef, af, ff, v, fw)); // llama 30B
+                                    test_cases.emplace_back(new test_rope(type, {128,  64, 2, 1}, 128, 0, 512, fs, ef, af, ff, v, fw)); // llama 65B
                                 }
 
                                 if (all) {
-                                    test_cases.emplace_back(new test_rope(type, { 64,   1, 2, 1},  64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 7B)
-                                    test_cases.emplace_back(new test_rope(type, { 64,  71, 2, 1},  64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 7B)
-                                    test_cases.emplace_back(new test_rope(type, { 64,   8, 2, 1},  64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 40B)
-                                    test_cases.emplace_back(new test_rope(type, { 80,  32, 2, 1},  20, 2, 512, fs, ef, af, ff, v)); // neox (stablelm)
-                                    test_cases.emplace_back(new test_rope(type, { 80,  32, 2, 1},  32, 2, 512, fs, ef, af, ff, v)); // neox (phi-2)
+                                    test_cases.emplace_back(new test_rope(type, { 64,   1, 2, 1},  64, 2, 512, fs, ef, af, ff, v, fw)); // neox (falcon 7B)
+                                    test_cases.emplace_back(new test_rope(type, { 64,  71, 2, 1},  64, 2, 512, fs, ef, af, ff, v, fw)); // neox (falcon 7B)
+                                    test_cases.emplace_back(new test_rope(type, { 64,   8, 2, 1},  64, 2, 512, fs, ef, af, ff, v, fw)); // neox (falcon 40B)
+                                    test_cases.emplace_back(new test_rope(type, { 80,  32, 2, 1},  20, 2, 512, fs, ef, af, ff, v, fw)); // neox (stablelm)
+                                    test_cases.emplace_back(new test_rope(type, { 80,  32, 2, 1},  32, 2, 512, fs, ef, af, ff, v, fw)); // neox (phi-2)
                                 }
 
-                                test_cases.emplace_back(new test_rope(type, { 64, 128, 2, 1},  64, 2, 512, fs, ef, af, ff, v)); // neox (falcon 40B)
+                                if (all) {
+                                    test_cases.emplace_back(new test_rope(type, {128,  12, 2, 1}, 128, GGML_ROPE_TYPE_MROPE,  512, fs, ef, af, ff, v, fw)); // rope_multi,m-rope (qwen2vl 2B)
+                                    test_cases.emplace_back(new test_rope(type, {128,  28, 2, 1}, 128, GGML_ROPE_TYPE_MROPE,  512, fs, ef, af, ff, v, fw)); // rope_multi,m-rope (qwen2vl 7B)
+                                    test_cases.emplace_back(new test_rope(type, { 80,  16, 2, 1},  80, GGML_ROPE_TYPE_VISION, 512, fs, ef, af, ff, v, fw)); // rope_multi,m-rope (qwen2vl ViT)
+                                }
+
+                                test_cases.emplace_back(new test_rope(type, { 64, 128, 2, 1},  64, 2, 512, fs, ef, af, ff, v, fw)); // neox (falcon 40B)
                             }
                         }
 
@@ -3726,12 +3920,15 @@ static std::vector> make_test_cases_eval() {
 
     test_cases.emplace_back(new test_sum());
     test_cases.emplace_back(new test_sum_rows());
+    test_cases.emplace_back(new test_mean());
     test_cases.emplace_back(new test_upscale());
     test_cases.emplace_back(new test_upscale(GGML_TYPE_F32, { 512, 512, 3, 1 }, 2, true));
     test_cases.emplace_back(new test_upscale_ext());
-    test_cases.emplace_back(new test_group_norm());
+    test_cases.emplace_back(new test_group_norm(GGML_TYPE_F32, {64, 64, 320, 1}));
+    test_cases.emplace_back(new test_group_norm(GGML_TYPE_F32, {9, 9, 1280, 1}));
     test_cases.emplace_back(new test_acc());
     test_cases.emplace_back(new test_pad());
+    test_cases.emplace_back(new test_pad_reflect_1d());
     test_cases.emplace_back(new test_arange());
     test_cases.emplace_back(new test_timestep_embedding());
     test_cases.emplace_back(new test_leaky_relu());
@@ -3745,7 +3942,7 @@ static std::vector> make_test_cases_eval() {
                     for (int nh : { 32, }) {
                         for (int kv : { 512, 1024, }) {
                             for (int nb : { 1, 3, 32, 35, }) {
-                                for (ggml_type type_KV : {GGML_TYPE_F16, GGML_TYPE_Q8_0, GGML_TYPE_Q4_0}) {
+                                for (ggml_type type_KV : {GGML_TYPE_F16, GGML_TYPE_BF16, GGML_TYPE_Q8_0, GGML_TYPE_Q4_0}) {
                                     test_cases.emplace_back(new test_flash_attn_ext(hs, nh, kv, nb, mask, max_bias, logit_softcap, type_KV));
                                 }
                             }
@@ -3757,9 +3954,7 @@ static std::vector> make_test_cases_eval() {
     }
 
     test_cases.emplace_back(new test_cross_entropy_loss());
-    for (float wd : {0.0f, 1e-2f}) {
-        test_cases.emplace_back(new test_opt_step_adamw(GGML_TYPE_F32, {10, 5, 4, 3}, 1.0f, 1e-3f, 0.9f, 0.999f, wd));
-    }
+    test_cases.emplace_back(new test_opt_step_adamw(GGML_TYPE_F32, {10, 5, 4, 3}));
 
     // these tests are disabled to save execution time, but they can be handy for debugging
 #if 0
@@ -3779,7 +3974,23 @@ static std::vector> make_test_cases_perf() {
     test_cases.emplace_back(new test_bin_bcast(ggml_add, GGML_TYPE_F32, {4096, 1, 1, 1}, {1,   1, 1, 1}));
     test_cases.emplace_back(new test_bin_bcast(ggml_add, GGML_TYPE_F32, {4096, 1, 1, 1}, {1, 512, 1, 1}));
 
-    for (int bs : {1, 512}) {
+    test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F16, {512, 3072, 1, 1}));
+    test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F32, {8192, 512, 2, 1}, {0, 2, 1, 3}));
+    test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F32, {3072, 512, 2, 1}, {0, 2, 1, 3}));
+
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {4096, 4096, 5, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {77, 4096, 5, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {1024, 1024, 10, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {77, 1024, 10, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {256, 256, 20, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {64, 64, 20, 1}, false, 1.0f, 0.0f));
+    test_cases.emplace_back(new test_soft_max(GGML_TYPE_F32, {77, 64, 20, 1}, false, 1.0f, 0.0f));
+
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {32, 10, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {1024, 10, 1, 1}));
+    test_cases.emplace_back(new test_argmax(GGML_TYPE_F32, {32000, 512, 1, 1}));
+
+    for (int bs : {1, 2, 3, 4, 5, 8, 512}) {
         for (ggml_type type_a : all_types) {
             for (ggml_type type_b : {GGML_TYPE_F32}) {
                 test_cases.emplace_back(new test_mul_mat(type_a, type_b, 4096, bs, 14336, {1,  1}, {1, 1}));
@@ -3787,13 +3998,29 @@ static std::vector> make_test_cases_perf() {
         }
     }
 
+    for (int K : {3, 5}) {
+        for (int IC : {256, 2560}) {
+            for (int IW_IH : {32, 64, 256}) {
+                if (IC == 2560 && IW_IH == 256) {
+                    // too big
+                    continue;
+                }
+                test_cases.emplace_back(new test_im2col(GGML_TYPE_F32, GGML_TYPE_F16, GGML_TYPE_F32, {IW_IH, IW_IH, IC, 1}, {K, K, IC, 1}, 1, 1, 1, 1, 1, 1, true));
+            }
+        }
+    }
+
     return test_cases;
 }
 
 static bool test_backend(ggml_backend_t backend, test_mode mode, const char * op_name) {
     if (mode == MODE_TEST) {
         auto test_cases = make_test_cases_eval();
-        ggml_backend_t backend_cpu = ggml_backend_cpu_init();
+        ggml_backend_t backend_cpu = ggml_backend_init_by_type(GGML_BACKEND_DEVICE_TYPE_CPU, NULL);
+        if (backend_cpu == NULL) {
+            printf("  Failed to initialize CPU backend\n");
+            return false;
+        }
 
         size_t n_ok = 0;
         for (auto & test : test_cases) {
@@ -3873,7 +4100,9 @@ int main(int argc, char ** argv) {
         }
     }
 
-    // enumerate backends
+    // load and enumerate backends
+    ggml_backend_load_all();
+
     printf("Testing %zu devices\n\n", ggml_backend_dev_count());
 
     size_t n_ok = 0;
@@ -3889,16 +4118,15 @@ int main(int argc, char ** argv) {
             continue;
         }
 
-        ggml_backend_t backend = ggml_backend_dev_init(dev, NULL);
-        GGML_ASSERT(backend != NULL);
-
-        if (backend_filter == NULL && ggml_backend_is_cpu(backend) && mode != MODE_GRAD) {
+        if (backend_filter == NULL && ggml_backend_dev_type(dev) == GGML_BACKEND_DEVICE_TYPE_CPU && mode != MODE_GRAD) {
             printf("  Skipping CPU backend\n");
-            ggml_backend_free(backend);
             n_ok++;
             continue;
         }
 
+        ggml_backend_t backend = ggml_backend_dev_init(dev, NULL);
+        GGML_ASSERT(backend != NULL);
+
         ggml_backend_reg_t reg = ggml_backend_dev_backend_reg(dev);
         auto ggml_backend_set_n_threads_fn = (ggml_backend_set_n_threads_t) ggml_backend_reg_get_proc_address(reg, "ggml_backend_set_n_threads");
         if (ggml_backend_set_n_threads_fn) {
@@ -3927,6 +4155,8 @@ int main(int argc, char ** argv) {
         ggml_backend_free(backend);
     }
 
+    ggml_quantize_free();
+
     printf("%zu/%zu backends passed\n", n_ok, ggml_backend_dev_count());
 
     if (n_ok != ggml_backend_dev_count()) {
@@ -3934,8 +4164,6 @@ int main(int argc, char ** argv) {
         return 1;
     }
 
-    ggml_quantize_free();
-
     printf("\033[1;32mOK\033[0m\n");
     return 0;
 }
diff --git a/tests/test-chat-template.cpp b/tests/test-chat-template.cpp
index 03e897e66..5ec318bd4 100644
--- a/tests/test-chat-template.cpp
+++ b/tests/test-chat-template.cpp
@@ -9,7 +9,7 @@
 #include "common.h"
 
 int main(void) {
-    llama_chat_message conversation[] = {
+    std::vector conversation {
         {"system", "You are a helpful assistant"},
         {"user", "Hello"},
         {"assistant", "Hi there"},
@@ -17,119 +17,187 @@ int main(void) {
         {"assistant", "   I am an assistant   "},
         {"user", "Another question"},
     };
-    size_t message_count = 6;
-    std::vector templates = {
-        // teknium/OpenHermes-2.5-Mistral-7B
-        "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\\n' + message['content'] + '<|im_end|>' + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\\n' }}{% endif %}",
-        // mistralai/Mistral-7B-Instruct-v0.2
-        "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
-        // TheBloke/FusionNet_34Bx2_MoE-AWQ
-        "{%- for idx in range(0, messages|length) -%}\\n{%- if messages[idx]['role'] == 'user' -%}\\n{%- if idx > 1 -%}\\n{{- bos_token + '[INST] ' + messages[idx]['content'] + ' [/INST]' -}}\\n{%- else -%}\\n{{- messages[idx]['content'] + ' [/INST]' -}}\\n{%- endif -%}\\n{% elif messages[idx]['role'] == 'system' %}\\n{{- '[INST] <>\\\\n' + messages[idx]['content'] + '\\\\n<>\\\\n\\\\n' -}}\\n{%- elif messages[idx]['role'] == 'assistant' -%}\\n{{- ' '  + messages[idx]['content'] + ' ' + eos_token -}}\\n{% endif %}\\n{% endfor %}",
-        // bofenghuang/vigogne-2-70b-chat
-        "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif true == true and not '<>' in messages[0]['content'] %}{% set loop_messages = messages %}{% set system_message = 'Vous êtes Vigogne, un assistant IA créé par Zaion Lab. Vous suivez extrêmement bien les instructions. Aidez autant que vous le pouvez.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<>\\\\n' + system_message + '\\\\n<>\\\\n\\\\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'system' %}{{ '<>\\\\n' + content.strip() + '\\\\n<>\\\\n\\\\n' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}",
-        // mlabonne/AlphaMonarch-7B
-        "{% for message in messages %}{{bos_token + message['role'] + '\\n' + message['content'] + eos_token + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ bos_token + 'assistant\\n' }}{% endif %}",
-        // google/gemma-7b-it
-        "{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '' + role + '\\n' + message['content'] | trim + '\\n' }}{% endfor %}{% if add_generation_prompt %}{{'model\\n'}}{% endif %}",
-        // OrionStarAI/Orion-14B-Chat
-        "{% for message in messages %}{% if loop.first %}{{ bos_token }}{% endif %}{% if message['role'] == 'user' %}{{ 'Human: ' + message['content'] + '\\n\\nAssistant: ' + eos_token }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token }}{% endif %}{% endfor %}",
-        // openchat/openchat-3.5-0106
-        // The included chat_template differs from the author's suggestions here: https://huggingface.co/openchat/openchat_3.5/discussions/5#65448109b4a3f3a2f486fd9d
-        // So we match against the included template but implement the suggested version.
-        "{{ bos_token }}{% for message in messages %}{{ 'GPT4 Correct ' + message['role'].title() + ': ' + message['content'] + '<|end_of_turn|>'}}{% endfor %}{% if add_generation_prompt %}{{ 'GPT4 Correct Assistant:' }}{% endif %}",
-        // deepseek-ai/deepseek-coder-33b-instruct
-        "{% if not add_generation_prompt is defined %}\n{% set add_generation_prompt = false %}\n{% endif %}\n{%- set ns = namespace(found=false) -%}\n{%- for message in messages -%}\n    {%- if message['role'] == 'system' -%}\n        {%- set ns.found = true -%}\n    {%- endif -%}\n{%- endfor -%}\n{{bos_token}}{%- if not ns.found -%}\n{{'You are an AI programming assistant, utilizing the Deepseek Coder model, developed by Deepseek Company, and you only answer questions related to computer science. For politically sensitive questions, security and privacy issues, and other non-computer science questions, you will refuse to answer\\n'}}\n{%- endif %}\n{%- for message in messages %}\n    {%- if message['role'] == 'system' %}\n{{ message['content'] }}\n    {%- else %}\n        {%- if message['role'] == 'user' %}\n{{'### Instruction:\\n' + message['content'] + '\\n'}}\n        {%- else %}\n{{'### Response:\\n' + message['content'] + '\\n<|EOT|>\\n'}}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{% if add_generation_prompt %}\n{{'### Response:'}}\n{% endif %}",
-        // eachadea/vicuna-13b-1.1
-        // No template included in tokenizer_config.json, so this template likely needs to be manually set.
-        "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{- '' + message['content'] + '\n\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
-        // Orca-Vicuna
-        // No template included in tokenizer_config.json, so this template likely needs to be manually set.
-        "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{-'SYSTEM: ' + message['content'] + '\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
-        // CohereForAI/c4ai-command-r-plus
-        "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are Command-R, a brilliant, sophisticated, AI-assistant trained to assist human users by providing thorough responses. You are trained by Cohere.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>'  + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}",
-        // Llama-3
-        "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}",
-        //Phi-3-mini
-        "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
-        //Phi-3-small
-        "{{ bos_token }}{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
-        //Phi-3-medium
-        "{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
-        //Phi-3-vision
-        "{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt and messages[-1]['role'] != 'assistant' %}{{- '<|assistant|>\n' -}}{% endif %}",
-        // ChatGLM3
-        "{% for message in messages %}{% if loop.first %}[gMASK]sop<|{{ message['role'] }}|>\n {{ message['content'] }}{% else %}<|{{ message['role'] }}|>\n {{ message['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}",
-        // ChatGLM4
-        u8"[gMASK]{% for item in messages %}{% if item['tools'] is defined %}<|system|>\n你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的，你的任务是针对用户的问题和要求提供适当的答复和支持。\n\n# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %}\n\n## {{ tool['function']['name'] }}\n\n{{ tool['function'] | tojson(indent=4) }}\n......{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }}\n{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}",
-        // MiniCPM-3B-OpenHermes-2.5-v2-GGUF
-        u8"{% for message in messages %}{% if message['role'] == 'user' %}{{'<用户>' + message['content'].strip() + ''}}{% else %}{{message['content'].strip()}}{% endif %}{% endfor %}",
-        // DeepSeek-V2
-        "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ 'User: ' + message['content'] + '\n\n' }}{% elif message['role'] == 'assistant' %}{{ 'Assistant: ' + message['content'] + eos_token }}{% elif message['role'] == 'system' %}{{ message['content'] + '\n\n' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}",
-        // ibm-granite/granite-3.0-8b-instruct
-        "{%- if tools %}\n    {{- '<|start_of_role|>available_tools<|end_of_role|>\n' }}\n    {%- for tool in tools %}\n    {{- tool | tojson(indent=4) }}\n    {%- if not loop.last %}\n        {{- '\n\n' }}\n    {%- endif %}\n    {%- endfor %}\n    {{- '<|end_of_text|>\n' }}\n{%- endif %}\n{%- for message in messages %}\n    {%- if message['role'] == 'system' %}\n    {{- '<|start_of_role|>system<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'user' %}\n    {{- '<|start_of_role|>user<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'assistant' %}\n    {{- '<|start_of_role|>assistant<|end_of_role|>'  + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'assistant_tool_call' %}\n    {{- '<|start_of_role|>assistant<|end_of_role|><|tool_call|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'tool_response' %}\n    {{- '<|start_of_role|>tool_response<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- endif %}\n    {%- if loop.last and add_generation_prompt %}\n    {{- '<|start_of_role|>assistant<|end_of_role|>' }}\n    {%- endif %}\n{%- endfor %}",
+    struct TestCase {
+        std::string name;
+        std::string template_str;
+        std::string expected_output;
     };
-    std::vector expected_output = {
-        // teknium/OpenHermes-2.5-Mistral-7B
-        "<|im_start|>system\nYou are a helpful assistant<|im_end|>\n<|im_start|>user\nHello<|im_end|>\n<|im_start|>assistant\nHi there<|im_end|>\n<|im_start|>user\nWho are you<|im_end|>\n<|im_start|>assistant\n   I am an assistant   <|im_end|>\n<|im_start|>user\nAnother question<|im_end|>\n<|im_start|>assistant\n",
-        // mistralai/Mistral-7B-Instruct-v0.2
-        "[INST] You are a helpful assistant\nHello [/INST]Hi there[INST] Who are you [/INST]   I am an assistant   [INST] Another question [/INST]",
-        // TheBloke/FusionNet_34Bx2_MoE-AWQ
-        "[INST] <>\nYou are a helpful assistant\n<>\n\nHello [/INST] Hi there [INST] Who are you [/INST]    I am an assistant    [INST] Another question [/INST]",
-        // bofenghuang/vigogne-2-70b-chat
-        "[INST] <>\nYou are a helpful assistant\n<>\n\nHello [/INST] Hi there [INST] Who are you [/INST] I am an assistant [INST] Another question [/INST]",
-        // mlabonne/AlphaMonarch-7B
-        "system\nYou are a helpful assistant\nuser\nHello\nassistant\nHi there\nuser\nWho are you\nassistant\n   I am an assistant   \nuser\nAnother question\nassistant\n",
-        // google/gemma-7b-it
-        "user\nYou are a helpful assistant\n\nHello\nmodel\nHi there\nuser\nWho are you\nmodel\nI am an assistant\nuser\nAnother question\nmodel\n",
-        // OrionStarAI/Orion-14B-Chat
-        "Human: You are a helpful assistant\n\nHello\n\nAssistant: Hi thereHuman: Who are you\n\nAssistant:    I am an assistant   Human: Another question\n\nAssistant: ",
-        // openchat/openchat-3.5-0106
-        "You are a helpful assistant<|end_of_turn|>GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi there<|end_of_turn|>GPT4 Correct User: Who are you<|end_of_turn|>GPT4 Correct Assistant:    I am an assistant   <|end_of_turn|>GPT4 Correct User: Another question<|end_of_turn|>GPT4 Correct Assistant:",
-        // deepseek-ai/deepseek-coder-33b-instruct
-        "You are a helpful assistant### Instruction:\nHello\n### Response:\nHi there\n<|EOT|>\n### Instruction:\nWho are you\n### Response:\n   I am an assistant   \n<|EOT|>\n### Instruction:\nAnother question\n### Response:\n",
-        // eachadea/vicuna-13b-1.1
-        "You are a helpful assistant\n\nUSER: Hello\nASSISTANT: Hi there\nUSER: Who are you\nASSISTANT:    I am an assistant   \nUSER: Another question\nASSISTANT:",
-        // Orca-Vicuna
-        "SYSTEM: You are a helpful assistant\nUSER: Hello\nASSISTANT: Hi there\nUSER: Who are you\nASSISTANT:    I am an assistant   \nUSER: Another question\nASSISTANT:",
-        // CohereForAI/c4ai-command-r-plus
-        "<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>You are a helpful assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Hello<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>Hi there<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Who are you<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>I am an assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Another question<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>",
-        // Llama 3
-        "<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHello<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nHi there<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWho are you<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nI am an assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nAnother question<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
-        //Phi-3-mini
-        "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
-        //Phi-3-small
-        "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
-        //Phi-3-medium
-        "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
-        //Phi-3-vision
-        "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
-        // ChatGLM3
-        "[gMASK]sop<|system|>\n You are a helpful assistant<|user|>\n Hello<|assistant|>\n Hi there<|user|>\n Who are you<|assistant|>\n    I am an assistant   <|user|>\n Another question<|assistant|>",
-        // ChatGLM4
-        "[gMASK]<|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n   I am an assistant   <|user|>\nAnother question<|assistant|>",
-        // MiniCPM-3B-OpenHermes-2.5-v2-GGUF
-        u8"You are a helpful assistant<用户>HelloHi there<用户>Who are youI am an assistant<用户>Another question",
-        // DeepSeek-V2
-        u8"You are a helpful assistant\n\nUser: Hello\n\nAssistant: Hi there<｜end▁of▁sentence｜>User: Who are you\n\nAssistant:    I am an assistant   <｜end▁of▁sentence｜>User: Another question\n\nAssistant:",
-        // ibm-granite/granite-3.0-8b-instruct
-        "<|start_of_role|>system<|end_of_role|>You are a helpful assistant<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Hello<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>Hi there<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Who are you<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>   I am an assistant   <|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Another question<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>\n",
+    std::vector test_cases {
+        {
+            /* .name= */ "teknium/OpenHermes-2.5-Mistral-7B",
+            /* .template_str= */ "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\\n' + message['content'] + '<|im_end|>' + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\\n' }}{% endif %}",
+            /* .expected_output= */ "<|im_start|>system\nYou are a helpful assistant<|im_end|>\n<|im_start|>user\nHello<|im_end|>\n<|im_start|>assistant\nHi there<|im_end|>\n<|im_start|>user\nWho are you<|im_end|>\n<|im_start|>assistant\n   I am an assistant   <|im_end|>\n<|im_start|>user\nAnother question<|im_end|>\n<|im_start|>assistant\n",
+        },
+        {
+            /* .name= */ "mistralai/Mistral-7B-Instruct-v0.2 (NOTE: Old pre-v1 without a system prompt)",
+            /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
+            /* .expected_output= */ "[INST] You are a helpful assistant\nHello [/INST]Hi there[INST] Who are you [/INST]   I am an assistant   [INST] Another question [/INST]",
+        },
+        {
+            /* .name= */ "TheBloke/FusionNet_34Bx2_MoE-AWQ",
+            /* .template_str= */ "{%- for idx in range(0, messages|length) -%}\\n{%- if messages[idx]['role'] == 'user' -%}\\n{%- if idx > 1 -%}\\n{{- bos_token + '[INST] ' + messages[idx]['content'] + ' [/INST]' -}}\\n{%- else -%}\\n{{- messages[idx]['content'] + ' [/INST]' -}}\\n{%- endif -%}\\n{% elif messages[idx]['role'] == 'system' %}\\n{{- '[INST] <>\\\\n' + messages[idx]['content'] + '\\\\n<>\\\\n\\\\n' -}}\\n{%- elif messages[idx]['role'] == 'assistant' -%}\\n{{- ' '  + messages[idx]['content'] + ' ' + eos_token -}}\\n{% endif %}\\n{% endfor %}",
+            /* .expected_output= */ "[INST] <>\nYou are a helpful assistant\n<>\n\nHello [/INST]Hi there[INST] Who are you [/INST]   I am an assistant   [INST] Another question [/INST]",
+        },
+        {
+            /* .name= */ "bofenghuang/vigogne-2-70b-chat",
+            /* .template_str= */ "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif true == true and not '<>' in messages[0]['content'] %}{% set loop_messages = messages %}{% set system_message = 'Vous êtes Vigogne, un assistant IA créé par Zaion Lab. Vous suivez extrêmement bien les instructions. Aidez autant que vous le pouvez.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<>\\\\n' + system_message + '\\\\n<>\\\\n\\\\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'system' %}{{ '<>\\\\n' + content.strip() + '\\\\n<>\\\\n\\\\n' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}",
+            /* .expected_output= */ "[INST] <>\nYou are a helpful assistant\n<>\n\nHello [/INST]Hi there[INST] Who are you [/INST]I am an assistant[INST] Another question [/INST]",
+        },
+        {
+            /* .name= */ "mlabonne/AlphaMonarch-7B",
+            /* .template_str= */ "{% for message in messages %}{{bos_token + message['role'] + '\\n' + message['content'] + eos_token + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ bos_token + 'assistant\\n' }}{% endif %}",
+            /* .expected_output= */ "system\nYou are a helpful assistant\nuser\nHello\nassistant\nHi there\nuser\nWho are you\nassistant\n   I am an assistant   \nuser\nAnother question\nassistant\n",
+        },
+        {
+            /* .name= */ "google/gemma-7b-it",
+            /* .template_str= */ "{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '' + role + '\\n' + message['content'] | trim + '\\n' }}{% endfor %}{% if add_generation_prompt %}{{'model\\n'}}{% endif %}",
+            /* .expected_output= */ "user\nYou are a helpful assistant\n\nHello\nmodel\nHi there\nuser\nWho are you\nmodel\nI am an assistant\nuser\nAnother question\nmodel\n",
+        },
+        {
+            /* .name= */ "OrionStarAI/Orion-14B-Chat",
+            /* .template_str= */ "{% for message in messages %}{% if loop.first %}{{ bos_token }}{% endif %}{% if message['role'] == 'user' %}{{ 'Human: ' + message['content'] + '\\n\\nAssistant: ' + eos_token }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token }}{% endif %}{% endfor %}",
+            /* .expected_output= */ "Human: You are a helpful assistant\n\nHello\n\nAssistant: Hi thereHuman: Who are you\n\nAssistant:    I am an assistant   Human: Another question\n\nAssistant: ",
+        },
+        {
+            /* .name= */ "openchat/openchat-3.5-0106",
+            // The included chat_template differs from the author's suggestions here: https://huggingface.co/openchat/openchat_3.5/discussions/5#65448109b4a3f3a2f486fd9d
+            // So we match against the included template but implement the suggested version.
+            /* .template_str= */ "{{ bos_token }}{% for message in messages %}{{ 'GPT4 Correct ' + message['role'].title() + ': ' + message['content'] + '<|end_of_turn|>'}}{% endfor %}{% if add_generation_prompt %}{{ 'GPT4 Correct Assistant:' }}{% endif %}",
+            /* .expected_output= */ "You are a helpful assistant<|end_of_turn|>GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi there<|end_of_turn|>GPT4 Correct User: Who are you<|end_of_turn|>GPT4 Correct Assistant:    I am an assistant   <|end_of_turn|>GPT4 Correct User: Another question<|end_of_turn|>GPT4 Correct Assistant:",
+        },
+        {
+            /* .name= */ "deepseek-ai/deepseek-coder-33b-instruct",
+            /* .template_str= */ "{% if not add_generation_prompt is defined %}\n{% set add_generation_prompt = false %}\n{% endif %}\n{%- set ns = namespace(found=false) -%}\n{%- for message in messages -%}\n    {%- if message['role'] == 'system' -%}\n        {%- set ns.found = true -%}\n    {%- endif -%}\n{%- endfor -%}\n{{bos_token}}{%- if not ns.found -%}\n{{'You are an AI programming assistant, utilizing the Deepseek Coder model, developed by Deepseek Company, and you only answer questions related to computer science. For politically sensitive questions, security and privacy issues, and other non-computer science questions, you will refuse to answer\\n'}}\n{%- endif %}\n{%- for message in messages %}\n    {%- if message['role'] == 'system' %}\n{{ message['content'] }}\n    {%- else %}\n        {%- if message['role'] == 'user' %}\n{{'### Instruction:\\n' + message['content'] + '\\n'}}\n        {%- else %}\n{{'### Response:\\n' + message['content'] + '\\n<|EOT|>\\n'}}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{% if add_generation_prompt %}\n{{'### Response:'}}\n{% endif %}",
+            /* .expected_output= */ "You are a helpful assistant### Instruction:\nHello\n### Response:\nHi there\n<|EOT|>\n### Instruction:\nWho are you\n### Response:\n   I am an assistant   \n<|EOT|>\n### Instruction:\nAnother question\n### Response:\n",
+        },
+        {
+            /* .name= */ "eachadea/vicuna-13b-1.1",
+            // No template included in tokenizer_config.json, so this template likely needs to be manually set.
+            /* .template_str= */ "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{- '' + message['content'] + '\n\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
+            /* .expected_output= */ "You are a helpful assistant\n\nUSER: Hello\nASSISTANT: Hi there\nUSER: Who are you\nASSISTANT:    I am an assistant   \nUSER: Another question\nASSISTANT:",
+        },
+        {
+            /* .name= */ "Orca-Vicuna",
+            // No template included in tokenizer_config.json, so this template likely needs to be manually set.
+            /* .template_str= */ "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{-'SYSTEM: ' + message['content'] + '\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
+            /* .expected_output= */ "SYSTEM: You are a helpful assistant\nUSER: Hello\nASSISTANT: Hi there\nUSER: Who are you\nASSISTANT:    I am an assistant   \nUSER: Another question\nASSISTANT:",
+        },
+        {
+            /* .name= */ "CohereForAI/c4ai-command-r-plus",
+            /* .template_str= */ "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are Command-R, a brilliant, sophisticated, AI-assistant trained to assist human users by providing thorough responses. You are trained by Cohere.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>'  + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}",
+            /* .expected_output= */ "<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>You are a helpful assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Hello<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>Hi there<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Who are you<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>I am an assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Another question<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>",
+        },
+        {
+            /* .name= */ "Llama-3",
+            /* .template_str= */ "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}",
+            /* .expected_output= */ "<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHello<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nHi there<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWho are you<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nI am an assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nAnother question<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
+        },
+        {
+            /* .name= */ "Phi-3-mini",
+            /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
+            /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
+        },
+        {
+            /* .name= */ "Phi-3-small",
+            /* .template_str= */ "{{ bos_token }}{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
+            /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
+        },
+        {
+            /* .name= */ "Phi-3-medium",
+            /* .template_str= */ "{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
+            /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
+        },
+        {
+            /* .name= */ "Phi-3-vision",
+            /* .template_str= */ "{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt and messages[-1]['role'] != 'assistant' %}{{- '<|assistant|>\n' -}}{% endif %}",
+            /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n   I am an assistant   <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
+        },
+        {
+            /* .name= */ "ChatGLM3",
+            /* .template_str= */ "{% for message in messages %}{% if loop.first %}[gMASK]sop<|{{ message['role'] }}|>\n {{ message['content'] }}{% else %}<|{{ message['role'] }}|>\n {{ message['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}",
+            /* .expected_output= */ "[gMASK]sop<|system|>\n You are a helpful assistant<|user|>\n Hello<|assistant|>\n Hi there<|user|>\n Who are you<|assistant|>\n    I am an assistant   <|user|>\n Another question<|assistant|>",
+        },
+        {
+            /* .name= */ "ChatGLM4",
+            /* .template_str= */ u8"[gMASK]{% for item in messages %}{% if item['tools'] is defined %}<|system|>\n你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的，你的任务是针对用户的问题和要求提供适当的答复和支持。\n\n# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %}\n\n## {{ tool['function']['name'] }}\n\n{{ tool['function'] | tojson(indent=4) }}\n......{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }}\n{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}",
+            /* .expected_output= */ "[gMASK]<|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n   I am an assistant   <|user|>\nAnother question<|assistant|>",
+        },
+        {
+            /* .name= */ "MiniCPM-3B-OpenHermes-2.5-v2-GGUF",
+            /* .template_str= */ u8"{% for message in messages %}{% if message['role'] == 'user' %}{{'<用户>' + message['content'].strip() + ''}}{% else %}{{message['content'].strip()}}{% endif %}{% endfor %}",
+            /* .expected_output= */ u8"You are a helpful assistant<用户>HelloHi there<用户>Who are youI am an assistant<用户>Another question",
+        },
+        {
+            /* .name= */ "DeepSeek-V2",
+            /* .template_str= */ "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ 'User: ' + message['content'] + '\n\n' }}{% elif message['role'] == 'assistant' %}{{ 'Assistant: ' + message['content'] + eos_token }}{% elif message['role'] == 'system' %}{{ message['content'] + '\n\n' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}",
+            /* .expected_output= */ u8"You are a helpful assistant\n\nUser: Hello\n\nAssistant: Hi there<｜end▁of▁sentence｜>User: Who are you\n\nAssistant:    I am an assistant   <｜end▁of▁sentence｜>User: Another question\n\nAssistant:",
+        },
+        {
+            /* .name= */ "ibm-granite/granite-3.0-8b-instruct",
+            /* .template_str= */ "{%- if tools %}\n    {{- '<|start_of_role|>available_tools<|end_of_role|>\n' }}\n    {%- for tool in tools %}\n    {{- tool | tojson(indent=4) }}\n    {%- if not loop.last %}\n        {{- '\n\n' }}\n    {%- endif %}\n    {%- endfor %}\n    {{- '<|end_of_text|>\n' }}\n{%- endif %}\n{%- for message in messages %}\n    {%- if message['role'] == 'system' %}\n    {{- '<|start_of_role|>system<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'user' %}\n    {{- '<|start_of_role|>user<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'assistant' %}\n    {{- '<|start_of_role|>assistant<|end_of_role|>'  + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'assistant_tool_call' %}\n    {{- '<|start_of_role|>assistant<|end_of_role|><|tool_call|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- elif message['role'] == 'tool_response' %}\n    {{- '<|start_of_role|>tool_response<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n    {%- endif %}\n    {%- if loop.last and add_generation_prompt %}\n    {{- '<|start_of_role|>assistant<|end_of_role|>' }}\n    {%- endif %}\n{%- endfor %}",
+            /* .expected_output= */ "<|start_of_role|>system<|end_of_role|>You are a helpful assistant<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Hello<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>Hi there<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Who are you<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>   I am an assistant   <|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Another question<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>\n",
+        },
+        {
+            /* .name= */ "mistralai/Mistral-7B-Instruct-v0.2 (mistralai 'v1' template with a system prompt)",
+            /* .template_str= */ "{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content'] %}\n    {%- set loop_messages = messages[1:] %}\n{%- else %}\n    {%- set loop_messages = messages %}\n{%- endif %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n        {{- raise_exception('After the optional system message, conversation roles must alternate user/assistant/user/assistant/...') }}\n    {%- endif %}\n    {%- if message['role'] == 'user' %}\n        {%- if loop.first and system_message is defined %}\n            {{- ' [INST] ' + system_message + '\\n\\n' + message['content'] + ' [/INST]' }}\n        {%- else %}\n            {{- ' [INST] ' + message['content'] + ' [/INST]' }}\n        {%- endif %}\n    {%- elif message['role'] == 'assistant' %}\n        {{- ' ' + message['content'] + eos_token}}\n    {%- else %}\n        {{- raise_exception('Only user and assistant roles are supported, with the exception of an initial optional system message!') }}\n    {%- endif %}\n{%- endfor %}\n",
+            /* .expected_output= */ " [INST] You are a helpful assistant\n\nHello [/INST] Hi there [INST] Who are you [/INST]    I am an assistant    [INST] Another question [/INST]",
+        },
+        {
+            /* .name= */ "Mistral-Large-Instruct-2407 (mistralai 'v3' template; modified to have system prompt at start)",
+            /* .template_str= */ "{%- if messages[0][\"role\"] == \"system\" %}\n    {%- set system_message = messages[0][\"content\"] %}\n    {%- set loop_messages = messages[1:] %}\n{%- else %}\n    {%- set loop_messages = messages %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n{%- set user_messages = loop_messages | selectattr(\"role\", \"equalto\", \"user\") | list %}\n\n{#- This block checks for alternating user/assistant messages, skipping tool calling messages #}\n{%- set ns = namespace() %}\n{%- set ns.index = 0 %}\n{%- for message in loop_messages %}\n    {%- if not (message.role == \"tool\" or message.role == \"tool_results\" or (message.tool_calls is defined and message.tool_calls is not none)) %}\n        {%- if (message[\"role\"] == \"user\") != (ns.index % 2 == 0) %}\n            {{- raise_exception(\"After the optional system message, conversation roles must alternate user/assistant/user/assistant/...\") }}\n        {%- endif %}\n        {%- set ns.index = ns.index + 1 %}\n    {%- endif %}\n{%- endfor %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n    {%- if message[\"role\"] == \"user\" %}\n        {%- if tools is not none and (message == user_messages[-1]) %}\n            {{- \"[AVAILABLE_TOOLS] [\" }}\n            {%- for tool in tools %}\n                {%- set tool = tool.function %}\n                {{- '{\"type\": \"function\", \"function\": {' }}\n                {%- for key, val in tool.items() if key != \"return\" %}\n                    {%- if val is string %}\n                        {{- '\"' + key + '\": \"' + val + '\"' }}\n                    {%- else %}\n                        {{- '\"' + key + '\": ' + val|tojson }}\n                    {%- endif %}\n                    {%- if not loop.last %}\n                        {{- \", \" }}\n                    {%- endif %}\n                {%- endfor %}\n                {{- \"}}\" }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- else %}\n                    {{- \"]\" }}\n                {%- endif %}\n            {%- endfor %}\n            {{- \"[/AVAILABLE_TOOLS]\" }}\n            {%- endif %}\n        {%- if loop.last and system_message is defined %}\n            {{- \"[INST] \" + system_message + \"\\n\\n\" + message[\"content\"] + \"[/INST]\" }}\n        {%- else %}\n            {{- \"[INST] \" + message[\"content\"] + \"[/INST]\" }}\n        {%- endif %}\n    {%- elif message.tool_calls is defined and message.tool_calls is not none %}\n        {{- \"[TOOL_CALLS] [\" }}\n        {%- for tool_call in message.tool_calls %}\n            {%- set out = tool_call.function|tojson %}\n            {{- out[:-1] }}\n            {%- if not tool_call.id is defined or tool_call.id|length != 9 %}\n                {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n            {%- endif %}\n            {{- ', \"id\": \"' + tool_call.id + '\"}' }}\n            {%- if not loop.last %}\n                {{- \", \" }}\n            {%- else %}\n                {{- \"]\" + eos_token }}\n            {%- endif %}\n        {%- endfor %}\n    {%- elif message[\"role\"] == \"assistant\" %}\n        {{- \" \" + message[\"content\"]|trim + eos_token}}\n    {%- elif message[\"role\"] == \"tool_results\" or message[\"role\"] == \"tool\" %}\n        {%- if message.content is defined and message.content.content is defined %}\n            {%- set content = message.content.content %}\n        {%- else %}\n            {%- set content = message.content %}\n        {%- endif %}\n        {{- '[TOOL_RESULTS] {\"content\": ' + content|string + \", \" }}\n        {%- if not message.tool_call_id is defined or message.tool_call_id|length != 9 %}\n            {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n        {%- endif %}\n        {{- '\"call_id\": \"' + message.tool_call_id + '\"}[/TOOL_RESULTS]' }}\n    {%- else %}\n        {{- raise_exception(\"Only user and assistant roles are supported, with the exception of an initial optional system message!\") }}\n    {%- endif %}\n{%- endfor %}\n",
+            /* .expected_output= */ "[INST] You are a helpful assistant\n\nHello[/INST] Hi there[INST] Who are you[/INST] I am an assistant[INST] Another question[/INST]",
+        },
+        {
+            /* .name= */ "Mistral-Nemo-Instruct-2407 (mistralai 'v3-tekken' template; modified to have system prompt at start)",
+            /* .template_str= */ "{%- if messages[0][\"role\"] == \"system\" %}\n    {%- set system_message = messages[0][\"content\"] %}\n    {%- set loop_messages = messages[1:] %}\n{%- else %}\n    {%- set loop_messages = messages %}\n{%- endif %}\n{%- if not tools is defined %}\n    {%- set tools = none %}\n{%- endif %}\n{%- set user_messages = loop_messages | selectattr(\"role\", \"equalto\", \"user\") | list %}\n\n{#- This block checks for alternating user/assistant messages, skipping tool calling messages #}\n{%- set ns = namespace() %}\n{%- set ns.index = 0 %}\n{%- for message in loop_messages %}\n    {%- if not (message.role == \"tool\" or message.role == \"tool_results\" or (message.tool_calls is defined and message.tool_calls is not none)) %}\n        {%- if (message[\"role\"] == \"user\") != (ns.index % 2 == 0) %}\n            {{- raise_exception(\"After the optional system message, conversation roles must alternate user/assistant/user/assistant/...\") }}\n        {%- endif %}\n        {%- set ns.index = ns.index + 1 %}\n    {%- endif %}\n{%- endfor %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n    {%- if message[\"role\"] == \"user\" %}\n        {%- if tools is not none and (message == user_messages[-1]) %}\n            {{- \"[AVAILABLE_TOOLS][\" }}\n            {%- for tool in tools %}\n                {%- set tool = tool.function %}\n                {{- '{\"type\": \"function\", \"function\": {' }}\n                {%- for key, val in tool.items() if key != \"return\" %}\n                    {%- if val is string %}\n                        {{- '\"' + key + '\": \"' + val + '\"' }}\n                    {%- else %}\n                        {{- '\"' + key + '\": ' + val|tojson }}\n                    {%- endif %}\n                    {%- if not loop.last %}\n                        {{- \", \" }}\n                    {%- endif %}\n                {%- endfor %}\n                {{- \"}}\" }}\n                {%- if not loop.last %}\n                    {{- \", \" }}\n                {%- else %}\n                    {{- \"]\" }}\n                {%- endif %}\n            {%- endfor %}\n            {{- \"[/AVAILABLE_TOOLS]\" }}\n            {%- endif %}\n        {%- if loop.last and system_message is defined %}\n            {{- \"[INST]\" + system_message + \"\\n\\n\" + message[\"content\"] + \"[/INST]\" }}\n        {%- else %}\n            {{- \"[INST]\" + message[\"content\"] + \"[/INST]\" }}\n        {%- endif %}\n    {%- elif (message.tool_calls is defined and message.tool_calls is not none) %}\n        {{- \"[TOOL_CALLS][\" }}\n        {%- for tool_call in message.tool_calls %}\n            {%- set out = tool_call.function|tojson %}\n            {{- out[:-1] }}\n            {%- if not tool_call.id is defined or tool_call.id|length != 9 %}\n                {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n            {%- endif %}\n            {{- ', \"id\": \"' + tool_call.id + '\"}' }}\n            {%- if not loop.last %}\n                {{- \", \" }}\n            {%- else %}\n                {{- \"]\" + eos_token }}\n            {%- endif %}\n        {%- endfor %}\n    {%- elif message[\"role\"] == \"assistant\" %}\n        {{- message[\"content\"] + eos_token}}\n    {%- elif message[\"role\"] == \"tool_results\" or message[\"role\"] == \"tool\" %}\n        {%- if message.content is defined and message.content.content is defined %}\n            {%- set content = message.content.content %}\n        {%- else %}\n            {%- set content = message.content %}\n        {%- endif %}\n        {{- '[TOOL_RESULTS]{\"content\": ' + content|string + \", \" }}\n        {%- if not message.tool_call_id is defined or message.tool_call_id|length != 9 %}\n            {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n        {%- endif %}\n        {{- '\"call_id\": \"' + message.tool_call_id + '\"}[/TOOL_RESULTS]' }}\n    {%- else %}\n        {{- raise_exception(\"Only user and assistant roles are supported, with the exception of an initial optional system message!\") }}\n    {%- endif %}\n{%- endfor %}\n",
+            /* .expected_output= */ "[INST]You are a helpful assistant\n\nHello[/INST]Hi there[INST]Who are you[/INST]   I am an assistant   [INST]Another question[/INST]",
+        },
+        {
+            /* .name= */ "mistralai/Mistral-Large-Instruct-2411 (mistralai 'v7' template)",
+            /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + '[/INST]' }}{% elif message['role'] == 'system' %}{{ '[SYSTEM_PROMPT] ' + message['content'] + '[/SYSTEM_PROMPT]' }}{% elif message['role'] == 'assistant' %}{{ ' ' + message['content'] + eos_token }}{% else %}{{ raise_exception('Only user, system and assistant roles are supported!') }}{% endif %}{% endfor %}",
+            /* .expected_output= */ "[SYSTEM_PROMPT] You are a helpful assistant[/SYSTEM_PROMPT][INST] Hello[/INST] Hi there[INST] Who are you[/INST]    I am an assistant   [INST] Another question[/INST]",
+        },
+        {
+            /* .name= */ "ai-sage/GigaChat-20B-A3B-instruct",
+            /* .template_str= */ "{% if messages[0]['role'] == 'system' -%}\n    {%- set loop_messages = messages[1:] -%}\n    {%- set system_message = bos_token + messages[0]['content'] + additional_special_tokens[1] -%}\n{%- else -%}\n    {%- set loop_messages = messages -%}\n    {%- set system_message = bos_token + '' -%}\n{%- endif -%}\n{%- for message in loop_messages %}\n    {% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n        {{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}\n    {% endif %}\n    \n    {%- if loop.index0 == 0 -%}\n        {{ system_message -}}\n    {%- endif -%}\n    {%- if message['role'] == 'user' -%}\n        {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n        {{ 'available functions' + additional_special_tokens[0] + additional_special_tokens[2] + additional_special_tokens[3]  + additional_special_tokens[1] -}}\n    {%- endif -%}\n    {%- if message['role'] == 'assistant' -%}\n        {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n    {%- endif -%}\n    {%- if loop.last and add_generation_prompt -%}\n        {{ 'assistant' + additional_special_tokens[0] -}}\n    {%- endif -%}\n{%- endfor %}",
+            /* .expected_output= */ "You are a helpful assistant<|message_sep|>user<|role_sep|>Hello<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>Hi there<|message_sep|>user<|role_sep|>Who are you<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>   I am an assistant   <|message_sep|>user<|role_sep|>Another question<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>",
+        },
+        {
+            /* .name= */ "Infinigence/Megrez-3B-Instruct",
+            /* .template_str= */ u8"{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|role_start|>system<|role_end|>你是Megrez-3B-Instruct，将针对用户的问题给出详细的、积极的回答。<|turn_end|>' }}{% endif %}{{ '<|role_start|>' + message['role'] + '<|role_end|>' + message['content'] + '<|turn_end|>' }}{% endfor %}{% if add_generation_prompt %}{{ '<|role_start|>assistant<|role_end|>' }}{% endif %}",
+            /* .expected_output= */ "<|role_start|>system<|role_end|>You are a helpful assistant<|turn_end|><|role_start|>user<|role_end|>Hello<|turn_end|><|role_start|>assistant<|role_end|>Hi there<|turn_end|><|role_start|>user<|role_end|>Who are you<|turn_end|><|role_start|>assistant<|role_end|>   I am an assistant   <|turn_end|><|role_start|>user<|role_end|>Another question<|turn_end|><|role_start|>assistant<|role_end|>",
+        },
+        {
+            /* .name= */ "phi-4",
+            /* .template_str= */ "{% for message in messages %}{% if (message['role'] == 'system') %}{{'<|im_start|>system<|im_sep|>' + message['content'] + '<|im_end|>'}}{% elif (message['role'] == 'user') %}{{'<|im_start|>user<|im_sep|>' + message['content'] + '<|im_end|><|im_start|>assistant<|im_sep|>'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|im_end|>'}}{% endif %}{% endfor %}",
+            /* .expected_output= */ "<|im_start|>system<|im_sep|>You are a helpful assistant<|im_end|><|im_start|>user<|im_sep|>Hello<|im_end|><|im_start|>assistant<|im_sep|>Hi there<|im_end|><|im_start|>user<|im_sep|>Who are you<|im_end|><|im_start|>assistant<|im_sep|>   I am an assistant   <|im_end|><|im_start|>user<|im_sep|>Another question<|im_end|><|im_start|>assistant<|im_sep|>",
+        },
     };
     std::vector formatted_chat(1024);
     int32_t res;
 
+    // list all supported templates
+    std::vector supported_tmpl;
+    res = llama_chat_builtin_templates(nullptr, 0);
+    assert(res > 0);
+    supported_tmpl.resize(res);
+    res = llama_chat_builtin_templates(supported_tmpl.data(), supported_tmpl.size());
+    printf("Built-in chat templates:\n");
+    for (auto tmpl : supported_tmpl) {
+        printf("  %s\n", tmpl);
+    }
+
     // test invalid chat template
-    res = llama_chat_apply_template(nullptr, "INVALID TEMPLATE", conversation, message_count, true, formatted_chat.data(), formatted_chat.size());
+    res = llama_chat_apply_template("INVALID TEMPLATE", conversation.data(), conversation.size(), true, formatted_chat.data(), formatted_chat.size());
     assert(res < 0);
 
-    for (size_t i = 0; i < templates.size(); i++) {
-        std::string custom_template = templates[i];
-        std::string expected = expected_output[i];
+    for (const auto & test_case : test_cases) {
+        printf("\n\n=== %s ===\n\n", test_case.name.c_str());
         formatted_chat.resize(1024);
         res = llama_chat_apply_template(
-            nullptr,
-            custom_template.c_str(),
-            conversation,
-            message_count,
+            test_case.template_str.c_str(),
+            conversation.data(),
+            conversation.size(),
             true,
             formatted_chat.data(),
             formatted_chat.size()
@@ -138,7 +206,7 @@ int main(void) {
         std::string output(formatted_chat.data(), formatted_chat.size());
         printf("%s\n", output.c_str());
         printf("-------------------------\n");
-        assert(output == expected);
+        assert(output == test_case.expected_output);
     }
 
 
@@ -154,9 +222,16 @@ int main(void) {
         return output;
     };
     assert(fmt_sys("chatml") == "<|im_start|>system\nYou are a helpful assistant<|im_end|>\n");
+    assert(fmt_sys("mistral-v1") == " [INST] You are a helpful assistant\n\n");
+    assert(fmt_sys("mistral-v3") == "[INST] You are a helpful assistant\n\n");
+    assert(fmt_sys("mistral-v3-tekken") == "[INST]You are a helpful assistant\n\n");
+    assert(fmt_sys("mistral-v7") == "[SYSTEM_PROMPT] You are a helpful assistant[/SYSTEM_PROMPT]");
     assert(fmt_sys("llama2") == "[INST] You are a helpful assistant\n");
+    assert(fmt_sys("llama2-sys") == "[INST] <>\nYou are a helpful assistant\n<>\n\n");
+    assert(fmt_sys("mistral") == "[INST] You are a helpful assistant\n"); // for old pre-v1 templates
     assert(fmt_sys("gemma")  == ""); // for gemma, system message is merged with user message
     assert(fmt_sys("llama3") == "<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|>");
+    assert(fmt_sys("gigachat") == "You are a helpful assistant<|message_sep|>");
 
 
     // test llama_chat_format_single for user message
@@ -173,9 +248,17 @@ int main(void) {
         return output;
     };
     assert(fmt_single("chatml") == "\n<|im_start|>user\nHow are you<|im_end|>\n<|im_start|>assistant\n");
+    assert(fmt_single("mistral-v1") == " [INST] How are you [/INST]");
+    assert(fmt_single("mistral-v3") == "[INST] How are you[/INST]");
+    assert(fmt_single("mistral-v3-tekken") == "[INST]How are you[/INST]");
+    assert(fmt_single("mistral-v7") == "[INST] How are you[/INST]");
     assert(fmt_single("llama2") == "[INST] How are you [/INST]");
+    assert(fmt_single("mistral") == "[INST] How are you [/INST]"); // for old pre-v1 templates
     assert(fmt_single("gemma")  == "\nuser\nHow are you\nmodel\n");
     assert(fmt_single("llama3") == "<|start_header_id|>user<|end_header_id|>\n\nHow are you<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n");
+    assert(fmt_single("gigachat") == "user<|role_sep|>How are you<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>");
+
+    printf("Test chat templates: OK\n");
 
     return 0;
 }
diff --git a/tests/test-gguf.cpp b/tests/test-gguf.cpp
new file mode 100644
index 000000000..611957ac0
--- /dev/null
+++ b/tests/test-gguf.cpp
@@ -0,0 +1,1334 @@
+#include "ggml.h"
+#include "ggml-backend.h"
+#include "../ggml/src/ggml-impl.h"
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+constexpr int offset_has_kv      = 1000;
+constexpr int offset_has_tensors = 2000;
+constexpr int offset_has_data    = 3000;
+
+enum handcrafted_file_type {
+    HANDCRAFTED_HEADER_BAD_MAGIC           =  10,
+    HANDCRAFTED_HEADER_BAD_VERSION_1       =  20,
+    HANDCRAFTED_HEADER_BAD_VERSION_FUTURE  =  30,
+    HANDCRAFTED_HEADER_BAD_N_TENSORS       =  40,
+    HANDCRAFTED_HEADER_BAD_N_KV            =  50,
+    HANDCRAFTED_HEADER_EMPTY               = 800,
+
+    HANDCRAFTED_KV_BAD_KEY_SIZE            =  10 + offset_has_kv,
+    HANDCRAFTED_KV_BAD_TYPE                =  20 + offset_has_kv,
+    // HANDCRAFTED_KV_BAD_VALUE_SIZE          =  30 + offset_has_kv, // removed because it can result in allocations > 1 TB (default sanitizer limit)
+    HANDCRAFTED_KV_DUPLICATE_KEY           =  40 + offset_has_kv,
+    HANDCRAFTED_KV_BAD_ALIGN               =  50 + offset_has_kv,
+    HANDCRAFTED_KV_SUCCESS                 = 800 + offset_has_kv,
+
+    HANDCRAFTED_TENSORS_BAD_NAME_SIZE      =  10 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_BAD_N_DIMS         =  20 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_BAD_SHAPE          =  30 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_NE_TOO_BIG         =  40 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_BAD_TYPE           =  50 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_BAD_OFFSET         =  60 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_DUPLICATE_NAME     =  70 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_BAD_ALIGN          =  75 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_INCONSISTENT_ALIGN =  80 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_SUCCESS            = 800 + offset_has_tensors,
+    HANDCRAFTED_TENSORS_CUSTOM_ALIGN       = 810 + offset_has_tensors,
+
+    HANDCRAFTED_DATA_NOT_ENOUGH_DATA       =  10 + offset_has_data,
+    HANDCRAFTED_DATA_BAD_ALIGN             =  15 + offset_has_data,
+    HANDCRAFTED_DATA_INCONSISTENT_ALIGN    =  20 + offset_has_data,
+    HANDCRAFTED_DATA_SUCCESS               = 800 + offset_has_data,
+    HANDCRAFTED_DATA_CUSTOM_ALIGN          = 810 + offset_has_data,
+};
+
+std::string handcrafted_file_type_name(const enum handcrafted_file_type hft) {
+    switch (hft) {
+        case HANDCRAFTED_HEADER_BAD_MAGIC:           return "HEADER_BAD_MAGIC";
+        case HANDCRAFTED_HEADER_BAD_VERSION_1:       return "HEADER_BAD_VERSION_1";
+        case HANDCRAFTED_HEADER_BAD_VERSION_FUTURE:  return "HEADER_BAD_VERSION_FUTURE";
+        case HANDCRAFTED_HEADER_BAD_N_KV:            return "HEADER_BAD_N_KV";
+        case HANDCRAFTED_HEADER_BAD_N_TENSORS:       return "HEADER_BAD_N_TENSORS";
+        case HANDCRAFTED_HEADER_EMPTY:               return "HEADER_EMPTY";
+
+        case HANDCRAFTED_KV_BAD_KEY_SIZE:            return "KV_BAD_KEY_SIZE";
+        case HANDCRAFTED_KV_BAD_TYPE:                return "KV_BAD_TYPE";
+        case HANDCRAFTED_KV_DUPLICATE_KEY:           return "KV_DUPLICATE_KEY";
+        case HANDCRAFTED_KV_BAD_ALIGN:               return "KV_BAD_ALIGN";
+        case HANDCRAFTED_KV_SUCCESS:                 return "KV_RANDOM_KV";
+
+        case HANDCRAFTED_TENSORS_BAD_NAME_SIZE:      return "TENSORS_BAD_NAME_SIZE";
+        case HANDCRAFTED_TENSORS_BAD_N_DIMS:         return "TENSORS_BAD_N_DIMS";
+        case HANDCRAFTED_TENSORS_BAD_SHAPE:          return "TENSORS_BAD_SHAPE";
+        case HANDCRAFTED_TENSORS_NE_TOO_BIG:         return "TENSORS_NE_TOO_BIG";
+        case HANDCRAFTED_TENSORS_BAD_TYPE:           return "TENSORS_BAD_TYPE";
+        case HANDCRAFTED_TENSORS_BAD_OFFSET:         return "TENSORS_BAD_OFFSET";
+        case HANDCRAFTED_TENSORS_DUPLICATE_NAME:     return "TENSORS_DUPLICATE_NAME";
+        case HANDCRAFTED_TENSORS_BAD_ALIGN:          return "TENSORS_BAD_ALIGN";
+        case HANDCRAFTED_TENSORS_INCONSISTENT_ALIGN: return "TENSORS_INCONSISTENT_ALIGN";
+        case HANDCRAFTED_TENSORS_SUCCESS:            return "TENSORS_SUCCESS";
+        case HANDCRAFTED_TENSORS_CUSTOM_ALIGN:       return "TENSORS_CUSTOM_ALIGN";
+
+        case HANDCRAFTED_DATA_NOT_ENOUGH_DATA:       return "DATA_NOT_ENOUGH_DATA";
+        case HANDCRAFTED_DATA_BAD_ALIGN:             return "DATA_BAD_ALIGN";
+        case HANDCRAFTED_DATA_INCONSISTENT_ALIGN:    return "DATA_INCONSISTENT_ALIGN";
+        case HANDCRAFTED_DATA_SUCCESS:               return "DATA_SUCCESS";
+        case HANDCRAFTED_DATA_CUSTOM_ALIGN:          return "DATA_CUSTOM_ALIGN";
+    }
+    GGML_ABORT("fatal error");
+}
+
+static bool expect_context_not_null(const enum handcrafted_file_type hft) {
+    if (hft < offset_has_kv) {
+        return hft >= HANDCRAFTED_HEADER_EMPTY;
+    }
+    if (hft < offset_has_tensors) {
+        return hft >= HANDCRAFTED_KV_SUCCESS;
+    }
+    if (hft < offset_has_data) {
+        return hft >= HANDCRAFTED_TENSORS_SUCCESS;
+    }
+    return hft >= HANDCRAFTED_DATA_SUCCESS;
+}
+
+typedef std::pair> tensor_config_t;
+
+std::vector get_tensor_configs(std::mt19937 & rng) {
+    std::vector tensor_configs;
+    tensor_configs.reserve(100);
+
+    for (int i = 0; i < 100; ++i) {
+        const enum ggml_type type = ggml_type(rng() % GGML_TYPE_COUNT);
+        if (ggml_type_size(type) == 0) {
+            continue;
+        }
+
+        std::array shape = {1, 1, 1, 1};
+        shape[0] = (1 + rng() % 10) * ggml_blck_size(type);
+        const int n_dims = 1 + rng() % GGML_MAX_DIMS;
+        for (int i = 1; i < n_dims; ++i) {
+            shape[i] = 1 + rng() % 10;
+        }
+
+        tensor_configs.push_back(std::make_pair(type, shape));
+    }
+
+    return tensor_configs;
+}
+
+std::vector> get_kv_types(std::mt19937 rng) {
+    std::vector> kv_types;
+    kv_types.reserve(100);
+
+    for (int i = 0; i < 100; ++i) {
+        const gguf_type type = gguf_type(rng() % GGUF_TYPE_COUNT);
+
+        if (type == GGUF_TYPE_ARRAY) {
+            const gguf_type type_arr = gguf_type(rng() % GGUF_TYPE_COUNT);
+            if (type_arr == GGUF_TYPE_ARRAY) {
+                continue;
+            }
+            kv_types.push_back(std::make_pair(type, type_arr));
+            continue;
+        }
+
+        kv_types.push_back(std::make_pair(type, gguf_type(-1)));
+    }
+    std::shuffle(kv_types.begin(), kv_types.end(), rng);
+
+    return kv_types;
+}
+
+template 
+static void helper_write(FILE * file, const T & val) {
+    GGML_ASSERT(fwrite(&val, 1, sizeof(val), file) == sizeof(val));
+}
+
+static void helper_write(FILE * file, const void * data, const size_t nbytes) {
+    GGML_ASSERT(fwrite(data, 1, nbytes, file) == nbytes);
+}
+
+static FILE * get_handcrafted_file(const unsigned int seed, const enum handcrafted_file_type hft, const int extra_bytes = 0) {
+    FILE * file = tmpfile();
+
+    if (!file) {
+        return file;
+    }
+
+    std::mt19937 rng(seed);
+    uint32_t alignment = GGUF_DEFAULT_ALIGNMENT;
+
+    if (hft == HANDCRAFTED_HEADER_BAD_MAGIC) {
+        const char bad_magic[4] = {'F', 'U', 'G', 'G'};
+        helper_write(file, bad_magic, sizeof(bad_magic));
+    } else {
+        helper_write(file, GGUF_MAGIC, 4);
+    }
+
+    if (hft == HANDCRAFTED_HEADER_BAD_VERSION_1) {
+        const uint32_t version = 1;
+        helper_write(file, version);
+    } else if (hft == HANDCRAFTED_HEADER_BAD_VERSION_FUTURE) {
+        const uint32_t version = GGUF_VERSION + 1;
+        helper_write(file, version);
+    } else {
+        const uint32_t version = GGUF_VERSION;
+        helper_write(file, version);
+    }
+
+    std::vector tensor_configs;
+    if (hft >= offset_has_tensors) {
+        tensor_configs = get_tensor_configs(rng);
+    }
+
+    if (hft == HANDCRAFTED_HEADER_BAD_N_TENSORS) {
+        const uint64_t n_tensors = -1;
+        helper_write(file, n_tensors);
+    } else {
+        const uint64_t n_tensors = tensor_configs.size();
+        helper_write(file, n_tensors);
+    }
+
+    std::vector> kv_types;
+    if (hft >= offset_has_kv) {
+        kv_types = get_kv_types(rng);
+    }
+    {
+        uint64_t n_kv = kv_types.size();
+        if (hft == HANDCRAFTED_KV_BAD_ALIGN      ||
+            hft == HANDCRAFTED_TENSORS_BAD_ALIGN || hft == HANDCRAFTED_TENSORS_CUSTOM_ALIGN ||
+            hft == HANDCRAFTED_DATA_BAD_ALIGN    || hft == HANDCRAFTED_DATA_CUSTOM_ALIGN) {
+
+            n_kv += 1;
+        } else if (hft == HANDCRAFTED_HEADER_BAD_N_KV) {
+            n_kv = -1;
+        }
+        helper_write(file, n_kv);
+    }
+
+    if (hft < offset_has_kv) {
+        while (ftell(file) % alignment != 0) {
+            const char pad = 0;
+            helper_write(file, pad);
+        }
+
+        for (int i = 0; i < extra_bytes; ++i) {
+            const char tmp = 0;
+            helper_write(file, tmp);
+        }
+        rewind(file);
+        return file;
+    }
+
+    for (int i = 0; i < int(kv_types.size()); ++i) {
+        const enum gguf_type type     = gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].first);
+        const enum gguf_type type_arr = gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].second);
+
+        const std::string key = "my_key_" + std::to_string((hft == HANDCRAFTED_KV_DUPLICATE_KEY ? i/2 : i));
+
+        if (hft == HANDCRAFTED_KV_BAD_KEY_SIZE) {
+            const uint64_t n = -1;
+            helper_write(file, n);
+        } else {
+            const uint64_t n = key.length();
+            helper_write(file, n);
+        }
+        helper_write(file, key.data(), key.length());
+
+        {
+            const int32_t type32 = int32_t(type);
+            helper_write(file, type32);
+        }
+
+        uint32_t data[16];
+        for (int j = 0; j < 16; ++j) {
+            data[j] = rng();
+            if (type == GGUF_TYPE_STRING || type_arr == GGUF_TYPE_STRING) {
+                data[j] |= 0x01010101; // avoid random null-termination of string
+            }
+        }
+
+        if (type == GGUF_TYPE_STRING) {
+            const uint64_t n = rng() % sizeof(data);
+            helper_write(file, n);
+            helper_write(file, data, n);
+            continue;
+        }
+
+        if (type == GGUF_TYPE_ARRAY) {
+            {
+                const int32_t type32 = int32_t(type_arr);
+                helper_write(file, type32);
+            }
+            if (type_arr == GGUF_TYPE_STRING) {
+                const uint64_t nstr = rng() % (16 + 1);
+                helper_write(file, nstr);
+                for (uint64_t istr = 0; istr < nstr; ++istr) {
+                    const uint64_t n = rng() % (sizeof(uint32_t) + 1);
+                    helper_write(file, n);
+                    helper_write(file, &data[istr], n);
+                }
+                continue;
+            }
+            const size_t type_size = gguf_type_size(type_arr);
+            const uint64_t n = (rng() % sizeof(data)) / type_size;
+            helper_write(file, n);
+            helper_write(file, &data, n*type_size);
+            continue;
+        }
+
+        helper_write(file, data, hft == HANDCRAFTED_KV_BAD_TYPE ? 1 : gguf_type_size(type));
+    }
+
+    if (hft == HANDCRAFTED_KV_BAD_ALIGN      ||
+        hft == HANDCRAFTED_TENSORS_BAD_ALIGN || hft == HANDCRAFTED_TENSORS_CUSTOM_ALIGN ||
+        hft == HANDCRAFTED_DATA_BAD_ALIGN    || hft == HANDCRAFTED_DATA_CUSTOM_ALIGN) {
+
+        const uint64_t n = strlen(GGUF_KEY_GENERAL_ALIGNMENT);
+        helper_write(file, n);
+        helper_write(file, GGUF_KEY_GENERAL_ALIGNMENT, n);
+
+        const int32_t type = gguf_type(GGUF_TYPE_UINT32);
+        helper_write(file, type);
+
+        alignment = expect_context_not_null(hft) ? 1 : 13;
+        helper_write(file, alignment);
+    }
+
+    if (hft < offset_has_tensors) {
+        while (ftell(file) % alignment != 0) {
+            const char pad = 0;
+            helper_write(file, pad);
+        }
+
+        for (int i = 0; i < extra_bytes; ++i) {
+            const char tmp = 0;
+            helper_write(file, tmp);
+        }
+        rewind(file);
+        return file;
+    }
+
+    if (hft == HANDCRAFTED_TENSORS_INCONSISTENT_ALIGN || hft == HANDCRAFTED_DATA_INCONSISTENT_ALIGN) {
+        alignment = 1;
+    }
+
+    uint64_t offset = 0;
+    for (int i = 0; i < int(tensor_configs.size()); ++i) {
+        const ggml_type                          type  = tensor_configs[i].first;
+        const std::array shape = tensor_configs[i].second;
+
+        std::string name = "my_tensor";
+        if (hft != HANDCRAFTED_TENSORS_DUPLICATE_NAME) {
+            name += "_" + std::to_string(i);
+        }
+        if (hft == HANDCRAFTED_TENSORS_BAD_NAME_SIZE) {
+            name += "_with_a_very_long_name_which_is_longer_than_what_is_allowed_for_ggml_tensors";
+            GGML_ASSERT(name.length() >= GGML_MAX_NAME);
+        }
+        {
+            const uint64_t n = name.length();
+            helper_write(file, n);
+        }
+        helper_write(file, name.data(), name.length());
+
+        uint32_t n_dims = hft == HANDCRAFTED_TENSORS_NE_TOO_BIG ? 2 : 1;
+        for (int i = GGML_MAX_DIMS-1; i >= 1; --i) {
+            if (shape[i] != 1) {
+                n_dims = i + 1;
+                break;
+            }
+        }
+        if (hft == HANDCRAFTED_TENSORS_BAD_N_DIMS) {
+            const uint32_t n_dims_bad = GGML_MAX_DIMS + 1;
+            helper_write(file, n_dims_bad);
+        } else {
+            helper_write(file, n_dims);
+        }
+
+        if (hft == HANDCRAFTED_TENSORS_BAD_SHAPE) {
+            for (uint32_t j = 0; j < n_dims; ++j) {
+                const int64_t bad_dim = -1;
+                helper_write(file, bad_dim);
+            }
+        } else if (hft == HANDCRAFTED_TENSORS_NE_TOO_BIG){
+            for (uint32_t j = 0; j < n_dims; ++j) {
+                const int64_t big_dim = 4*int64_t(INT32_MAX);
+                helper_write(file, big_dim);
+            }
+        } else {
+            helper_write(file, shape.data(), n_dims*sizeof(int64_t));
+        }
+
+        {
+            const int32_t type32 = hft == HANDCRAFTED_TENSORS_BAD_TYPE ? GGML_TYPE_COUNT : int32_t(type);
+            helper_write(file, type32);
+        }
+
+        if (hft == HANDCRAFTED_TENSORS_BAD_OFFSET) {
+            const uint64_t bad_offset = -1;
+            helper_write(file, bad_offset);
+        } else {
+            helper_write(file, offset);
+        }
+
+        int64_t ne = shape[0];
+        for (uint32_t i = 1; i < n_dims; ++i) {
+            ne *= shape[i];
+        }
+        offset += GGML_PAD(ggml_row_size(type, ne), alignment);
+    }
+
+    while (ftell(file) % alignment != 0) {
+        const char pad = 0;
+        helper_write(file, pad);
+    }
+
+    if (hft >= offset_has_data) {
+        rng.seed(seed + 1);
+        uint64_t nbytes = offset;
+        if (hft == HANDCRAFTED_DATA_NOT_ENOUGH_DATA) {
+            nbytes -= 1;
+        }
+        for (uint64_t i = 0; i < nbytes; ++i) {
+            const uint8_t random_byte = i % 256;
+            helper_write(file, random_byte);
+        }
+    }
+
+    for (int i = 0; i < extra_bytes; ++i) {
+        const char tmp = 0;
+        helper_write(file, tmp);
+    }
+    rewind(file);
+    return file;
+}
+
+static bool handcrafted_check_header(const gguf_context * gguf_ctx, const unsigned int seed, const bool has_kv, const bool has_tensors, const bool alignment_defined) {
+    if (!gguf_ctx) {
+        return false;
+    }
+
+    std::mt19937 rng(seed);
+
+    std::vector tensor_configs;
+    if (has_tensors) {
+        tensor_configs = get_tensor_configs(rng);
+    }
+    std::vector> kv_types;
+    if (has_kv) {
+        kv_types = get_kv_types(rng);
+    }
+
+    bool ok = true;
+
+    if (gguf_get_version(gguf_ctx) != GGUF_VERSION) {
+        ok = false;
+    }
+    if (gguf_get_n_tensors(gguf_ctx) != int(tensor_configs.size())) {
+        ok = false;
+    }
+    if (gguf_get_n_kv(gguf_ctx) != int(alignment_defined ? kv_types.size() + 1 : kv_types.size())) {
+        ok = false;
+    }
+
+    return ok;
+}
+
+static bool handcrafted_check_kv(const gguf_context * gguf_ctx, const unsigned int seed, const bool has_tensors, const bool alignment_defined) {
+    if (!gguf_ctx) {
+        return false;
+    }
+
+    std::mt19937 rng(seed);
+
+    std::vector tensor_configs;
+    if (has_tensors) {
+        tensor_configs = get_tensor_configs(rng);
+    }
+
+    std::vector> kv_types = get_kv_types(rng);
+
+    bool ok = true;
+
+    for (int i = 0; i < int(kv_types.size()); ++i) {
+        const enum gguf_type type     = gguf_type(kv_types[i].first);
+        const enum gguf_type type_arr = gguf_type(kv_types[i].second);
+
+        const std::string key = "my_key_" + std::to_string(i);
+
+        uint32_t data[16];
+        for (int j = 0; j < 16; ++j) {
+            data[j] = rng();
+            if (type == GGUF_TYPE_STRING || type_arr == GGUF_TYPE_STRING) {
+                data[j] |= 0x01010101; // avoid random null-termination of string
+            }
+        }
+
+        const char * data8 = reinterpret_cast(data);
+        const int id = gguf_find_key(gguf_ctx, key.c_str());
+
+        if (type == GGUF_TYPE_STRING) {
+            const char * str = gguf_get_val_str(gguf_ctx, id);
+            const uint64_t n = strlen(str);
+            const uint64_t n_expected = rng() % sizeof(data);
+            if (n != n_expected) {
+                ok = false;
+                continue;
+            }
+            if (!std::equal(str, str + n, data8)) {
+                ok = false;
+            }
+            continue;
+        }
+
+        if (type == GGUF_TYPE_ARRAY) {
+            const size_t type_size = gguf_type_size(type_arr);
+            const uint64_t arr_n = gguf_get_arr_n(gguf_ctx, id);
+
+            if (type_arr == GGUF_TYPE_STRING) {
+                const uint64_t nstr_expected = rng() % (16 + 1);
+                if (arr_n != nstr_expected) {
+                    ok = false;
+                    continue;
+                }
+                for (uint64_t istr = 0; istr < nstr_expected; ++istr) {
+                    const char * str = gguf_get_arr_str(gguf_ctx, id, istr);
+                    const uint64_t n = strlen(str);
+                    const uint64_t n_expected = rng() % (sizeof(uint32_t) + 1);
+
+                    if (n != n_expected) {
+                        ok = false;
+                        continue;
+                    }
+                    const char * str_expected = reinterpret_cast(&data[istr]);
+                    if (strncmp(str, str_expected, n) != 0) {
+                        ok = false;
+                        continue;
+                    }
+                }
+                continue;
+            }
+
+            const uint64_t arr_n_expected = (rng() % sizeof(data)) / type_size;
+            if (arr_n != arr_n_expected) {
+                ok = false;
+                continue;
+            }
+
+            const char * data_gguf = reinterpret_cast(gguf_get_arr_data(gguf_ctx, id));
+
+            if (type_arr == GGUF_TYPE_BOOL) {
+                for (size_t arr_i = 0; arr_i < arr_n; ++arr_i) {
+                    if (bool(data8[arr_i]) != bool(data_gguf[arr_i])) {
+                        ok = false;
+                    }
+                }
+                continue;
+            }
+
+            if (!std::equal(data8, data8 + arr_n*type_size, data_gguf)) {
+                ok = false;
+            }
+            continue;
+        }
+
+        const char * data_gguf = reinterpret_cast(gguf_get_val_data(gguf_ctx, id));
+
+        if (type == GGUF_TYPE_BOOL) {
+            if (bool(*data8) != bool(*data_gguf)) {
+                ok = false;
+            }
+            continue;
+        }
+
+        if (!std::equal(data8, data8 + gguf_type_size(type), data_gguf)) {
+            ok = false;
+        }
+    }
+
+    const uint32_t expected_alignment = alignment_defined ? 1 : GGUF_DEFAULT_ALIGNMENT;
+    if (gguf_get_alignment(gguf_ctx) != expected_alignment) {
+        ok = false;
+    }
+
+    return ok;
+}
+
+static bool handcrafted_check_tensors(const gguf_context * gguf_ctx, const unsigned int seed) {
+    if (!gguf_ctx) {
+        return false;
+    }
+
+    std::mt19937 rng(seed);
+
+    std::vector tensor_configs = get_tensor_configs(rng);
+
+    // Call get_kv_types to get the same RNG state:
+    get_kv_types(rng);
+
+    bool ok = true;
+
+    const int id_alignment = gguf_find_key(gguf_ctx, GGUF_KEY_GENERAL_ALIGNMENT);
+    const uint32_t alignment = id_alignment >= 0 ? gguf_get_val_u32(gguf_ctx, id_alignment) : GGUF_DEFAULT_ALIGNMENT;
+
+    uint64_t expected_offset = 0;
+    for (int i = 0; i < int(tensor_configs.size()); ++i) {
+        const ggml_type                          type  = tensor_configs[i].first;
+        const std::array shape = tensor_configs[i].second;
+
+        const std::string name = "my_tensor_" + std::to_string(i);
+        const int id = gguf_find_tensor(gguf_ctx, name.c_str());
+
+        if (id >= 0) {
+            if (std::string(gguf_get_tensor_name(gguf_ctx, id)) != name) {
+                ok = false;
+            }
+
+            if (gguf_get_tensor_type(gguf_ctx, id) != type) {
+                ok = false;
+            }
+        } else {
+            ok = false;
+            continue;
+        }
+
+        const size_t offset = gguf_get_tensor_offset(gguf_ctx, id);
+
+        if (offset != expected_offset) {
+            ok = false;
+        }
+
+        int64_t ne = shape[0];
+        for (size_t j = 1; j < GGML_MAX_DIMS; ++j) {
+            ne *= shape[j];
+        }
+        expected_offset += GGML_PAD(ggml_row_size(type, ne), alignment);
+    }
+
+    return ok;
+}
+
+static bool handcrafted_check_tensor_data(const gguf_context * gguf_ctx, const unsigned int seed, FILE * file) {
+    if (!gguf_ctx) {
+        return false;
+    }
+
+    std::mt19937 rng(seed);
+
+    std::vector tensor_configs = get_tensor_configs(rng);
+
+    bool ok = true;
+
+    const uint32_t alignment = GGUF_DEFAULT_ALIGNMENT;
+
+    for (int i = 0; i < int(tensor_configs.size()); ++i) {
+        const ggml_type                          type  = tensor_configs[i].first;
+        const std::array shape = tensor_configs[i].second;
+
+        int64_t ne = shape[0];
+        for (size_t j = 1; j < GGML_MAX_DIMS; ++j) {
+            ne *= shape[j];
+        }
+        const size_t size = ggml_row_size(type, ne);
+
+        const std::string name = "my_tensor_" + std::to_string(i);
+        const size_t offset = gguf_get_tensor_offset(gguf_ctx, gguf_find_tensor(gguf_ctx, name.c_str()));
+
+        std::vector data(size);
+        GGML_ASSERT(fseek(file, gguf_get_data_offset(gguf_ctx) + offset, SEEK_SET) == 0);
+        GGML_ASSERT(fread(data.data(), 1, data.size(), file) == data.size());
+
+        for (size_t j = 0; j < size; ++j) {
+            const uint8_t expected_byte = (j + offset) % 256;
+            if (data[j] != expected_byte) {
+                ok = false;
+            }
+        }
+    }
+
+    return ok;
+}
+
+static std::pair test_handcrafted_file(const unsigned int seed) {
+    int npass = 0;
+    int ntest = 0;
+
+    const std::vector hfts = {
+        HANDCRAFTED_HEADER_BAD_MAGIC,
+        HANDCRAFTED_HEADER_BAD_VERSION_1,
+        HANDCRAFTED_HEADER_BAD_VERSION_FUTURE,
+        HANDCRAFTED_HEADER_BAD_N_KV,
+        HANDCRAFTED_HEADER_BAD_N_TENSORS,
+        HANDCRAFTED_HEADER_EMPTY,
+
+        HANDCRAFTED_KV_BAD_KEY_SIZE,
+        HANDCRAFTED_KV_BAD_TYPE,
+        HANDCRAFTED_KV_DUPLICATE_KEY,
+        HANDCRAFTED_KV_BAD_ALIGN,
+        HANDCRAFTED_KV_SUCCESS,
+
+        HANDCRAFTED_TENSORS_BAD_NAME_SIZE,
+        HANDCRAFTED_TENSORS_BAD_N_DIMS,
+        HANDCRAFTED_TENSORS_BAD_SHAPE,
+        HANDCRAFTED_TENSORS_NE_TOO_BIG,
+        HANDCRAFTED_TENSORS_BAD_TYPE,
+        HANDCRAFTED_TENSORS_BAD_OFFSET,
+        HANDCRAFTED_TENSORS_DUPLICATE_NAME,
+        HANDCRAFTED_TENSORS_BAD_ALIGN,
+        HANDCRAFTED_TENSORS_INCONSISTENT_ALIGN,
+        HANDCRAFTED_TENSORS_SUCCESS,
+        HANDCRAFTED_TENSORS_CUSTOM_ALIGN,
+
+        HANDCRAFTED_DATA_NOT_ENOUGH_DATA,
+        HANDCRAFTED_DATA_BAD_ALIGN,
+        HANDCRAFTED_DATA_INCONSISTENT_ALIGN,
+        HANDCRAFTED_DATA_SUCCESS,
+        HANDCRAFTED_DATA_CUSTOM_ALIGN,
+    };
+
+    for (enum handcrafted_file_type hft : hfts) {
+        printf("%s: handcrafted_file_type=%s\n", __func__, handcrafted_file_type_name(hft).c_str());
+        FILE * file = get_handcrafted_file(seed, hft);
+
+#ifdef _WIN32
+        if (!file) {
+            printf("%s: failed to create tmpfile(), needs elevated privileges on Windows");
+            printf("%s: skipping tests");
+            continue;
+        }
+#else
+        GGML_ASSERT(file);
+#endif // _WIN32
+
+        struct ggml_context * ctx = nullptr;
+        struct gguf_init_params gguf_params = {
+            /*no_alloc =*/ false,
+            /*ctx      =*/ hft >= offset_has_data ? &ctx : nullptr,
+        };
+
+        struct gguf_context * gguf_ctx = gguf_init_from_file_impl(file, gguf_params);
+
+        if (expect_context_not_null(hft)) {
+            printf("%s:   - context_not_null: ", __func__);
+        } else {
+            printf("%s:   - context_null: ", __func__);
+        }
+        if (bool(gguf_ctx) == expect_context_not_null(hft)) {
+            printf("\033[1;32mOK\033[0m\n");
+            npass++;
+        } else {
+            printf("\033[1;31mFAIL\033[0m\n");
+        }
+        ntest++;
+
+        if (hft >= offset_has_data && !expect_context_not_null(hft)) {
+            printf("%s:   - no_dangling_ggml_context_pointer: ", __func__);
+            if (ctx) {
+                printf("\033[1;31mFAIL\033[0m\n");
+            } else {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            }
+            ntest++;
+        }
+
+        const bool alignment_defined = hft == HANDCRAFTED_TENSORS_CUSTOM_ALIGN || hft == HANDCRAFTED_DATA_CUSTOM_ALIGN;
+
+        if (expect_context_not_null(hft)) {
+            printf("%s:   - check_header: ", __func__);
+            if (handcrafted_check_header(gguf_ctx, seed, hft >= offset_has_kv, hft >= offset_has_tensors, alignment_defined)) {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            } else {
+                printf("\033[1;31mFAIL\033[0m\n");
+            }
+            ntest++;
+        }
+
+        if (expect_context_not_null(hft) && hft >= offset_has_kv) {
+            printf("%s:   - check_kv: ", __func__);
+            if (handcrafted_check_kv(gguf_ctx, seed, hft >= offset_has_tensors, alignment_defined)) {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            } else {
+                printf("\033[1;31mFAIL\033[0m\n");
+            }
+            ntest++;
+        }
+
+        if (expect_context_not_null(hft) && hft >= offset_has_tensors) {
+            printf("%s:   - check_tensors: ", __func__);
+            if (handcrafted_check_tensors(gguf_ctx, seed)) {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            } else {
+                printf("\033[1;31mFAIL\033[0m\n");
+            }
+            ntest++;
+        }
+
+        if (expect_context_not_null(hft) && hft >= offset_has_data) {
+            printf("%s:   - check_tensor_data: ", __func__);
+            if (handcrafted_check_tensor_data(gguf_ctx, seed, file)) {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            } else {
+                printf("\033[1;31mFAIL\033[0m\n");
+            }
+            ntest++;
+        }
+
+        fclose(file);
+        if (gguf_ctx) {
+            ggml_free(ctx);
+            gguf_free(gguf_ctx);
+        }
+        printf("\n");
+    }
+
+
+    return std::make_pair(npass, ntest);
+}
+
+struct random_gguf_context_result {
+    struct gguf_context * gguf_ctx;
+    struct ggml_context * ctx;
+    ggml_backend_buffer_t buffer;
+};
+
+static struct random_gguf_context_result get_random_gguf_context(ggml_backend_t backend, const unsigned int seed) {
+    std::mt19937 rng(seed);
+
+    struct gguf_context * gguf_ctx = gguf_init_empty();
+
+    for (int i = 0; i < 256; ++i) {
+        const std::string key = "my_key_" + std::to_string(rng() % 1024);
+        const enum gguf_type type = gguf_type(rng() % GGUF_TYPE_COUNT);
+
+        switch (type) {
+            case GGUF_TYPE_UINT8:   gguf_set_val_u8  (gguf_ctx, key.c_str(), rng() % (1 <<  7));             break;
+            case GGUF_TYPE_INT8:    gguf_set_val_i8  (gguf_ctx, key.c_str(), rng() % (1 <<  7) - (1 <<  6)); break;
+            case GGUF_TYPE_UINT16:  gguf_set_val_u16 (gguf_ctx, key.c_str(), rng() % (1 << 15));             break;
+            case GGUF_TYPE_INT16:   gguf_set_val_i16 (gguf_ctx, key.c_str(), rng() % (1 << 15) - (1 << 14)); break;
+            case GGUF_TYPE_UINT32:  gguf_set_val_u32 (gguf_ctx, key.c_str(), rng());                         break;
+            case GGUF_TYPE_INT32:   gguf_set_val_i32 (gguf_ctx, key.c_str(), rng()             - (1 << 30)); break;
+            case GGUF_TYPE_FLOAT32: gguf_set_val_f32 (gguf_ctx, key.c_str(), rng() % 1024      - 512);       break;
+            case GGUF_TYPE_BOOL:    gguf_set_val_bool(gguf_ctx, key.c_str(), rng() % 2 == 0);                break;
+            case GGUF_TYPE_STRING:  gguf_set_val_str (gguf_ctx, key.c_str(), std::to_string(rng()).c_str()); break;
+            case GGUF_TYPE_UINT64:  gguf_set_val_u64 (gguf_ctx, key.c_str(), rng());                         break;
+            case GGUF_TYPE_INT64:   gguf_set_val_i64 (gguf_ctx, key.c_str(), rng()             - (1 << 30)); break;
+            case GGUF_TYPE_FLOAT64: gguf_set_val_f32 (gguf_ctx, key.c_str(), rng() % 1024      - 512);       break;
+            case GGUF_TYPE_ARRAY: {
+                const enum gguf_type type_arr = gguf_type(rng() % GGUF_TYPE_COUNT);
+                const uint64_t ne = rng() % 1024;
+
+                switch (type_arr) {
+                    case GGUF_TYPE_UINT8:
+                    case GGUF_TYPE_INT8:
+                    case GGUF_TYPE_UINT16:
+                    case GGUF_TYPE_INT16:
+                    case GGUF_TYPE_UINT32:
+                    case GGUF_TYPE_INT32:
+                    case GGUF_TYPE_FLOAT32:
+                    case GGUF_TYPE_BOOL:
+                    case GGUF_TYPE_UINT64:
+                    case GGUF_TYPE_INT64:
+                    case GGUF_TYPE_FLOAT64: {
+                        const size_t nbytes = ne*gguf_type_size(type_arr);
+                        std::vector random_data((nbytes + sizeof(uint32_t) - 1) / sizeof(uint32_t));
+                        for (size_t j = 0; j < random_data.size(); ++j) {
+                            random_data[j] = rng();
+                            if (type_arr == GGUF_TYPE_BOOL) {
+                                random_data[j] &= 0x01010101; // the sanitizer complains if booleans are not 0 or 1
+                            }
+                        }
+                        gguf_set_arr_data(gguf_ctx, key.c_str(), type_arr, random_data.data(), ne);
+                    } break;
+                    case GGUF_TYPE_STRING: {
+                        std::vector  data_cpp(ne);
+                        std::vector data_c(ne);
+                        for (size_t j = 0; j < data_cpp.size(); ++j) {
+                            data_cpp[j] = std::to_string(rng());
+                            data_c[j]   = data_cpp[j].c_str();
+                        }
+                        gguf_set_arr_str(gguf_ctx, key.c_str(), data_c.data(), ne);
+                    } break;
+                    case GGUF_TYPE_ARRAY: {
+                        break; // not supported
+                    }
+                    case GGUF_TYPE_COUNT:
+                    default: {
+                        GGML_ABORT("fatal error");
+                    } break;
+                }
+            } break;
+            case GGUF_TYPE_COUNT:
+            default: {
+                GGML_ABORT("fatal error");
+            } break;
+        }
+    }
+
+    struct ggml_init_params ggml_params = {
+        /*.mem_size   =*/ 256*ggml_tensor_overhead(),
+        /*.mem_buffer =*/ nullptr,
+        /*.no_alloc   =*/ true,
+    };
+    struct ggml_context * ctx = ggml_init(ggml_params);
+
+    for (int i = 0; i < 256; ++i) {
+        const std::string name = "my_tensor_" + std::to_string(i);
+        const enum ggml_type type = ggml_type(rng() % GGML_TYPE_COUNT);
+        const size_t type_size = ggml_type_size(type);
+
+        if (type_size == 0) {
+            continue;
+        }
+
+        const int n_dims = 1 + rng() % GGML_MAX_DIMS;
+        int64_t ne[GGML_MAX_DIMS];
+        ne[0] = (1 + rng() % 10) * ggml_blck_size(type);
+        for (int j = 1; j < n_dims; ++j) {
+            ne[j] = 1 + rng() % 10;
+        }
+
+        struct ggml_tensor * tensor = ggml_new_tensor(ctx, type, n_dims, ne);
+        ggml_set_name(tensor, name.c_str());
+    }
+
+    ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors(ctx, backend);
+    for (struct ggml_tensor * t = ggml_get_first_tensor(ctx); t != nullptr; t = ggml_get_next_tensor(ctx, t)) {
+        const size_t nbytes = ggml_nbytes(t);
+        std::vector random_data((nbytes + sizeof(uint32_t) - 1) / sizeof(uint32_t));
+        for (size_t j = 0; j < random_data.size(); ++j) {
+            random_data[j] = rng();
+        }
+        ggml_backend_tensor_set(t, random_data.data(), 0, nbytes);
+
+        gguf_add_tensor(gguf_ctx, t);
+    }
+
+    return {gguf_ctx, ctx, buf};
+}
+
+static bool all_kv_in_other(const gguf_context * ctx, const gguf_context * other) {
+    bool ok = true;
+
+    const int n_kv = gguf_get_n_kv(ctx);
+    for (int id = 0; id < n_kv; ++id) {
+        const char * name = gguf_get_key(ctx, id);
+
+        const int idx_other = gguf_find_key(other, name);
+        if (idx_other < 0) {
+            ok = false;
+            continue;
+        }
+
+        const gguf_type type = gguf_get_kv_type(ctx, id);
+        if (type != gguf_get_kv_type(other, idx_other)) {
+            ok = false;
+            continue;
+        }
+
+        if (type == GGUF_TYPE_ARRAY) {
+            const int arr_n = gguf_get_arr_n(ctx, id);
+            if (arr_n != gguf_get_arr_n(other, idx_other)) {
+                ok = false;
+                continue;
+            }
+
+            const gguf_type type_arr = gguf_get_arr_type(ctx, id);
+            if (type_arr != gguf_get_arr_type(other, idx_other)) {
+                ok = false;
+                continue;
+            }
+
+            if (type_arr == GGUF_TYPE_BOOL) {
+                const int8_t * data       = reinterpret_cast(gguf_get_arr_data(ctx,   id));
+                const int8_t * data_other = reinterpret_cast(gguf_get_arr_data(other, idx_other));
+                for (int arr_i = 0; arr_i < arr_n; ++arr_i) {
+                    if (bool(data[arr_i]) != bool(data_other[arr_i])) {
+                        ok = false;
+                    }
+                }
+                continue;
+            }
+
+            if (type_arr == GGUF_TYPE_STRING) {
+                for (int arr_i = 0; arr_i < arr_n; ++arr_i) {
+                    const std::string str       = gguf_get_arr_str(ctx,   id,       arr_i);
+                    const std::string str_other = gguf_get_arr_str(other, idx_other, arr_i);
+                    if (str != str_other) {
+                        ok = false;
+                    }
+                }
+                continue;
+            }
+
+            const int8_t * data       = reinterpret_cast(gguf_get_arr_data(ctx,   id));
+            const int8_t * data_other = reinterpret_cast(gguf_get_arr_data(other, idx_other));
+            if (!std::equal(data, data + arr_n*gguf_type_size(type_arr), data_other)) {
+                ok = false;
+            }
+            continue;
+        }
+
+        if (type == GGUF_TYPE_STRING) {
+            const std::string str       = gguf_get_val_str(ctx,   id);
+            const std::string str_other = gguf_get_val_str(other, idx_other);
+            if (str != str_other) {
+                ok = false;
+            }
+            continue;
+        }
+
+        const char * data       = reinterpret_cast(gguf_get_val_data(ctx,   id));
+        const char * data_other = reinterpret_cast(gguf_get_val_data(other, idx_other));
+        if (!std::equal(data, data + gguf_type_size(type), data_other)) {
+            ok = false;
+        }
+    }
+
+    return ok;
+}
+
+static bool all_tensors_in_other(const gguf_context * ctx, const gguf_context * other) {
+    bool ok = true;
+
+    const int n_tensors = gguf_get_n_tensors(ctx);
+    for (int id = 0; id < n_tensors; ++id) {
+        const std::string name = gguf_get_tensor_name(ctx, id);
+
+        const int idx_other = gguf_find_tensor(other, name.c_str());
+        if (id != idx_other) {
+            ok = false;
+            if (idx_other < 0) {
+                continue;
+            }
+        }
+
+        const ggml_type type = gguf_get_tensor_type(ctx, id);
+        if (type != gguf_get_tensor_type(other, id)) {
+            ok = false;
+        }
+
+        const size_t offset = gguf_get_tensor_offset(ctx, id);
+        if (offset != gguf_get_tensor_offset(other, id)) {
+            ok = false;
+        }
+    }
+
+    return ok;
+}
+
+static bool same_tensor_data(const struct ggml_context * orig, const struct ggml_context * read) {
+    bool ok = true;
+
+    struct ggml_tensor * t_orig = ggml_get_first_tensor(orig);
+    struct ggml_tensor * t_read = ggml_get_first_tensor(read);
+    while (t_orig) {
+        if (!t_read) {
+            ok = false;
+            break;
+        }
+
+        const size_t nbytes = ggml_nbytes(t_orig);
+        if (ggml_nbytes(t_read) != nbytes) {
+            ok = false;
+            break;
+        }
+        std::vector data_orig(nbytes);
+        ggml_backend_tensor_get(t_orig, data_orig.data(), 0, nbytes);
+        if (!std::equal(data_orig.data(), data_orig.data() + nbytes, reinterpret_cast(t_read->data))) {
+            ok = false;
+        }
+
+        t_orig = ggml_get_next_tensor(orig, t_orig);
+        t_read = ggml_get_next_tensor(orig, t_read);
+    }
+    if (t_read) {
+        ok = false;
+    }
+
+    return true;
+}
+
+static std::pair test_roundtrip(ggml_backend_dev_t dev, const unsigned int seed, const bool only_meta) {
+    ggml_backend_t backend = ggml_backend_dev_init(dev, nullptr);
+    printf("%s: device=%s, backend=%s, only_meta=%s\n",
+        __func__, ggml_backend_dev_description(dev), ggml_backend_name(backend), only_meta ? "yes" : "no");
+
+    int npass = 0;
+    int ntest = 0;
+
+    struct gguf_context * gguf_ctx_0;
+    struct ggml_context * ctx_0;
+    ggml_backend_buffer_t bbuf;
+    {
+        struct random_gguf_context_result result = get_random_gguf_context(backend, seed);
+        gguf_ctx_0 = result.gguf_ctx;
+        ctx_0      = result.ctx;
+        bbuf       = result.buffer;
+    }
+
+    FILE * file = tmpfile();
+
+#ifdef _WIN32
+    if (!file) {
+        printf("%s: failed to create tmpfile(), needs elevated privileges on Windows");
+        printf("%s: skipping tests");
+        return std::make_pair(0, 0);
+    }
+#else
+    GGML_ASSERT(file);
+#endif // _WIN32
+
+    {
+        std::vector buf;
+        gguf_write_to_buf(gguf_ctx_0, buf, only_meta);
+        GGML_ASSERT(fwrite(buf.data(), 1, buf.size(), file) == buf.size());
+        rewind(file);
+    }
+
+    struct ggml_context * ctx_1 = nullptr;
+    struct gguf_init_params gguf_params = {
+        /*no_alloc =*/ false,
+        /*ctx      =*/ only_meta ? nullptr : &ctx_1,
+    };
+    struct gguf_context * gguf_ctx_1 = gguf_init_from_file_impl(file, gguf_params);
+
+    printf("%s: same_version: ", __func__);
+    if (gguf_get_version(gguf_ctx_0) == gguf_get_version(gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: same_n_kv: ", __func__);
+    if (gguf_get_n_kv(gguf_ctx_0) == gguf_get_n_kv(gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: same_n_tensors: ", __func__);
+    if (gguf_get_n_tensors(gguf_ctx_0) == gguf_get_n_tensors(gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_orig_kv_in_read: ", __func__);
+    if (all_kv_in_other(gguf_ctx_0, gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_read_kv_in_orig: ", __func__);
+    if (all_kv_in_other(gguf_ctx_1, gguf_ctx_0)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_orig_tensors_in_read: ", __func__);
+    if (all_tensors_in_other(gguf_ctx_0, gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_read_tensors_in_orig: ", __func__);
+    if (all_tensors_in_other(gguf_ctx_1, gguf_ctx_0)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    if (!only_meta) {
+        printf("%s: same_tensor_data: ", __func__);
+        if (same_tensor_data(ctx_0, ctx_1)) {
+            printf("\033[1;32mOK\033[0m\n");
+            npass++;
+        } else {
+            printf("\033[1;31mFAIL\033[0m\n");
+        }
+        ntest++;
+    }
+
+    ggml_backend_buffer_free(bbuf);
+    ggml_free(ctx_0);
+    ggml_free(ctx_1);
+    gguf_free(gguf_ctx_0);
+    gguf_free(gguf_ctx_1);
+    ggml_backend_free(backend);
+    fclose(file);
+
+    printf("\n");
+    return std::make_pair(npass, ntest);
+}
+
+static std::pair test_gguf_set_kv(ggml_backend_dev_t dev, const unsigned int seed) {
+    ggml_backend_t backend = ggml_backend_dev_init(dev, nullptr);
+    printf("%s: device=%s, backend=%s\n", __func__, ggml_backend_dev_description(dev), ggml_backend_name(backend));
+
+    int npass = 0;
+    int ntest = 0;
+
+    struct gguf_context * gguf_ctx_0;
+    struct ggml_context * ctx_0;
+    ggml_backend_buffer_t bbuf_0;
+    {
+        struct random_gguf_context_result result = get_random_gguf_context(backend, seed);
+        gguf_ctx_0 = result.gguf_ctx;
+        ctx_0      = result.ctx;
+        bbuf_0     = result.buffer;
+    }
+
+    struct gguf_context * gguf_ctx_1;
+    struct ggml_context * ctx_1;
+    ggml_backend_buffer_t bbuf_1;
+    {
+        struct random_gguf_context_result result = get_random_gguf_context(backend, seed + 1);
+        gguf_ctx_1 = result.gguf_ctx;
+        ctx_1      = result.ctx;
+        bbuf_1     = result.buffer;
+    }
+
+    struct gguf_context * gguf_ctx_2 = gguf_init_empty();
+
+    gguf_set_kv(gguf_ctx_1, gguf_ctx_0);
+    gguf_set_kv(gguf_ctx_2, gguf_ctx_0);
+
+    printf("%s: same_n_kv: ", __func__);
+    if (gguf_get_n_kv(gguf_ctx_0) == gguf_get_n_kv(gguf_ctx_2)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_kv_0_in_1: ", __func__);
+    if (all_kv_in_other(gguf_ctx_0, gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_kv_0_in_2: ", __func__);
+    if (all_kv_in_other(gguf_ctx_0, gguf_ctx_2)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    gguf_set_kv(gguf_ctx_0, gguf_ctx_1);
+
+    printf("%s: same_n_kv_after_double_copy: ", __func__);
+    if (gguf_get_n_kv(gguf_ctx_0) == gguf_get_n_kv(gguf_ctx_1)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    printf("%s: all_kv_1_in_0_after_double_copy: ", __func__);
+    if (all_kv_in_other(gguf_ctx_1, gguf_ctx_0)) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    ggml_backend_buffer_free(bbuf_0);
+    ggml_backend_buffer_free(bbuf_1);
+    ggml_free(ctx_0);
+    ggml_free(ctx_1);
+    gguf_free(gguf_ctx_0);
+    gguf_free(gguf_ctx_1);
+    gguf_free(gguf_ctx_2);
+    ggml_backend_free(backend);
+
+    printf("\n");
+    return std::make_pair(npass, ntest);
+}
+
+static void print_usage() {
+    printf("usage: test-gguf [seed]\n");
+    printf("  if no seed is unspecified then a random seed is used\n");
+}
+
+int main(int argc, char ** argv) {
+    if (argc > 2) {
+        print_usage();
+        return 1;
+    }
+
+    std::random_device rd;
+    const unsigned int seed = argc < 2 ? rd() : std::stoi(argv[1]);
+
+    // Initialize ggml backends early so the prints aren't interleaved with the test results:
+    ggml_backend_dev_count();
+    fprintf(stderr, "\n");
+
+    int npass = 0;
+    int ntest = 0;
+    {
+        std::pair result = test_handcrafted_file(seed);
+        npass += result.first;
+        ntest += result.second;
+    }
+
+    for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
+        ggml_backend_dev_t dev = ggml_backend_dev_get(i);
+
+        for (bool only_meta : {true, false}) {
+            std::pair result = test_roundtrip(dev, seed, only_meta);
+            npass += result.first;
+            ntest += result.second;
+        }
+
+        {
+            std::pair result = test_gguf_set_kv(dev, seed);
+            npass += result.first;
+            ntest += result.second;
+        }
+    }
+
+    printf("%d/%d tests passed\n", npass, ntest);
+    if (npass != ntest) {
+        printf("\033[1;31mFAIL\033[0m\n");
+        return 1;
+    }
+    printf("\033[1;32mOK\033[0m\n");
+    return 0;
+}
diff --git a/tests/test-grad0.cpp b/tests/test-grad0.cpp
deleted file mode 100644
index c712dba7f..000000000
--- a/tests/test-grad0.cpp
+++ /dev/null
@@ -1,1684 +0,0 @@
-#define _CRT_SECURE_NO_DEPRECATE // Disables ridiculous "unsafe" warnings on Windows
-#include "ggml.h"
-#include "ggml-cpu.h"
-
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-#include 
-
-#if defined(_MSC_VER)
-#pragma warning(disable: 4244 4267) // possible loss of data
-#endif
-
-#if defined(__GNUC__)
-#pragma GCC diagnostic ignored "-Wdouble-promotion"
-#endif
-
-#define MAX_NARGS 3
-
-#undef MIN
-#undef MAX
-#define MIN(a, b) ((a) < (b) ? (a) : (b))
-#define MAX(a, b) ((a) > (b) ? (a) : (b))
-
-#define GGML_SILU_FP16
-
-//
-// logging
-//
-
-#if (GGML_DEBUG >= 1)
-#define GGML_PRINT_DEBUG(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG(...)
-#endif
-
-#if (GGML_DEBUG >= 5)
-#define GGML_PRINT_DEBUG_5(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_5(...)
-#endif
-
-#if (GGML_DEBUG >= 10)
-#define GGML_PRINT_DEBUG_10(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_10(...)
-#endif
-
-#define GGML_PRINT(...) printf(__VA_ARGS__)
-
-static float frand(void) {
-    return (float)rand()/(float)RAND_MAX;
-}
-
-static int irand(int n) {
-    if (n == 0) return 0;
-    return rand()%n;
-}
-
-static void get_random_dims(int64_t * dims, int ndims) {
-    dims[0] = dims[1] = dims[2] = dims[3] = 1;
-
-    for (int i = 0; i < ndims; i++) {
-        dims[i] = 1 + irand(4);
-    }
-}
-
-static struct ggml_tensor * get_random_tensor_f32(
-        struct ggml_context * ctx0,
-        int ndims,
-        int64_t ne[],
-        float fmin,
-        float fmax) {
-    struct ggml_tensor * result = ggml_new_tensor(ctx0, GGML_TYPE_F32, ndims, ne);
-
-    switch (ndims) {
-        case 1:
-            for (int i0 = 0; i0 < ne[0]; i0++) {
-                ((float *)result->data)[i0] = frand()*(fmax - fmin) + fmin;
-            }
-            break;
-        case 2:
-            for (int i1 = 0; i1 < ne[1]; i1++) {
-                for (int i0 = 0; i0 < ne[0]; i0++) {
-                    ((float *)result->data)[i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
-                }
-            }
-            break;
-        case 3:
-            for (int i2 = 0; i2 < ne[2]; i2++) {
-                for (int i1 = 0; i1 < ne[1]; i1++) {
-                    for (int i0 = 0; i0 < ne[0]; i0++) {
-                        ((float *)result->data)[i2*ne[1]*ne[0] + i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
-                    }
-                }
-            }
-            break;
-        case 4:
-            for (int i3 = 0; i3 < ne[3]; i3++) {
-                for (int i2 = 0; i2 < ne[2]; i2++) {
-                    for (int i1 = 0; i1 < ne[1]; i1++) {
-                        for (int i0 = 0; i0 < ne[0]; i0++) {
-                            ((float *)result->data)[i3*ne[2]*ne[1]*ne[0] + i2*ne[1]*ne[0] + i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
-                        }
-                    }
-                }
-            }
-            break;
-        default:
-            assert(false);
-    }
-
-    return result;
-}
-
-static struct ggml_tensor * get_random_tensor_f16(
-        struct ggml_context * ctx0,
-        int ndims,
-        int64_t ne[],
-        float fmin,
-        float fmax) {
-    struct ggml_tensor * result = ggml_new_tensor(ctx0, GGML_TYPE_F16, ndims, ne);
-
-    switch (ndims) {
-        case 1:
-            for (int i0 = 0; i0 < ne[0]; i0++) {
-                ((ggml_fp16_t *)result->data)[i0] = ggml_fp32_to_fp16(frand()*(fmax - fmin) + fmin);
-            }
-            break;
-        case 2:
-            for (int i1 = 0; i1 < ne[1]; i1++) {
-                for (int i0 = 0; i0 < ne[0]; i0++) {
-                    ((ggml_fp16_t *)result->data)[i1*ne[0] + i0] = ggml_fp32_to_fp16(frand()*(fmax - fmin) + fmin);
-                }
-            }
-            break;
-        case 3:
-            for (int i2 = 0; i2 < ne[2]; i2++) {
-                for (int i1 = 0; i1 < ne[1]; i1++) {
-                    for (int i0 = 0; i0 < ne[0]; i0++) {
-                        ((ggml_fp16_t *)result->data)[i2*ne[1]*ne[0] + i1*ne[0] + i0] = ggml_fp32_to_fp16(frand()*(fmax - fmin) + fmin);
-                    }
-                }
-            }
-            break;
-        case 4:
-            for (int i3 = 0; i3 < ne[3]; i3++) {
-                for (int i2 = 0; i2 < ne[2]; i2++) {
-                    for (int i1 = 0; i1 < ne[1]; i1++) {
-                        for (int i0 = 0; i0 < ne[0]; i0++) {
-                            ((ggml_fp16_t *)result->data)[i3*ne[2]*ne[1]*ne[0] + i2*ne[1]*ne[0] + i1*ne[0] + i0] = ggml_fp32_to_fp16(frand()*(fmax - fmin) + fmin);
-                        }
-                    }
-                }
-            }
-            break;
-        default:
-            assert(false);
-    }
-
-    return result;
-}
-
-static struct ggml_tensor * get_random_tensor_i32(
-        struct ggml_context * ctx0,
-        int ndims,
-        int64_t ne[],
-        int32_t imin,
-        int32_t imax) {
-    struct ggml_tensor * result = ggml_new_tensor(ctx0, GGML_TYPE_I32, ndims, ne);
-
-    switch (ndims) {
-        case 1:
-            for (int i0 = 0; i0 < ne[0]; i0++) {
-                ((int32_t *)result->data)[i0] = irand(imax - imin) + imin;
-            }
-            break;
-        case 2:
-            for (int i1 = 0; i1 < ne[1]; i1++) {
-                for (int i0 = 0; i0 < ne[0]; i0++) {
-                    ((int32_t *)result->data)[i1*ne[0] + i0] = irand(imax - imin) + imin;
-                }
-            }
-            break;
-        case 3:
-            for (int i2 = 0; i2 < ne[2]; i2++) {
-                for (int i1 = 0; i1 < ne[1]; i1++) {
-                    for (int i0 = 0; i0 < ne[0]; i0++) {
-                        ((int32_t *)result->data)[i2*ne[1]*ne[0] + i1*ne[0] + i0] = irand(imax - imin) + imin;
-                    }
-                }
-            }
-            break;
-        case 4:
-            for (int i3 = 0; i3 < ne[3]; i3++) {
-                for (int i2 = 0; i2 < ne[2]; i2++) {
-                    for (int i1 = 0; i1 < ne[1]; i1++) {
-                        for (int i0 = 0; i0 < ne[0]; i0++) {
-                            ((int32_t *)result->data)[i3*ne[2]*ne[1]*ne[0] + i2*ne[1]*ne[0] + i1*ne[0] + i0] = irand(imax - imin) + imin;
-                        }
-                    }
-                }
-            }
-            break;
-        default:
-            assert(false);
-    }
-
-    return result;
-}
-
-static bool check_gradient(
-        const char * op_name,
-        struct ggml_context * ctx0,
-        struct ggml_tensor * x[],
-        struct ggml_tensor * f,
-        int ndims,
-        int nargs,
-        float eps,
-        float max_error_abs,
-        float max_error_rel,
-        std::vector expected_vals) {
-
-    static int n_threads = -1;
-    if (n_threads < 0) {
-        n_threads = GGML_DEFAULT_N_THREADS;
-
-        const char *env = getenv("GGML_N_THREADS");
-        if (env) {
-            n_threads = atoi(env);
-        }
-
-        printf("GGML_N_THREADS = %d\n", n_threads);
-    }
-
-    struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, GGML_DEFAULT_GRAPH_SIZE, true);
-    struct ggml_cgraph * gb = ggml_new_graph_custom(ctx0, GGML_DEFAULT_GRAPH_SIZE, true);
-    ggml_build_forward_expand(gf, f);
-    ggml_graph_cpy(gf, gb);
-    ggml_build_backward_expand(ctx0, gf, gb, false);
-
-    ggml_graph_compute_with_ctx(ctx0, gf, n_threads);
-
-    ggml_graph_reset(gb);
-    if (f->grad) {
-        ggml_set_f32(f->grad, 1.0f);
-    }
-
-    ggml_graph_compute_with_ctx(ctx0, gb, n_threads);
-
-    // ggml_graph_dump_dot(gf, NULL, "test-grad0-forward.dot");
-    // ggml_graph_dump_dot(gb, gf,  "test-grad0-backward.dot");
-
-    for (int i = 0; i < nargs; ++i) {
-        bool all_g0_bad = true;
-        const int nelements = ggml_nelements(x[i]);
-        for (int k = 0; k < nelements; ++k) {
-            // Calculate gradient numerically:
-            const float x0 = ggml_get_f32_1d(x[i], k);
-            const float xm = x0 - eps;
-            const float xp = x0 + eps;
-            ggml_set_f32_1d(x[i], k, xp);
-
-            ggml_graph_compute_with_ctx(ctx0, gf, n_threads);
-
-            const double f0 = ggml_get_f32_1d(f, 0);
-
-            ggml_set_f32_1d(x[i], k, xm);
-
-            ggml_graph_compute_with_ctx(ctx0, gf, n_threads);
-
-            const double f1 = ggml_get_f32_1d(f, 0);
-            const double g0 = (f0 - f1)/(2.0*(double) eps);
-
-            // The numerical calculation of the gradient fails around noncontinuities (e.g. 0 for ReLU).
-            // In such cases, provide a vector of expected values and skip the comparison for failed calculations.
-            if (!expected_vals.empty()) {
-                bool matches_any = false;
-                for (const double & ev : expected_vals) {
-                    const double error_abs = std::fabs(g0 - ev);
-                    if (error_abs > max_error_abs) {
-                        continue;
-                    }
-                    const double error_rel = g0 != 0.0 ? fabs(g0 - ev)/fabs(g0) : 0.0;
-                    if (error_rel > max_error_rel) {
-                        continue;
-                    }
-                    matches_any = true;
-                    break;
-                }
-                if (!matches_any) {
-                    continue;
-                }
-            }
-            all_g0_bad = false;
-
-            ggml_set_f32_1d(x[i], k, x0);
-
-            // compute gradient using backward graph
-            ggml_graph_reset(gb);
-            if (f->grad) {
-                ggml_set_f32(f->grad, 1.0f);
-            }
-
-            ggml_graph_compute_with_ctx(ctx0, gb, n_threads);
-
-            const double g1 = ggml_get_f32_1d(x[i]->grad, k);
-
-            const double error_abs = fabs(g0 - g1);
-            const double error_rel = g0 != 0.0 ? fabs(g0 - g1)/fabs(g0) : 0.0;
-
-            if (error_abs > max_error_abs || error_rel > max_error_rel) {
-                printf("%s: ndims=%d, i=%d, k=%d, x0=%f, xm=%f, xp=%f, f0=%f, f1=%f, g0=%f, g1=%f, eps=%f, error_abs=%f, error_rel=%f\n",
-                            op_name, ndims, i, k, x0, xm, xp, f0, f1, g0, g1, eps, error_abs, error_rel);
-                //assert(false);
-                return false;
-            }
-        }
-        if (all_g0_bad) {
-            printf("%s: numerical calculation of the gradient failed for all values\n", op_name);
-            return false;
-        }
-    }
-
-    return true;
-}
-
-// TODO: clean-up this ..
-static bool check_mat_mul(
-        const struct ggml_tensor * y,
-        const struct ggml_tensor * x0,
-        const struct ggml_tensor * x1) {
-    float * dst  = (float *) y->data;
-    float * src0 = (float *) x0->data;
-    float * src1 = (float *) x1->data;
-
-    const int nc = x0->ne[1];
-    const int nr = x1->ne[1];
-    const int nk = x0->ne[0];
-
-    GGML_PRINT_DEBUG("check_mat_mul: nc=%d, nr=%d, nk=%d\n", nc, nr, nk);
-
-    GGML_PRINT_DEBUG("x0:\n");
-    for (int j = 0; j < x0->ne[1]; ++j) {
-        for (int i = 0; i < x0->ne[0]; ++i) {
-            GGML_PRINT_DEBUG("%6.3f ", src0[j*nk + i]);
-        }
-        GGML_PRINT_DEBUG("\n");
-    }
-    GGML_PRINT_DEBUG("\n");
-
-    GGML_PRINT_DEBUG("x1:\n");
-    for (int j = 0; j < x1->ne[1]; ++j) {
-        for (int i = 0; i < x1->ne[0]; ++i) {
-            GGML_PRINT_DEBUG("%6.3f ", src1[j*nk + i]);
-        }
-        GGML_PRINT_DEBUG("\n");
-    }
-    GGML_PRINT_DEBUG("\n");
-
-    GGML_PRINT_DEBUG("y: n_dims = %d, (%lld, %lld)\n", y->n_dims, y->ne[0], y->ne[1]);
-    for (int j = 0; j < y->ne[1]; ++j) {
-        for (int i = 0; i < y->ne[0]; ++i) {
-            GGML_PRINT_DEBUG("%6.3f ", dst[j*nr + i]);
-        }
-        GGML_PRINT_DEBUG("\n");
-    }
-
-    for (int i = 0; i < nr; ++i) {
-        for (int j = 0; j < nc; ++j) {
-            float sum = 0.0f;
-
-            for (int k = 0; k < nk; ++k) {
-                sum += src0[j*nk + k]*src1[i*nk + k];
-            }
-
-            if (fabsf(dst[i*nc + j] - sum) > 1e-5f) {
-                fprintf(stderr, "check_mat_mul: dst[%d] = %f, sum = %f\n", i*nc + j, dst[i*nc + j], sum);
-                assert(false);
-                return false;
-            }
-        }
-    }
-
-    return true;
-}
-
-#define NUM_PERMUTATIONS (4*3*2*1)
-
-int main(int argc, const char ** argv) {
-    struct ggml_init_params params = {
-        /* .mem_size   = */ 256*1024*1024,
-        /* .mem_buffer = */ NULL,
-        /* .no_alloc   = */ false,
-    };
-
-    int64_t ne[4];
-
-    int all_permutations[4 * NUM_PERMUTATIONS];
-    {
-        int count = 0;
-        for (int ax0=0; ax0<4; ++ax0) {
-            for (int ax1=0; ax1<4; ++ax1) {
-                if (ax1 == ax0) continue;
-                for (int ax2=0; ax2<4; ++ax2) {
-                    if (ax2 == ax0) continue;
-                    if (ax2 == ax1) continue;
-                    for (int ax3=0; ax3<4; ++ax3) {
-                        if (ax3 == ax0) continue;
-                        if (ax3 == ax1) continue;
-                        if (ax3 == ax2) continue;
-                        assert(count < NUM_PERMUTATIONS);
-                        all_permutations[count*4+0] = ax0;
-                        all_permutations[count*4+1] = ax1;
-                        all_permutations[count*4+2] = ax2;
-                        all_permutations[count*4+3] = ax3;
-                        ++count;
-                    }
-                }
-            }
-        }
-    }
-
-    unsigned seed_iter = 1;
-
-    // original loop: 1000
-    int niter = 4;
-    const char *env = getenv("GGML_NLOOP");
-    if (env != NULL) {
-        niter = atoi(env);
-    }
-    if (argc > 1) {
-        niter = atoi(argv[1]);
-    }
-    for (int iter = 0; iter < niter; ++iter) {
-        srand(seed_iter);
-        seed_iter = rand();
-        unsigned seed = rand();
-
-        printf("test-grad0: iter:%d/%d\n", (iter+1), niter);
-        struct ggml_context * ctx0 = ggml_init(params);
-
-        get_random_dims(ne, 4);
-
-        struct ggml_tensor * x[MAX_NARGS];
-
-        // add f32
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_add(ctx0, x[0], x[1]));
-
-                check_gradient("add f32", ctx0, x, f, ndims, nargs, 1e-3f, 2e-3f, 2e-3f, {});
-            }
-        }
-
-        // add f16
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f16(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_add(ctx0, x[0], x[1]));
-
-                check_gradient("add f16", ctx0, x, f, ndims, nargs, 1e-1f, 2e-1f, 2e-1f, {});
-            }
-        }
-
-        // sub
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sub(ctx0, x[0], x[1]));
-
-                check_gradient("sub", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // mul
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_mul(ctx0, x[0], x[1]));
-
-                check_gradient("mul", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // div
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, 0.5f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_div(ctx0, x[0], x[1]));
-
-                check_gradient("div", ctx0, x, f, ndims, nargs, 1e-3f, 1e-1f, 1e-1f, {});
-            }
-        }
-
-        // sqr
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sqr(ctx0, x[0]));
-
-                check_gradient("sqr", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // sqrt
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, 2.0f*1e-3f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sqrt(ctx0, x[0]));
-
-                check_gradient("sqrt", ctx0, x, f, ndims, nargs, 1e-3f, 2e-2f, 1e-1f, {});
-            }
-        }
-
-        // log
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, 2.0f*1e-3f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_log(ctx0, x[0]));
-
-                check_gradient("log", ctx0, x, f, ndims, nargs, 1e-3f, INFINITY, 1e-1f, {});
-            }
-        }
-
-        // sum
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, x[0]);
-
-                check_gradient("sum", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-
-        // sum_rows
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sqr(ctx0, ggml_sum_rows(ctx0, x[0])));
-
-                check_gradient("sum_rows", ctx0, x, f, ndims, nargs, 1e-3f, 1e-2f, INFINITY, {});
-            }
-        }
-
-        // mean, not yet fully implemented
-        if(0)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_mean(ctx0, x[0]));
-
-                check_gradient("mean", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // argmax
-        if (0)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_argmax(ctx0, x[0]));
-
-                check_gradient("argmax", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // repeat
-        {
-            srand(seed);
-            int64_t ne2[4];
-            get_random_dims(ne2, 4);
-
-            ne2[0] = ne[0] * ne2[0];
-            ne2[1] = ne[1] * ne2[1];
-            ne2[2] = 1;
-            ne2[3] = 1;
-
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                x[1] = get_random_tensor_f32(ctx0, ndims, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sqr(ctx0, ggml_sub(ctx0, x[1], ggml_repeat(ctx0, x[0], x[1]))));
-
-                check_gradient("repeat", ctx0, x, f, ndims, nargs, 1e-3f, 1e-2f, INFINITY, {});
-            }
-        }
-
-        // repeat back
-        {
-            srand(seed);
-            int64_t ne2[4];
-            get_random_dims(ne2, 4);
-
-            ne2[0] = ne[0] * ne2[0];
-            ne2[1] = ne[1] * ne2[1];
-            ne2[2] = 1;
-            ne2[3] = 1;
-
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                x[1] = get_random_tensor_f32(ctx0, ndims, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_sqr(ctx0, ggml_sub(ctx0, x[0], ggml_repeat_back(ctx0, x[1], x[0]))));
-
-                check_gradient("repeat back", ctx0, x, f, ndims, nargs, 1e-3f, 1e-2f, INFINITY, {});
-            }
-        }
-
-        // abs
-        {
-           const int nargs = 1;
-
-           for (int ndims = 1; ndims <= 4; ++ndims) {
-               for (int i = 0; i < nargs; ++i) {
-                   x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                   ggml_set_param(ctx0, x[i]);
-               }
-
-               struct ggml_tensor * f = ggml_sum(ctx0, ggml_abs(ctx0, x[0]));
-
-               check_gradient("abs", ctx0, x, f, ndims, nargs, 1e-3f, INFINITY, 1e-3f, {-1.0, 1.0});
-           }
-        }
-
-        // sgn
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_sgn(ctx0, x[0]));
-
-                check_gradient("sgn", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {0.0});
-            }
-        }
-
-        // neg
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_neg(ctx0, x[0]));
-
-                check_gradient("neg", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // step
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_step(ctx0, x[0]));
-
-                check_gradient("step", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {0.0});
-            }
-        }
-
-        // tanh, not yet fully implemented
-        if(0)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_tanh(ctx0, x[0]));
-
-                check_gradient("tanh", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // mul_mat
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 2; ndims <= 4; ++ndims) {
-                int max_nrep = (ndims >= 3) ? 2 : 1;
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                for (int nrep2 = 1; nrep2 < max_nrep; ++nrep2) {
-                    for (int nrep3 = 1; nrep3 < max_nrep; ++nrep3) {
-                        {
-                            int64_t ne2[4];
-                            get_random_dims(ne2, 4);
-                            ne2[0] = ne[0];
-                            ne2[2] = nrep2 * ne[2];
-                            ne2[3] = nrep3 * ne[3];
-                            x[1] = get_random_tensor_f32(ctx0, ndims, ne2, -1.0f, 1.0f);
-                        }
-
-                        ggml_set_param(ctx0, x[0]);
-                        ggml_set_param(ctx0, x[1]);
-
-                        struct ggml_tensor * m = ggml_mul_mat(ctx0, x[1], x[0]);
-                        struct ggml_tensor * f = ggml_sum(ctx0, m);
-
-                        GGML_PRINT_DEBUG("testing: mul_mat, [%lld, %lld] (%d) * [%lld, %lld] (%d)\n", x[1]->ne[0], x[1]->ne[1], x[1]->n_dims, x[0]->ne[0], x[0]->ne[1], x[0]->n_dims);
-
-                        check_gradient("mul_mat", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-                        if (ndims == 2) {
-                            // check_mat_mul does not support ndims > 2
-                            check_mat_mul(m, x[1], x[0]);
-                        }
-                    }
-                }
-            }
-        }
-
-        // elu, not yet fully implemented
-        if(0)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_elu(ctx0, x[0]));
-
-                check_gradient("elu", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // relu
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_relu(ctx0, x[0]));
-
-                check_gradient("relu", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {0.0, 1.0});
-            }
-        }
-
-        // gelu, not yet fully implemented
-        if(0)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor* f = ggml_sum(ctx0, ggml_gelu(ctx0, x[0]));
-
-                check_gradient("gelu", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, 1e-3f, {});
-            }
-        }
-
-        // silu
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_silu(ctx0, x[0]));
-
-#ifdef GGML_SILU_FP16
-                // due to GGML_SILU_FP16 the finite difference method will be slightly wrong -> increase error bounds.
-                check_gradient("silu", ctx0, x, f, ndims, nargs, 1e-3f, 0.5, INFINITY, {});
-#else
-                check_gradient("silu", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-#endif
-            }
-        }
-
-        // rms_norm
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_rms_norm(ctx0, x[0], 1e-6f));
-
-                check_gradient("rms_norm", ctx0, x, f, ndims, nargs, 1e-4f, 1.0f, INFINITY, {});
-            }
-        }
-
-        // scale
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-
-                const float s = -1.0f + 2.0f*frand();
-
-                ggml_set_param(ctx0, x[0]);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_scale(ctx0, x[0], s));
-
-                check_gradient("scale", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // cpy f32
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-                // x[1] is overwritten by x[0], so the gradients don't propagate to x[1]
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_cpy(ctx0, x[0], x[1]));
-
-                check_gradient("cpy f32", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // cpy f16
-        {
-            srand(seed);
-            const int nargs = 2;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                for (int i = 0; i < nargs; ++i) {
-                    x[i] = get_random_tensor_f16(ctx0, ndims, ne, -1.0f, 1.0f);
-                    ggml_set_param(ctx0, x[i]);
-                }
-                // x[1] is overwritten by x[0], so the gradients don't propagate to x[1]
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_cpy(ctx0, x[0], x[1]));
-
-                check_gradient("cpy f16", ctx0, x, f, ndims, nargs, 1e-1f, 1e-1f, INFINITY, {});
-            }
-        }
-
-        // reshape (1d->nd)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                int64_t ne2[4];
-                ne2[0] = 1;
-                ne2[1] = 1;
-                ne2[2] = 1;
-                ne2[3] = 1;
-                for (int i = 0; i < ndims; ++i) {
-                    ne2[0] *= ne[i];
-                }
-                x[0] = get_random_tensor_f32(ctx0, 1, ne2, -1.0f, 1.0f);
-                x[1] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_reshape(ctx0, x[0], x[1]));
-                check_gradient("reshape", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // reshape (nd->1d)
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            for (int ndims = 1; ndims <= 2; ++ndims) {
-                int64_t ne2[4];
-                ne2[0] = 1;
-                ne2[1] = 1;
-                ne2[2] = 1;
-                ne2[3] = 1;
-                for (int i = 0; i < ndims; ++i) {
-                    ne2[0] *= ne[i];
-                }
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                x[1] = get_random_tensor_f32(ctx0, 1, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_reshape(ctx0, x[0], x[1]));
-                check_gradient("reshape", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // acc 1d
-        {
-            srand(seed);
-            int64_t ne2[4] = { 1, 1, 1, 1 };
-
-            const int nargs = 2;
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 1);
-                while ((ne2[0] > ne[0]) || (ne2[0] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 1);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 1, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                const int max_offset = MAX(0, ggml_nelements(x[0]) - ggml_nelements(x[1]));
-                const int offset = irand(max_offset) * ggml_element_size(x[0]);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_acc(ctx0, x[0], x[1], x[0]->nb[1], x[0]->nb[2], x[0]->nb[3], offset));
-
-                check_gradient("acc 1d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // acc 2d
-        {
-            srand(seed);
-            int64_t ne2[4]         = { 1, 1, 1, 1 };
-            int64_t max_offsets[4] = { 0, 0, 0, 0 };
-            int64_t offsets[4]     = { 0, 0, 0, 0 };
-
-            const int nargs = 2;
-            for (int ndims = 2; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 2);
-                while ((ne2[0] > ne[0]) || (ne2[1] > ne[1]) || (ne2[0]*ne2[1] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 2);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 2, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                max_offsets[0] = MAX(0, x[0]->ne[0] - x[1]->ne[0]);
-                max_offsets[1] = MAX(0, x[0]->ne[1] - x[1]->ne[1]);
-                offsets[0] = irand(max_offsets[0]) * x[0]->nb[0];
-                offsets[1] = irand(max_offsets[1]) * x[0]->nb[1];
-                const int offset = offsets[0] + offsets[1];
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_acc(ctx0, x[0], x[1], x[0]->nb[1], x[0]->nb[2], x[0]->nb[3], offset));
-
-                check_gradient("acc 2d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // acc 3d
-        {
-            srand(seed);
-            int64_t ne2[4]         = { 1, 1, 1, 1 };
-            int64_t max_offsets[4] = { 0, 0, 0, 0 };
-            int64_t offsets[4]     = { 0, 0, 0, 0 };
-
-            const int nargs = 2;
-            for (int ndims = 3; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 3);
-                while ((ne2[0] > ne[0]) || (ne2[1] > ne[1]) || (ne2[2] > ne[2]) || (ne2[0]*ne2[1]*ne2[2] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 3);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 3, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                max_offsets[0] = MAX(0, x[0]->ne[0] - x[1]->ne[0]);
-                max_offsets[1] = MAX(0, x[0]->ne[1] - x[1]->ne[1]);
-                max_offsets[2] = MAX(0, x[0]->ne[2] - x[1]->ne[2]);
-                offsets[0] = irand(max_offsets[0]) * x[0]->nb[0];
-                offsets[1] = irand(max_offsets[1]) * x[0]->nb[1];
-                offsets[2] = irand(max_offsets[2]) * x[0]->nb[2];
-                const int offset = offsets[0] + offsets[1] + offsets[2];
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_acc(ctx0, x[0], x[1], x[0]->nb[1], x[0]->nb[2], x[0]->nb[3], offset));
-
-                check_gradient("acc 3d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // acc 4d
-        {
-            srand(seed);
-            int64_t ne2[4]         = { 1, 1, 1, 1 };
-            int64_t max_offsets[4] = { 0, 0, 0, 0 };
-            int64_t offsets[4]     = { 0, 0, 0, 0 };
-
-            const int nargs = 2;
-            for (int ndims = 4; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 4);
-                while ((ne2[0] > ne[0]) || (ne2[1] > ne[1]) || (ne2[2] > ne[2]) || (ne2[3] > ne[3]) || (ne2[0]*ne2[1]*ne2[2]*ne2[3] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 4);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 4, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                max_offsets[0] = MAX(0, x[0]->ne[0] - x[1]->ne[0]);
-                max_offsets[1] = MAX(0, x[0]->ne[1] - x[1]->ne[1]);
-                max_offsets[2] = MAX(0, x[0]->ne[2] - x[1]->ne[2]);
-                max_offsets[3] = MAX(0, x[0]->ne[3] - x[1]->ne[3]);
-                offsets[0] = irand(max_offsets[0]) * x[0]->nb[0];
-                offsets[1] = irand(max_offsets[1]) * x[0]->nb[1];
-                offsets[2] = irand(max_offsets[2]) * x[0]->nb[2];
-                offsets[3] = irand(max_offsets[3]) * x[0]->nb[3];
-                const int offset = offsets[0] + offsets[1] + offsets[2] + offsets[3];
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_acc(ctx0, x[0], x[1], x[0]->nb[1], x[0]->nb[2], x[0]->nb[3], offset));
-
-                check_gradient("acc 4d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // set_1d
-        {
-            srand(seed);
-            int64_t ne2[4];
-
-            const int nargs = 2;
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 1);
-                while ((ne2[0] > ne[0]) || (ne2[0] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 1);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 1, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                const int max_offset = MAX(0, ggml_nelements(x[0]) - ggml_nelements(x[1]));
-                const int offset = irand(max_offset) * ggml_element_size(x[0]);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_set_1d(ctx0, x[0], x[1], offset));
-
-                check_gradient("set_1d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // set_2d
-        {
-            srand(seed);
-            int64_t ne2[4];
-            int64_t max_offsets[4] = { 0, 0, 0, 0 };
-            int64_t offsets[4]     = { 0, 0, 0, 0 };
-
-            const int nargs = 1;
-            for (int ndims = 2; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[0]);
-
-                get_random_dims(ne2, 2);
-                while ((ne2[0] > ne[0]) || (ne2[1] > ne[1]) || (ne2[0]*ne2[1] > ggml_nelements(x[0]))) {
-                    get_random_dims(ne2, 2);
-                }
-
-                x[1] = get_random_tensor_f32(ctx0, 2, ne2, -1.0f, 1.0f);
-                ggml_set_param(ctx0, x[1]);
-
-                max_offsets[0] = MAX(0, x[0]->ne[0] - x[1]->ne[0]);
-                max_offsets[1] = MAX(0, x[0]->ne[1] - x[1]->ne[1]);
-                offsets[0] = irand(max_offsets[0]) * x[0]->nb[0];
-                offsets[1] = irand(max_offsets[1]) * x[0]->nb[1];
-                const int offset = offsets[0] + offsets[1];
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_set_2d(ctx0, x[0], x[1], x[1]->nb[1], offset));
-
-                check_gradient("set_2d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // view_1d
-        {
-            srand(seed);
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-
-                ggml_set_param(ctx0, x[0]);
-
-                const int k0 = irand(ggml_nelements(x[0]));
-                const int k1 = irand(ggml_nelements(x[0]));
-                const int i0 = MIN(k0, k1);
-                const int i1 = MAX(k0, k1);
-
-                const int offset = i0 * sizeof(float);
-                const int nelem  = i1 - i0;
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_view_1d(ctx0, x[0], nelem, offset));
-
-                check_gradient("view_1d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // view_2d
-        {
-            srand(seed);
-            int64_t ne2[4];
-            int64_t nb2[4];
-
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-
-                get_random_dims(ne2, 2);
-                while (ne2[0]*ne2[1] > ggml_nelements(x[0])) {
-                    get_random_dims(ne2, 2);
-                }
-                const int count = ne2[0]*ne2[1];
-
-                nb2[0] = sizeof(float);
-                nb2[1] = nb2[0]*ne2[0];
-
-                ggml_set_param(ctx0, x[0]);
-
-                const int max_offset = ggml_nelements(x[0]) - count;
-                const int offset = irand(max_offset+1) * sizeof(float);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_view_2d(ctx0, x[0], ne2[0], ne2[1], nb2[1], offset));
-
-                check_gradient("view_2d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // view_3d
-        {
-            srand(seed);
-            int64_t ne2[4] = {1,1,1,1};
-            int64_t nb2[4] = {0,0,0,0};
-
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
-
-                get_random_dims(ne2, 3);
-                while (ne2[0]*ne2[1]*ne2[2] > ggml_nelements(x[0])) {
-                    get_random_dims(ne2, 3);
-                }
-                const int count = ne2[0]*ne2[1]*ne2[2];
-
-                nb2[0] = sizeof(float);
-                nb2[1] = nb2[0]*ne2[0];
-                nb2[2] = nb2[1]*ne2[1];
-
-                ggml_set_param(ctx0, x[0]);
-
-                const int max_offset = ggml_nelements(x[0]) - count;
-                const int offset = irand(max_offset+1) * sizeof(float);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_view_3d(ctx0, x[0], ne2[0], ne2[1], ne2[2], nb2[1], nb2[2], offset));
-
-                check_gradient("view_3d", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // permute
-        {
-            srand(seed);
-            int64_t ne2[4];
-
-            const int nargs = 1;
-            for (int ndims = 1; ndims <= 4; ++ndims)
-            {
-                // ggml_permute will set axes of dimensions below n_dims to 1.
-                // to make ggml_permute work correctly on all axes,
-                // the input tensor needs maximal n_dim of 4.
-                for (int i=0; i finite differences should not work
-                // instead use sum(log(soft_max()*(1-eps)+eps)); use eps to avoid log(0)
-                struct ggml_tensor * f = ggml_sum(ctx0,
-                                            ggml_log(ctx0,
-                                                ggml_add1(ctx0,
-                                                    ggml_scale(ctx0,
-                                                        ggml_soft_max(ctx0, x[0]),
-                                                        1.0f - eps),
-                                                    ggml_new_f32(ctx0, eps))));
-
-                check_gradient("softmax", ctx0, x, f, ndims, nargs, 1e-3f, 2e-1f, INFINITY, {});
-                // NOTE: softmax forward is computed using f16 table lookup instead of using actual expf, but backward assumes actual expf.
-                // this may result in different gradients too finite differences.
-                // when this test reports errors, first try to replace the table lookup with actual expf and test again to see if just that was the cause.
-                // if only the table lookup causes gradients to differ this is acceptable.
-            }
-        }
-
-        // cross_entropy_loss
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            int64_t ne2[4];
-            get_random_dims(ne2, 4);
-
-            for (int ndims = 1; ndims <= 4; ++ndims) {
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne2, -1.0f, 1.0f);
-                x[1] = get_random_tensor_f32(ctx0, ndims, ne2, 0.0f, 1.0f);
-                // the second argument to cross_entropy_loss must sum up to 1 for each row
-                int nr = ggml_nrows(x[1]);
-                int nc = ggml_nelements(x[1]) / nr;
-                for (int ir = 0; ir < nr; ++ir) {
-                    float sum = 0;
-                    for (int ic = 0; ic < nc; ++ic) {
-                        sum += ((float *) x[1]->data)[ic + ir*nc];
-                    }
-                    for (int ic = 0; ic < nc; ++ic) {
-                        ((float *) x[1]->data)[ic + ir*nc] /= sum;
-                    }
-                }
-                ggml_set_param(ctx0, x[0]);
-
-                struct ggml_tensor * f = ggml_cross_entropy_loss(ctx0, x[0], x[1]);
-
-                check_gradient("cross_entropy_loss", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // rope f32
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            int64_t ne2[4];
-            get_random_dims(ne2, 4);
-            ne2[0] += ne2[0] % 2;
-            int n_rot = ne2[0];
-
-            for (int ndims = 3; ndims <= 4; ++ndims) {
-                for (int mode = 0; mode < 4; ++mode) {
-                    for (int n_past = 1; n_past < ne2[2]; ++n_past) {
-                        x[0] = get_random_tensor_f32(ctx0, ndims, ne2, -1.0f, 1.0f);
-
-                        struct ggml_tensor * p = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne2[2]);
-                        for (int i = 0; i < ne2[2]; ++i) {
-                            ((int32_t *) p->data)[i] = n_past + i;
-                        }
-
-                        ggml_set_param(ctx0, x[0]);
-
-                        const bool skip_past = (mode & 1);
-                        if (skip_past) {
-                            // we have no past, so this would have to work on uninitialized memory.
-                            // we only test the gradients here;
-                            // skip_past should have no influence on gradient computation.
-                            // so when other modes work, we assume that this does as well.
-                            continue;
-                        }
-
-                        struct ggml_tensor * f = ggml_sum(ctx0, ggml_rope(ctx0, x[0], p, n_rot, mode));
-
-                        GGML_PRINT_DEBUG("rope f32: n_past: %d n_rot: %d mode: %d\n", n_past, n_rot, mode);
-                        check_gradient("rope f32", ctx0, x, f, ndims, nargs, 1e-2f, 1e-3f, INFINITY, {});
-                    }
-                }
-            }
-        }
-
-        // rope f16
-        {
-            srand(seed);
-            const int nargs = 1;
-
-            int64_t ne2[4];
-            get_random_dims(ne2, 4);
-            ne2[0] += ne2[0] % 2;
-            int n_rot = ne2[0];
-
-            for (int ndims = 3; ndims <= 4; ++ndims) {
-                for (int mode = 0; mode < 4; ++mode) {
-                    for (int n_past = 1; n_past < ne2[2]; ++n_past) {
-                        x[0] = get_random_tensor_f16(ctx0, ndims, ne2, -1.0f, 1.0f);
-
-                        struct ggml_tensor * p = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne2[2]);
-                        for (int i = 0; i < ne2[2]; ++i) {
-                            ((int32_t *) p->data)[i] = n_past + i;
-                        }
-
-                        ggml_set_param(ctx0, x[0]);
-
-                        const bool skip_past = (mode & 1);
-                        if (skip_past) {
-                            // we have no past, so this would have to work on uninitialized memory.
-                            // we only test the gradients here;
-                            // skip_past should have no influence on gradient computation.
-                            // so when other modes work, we assume that this does as well.
-                            continue;
-                        }
-
-                        struct ggml_tensor * f = ggml_sum(ctx0, ggml_rope(ctx0, x[0], p, n_rot, mode));
-
-                        GGML_PRINT_DEBUG("rope f16: n_past: %d n_rot: %d mode: %d\n", n_past, n_rot, mode);
-                        check_gradient("rope f16", ctx0, x, f, ndims, nargs, 1e-1f, 1e-1f, INFINITY, {});
-                    }
-                }
-            }
-        }
-
-        // im2col f32
-        {
-            srand(seed);
-            const int nargs = 1;
-            const int ndims = 4;
-
-            for (const bool is_2D : {false, true}) {
-                int64_t ne0[ndims];
-                int64_t ne1[ndims];
-                get_random_dims(ne0, ndims);
-                get_random_dims(ne1, ndims);
-
-                // // Ensure that the output is not zero-sized:
-                ne1[0] += 8;
-                ne1[1] += 8;
-
-                if (is_2D) {
-                    ne1[2] = ne0[2];
-                } else {
-                    ne1[1] = ne0[1];
-                    ne0[3] = 1;
-                    ne1[3] = 1;
-                }
-
-                // The order of arguments is swapped because the first tensor is only used for its shape.
-                x[1] = get_random_tensor_f16(ctx0, ndims, ne0, -1.0f, 1.0f);
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne1, -1.0f, 1.0f);
-
-                ggml_set_param(ctx0, x[0]);
-
-                const int s0 =         1 + irand(2);
-                const int s1 = is_2D ? 1 + irand(2) : 0;
-                const int p0 =         0 + irand(2);
-                const int p1 = is_2D ? 0 + irand(2) : 0;
-                const int d0 =         1 + irand(2);
-                const int d1 = is_2D ? 1 + irand(2) : 0;
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_im2col(ctx0, x[1], x[0], s0, s1, p0, p1, d0, d1, is_2D, GGML_TYPE_F32));
-
-                GGML_PRINT_DEBUG("im2col f32: is_2D=%s, s0=%d, s1=%d, p0=%d, p1=%d, d0=%d, d1=%d\n", is_2D ? "yes" : "no", s0, s1, p0, p1, d0, d1);
-                check_gradient("im2col f32", ctx0, x, f, ndims, nargs, 1e-2f, 1e-3f, INFINITY, {});
-            }
-        }
-
-        // pool_2d f32
-        {
-            srand(seed);
-            const int nargs = 1;
-            const int ndims = 4;
-
-            for (const enum ggml_op_pool op : {GGML_OP_POOL_AVG, GGML_OP_POOL_MAX}) {
-                int64_t ne0[ndims];
-                get_random_dims(ne0, ndims);
-
-                ne0[0] += 8;
-                ne0[1] += 8;
-
-                x[0] = get_random_tensor_f32(ctx0, ndims, ne0, -1.0f, 1.0f);
-
-                ggml_set_param(ctx0, x[0]);
-
-                const int k0 = 2 + irand(2);
-                const int k1 = 2 + irand(2);
-                const int s0 = 2 + irand(2);
-                const int s1 = 2 + irand(2);
-                const int p0 = 0 + irand(2);
-                const int p1 = 0 + irand(2);
-
-                struct ggml_tensor * f = ggml_sum(ctx0, ggml_pool_2d(ctx0, x[0], op, k0, k1, s0, s1, p0, p1));
-
-                GGML_PRINT_DEBUG("ggml_pool_2d f32: op=%s k0=%d, k1=%d, s0=%d, s1=%d, p0=%d, p1=%d\n",
-                                 op == GGML_OP_POOL_MAX ? "max" : "avg", k0, k1, s0, s1, p0, p1);
-                std::vector expected_vals;
-                if (op == GGML_OP_POOL_MAX) {
-                    expected_vals.push_back(0.0);
-                    expected_vals.push_back(1.0);
-                }
-                check_gradient("ggml_pool_2d f32", ctx0, x, f, ndims, nargs, 1e-3f, 1e-3f, INFINITY, expected_vals);
-            }
-        }
-
-        // flash_attn f32
-        // TODO: adapt to ggml_flash_attn_ext() changes
-        //{
-        //    srand(seed);
-        //    const int nargs = 3;
-
-        //    int64_t ne2[4];
-
-        //    get_random_dims(ne2, 4);
-        //    int64_t D = ne2[0];
-        //    int64_t N = ne2[1];
-        //    int64_t M = ne2[2] + N;
-        //    int64_t B = ne2[3];
-
-        //    for (int masked = 0; masked <= 1; ++masked) {
-        //        for (int ndims = 2; ndims <= 4; ++ndims) {
-        //            int max_nrep = (ndims >= 3) ? 2 : 1;
-        //            for (int nrep = 1; nrep < max_nrep; ++nrep) {
-        //                int64_t neq[4] = { D, N, B*nrep, ne[3] };
-        //                int64_t nek[4] = { D, M, B, ne[3] };
-        //                int64_t nev[4] = { M, D, B, ne[3] };
-        //                if (ndims == 2) {
-        //                    neq[2] = 1; neq[3] = 1;
-        //                    nek[2] = 1; nek[3] = 1;
-        //                    nev[2] = 1; nev[3] = 1;
-        //                } else if (ndims == 3) {
-        //                    neq[3] = 1;
-        //                    nek[3] = 1;
-        //                    nev[3] = 1;
-        //                }
-        //                x[0] = get_random_tensor_f32(ctx0, ndims, neq, -0.1250f, 0.1250f);
-        //                x[1] = get_random_tensor_f32(ctx0, ndims, nek, -0.1250f, 0.1250f);
-        //                x[2] = get_random_tensor_f32(ctx0, ndims, nev, -0.1250f, 0.1250f);
-        //                ggml_set_param(ctx0, x[0]);
-        //                ggml_set_param(ctx0, x[1]);
-        //                ggml_set_param(ctx0, x[2]);
-
-        //                struct ggml_tensor * f = ggml_sum(ctx0, ggml_flash_attn(ctx0, x[0], x[1], x[2], (masked == 0)));
-
-        //                check_gradient("flash_attn f32", ctx0, x, f, ndims, nargs, 1.5e-4f, 1e-3f, INFINITY, {});
-        //            }
-        //        }
-        //    }
-        //}
-
-        ggml_free(ctx0);
-    }
-
-    return 0;
-}
diff --git a/tests/test-grammar-integration.cpp b/tests/test-grammar-integration.cpp
index 5cc0cdb04..e1bdbb925 100644
--- a/tests/test-grammar-integration.cpp
+++ b/tests/test-grammar-integration.cpp
@@ -32,13 +32,10 @@ static bool test_build_grammar_fails(const std::string & grammar_str) {
 static bool match_string(const std::string & input, llama_grammar * grammar) {
     const auto cpts = unicode_cpts_from_utf8(input);
 
-    const llama_grammar_rules  & rules      = llama_grammar_get_rules (grammar);
-          llama_grammar_stacks & stacks_cur = llama_grammar_get_stacks(grammar);
+    auto & stacks_cur = llama_grammar_get_stacks(grammar);
 
     for (const auto & cpt : cpts) {
-        const llama_grammar_stacks stacks_prev = llama_grammar_get_stacks(grammar); // copy
-
-        llama_grammar_accept(rules, stacks_prev, cpt, stacks_cur);
+        llama_grammar_accept(grammar, cpt);
 
         if (stacks_cur.empty()) {
             // no stacks means that the grammar failed to match at this point
@@ -63,7 +60,7 @@ static void test(const std::string & test_desc, const std::string & grammar_str,
     auto * grammar = build_grammar(grammar_str);
 
     // Save the original grammar stacks so that we can reset after every new string we want to test
-    const llama_grammar_stacks stacks_org = llama_grammar_get_stacks(grammar);
+    const llama_grammar_stacks stacks_org = llama_grammar_get_stacks(grammar); // copy
 
     llama_grammar_stacks & stacks_cur = llama_grammar_get_stacks(grammar);
 
diff --git a/tests/test-llama-grammar.cpp b/tests/test-llama-grammar.cpp
index 6f1374ca8..e2129206b 100644
--- a/tests/test-llama-grammar.cpp
+++ b/tests/test-llama-grammar.cpp
@@ -113,12 +113,10 @@ int main()
         }
     }
 
-    llama_grammar * grammar = NULL;
     std::vector grammar_rules(parsed_grammar.c_rules());
 
-    grammar = llama_grammar_init_impl(nullptr, grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
-    if (grammar == nullptr)
-    {
+    llama_grammar * grammar = llama_grammar_init_impl(nullptr, grammar_rules.data(), grammar_rules.size(), parsed_grammar.symbol_ids.at("root"));
+    if (grammar == nullptr) {
         throw std::runtime_error("Failed to initialize llama_grammar");
     }
 
diff --git a/tests/test-lora-conversion-inference.sh b/tests/test-lora-conversion-inference.sh
index fe90ce0d1..1d1f4886c 100755
--- a/tests/test-lora-conversion-inference.sh
+++ b/tests/test-lora-conversion-inference.sh
@@ -10,11 +10,16 @@ declare -a params=(
 
 MODELS_REPO=lora-tests
 MODELS_REPO_URL=https://huggingface.co/ggml-org/$MODELS_REPO
+COMMIT=c26d5fb85b4070a9e9c4e65d132c783b98086890
 
 # Clone the Hugging Face repository if the directory does not exist
 if [ ! -d "$MODELS_REPO" ]; then
     echo "Cloning the Hugging Face repository..."
     git clone $MODELS_REPO_URL --depth 1
+    cd $MODELS_REPO
+    git fetch --depth=1 origin $COMMIT
+    git reset --hard $COMMIT
+    cd -
 else
     echo "Repository already exists. Skipping clone."
 fi
@@ -75,18 +80,18 @@ run_conversion_and_inference_lora() {
     # Run inference
     echo -e "\n\n---------------------------\n\n"
     echo "Running llama-cli without lora for $model_name with hidden_size $hidden_size..."
-    OUTPUT_BASE=$(./llama-cli -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32.gguf \
+    OUTPUT_BASE=$(./llama-cli -no-cnv -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32.gguf \
         -p "$EXPECTED_BASE_FIRST_WORD" -n 50 --seed 42 --temp 0)
 
     echo -e "\n\n---------------------------\n\n"
     echo "Running llama-cli with hot lora for $model_name with hidden_size $hidden_size..."
-    OUTPUT_LORA_HOT=$(./llama-cli -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32.gguf \
+    OUTPUT_LORA_HOT=$(./llama-cli -no-cnv -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32.gguf \
         --lora $MODELS_REPO/$model_name/hidden_size=$hidden_size/lora/Lora-F32-LoRA.gguf \
         -p "$EXPECTED_LORA_FIRST_WORD" -n 50 --seed 42 --temp 0)
 
     echo -e "\n\n---------------------------\n\n"
     echo "Running llama-cli with merged lora for $model_name with hidden_size $hidden_size..."
-    OUTPUT_LORA_MERGED=$(./llama-cli -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32-lora-merged.gguf \
+    OUTPUT_LORA_MERGED=$(./llama-cli -no-cnv -m $MODELS_REPO/$model_name/hidden_size=$hidden_size/base/Base-F32-lora-merged.gguf \
         -p "$EXPECTED_LORA_FIRST_WORD" -n 50 --seed 42 --temp 0)
 
     # Remove any initial white space
diff --git a/tests/test-model-load-cancel.cpp b/tests/test-model-load-cancel.cpp
index 858535c3c..9095826fa 100644
--- a/tests/test-model-load-cancel.cpp
+++ b/tests/test-model-load-cancel.cpp
@@ -21,7 +21,7 @@ int main(int argc, char *argv[] ) {
         (void) ctx;
         return progress > 0.50;
     };
-    auto * model = llama_load_model_from_file(model_path, params);
+    auto * model = llama_model_load_from_file(model_path, params);
     llama_backend_free();
     return model == nullptr ? EXIT_SUCCESS : EXIT_FAILURE;
 }
diff --git a/tests/test-opt.cpp b/tests/test-opt.cpp
index 546ca230b..f90c92b4b 100644
--- a/tests/test-opt.cpp
+++ b/tests/test-opt.cpp
@@ -1,181 +1,892 @@
 #include "ggml.h"
+#include "ggml-alloc.h"
+#include "ggml-backend.h"
+#include "ggml-cpu.h"
+#include "ggml-opt.h"
 
 #include 
-#include 
-#include 
-#include 
+#include 
+#include 
+#include 
+#include 
+#include 
 
-#define MAX_NARGS 2
-
-#if defined(__GNUC__)
-#pragma GCC diagnostic ignored "-Wdouble-promotion"
-#endif
-
-//
-// logging
-//
-#define GGML_DEBUG 0
-#if (GGML_DEBUG >= 1)
-#define GGML_PRINT_DEBUG(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG(...)
-#endif
-
-#if (GGML_DEBUG >= 5)
-#define GGML_PRINT_DEBUG_5(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_5(...)
-#endif
-
-#if (GGML_DEBUG >= 10)
-#define GGML_PRINT_DEBUG_10(...) printf(__VA_ARGS__)
-#else
-#define GGML_PRINT_DEBUG_10(...)
-#endif
-
-#define GGML_PRINT(...) printf(__VA_ARGS__)
-
-
-static float frand(void) {
-    return (float)rand()/(float)RAND_MAX;
+static bool almost_equal(const double a, const double b, const double atol) {
+    return fabs(a - b) < atol;
 }
 
-static struct ggml_tensor * get_random_tensor(
-    struct ggml_context * ctx0, int ndims, int64_t ne[], float fmin, float fmax
-) {
-    struct ggml_tensor * result = ggml_new_tensor(ctx0, GGML_TYPE_F32, ndims, ne);
+constexpr int64_t ne_datapoint = 2;
+constexpr int64_t ne_label     = 1;
+constexpr int64_t ndata        = 6;
 
-    switch (ndims) {
-        case 1:
-            for (int i0 = 0; i0 < ne[0]; i0++) {
-                ((float *)result->data)[i0] = frand()*(fmax - fmin) + fmin;
+struct helper_ctx_data {
+    std::vector   datasets_supervised;
+    std::vector data_batch;
+    std::vector labels_batch;
+
+    ggml_opt_dataset_t       dataset_unsupervised;
+    struct ggml_context    * ctx_static;
+    struct ggml_context    * ctx_compute;
+    struct ggml_opt_params   opt_params;
+    ggml_opt_context_t       opt_ctx;
+    struct ggml_tensor     * inputs;
+    struct ggml_tensor     * weights;
+    struct ggml_tensor     * outputs;
+    ggml_backend_buffer_t    buf;
+    ggml_opt_result_t        result;
+    ggml_opt_result_t        result2;
+};
+
+// These default values make it easier to check optimization results vs. expected values.
+static ggml_opt_optimizer_params helper_get_test_opt_pars(void * userdata) {
+    ggml_opt_optimizer_params result = ggml_opt_get_default_optimizer_params(userdata);
+    result.adamw.alpha = 1.0f;
+    result.adamw.beta1 = 0.0f;
+    result.adamw.beta2 = 0.0f;
+    result.adamw.eps   = 0.0f;
+    return result;
+}
+
+static helper_ctx_data helper_get_ctx_data(
+        ggml_backend_sched_t    backend_sched,
+        ggml_backend_t          backend,
+        const bool              init_opt_ctx       = true,
+        const bool              optimizer_defaults = true,
+        int64_t                 nbatch_logical     = 1,
+        int64_t                 nbatch_physical    = 1,
+        enum ggml_opt_loss_type loss_type          = GGML_OPT_LOSS_TYPE_SUM) {
+    std::vector datasets(ndata);
+    for (int64_t ndata_shard = 1; ndata_shard <= ndata; ++ndata_shard) {
+        ggml_opt_dataset_t dataset = ggml_opt_dataset_init(ne_datapoint, ne_label, ndata, ndata_shard);
+
+        float * data   = ggml_get_data_f32(ggml_opt_dataset_data(  dataset));
+        float * labels = ggml_get_data_f32(ggml_opt_dataset_labels(dataset));
+
+        for (int64_t idata = 0; idata < ndata; ++idata) {
+            for (int64_t id = 0; id < ne_datapoint; ++id) {
+                data[  idata*ne_datapoint + id] =     16*idata + id;
             }
-            break;
-        case 2:
-            for (int i1 = 0; i1 < ne[1]; i1++) {
-                for (int i0 = 0; i0 < ne[0]; i0++) {
-                    ((float *)result->data)[i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
-                }
+            for (int64_t il = 0; il < ne_label;     ++il) {
+                labels[idata*ne_label     + il] = 16*(16*idata + il);
             }
-            break;
-        case 3:
-            for (int i2 = 0; i2 < ne[2]; i2++) {
-                for (int i1 = 0; i1 < ne[1]; i1++) {
-                    for (int i0 = 0; i0 < ne[0]; i0++) {
-                        ((float *)result->data)[i2*ne[1]*ne[0] + i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
+        }
+
+        datasets[ndata_shard-1] = dataset;
+    }
+
+    ggml_opt_dataset_t dataset_unsupervised = ggml_opt_dataset_init(1, 0, ndata, /*ndata_shard =*/ 1);
+
+    float * data = ggml_get_data_f32(ggml_opt_dataset_data(dataset_unsupervised));
+
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        data[idata] = idata;
+    }
+
+    struct ggml_context * ctx_static;
+    struct ggml_context * ctx_compute;
+    {
+        struct ggml_init_params params = {
+            /*.mem_size   =*/ (2*ndata + 2)*ggml_tensor_overhead(),
+            /*.mem_buffer =*/ nullptr,
+            /*.no_alloc   =*/ true,
+        };
+        ctx_static = ggml_init(params);
+    }
+    {
+        struct ggml_init_params params = {
+            /*.mem_size   =*/ GGML_DEFAULT_GRAPH_SIZE*ggml_tensor_overhead() + 3*ggml_graph_overhead(),
+            /*.mem_buffer =*/ nullptr,
+            /*.no_alloc   =*/ true,
+        };
+        ctx_compute = ggml_init(params);
+    }
+
+    std::vector   data_batch(ndata);
+    std::vector labels_batch(ndata);
+    for (int64_t ndata_batch = 1; ndata_batch <= ndata; ++ndata_batch) {
+        data_batch[ndata_batch-1]   = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, ndata_batch*ne_datapoint);
+        labels_batch[ndata_batch-1] = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, ndata_batch*ne_label);
+    }
+
+    struct ggml_tensor * inputs = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, nbatch_physical);
+    ggml_set_name(inputs, "inputs");
+
+    struct ggml_tensor * weights = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, 1);
+    ggml_set_name(weights, "weights");
+    ggml_set_param(ctx_static, weights);
+
+    struct ggml_tensor * intermediary = ggml_add(ctx_compute, inputs, weights);
+
+    struct ggml_tensor * outputs = ggml_scale(ctx_compute, intermediary, 1.0f);
+    ggml_set_name(outputs, "outputs");
+
+    ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors(ctx_static, backend);
+    const float w0 = float(ndata)/2;
+    ggml_backend_tensor_set(weights, &w0, 0, sizeof(float));
+
+    GGML_ASSERT(nbatch_logical % nbatch_physical == 0);
+    const int32_t opt_period = nbatch_logical / nbatch_physical;
+
+    struct ggml_opt_params opt_params = ggml_opt_default_params(backend_sched, ctx_compute, inputs, outputs, loss_type);
+    opt_params.opt_period = opt_period;
+    if (!optimizer_defaults) {
+        opt_params.get_opt_pars = helper_get_test_opt_pars;
+    }
+    ggml_opt_context_t opt_ctx = init_opt_ctx ? ggml_opt_init(opt_params) : nullptr;
+
+    ggml_opt_result_t result  = ggml_opt_result_init();
+    ggml_opt_result_t result2 = ggml_opt_result_init();
+
+    return {datasets, data_batch, labels_batch, dataset_unsupervised, ctx_static, ctx_compute, opt_params, opt_ctx, inputs, weights, outputs, buf, result, result2};
+}
+
+static void helper_free_ctx_data(struct helper_ctx_data ctx_data) {
+    ggml_opt_result_free(ctx_data.result);
+    ggml_opt_result_free(ctx_data.result2);
+    ggml_opt_free(ctx_data.opt_ctx);
+    ggml_backend_buffer_free(ctx_data.buf);
+    ggml_free(ctx_data.ctx_static);
+    ggml_free(ctx_data.ctx_compute);
+    for (ggml_opt_dataset_t dataset : ctx_data.datasets_supervised) {
+        ggml_opt_dataset_free(dataset);
+    }
+    ggml_opt_dataset_free(ctx_data.dataset_unsupervised);
+}
+
+static void helper_after_test(
+        const char * func, const bool high_level, const std::string options,
+        const std::string subtest, const bool subtest_ok, int & ntest, int & npass) {
+    printf("  %s(high_level=%s%s, subtest=%s): ",
+           func, high_level ? "yes" : "no", options.c_str(), subtest.c_str());
+    if (subtest_ok) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+}
+
+static std::pair test_dataset(ggml_backend_sched_t backend_sched, ggml_backend_t backend, const bool shuffle) {
+    int ntest = 0;
+    int npass = 0;
+
+    struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend);
+
+    for (int64_t ndata_shard = 1; ndata_shard <= ndata; ++ndata_shard) {
+        ggml_opt_dataset_t dataset = cd.datasets_supervised[ndata_shard-1];
+
+        if (shuffle) {
+            ggml_opt_dataset_shuffle(cd.opt_ctx, dataset, -1);
+        }
+
+        for (int64_t ndata_batch = 1; ndata_batch <= ndata; ++ndata_batch) {
+            if (ndata_batch % ndata_shard != 0) {
+                continue;
+            }
+            bool subtest_ok = true;
+
+            struct ggml_tensor *   data_batch =   cd.data_batch[ndata_batch-1];
+            struct ggml_tensor * labels_batch = cd.labels_batch[ndata_batch-1];
+
+            std::vector   data(ggml_nelements(  data_batch));
+            std::vector labels(ggml_nelements(labels_batch));
+
+            std::vector idata_shuffled;
+            const int64_t nbatches = ndata / ndata_batch;
+            for (int64_t ibatch = 0; ibatch < nbatches; ++ibatch) {
+                ggml_opt_dataset_get_batch(dataset, data_batch, labels_batch, ibatch);
+
+                ggml_backend_tensor_get(  data_batch,   data.data(), 0, ggml_nbytes(  data_batch));
+                ggml_backend_tensor_get(labels_batch, labels.data(), 0, ggml_nbytes(labels_batch));
+
+                for (int64_t idata_batch = 0; idata_batch < ndata_batch; ++idata_batch) {
+                    const int64_t idata = ibatch*ndata_batch + idata_batch;
+                    const int64_t idata_found = data[idata_batch*ne_datapoint] / 16;
+                    subtest_ok = subtest_ok && (shuffle || idata_found == idata);
+                    idata_shuffled.push_back(idata_found);
+
+                    for (int64_t id = 0; id < ne_datapoint; ++id) {
+                        if (data[  idata_batch*ne_datapoint + id] != 16*idata_found + id) {
+                            subtest_ok = false;
+                        }
                     }
-                }
-            }
-            break;
-        case 4:
-            for (int i3 = 0; i3 < ne[3]; i3++) {
-                for (int i2 = 0; i2 < ne[2]; i2++) {
-                    for (int i1 = 0; i1 < ne[1]; i1++) {
-                        for (int i0 = 0; i0 < ne[0]; i0++) {
-                            ((float *)result->data)[i3*ne[2]*ne[1]*ne[0] + i2*ne[1]*ne[0] + i1*ne[0] + i0] = frand()*(fmax - fmin) + fmin;
+                    for (int64_t il = 0; il < ne_label;     ++il) {
+                        if (labels[idata_batch*ne_label     + il] != 16*(16*idata_found + il)) {
+                            subtest_ok = false;
                         }
                     }
                 }
             }
-            break;
-        default:
-            assert(false);
+
+            if (!shuffle || ndata % ndata_batch == 0) {
+                const int ndata_max = (ndata / ndata_batch) * ndata_batch;
+
+                for (int64_t idata = 0; subtest_ok && idata < ndata_max; ++idata) {
+                    int ninstances = 0;
+                    for (int64_t id : idata_shuffled) {
+                        ninstances += id == idata;
+                    }
+                    if (ninstances != 1) {
+                        subtest_ok = false;
+                    }
+                }
+            }
+
+            printf("  %s(shuffle=%s, ndata_shard=%" PRId64 ", ndata_batch=%" PRId64 "): ",
+                   __func__, shuffle ? "yes" : "no", ndata_shard, ndata_batch);
+            if (subtest_ok) {
+                printf("\033[1;32mOK\033[0m\n");
+                npass++;
+            } else {
+                printf("\033[1;31mFAIL\033[0m\n");
+            }
+            ntest++;
+        }
     }
 
+    helper_free_ctx_data(cd);
+
+    return std::make_pair(npass, ntest);
+}
+
+static std::pair test_grad(ggml_backend_sched_t backend_sched, ggml_backend_t backend) {
+    int ntest = 0;
+    int npass = 0;
+
+    struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend, /*init_opt_ctx =*/ true, /*optimizer_defaults =*/ false,
+    /*nbatch_logical =*/ 999999, /*nbatch_physical =*/ 1);
+
+    std::vector grad_history(ndata);
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        grad_history[idata] = NAN;
+    }
+
+    for (int idata = 0; idata < ndata; ++idata) {
+        const float idataf = idata;
+        ggml_backend_tensor_set(cd.inputs, &idataf, 0, ggml_nbytes(cd.inputs));
+        ggml_opt_forward_backward(cd.opt_ctx, cd.result);
+        ggml_backend_tensor_get(ggml_opt_grad_acc(cd.opt_ctx, cd.weights), grad_history.data() + idata, 0, sizeof(float));
+    }
+
+    {
+        bool subtest_ok = true;
+        for (int idata = 0; idata < ndata; ++idata) {
+            if (grad_history[idata] != idata + 1) {
+                subtest_ok = false;
+            }
+        }
+        printf("  %s(): ", __func__);
+        if (subtest_ok) {
+            printf("\033[1;32mOK\033[0m\n");
+            npass++;
+        } else {
+            printf("\033[1;31mFAIL\033[0m\n");
+        }
+        ntest++;
+    }
+
+    helper_free_ctx_data(cd);
+
+    return std::make_pair(npass, ntest);
+}
+
+static void helper_after_test_forward_backward(
+        const char * func, const bool high_level, const bool shuffle,
+        const std::string subtest, const bool subtest_ok, int & ntest, int & npass) {
+    std::string options = ", shuffle=";
+    options += shuffle ? "yes" : "no";
+    helper_after_test(func, high_level, options, subtest, subtest_ok, ntest, npass);
+}
+
+static std::pair test_forward_backward(
+        ggml_backend_sched_t backend_sched, ggml_backend_t backend, const bool high_level, const bool shuffle) {
+    int ntest = 0;
+    int npass = 0;
+
+    struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend, /*init_opt_ctx =*/ true, /*optimizer_defaults =*/ false);
+    struct ggml_tensor * loss = ggml_opt_loss(cd.opt_ctx);
+
+    std::vector loss_history(ndata);
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        loss_history[idata] = NAN;
+    }
+
+    {
+        int64_t ndata;
+        ggml_opt_result_ndata(cd.result, &ndata);
+        double loss;
+        double loss_unc;
+        ggml_opt_result_loss(cd.result, &loss, &loss_unc);
+        double accuracy;
+        double accuracy_unc;
+        ggml_opt_result_accuracy(cd.result, &accuracy, &accuracy_unc);
+        const bool subtest_ok = ndata == 0 && loss == 0.0 && std::isnan(loss_unc) && std::isnan(accuracy) && std::isnan(accuracy_unc);
+        helper_after_test_forward_backward(__func__, high_level, shuffle, "results_initial", subtest_ok, ntest, npass);
+    }
+
+    if (high_level) {
+        ggml_opt_dataset_t dataset = cd.dataset_unsupervised;
+        if (shuffle) {
+            ggml_opt_dataset_shuffle(cd.opt_ctx, dataset, -1);
+        }
+        ggml_opt_epoch(cd.opt_ctx, dataset, nullptr, cd.result, 0, nullptr, nullptr);
+    } else {
+        for (int idata = 0; idata < ndata; ++idata) {
+            const float idataf = idata;
+            ggml_backend_tensor_set(cd.inputs, &idataf, 0, ggml_nbytes(cd.inputs));
+            ggml_opt_forward(cd.opt_ctx, cd.result);
+            ggml_backend_tensor_get(loss, loss_history.data() + idata, 0, sizeof(float));
+        }
+    }
+
+    {
+        float weights;
+        ggml_backend_tensor_get(cd.weights, &weights, 0, sizeof(float));
+        const bool subtest_ok = weights == ndata/2;
+        helper_after_test_forward_backward(__func__, high_level, shuffle, "weights_after_forward", subtest_ok, ntest, npass);
+    }
+    {
+        int64_t ndata;
+        ggml_opt_result_ndata(cd.result, &ndata);
+        bool subtest_ok = ndata == 6;
+
+        double loss;
+        double loss_unc;
+        ggml_opt_result_loss(cd.result, &loss, &loss_unc);
+        subtest_ok = subtest_ok && loss == 33.0 && almost_equal(loss_unc, sqrt(3.5), 1e-10);
+
+        double accuracy;
+        double accuracy_unc;
+        ggml_opt_result_accuracy(cd.result, &accuracy, &accuracy_unc);
+        subtest_ok = subtest_ok && std::isnan(accuracy) && std::isnan(accuracy_unc);
+
+        helper_after_test_forward_backward(__func__, high_level, shuffle, "results_after_forward", subtest_ok, ntest, npass);
+    }
+
+    float w0;
+    ggml_backend_tensor_get(cd.weights, &w0, 0, sizeof(float));
+    for (int i = 0; i < 10; ++i) {
+        ggml_opt_forward_backward(cd.opt_ctx, nullptr);
+    }
+    ggml_backend_tensor_set(cd.weights, &w0, 0, sizeof(float));
+
+    ggml_opt_reset(cd.opt_ctx, /*optimizer =*/ false);
+    ggml_opt_result_reset(cd.result);
+
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        loss_history[idata] = NAN;
+    }
+
+    if (high_level) {
+        ggml_opt_dataset_t dataset = cd.dataset_unsupervised;
+        if (shuffle) {
+            ggml_opt_dataset_shuffle(cd.opt_ctx, dataset, -1);
+        }
+        ggml_opt_epoch(cd.opt_ctx, dataset, cd.result, nullptr, ndata, nullptr, nullptr);
+    } else {
+        for (int idata = 0; idata < ndata; ++idata) {
+            const float idataf = idata;
+            ggml_backend_tensor_set(cd.inputs, &idataf, 0, ggml_nbytes(cd.inputs));
+            ggml_opt_forward_backward(cd.opt_ctx, cd.result);
+            ggml_backend_tensor_get(loss, loss_history.data() + idata, 0, sizeof(float));
+        }
+    }
+
+    {
+        float weights;
+        ggml_backend_tensor_get(cd.weights, &weights, 0, sizeof(float));
+        const bool subtest_ok = weights == -ndata/2;
+        helper_after_test_forward_backward(__func__, high_level, shuffle, "weights_after_forward_backward", subtest_ok, ntest, npass);
+    }
+    {
+        int64_t ndata;
+        ggml_opt_result_ndata(cd.result, &ndata);
+        bool subtest_ok = ndata == 6;
+
+        double loss;
+        double loss_unc;
+        ggml_opt_result_loss(cd.result, &loss, &loss_unc);
+        subtest_ok = subtest_ok && loss == 18.0 && (shuffle || loss_unc == 0.0);
+
+        double accuracy;
+        double accuracy_unc;
+        ggml_opt_result_accuracy(cd.result, &accuracy, &accuracy_unc);
+        subtest_ok = subtest_ok && std::isnan(accuracy) && std::isnan(accuracy_unc);
+
+        helper_after_test_forward_backward(__func__, high_level, shuffle, "result_after_forward_backward", subtest_ok, ntest, npass);
+    }
+
+    helper_free_ctx_data(cd);
+
+    return std::make_pair(npass, ntest);
+}
+
+static std::pair test_epoch_vs_fit(ggml_backend_sched_t backend_sched, ggml_backend_t backend) {
+    int ntest = 0;
+    int npass = 0;
+
+    float weights_epoch;
+    float weights_fit;
+
+    {
+        struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend, /*init_opt_ctx =*/ true);
+        ggml_opt_dataset_t dataset = cd.dataset_unsupervised;
+
+        ggml_opt_dataset_shuffle(cd.opt_ctx, dataset, -1);
+        ggml_opt_epoch(cd.opt_ctx, dataset, cd.result, nullptr, ndata, nullptr, nullptr);
+
+        ggml_backend_tensor_get(cd.weights, &weights_epoch, 0, ggml_nbytes(cd.weights));
+        helper_free_ctx_data(cd);
+    }
+    {
+        struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend, /*init_opt_ctx =*/ false);
+        ggml_opt_dataset_t dataset = cd.dataset_unsupervised;
+
+        ggml_opt_fit(backend_sched, cd.ctx_compute, cd.inputs, cd.outputs, dataset,
+            GGML_OPT_LOSS_TYPE_SUM, ggml_opt_get_default_optimizer_params, 1, 1, 0.0f, true);
+
+        ggml_backend_tensor_get(cd.weights, &weights_fit, 0, ggml_nbytes(cd.weights));
+        helper_free_ctx_data(cd);
+    }
+
+    const bool subtest_ok = weights_epoch == weights_fit;
+
+    printf("  %s(): ", __func__);
+    if (subtest_ok) {
+        printf("\033[1;32mOK\033[0m\n");
+        npass++;
+    } else {
+        printf("\033[1;31mFAIL\033[0m\n");
+    }
+    ntest++;
+
+    return std::make_pair(npass, ntest);
+}
+
+static void helper_after_test_idata_split(
+        const char * func, const bool high_level, const int epoch,
+        const std::string subtest, const bool subtest_ok, int & ntest, int & npass) {
+    std::string options = ", epoch=";
+    options += std::to_string(epoch);
+    helper_after_test(func, high_level, options, subtest, subtest_ok, ntest, npass);
+}
+
+static std::pair test_idata_split(ggml_backend_sched_t backend_sched, ggml_backend_t backend, const bool high_level) {
+    int ntest = 0;
+    int npass = 0;
+
+    struct helper_ctx_data cd = helper_get_ctx_data(backend_sched, backend, /*init_opt_ctx =*/ true, /*optimizer_defaults =*/ false);
+    struct ggml_tensor * loss = ggml_opt_loss(cd.opt_ctx);
+    const int idata_split = ndata * 2/3;
+
+    std::vector loss_history(ndata);
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        loss_history[idata] = NAN;
+    }
+
+    for (int epoch = 1; epoch <= 4; ++epoch) {
+        if (high_level) {
+            ggml_opt_epoch(cd.opt_ctx, cd.dataset_unsupervised, cd.result, cd.result2, idata_split, nullptr, nullptr);
+        } else {
+            int idata = 0;
+            for (; idata < idata_split; ++idata) {
+                const float idataf = idata;
+                ggml_backend_tensor_set(cd.inputs, &idataf, 0, ggml_nbytes(cd.inputs));
+                ggml_opt_forward_backward(cd.opt_ctx, cd.result);
+                ggml_backend_tensor_get(loss, loss_history.data() + idata, 0, sizeof(float));
+            }
+            for (; idata < ndata; ++idata) {
+                const float idataf = idata;
+                ggml_backend_tensor_set(cd.inputs, &idataf, 0, ggml_nbytes(cd.inputs));
+                ggml_opt_forward(cd.opt_ctx, cd.result2);
+                ggml_backend_tensor_get(loss, loss_history.data() + idata, 0, sizeof(float));
+            }
+        }
+
+        {
+            float weights;
+            ggml_backend_tensor_get(cd.weights, &weights, 0, sizeof(float));
+            const bool subtest_ok = weights == ndata/2 - epoch*idata_split;
+            helper_after_test_idata_split(__func__, high_level, epoch, "weights", subtest_ok, ntest, npass);
+        }
+        {
+            int64_t ndata_result;
+            ggml_opt_result_ndata(cd.result, &ndata_result);
+            bool subtest_ok = ndata_result == idata_split;
+
+            double loss;
+            double loss_unc;
+            ggml_opt_result_loss(cd.result, &loss, &loss_unc);
+            subtest_ok = subtest_ok && loss == 28.0 - epoch*16.0 && loss_unc == 0.0;
+
+            double accuracy;
+            double accuracy_unc;
+            ggml_opt_result_accuracy(cd.result, &accuracy, &accuracy_unc);
+            subtest_ok = subtest_ok && std::isnan(accuracy) && std::isnan(accuracy_unc);
+
+            helper_after_test_idata_split(__func__, high_level, epoch, "results_backward", subtest_ok, ntest, npass);
+        }
+        {
+            int64_t ndata_result;
+            ggml_opt_result_ndata(cd.result2, &ndata_result);
+            bool subtest_ok = ndata_result == ndata - idata_split;
+
+            double loss;
+            double loss_unc;
+            ggml_opt_result_loss(cd.result2, &loss, &loss_unc);
+            subtest_ok = subtest_ok && loss == 15.0 - epoch*8 && almost_equal(loss_unc, sqrt(0.5), 1e-10);
+
+            double accuracy;
+            double accuracy_unc;
+            ggml_opt_result_accuracy(cd.result2, &accuracy, &accuracy_unc);
+            subtest_ok = subtest_ok && std::isnan(accuracy) && std::isnan(accuracy_unc);
+
+            helper_after_test_idata_split(__func__, high_level, epoch, "results_forward", subtest_ok, ntest, npass);
+        }
+
+        ggml_opt_result_reset(cd.result);
+        ggml_opt_result_reset(cd.result2);
+    }
+
+    helper_free_ctx_data(cd);
+
+    return std::make_pair(npass, ntest);
+}
+
+static void helper_after_test_gradient_accumulation(
+        const char * func, const int nbatch_physical, const enum ggml_opt_loss_type loss_type, const int epoch,
+        const std::string subtest, const bool subtest_ok, int & ntest, int & npass) {
+    std::string options = ", nbatch_physical=";
+    options += std::to_string(nbatch_physical);
+    options += ", loss_type=";
+    options += loss_type == GGML_OPT_LOSS_TYPE_MEAN ? "mean" : "sum";
+    options += ", epoch=";
+    options += std::to_string(epoch);
+    helper_after_test(func, false, options, subtest, subtest_ok, ntest, npass);
+}
+
+static std::pair test_gradient_accumulation(
+        ggml_backend_sched_t backend_sched, ggml_backend_t backend, const int32_t nbatch_physical, const enum ggml_opt_loss_type loss_type) {
+    int ntest = 0;
+    int npass = 0;
+
+    struct helper_ctx_data cd = helper_get_ctx_data(
+        backend_sched, backend, /*init_opt_ctx =*/ true, /*optimizer_defaults =*/ false, /*nbatch_logical =*/ 6, nbatch_physical, loss_type);
+    struct ggml_tensor * loss = ggml_opt_loss(cd.opt_ctx);
+
+    std::vector grad_history(ndata);
+    for (int64_t idata = 0; idata < ndata; ++idata) {
+        grad_history[idata] = NAN;
+    }
+
+    for (int epoch = 1; epoch <= 4; ++epoch) {
+        if (nbatch_physical == 1) {
+            for (int idata = 0; idata < ndata; ++idata) {
+                const float idataf = idata;
+                ggml_backend_tensor_set(cd.inputs, &idataf, 0, 1*sizeof(float));
+                ggml_opt_forward_backward(cd.opt_ctx, cd.result);
+                ggml_backend_tensor_get(ggml_opt_grad_acc(cd.opt_ctx, cd.weights), grad_history.data() + idata, 0, 1*sizeof(float));
+            }
+        } else if (nbatch_physical == 2) {
+            for (int idata = 0; idata < ndata; idata += 2) {
+                const float idataf[2] = {float(idata + 0), float(idata + 1)};
+                ggml_backend_tensor_set(cd.inputs, idataf, 0, 2*sizeof(float));
+                ggml_opt_forward_backward(cd.opt_ctx, cd.result);
+
+                grad_history[idata + 0] = 0.0f;
+                ggml_backend_tensor_get(ggml_opt_grad_acc(cd.opt_ctx, cd.weights), grad_history.data() + idata + 1, 0, 1*sizeof(float));
+            }
+        } else {
+            GGML_ASSERT(false);
+        }
+
+        {
+            GGML_ASSERT(ndata == 6);
+            constexpr double atol = 1e-6;
+            bool subtest_ok = true;
+            if (loss_type == GGML_OPT_LOSS_TYPE_SUM) {
+                if (nbatch_physical == 1) {
+                    subtest_ok = subtest_ok && almost_equal(grad_history[0], 1.0, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[2], 3.0, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[4], 5.0, atol);
+                } else {
+                    subtest_ok = subtest_ok && almost_equal(grad_history[0], 0.0, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[2], 0.0, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[4], 0.0, atol);
+                }
+                subtest_ok = subtest_ok && almost_equal(grad_history[1], 2.0, atol);
+                subtest_ok = subtest_ok && almost_equal(grad_history[3], 4.0, atol);
+                subtest_ok = subtest_ok && almost_equal(grad_history[5], 0.0, atol);
+            } else if (loss_type == GGML_OPT_LOSS_TYPE_MEAN) {
+                if (nbatch_physical == 1) {
+                    subtest_ok = subtest_ok && almost_equal(grad_history[0], 1.0/ndata, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[2], 3.0/ndata, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[4], 5.0/ndata, atol);
+                } else {
+                    subtest_ok = subtest_ok && almost_equal(grad_history[0], 0.0/ndata, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[2], 0.0/ndata, atol);
+                    subtest_ok = subtest_ok && almost_equal(grad_history[4], 0.0/ndata, atol);
+                }
+                subtest_ok = subtest_ok && almost_equal(grad_history[1], 2.0/ndata, atol);
+                subtest_ok = subtest_ok && almost_equal(grad_history[3], 4.0/ndata, atol);
+                subtest_ok = subtest_ok && almost_equal(grad_history[5], 0.0/ndata, atol);
+            } else {
+                GGML_ASSERT(false);
+            }
+            helper_after_test_gradient_accumulation(__func__, nbatch_physical, loss_type, epoch, "grads", subtest_ok, ntest, npass);
+        }
+        {
+            float weights;
+            ggml_backend_tensor_get(cd.weights, &weights, 0, sizeof(float));
+            const bool subtest_ok = weights == (ndata/2) - epoch;
+            helper_after_test_gradient_accumulation(__func__, nbatch_physical, loss_type, epoch, "weights", subtest_ok, ntest, npass);
+        }
+        {
+            int64_t ndata_result;
+            ggml_opt_result_ndata(cd.result, &ndata_result);
+            bool subtest_ok = ndata_result == ndata/nbatch_physical;
+
+            double loss;
+            ggml_opt_result_loss(cd.result, &loss, /*loss_unc =*/ nullptr);
+            if (loss_type == GGML_OPT_LOSS_TYPE_SUM) {
+                subtest_ok = subtest_ok && loss == (39.0 - epoch*6.0);
+            } else if (loss_type == GGML_OPT_LOSS_TYPE_MEAN) {
+                subtest_ok = subtest_ok && almost_equal(loss, (39.0 - epoch*6.0) / ndata, 1e-6);
+            } else {
+                GGML_ASSERT(false);
+            }
+
+            double accuracy;
+            double accuracy_unc;
+            ggml_opt_result_accuracy(cd.result, &accuracy, &accuracy_unc);
+            subtest_ok = subtest_ok && std::isnan(accuracy) && std::isnan(accuracy_unc);
+
+            helper_after_test_gradient_accumulation(__func__, nbatch_physical, loss_type, epoch, "results", subtest_ok, ntest, npass);
+        }
+
+        ggml_opt_result_reset(cd.result);
+    }
+
+    helper_free_ctx_data(cd);
+
+    return std::make_pair(npass, ntest);
+}
+
+static ggml_opt_optimizer_params helper_get_regression_opt_pars(void * userdata) {
+    ggml_opt_optimizer_params result = ggml_opt_get_default_optimizer_params(userdata);
+    result.adamw.alpha = 0.1f;
     return result;
 }
 
-int main(void) {
-    struct ggml_init_params params = {
-        /* .mem_size   = */ 1024*1024*1024,
-        /* .mem_buffer = */ NULL,
-        /* .no_alloc   = */ false,
-    };
+static std::pair test_regression(ggml_backend_sched_t backend_sched, ggml_backend_t backend) {
+    int ntest = 0;
+    int npass = 0;
 
-    struct ggml_context * ctx = ggml_init(params);
+    // Test for simple regression with f(x) = a*x + b
 
-    int64_t ne1[4] = {4, 128, 1, 1};
-    int64_t ne2[4] = {4, 256, 1, 1};
-    int64_t ne3[4] = {128, 256, 1, 1};
+    constexpr int64_t ndata_regression = 201;
+    constexpr float a_true = 1.2f;
+    constexpr float b_true = 3.4f;
 
-    struct ggml_tensor * a = get_random_tensor(ctx, 2, ne1, -1, +1);
-    struct ggml_tensor * b = get_random_tensor(ctx, 2, ne2, -1, +1);
-    ggml_set_param(ctx, a);
-    ggml_set_param(ctx, b);
+    std::mt19937 gen(12345);
+    std::normal_distribution nd{0.0f, 0.1f};
 
-    struct ggml_tensor * c = get_random_tensor(ctx, 2, ne3, -1, +1);
+    ggml_opt_dataset_t dataset = ggml_opt_dataset_init(1, 1, ndata_regression, ndata_regression);
 
-    struct ggml_tensor * ab = ggml_mul_mat(ctx, a, b);
-    struct ggml_tensor * d  = ggml_sub(ctx, c, ab);
-    struct ggml_tensor * e  = ggml_sum(ctx, ggml_sqr(ctx, d));
+    float * data   = ggml_get_data_f32(ggml_opt_dataset_data(  dataset));
+    float * labels = ggml_get_data_f32(ggml_opt_dataset_labels(dataset));
 
-    struct ggml_cgraph * ge = ggml_new_graph_custom(ctx, GGML_DEFAULT_GRAPH_SIZE, true);
-    ggml_build_forward_expand(ge, e);
-    ggml_graph_reset(ge);
+    constexpr float x_min = -100.0f;
+    constexpr float x_max =  100.0f;
 
-    ggml_graph_compute_with_ctx(ctx, ge, /*n_threads*/ 1);
+    for (int64_t idata = 0; idata < ndata_regression; ++idata) {
+        const float x = x_min + (x_max - x_min) * idata/(ndata_regression-1);
+        const float y = a_true*x + b_true + nd(gen);
 
-    const float fe = ggml_get_f32_1d(e, 0);
-    printf("%s: e = %.4f\n", __func__, fe);
+        data[idata]   = x;
+        labels[idata] = y;
+    }
 
-    struct ggml_opt_params opt_params = ggml_opt_default_params(GGML_OPT_TYPE_ADAM);
+    struct ggml_context * ctx_static;
+    struct ggml_context * ctx_compute;
+    {
+        struct ggml_init_params params = {
+            /*.mem_size   =*/ 3*ggml_tensor_overhead(),
+            /*.mem_buffer =*/ nullptr,
+            /*.no_alloc   =*/ true,
+        };
+        ctx_static = ggml_init(params);
+    }
+    {
+        struct ggml_init_params params = {
+            /*.mem_size   =*/ GGML_DEFAULT_GRAPH_SIZE*ggml_tensor_overhead() + 3*ggml_graph_overhead(),
+            /*.mem_buffer =*/ nullptr,
+            /*.no_alloc   =*/ true,
+        };
+        ctx_compute = ggml_init(params);
+    }
 
-    ggml_opt(ctx, opt_params, e);
+    // The first dimension is the dimension of the datapoints, the second dimension is the number of datapoints.
+    struct ggml_tensor * x = ggml_new_tensor_2d(ctx_static, GGML_TYPE_F32, 1, ndata_regression);
+    ggml_set_name(x, "x");
 
-    ggml_graph_reset(ge);
+    struct ggml_tensor * a = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, 1);
+    ggml_set_name(a, "a");
+    ggml_set_param(ctx_static, a);
 
-    ggml_graph_compute_with_ctx(ctx, ge, /*n_threads*/ 1);
+    struct ggml_tensor * b = ggml_new_tensor_1d(ctx_static, GGML_TYPE_F32, 1);
+    ggml_set_name(b, "b");
+    ggml_set_param(ctx_static, b);
 
-    const float fe_opt = ggml_get_f32_1d(e, 0);
-    printf("%s: original  e = %.4f\n", __func__, fe);
-    printf("%s: optimized e = %.4f\n", __func__, fe_opt);
+    struct ggml_tensor * f = ggml_add(ctx_compute, ggml_mul(ctx_compute, x, a), b);
+    ggml_set_name(f, "f");
+    ggml_set_param(ctx_static, f);
 
-    const bool success = (fe_opt <= fe);
-    assert(success);
+    ggml_backend_buffer_t buf = ggml_backend_alloc_ctx_tensors(ctx_static, backend);
+    const float a0 = 1.0f;
+    const float b0 = 3.0f;
+    ggml_backend_tensor_set(a, &a0, 0, sizeof(float));
+    ggml_backend_tensor_set(b, &b0, 0, sizeof(float));
 
-    ggml_free(ctx);
-    return success ? 0 : -1;
+    ggml_opt_fit(backend_sched, ctx_compute, x, f, dataset, GGML_OPT_LOSS_TYPE_MEAN_SQUARED_ERROR,
+        helper_get_regression_opt_pars, 100, ndata_regression, 0.0f, true);
+
+    {
+        float a_fit;
+        ggml_backend_tensor_get(a, &a_fit, 0, sizeof(float));
+        float b_fit;
+        ggml_backend_tensor_get(b, &b_fit, 0, sizeof(float));
+        const bool subtest_ok = almost_equal(a_fit, a_true, 1e-2) && almost_equal(b_fit, b_true, 1e-2);
+        printf("  %s(subtest=weights): ", __func__);
+        if (subtest_ok) {
+            printf("\033[1;32mOK\033[0m\n");
+            npass++;
+        } else {
+            printf("\033[1;31mFAIL\033[0m\n");
+        }
+        ntest++;
+    }
+
+    ggml_backend_buffer_free(buf);
+    ggml_free(ctx_static);
+    ggml_opt_dataset_free(dataset);
+
+    return std::make_pair(npass, ntest);
 }
-// int64_t ne1[4] = {4, 128, 1, 1};
-// int64_t ne2[4] = {4, 256, 1, 1};;
-// int64_t ne3[4] = {128, 256, 1, 1};
-// main: original  e = 25890.9375
-// main: optimized e = 10094.7031
 
-// int64_t ne1[4] = {8, 128, 1, 1};
-// int64_t ne2[4] = {8, 256, 1, 1};;
-// int64_t ne3[4] = {128, 256, 1, 1};
-// main: original  e = 39429.5078
-// main: optimized e = 9275.8936
+static std::pair test_backend(ggml_backend_sched_t backend_sched, ggml_backend_t backend) {
+    int npass = 0;
+    int ntest = 0;
 
-// int64_t ne1[4] = {16, 128, 1, 1};
-// int64_t ne2[4] = {16, 256, 1, 1};;
-// int64_t ne3[4] = {128, 256, 1, 1};
-// main: original  e = 68371.1328
-// main: optimized e = 7854.4502
+    for (bool shuffle : {false, true}) {
+        std::pair partial = test_dataset(backend_sched, backend, shuffle);
+        npass += partial.first;
+        ntest += partial.second;
+    }
+    {
+        std::pair partial = test_grad(backend_sched, backend);
+        npass += partial.first;
+        ntest += partial.second;
+    }
+    for (bool high_level : {false, true}){
+        for (bool shuffle : {false, true}) {
+            if (!high_level && shuffle) {
+                continue;
+            }
 
+            std::pair partial = test_forward_backward(backend_sched, backend, high_level, shuffle);
+            npass += partial.first;
+            ntest += partial.second;
+        }
+    }
+    {
+        std::pair partial = test_epoch_vs_fit(backend_sched, backend);
+        npass += partial.first;
+        ntest += partial.second;
+    }
+    for (bool high_level : {false, true}){
+        std::pair partial = test_idata_split(backend_sched, backend, high_level);
+        npass += partial.first;
+        ntest += partial.second;
+    }
+    for (int32_t nbatch_physical : {2, 1}) {
+        for (enum ggml_opt_loss_type loss_type : {GGML_OPT_LOSS_TYPE_SUM, GGML_OPT_LOSS_TYPE_MEAN}) {
+            std::pair partial = test_gradient_accumulation(backend_sched, backend, nbatch_physical, loss_type);
+            npass += partial.first;
+            ntest += partial.second;
+        }
+    }
+    {
+        std::pair partial = test_regression(backend_sched, backend);
+        npass += partial.first;
+        ntest += partial.second;
+    }
 
-// int64_t ne1[4] = {32, 128, 1, 1};
-// int64_t ne2[4] = {32, 256, 1, 1};;
-// int64_t ne3[4] = {128, 256, 1, 1};
-// main: original  e = 126061.1953
-// main: optimized e = 5451.0166
+    return std::make_pair(npass, ntest);
+}
 
-// int64_t ne1[4] = {4, 1024, 1, 1};
-// int64_t ne2[4] = {4, 2048, 1, 1};;
-// int64_t ne3[4] = {1024, 2048, 1, 1};
-// main: original  e = 1620817.8750
-// main: optimized e = 698387.6875
+int main(void) {
+    const size_t dev_count = ggml_backend_dev_count();
+    printf("Testing %zu devices\n\n", dev_count);
+    size_t n_ok = 0;
 
-// another run on M1
-// int64_t ne1[4] = {4, 1024, 1, 1};
-// int64_t ne2[4] = {4, 2048, 1, 1};;
-// int64_t ne3[4] = {1024, 2048, 1, 1};
-// main: original  e = 1629595.6250
-// main: optimized e = 698169.1250
+    std::vector devs;
+    std::vector     backends;
 
-// int64_t ne1[4] = {32, 1024, 1, 1};
-// int64_t ne2[4] = {32, 2048, 1, 1};;
-// int64_t ne3[4] = {1024, 2048, 1, 1};
-// main: original  e = 8146770.5000
-// main: optimized e = 651119.1250
+    for (size_t i = 0; i < dev_count; ++i) {
+        devs.push_back(ggml_backend_dev_get(i));
+
+        ggml_backend_t backend = ggml_backend_dev_init(devs[i], NULL);
+        GGML_ASSERT(backend != NULL);
+
+        if (ggml_backend_is_cpu(backend)) {
+            ggml_backend_cpu_set_n_threads(backend, std::thread::hardware_concurrency() / 2);
+        }
+
+        backends.push_back(backend);
+    }
+
+    for (size_t i = 0; i < dev_count; ++i) {
+        // Put the backend to be tested in front so that it's prioritized:
+        std::vector backends_modded = {backends[i]};
+        backends_modded.insert(backends_modded.end(), backends.begin(), backends.end());
+
+        ggml_backend_sched_t backend_sched = ggml_backend_sched_new(
+            backends_modded.data(), nullptr, backends_modded.size(), GGML_DEFAULT_GRAPH_SIZE, false);
+
+        printf("Backend %zu/%zu: %s\n", i + 1, dev_count, ggml_backend_dev_name(devs[i]));
+        printf("  Device description: %s\n", ggml_backend_dev_description(devs[i]));
+        size_t free, total; // NOLINT
+        ggml_backend_dev_memory(devs[i], &free, &total);
+        printf("  Device memory: %zu MB (%zu MB free)\n", total / 1024 / 1024, free / 1024 / 1024);
+        printf("\n");
+
+        std::pair result = test_backend(backend_sched, backends[i]);
+
+        printf("  %d/%d tests passed\n", result.first, result.second);
+        printf("  Backend %s: ", ggml_backend_name(backends[i]));
+        if (result.first == result.second) {
+            printf("\033[1;32mOK\033[0m\n");
+            n_ok++;
+        } else {
+            printf("\033[1;31mFAIL\033[0m\n");
+        }
+
+        printf("\n");
+
+        ggml_backend_sched_free(backend_sched);
+    }
+
+    for (ggml_backend_t backend : backends) {
+        ggml_backend_free(backend);
+    }
+
+    printf("%zu/%zu backends passed\n", n_ok, dev_count);
+    if (n_ok != dev_count) {
+        printf("\033[1;31mFAIL\033[0m\n");
+        return 1;
+    }
+    printf("\033[1;32mOK\033[0m\n");
+    return 0;
+}
diff --git a/tests/test-quantize-fns.cpp b/tests/test-quantize-fns.cpp
index 000e60adf..c77c8ed13 100644
--- a/tests/test-quantize-fns.cpp
+++ b/tests/test-quantize-fns.cpp
@@ -45,22 +45,23 @@ static float array_rmse(const float * a1, const float * a2, size_t n) {
 }
 
 // Total quantization error on test data
-static float total_quantization_error(const ggml_type_traits * qfns, size_t test_size, const float * test_data) {
+static float total_quantization_error(const ggml_type_traits * qfns, const ggml_type_traits_cpu * qfns_cpu, size_t test_size, const float * test_data) {
     std::vector tmp_q(2*test_size);
     std::vector tmp_out(test_size);
 
-    qfns->from_float(test_data, tmp_q.data(), test_size);
+    qfns_cpu->from_float(test_data, tmp_q.data(), test_size);
     qfns->to_float(tmp_q.data(), tmp_out.data(), test_size);
     return array_rmse(test_data, tmp_out.data(), test_size);
 }
 
 // Total quantization error on test data
-static float reference_quantization_error(const ggml_type_traits * qfns, size_t test_size, const float * test_data) {
+static float reference_quantization_error(const ggml_type_traits * qfns, const ggml_type_traits_cpu * qfns_cpu, size_t test_size, const float * test_data) {
     std::vector tmp_q(2*test_size);
     std::vector tmp_out(test_size);
     std::vector tmp_out_ref(test_size);
 
-    qfns->from_float(test_data, tmp_q.data(), test_size);
+    // FIXME: why is done twice?
+    qfns_cpu->from_float(test_data, tmp_q.data(), test_size);
     qfns->to_float(tmp_q.data(), tmp_out.data(), test_size);
 
     qfns->from_float_ref(test_data, tmp_q.data(), test_size);
@@ -78,15 +79,15 @@ static float dot_product(const float * a1, const float * a2, size_t test_size) {
 }
 
 // Total dot product error
-static float dot_product_error(
-    const ggml_type_traits * qfns, const ggml_type_traits_cpu * qfns_cpu, size_t test_size, const float * test_data1, const float *test_data2
-) {
+static float dot_product_error(const ggml_type_traits * qfns, const ggml_type_traits_cpu * qfns_cpu, size_t test_size, const float * test_data1, const float * test_data2) {
+    GGML_UNUSED(qfns);
+
     std::vector tmp_q1(2*test_size);
     std::vector tmp_q2(2*test_size);
 
-    const auto * vdot = ggml_get_type_traits(qfns_cpu->vec_dot_type);
+    const auto * vdot = ggml_get_type_traits_cpu(qfns_cpu->vec_dot_type);
 
-    qfns->from_float(test_data1, tmp_q1.data(), test_size);
+    qfns_cpu->from_float(test_data1, tmp_q1.data(), test_size);
     vdot->from_float(test_data2, tmp_q2.data(), test_size);
 
     float result = INFINITY;
@@ -145,8 +146,8 @@ int main(int argc, char * argv[]) {
         printf("Testing %s\n", ggml_type_name((ggml_type) i));
         ggml_quantize_init(ei);
 
-        if (qfns->from_float && qfns->to_float) {
-            const float total_error = total_quantization_error(qfns, test_size, test_data.data());
+        if (qfns_cpu->from_float && qfns->to_float) {
+            const float total_error = total_quantization_error(qfns, qfns_cpu, test_size, test_data.data());
             const float max_quantization_error =
                 type == GGML_TYPE_TQ1_0   ? MAX_QUANTIZATION_TOTAL_ERROR_TERNARY :
                 type == GGML_TYPE_TQ2_0   ? MAX_QUANTIZATION_TOTAL_ERROR_TERNARY :
@@ -161,7 +162,7 @@ int main(int argc, char * argv[]) {
                 printf("%5s absolute quantization error:    %s (%f)\n", ggml_type_name(type), RESULT_STR[failed], total_error);
             }
 
-            const float reference_error = reference_quantization_error(qfns, test_size, test_data.data());
+            const float reference_error = reference_quantization_error(qfns, qfns_cpu, test_size, test_data.data());
             failed = !(reference_error < MAX_QUANTIZATION_REFERENCE_ERROR);
             num_failed += failed;
             if (failed || verbose) {
diff --git a/tests/test-quantize-perf.cpp b/tests/test-quantize-perf.cpp
index 221424de8..288288493 100644
--- a/tests/test-quantize-perf.cpp
+++ b/tests/test-quantize-perf.cpp
@@ -7,7 +7,6 @@
 #include 
 #include 
 #include 
-#include 
 #include 
 #include 
 #include 
@@ -123,9 +122,10 @@ static void usage(char * argv[]) {
     printf("  --type TYPE           set test type as");
     for (int i = 0; i < GGML_TYPE_COUNT; i++) {
         ggml_type type = (ggml_type) i;
-        const auto * qfns = ggml_get_type_traits(type);
+        const auto * qfns     = ggml_get_type_traits(type);
+        const auto * qfns_cpu = ggml_get_type_traits_cpu(type);
         if (ggml_type_name(type) != NULL) {
-            if (qfns->from_float && qfns->to_float) {
+            if (qfns_cpu->from_float && qfns->to_float) {
                 printf(" %s", ggml_type_name(type));
             }
         }
@@ -277,7 +277,7 @@ int main(int argc, char * argv[]) {
             continue;
         }
 
-        if (qfns->from_float && qfns->to_float) {
+        if (qfns_cpu->from_float && qfns->to_float) {
             printf("%s\n", ggml_type_name(type));
 
             ggml_quantize_init(type);
@@ -301,7 +301,7 @@ int main(int argc, char * argv[]) {
                 for (size_t size : params.test_sizes) {
                     printf("    %zu values (%.2f MB)\n", size, 4*size/(float)(1024*1024));
                     auto quantize_fn = [&](void) -> float {
-                        qfns->from_float(test_data1, test_q1, size);
+                        qfns_cpu->from_float(test_data1, test_q1, size);
                         return test_q1[0];
                     };
                     size_t quantized_size = ggml_row_size(type, size);
@@ -312,7 +312,7 @@ int main(int argc, char * argv[]) {
 
             if (params.op_dequantize_row_q) {
                 printf("  dequantize_row_q\n");
-                qfns->from_float(test_data1, test_q1, largest);
+                qfns_cpu->from_float(test_data1, test_q1, largest);
                 for (size_t size : params.test_sizes) {
                     printf("    %zu values (%.2f MB)\n", size, 4*size/(float)(1024*1024));
                     auto quantize_fn = [&](void) -> float {
@@ -330,7 +330,7 @@ int main(int argc, char * argv[]) {
                 for (size_t size : params.test_sizes) {
                     printf("    %zu values (%.2f MB)\n", size, 4*size/(float)(1024*1024));
                     auto quantize_fn = [&](void) -> float {
-                        const auto * vdot = ggml_get_type_traits(qfns_cpu->vec_dot_type);
+                        const auto * vdot = ggml_get_type_traits_cpu(qfns_cpu->vec_dot_type);
                         vdot->from_float(test_data1, test_q1, size);
                         return test_q1[0];
                     };
@@ -342,8 +342,8 @@ int main(int argc, char * argv[]) {
 
             if (params.op_vec_dot_q) {
                 printf("  vec_dot_q\n");
-                qfns->from_float(test_data1, test_q1, largest);
-                qfns->from_float(test_data2, test_q2, largest);
+                qfns_cpu->from_float(test_data1, test_q1, largest);
+                qfns_cpu->from_float(test_data2, test_q2, largest);
                 for (size_t size : params.test_sizes) {
                     printf("    %zu values (%.2f MB)\n", size, 4*size/(float)(1024*1024));
                     auto quantize_fn = [&](void) -> float {
diff --git a/tests/test-rope.cpp b/tests/test-rope.cpp
index 4656b30f0..322b8bb99 100644
--- a/tests/test-rope.cpp
+++ b/tests/test-rope.cpp
@@ -138,7 +138,7 @@ int main(int /*argc*/, const char ** /*argv*/) {
     struct ggml_tensor * x;
 
     // rope f32
-    for (int m = 0; m < 3; ++m) {
+    for (int m = 0; m < 5; ++m) {
         const int ndims = 4;
 
         const int64_t n_rot = 128;
@@ -147,28 +147,69 @@ int main(int /*argc*/, const char ** /*argv*/) {
         const int n_past_0 = 100;
         const int n_past_2 = 33;
 
-        struct ggml_tensor * p0 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
-        struct ggml_tensor * p1 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
-        struct ggml_tensor * p2 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
-
-        for (int i = 0; i < ne[2]; ++i) {
-            ((int32_t *) p0->data)[i] = n_past_0 + i;
-            ((int32_t *) p1->data)[i] = n_past_2 - n_past_0;
-            ((int32_t *) p2->data)[i] = n_past_2 + i;
-        }
-
-        // test mode 0, 2, 4 (standard, GPT-NeoX, GLM)
-        const int mode = m == 0 ? 0 : m == 1 ? 2 : 4;
-
+        struct ggml_tensor * r0;
+        struct ggml_tensor * r1;
+        struct ggml_tensor * r2;
         x = get_random_tensor_f32(ctx0, ndims, ne, -1.0f, 1.0f);
+        int mode = -1;
 
-        // 100, 101, 102, ..., 172
-        struct ggml_tensor * r0 = ggml_rope(ctx0, x,  p0, n_rot, mode);
-        // -67, -67, -67, ..., -67
-        struct ggml_tensor * r1 = ggml_rope(ctx0, r0, p1, n_rot, mode); // "context swap", i.e. forget n_past_0 - n_past_2 tokens
+        if (m < 3) {
+            struct ggml_tensor * p0 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
+            struct ggml_tensor * p1 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
+            struct ggml_tensor * p2 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2]);
 
-        //  33,  34,  35, ..., 105
-        struct ggml_tensor * r2 = ggml_rope(ctx0, x,  p2, n_rot, mode);
+            for (int i = 0; i < ne[2]; ++i) {
+                ((int32_t *) p0->data)[i] = n_past_0 + i;
+                ((int32_t *) p1->data)[i] = n_past_2 - n_past_0;
+                ((int32_t *) p2->data)[i] = n_past_2 + i;
+            }
+            // test mode 0, 2, 4 (standard, GPT-NeoX, GLM)
+            mode = m == 0 ? 0 : m == 1 ? 2 : 4;
+
+            // 100, 101, 102, ..., 172
+            r0 = ggml_rope(ctx0, x,  p0, n_rot, mode);
+            // -67, -67, -67, ..., -67
+            r1 = ggml_rope(ctx0, r0, p1, n_rot, mode); // "context swap", i.e. forget n_past_0 - n_past_2 tokens
+
+            //  33,  34,  35, ..., 105
+            r2 = ggml_rope(ctx0, x,  p2, n_rot, mode);
+        } else {
+            // testing multi-dimension rope position embedding mode
+            struct ggml_tensor * p0 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2] * 4);
+            struct ggml_tensor * p1 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2] * 4);
+            struct ggml_tensor * p2 = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, ne[2] * 4);
+
+            int sections[4] = {16, 24, 24, 0};
+            mode = (m == 3) ? GGML_ROPE_TYPE_MROPE : GGML_ROPE_TYPE_VISION;
+
+            for (int i = 0; i < ne[2]; ++i) {
+                for (int j = 0; j < 4; ++j) {
+                    ((int32_t *) p0->data)[i + ne[2] * j] = n_past_0 + i + j;
+                    ((int32_t *) p1->data)[i + ne[2] * j] = n_past_2 - n_past_0;
+                    ((int32_t *) p2->data)[i + ne[2] * j] = n_past_2 + i + j;
+                }
+            }
+
+            // [[100, 101, 102, ..., 172],
+            // [101, 102, 103, ..., 173],
+            // [102, 103, 104, ..., 174]]
+            r0 = ggml_rope_multi(
+                ctx0, x, p0, nullptr,
+                n_rot, sections, mode, 32768, 1000000, 1, 0, 1, 32, 1);
+            // [[-67, -67, -67, ..., -67]
+            // [-67, -67, -67, ..., -67]
+            // [-67, -67, -67, ..., -67]]
+            r1 = ggml_rope_multi(
+                ctx0, r0, p1, nullptr,
+                n_rot, sections, mode, 32768, 1000000, 1, 0, 1, 32, 1);
+
+            //  [[33,  34,  35, ..., 105]
+            //  [34,  35,  36, ..., 106]
+            //  [35,  36,  37, ..., 107]]
+            r2 = ggml_rope_multi(
+                ctx0, x, p2, nullptr,
+                n_rot, sections, mode, 32768, 1000000, 1, 0, 1, 32, 1);
+        }
 
         ggml_cgraph * gf = ggml_new_graph(ctx0);
 
diff --git a/tests/test-sampling.cpp b/tests/test-sampling.cpp
index be370044d..c0dcb4848 100644
--- a/tests/test-sampling.cpp
+++ b/tests/test-sampling.cpp
@@ -145,7 +145,7 @@ static void test_penalties(
     sampler_tester tester(probs, probs_expected);
 
     const size_t n_vocab = probs.size();
-    auto * sampler = llama_sampler_init_penalties(n_vocab, LLAMA_TOKEN_NULL, LLAMA_TOKEN_NULL, last_tokens.size(), repeat_penalty, alpha_frequency, alpha_presence, false, false);
+    auto * sampler = llama_sampler_init_penalties(last_tokens.size(), repeat_penalty, alpha_frequency, alpha_presence);
 
     for (size_t i = 0; i < last_tokens.size(); i++) {
         llama_sampler_accept(sampler, last_tokens[i]);
@@ -284,7 +284,7 @@ static void test_perf() {
 
     data.reserve(n_vocab);
     for (int i = 0; i < n_vocab; i++) {
-        const float logit = 2.0f*((float)(rand())/RAND_MAX - 0.5f);
+        const float logit = 2.0f*((double)(rand())/RAND_MAX - 0.5);
         data.emplace_back(llama_token_data{i, logit, 0.0f});
     }
 
diff --git a/tests/test-tokenizer-0.cpp b/tests/test-tokenizer-0.cpp
index 0af85f002..59dda4877 100644
--- a/tests/test-tokenizer-0.cpp
+++ b/tests/test-tokenizer-0.cpp
@@ -152,7 +152,7 @@ int main(int argc, char **argv) {
 
         mparams.vocab_only = true;
 
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
 
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -161,11 +161,11 @@ int main(int argc, char **argv) {
 
         auto cparams = llama_context_default_params();
 
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
 
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
@@ -300,7 +300,7 @@ int main(int argc, char **argv) {
         fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());
     }
 
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
 
     llama_backend_free();
diff --git a/tests/test-tokenizer-1-bpe.cpp b/tests/test-tokenizer-1-bpe.cpp
index 0ff7fc833..55425d88a 100644
--- a/tests/test-tokenizer-1-bpe.cpp
+++ b/tests/test-tokenizer-1-bpe.cpp
@@ -46,7 +46,7 @@ int main(int argc, char **argv) {
 
         mparams.vocab_only = true;
 
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
 
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -55,17 +55,19 @@ int main(int argc, char **argv) {
 
         auto cparams = llama_context_default_params();
 
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
 
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
 
-    //GGML_ASSERT(llama_vocab_type(model) == LLAMA_VOCAB_TYPE_BPE);
-    if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_BPE) {
+    const llama_vocab * vocab = llama_model_get_vocab(model);
+
+    //GGML_ASSERT(llama_vocab_type(vocab) == LLAMA_VOCAB_TYPE_BPE);
+    if (llama_vocab_type(vocab) != LLAMA_VOCAB_TYPE_BPE) {
         return 99;
     }
 
@@ -75,7 +77,7 @@ int main(int argc, char **argv) {
     atexit([]() { console::cleanup(); });
 #endif
 
-    const int n_vocab = llama_n_vocab(model);
+    const int n_vocab = llama_vocab_n_tokens(vocab);
 
     for (int i = 0; i < n_vocab; ++i) {
         std::string str = common_detokenize(ctx, std::vector(1, i));
@@ -143,7 +145,7 @@ int main(int argc, char **argv) {
         }
     }
 
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
 
     llama_backend_free();
diff --git a/tests/test-tokenizer-1-spm.cpp b/tests/test-tokenizer-1-spm.cpp
index 9b0716a43..9e7b77f31 100644
--- a/tests/test-tokenizer-1-spm.cpp
+++ b/tests/test-tokenizer-1-spm.cpp
@@ -34,7 +34,7 @@ int main(int argc, char ** argv) {
 
         mparams.vocab_only = true;
 
-        model = llama_load_model_from_file(fname.c_str(), mparams);
+        model = llama_model_load_from_file(fname.c_str(), mparams);
 
         if (model == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
@@ -43,17 +43,19 @@ int main(int argc, char ** argv) {
 
         auto cparams = llama_context_default_params();
 
-        ctx = llama_new_context_with_model(model, cparams);
+        ctx = llama_init_from_model(model, cparams);
 
         if (ctx == NULL) {
             fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
-            llama_free_model(model);
+            llama_model_free(model);
             return 1;
         }
     }
 
+    const llama_vocab * vocab = llama_model_get_vocab(model);
+
     //GGML_ASSERT(llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM);
-    if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_SPM) {
+    if (llama_vocab_type(vocab) != LLAMA_VOCAB_TYPE_SPM) {
         return 99;
     }
 
@@ -63,7 +65,7 @@ int main(int argc, char ** argv) {
     atexit([]() { console::cleanup(); });
 #endif
 
-    const int n_vocab = llama_n_vocab(model);
+    const int n_vocab = llama_vocab_n_tokens(vocab);
 
     for (int i = 0; i < n_vocab; ++i) {
         std::string str = common_detokenize(ctx, std::vector(1, i), true);
@@ -113,7 +115,7 @@ int main(int argc, char ** argv) {
         }
     }
 
-    llama_free_model(model);
+    llama_model_free(model);
     llama_free(ctx);
 
     llama_backend_free();
diff --git a/tests/test-tokenizer-random.py b/tests/test-tokenizer-random.py
index 9ebe6c891..c6cdcb554 100644
--- a/tests/test-tokenizer-random.py
+++ b/tests/test-tokenizer-random.py
@@ -76,7 +76,7 @@ class LibLlamaModel:
         self.ffi = libllama.ffi
         if isinstance(mparams, dict):
             mparams = libllama.model_default_params(**mparams)
-        self.model = self.lib.llama_load_model_from_file(path_model.encode(), mparams)
+        self.model = self.lib.llama_model_load_from_file(path_model.encode(), mparams)
         if not self.model:
             raise RuntimeError("error: failed to load model '%s'" % path_model)
         if isinstance(cparams, dict):
@@ -92,7 +92,7 @@ class LibLlamaModel:
         if self.ctx:
             self.lib.llama_free(self.ctx)
         if self.model:
-            self.lib.llama_free_model(self.model)
+            self.lib.llama_model_free(self.model)
         self.ctx = None
         self.model = None
         self.lib = None